[GIT PULL de-alpha] Make core code not need to know about Alpha for v4.15

[GIT PULL de-alpha] Make core code not need to know about Alpha for v4.15

[Paul E. McKenney]

Hello, Ingo,

This series is a first step towards making the core kernel no longer
need to consider DEC Alpha as a special case.  This is accomplished
by two sets of patches, followed by a Coccinelle script:

1.	Patches 1/19 through 15/19 in the following patches, which
	change non-Coccinelle-susceptible instances of ACCESS_ONCE()
	to either READ_ONCE() or WRITE_ONCE(), as appropriate.	Most of
	these patches are courtesy of Mark Rutland.

2.	Patches 16/19 through 18/19 in the following patches, which
	add smp_read_barrier_depends() to READ_ONCE() and to Alpha's
	value-returning _release and _relaxed atomic read-modify-write
	operations, while also removing lockless_dereference() in favor
	of the modified READ_ONCE().

3.	A patch generated by Coccinelle (script in commit log) that
	converts ACCESS_ONCE() to either READ_ONCE() or WRITE_ONCE(),
	as appropriate.  Of course, this patch should not be applied
	as-is to mainline, instead, the Coccinelle script should be
	re-run in order to account for any changes between now and
	the time of merging into mainline.

A later series (most likely targeted to v4.16) will remove instances of
smp_read_barrier_depends() that are made redundant by this series, that
is to say, almost all of them.

These changes are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git rcu/alpha-cocci

for you to fetch changes up to d19fa5187a69bd7685e9c7452ce9e339b1b079fc:

  COCCINELLE: treewide: kill off ACCESS_ONCE() (2017-10-21 12:52:09 -0700)

----------------------------------------------------------------
Mark Rutland (14):
      dm integrity: Kill off ACCESS_ONCE()
      EDAC, altera: Kill off ACCESS_ONCE()
      firmware/ivc: Kill off ACCESS_ONCE()
      fs: dcache: Kill off ACCESS_ONCE()
      fs: ncpfs: Kill off ACCESS_ONCE()
      media: dvb_ringbuffer: Kill off ACCESS_ONCE()
      net: netlink/netfilter: Kill off ACCESS_ONCE()
      net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()
      net: average: Kill off ACCESS_ONCE()
      samples: mic/mpssd/mpssd.c: Kill off ACCESS_ONCE()
      selftests/powerpc: Kill off ACCESS_ONCE()
      workqueue: Kill off ACCESS_ONCE()
      rcutorture: formal: Prepare for ACCESS_ONCE() removal
      COCCINELLE: treewide: kill off ACCESS_ONCE()

Paul E. McKenney (3):
      doc: filesystems: Kill off ACCESS_ONCE()
      mm: Kill off ACCESS_ONCE()
      Merge commit 'c78bcafd0d3b3835f23369aea7ed3686f25750c2'; commit '88f35a0d24b00b80538ae7083cfcab8eb00db3cb' into HEAD

Will Deacon (3):
      linux/compiler.h: Split into compiler.h and compiler-types.h
      locking/barriers: Kill lockless_dereference
      alpha: atomics: Add smp_read_barrier_depends() to release/relaxed atomics

 Documentation/filesystems/path-lookup.md           |   6 +-
 Documentation/memory-barriers.txt                  |  12 -
 .../translations/ko_KR/memory-barriers.txt         |  12 -
 arch/alpha/include/asm/atomic.h                    |  13 +
 arch/arc/kernel/smp.c                              |   2 +-
 arch/arm/include/asm/ptrace.h                      |   3 +-
 arch/arm/include/asm/spinlock.h                    |   6 +-
 arch/arm/mach-tegra/cpuidle-tegra20.c              |   2 +-
 arch/arm/vdso/vgettimeofday.c                      |   2 +-
 arch/ia64/include/asm/spinlock.h                   |   8 +-
 arch/mips/include/asm/vdso.h                       |   2 +-
 arch/mips/kernel/pm-cps.c                          |   2 +-
 arch/mn10300/kernel/mn10300-serial.c               |   4 +-
 arch/parisc/include/asm/atomic.h                   |   2 +-
 arch/powerpc/platforms/powernv/opal-msglog.c       |   2 +-
 arch/s390/include/asm/spinlock.h                   |   6 +-
 arch/s390/lib/spinlock.c                           |  16 +-
 arch/sparc/include/asm/atomic_32.h                 |   2 +-
 arch/sparc/include/asm/ptrace.h                    |   1 +
 arch/tile/gxio/dma_queue.c                         |   4 +-
 arch/tile/include/gxio/dma_queue.h                 |   2 +-
 arch/tile/kernel/ptrace.c                          |   2 +-
 arch/um/include/shared/init.h                      |   2 +-
 arch/x86/entry/common.c                            |   2 +-
 arch/x86/entry/vdso/vclock_gettime.c               |   2 +-
 arch/x86/events/core.c                             |   4 +-
 arch/x86/include/asm/mmu_context.h                 |   4 +-
 arch/x86/include/asm/vgtod.h                       |   2 +-
 arch/x86/kernel/espfix_64.c                        |   6 +-
 arch/x86/kernel/ldt.c                              |   2 +-
 arch/x86/kernel/nmi.c                              |   2 +-
 arch/x86/kvm/mmu.c                                 |   4 +-
 arch/x86/kvm/page_track.c                          |   2 +-
 arch/x86/xen/p2m.c                                 |   2 +-
 arch/xtensa/platforms/xtfpga/lcd.c                 |  14 +-
 block/blk-wbt.c                                    |   2 +-
 drivers/base/core.c                                |   2 +-
 drivers/base/power/runtime.c                       |   4 +-
 drivers/char/random.c                              |   4 +-
 drivers/clocksource/bcm2835_timer.c                |   2 +-
 drivers/crypto/caam/jr.c                           |   4 +-
 drivers/crypto/nx/nx-842-powernv.c                 |   2 +-
 drivers/edac/altera_edac.c                         |  10 +-
 drivers/firewire/ohci.c                            |  10 +-
 drivers/firmware/tegra/ivc.c                       |  24 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c          |   4 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c            |   4 +-
 drivers/gpu/drm/amd/scheduler/gpu_scheduler.c      |   2 +-
 drivers/gpu/drm/radeon/radeon_gem.c                |   4 +-
 drivers/gpu/drm/vmwgfx/vmwgfx_surface.c            |   2 +-
 drivers/infiniband/hw/hfi1/file_ops.c              |   2 +-
 drivers/infiniband/hw/hfi1/pio.c                   |   6 +-
 drivers/infiniband/hw/hfi1/ruc.c                   |   2 +-
 drivers/infiniband/hw/hfi1/sdma.c                  |   8 +-
 drivers/infiniband/hw/hfi1/sdma.h                  |   2 +-
 drivers/infiniband/hw/hfi1/uc.c                    |   4 +-
 drivers/infiniband/hw/hfi1/ud.c                    |   4 +-
 drivers/infiniband/hw/hfi1/user_sdma.c             |   8 +-
 drivers/infiniband/hw/qib/qib_ruc.c                |   2 +-
 drivers/infiniband/hw/qib/qib_uc.c                 |   4 +-
 drivers/infiniband/hw/qib/qib_ud.c                 |   4 +-
 drivers/infiniband/sw/rdmavt/qp.c                  |   6 +-
 drivers/input/misc/regulator-haptic.c              |   2 +-
 drivers/md/dm-bufio.c                              |  10 +-
 drivers/md/dm-integrity.c                          |  15 +-
 drivers/md/dm-kcopyd.c                             |   4 +-
 drivers/md/dm-mpath.c                              |  20 +-
 drivers/md/dm-stats.c                              |  36 +--
 drivers/md/dm-switch.c                             |   2 +-
 drivers/md/dm-thin.c                               |   2 +-
 drivers/md/dm-verity-target.c                      |   2 +-
 drivers/md/dm.c                                    |   4 +-
 drivers/md/md.c                                    |   2 +-
 drivers/md/raid5.c                                 |   2 +-
 drivers/media/dvb-core/dvb_ringbuffer.c            |   8 +-
 drivers/misc/mic/scif/scif_rb.c                    |   8 +-
 drivers/misc/mic/scif/scif_rma_list.c              |   2 +-
 drivers/net/bonding/bond_alb.c                     |   2 +-
 drivers/net/bonding/bond_main.c                    |   6 +-
 drivers/net/ethernet/chelsio/cxgb4/sge.c           |   4 +-
 drivers/net/ethernet/emulex/benet/be_main.c        |   2 +-
 drivers/net/ethernet/hisilicon/hip04_eth.c         |   4 +-
 drivers/net/ethernet/intel/i40e/i40e_debugfs.c     |   4 +-
 drivers/net/ethernet/intel/i40e/i40e_ethtool.c     |   2 +-
 drivers/net/ethernet/intel/i40e/i40e_main.c        |   4 +-
 drivers/net/ethernet/intel/i40e/i40e_ptp.c         |   4 +-
 drivers/net/ethernet/intel/igb/e1000_regs.h        |   2 +-
 drivers/net/ethernet/intel/igb/igb_main.c          |   2 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_common.h    |   4 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c      |   8 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c       |   4 +-
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c  |   2 +-
 drivers/net/ethernet/intel/ixgbevf/vf.h            |   2 +-
 drivers/net/ethernet/mellanox/mlx4/en_tx.c         |  12 +-
 drivers/net/ethernet/neterion/vxge/vxge-main.c     |   2 +-
 drivers/net/ethernet/sfc/ef10.c                    |  10 +-
 drivers/net/ethernet/sfc/efx.c                     |   4 +-
 drivers/net/ethernet/sfc/falcon/efx.c              |   4 +-
 drivers/net/ethernet/sfc/falcon/falcon.c           |   4 +-
 drivers/net/ethernet/sfc/falcon/farch.c            |   8 +-
 drivers/net/ethernet/sfc/falcon/nic.h              |   6 +-
 drivers/net/ethernet/sfc/falcon/tx.c               |   6 +-
 drivers/net/ethernet/sfc/farch.c                   |   8 +-
 drivers/net/ethernet/sfc/nic.h                     |   6 +-
 drivers/net/ethernet/sfc/ptp.c                     |  10 +-
 drivers/net/ethernet/sfc/tx.c                      |   6 +-
 drivers/net/ethernet/sun/niu.c                     |   4 +-
 drivers/net/tap.c                                  |   2 +-
 drivers/net/tun.c                                  |   4 +-
 drivers/net/wireless/ath/ath5k/desc.c              |   8 +-
 .../wireless/broadcom/brcm80211/brcmfmac/sdio.c    |   2 +-
 drivers/net/wireless/intel/iwlwifi/mvm/ops.c       |   2 +-
 drivers/net/wireless/intel/iwlwifi/mvm/tx.c        |   4 +-
 drivers/net/wireless/intel/iwlwifi/pcie/rx.c       |   2 +-
 drivers/net/wireless/intel/iwlwifi/pcie/trans.c    |  10 +-
 drivers/net/wireless/mac80211_hwsim.c              |   4 +-
 drivers/scsi/qla2xxx/qla_target.c                  |   2 +-
 drivers/target/target_core_user.c                  |   2 +-
 drivers/usb/class/cdc-wdm.c                        |   2 +-
 drivers/usb/core/devio.c                           |   2 +-
 drivers/usb/core/sysfs.c                           |   4 +-
 drivers/usb/gadget/udc/gr_udc.c                    |   4 +-
 drivers/usb/host/ohci-hcd.c                        |   2 +-
 drivers/usb/host/uhci-hcd.h                        |   4 +-
 drivers/vfio/vfio.c                                |   2 +-
 drivers/vhost/scsi.c                               |   2 +-
 fs/aio.c                                           |   2 +-
 fs/buffer.c                                        |   3 +-
 fs/crypto/keyinfo.c                                |   2 +-
 fs/dcache.c                                        |  22 +-
 fs/direct-io.c                                     |   2 +-
 fs/exec.c                                          |   2 +-
 fs/fcntl.c                                         |   2 +-
 fs/fs_pin.c                                        |   4 +-
 fs/fuse/dev.c                                      |   2 +-
 fs/inode.c                                         |   2 +-
 fs/namei.c                                         |   4 +-
 fs/namespace.c                                     |   2 +-
 fs/ncpfs/dir.c                                     |   9 -
 fs/nfs/dir.c                                       |   8 +-
 fs/overlayfs/ovl_entry.h                           |   2 +-
 fs/overlayfs/readdir.c                             |   2 +-
 fs/proc/array.c                                    |   2 +-
 fs/proc_namespace.c                                |   2 +-
 fs/splice.c                                        |   2 +-
 fs/userfaultfd.c                                   |   8 +-
 fs/xfs/xfs_log_priv.h                              |   4 +-
 include/linux/average.h                            |  10 +-
 include/linux/bitops.h                             |   4 +-
 include/linux/compiler-clang.h                     |   2 +-
 include/linux/compiler-gcc.h                       |   2 +-
 include/linux/compiler-intel.h                     |   2 +-
 include/linux/compiler-types.h                     | 274 ++++++++++++++++++++
 include/linux/compiler.h                           | 286 +--------------------
 include/linux/dcache.h                             |   4 +-
 include/linux/dynamic_queue_limits.h               |   2 +-
 include/linux/genetlink.h                          |   2 +-
 include/linux/huge_mm.h                            |   2 +-
 include/linux/if_team.h                            |   2 +-
 include/linux/linkage.h                            |   2 +-
 include/linux/llist.h                              |   2 +-
 include/linux/netfilter/nfnetlink.h                |   2 +-
 include/linux/pm_runtime.h                         |   2 +-
 include/linux/rculist.h                            |   4 +-
 include/linux/rcupdate.h                           |   4 +-
 include/linux/rtnetlink.h                          |   2 +-
 include/net/ip_vs.h                                |   6 +-
 include/net/netfilter/nf_tables.h                  |   4 +-
 include/uapi/linux/stddef.h                        |   2 +-
 kernel/acct.c                                      |   4 +-
 kernel/events/core.c                               |  10 +-
 kernel/events/ring_buffer.c                        |   2 +-
 kernel/exit.c                                      |   2 +-
 kernel/seccomp.c                                   |   2 +-
 kernel/task_work.c                                 |   2 +-
 kernel/trace/ring_buffer.c                         |   2 +-
 kernel/trace/trace.h                               |   2 +-
 kernel/trace/trace_stack.c                         |   2 +-
 kernel/user_namespace.c                            |   2 +-
 kernel/workqueue.c                                 |   4 +-
 lib/assoc_array.c                                  |  20 +-
 lib/dynamic_queue_limits.c                         |   2 +-
 lib/llist.c                                        |   2 +-
 lib/vsprintf.c                                     |   4 +-
 mm/huge_memory.c                                   |   2 +-
 mm/memory.c                                        |   6 +-
 mm/slab.h                                          |   2 +-
 net/core/dev.c                                     |   2 +-
 net/core/pktgen.c                                  |   2 +-
 net/ipv4/inet_fragment.c                           |   2 +-
 net/ipv4/route.c                                   |   2 +-
 net/ipv4/tcp_input.c                               |   6 +-
 net/ipv4/tcp_output.c                              |   2 +-
 net/ipv4/udp.c                                     |   4 +-
 net/ipv6/ip6_tunnel.c                              |   8 +-
 net/ipv6/udp.c                                     |   4 +-
 net/llc/llc_input.c                                |   4 +-
 net/mac80211/sta_info.c                            |   2 +-
 net/netfilter/ipvs/ip_vs_sync.c                    |   2 +-
 net/netfilter/nfnetlink_queue.c                    |   4 +-
 net/netlabel/netlabel_calipso.c                    |   2 +-
 net/wireless/nl80211.c                             |   2 +-
 samples/mic/mpssd/mpssd.c                          |   6 +-
 scripts/headers_install.sh                         |   2 +-
 sound/firewire/amdtp-am824.c                       |   6 +-
 sound/firewire/amdtp-stream.c                      |  23 +-
 sound/firewire/amdtp-stream.h                      |   2 +-
 sound/firewire/digi00x/amdtp-dot.c                 |   6 +-
 sound/firewire/fireface/amdtp-ff.c                 |   4 +-
 sound/firewire/fireface/ff-midi.c                  |  10 +-
 sound/firewire/fireface/ff-transaction.c           |   8 +-
 sound/firewire/isight.c                            |  18 +-
 sound/firewire/motu/amdtp-motu.c                   |   4 +-
 sound/firewire/oxfw/oxfw-scs1x.c                   |  12 +-
 sound/firewire/tascam/amdtp-tascam.c               |   4 +-
 sound/firewire/tascam/tascam-transaction.c         |   6 +-
 sound/soc/xtensa/xtfpga-i2s.c                      |   6 +-
 sound/usb/bcd2000/bcd2000.c                        |   4 +-
 tools/arch/x86/include/asm/atomic.h                |   2 +-
 tools/include/asm-generic/atomic-gcc.h             |   2 +-
 tools/perf/util/auxtrace.h                         |   4 +-
 tools/perf/util/evlist.h                           |   2 +-
 tools/perf/util/session.h                          |   2 +-
 tools/testing/selftests/powerpc/dscr/dscr.h        |   2 +-
 .../selftests/powerpc/dscr/dscr_default_test.c     |   2 +-
 .../rcutorture/formal/srcu-cbmc/src/barriers.h     |   5 +-
 virt/kvm/kvm_main.c                                |   2 +-
 227 files changed, 786 insertions(+), 808 deletions(-)
 create mode 100644 include/linux/compiler-types.h

[PATCH 01/19] dm integrity: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't pick up some uses, including those
in dm-integrity.c. As a preparatory step, this patch converts the driver
to use {READ,WRITE}_ONCE() consistently.

At the same time, this patch adds the missing include of
<linux/compiler.h> necessary for the {READ,WRITE}_ONCE() definitions.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Snitzer <snitzer@redhat.com>
---
 drivers/md/dm-integrity.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/drivers/md/dm-integrity.c b/drivers/md/dm-integrity.c
index 096fe9b66c50..8c5756e1df94 100644
--- a/drivers/md/dm-integrity.c
+++ b/drivers/md/dm-integrity.c
@@ -6,6 +6,7 @@
  * This file is released under the GPL.
  */
 
+#include <linux/compiler.h>
 #include <linux/module.h>
 #include <linux/device-mapper.h>
 #include <linux/dm-io.h>
@@ -80,13 +81,13 @@ struct journal_entry {
 #define journal_entry_tag(ic, je)		((__u8 *)&(je)->last_bytes[(ic)->sectors_per_block])
 
 #if BITS_PER_LONG == 64
-#define journal_entry_set_sector(je, x)		do { smp_wmb(); ACCESS_ONCE((je)->u.sector) = cpu_to_le64(x); } while (0)
+#define journal_entry_set_sector(je, x)		do { smp_wmb(); WRITE_ONCE((je)->u.sector, cpu_to_le64(x)); } while (0)
 #define journal_entry_get_sector(je)		le64_to_cpu((je)->u.sector)
 #elif defined(CONFIG_LBDAF)
-#define journal_entry_set_sector(je, x)		do { (je)->u.s.sector_lo = cpu_to_le32(x); smp_wmb(); ACCESS_ONCE((je)->u.s.sector_hi) = cpu_to_le32((x) >> 32); } while (0)
+#define journal_entry_set_sector(je, x)		do { (je)->u.s.sector_lo = cpu_to_le32(x); smp_wmb(); WRITE_ONCE((je)->u.s.sector_hi, cpu_to_le32((x) >> 32)); } while (0)
 #define journal_entry_get_sector(je)		le64_to_cpu((je)->u.sector)
 #else
-#define journal_entry_set_sector(je, x)		do { (je)->u.s.sector_lo = cpu_to_le32(x); smp_wmb(); ACCESS_ONCE((je)->u.s.sector_hi) = cpu_to_le32(0); } while (0)
+#define journal_entry_set_sector(je, x)		do { (je)->u.s.sector_lo = cpu_to_le32(x); smp_wmb(); WRITE_ONCE((je)->u.s.sector_hi, cpu_to_le32(0)); } while (0)
 #define journal_entry_get_sector(je)		le32_to_cpu((je)->u.s.sector_lo)
 #endif
 #define journal_entry_is_unused(je)		((je)->u.s.sector_hi == cpu_to_le32(-1))
@@ -320,7 +321,7 @@ static void dm_integrity_io_error(struct dm_integrity_c *ic, const char *msg, in
 
 static int dm_integrity_failed(struct dm_integrity_c *ic)
 {
-	return ACCESS_ONCE(ic->failed);
+	return READ_ONCE(ic->failed);
 }
 
 static commit_id_t dm_integrity_commit_id(struct dm_integrity_c *ic, unsigned i,
@@ -1545,7 +1546,7 @@ static bool __journal_read_write(struct dm_integrity_io *dio, struct bio *bio,
 		smp_mb();
 		if (unlikely(waitqueue_active(&ic->copy_to_journal_wait)))
 			wake_up(&ic->copy_to_journal_wait);
-		if (ACCESS_ONCE(ic->free_sectors) <= ic->free_sectors_threshold) {
+		if (READ_ONCE(ic->free_sectors) <= ic->free_sectors_threshold) {
 			queue_work(ic->commit_wq, &ic->commit_work);
 		} else {
 			schedule_autocommit(ic);
@@ -1798,7 +1799,7 @@ static void integrity_commit(struct work_struct *w)
 	ic->n_committed_sections += commit_sections;
 	spin_unlock_irq(&ic->endio_wait.lock);
 
-	if (ACCESS_ONCE(ic->free_sectors) <= ic->free_sectors_threshold)
+	if (READ_ONCE(ic->free_sectors) <= ic->free_sectors_threshold)
 		queue_work(ic->writer_wq, &ic->writer_work);
 
 release_flush_bios:
@@ -1980,7 +1981,7 @@ static void integrity_writer(struct work_struct *w)
 	unsigned prev_free_sectors;
 
 	/* the following test is not needed, but it tests the replay code */
-	if (ACCESS_ONCE(ic->suspending))
+	if (READ_ONCE(ic->suspending))
 		return;
 
 	spin_lock_irq(&ic->endio_wait.lock);
-- 
2.5.2

[PATCH 01/19] dm integrity: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't pick up some uses, including those
in dm-integrity.c. As a preparatory step, this patch converts the driver
to use {READ,WRITE}_ONCE() consistently.

At the same time, this patch adds the missing include of
<linux/compiler.h> necessary for the {READ,WRITE}_ONCE() definitions.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Snitzer <snitzer@redhat.com>
---
 drivers/md/dm-integrity.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/drivers/md/dm-integrity.c b/drivers/md/dm-integrity.c
index 096fe9b66c50..8c5756e1df94 100644
--- a/drivers/md/dm-integrity.c
+++ b/drivers/md/dm-integrity.c
@@ -6,6 +6,7 @@
  * This file is released under the GPL.
  */
 
+#include <linux/compiler.h>
 #include <linux/module.h>
 #include <linux/device-mapper.h>
 #include <linux/dm-io.h>
@@ -80,13 +81,13 @@ struct journal_entry {
 #define journal_entry_tag(ic, je)		((__u8 *)&(je)->last_bytes[(ic)->sectors_per_block])
 
 #if BITS_PER_LONG == 64
-#define journal_entry_set_sector(je, x)		do { smp_wmb(); ACCESS_ONCE((je)->u.sector) = cpu_to_le64(x); } while (0)
+#define journal_entry_set_sector(je, x)		do { smp_wmb(); WRITE_ONCE((je)->u.sector, cpu_to_le64(x)); } while (0)
 #define journal_entry_get_sector(je)		le64_to_cpu((je)->u.sector)
 #elif defined(CONFIG_LBDAF)
-#define journal_entry_set_sector(je, x)		do { (je)->u.s.sector_lo = cpu_to_le32(x); smp_wmb(); ACCESS_ONCE((je)->u.s.sector_hi) = cpu_to_le32((x) >> 32); } while (0)
+#define journal_entry_set_sector(je, x)		do { (je)->u.s.sector_lo = cpu_to_le32(x); smp_wmb(); WRITE_ONCE((je)->u.s.sector_hi, cpu_to_le32((x) >> 32)); } while (0)
 #define journal_entry_get_sector(je)		le64_to_cpu((je)->u.sector)
 #else
-#define journal_entry_set_sector(je, x)		do { (je)->u.s.sector_lo = cpu_to_le32(x); smp_wmb(); ACCESS_ONCE((je)->u.s.sector_hi) = cpu_to_le32(0); } while (0)
+#define journal_entry_set_sector(je, x)		do { (je)->u.s.sector_lo = cpu_to_le32(x); smp_wmb(); WRITE_ONCE((je)->u.s.sector_hi, cpu_to_le32(0)); } while (0)
 #define journal_entry_get_sector(je)		le32_to_cpu((je)->u.s.sector_lo)
 #endif
 #define journal_entry_is_unused(je)		((je)->u.s.sector_hi == cpu_to_le32(-1))
@@ -320,7 +321,7 @@ static void dm_integrity_io_error(struct dm_integrity_c *ic, const char *msg, in
 
 static int dm_integrity_failed(struct dm_integrity_c *ic)
 {
-	return ACCESS_ONCE(ic->failed);
+	return READ_ONCE(ic->failed);
 }
 
 static commit_id_t dm_integrity_commit_id(struct dm_integrity_c *ic, unsigned i,
@@ -1545,7 +1546,7 @@ static bool __journal_read_write(struct dm_integrity_io *dio, struct bio *bio,
 		smp_mb();
 		if (unlikely(waitqueue_active(&ic->copy_to_journal_wait)))
 			wake_up(&ic->copy_to_journal_wait);
-		if (ACCESS_ONCE(ic->free_sectors) <= ic->free_sectors_threshold) {
+		if (READ_ONCE(ic->free_sectors) <= ic->free_sectors_threshold) {
 			queue_work(ic->commit_wq, &ic->commit_work);
 		} else {
 			schedule_autocommit(ic);
@@ -1798,7 +1799,7 @@ static void integrity_commit(struct work_struct *w)
 	ic->n_committed_sections += commit_sections;
 	spin_unlock_irq(&ic->endio_wait.lock);
 
-	if (ACCESS_ONCE(ic->free_sectors) <= ic->free_sectors_threshold)
+	if (READ_ONCE(ic->free_sectors) <= ic->free_sectors_threshold)
 		queue_work(ic->writer_wq, &ic->writer_work);
 
 release_flush_bios:
@@ -1980,7 +1981,7 @@ static void integrity_writer(struct work_struct *w)
 	unsigned prev_free_sectors;
 
 	/* the following test is not needed, but it tests the replay code */
-	if (ACCESS_ONCE(ic->suspending))
+	if (READ_ONCE(ic->suspending))
 		return;
 
 	spin_lock_irq(&ic->endio_wait.lock);
-- 
2.5.2

[PATCH 02/19] EDAC, altera: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the Altera EDAC code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Thor Thayer <thor.thayer@linux.intel.com>
Cc: Borislav Petkov <bp@alien8.de>
---
 drivers/edac/altera_edac.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c
index 346c4987b284..11d6419788c2 100644
--- a/drivers/edac/altera_edac.c
+++ b/drivers/edac/altera_edac.c
@@ -175,11 +175,11 @@ static ssize_t altr_sdr_mc_err_inject_write(struct file *file,
 	/*
 	 * To trigger the error, we need to read the data back
 	 * (the data was written with errors above).
-	 * The ACCESS_ONCE macros and printk are used to prevent the
+	 * The READ_ONCE macros and printk are used to prevent the
 	 * the compiler optimizing these reads out.
 	 */
-	reg = ACCESS_ONCE(ptemp[0]);
-	read_reg = ACCESS_ONCE(ptemp[1]);
+	reg = READ_ONCE(ptemp[0]);
+	read_reg = READ_ONCE(ptemp[1]);
 	/* Force Read */
 	rmb();
 
@@ -618,7 +618,7 @@ static ssize_t altr_edac_device_trig(struct file *file,
 	for (i = 0; i < (priv->trig_alloc_sz / sizeof(*ptemp)); i++) {
 		/* Read data so we're in the correct state */
 		rmb();
-		if (ACCESS_ONCE(ptemp[i]))
+		if (READ_ONCE(ptemp[i]))
 			result = -1;
 		/* Toggle Error bit (it is latched), leave ECC enabled */
 		writel(error_mask, (drvdata->base + priv->set_err_ofst));
@@ -635,7 +635,7 @@ static ssize_t altr_edac_device_trig(struct file *file,
 
 	/* Read out written data. ECC error caused here */
 	for (i = 0; i < ALTR_TRIGGER_READ_WRD_CNT; i++)
-		if (ACCESS_ONCE(ptemp[i]) != i)
+		if (READ_ONCE(ptemp[i]) != i)
 			edac_printk(KERN_ERR, EDAC_DEVICE,
 				    "Read doesn't match written data\n");
 
-- 
2.5.2

[PATCH 02/19] EDAC, altera: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the Altera EDAC code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Thor Thayer <thor.thayer@linux.intel.com>
Cc: Borislav Petkov <bp@alien8.de>
---
 drivers/edac/altera_edac.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c
index 346c4987b284..11d6419788c2 100644
--- a/drivers/edac/altera_edac.c
+++ b/drivers/edac/altera_edac.c
@@ -175,11 +175,11 @@ static ssize_t altr_sdr_mc_err_inject_write(struct file *file,
 	/*
 	 * To trigger the error, we need to read the data back
 	 * (the data was written with errors above).
-	 * The ACCESS_ONCE macros and printk are used to prevent the
+	 * The READ_ONCE macros and printk are used to prevent the
 	 * the compiler optimizing these reads out.
 	 */
-	reg = ACCESS_ONCE(ptemp[0]);
-	read_reg = ACCESS_ONCE(ptemp[1]);
+	reg = READ_ONCE(ptemp[0]);
+	read_reg = READ_ONCE(ptemp[1]);
 	/* Force Read */
 	rmb();
 
@@ -618,7 +618,7 @@ static ssize_t altr_edac_device_trig(struct file *file,
 	for (i = 0; i < (priv->trig_alloc_sz / sizeof(*ptemp)); i++) {
 		/* Read data so we're in the correct state */
 		rmb();
-		if (ACCESS_ONCE(ptemp[i]))
+		if (READ_ONCE(ptemp[i]))
 			result = -1;
 		/* Toggle Error bit (it is latched), leave ECC enabled */
 		writel(error_mask, (drvdata->base + priv->set_err_ofst));
@@ -635,7 +635,7 @@ static ssize_t altr_edac_device_trig(struct file *file,
 
 	/* Read out written data. ECC error caused here */
 	for (i = 0; i < ALTR_TRIGGER_READ_WRD_CNT; i++)
-		if (ACCESS_ONCE(ptemp[i]) != i)
+		if (READ_ONCE(ptemp[i]) != i)
 			edac_printk(KERN_ERR, EDAC_DEVICE,
 				    "Read doesn't match written data\n");
 
-- 
2.5.2

[PATCH 03/19] firmware/ivc: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

workqueue: kill off ACCESS_ONCE()

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the Tegra IVC code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Jonathan Hunter <jonathanh@nvidia.com>
Cc: Thierry Reding <thierry.reding@gmail.com>
---
 drivers/firmware/tegra/ivc.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/drivers/firmware/tegra/ivc.c b/drivers/firmware/tegra/ivc.c
index a01461d63f68..00de793e6423 100644
--- a/drivers/firmware/tegra/ivc.c
+++ b/drivers/firmware/tegra/ivc.c
@@ -99,11 +99,11 @@ static inline bool tegra_ivc_empty(struct tegra_ivc *ivc,
 {
 	/*
 	 * This function performs multiple checks on the same values with
-	 * security implications, so create snapshots with ACCESS_ONCE() to
+	 * security implications, so create snapshots with READ_ONCE() to
 	 * ensure that these checks use the same values.
 	 */
-	u32 tx = ACCESS_ONCE(header->tx.count);
-	u32 rx = ACCESS_ONCE(header->rx.count);
+	u32 tx = READ_ONCE(header->tx.count);
+	u32 rx = READ_ONCE(header->rx.count);
 
 	/*
 	 * Perform an over-full check to prevent denial of service attacks
@@ -124,8 +124,8 @@ static inline bool tegra_ivc_empty(struct tegra_ivc *ivc,
 static inline bool tegra_ivc_full(struct tegra_ivc *ivc,
 				  struct tegra_ivc_header *header)
 {
-	u32 tx = ACCESS_ONCE(header->tx.count);
-	u32 rx = ACCESS_ONCE(header->rx.count);
+	u32 tx = READ_ONCE(header->tx.count);
+	u32 rx = READ_ONCE(header->rx.count);
 
 	/*
 	 * Invalid cases where the counters indicate that the queue is over
@@ -137,8 +137,8 @@ static inline bool tegra_ivc_full(struct tegra_ivc *ivc,
 static inline u32 tegra_ivc_available(struct tegra_ivc *ivc,
 				      struct tegra_ivc_header *header)
 {
-	u32 tx = ACCESS_ONCE(header->tx.count);
-	u32 rx = ACCESS_ONCE(header->rx.count);
+	u32 tx = READ_ONCE(header->tx.count);
+	u32 rx = READ_ONCE(header->rx.count);
 
 	/*
 	 * This function isn't expected to be used in scenarios where an
@@ -151,8 +151,8 @@ static inline u32 tegra_ivc_available(struct tegra_ivc *ivc,
 
 static inline void tegra_ivc_advance_tx(struct tegra_ivc *ivc)
 {
-	ACCESS_ONCE(ivc->tx.channel->tx.count) =
-		ACCESS_ONCE(ivc->tx.channel->tx.count) + 1;
+	WRITE_ONCE(ivc->tx.channel->tx.count,
+		   READ_ONCE(ivc->tx.channel->tx.count) + 1);
 
 	if (ivc->tx.position == ivc->num_frames - 1)
 		ivc->tx.position = 0;
@@ -162,8 +162,8 @@ static inline void tegra_ivc_advance_tx(struct tegra_ivc *ivc)
 
 static inline void tegra_ivc_advance_rx(struct tegra_ivc *ivc)
 {
-	ACCESS_ONCE(ivc->rx.channel->rx.count) =
-		ACCESS_ONCE(ivc->rx.channel->rx.count) + 1;
+	WRITE_ONCE(ivc->rx.channel->rx.count,
+		   READ_ONCE(ivc->rx.channel->rx.count) + 1);
 
 	if (ivc->rx.position == ivc->num_frames - 1)
 		ivc->rx.position = 0;
@@ -428,7 +428,7 @@ int tegra_ivc_notified(struct tegra_ivc *ivc)
 
 	/* Copy the receiver's state out of shared memory. */
 	tegra_ivc_invalidate(ivc, ivc->rx.phys + offset);
-	state = ACCESS_ONCE(ivc->rx.channel->tx.state);
+	state = READ_ONCE(ivc->rx.channel->tx.state);
 
 	if (state == TEGRA_IVC_STATE_SYNC) {
 		offset = offsetof(struct tegra_ivc_header, tx.count);
-- 
2.5.2

[PATCH 03/19] firmware/ivc: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

workqueue: kill off ACCESS_ONCE()

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the Tegra IVC code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Jonathan Hunter <jonathanh@nvidia.com>
Cc: Thierry Reding <thierry.reding@gmail.com>
---
 drivers/firmware/tegra/ivc.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/drivers/firmware/tegra/ivc.c b/drivers/firmware/tegra/ivc.c
index a01461d63f68..00de793e6423 100644
--- a/drivers/firmware/tegra/ivc.c
+++ b/drivers/firmware/tegra/ivc.c
@@ -99,11 +99,11 @@ static inline bool tegra_ivc_empty(struct tegra_ivc *ivc,
 {
 	/*
 	 * This function performs multiple checks on the same values with
-	 * security implications, so create snapshots with ACCESS_ONCE() to
+	 * security implications, so create snapshots with READ_ONCE() to
 	 * ensure that these checks use the same values.
 	 */
-	u32 tx = ACCESS_ONCE(header->tx.count);
-	u32 rx = ACCESS_ONCE(header->rx.count);
+	u32 tx = READ_ONCE(header->tx.count);
+	u32 rx = READ_ONCE(header->rx.count);
 
 	/*
 	 * Perform an over-full check to prevent denial of service attacks
@@ -124,8 +124,8 @@ static inline bool tegra_ivc_empty(struct tegra_ivc *ivc,
 static inline bool tegra_ivc_full(struct tegra_ivc *ivc,
 				  struct tegra_ivc_header *header)
 {
-	u32 tx = ACCESS_ONCE(header->tx.count);
-	u32 rx = ACCESS_ONCE(header->rx.count);
+	u32 tx = READ_ONCE(header->tx.count);
+	u32 rx = READ_ONCE(header->rx.count);
 
 	/*
 	 * Invalid cases where the counters indicate that the queue is over
@@ -137,8 +137,8 @@ static inline bool tegra_ivc_full(struct tegra_ivc *ivc,
 static inline u32 tegra_ivc_available(struct tegra_ivc *ivc,
 				      struct tegra_ivc_header *header)
 {
-	u32 tx = ACCESS_ONCE(header->tx.count);
-	u32 rx = ACCESS_ONCE(header->rx.count);
+	u32 tx = READ_ONCE(header->tx.count);
+	u32 rx = READ_ONCE(header->rx.count);
 
 	/*
 	 * This function isn't expected to be used in scenarios where an
@@ -151,8 +151,8 @@ static inline u32 tegra_ivc_available(struct tegra_ivc *ivc,
 
 static inline void tegra_ivc_advance_tx(struct tegra_ivc *ivc)
 {
-	ACCESS_ONCE(ivc->tx.channel->tx.count) =
-		ACCESS_ONCE(ivc->tx.channel->tx.count) + 1;
+	WRITE_ONCE(ivc->tx.channel->tx.count,
+		   READ_ONCE(ivc->tx.channel->tx.count) + 1);
 
 	if (ivc->tx.position == ivc->num_frames - 1)
 		ivc->tx.position = 0;
@@ -162,8 +162,8 @@ static inline void tegra_ivc_advance_tx(struct tegra_ivc *ivc)
 
 static inline void tegra_ivc_advance_rx(struct tegra_ivc *ivc)
 {
-	ACCESS_ONCE(ivc->rx.channel->rx.count) =
-		ACCESS_ONCE(ivc->rx.channel->rx.count) + 1;
+	WRITE_ONCE(ivc->rx.channel->rx.count,
+		   READ_ONCE(ivc->rx.channel->rx.count) + 1);
 
 	if (ivc->rx.position == ivc->num_frames - 1)
 		ivc->rx.position = 0;
@@ -428,7 +428,7 @@ int tegra_ivc_notified(struct tegra_ivc *ivc)
 
 	/* Copy the receiver's state out of shared memory. */
 	tegra_ivc_invalidate(ivc, ivc->rx.phys + offset);
-	state = ACCESS_ONCE(ivc->rx.channel->tx.state);
+	state = READ_ONCE(ivc->rx.channel->tx.state);
 
 	if (state == TEGRA_IVC_STATE_SYNC) {
 		offset = offsetof(struct tegra_ivc_header, tx.count);
-- 
2.5.2

[PATCH 04/19] fs: dcache: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the dcache code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Andrew Morton <akpm@linux-foundation.org>
---
 fs/dcache.c            | 18 +++++++++---------
 include/linux/dcache.h |  4 ++--
 2 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/fs/dcache.c b/fs/dcache.c
index f90141387f01..f283cd31a491 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -630,7 +630,7 @@ static inline struct dentry *lock_parent(struct dentry *dentry)
 	rcu_read_lock();
 	spin_unlock(&dentry->d_lock);
 again:
-	parent = ACCESS_ONCE(dentry->d_parent);
+	parent = READ_ONCE(dentry->d_parent);
 	spin_lock(&parent->d_lock);
 	/*
 	 * We can't blindly lock dentry until we are sure
@@ -721,7 +721,7 @@ static inline bool fast_dput(struct dentry *dentry)
 	 * around with a zero refcount.
 	 */
 	smp_rmb();
-	d_flags = ACCESS_ONCE(dentry->d_flags);
+	d_flags = READ_ONCE(dentry->d_flags);
 	d_flags &= DCACHE_REFERENCED | DCACHE_LRU_LIST | DCACHE_DISCONNECTED;
 
 	/* Nothing to do? Dropping the reference was all we needed? */
@@ -850,11 +850,11 @@ struct dentry *dget_parent(struct dentry *dentry)
 	 * locking.
 	 */
 	rcu_read_lock();
-	ret = ACCESS_ONCE(dentry->d_parent);
+	ret = READ_ONCE(dentry->d_parent);
 	gotref = lockref_get_not_zero(&ret->d_lockref);
 	rcu_read_unlock();
 	if (likely(gotref)) {
-		if (likely(ret == ACCESS_ONCE(dentry->d_parent)))
+		if (likely(ret == READ_ONCE(dentry->d_parent)))
 			return ret;
 		dput(ret);
 	}
@@ -3040,7 +3040,7 @@ static int prepend(char **buffer, int *buflen, const char *str, int namelen)
  * @buflen: allocated length of the buffer
  * @name:   name string and length qstr structure
  *
- * With RCU path tracing, it may race with d_move(). Use ACCESS_ONCE() to
+ * With RCU path tracing, it may race with d_move(). Use READ_ONCE() to
  * make sure that either the old or the new name pointer and length are
  * fetched. However, there may be mismatch between length and pointer.
  * The length cannot be trusted, we need to copy it byte-by-byte until
@@ -3054,8 +3054,8 @@ static int prepend(char **buffer, int *buflen, const char *str, int namelen)
  */
 static int prepend_name(char **buffer, int *buflen, const struct qstr *name)
 {
-	const char *dname = ACCESS_ONCE(name->name);
-	u32 dlen = ACCESS_ONCE(name->len);
+	const char *dname = READ_ONCE(name->name);
+	u32 dlen = READ_ONCE(name->len);
 	char *p;
 
 	smp_read_barrier_depends();
@@ -3120,7 +3120,7 @@ static int prepend_path(const struct path *path,
 		struct dentry * parent;
 
 		if (dentry == vfsmnt->mnt_root || IS_ROOT(dentry)) {
-			struct mount *parent = ACCESS_ONCE(mnt->mnt_parent);
+			struct mount *parent = READ_ONCE(mnt->mnt_parent);
 			/* Escaped? */
 			if (dentry != vfsmnt->mnt_root) {
 				bptr = *buffer;
@@ -3130,7 +3130,7 @@ static int prepend_path(const struct path *path,
 			}
 			/* Global root? */
 			if (mnt != parent) {
-				dentry = ACCESS_ONCE(mnt->mnt_mountpoint);
+				dentry = READ_ONCE(mnt->mnt_mountpoint);
 				mnt = parent;
 				vfsmnt = &mnt->mnt;
 				continue;
diff --git a/include/linux/dcache.h b/include/linux/dcache.h
index ed1a7cf6923a..1d8f5818f647 100644
--- a/include/linux/dcache.h
+++ b/include/linux/dcache.h
@@ -519,7 +519,7 @@ static inline struct inode *d_inode(const struct dentry *dentry)
 }
 
 /**
- * d_inode_rcu - Get the actual inode of this dentry with ACCESS_ONCE()
+ * d_inode_rcu - Get the actual inode of this dentry with READ_ONCE()
  * @dentry: The dentry to query
  *
  * This is the helper normal filesystems should use to get at their own inodes
@@ -527,7 +527,7 @@ static inline struct inode *d_inode(const struct dentry *dentry)
  */
 static inline struct inode *d_inode_rcu(const struct dentry *dentry)
 {
-	return ACCESS_ONCE(dentry->d_inode);
+	return READ_ONCE(dentry->d_inode);
 }
 
 /**
-- 
2.5.2

[PATCH 04/19] fs: dcache: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the dcache code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Andrew Morton <akpm@linux-foundation.org>
---
 fs/dcache.c            | 18 +++++++++---------
 include/linux/dcache.h |  4 ++--
 2 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/fs/dcache.c b/fs/dcache.c
index f90141387f01..f283cd31a491 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -630,7 +630,7 @@ static inline struct dentry *lock_parent(struct dentry *dentry)
 	rcu_read_lock();
 	spin_unlock(&dentry->d_lock);
 again:
-	parent = ACCESS_ONCE(dentry->d_parent);
+	parent = READ_ONCE(dentry->d_parent);
 	spin_lock(&parent->d_lock);
 	/*
 	 * We can't blindly lock dentry until we are sure
@@ -721,7 +721,7 @@ static inline bool fast_dput(struct dentry *dentry)
 	 * around with a zero refcount.
 	 */
 	smp_rmb();
-	d_flags = ACCESS_ONCE(dentry->d_flags);
+	d_flags = READ_ONCE(dentry->d_flags);
 	d_flags &= DCACHE_REFERENCED | DCACHE_LRU_LIST | DCACHE_DISCONNECTED;
 
 	/* Nothing to do? Dropping the reference was all we needed? */
@@ -850,11 +850,11 @@ struct dentry *dget_parent(struct dentry *dentry)
 	 * locking.
 	 */
 	rcu_read_lock();
-	ret = ACCESS_ONCE(dentry->d_parent);
+	ret = READ_ONCE(dentry->d_parent);
 	gotref = lockref_get_not_zero(&ret->d_lockref);
 	rcu_read_unlock();
 	if (likely(gotref)) {
-		if (likely(ret == ACCESS_ONCE(dentry->d_parent)))
+		if (likely(ret == READ_ONCE(dentry->d_parent)))
 			return ret;
 		dput(ret);
 	}
@@ -3040,7 +3040,7 @@ static int prepend(char **buffer, int *buflen, const char *str, int namelen)
  * @buflen: allocated length of the buffer
  * @name:   name string and length qstr structure
  *
- * With RCU path tracing, it may race with d_move(). Use ACCESS_ONCE() to
+ * With RCU path tracing, it may race with d_move(). Use READ_ONCE() to
  * make sure that either the old or the new name pointer and length are
  * fetched. However, there may be mismatch between length and pointer.
  * The length cannot be trusted, we need to copy it byte-by-byte until
@@ -3054,8 +3054,8 @@ static int prepend(char **buffer, int *buflen, const char *str, int namelen)
  */
 static int prepend_name(char **buffer, int *buflen, const struct qstr *name)
 {
-	const char *dname = ACCESS_ONCE(name->name);
-	u32 dlen = ACCESS_ONCE(name->len);
+	const char *dname = READ_ONCE(name->name);
+	u32 dlen = READ_ONCE(name->len);
 	char *p;
 
 	smp_read_barrier_depends();
@@ -3120,7 +3120,7 @@ static int prepend_path(const struct path *path,
 		struct dentry * parent;
 
 		if (dentry == vfsmnt->mnt_root || IS_ROOT(dentry)) {
-			struct mount *parent = ACCESS_ONCE(mnt->mnt_parent);
+			struct mount *parent = READ_ONCE(mnt->mnt_parent);
 			/* Escaped? */
 			if (dentry != vfsmnt->mnt_root) {
 				bptr = *buffer;
@@ -3130,7 +3130,7 @@ static int prepend_path(const struct path *path,
 			}
 			/* Global root? */
 			if (mnt != parent) {
-				dentry = ACCESS_ONCE(mnt->mnt_mountpoint);
+				dentry = READ_ONCE(mnt->mnt_mountpoint);
 				mnt = parent;
 				vfsmnt = &mnt->mnt;
 				continue;
diff --git a/include/linux/dcache.h b/include/linux/dcache.h
index ed1a7cf6923a..1d8f5818f647 100644
--- a/include/linux/dcache.h
+++ b/include/linux/dcache.h
@@ -519,7 +519,7 @@ static inline struct inode *d_inode(const struct dentry *dentry)
 }
 
 /**
- * d_inode_rcu - Get the actual inode of this dentry with ACCESS_ONCE()
+ * d_inode_rcu - Get the actual inode of this dentry with READ_ONCE()
  * @dentry: The dentry to query
  *
  * This is the helper normal filesystems should use to get at their own inodes
@@ -527,7 +527,7 @@ static inline struct inode *d_inode(const struct dentry *dentry)
  */
 static inline struct inode *d_inode_rcu(const struct dentry *dentry)
 {
-	return ACCESS_ONCE(dentry->d_inode);
+	return READ_ONCE(dentry->d_inode);
 }
 
 /**
-- 
2.5.2

[PATCH 05/19] fs: ncpfs: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

The NCPFS code has some stale comments regarding ACCESS_ONCE() uses
which were removed a long time ago.

Let's remove the stale comments.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Petr Vandrovec <petr@vandrovec.name>
---
 fs/ncpfs/dir.c | 9 ---------
 1 file changed, 9 deletions(-)

diff --git a/fs/ncpfs/dir.c b/fs/ncpfs/dir.c
index 088f52484d6e..72cfaa253a8f 100644
--- a/fs/ncpfs/dir.c
+++ b/fs/ncpfs/dir.c
@@ -119,10 +119,6 @@ static inline int ncp_case_sensitive(const struct inode *i)
 /*
  * Note: leave the hash unchanged if the directory
  * is case-sensitive.
- *
- * Accessing the parent inode can be racy under RCU pathwalking.
- * Use ACCESS_ONCE() to make sure we use _one_ particular inode,
- * the callers will handle races.
  */
 static int 
 ncp_hash_dentry(const struct dentry *dentry, struct qstr *this)
@@ -147,11 +143,6 @@ ncp_hash_dentry(const struct dentry *dentry, struct qstr *this)
 	return 0;
 }
 
-/*
- * Accessing the parent inode can be racy under RCU pathwalking.
- * Use ACCESS_ONCE() to make sure we use _one_ particular inode,
- * the callers will handle races.
- */
 static int
 ncp_compare_dentry(const struct dentry *dentry,
 		unsigned int len, const char *str, const struct qstr *name)
-- 
2.5.2

[PATCH 06/19] media: dvb_ringbuffer: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the DVB ringbuffer code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Sakari Ailus <sakari.ailus@linux.intel.com>
---
 drivers/media/dvb-core/dvb_ringbuffer.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/media/dvb-core/dvb_ringbuffer.c b/drivers/media/dvb-core/dvb_ringbuffer.c
index 2322af1b8742..53011629c9ad 100644
--- a/drivers/media/dvb-core/dvb_ringbuffer.c
+++ b/drivers/media/dvb-core/dvb_ringbuffer.c
@@ -66,12 +66,12 @@ ssize_t dvb_ringbuffer_free(struct dvb_ringbuffer *rbuf)
 {
 	ssize_t free;
 
-	/* ACCESS_ONCE() to load read pointer on writer side
+	/* READ_ONCE() to load read pointer on writer side
 	 * this pairs with smp_store_release() in dvb_ringbuffer_read(),
 	 * dvb_ringbuffer_read_user(), dvb_ringbuffer_flush(),
 	 * or dvb_ringbuffer_reset()
 	 */
-	free = ACCESS_ONCE(rbuf->pread) - rbuf->pwrite;
+	free = READ_ONCE(rbuf->pread) - rbuf->pwrite;
 	if (free <= 0)
 		free += rbuf->size;
 	return free-1;
@@ -143,7 +143,7 @@ ssize_t dvb_ringbuffer_read_user(struct dvb_ringbuffer *rbuf, u8 __user *buf, si
 		todo -= split;
 		/* smp_store_release() for read pointer update to ensure
 		 * that buf is not overwritten until read is complete,
-		 * this pairs with ACCESS_ONCE() in dvb_ringbuffer_free()
+		 * this pairs with READ_ONCE() in dvb_ringbuffer_free()
 		 */
 		smp_store_release(&rbuf->pread, 0);
 	}
@@ -168,7 +168,7 @@ void dvb_ringbuffer_read(struct dvb_ringbuffer *rbuf, u8 *buf, size_t len)
 		todo -= split;
 		/* smp_store_release() for read pointer update to ensure
 		 * that buf is not overwritten until read is complete,
-		 * this pairs with ACCESS_ONCE() in dvb_ringbuffer_free()
+		 * this pairs with READ_ONCE() in dvb_ringbuffer_free()
 		 */
 		smp_store_release(&rbuf->pread, 0);
 	}
-- 
2.5.2

[PATCH 06/19] media: dvb_ringbuffer: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the DVB ringbuffer code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Sakari Ailus <sakari.ailus@linux.intel.com>
---
 drivers/media/dvb-core/dvb_ringbuffer.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/media/dvb-core/dvb_ringbuffer.c b/drivers/media/dvb-core/dvb_ringbuffer.c
index 2322af1b8742..53011629c9ad 100644
--- a/drivers/media/dvb-core/dvb_ringbuffer.c
+++ b/drivers/media/dvb-core/dvb_ringbuffer.c
@@ -66,12 +66,12 @@ ssize_t dvb_ringbuffer_free(struct dvb_ringbuffer *rbuf)
 {
 	ssize_t free;
 
-	/* ACCESS_ONCE() to load read pointer on writer side
+	/* READ_ONCE() to load read pointer on writer side
 	 * this pairs with smp_store_release() in dvb_ringbuffer_read(),
 	 * dvb_ringbuffer_read_user(), dvb_ringbuffer_flush(),
 	 * or dvb_ringbuffer_reset()
 	 */
-	free = ACCESS_ONCE(rbuf->pread) - rbuf->pwrite;
+	free = READ_ONCE(rbuf->pread) - rbuf->pwrite;
 	if (free <= 0)
 		free += rbuf->size;
 	return free-1;
@@ -143,7 +143,7 @@ ssize_t dvb_ringbuffer_read_user(struct dvb_ringbuffer *rbuf, u8 __user *buf, si
 		todo -= split;
 		/* smp_store_release() for read pointer update to ensure
 		 * that buf is not overwritten until read is complete,
-		 * this pairs with ACCESS_ONCE() in dvb_ringbuffer_free()
+		 * this pairs with READ_ONCE() in dvb_ringbuffer_free()
 		 */
 		smp_store_release(&rbuf->pread, 0);
 	}
@@ -168,7 +168,7 @@ void dvb_ringbuffer_read(struct dvb_ringbuffer *rbuf, u8 *buf, size_t len)
 		todo -= split;
 		/* smp_store_release() for read pointer update to ensure
 		 * that buf is not overwritten until read is complete,
-		 * this pairs with ACCESS_ONCE() in dvb_ringbuffer_free()
+		 * this pairs with READ_ONCE() in dvb_ringbuffer_free()
 		 */
 		smp_store_release(&rbuf->pread, 0);
 	}
-- 
2.5.2

[PATCH 07/19] net: netlink/netfilter: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts netlink and netfilter code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Florian Westphal <fw@strlen.de>
Cc: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Cc: Pablo Neira Ayuso <pablo@netfilter.org>
---
 include/linux/genetlink.h           | 2 +-
 include/linux/netfilter/nfnetlink.h | 2 +-
 include/linux/rtnetlink.h           | 2 +-
 include/net/netfilter/nf_tables.h   | 4 ++--
 net/netfilter/ipvs/ip_vs_sync.c     | 2 +-
 net/netfilter/nfnetlink_queue.c     | 4 ++--
 6 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/include/linux/genetlink.h b/include/linux/genetlink.h
index a4c61cbce777..0e694cf62414 100644
--- a/include/linux/genetlink.h
+++ b/include/linux/genetlink.h
@@ -30,7 +30,7 @@ extern wait_queue_head_t genl_sk_destructing_waitq;
  * @p: The pointer to read, prior to dereferencing
  *
  * Return the value of the specified RCU-protected pointer, but omit
- * both the smp_read_barrier_depends() and the ACCESS_ONCE(), because
+ * both the smp_read_barrier_depends() and the READ_ONCE(), because
  * caller holds genl mutex.
  */
 #define genl_dereference(p)					\
diff --git a/include/linux/netfilter/nfnetlink.h b/include/linux/netfilter/nfnetlink.h
index 41d04e9d088a..0f47a4aa7fc4 100644
--- a/include/linux/netfilter/nfnetlink.h
+++ b/include/linux/netfilter/nfnetlink.h
@@ -66,7 +66,7 @@ static inline bool lockdep_nfnl_is_held(__u8 subsys_id)
  * @ss: The nfnetlink subsystem ID
  *
  * Return the value of the specified RCU-protected pointer, but omit
- * both the smp_read_barrier_depends() and the ACCESS_ONCE(), because
+ * both the smp_read_barrier_depends() and the READ_ONCE(), because
  * caller holds the NFNL subsystem mutex.
  */
 #define nfnl_dereference(p, ss)					\
diff --git a/include/linux/rtnetlink.h b/include/linux/rtnetlink.h
index dea59c8eec54..765f7b915475 100644
--- a/include/linux/rtnetlink.h
+++ b/include/linux/rtnetlink.h
@@ -67,7 +67,7 @@ static inline bool lockdep_rtnl_is_held(void)
  * @p: The pointer to read, prior to dereferencing
  *
  * Return the value of the specified RCU-protected pointer, but omit
- * both the smp_read_barrier_depends() and the ACCESS_ONCE(), because
+ * both the smp_read_barrier_depends() and the READ_ONCE(), because
  * caller holds RTNL.
  */
 #define rtnl_dereference(p)					\
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index 0f5b12a4ad09..5c68e279eaea 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -1164,8 +1164,8 @@ static inline u8 nft_genmask_next(const struct net *net)
 
 static inline u8 nft_genmask_cur(const struct net *net)
 {
-	/* Use ACCESS_ONCE() to prevent refetching the value for atomicity */
-	return 1 << ACCESS_ONCE(net->nft.gencursor);
+	/* Use READ_ONCE() to prevent refetching the value for atomicity */
+	return 1 << READ_ONCE(net->nft.gencursor);
 }
 
 #define NFT_GENMASK_ANY		((1 << 0) | (1 << 1))
diff --git a/net/netfilter/ipvs/ip_vs_sync.c b/net/netfilter/ipvs/ip_vs_sync.c
index 0e5b64a75da0..1cfffd42d1e2 100644
--- a/net/netfilter/ipvs/ip_vs_sync.c
+++ b/net/netfilter/ipvs/ip_vs_sync.c
@@ -457,7 +457,7 @@ static inline bool in_persistence(struct ip_vs_conn *cp)
 static int ip_vs_sync_conn_needed(struct netns_ipvs *ipvs,
 				  struct ip_vs_conn *cp, int pkts)
 {
-	unsigned long orig = ACCESS_ONCE(cp->sync_endtime);
+	unsigned long orig = READ_ONCE(cp->sync_endtime);
 	unsigned long now = jiffies;
 	unsigned long n = (now + cp->timeout) & ~3UL;
 	unsigned int sync_refresh_period;
diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index c9796629858f..a16356cacec3 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -401,7 +401,7 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue,
 
 	outdev = entry->state.out;
 
-	switch ((enum nfqnl_config_mode)ACCESS_ONCE(queue->copy_mode)) {
+	switch ((enum nfqnl_config_mode)READ_ONCE(queue->copy_mode)) {
 	case NFQNL_COPY_META:
 	case NFQNL_COPY_NONE:
 		break;
@@ -412,7 +412,7 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue,
 		    skb_checksum_help(entskb))
 			return NULL;
 
-		data_len = ACCESS_ONCE(queue->copy_range);
+		data_len = READ_ONCE(queue->copy_range);
 		if (data_len > entskb->len)
 			data_len = entskb->len;
 
-- 
2.5.2

[PATCH 08/19] net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the IPv4 TCP input code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: David S. Miller <davem@davemloft.net>
---
 net/ipv4/tcp_input.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index c5d7656beeee..133e7e4d94b5 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -815,12 +815,12 @@ static void tcp_update_pacing_rate(struct sock *sk)
 	if (likely(tp->srtt_us))
 		do_div(rate, tp->srtt_us);
 
-	/* ACCESS_ONCE() is needed because sch_fq fetches sk_pacing_rate
+	/* WRITE_ONCE() is needed because sch_fq fetches sk_pacing_rate
 	 * without any lock. We want to make sure compiler wont store
 	 * intermediate values in this location.
 	 */
-	ACCESS_ONCE(sk->sk_pacing_rate) = min_t(u64, rate,
-						sk->sk_max_pacing_rate);
+	WRITE_ONCE(sk->sk_pacing_rate, min_t(u64, rate,
+					     sk->sk_max_pacing_rate));
 }
 
 /* Calculate rto without backoff.  This is the second half of Van Jacobson's
-- 
2.5.2

[PATCH 08/19] net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the IPv4 TCP input code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: David S. Miller <davem@davemloft.net>
---
 net/ipv4/tcp_input.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index c5d7656beeee..133e7e4d94b5 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -815,12 +815,12 @@ static void tcp_update_pacing_rate(struct sock *sk)
 	if (likely(tp->srtt_us))
 		do_div(rate, tp->srtt_us);
 
-	/* ACCESS_ONCE() is needed because sch_fq fetches sk_pacing_rate
+	/* WRITE_ONCE() is needed because sch_fq fetches sk_pacing_rate
 	 * without any lock. We want to make sure compiler wont store
 	 * intermediate values in this location.
 	 */
-	ACCESS_ONCE(sk->sk_pacing_rate) = min_t(u64, rate,
-						sk->sk_max_pacing_rate);
+	WRITE_ONCE(sk->sk_pacing_rate, min_t(u64, rate,
+					     sk->sk_max_pacing_rate));
 }
 
 /* Calculate rto without backoff.  This is the second half of Van Jacobson's
-- 
2.5.2

[PATCH 09/19] net: average: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't pick up some uses, including those
in <linux/average.h>. As a preparatory step, this patch converts the
file to use {READ,WRITE}_ONCE() consistently.

At the same time, this patch addds missing includes necessary for
{READ,WRITE}_ONCE(), *BUG_ON*(), and ilog2().

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Johannes Berg <johannes.berg@intel.com>
Cc: David S. Miller <davem@davemloft.net>
---
 include/linux/average.h | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/include/linux/average.h b/include/linux/average.h
index 7ddaf340d2ac..3f462292269c 100644
--- a/include/linux/average.h
+++ b/include/linux/average.h
@@ -1,6 +1,10 @@
 #ifndef _LINUX_AVERAGE_H
 #define _LINUX_AVERAGE_H
 
+#include <linux/bug.h>
+#include <linux/compiler.h>
+#include <linux/log2.h>
+
 /*
  * Exponentially weighted moving average (EWMA)
  *
@@ -48,7 +52,7 @@
 	static inline void ewma_##name##_add(struct ewma_##name *e,	\
 					     unsigned long val)		\
 	{								\
-		unsigned long internal = ACCESS_ONCE(e->internal);	\
+		unsigned long internal = READ_ONCE(e->internal);	\
 		unsigned long weight_rcp = ilog2(_weight_rcp);		\
 		unsigned long precision = _precision;			\
 									\
@@ -57,10 +61,10 @@
 		BUILD_BUG_ON((_precision) > 30);			\
 		BUILD_BUG_ON_NOT_POWER_OF_2(_weight_rcp);		\
 									\
-		ACCESS_ONCE(e->internal) = internal ?			\
+		WRITE_ONCE(e->internal, internal ?			\
 			(((internal << weight_rcp) - internal) +	\
 				(val << precision)) >> weight_rcp :	\
-			(val << precision);				\
+			(val << precision));				\
 	}
 
 #endif /* _LINUX_AVERAGE_H */
-- 
2.5.2

Re: [PATCH 09/19] net: average: Kill off ACCESS_ONCE()

[Johannes Berg]

On Mon, 2017-10-23 at 21:07 +0000, Paul E. McKenney wrote:
> From: Mark Rutland <mark.rutland@arm.com>
> 
> For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
> preference to ACCESS_ONCE(), and new code is expected to use one of the
> former. So far, there's been no reason to change most existing uses of
> ACCESS_ONCE(), as these aren't currently harmful.
> 
> However, for some features it is necessary to instrument reads and
> writes separately, which is not possible with ACCESS_ONCE(). This
> distinction is critical to correct operation.
> 
> It's possible to transform the bulk of kernel code using the Coccinelle
> script below. However, this doesn't pick up some uses, including those
> in <linux/average.h>. As a preparatory step, this patch converts the
> file to use {READ,WRITE}_ONCE() consistently.
> 
> At the same time, this patch addds missing includes necessary for
> {READ,WRITE}_ONCE(), *BUG_ON*(), and ilog2().
> 
> ----
> virtual patch
> 
> @ depends on patch @
> expression E1, E2;
> @@
> 
> - ACCESS_ONCE(E1) = E2
> + WRITE_ONCE(E1, E2)
> 
> @ depends on patch @
> expression E;
> @@
> 
> - ACCESS_ONCE(E)
> + READ_ONCE(E)
> ----
> 
> Signed-off-by: Mark Rutland <mark.rutland@arm.com>
> Cc: Johannes Berg <johannes.berg@intel.com>

Reviewed-by: Johannes Berg <johannes@sipsolutions.net>

Let me know if you want me to apply this, since I seem to be the
average.h maintainer :-)

johannes

Re: [PATCH 09/19] net: average: Kill off ACCESS_ONCE()

[Johannes Berg]

On Mon, 2017-10-23 at 21:07 +0000, Paul E. McKenney wrote:
> From: Mark Rutland <mark.rutland@arm.com>
> 
> For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
> preference to ACCESS_ONCE(), and new code is expected to use one of the
> former. So far, there's been no reason to change most existing uses of
> ACCESS_ONCE(), as these aren't currently harmful.
> 
> However, for some features it is necessary to instrument reads and
> writes separately, which is not possible with ACCESS_ONCE(). This
> distinction is critical to correct operation.
> 
> It's possible to transform the bulk of kernel code using the Coccinelle
> script below. However, this doesn't pick up some uses, including those
> in <linux/average.h>. As a preparatory step, this patch converts the
> file to use {READ,WRITE}_ONCE() consistently.
> 
> At the same time, this patch addds missing includes necessary for
> {READ,WRITE}_ONCE(), *BUG_ON*(), and ilog2().
> 
> ----
> virtual patch
> 
> @ depends on patch @
> expression E1, E2;
> @@
> 
> - ACCESS_ONCE(E1) = E2
> + WRITE_ONCE(E1, E2)
> 
> @ depends on patch @
> expression E;
> @@
> 
> - ACCESS_ONCE(E)
> + READ_ONCE(E)
> ----
> 
> Signed-off-by: Mark Rutland <mark.rutland@arm.com>
> Cc: Johannes Berg <johannes.berg@intel.com>

Reviewed-by: Johannes Berg <johannes@sipsolutions.net>

Let me know if you want me to apply this, since I seem to be the
average.h maintainer :-)

johannes

Re: [PATCH 09/19] net: average: Kill off ACCESS_ONCE()

[Ingo Molnar]

* Johannes Berg <johannes@sipsolutions.net> wrote:

> On Mon, 2017-10-23 at 21:07 +0000, Paul E. McKenney wrote:
> > From: Mark Rutland <mark.rutland@arm.com>
> > 
> > For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
> > preference to ACCESS_ONCE(), and new code is expected to use one of the
> > former. So far, there's been no reason to change most existing uses of
> > ACCESS_ONCE(), as these aren't currently harmful.
> > 
> > However, for some features it is necessary to instrument reads and
> > writes separately, which is not possible with ACCESS_ONCE(). This
> > distinction is critical to correct operation.
> > 
> > It's possible to transform the bulk of kernel code using the Coccinelle
> > script below. However, this doesn't pick up some uses, including those
> > in <linux/average.h>. As a preparatory step, this patch converts the
> > file to use {READ,WRITE}_ONCE() consistently.
> > 
> > At the same time, this patch addds missing includes necessary for
> > {READ,WRITE}_ONCE(), *BUG_ON*(), and ilog2().
> > 
> > ----
> > virtual patch
> > 
> > @ depends on patch @
> > expression E1, E2;
> > @@
> > 
> > - ACCESS_ONCE(E1) = E2
> > + WRITE_ONCE(E1, E2)
> > 
> > @ depends on patch @
> > expression E;
> > @@
> > 
> > - ACCESS_ONCE(E)
> > + READ_ONCE(E)
> > ----
> > 
> > Signed-off-by: Mark Rutland <mark.rutland@arm.com>
> > Cc: Johannes Berg <johannes.berg@intel.com>
> 
> Reviewed-by: Johannes Berg <johannes@sipsolutions.net>

Thanks!

> Let me know if you want me to apply this, since I seem to be the
> average.h maintainer :-)

Would be nice to keep these patches together if possible, so that we can remove 
ACCESS_ONCE() - or at least add a build warning.

Thanks,

	Ingo

Re: [PATCH 09/19] net: average: Kill off ACCESS_ONCE()

[Ingo Molnar]

* Johannes Berg <johannes@sipsolutions.net> wrote:

> On Mon, 2017-10-23 at 21:07 +0000, Paul E. McKenney wrote:
> > From: Mark Rutland <mark.rutland@arm.com>
> > 
> > For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
> > preference to ACCESS_ONCE(), and new code is expected to use one of the
> > former. So far, there's been no reason to change most existing uses of
> > ACCESS_ONCE(), as these aren't currently harmful.
> > 
> > However, for some features it is necessary to instrument reads and
> > writes separately, which is not possible with ACCESS_ONCE(). This
> > distinction is critical to correct operation.
> > 
> > It's possible to transform the bulk of kernel code using the Coccinelle
> > script below. However, this doesn't pick up some uses, including those
> > in <linux/average.h>. As a preparatory step, this patch converts the
> > file to use {READ,WRITE}_ONCE() consistently.
> > 
> > At the same time, this patch addds missing includes necessary for
> > {READ,WRITE}_ONCE(), *BUG_ON*(), and ilog2().
> > 
> > ----
> > virtual patch
> > 
> > @ depends on patch @
> > expression E1, E2;
> > @@
> > 
> > - ACCESS_ONCE(E1) = E2
> > + WRITE_ONCE(E1, E2)
> > 
> > @ depends on patch @
> > expression E;
> > @@
> > 
> > - ACCESS_ONCE(E)
> > + READ_ONCE(E)
> > ----
> > 
> > Signed-off-by: Mark Rutland <mark.rutland@arm.com>
> > Cc: Johannes Berg <johannes.berg@intel.com>
> 
> Reviewed-by: Johannes Berg <johannes@sipsolutions.net>

Thanks!

> Let me know if you want me to apply this, since I seem to be the
> average.h maintainer :-)

Would be nice to keep these patches together if possible, so that we can remove 
ACCESS_ONCE() - or at least add a build warning.

Thanks,

	Ingo

Re: [PATCH 09/19] net: average: Kill off ACCESS_ONCE()

[Paul E. McKenney]

On Tue, Oct 24, 2017 at 11:34:13AM +0200, Ingo Molnar wrote:
> 
> * Johannes Berg <johannes@sipsolutions.net> wrote:
> 
> > On Mon, 2017-10-23 at 21:07 +0000, Paul E. McKenney wrote:
> > > From: Mark Rutland <mark.rutland@arm.com>
> > > 
> > > For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
> > > preference to ACCESS_ONCE(), and new code is expected to use one of the
> > > former. So far, there's been no reason to change most existing uses of
> > > ACCESS_ONCE(), as these aren't currently harmful.
> > > 
> > > However, for some features it is necessary to instrument reads and
> > > writes separately, which is not possible with ACCESS_ONCE(). This
> > > distinction is critical to correct operation.
> > > 
> > > It's possible to transform the bulk of kernel code using the Coccinelle
> > > script below. However, this doesn't pick up some uses, including those
> > > in <linux/average.h>. As a preparatory step, this patch converts the
> > > file to use {READ,WRITE}_ONCE() consistently.
> > > 
> > > At the same time, this patch addds missing includes necessary for
> > > {READ,WRITE}_ONCE(), *BUG_ON*(), and ilog2().
> > > 
> > > ----
> > > virtual patch
> > > 
> > > @ depends on patch @
> > > expression E1, E2;
> > > @@
> > > 
> > > - ACCESS_ONCE(E1) = E2
> > > + WRITE_ONCE(E1, E2)
> > > 
> > > @ depends on patch @
> > > expression E;
> > > @@
> > > 
> > > - ACCESS_ONCE(E)
> > > + READ_ONCE(E)
> > > ----
> > > 
> > > Signed-off-by: Mark Rutland <mark.rutland@arm.com>
> > > Cc: Johannes Berg <johannes.berg@intel.com>
> > 
> > Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
> 
> Thanks!
> 
> > Let me know if you want me to apply this, since I seem to be the
> > average.h maintainer :-)
> 
> Would be nice to keep these patches together if possible, so that we can remove 
> ACCESS_ONCE() - or at least add a build warning.

My goal is to remove it.  After all, I am the guy who added it.  ;-)

							Thanx, Paul

[PATCH 10/19] samples: mic/mpssd/mpssd.c: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

The bulk of the kernel code can be transformed via Coccinelle to use
{READ,WRITE}_ONCE(), though this only modifies users of ACCESS_ONCE(),
and not the implementation itself. As such, it has the potential to
break homebrew ACCESS_ONCE() macros seen in some user code in the kernel
tree (e.g. the virtio code, as fixed in commit ea9156fb3b71d9f7).

To avoid fragility if/when that transformation occurs, and to align with
the preferred usage of {READ,WRITE}_ONCE(), this patch updates the MPSSD
sample code to use READ_ONCE() rather than ACCESS_ONCE(). There should
be no functional change as a result of this patch.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Shuah Khan <shuah@kernel.org>
---
 samples/mic/mpssd/mpssd.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/samples/mic/mpssd/mpssd.c b/samples/mic/mpssd/mpssd.c
index 49db1def1721..f42ce551bb48 100644
--- a/samples/mic/mpssd/mpssd.c
+++ b/samples/mic/mpssd/mpssd.c
@@ -65,7 +65,7 @@ static struct mic_info mic_list;
 /* to align the pointer to the (next) page boundary */
 #define PAGE_ALIGN(addr)        _ALIGN(addr, PAGE_SIZE)
 
-#define ACCESS_ONCE(x) (*(volatile typeof(x) *)&(x))
+#define READ_ONCE(x) (*(volatile typeof(x) *)&(x))
 
 #define GSO_ENABLED		1
 #define MAX_GSO_SIZE		(64 * 1024)
@@ -382,7 +382,7 @@ disp_iovec(struct mic_info *mic, struct mic_copy_desc *copy,
 
 static inline __u16 read_avail_idx(struct mic_vring *vr)
 {
-	return ACCESS_ONCE(vr->info->avail_idx);
+	return READ_ONCE(vr->info->avail_idx);
 }
 
 static inline void txrx_prepare(int type, bool tx, struct mic_vring *vr,
@@ -523,7 +523,7 @@ spin_for_descriptors(struct mic_info *mic, struct mic_vring *vr)
 {
 	__u16 avail_idx = read_avail_idx(vr);
 
-	while (avail_idx == le16toh(ACCESS_ONCE(vr->vr.avail->idx))) {
+	while (avail_idx == le16toh(READ_ONCE(vr->vr.avail->idx))) {
 #ifdef DEBUG
 		mpsslog("%s %s waiting for desc avail %d info_avail %d\n",
 			mic->name, __func__,
-- 
2.5.2

[PATCH 11/19] selftests/powerpc: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

The bulk of the kernel code can be transformed via Coccinelle to use
{READ,WRITE}_ONCE(), though this only modifies users of ACCESS_ONCE(),
and not the implementation itself. As such, it has the potential to
break homebrew ACCESS_ONCE() macros seen in some user code in the kernel
tree (e.g. the virtio code, as fixed in commit ea9156fb3b71d9f7).

To avoid fragility if/when that transformation occurs, and to align with
the preferred usage of {READ,WRITE}_ONCE(), this patch updates the DSCR
selftest code to use READ_ONCE() rather than ACCESS_ONCE(). There should
be no functional change as a result of this patch.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Michael Ellerman <mpe@ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Shuah Khan <shuah@kernel.org>
---
 tools/testing/selftests/powerpc/dscr/dscr.h              | 2 +-
 tools/testing/selftests/powerpc/dscr/dscr_default_test.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/powerpc/dscr/dscr.h b/tools/testing/selftests/powerpc/dscr/dscr.h
index 18ea223bd398..cdb840bc54f2 100644
--- a/tools/testing/selftests/powerpc/dscr/dscr.h
+++ b/tools/testing/selftests/powerpc/dscr/dscr.h
@@ -39,7 +39,7 @@
 #define rmb()  asm volatile("lwsync":::"memory")
 #define wmb()  asm volatile("lwsync":::"memory")
 
-#define ACCESS_ONCE(x) (*(volatile typeof(x) *)&(x))
+#define READ_ONCE(x) (*(volatile typeof(x) *)&(x))
 
 /* Prilvilege state DSCR access */
 inline unsigned long get_dscr(void)
diff --git a/tools/testing/selftests/powerpc/dscr/dscr_default_test.c b/tools/testing/selftests/powerpc/dscr/dscr_default_test.c
index df17c3bab0a7..9e1a37e93b63 100644
--- a/tools/testing/selftests/powerpc/dscr/dscr_default_test.c
+++ b/tools/testing/selftests/powerpc/dscr/dscr_default_test.c
@@ -27,7 +27,7 @@ static void *do_test(void *in)
 		unsigned long d, cur_dscr, cur_dscr_usr;
 		unsigned long s1, s2;
 
-		s1 = ACCESS_ONCE(sequence);
+		s1 = READ_ONCE(sequence);
 		if (s1 & 1)
 			continue;
 		rmb();
-- 
2.5.2

[PATCH 11/19] selftests/powerpc: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

The bulk of the kernel code can be transformed via Coccinelle to use
{READ,WRITE}_ONCE(), though this only modifies users of ACCESS_ONCE(),
and not the implementation itself. As such, it has the potential to
break homebrew ACCESS_ONCE() macros seen in some user code in the kernel
tree (e.g. the virtio code, as fixed in commit ea9156fb3b71d9f7).

To avoid fragility if/when that transformation occurs, and to align with
the preferred usage of {READ,WRITE}_ONCE(), this patch updates the DSCR
selftest code to use READ_ONCE() rather than ACCESS_ONCE(). There should
be no functional change as a result of this patch.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Michael Ellerman <mpe@ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Shuah Khan <shuah@kernel.org>
---
 tools/testing/selftests/powerpc/dscr/dscr.h              | 2 +-
 tools/testing/selftests/powerpc/dscr/dscr_default_test.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/powerpc/dscr/dscr.h b/tools/testing/selftests/powerpc/dscr/dscr.h
index 18ea223bd398..cdb840bc54f2 100644
--- a/tools/testing/selftests/powerpc/dscr/dscr.h
+++ b/tools/testing/selftests/powerpc/dscr/dscr.h
@@ -39,7 +39,7 @@
 #define rmb()  asm volatile("lwsync":::"memory")
 #define wmb()  asm volatile("lwsync":::"memory")
 
-#define ACCESS_ONCE(x) (*(volatile typeof(x) *)&(x))
+#define READ_ONCE(x) (*(volatile typeof(x) *)&(x))
 
 /* Prilvilege state DSCR access */
 inline unsigned long get_dscr(void)
diff --git a/tools/testing/selftests/powerpc/dscr/dscr_default_test.c b/tools/testing/selftests/powerpc/dscr/dscr_default_test.c
index df17c3bab0a7..9e1a37e93b63 100644
--- a/tools/testing/selftests/powerpc/dscr/dscr_default_test.c
+++ b/tools/testing/selftests/powerpc/dscr/dscr_default_test.c
@@ -27,7 +27,7 @@ static void *do_test(void *in)
 		unsigned long d, cur_dscr, cur_dscr_usr;
 		unsigned long s1, s2;
 
-		s1 = ACCESS_ONCE(sequence);
+		s1 = READ_ONCE(sequence);
 		if (s1 & 1)
 			continue;
 		rmb();
-- 
2.5.2

[PATCH 12/19] workqueue: Kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the workqueue code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
---
 kernel/workqueue.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 64d0edf428f8..39831b2f3c5f 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -4647,7 +4647,7 @@ static void rebind_workers(struct worker_pool *pool)
 		 * concurrency management.  Note that when or whether
 		 * @worker clears REBOUND doesn't affect correctness.
 		 *
-		 * ACCESS_ONCE() is necessary because @worker->flags may be
+		 * WRITE_ONCE() is necessary because @worker->flags may be
 		 * tested without holding any lock in
 		 * wq_worker_waking_up().  Without it, NOT_RUNNING test may
 		 * fail incorrectly leading to premature concurrency
@@ -4656,7 +4656,7 @@ static void rebind_workers(struct worker_pool *pool)
 		WARN_ON_ONCE(!(worker_flags & WORKER_UNBOUND));
 		worker_flags |= WORKER_REBOUND;
 		worker_flags &= ~WORKER_UNBOUND;
-		ACCESS_ONCE(worker->flags) = worker_flags;
+		WRITE_ONCE(worker->flags, worker_flags);
 	}
 
 	spin_unlock_irq(&pool->lock);
-- 
2.5.2

[PATCH 13/19] rcutorture: formal: Prepare for ACCESS_ONCE() removal

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

The bulk of the kernel code can be transformed via Coccinelle to use
{READ,WRITE}_ONCE(), though this only modifies users of ACCESS_ONCE(),
and not the implementation itself. As such, it has the potential to
break homebrew ACCESS_ONCE() macros seen in some user code in the kernel
tree (e.g. the virtio code, as fixed in commit ea9156fb3b71d9f7).

To avoid fragility if/when that transformation occurs, this patch
reworks the definitions of {READ,WRITE}_ONCE() in the rcutorture formal
tests, and removes the unused ACCESS_ONCE() helper. There should be no
functional change as a result of this patch.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
---
 tools/testing/selftests/rcutorture/formal/srcu-cbmc/src/barriers.h | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/tools/testing/selftests/rcutorture/formal/srcu-cbmc/src/barriers.h b/tools/testing/selftests/rcutorture/formal/srcu-cbmc/src/barriers.h
index 6687acc08e6d..cc27b9ebcf20 100644
--- a/tools/testing/selftests/rcutorture/formal/srcu-cbmc/src/barriers.h
+++ b/tools/testing/selftests/rcutorture/formal/srcu-cbmc/src/barriers.h
@@ -34,8 +34,7 @@
 #define rs_smp_mb() do {} while (0)
 #endif
 
-#define ACCESS_ONCE(x) (*(volatile typeof(x) *) &(x))
-#define READ_ONCE(x) ACCESS_ONCE(x)
-#define WRITE_ONCE(x, val) (ACCESS_ONCE(x) = (val))
+#define READ_ONCE(x) (*(volatile typeof(x) *) &(x))
+#define WRITE_ONCE(x) ((*(volatile typeof(x) *) &(x)) = (val))
 
 #endif
-- 
2.5.2

[PATCH 14/19] doc: filesystems: Kill off ACCESS_ONCE()

[Paul E. McKenney]

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle documentation, leaving
references to ACCESS_ONCE() instances which have been removed. As a
preparatory step, this patch converts the filesystems documentation to
use {READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
---
 Documentation/filesystems/path-lookup.md | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/Documentation/filesystems/path-lookup.md b/Documentation/filesystems/path-lookup.md
index 1b39e084a2b2..1933ef734e63 100644
--- a/Documentation/filesystems/path-lookup.md
+++ b/Documentation/filesystems/path-lookup.md
@@ -826,9 +826,9 @@ If the filesystem may need to revalidate dcache entries, then
 *is* passed the dentry but does not have access to the `inode` or the
 `seq` number from the `nameidata`, so it needs to be extra careful
 when accessing fields in the dentry.  This "extra care" typically
-involves using `ACCESS_ONCE()` or the newer [`READ_ONCE()`] to access
-fields, and verifying the result is not NULL before using it.  This
-pattern can be see in `nfs_lookup_revalidate()`.
+involves using [`READ_ONCE()`] to access fields, and verifying the
+result is not NULL before using it.  This pattern can be seen in
+`nfs_lookup_revalidate()`.
 
 A pair of patterns
 ------------------
-- 
2.5.2

[PATCH 15/19] mm: Kill off ACCESS_ONCE()

[Paul E. McKenney]

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't currently harmful.

However, for some features it is necessary to instrument reads and
writes separately, which is not possible with ACCESS_ONCE(). This
distinction is critical to correct operation.

It's possible to transform the bulk of kernel code using the Coccinelle
script below. However, this doesn't handle comments, leaving references
to ACCESS_ONCE() instances which have been removed. As a preparatory
step, this patch converts the mm code and comments to use
{READ,WRITE}_ONCE() consistently.

----
virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
---
 mm/memory.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index a728bed16c20..cae514e7dcfc 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3891,9 +3891,9 @@ static int handle_pte_fault(struct vm_fault *vmf)
 		/*
 		 * some architectures can have larger ptes than wordsize,
 		 * e.g.ppc44x-defconfig has CONFIG_PTE_64BIT=y and
-		 * CONFIG_32BIT=y, so READ_ONCE or ACCESS_ONCE cannot guarantee
-		 * atomic accesses.  The code below just needs a consistent
-		 * view for the ifs and we later double check anyway with the
+		 * CONFIG_32BIT=y, so READ_ONCE cannot guarantee atomic
+		 * accesses.  The code below just needs a consistent view
+		 * for the ifs and we later double check anyway with the
 		 * ptl lock held. So here a barrier will do.
 		 */
 		barrier();
-- 
2.5.2

[PATCH 16/19] linux/compiler.h: Split into compiler.h and compiler-types.h

[Paul E. McKenney]

From: Will Deacon <will.deacon@arm.com>

linux/compiler.h is included indirectly by linux/types.h via
uapi/linux/types.h -> uapi/linux/posix_types.h -> linux/stddef.h
-> uapi/linux/stddef.h and is needed to provide a proper definition of
offsetof.

Unfortunately, compiler.h requires a definition of
smp_read_barrier_depends() for defining lockless_dereference and soon
for defining READ_ONCE, which means that all
users of READ_ONCE will need to include asm/barrier.h to avoid splats
such as:

   In file included from include/uapi/linux/stddef.h:1:0,
                    from include/linux/stddef.h:4,
                    from arch/h8300/kernel/asm-offsets.c:11:
   include/linux/list.h: In function 'list_empty':
>> include/linux/compiler.h:343:2: error: implicit declaration of function 'smp_read_barrier_depends' [-Werror=implicit-function-declaration]
     smp_read_barrier_depends(); /* Enforce dependency ordering from x */ \
     ^

A better alternative is to include asm/barrier.h in linux/compiler.h,
but this requires a type definition for "bool" on some architectures
(e.g. x86), which is defined later by linux/types.h. Type "bool" is also
used directly in linux/compiler.h, so the whole thing is pretty fragile.

This patch splits compiler.h in two: compiler-types.h contains type
annotations, definitions and the compiler-specific parts, whereas
compiler.h #includes compiler-types.h and additionally defines macros
such as {READ,WRITE.ACCESS}_ONCE.

uapi/linux/stddef.h and linux/linkage.h are then moved over to include
linux/compiler-types.h, which fixes the build for h8 and blackfin.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Add Will's fixes for issues reported by kbuild test robot. ]
---
 arch/arm/include/asm/ptrace.h   |   3 +-
 arch/sparc/include/asm/ptrace.h |   1 +
 arch/um/include/shared/init.h   |   2 +-
 include/linux/compiler-clang.h  |   2 +-
 include/linux/compiler-gcc.h    |   2 +-
 include/linux/compiler-intel.h  |   2 +-
 include/linux/compiler-types.h  | 274 ++++++++++++++++++++++++++++++++++++++++
 include/linux/compiler.h        | 265 +-------------------------------------
 include/linux/linkage.h         |   2 +-
 include/uapi/linux/stddef.h     |   2 +-
 scripts/headers_install.sh      |   2 +-
 11 files changed, 286 insertions(+), 271 deletions(-)
 create mode 100644 include/linux/compiler-types.h

diff --git a/arch/arm/include/asm/ptrace.h b/arch/arm/include/asm/ptrace.h
index e9c9a117bd25..c7cdbb43ae7c 100644
--- a/arch/arm/include/asm/ptrace.h
+++ b/arch/arm/include/asm/ptrace.h
@@ -126,8 +126,7 @@ extern unsigned long profile_pc(struct pt_regs *regs);
 /*
  * kprobe-based event tracer support
  */
-#include <linux/stddef.h>
-#include <linux/types.h>
+#include <linux/compiler.h>
 #define MAX_REG_OFFSET (offsetof(struct pt_regs, ARM_ORIG_r0))
 
 extern int regs_query_register_offset(const char *name);
diff --git a/arch/sparc/include/asm/ptrace.h b/arch/sparc/include/asm/ptrace.h
index d73428e4333c..b383484edcd3 100644
--- a/arch/sparc/include/asm/ptrace.h
+++ b/arch/sparc/include/asm/ptrace.h
@@ -6,6 +6,7 @@
 #if defined(__sparc__) && defined(__arch64__)
 #ifndef __ASSEMBLY__
 
+#include <linux/compiler.h>
 #include <linux/threads.h>
 #include <asm/switch_to.h>
 
diff --git a/arch/um/include/shared/init.h b/arch/um/include/shared/init.h
index 233e2593eee0..3ad7de77aad0 100644
--- a/arch/um/include/shared/init.h
+++ b/arch/um/include/shared/init.h
@@ -40,7 +40,7 @@
 typedef int (*initcall_t)(void);
 typedef void (*exitcall_t)(void);
 
-#include <linux/compiler.h>
+#include <linux/compiler-types.h>
 
 /* These are for everybody (although not all archs will actually
    discard it in modules) */
diff --git a/include/linux/compiler-clang.h b/include/linux/compiler-clang.h
index de179993e039..5947a3e6c0e6 100644
--- a/include/linux/compiler-clang.h
+++ b/include/linux/compiler-clang.h
@@ -1,4 +1,4 @@
-#ifndef __LINUX_COMPILER_H
+#ifndef __LINUX_COMPILER_TYPES_H
 #error "Please don't include <linux/compiler-clang.h> directly, include <linux/compiler.h> instead."
 #endif
 
diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h
index 16d41de92ee3..ce8e965646ef 100644
--- a/include/linux/compiler-gcc.h
+++ b/include/linux/compiler-gcc.h
@@ -1,4 +1,4 @@
-#ifndef __LINUX_COMPILER_H
+#ifndef __LINUX_COMPILER_TYPES_H
 #error "Please don't include <linux/compiler-gcc.h> directly, include <linux/compiler.h> instead."
 #endif
 
diff --git a/include/linux/compiler-intel.h b/include/linux/compiler-intel.h
index d4c71132d07f..e438ac89c692 100644
--- a/include/linux/compiler-intel.h
+++ b/include/linux/compiler-intel.h
@@ -1,4 +1,4 @@
-#ifndef __LINUX_COMPILER_H
+#ifndef __LINUX_COMPILER_TYPES_H
 #error "Please don't include <linux/compiler-intel.h> directly, include <linux/compiler.h> instead."
 #endif
 
diff --git a/include/linux/compiler-types.h b/include/linux/compiler-types.h
new file mode 100644
index 000000000000..6b79a9bba9a7
--- /dev/null
+++ b/include/linux/compiler-types.h
@@ -0,0 +1,274 @@
+#ifndef __LINUX_COMPILER_TYPES_H
+#define __LINUX_COMPILER_TYPES_H
+
+#ifndef __ASSEMBLY__
+
+#ifdef __CHECKER__
+# define __user		__attribute__((noderef, address_space(1)))
+# define __kernel	__attribute__((address_space(0)))
+# define __safe		__attribute__((safe))
+# define __force	__attribute__((force))
+# define __nocast	__attribute__((nocast))
+# define __iomem	__attribute__((noderef, address_space(2)))
+# define __must_hold(x)	__attribute__((context(x,1,1)))
+# define __acquires(x)	__attribute__((context(x,0,1)))
+# define __releases(x)	__attribute__((context(x,1,0)))
+# define __acquire(x)	__context__(x,1)
+# define __release(x)	__context__(x,-1)
+# define __cond_lock(x,c)	((c) ? ({ __acquire(x); 1; }) : 0)
+# define __percpu	__attribute__((noderef, address_space(3)))
+# define __rcu		__attribute__((noderef, address_space(4)))
+# define __private	__attribute__((noderef))
+extern void __chk_user_ptr(const volatile void __user *);
+extern void __chk_io_ptr(const volatile void __iomem *);
+# define ACCESS_PRIVATE(p, member) (*((typeof((p)->member) __force *) &(p)->member))
+#else /* __CHECKER__ */
+# ifdef STRUCTLEAK_PLUGIN
+#  define __user __attribute__((user))
+# else
+#  define __user
+# endif
+# define __kernel
+# define __safe
+# define __force
+# define __nocast
+# define __iomem
+# define __chk_user_ptr(x) (void)0
+# define __chk_io_ptr(x) (void)0
+# define __builtin_warning(x, y...) (1)
+# define __must_hold(x)
+# define __acquires(x)
+# define __releases(x)
+# define __acquire(x) (void)0
+# define __release(x) (void)0
+# define __cond_lock(x,c) (c)
+# define __percpu
+# define __rcu
+# define __private
+# define ACCESS_PRIVATE(p, member) ((p)->member)
+#endif /* __CHECKER__ */
+
+/* Indirect macros required for expanded argument pasting, eg. __LINE__. */
+#define ___PASTE(a,b) a##b
+#define __PASTE(a,b) ___PASTE(a,b)
+
+#ifdef __KERNEL__
+
+#ifdef __GNUC__
+#include <linux/compiler-gcc.h>
+#endif
+
+#if defined(CC_USING_HOTPATCH) && !defined(__CHECKER__)
+#define notrace __attribute__((hotpatch(0,0)))
+#else
+#define notrace __attribute__((no_instrument_function))
+#endif
+
+/* Intel compiler defines __GNUC__. So we will overwrite implementations
+ * coming from above header files here
+ */
+#ifdef __INTEL_COMPILER
+# include <linux/compiler-intel.h>
+#endif
+
+/* Clang compiler defines __GNUC__. So we will overwrite implementations
+ * coming from above header files here
+ */
+#ifdef __clang__
+#include <linux/compiler-clang.h>
+#endif
+
+/*
+ * Generic compiler-dependent macros required for kernel
+ * build go below this comment. Actual compiler/compiler version
+ * specific implementations come from the above header files
+ */
+
+struct ftrace_branch_data {
+	const char *func;
+	const char *file;
+	unsigned line;
+	union {
+		struct {
+			unsigned long correct;
+			unsigned long incorrect;
+		};
+		struct {
+			unsigned long miss;
+			unsigned long hit;
+		};
+		unsigned long miss_hit[2];
+	};
+};
+
+struct ftrace_likely_data {
+	struct ftrace_branch_data	data;
+	unsigned long			constant;
+};
+
+#endif /* __KERNEL__ */
+
+#endif /* __ASSEMBLY__ */
+
+#ifdef __KERNEL__
+/*
+ * Allow us to mark functions as 'deprecated' and have gcc emit a nice
+ * warning for each use, in hopes of speeding the functions removal.
+ * Usage is:
+ * 		int __deprecated foo(void)
+ */
+#ifndef __deprecated
+# define __deprecated		/* unimplemented */
+#endif
+
+#ifdef MODULE
+#define __deprecated_for_modules __deprecated
+#else
+#define __deprecated_for_modules
+#endif
+
+#ifndef __must_check
+#define __must_check
+#endif
+
+#ifndef CONFIG_ENABLE_MUST_CHECK
+#undef __must_check
+#define __must_check
+#endif
+#ifndef CONFIG_ENABLE_WARN_DEPRECATED
+#undef __deprecated
+#undef __deprecated_for_modules
+#define __deprecated
+#define __deprecated_for_modules
+#endif
+
+#ifndef __malloc
+#define __malloc
+#endif
+
+/*
+ * Allow us to avoid 'defined but not used' warnings on functions and data,
+ * as well as force them to be emitted to the assembly file.
+ *
+ * As of gcc 3.4, static functions that are not marked with attribute((used))
+ * may be elided from the assembly file.  As of gcc 3.4, static data not so
+ * marked will not be elided, but this may change in a future gcc version.
+ *
+ * NOTE: Because distributions shipped with a backported unit-at-a-time
+ * compiler in gcc 3.3, we must define __used to be __attribute__((used))
+ * for gcc >=3.3 instead of 3.4.
+ *
+ * In prior versions of gcc, such functions and data would be emitted, but
+ * would be warned about except with attribute((unused)).
+ *
+ * Mark functions that are referenced only in inline assembly as __used so
+ * the code is emitted even though it appears to be unreferenced.
+ */
+#ifndef __used
+# define __used			/* unimplemented */
+#endif
+
+#ifndef __maybe_unused
+# define __maybe_unused		/* unimplemented */
+#endif
+
+#ifndef __always_unused
+# define __always_unused	/* unimplemented */
+#endif
+
+#ifndef noinline
+#define noinline
+#endif
+
+/*
+ * Rather then using noinline to prevent stack consumption, use
+ * noinline_for_stack instead.  For documentation reasons.
+ */
+#define noinline_for_stack noinline
+
+#ifndef __always_inline
+#define __always_inline inline
+#endif
+
+#endif /* __KERNEL__ */
+
+/*
+ * From the GCC manual:
+ *
+ * Many functions do not examine any values except their arguments,
+ * and have no effects except the return value.  Basically this is
+ * just slightly more strict class than the `pure' attribute above,
+ * since function is not allowed to read global memory.
+ *
+ * Note that a function that has pointer arguments and examines the
+ * data pointed to must _not_ be declared `const'.  Likewise, a
+ * function that calls a non-`const' function usually must not be
+ * `const'.  It does not make sense for a `const' function to return
+ * `void'.
+ */
+#ifndef __attribute_const__
+# define __attribute_const__	/* unimplemented */
+#endif
+
+#ifndef __designated_init
+# define __designated_init
+#endif
+
+#ifndef __latent_entropy
+# define __latent_entropy
+#endif
+
+#ifndef __randomize_layout
+# define __randomize_layout __designated_init
+#endif
+
+#ifndef __no_randomize_layout
+# define __no_randomize_layout
+#endif
+
+#ifndef randomized_struct_fields_start
+# define randomized_struct_fields_start
+# define randomized_struct_fields_end
+#endif
+
+/*
+ * Tell gcc if a function is cold. The compiler will assume any path
+ * directly leading to the call is unlikely.
+ */
+
+#ifndef __cold
+#define __cold
+#endif
+
+/* Simple shorthand for a section definition */
+#ifndef __section
+# define __section(S) __attribute__ ((__section__(#S)))
+#endif
+
+#ifndef __visible
+#define __visible
+#endif
+
+#ifndef __nostackprotector
+# define __nostackprotector
+#endif
+
+/*
+ * Assume alignment of return value.
+ */
+#ifndef __assume_aligned
+#define __assume_aligned(a, ...)
+#endif
+
+
+/* Are two types/vars the same type (ignoring qualifiers)? */
+#ifndef __same_type
+# define __same_type(a, b) __builtin_types_compatible_p(typeof(a), typeof(b))
+#endif
+
+/* Is this type a native word size -- useful for atomic operations */
+#ifndef __native_word
+# define __native_word(t) (sizeof(t) == sizeof(char) || sizeof(t) == sizeof(short) || sizeof(t) == sizeof(int) || sizeof(t) == sizeof(long))
+#endif
+
+#endif /* __LINUX_COMPILER_TYPES_H */
diff --git a/include/linux/compiler.h b/include/linux/compiler.h
index e95a2631e545..c714239f453f 100644
--- a/include/linux/compiler.h
+++ b/include/linux/compiler.h
@@ -1,111 +1,12 @@
 #ifndef __LINUX_COMPILER_H
 #define __LINUX_COMPILER_H
 
-#ifndef __ASSEMBLY__
+#include <linux/compiler-types.h>
 
-#ifdef __CHECKER__
-# define __user		__attribute__((noderef, address_space(1)))
-# define __kernel	__attribute__((address_space(0)))
-# define __safe		__attribute__((safe))
-# define __force	__attribute__((force))
-# define __nocast	__attribute__((nocast))
-# define __iomem	__attribute__((noderef, address_space(2)))
-# define __must_hold(x)	__attribute__((context(x,1,1)))
-# define __acquires(x)	__attribute__((context(x,0,1)))
-# define __releases(x)	__attribute__((context(x,1,0)))
-# define __acquire(x)	__context__(x,1)
-# define __release(x)	__context__(x,-1)
-# define __cond_lock(x,c)	((c) ? ({ __acquire(x); 1; }) : 0)
-# define __percpu	__attribute__((noderef, address_space(3)))
-# define __rcu		__attribute__((noderef, address_space(4)))
-# define __private	__attribute__((noderef))
-extern void __chk_user_ptr(const volatile void __user *);
-extern void __chk_io_ptr(const volatile void __iomem *);
-# define ACCESS_PRIVATE(p, member) (*((typeof((p)->member) __force *) &(p)->member))
-#else /* __CHECKER__ */
-# ifdef STRUCTLEAK_PLUGIN
-#  define __user __attribute__((user))
-# else
-#  define __user
-# endif
-# define __kernel
-# define __safe
-# define __force
-# define __nocast
-# define __iomem
-# define __chk_user_ptr(x) (void)0
-# define __chk_io_ptr(x) (void)0
-# define __builtin_warning(x, y...) (1)
-# define __must_hold(x)
-# define __acquires(x)
-# define __releases(x)
-# define __acquire(x) (void)0
-# define __release(x) (void)0
-# define __cond_lock(x,c) (c)
-# define __percpu
-# define __rcu
-# define __private
-# define ACCESS_PRIVATE(p, member) ((p)->member)
-#endif /* __CHECKER__ */
-
-/* Indirect macros required for expanded argument pasting, eg. __LINE__. */
-#define ___PASTE(a,b) a##b
-#define __PASTE(a,b) ___PASTE(a,b)
+#ifndef __ASSEMBLY__
 
 #ifdef __KERNEL__
 
-#ifdef __GNUC__
-#include <linux/compiler-gcc.h>
-#endif
-
-#if defined(CC_USING_HOTPATCH) && !defined(__CHECKER__)
-#define notrace __attribute__((hotpatch(0,0)))
-#else
-#define notrace __attribute__((no_instrument_function))
-#endif
-
-/* Intel compiler defines __GNUC__. So we will overwrite implementations
- * coming from above header files here
- */
-#ifdef __INTEL_COMPILER
-# include <linux/compiler-intel.h>
-#endif
-
-/* Clang compiler defines __GNUC__. So we will overwrite implementations
- * coming from above header files here
- */
-#ifdef __clang__
-#include <linux/compiler-clang.h>
-#endif
-
-/*
- * Generic compiler-dependent macros required for kernel
- * build go below this comment. Actual compiler/compiler version
- * specific implementations come from the above header files
- */
-
-struct ftrace_branch_data {
-	const char *func;
-	const char *file;
-	unsigned line;
-	union {
-		struct {
-			unsigned long correct;
-			unsigned long incorrect;
-		};
-		struct {
-			unsigned long miss;
-			unsigned long hit;
-		};
-		unsigned long miss_hit[2];
-	};
-};
-
-struct ftrace_likely_data {
-	struct ftrace_branch_data	data;
-	unsigned long			constant;
-};
-
 /*
  * Note: DISABLE_BRANCH_PROFILING can be used by special lowlevel code
  * to disable branch tracing on a per file basis.
@@ -332,6 +233,7 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
  * with an explicit memory barrier or atomic instruction that provides the
  * required ordering.
  */
+#include <asm/barrier.h>
 
 #define __READ_ONCE(x, check)						\
 ({									\
@@ -362,167 +264,6 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
 
 #endif /* __ASSEMBLY__ */
 
-#ifdef __KERNEL__
-/*
- * Allow us to mark functions as 'deprecated' and have gcc emit a nice
- * warning for each use, in hopes of speeding the functions removal.
- * Usage is:
- * 		int __deprecated foo(void)
- */
-#ifndef __deprecated
-# define __deprecated		/* unimplemented */
-#endif
-
-#ifdef MODULE
-#define __deprecated_for_modules __deprecated
-#else
-#define __deprecated_for_modules
-#endif
-
-#ifndef __must_check
-#define __must_check
-#endif
-
-#ifndef CONFIG_ENABLE_MUST_CHECK
-#undef __must_check
-#define __must_check
-#endif
-#ifndef CONFIG_ENABLE_WARN_DEPRECATED
-#undef __deprecated
-#undef __deprecated_for_modules
-#define __deprecated
-#define __deprecated_for_modules
-#endif
-
-#ifndef __malloc
-#define __malloc
-#endif
-
-/*
- * Allow us to avoid 'defined but not used' warnings on functions and data,
- * as well as force them to be emitted to the assembly file.
- *
- * As of gcc 3.4, static functions that are not marked with attribute((used))
- * may be elided from the assembly file.  As of gcc 3.4, static data not so
- * marked will not be elided, but this may change in a future gcc version.
- *
- * NOTE: Because distributions shipped with a backported unit-at-a-time
- * compiler in gcc 3.3, we must define __used to be __attribute__((used))
- * for gcc >=3.3 instead of 3.4.
- *
- * In prior versions of gcc, such functions and data would be emitted, but
- * would be warned about except with attribute((unused)).
- *
- * Mark functions that are referenced only in inline assembly as __used so
- * the code is emitted even though it appears to be unreferenced.
- */
-#ifndef __used
-# define __used			/* unimplemented */
-#endif
-
-#ifndef __maybe_unused
-# define __maybe_unused		/* unimplemented */
-#endif
-
-#ifndef __always_unused
-# define __always_unused	/* unimplemented */
-#endif
-
-#ifndef noinline
-#define noinline
-#endif
-
-/*
- * Rather then using noinline to prevent stack consumption, use
- * noinline_for_stack instead.  For documentation reasons.
- */
-#define noinline_for_stack noinline
-
-#ifndef __always_inline
-#define __always_inline inline
-#endif
-
-#endif /* __KERNEL__ */
-
-/*
- * From the GCC manual:
- *
- * Many functions do not examine any values except their arguments,
- * and have no effects except the return value.  Basically this is
- * just slightly more strict class than the `pure' attribute above,
- * since function is not allowed to read global memory.
- *
- * Note that a function that has pointer arguments and examines the
- * data pointed to must _not_ be declared `const'.  Likewise, a
- * function that calls a non-`const' function usually must not be
- * `const'.  It does not make sense for a `const' function to return
- * `void'.
- */
-#ifndef __attribute_const__
-# define __attribute_const__	/* unimplemented */
-#endif
-
-#ifndef __designated_init
-# define __designated_init
-#endif
-
-#ifndef __latent_entropy
-# define __latent_entropy
-#endif
-
-#ifndef __randomize_layout
-# define __randomize_layout __designated_init
-#endif
-
-#ifndef __no_randomize_layout
-# define __no_randomize_layout
-#endif
-
-#ifndef randomized_struct_fields_start
-# define randomized_struct_fields_start
-# define randomized_struct_fields_end
-#endif
-
-/*
- * Tell gcc if a function is cold. The compiler will assume any path
- * directly leading to the call is unlikely.
- */
-
-#ifndef __cold
-#define __cold
-#endif
-
-/* Simple shorthand for a section definition */
-#ifndef __section
-# define __section(S) __attribute__ ((__section__(#S)))
-#endif
-
-#ifndef __visible
-#define __visible
-#endif
-
-#ifndef __nostackprotector
-# define __nostackprotector
-#endif
-
-/*
- * Assume alignment of return value.
- */
-#ifndef __assume_aligned
-#define __assume_aligned(a, ...)
-#endif
-
-
-/* Are two types/vars the same type (ignoring qualifiers)? */
-#ifndef __same_type
-# define __same_type(a, b) __builtin_types_compatible_p(typeof(a), typeof(b))
-#endif
-
-/* Is this type a native word size -- useful for atomic operations */
-#ifndef __native_word
-# define __native_word(t) (sizeof(t) == sizeof(char) || sizeof(t) == sizeof(short) || sizeof(t) == sizeof(int) || sizeof(t) == sizeof(long))
-#endif
-
 /* Compile time object size, -1 for unknown */
 #ifndef __compiletime_object_size
 # define __compiletime_object_size(obj) -1
diff --git a/include/linux/linkage.h b/include/linux/linkage.h
index a6a42dd02466..9599d13cea0c 100644
--- a/include/linux/linkage.h
+++ b/include/linux/linkage.h
@@ -1,7 +1,7 @@
 #ifndef _LINUX_LINKAGE_H
 #define _LINUX_LINKAGE_H
 
-#include <linux/compiler.h>
+#include <linux/compiler-types.h>
 #include <linux/stringify.h>
 #include <linux/export.h>
 #include <asm/linkage.h>
diff --git a/include/uapi/linux/stddef.h b/include/uapi/linux/stddef.h
index 621fa8ac4425..1ec6a7867c46 100644
--- a/include/uapi/linux/stddef.h
+++ b/include/uapi/linux/stddef.h
@@ -1,4 +1,4 @@
-#include <linux/compiler.h>
+#include <linux/compiler-types.h>
 
 #ifndef __always_inline
 #define __always_inline inline
diff --git a/scripts/headers_install.sh b/scripts/headers_install.sh
index fdebd66f8fc1..bbb7f3d35289 100755
--- a/scripts/headers_install.sh
+++ b/scripts/headers_install.sh
@@ -33,7 +33,7 @@ do
 	sed -r \
 		-e 's/([ \t(])(__user|__force|__iomem)[ \t]/\1/g' \
 		-e 's/__attribute_const__([ \t]|$)/\1/g' \
-		-e 's@^#include <linux/compiler.h>@@' \
+		-e 's@^#include <linux/compiler(|-types).h>@@' \
 		-e 's/(^|[^a-zA-Z0-9])__packed([^a-zA-Z0-9_]|$)/\1__attribute__((packed))\2/g' \
 		-e 's/(^|[ \t(])(inline|asm|volatile)([ \t(]|$)/\1__\2__\3/g' \
 		-e 's@#(ifndef|define|endif[ \t]*/[*])[ \t]*_UAPI@#\1 @' \
-- 
2.5.2

[PATCH 16/19] linux/compiler.h: Split into compiler.h and compiler-types.h

[Paul E. McKenney]

From: Will Deacon <will.deacon@arm.com>

linux/compiler.h is included indirectly by linux/types.h via
uapi/linux/types.h -> uapi/linux/posix_types.h -> linux/stddef.h
-> uapi/linux/stddef.h and is needed to provide a proper definition of
offsetof.

Unfortunately, compiler.h requires a definition of
smp_read_barrier_depends() for defining lockless_dereference and soon
for defining READ_ONCE, which means that all
users of READ_ONCE will need to include asm/barrier.h to avoid splats
such as:

   In file included from include/uapi/linux/stddef.h:1:0,
                    from include/linux/stddef.h:4,
                    from arch/h8300/kernel/asm-offsets.c:11:
   include/linux/list.h: In function 'list_empty':
>> include/linux/compiler.h:343:2: error: implicit declaration of function 'smp_read_barrier_depends' [-Werror=implicit-function-declaration]
     smp_read_barrier_depends(); /* Enforce dependency ordering from x */ \
     ^

A better alternative is to include asm/barrier.h in linux/compiler.h,
but this requires a type definition for "bool" on some architectures
(e.g. x86), which is defined later by linux/types.h. Type "bool" is also
used directly in linux/compiler.h, so the whole thing is pretty fragile.

This patch splits compiler.h in two: compiler-types.h contains type
annotations, definitions and the compiler-specific parts, whereas
compiler.h #includes compiler-types.h and additionally defines macros
such as {READ,WRITE.ACCESS}_ONCE.

uapi/linux/stddef.h and linux/linkage.h are then moved over to include
linux/compiler-types.h, which fixes the build for h8 and blackfin.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Add Will's fixes for issues reported by kbuild test robot. ]
---
 arch/arm/include/asm/ptrace.h   |   3 +-
 arch/sparc/include/asm/ptrace.h |   1 +
 arch/um/include/shared/init.h   |   2 +-
 include/linux/compiler-clang.h  |   2 +-
 include/linux/compiler-gcc.h    |   2 +-
 include/linux/compiler-intel.h  |   2 +-
 include/linux/compiler-types.h  | 274 ++++++++++++++++++++++++++++++++++++++++
 include/linux/compiler.h        | 265 +-------------------------------------
 include/linux/linkage.h         |   2 +-
 include/uapi/linux/stddef.h     |   2 +-
 scripts/headers_install.sh      |   2 +-
 11 files changed, 286 insertions(+), 271 deletions(-)
 create mode 100644 include/linux/compiler-types.h

diff --git a/arch/arm/include/asm/ptrace.h b/arch/arm/include/asm/ptrace.h
index e9c9a117bd25..c7cdbb43ae7c 100644
--- a/arch/arm/include/asm/ptrace.h
+++ b/arch/arm/include/asm/ptrace.h
@@ -126,8 +126,7 @@ extern unsigned long profile_pc(struct pt_regs *regs);
 /*
  * kprobe-based event tracer support
  */
-#include <linux/stddef.h>
-#include <linux/types.h>
+#include <linux/compiler.h>
 #define MAX_REG_OFFSET (offsetof(struct pt_regs, ARM_ORIG_r0))
 
 extern int regs_query_register_offset(const char *name);
diff --git a/arch/sparc/include/asm/ptrace.h b/arch/sparc/include/asm/ptrace.h
index d73428e4333c..b383484edcd3 100644
--- a/arch/sparc/include/asm/ptrace.h
+++ b/arch/sparc/include/asm/ptrace.h
@@ -6,6 +6,7 @@
 #if defined(__sparc__) && defined(__arch64__)
 #ifndef __ASSEMBLY__
 
+#include <linux/compiler.h>
 #include <linux/threads.h>
 #include <asm/switch_to.h>
 
diff --git a/arch/um/include/shared/init.h b/arch/um/include/shared/init.h
index 233e2593eee0..3ad7de77aad0 100644
--- a/arch/um/include/shared/init.h
+++ b/arch/um/include/shared/init.h
@@ -40,7 +40,7 @@
 typedef int (*initcall_t)(void);
 typedef void (*exitcall_t)(void);
 
-#include <linux/compiler.h>
+#include <linux/compiler-types.h>
 
 /* These are for everybody (although not all archs will actually
    discard it in modules) */
diff --git a/include/linux/compiler-clang.h b/include/linux/compiler-clang.h
index de179993e039..5947a3e6c0e6 100644
--- a/include/linux/compiler-clang.h
+++ b/include/linux/compiler-clang.h
@@ -1,4 +1,4 @@
-#ifndef __LINUX_COMPILER_H
+#ifndef __LINUX_COMPILER_TYPES_H
 #error "Please don't include <linux/compiler-clang.h> directly, include <linux/compiler.h> instead."
 #endif
 
diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h
index 16d41de92ee3..ce8e965646ef 100644
--- a/include/linux/compiler-gcc.h
+++ b/include/linux/compiler-gcc.h
@@ -1,4 +1,4 @@
-#ifndef __LINUX_COMPILER_H
+#ifndef __LINUX_COMPILER_TYPES_H
 #error "Please don't include <linux/compiler-gcc.h> directly, include <linux/compiler.h> instead."
 #endif
 
diff --git a/include/linux/compiler-intel.h b/include/linux/compiler-intel.h
index d4c71132d07f..e438ac89c692 100644
--- a/include/linux/compiler-intel.h
+++ b/include/linux/compiler-intel.h
@@ -1,4 +1,4 @@
-#ifndef __LINUX_COMPILER_H
+#ifndef __LINUX_COMPILER_TYPES_H
 #error "Please don't include <linux/compiler-intel.h> directly, include <linux/compiler.h> instead."
 #endif
 
diff --git a/include/linux/compiler-types.h b/include/linux/compiler-types.h
new file mode 100644
index 000000000000..6b79a9bba9a7
--- /dev/null
+++ b/include/linux/compiler-types.h
@@ -0,0 +1,274 @@
+#ifndef __LINUX_COMPILER_TYPES_H
+#define __LINUX_COMPILER_TYPES_H
+
+#ifndef __ASSEMBLY__
+
+#ifdef __CHECKER__
+# define __user		__attribute__((noderef, address_space(1)))
+# define __kernel	__attribute__((address_space(0)))
+# define __safe		__attribute__((safe))
+# define __force	__attribute__((force))
+# define __nocast	__attribute__((nocast))
+# define __iomem	__attribute__((noderef, address_space(2)))
+# define __must_hold(x)	__attribute__((context(x,1,1)))
+# define __acquires(x)	__attribute__((context(x,0,1)))
+# define __releases(x)	__attribute__((context(x,1,0)))
+# define __acquire(x)	__context__(x,1)
+# define __release(x)	__context__(x,-1)
+# define __cond_lock(x,c)	((c) ? ({ __acquire(x); 1; }) : 0)
+# define __percpu	__attribute__((noderef, address_space(3)))
+# define __rcu		__attribute__((noderef, address_space(4)))
+# define __private	__attribute__((noderef))
+extern void __chk_user_ptr(const volatile void __user *);
+extern void __chk_io_ptr(const volatile void __iomem *);
+# define ACCESS_PRIVATE(p, member) (*((typeof((p)->member) __force *) &(p)->member))
+#else /* __CHECKER__ */
+# ifdef STRUCTLEAK_PLUGIN
+#  define __user __attribute__((user))
+# else
+#  define __user
+# endif
+# define __kernel
+# define __safe
+# define __force
+# define __nocast
+# define __iomem
+# define __chk_user_ptr(x) (void)0
+# define __chk_io_ptr(x) (void)0
+# define __builtin_warning(x, y...) (1)
+# define __must_hold(x)
+# define __acquires(x)
+# define __releases(x)
+# define __acquire(x) (void)0
+# define __release(x) (void)0
+# define __cond_lock(x,c) (c)
+# define __percpu
+# define __rcu
+# define __private
+# define ACCESS_PRIVATE(p, member) ((p)->member)
+#endif /* __CHECKER__ */
+
+/* Indirect macros required for expanded argument pasting, eg. __LINE__. */
+#define ___PASTE(a,b) a##b
+#define __PASTE(a,b) ___PASTE(a,b)
+
+#ifdef __KERNEL__
+
+#ifdef __GNUC__
+#include <linux/compiler-gcc.h>
+#endif
+
+#if defined(CC_USING_HOTPATCH) && !defined(__CHECKER__)
+#define notrace __attribute__((hotpatch(0,0)))
+#else
+#define notrace __attribute__((no_instrument_function))
+#endif
+
+/* Intel compiler defines __GNUC__. So we will overwrite implementations
+ * coming from above header files here
+ */
+#ifdef __INTEL_COMPILER
+# include <linux/compiler-intel.h>
+#endif
+
+/* Clang compiler defines __GNUC__. So we will overwrite implementations
+ * coming from above header files here
+ */
+#ifdef __clang__
+#include <linux/compiler-clang.h>
+#endif
+
+/*
+ * Generic compiler-dependent macros required for kernel
+ * build go below this comment. Actual compiler/compiler version
+ * specific implementations come from the above header files
+ */
+
+struct ftrace_branch_data {
+	const char *func;
+	const char *file;
+	unsigned line;
+	union {
+		struct {
+			unsigned long correct;
+			unsigned long incorrect;
+		};
+		struct {
+			unsigned long miss;
+			unsigned long hit;
+		};
+		unsigned long miss_hit[2];
+	};
+};
+
+struct ftrace_likely_data {
+	struct ftrace_branch_data	data;
+	unsigned long			constant;
+};
+
+#endif /* __KERNEL__ */
+
+#endif /* __ASSEMBLY__ */
+
+#ifdef __KERNEL__
+/*
+ * Allow us to mark functions as 'deprecated' and have gcc emit a nice
+ * warning for each use, in hopes of speeding the functions removal.
+ * Usage is:
+ * 		int __deprecated foo(void)
+ */
+#ifndef __deprecated
+# define __deprecated		/* unimplemented */
+#endif
+
+#ifdef MODULE
+#define __deprecated_for_modules __deprecated
+#else
+#define __deprecated_for_modules
+#endif
+
+#ifndef __must_check
+#define __must_check
+#endif
+
+#ifndef CONFIG_ENABLE_MUST_CHECK
+#undef __must_check
+#define __must_check
+#endif
+#ifndef CONFIG_ENABLE_WARN_DEPRECATED
+#undef __deprecated
+#undef __deprecated_for_modules
+#define __deprecated
+#define __deprecated_for_modules
+#endif
+
+#ifndef __malloc
+#define __malloc
+#endif
+
+/*
+ * Allow us to avoid 'defined but not used' warnings on functions and data,
+ * as well as force them to be emitted to the assembly file.
+ *
+ * As of gcc 3.4, static functions that are not marked with attribute((used))
+ * may be elided from the assembly file.  As of gcc 3.4, static data not so
+ * marked will not be elided, but this may change in a future gcc version.
+ *
+ * NOTE: Because distributions shipped with a backported unit-at-a-time
+ * compiler in gcc 3.3, we must define __used to be __attribute__((used))
+ * for gcc >=3.3 instead of 3.4.
+ *
+ * In prior versions of gcc, such functions and data would be emitted, but
+ * would be warned about except with attribute((unused)).
+ *
+ * Mark functions that are referenced only in inline assembly as __used so
+ * the code is emitted even though it appears to be unreferenced.
+ */
+#ifndef __used
+# define __used			/* unimplemented */
+#endif
+
+#ifndef __maybe_unused
+# define __maybe_unused		/* unimplemented */
+#endif
+
+#ifndef __always_unused
+# define __always_unused	/* unimplemented */
+#endif
+
+#ifndef noinline
+#define noinline
+#endif
+
+/*
+ * Rather then using noinline to prevent stack consumption, use
+ * noinline_for_stack instead.  For documentation reasons.
+ */
+#define noinline_for_stack noinline
+
+#ifndef __always_inline
+#define __always_inline inline
+#endif
+
+#endif /* __KERNEL__ */
+
+/*
+ * From the GCC manual:
+ *
+ * Many functions do not examine any values except their arguments,
+ * and have no effects except the return value.  Basically this is
+ * just slightly more strict class than the `pure' attribute above,
+ * since function is not allowed to read global memory.
+ *
+ * Note that a function that has pointer arguments and examines the
+ * data pointed to must _not_ be declared `const'.  Likewise, a
+ * function that calls a non-`const' function usually must not be
+ * `const'.  It does not make sense for a `const' function to return
+ * `void'.
+ */
+#ifndef __attribute_const__
+# define __attribute_const__	/* unimplemented */
+#endif
+
+#ifndef __designated_init
+# define __designated_init
+#endif
+
+#ifndef __latent_entropy
+# define __latent_entropy
+#endif
+
+#ifndef __randomize_layout
+# define __randomize_layout __designated_init
+#endif
+
+#ifndef __no_randomize_layout
+# define __no_randomize_layout
+#endif
+
+#ifndef randomized_struct_fields_start
+# define randomized_struct_fields_start
+# define randomized_struct_fields_end
+#endif
+
+/*
+ * Tell gcc if a function is cold. The compiler will assume any path
+ * directly leading to the call is unlikely.
+ */
+
+#ifndef __cold
+#define __cold
+#endif
+
+/* Simple shorthand for a section definition */
+#ifndef __section
+# define __section(S) __attribute__ ((__section__(#S)))
+#endif
+
+#ifndef __visible
+#define __visible
+#endif
+
+#ifndef __nostackprotector
+# define __nostackprotector
+#endif
+
+/*
+ * Assume alignment of return value.
+ */
+#ifndef __assume_aligned
+#define __assume_aligned(a, ...)
+#endif
+
+
+/* Are two types/vars the same type (ignoring qualifiers)? */
+#ifndef __same_type
+# define __same_type(a, b) __builtin_types_compatible_p(typeof(a), typeof(b))
+#endif
+
+/* Is this type a native word size -- useful for atomic operations */
+#ifndef __native_word
+# define __native_word(t) (sizeof(t) == sizeof(char) || sizeof(t) == sizeof(short) || sizeof(t) == sizeof(int) || sizeof(t) == sizeof(long))
+#endif
+
+#endif /* __LINUX_COMPILER_TYPES_H */
diff --git a/include/linux/compiler.h b/include/linux/compiler.h
index e95a2631e545..c714239f453f 100644
--- a/include/linux/compiler.h
+++ b/include/linux/compiler.h
@@ -1,111 +1,12 @@
 #ifndef __LINUX_COMPILER_H
 #define __LINUX_COMPILER_H
 
-#ifndef __ASSEMBLY__
+#include <linux/compiler-types.h>
 
-#ifdef __CHECKER__
-# define __user		__attribute__((noderef, address_space(1)))
-# define __kernel	__attribute__((address_space(0)))
-# define __safe		__attribute__((safe))
-# define __force	__attribute__((force))
-# define __nocast	__attribute__((nocast))
-# define __iomem	__attribute__((noderef, address_space(2)))
-# define __must_hold(x)	__attribute__((context(x,1,1)))
-# define __acquires(x)	__attribute__((context(x,0,1)))
-# define __releases(x)	__attribute__((context(x,1,0)))
-# define __acquire(x)	__context__(x,1)
-# define __release(x)	__context__(x,-1)
-# define __cond_lock(x,c)	((c) ? ({ __acquire(x); 1; }) : 0)
-# define __percpu	__attribute__((noderef, address_space(3)))
-# define __rcu		__attribute__((noderef, address_space(4)))
-# define __private	__attribute__((noderef))
-extern void __chk_user_ptr(const volatile void __user *);
-extern void __chk_io_ptr(const volatile void __iomem *);
-# define ACCESS_PRIVATE(p, member) (*((typeof((p)->member) __force *) &(p)->member))
-#else /* __CHECKER__ */
-# ifdef STRUCTLEAK_PLUGIN
-#  define __user __attribute__((user))
-# else
-#  define __user
-# endif
-# define __kernel
-# define __safe
-# define __force
-# define __nocast
-# define __iomem
-# define __chk_user_ptr(x) (void)0
-# define __chk_io_ptr(x) (void)0
-# define __builtin_warning(x, y...) (1)
-# define __must_hold(x)
-# define __acquires(x)
-# define __releases(x)
-# define __acquire(x) (void)0
-# define __release(x) (void)0
-# define __cond_lock(x,c) (c)
-# define __percpu
-# define __rcu
-# define __private
-# define ACCESS_PRIVATE(p, member) ((p)->member)
-#endif /* __CHECKER__ */
-
-/* Indirect macros required for expanded argument pasting, eg. __LINE__. */
-#define ___PASTE(a,b) a##b
-#define __PASTE(a,b) ___PASTE(a,b)
+#ifndef __ASSEMBLY__
 
 #ifdef __KERNEL__
 
-#ifdef __GNUC__
-#include <linux/compiler-gcc.h>
-#endif
-
-#if defined(CC_USING_HOTPATCH) && !defined(__CHECKER__)
-#define notrace __attribute__((hotpatch(0,0)))
-#else
-#define notrace __attribute__((no_instrument_function))
-#endif
-
-/* Intel compiler defines __GNUC__. So we will overwrite implementations
- * coming from above header files here
- */
-#ifdef __INTEL_COMPILER
-# include <linux/compiler-intel.h>
-#endif
-
-/* Clang compiler defines __GNUC__. So we will overwrite implementations
- * coming from above header files here
- */
-#ifdef __clang__
-#include <linux/compiler-clang.h>
-#endif
-
-/*
- * Generic compiler-dependent macros required for kernel
- * build go below this comment. Actual compiler/compiler version
- * specific implementations come from the above header files
- */
-
-struct ftrace_branch_data {
-	const char *func;
-	const char *file;
-	unsigned line;
-	union {
-		struct {
-			unsigned long correct;
-			unsigned long incorrect;
-		};
-		struct {
-			unsigned long miss;
-			unsigned long hit;
-		};
-		unsigned long miss_hit[2];
-	};
-};
-
-struct ftrace_likely_data {
-	struct ftrace_branch_data	data;
-	unsigned long			constant;
-};
-
 /*
  * Note: DISABLE_BRANCH_PROFILING can be used by special lowlevel code
  * to disable branch tracing on a per file basis.
@@ -332,6 +233,7 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
  * with an explicit memory barrier or atomic instruction that provides the
  * required ordering.
  */
+#include <asm/barrier.h>
 
 #define __READ_ONCE(x, check)						\
 ({									\
@@ -362,167 +264,6 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
 
 #endif /* __ASSEMBLY__ */
 
-#ifdef __KERNEL__
-/*
- * Allow us to mark functions as 'deprecated' and have gcc emit a nice
- * warning for each use, in hopes of speeding the functions removal.
- * Usage is:
- * 		int __deprecated foo(void)
- */
-#ifndef __deprecated
-# define __deprecated		/* unimplemented */
-#endif
-
-#ifdef MODULE
-#define __deprecated_for_modules __deprecated
-#else
-#define __deprecated_for_modules
-#endif
-
-#ifndef __must_check
-#define __must_check
-#endif
-
-#ifndef CONFIG_ENABLE_MUST_CHECK
-#undef __must_check
-#define __must_check
-#endif
-#ifndef CONFIG_ENABLE_WARN_DEPRECATED
-#undef __deprecated
-#undef __deprecated_for_modules
-#define __deprecated
-#define __deprecated_for_modules
-#endif
-
-#ifndef __malloc
-#define __malloc
-#endif
-
-/*
- * Allow us to avoid 'defined but not used' warnings on functions and data,
- * as well as force them to be emitted to the assembly file.
- *
- * As of gcc 3.4, static functions that are not marked with attribute((used))
- * may be elided from the assembly file.  As of gcc 3.4, static data not so
- * marked will not be elided, but this may change in a future gcc version.
- *
- * NOTE: Because distributions shipped with a backported unit-at-a-time
- * compiler in gcc 3.3, we must define __used to be __attribute__((used))
- * for gcc >=3.3 instead of 3.4.
- *
- * In prior versions of gcc, such functions and data would be emitted, but
- * would be warned about except with attribute((unused)).
- *
- * Mark functions that are referenced only in inline assembly as __used so
- * the code is emitted even though it appears to be unreferenced.
- */
-#ifndef __used
-# define __used			/* unimplemented */
-#endif
-
-#ifndef __maybe_unused
-# define __maybe_unused		/* unimplemented */
-#endif
-
-#ifndef __always_unused
-# define __always_unused	/* unimplemented */
-#endif
-
-#ifndef noinline
-#define noinline
-#endif
-
-/*
- * Rather then using noinline to prevent stack consumption, use
- * noinline_for_stack instead.  For documentation reasons.
- */
-#define noinline_for_stack noinline
-
-#ifndef __always_inline
-#define __always_inline inline
-#endif
-
-#endif /* __KERNEL__ */
-
-/*
- * From the GCC manual:
- *
- * Many functions do not examine any values except their arguments,
- * and have no effects except the return value.  Basically this is
- * just slightly more strict class than the `pure' attribute above,
- * since function is not allowed to read global memory.
- *
- * Note that a function that has pointer arguments and examines the
- * data pointed to must _not_ be declared `const'.  Likewise, a
- * function that calls a non-`const' function usually must not be
- * `const'.  It does not make sense for a `const' function to return
- * `void'.
- */
-#ifndef __attribute_const__
-# define __attribute_const__	/* unimplemented */
-#endif
-
-#ifndef __designated_init
-# define __designated_init
-#endif
-
-#ifndef __latent_entropy
-# define __latent_entropy
-#endif
-
-#ifndef __randomize_layout
-# define __randomize_layout __designated_init
-#endif
-
-#ifndef __no_randomize_layout
-# define __no_randomize_layout
-#endif
-
-#ifndef randomized_struct_fields_start
-# define randomized_struct_fields_start
-# define randomized_struct_fields_end
-#endif
-
-/*
- * Tell gcc if a function is cold. The compiler will assume any path
- * directly leading to the call is unlikely.
- */
-
-#ifndef __cold
-#define __cold
-#endif
-
-/* Simple shorthand for a section definition */
-#ifndef __section
-# define __section(S) __attribute__ ((__section__(#S)))
-#endif
-
-#ifndef __visible
-#define __visible
-#endif
-
-#ifndef __nostackprotector
-# define __nostackprotector
-#endif
-
-/*
- * Assume alignment of return value.
- */
-#ifndef __assume_aligned
-#define __assume_aligned(a, ...)
-#endif
-
-
-/* Are two types/vars the same type (ignoring qualifiers)? */
-#ifndef __same_type
-# define __same_type(a, b) __builtin_types_compatible_p(typeof(a), typeof(b))
-#endif
-
-/* Is this type a native word size -- useful for atomic operations */
-#ifndef __native_word
-# define __native_word(t) (sizeof(t) == sizeof(char) || sizeof(t) == sizeof(short) || sizeof(t) == sizeof(int) || sizeof(t) == sizeof(long))
-#endif
-
 /* Compile time object size, -1 for unknown */
 #ifndef __compiletime_object_size
 # define __compiletime_object_size(obj) -1
diff --git a/include/linux/linkage.h b/include/linux/linkage.h
index a6a42dd02466..9599d13cea0c 100644
--- a/include/linux/linkage.h
+++ b/include/linux/linkage.h
@@ -1,7 +1,7 @@
 #ifndef _LINUX_LINKAGE_H
 #define _LINUX_LINKAGE_H
 
-#include <linux/compiler.h>
+#include <linux/compiler-types.h>
 #include <linux/stringify.h>
 #include <linux/export.h>
 #include <asm/linkage.h>
diff --git a/include/uapi/linux/stddef.h b/include/uapi/linux/stddef.h
index 621fa8ac4425..1ec6a7867c46 100644
--- a/include/uapi/linux/stddef.h
+++ b/include/uapi/linux/stddef.h
@@ -1,4 +1,4 @@
-#include <linux/compiler.h>
+#include <linux/compiler-types.h>
 
 #ifndef __always_inline
 #define __always_inline inline
diff --git a/scripts/headers_install.sh b/scripts/headers_install.sh
index fdebd66f8fc1..bbb7f3d35289 100755
--- a/scripts/headers_install.sh
+++ b/scripts/headers_install.sh
@@ -33,7 +33,7 @@ do
 	sed -r \
 		-e 's/([ \t(])(__user|__force|__iomem)[ \t]/\1/g' \
 		-e 's/__attribute_const__([ \t]|$)/\1/g' \
-		-e 's@^#include <linux/compiler.h>@@' \
+		-e 's@^#include <linux/compiler(|-types).h>@@' \
 		-e 's/(^|[^a-zA-Z0-9])__packed([^a-zA-Z0-9_]|$)/\1__attribute__((packed))\2/g' \
 		-e 's/(^|[ \t(])(inline|asm|volatile)([ \t(]|$)/\1__\2__\3/g' \
 		-e 's@#(ifndef|define|endif[ \t]*/[*])[ \t]*_UAPI@#\1 @' \
-- 
2.5.2

Re: [PATCH 16/19] linux/compiler.h: Split into compiler.h and compiler-types.h

[Ingo Molnar]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> From: Will Deacon <will.deacon@arm.com>
> 
> linux/compiler.h is included indirectly by linux/types.h via
> uapi/linux/types.h -> uapi/linux/posix_types.h -> linux/stddef.h
> -> uapi/linux/stddef.h and is needed to provide a proper definition of
> offsetof.
> 
> Unfortunately, compiler.h requires a definition of
> smp_read_barrier_depends() for defining lockless_dereference and soon
> for defining READ_ONCE, which means that all
> users of READ_ONCE will need to include asm/barrier.h to avoid splats
> such as:
> 
>    In file included from include/uapi/linux/stddef.h:1:0,
>                     from include/linux/stddef.h:4,
>                     from arch/h8300/kernel/asm-offsets.c:11:
>    include/linux/list.h: In function 'list_empty':
> >> include/linux/compiler.h:343:2: error: implicit declaration of function 'smp_read_barrier_depends' [-Werror=implicit-function-declaration]
>      smp_read_barrier_depends(); /* Enforce dependency ordering from x */ \
>      ^
> 
> A better alternative is to include asm/barrier.h in linux/compiler.h,
> but this requires a type definition for "bool" on some architectures
> (e.g. x86), which is defined later by linux/types.h. Type "bool" is also
> used directly in linux/compiler.h, so the whole thing is pretty fragile.
> 
> This patch splits compiler.h in two: compiler-types.h contains type
> annotations, definitions and the compiler-specific parts, whereas
> compiler.h #includes compiler-types.h and additionally defines macros
> such as {READ,WRITE.ACCESS}_ONCE.
> 
> uapi/linux/stddef.h and linux/linkage.h are then moved over to include
> linux/compiler-types.h, which fixes the build for h8 and blackfin.
> 
> Reported-by: kbuild test robot <fengguang.wu@intel.com>
> Signed-off-by: Will Deacon <will.deacon@arm.com>
> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> [ paulmck: Add Will's fixes for issues reported by kbuild test robot. ]
> ---
>  arch/arm/include/asm/ptrace.h   |   3 +-
>  arch/sparc/include/asm/ptrace.h |   1 +
>  arch/um/include/shared/init.h   |   2 +-
>  include/linux/compiler-clang.h  |   2 +-
>  include/linux/compiler-gcc.h    |   2 +-
>  include/linux/compiler-intel.h  |   2 +-
>  include/linux/compiler-types.h  | 274 ++++++++++++++++++++++++++++++++++++++++
>  include/linux/compiler.h        | 265 +-------------------------------------

The common pattern is to use an underscore (compiler_types.h), not a hyphen 
(compiler-types.h) ...

Even the patch description cites such an example, posix_types.h:

> linux/compiler.h is included indirectly by linux/types.h via
> uapi/linux/types.h -> uapi/linux/posix_types.h -> linux/stddef.h

Thanks,

	Ingo

Re: [PATCH 16/19] linux/compiler.h: Split into compiler.h and compiler-types.h

[Will Deacon]

On Tue, Oct 24, 2017 at 11:25:12AM +0200, Ingo Molnar wrote:
> 
> * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> 
> > From: Will Deacon <will.deacon@arm.com>
> > 
> > linux/compiler.h is included indirectly by linux/types.h via
> > uapi/linux/types.h -> uapi/linux/posix_types.h -> linux/stddef.h
> > -> uapi/linux/stddef.h and is needed to provide a proper definition of
> > offsetof.
> > 
> > Unfortunately, compiler.h requires a definition of
> > smp_read_barrier_depends() for defining lockless_dereference and soon
> > for defining READ_ONCE, which means that all
> > users of READ_ONCE will need to include asm/barrier.h to avoid splats
> > such as:
> > 
> >    In file included from include/uapi/linux/stddef.h:1:0,
> >                     from include/linux/stddef.h:4,
> >                     from arch/h8300/kernel/asm-offsets.c:11:
> >    include/linux/list.h: In function 'list_empty':
> > >> include/linux/compiler.h:343:2: error: implicit declaration of function 'smp_read_barrier_depends' [-Werror=implicit-function-declaration]
> >      smp_read_barrier_depends(); /* Enforce dependency ordering from x */ \
> >      ^
> > 
> > A better alternative is to include asm/barrier.h in linux/compiler.h,
> > but this requires a type definition for "bool" on some architectures
> > (e.g. x86), which is defined later by linux/types.h. Type "bool" is also
> > used directly in linux/compiler.h, so the whole thing is pretty fragile.
> > 
> > This patch splits compiler.h in two: compiler-types.h contains type
> > annotations, definitions and the compiler-specific parts, whereas
> > compiler.h #includes compiler-types.h and additionally defines macros
> > such as {READ,WRITE.ACCESS}_ONCE.
> > 
> > uapi/linux/stddef.h and linux/linkage.h are then moved over to include
> > linux/compiler-types.h, which fixes the build for h8 and blackfin.
> > 
> > Reported-by: kbuild test robot <fengguang.wu@intel.com>
> > Signed-off-by: Will Deacon <will.deacon@arm.com>
> > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > [ paulmck: Add Will's fixes for issues reported by kbuild test robot. ]
> > ---
> >  arch/arm/include/asm/ptrace.h   |   3 +-
> >  arch/sparc/include/asm/ptrace.h |   1 +
> >  arch/um/include/shared/init.h   |   2 +-
> >  include/linux/compiler-clang.h  |   2 +-
> >  include/linux/compiler-gcc.h    |   2 +-
> >  include/linux/compiler-intel.h  |   2 +-
> >  include/linux/compiler-types.h  | 274 ++++++++++++++++++++++++++++++++++++++++
> >  include/linux/compiler.h        | 265 +-------------------------------------
> 
> The common pattern is to use an underscore (compiler_types.h), not a hyphen 
> (compiler-types.h) ...

I just followed the existing examples of compiler-{clang,gcc,intel}.h!
I can respin if you like, but it will be inconsistent with the existing
compiler* headers and it's not like there aren't other hyphenated header
files in include/linux/ (I count 433).

Will

Re: [PATCH 16/19] linux/compiler.h: Split into compiler.h and compiler-types.h

[Will Deacon]

On Tue, Oct 24, 2017 at 11:25:12AM +0200, Ingo Molnar wrote:
> 
> * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> 
> > From: Will Deacon <will.deacon@arm.com>
> > 
> > linux/compiler.h is included indirectly by linux/types.h via
> > uapi/linux/types.h -> uapi/linux/posix_types.h -> linux/stddef.h
> > -> uapi/linux/stddef.h and is needed to provide a proper definition of
> > offsetof.
> > 
> > Unfortunately, compiler.h requires a definition of
> > smp_read_barrier_depends() for defining lockless_dereference and soon
> > for defining READ_ONCE, which means that all
> > users of READ_ONCE will need to include asm/barrier.h to avoid splats
> > such as:
> > 
> >    In file included from include/uapi/linux/stddef.h:1:0,
> >                     from include/linux/stddef.h:4,
> >                     from arch/h8300/kernel/asm-offsets.c:11:
> >    include/linux/list.h: In function 'list_empty':
> > >> include/linux/compiler.h:343:2: error: implicit declaration of function 'smp_read_barrier_depends' [-Werror=implicit-function-declaration]
> >      smp_read_barrier_depends(); /* Enforce dependency ordering from x */ \
> >      ^
> > 
> > A better alternative is to include asm/barrier.h in linux/compiler.h,
> > but this requires a type definition for "bool" on some architectures
> > (e.g. x86), which is defined later by linux/types.h. Type "bool" is also
> > used directly in linux/compiler.h, so the whole thing is pretty fragile.
> > 
> > This patch splits compiler.h in two: compiler-types.h contains type
> > annotations, definitions and the compiler-specific parts, whereas
> > compiler.h #includes compiler-types.h and additionally defines macros
> > such as {READ,WRITE.ACCESS}_ONCE.
> > 
> > uapi/linux/stddef.h and linux/linkage.h are then moved over to include
> > linux/compiler-types.h, which fixes the build for h8 and blackfin.
> > 
> > Reported-by: kbuild test robot <fengguang.wu@intel.com>
> > Signed-off-by: Will Deacon <will.deacon@arm.com>
> > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > [ paulmck: Add Will's fixes for issues reported by kbuild test robot. ]
> > ---
> >  arch/arm/include/asm/ptrace.h   |   3 +-
> >  arch/sparc/include/asm/ptrace.h |   1 +
> >  arch/um/include/shared/init.h   |   2 +-
> >  include/linux/compiler-clang.h  |   2 +-
> >  include/linux/compiler-gcc.h    |   2 +-
> >  include/linux/compiler-intel.h  |   2 +-
> >  include/linux/compiler-types.h  | 274 ++++++++++++++++++++++++++++++++++++++++
> >  include/linux/compiler.h        | 265 +-------------------------------------
> 
> The common pattern is to use an underscore (compiler_types.h), not a hyphen 
> (compiler-types.h) ...

I just followed the existing examples of compiler-{clang,gcc,intel}.h!
I can respin if you like, but it will be inconsistent with the existing
compiler* headers and it's not like there aren't other hyphenated header
files in include/linux/ (I count 433).

Will

Re: [PATCH 16/19] linux/compiler.h: Split into compiler.h and compiler-types.h

[Ingo Molnar]

* Will Deacon <will.deacon@arm.com> wrote:

> On Tue, Oct 24, 2017 at 11:25:12AM +0200, Ingo Molnar wrote:
> > 
> > * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> > 
> > > From: Will Deacon <will.deacon@arm.com>
> > > 
> > > linux/compiler.h is included indirectly by linux/types.h via
> > > uapi/linux/types.h -> uapi/linux/posix_types.h -> linux/stddef.h
> > > -> uapi/linux/stddef.h and is needed to provide a proper definition of
> > > offsetof.
> > > 
> > > Unfortunately, compiler.h requires a definition of
> > > smp_read_barrier_depends() for defining lockless_dereference and soon
> > > for defining READ_ONCE, which means that all
> > > users of READ_ONCE will need to include asm/barrier.h to avoid splats
> > > such as:
> > > 
> > >    In file included from include/uapi/linux/stddef.h:1:0,
> > >                     from include/linux/stddef.h:4,
> > >                     from arch/h8300/kernel/asm-offsets.c:11:
> > >    include/linux/list.h: In function 'list_empty':
> > > >> include/linux/compiler.h:343:2: error: implicit declaration of function 'smp_read_barrier_depends' [-Werror=implicit-function-declaration]
> > >      smp_read_barrier_depends(); /* Enforce dependency ordering from x */ \
> > >      ^
> > > 
> > > A better alternative is to include asm/barrier.h in linux/compiler.h,
> > > but this requires a type definition for "bool" on some architectures
> > > (e.g. x86), which is defined later by linux/types.h. Type "bool" is also
> > > used directly in linux/compiler.h, so the whole thing is pretty fragile.
> > > 
> > > This patch splits compiler.h in two: compiler-types.h contains type
> > > annotations, definitions and the compiler-specific parts, whereas
> > > compiler.h #includes compiler-types.h and additionally defines macros
> > > such as {READ,WRITE.ACCESS}_ONCE.
> > > 
> > > uapi/linux/stddef.h and linux/linkage.h are then moved over to include
> > > linux/compiler-types.h, which fixes the build for h8 and blackfin.
> > > 
> > > Reported-by: kbuild test robot <fengguang.wu@intel.com>
> > > Signed-off-by: Will Deacon <will.deacon@arm.com>
> > > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > > [ paulmck: Add Will's fixes for issues reported by kbuild test robot. ]
> > > ---
> > >  arch/arm/include/asm/ptrace.h   |   3 +-
> > >  arch/sparc/include/asm/ptrace.h |   1 +
> > >  arch/um/include/shared/init.h   |   2 +-
> > >  include/linux/compiler-clang.h  |   2 +-
> > >  include/linux/compiler-gcc.h    |   2 +-
> > >  include/linux/compiler-intel.h  |   2 +-
> > >  include/linux/compiler-types.h  | 274 ++++++++++++++++++++++++++++++++++++++++
> > >  include/linux/compiler.h        | 265 +-------------------------------------
> > 
> > The common pattern is to use an underscore (compiler_types.h), not a hyphen 
> > (compiler-types.h) ...
> 
> I just followed the existing examples of compiler-{clang,gcc,intel}.h!
> I can respin if you like, but it will be inconsistent with the existing
> compiler* headers and it's not like there aren't other hyphenated header
> files in include/linux/ (I count 433).

The pattern to follow here is the _types.h header file postfix that denotes data 
type definitions of a subsystem, and those are consistently _types.h, in 90%+ of 
the cases:

  triton:~/tip> find include/ arch/ | grep -e _types.h | wc -l
  90

  triton:~/tip> find include/ arch/ | grep -e -types.h | wc -l
  6

Thanks,

	Ingo

Re: [PATCH 16/19] linux/compiler.h: Split into compiler.h and compiler-types.h

[Ingo Molnar]

* Will Deacon <will.deacon@arm.com> wrote:

> On Tue, Oct 24, 2017 at 11:25:12AM +0200, Ingo Molnar wrote:
> > 
> > * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> > 
> > > From: Will Deacon <will.deacon@arm.com>
> > > 
> > > linux/compiler.h is included indirectly by linux/types.h via
> > > uapi/linux/types.h -> uapi/linux/posix_types.h -> linux/stddef.h
> > > -> uapi/linux/stddef.h and is needed to provide a proper definition of
> > > offsetof.
> > > 
> > > Unfortunately, compiler.h requires a definition of
> > > smp_read_barrier_depends() for defining lockless_dereference and soon
> > > for defining READ_ONCE, which means that all
> > > users of READ_ONCE will need to include asm/barrier.h to avoid splats
> > > such as:
> > > 
> > >    In file included from include/uapi/linux/stddef.h:1:0,
> > >                     from include/linux/stddef.h:4,
> > >                     from arch/h8300/kernel/asm-offsets.c:11:
> > >    include/linux/list.h: In function 'list_empty':
> > > >> include/linux/compiler.h:343:2: error: implicit declaration of function 'smp_read_barrier_depends' [-Werror=implicit-function-declaration]
> > >      smp_read_barrier_depends(); /* Enforce dependency ordering from x */ \
> > >      ^
> > > 
> > > A better alternative is to include asm/barrier.h in linux/compiler.h,
> > > but this requires a type definition for "bool" on some architectures
> > > (e.g. x86), which is defined later by linux/types.h. Type "bool" is also
> > > used directly in linux/compiler.h, so the whole thing is pretty fragile.
> > > 
> > > This patch splits compiler.h in two: compiler-types.h contains type
> > > annotations, definitions and the compiler-specific parts, whereas
> > > compiler.h #includes compiler-types.h and additionally defines macros
> > > such as {READ,WRITE.ACCESS}_ONCE.
> > > 
> > > uapi/linux/stddef.h and linux/linkage.h are then moved over to include
> > > linux/compiler-types.h, which fixes the build for h8 and blackfin.
> > > 
> > > Reported-by: kbuild test robot <fengguang.wu@intel.com>
> > > Signed-off-by: Will Deacon <will.deacon@arm.com>
> > > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> > > [ paulmck: Add Will's fixes for issues reported by kbuild test robot. ]
> > > ---
> > >  arch/arm/include/asm/ptrace.h   |   3 +-
> > >  arch/sparc/include/asm/ptrace.h |   1 +
> > >  arch/um/include/shared/init.h   |   2 +-
> > >  include/linux/compiler-clang.h  |   2 +-
> > >  include/linux/compiler-gcc.h    |   2 +-
> > >  include/linux/compiler-intel.h  |   2 +-
> > >  include/linux/compiler-types.h  | 274 ++++++++++++++++++++++++++++++++++++++++
> > >  include/linux/compiler.h        | 265 +-------------------------------------
> > 
> > The common pattern is to use an underscore (compiler_types.h), not a hyphen 
> > (compiler-types.h) ...
> 
> I just followed the existing examples of compiler-{clang,gcc,intel}.h!
> I can respin if you like, but it will be inconsistent with the existing
> compiler* headers and it's not like there aren't other hyphenated header
> files in include/linux/ (I count 433).

The pattern to follow here is the _types.h header file postfix that denotes data 
type definitions of a subsystem, and those are consistently _types.h, in 90%+ of 
the cases:

  triton:~/tip> find include/ arch/ | grep -e _types.h | wc -l
  90

  triton:~/tip> find include/ arch/ | grep -e -types.h | wc -l
  6

Thanks,

	Ingo

[PATCH 17/19] locking/barriers: Kill lockless_dereference

[Paul E. McKenney]

From: Will Deacon <will.deacon@arm.com>

lockless_dereference is a nice idea, but its gained little traction in
kernel code since it's introduction three years ago. This is partly
because it's a pain to type, but also because using READ_ONCE instead
will work correctly on all architectures apart from Alpha, which is a
fully supported but somewhat niche architecture these days.

This patch moves smp_read_barrier_depends() (a NOP on all architectures
other than Alpha) from lockless_dereference into READ_ONCE, converts
the few actual users over to READ_ONCE and then finally removes
lockless_dereference altogether.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
---
 Documentation/memory-barriers.txt                   | 12 ------------
 .../translations/ko_KR/memory-barriers.txt          | 12 ------------
 arch/x86/events/core.c                              |  2 +-
 arch/x86/include/asm/mmu_context.h                  |  4 ++--
 arch/x86/kernel/ldt.c                               |  2 +-
 drivers/md/dm-mpath.c                               | 20 ++++++++++----------
 fs/dcache.c                                         |  4 ++--
 fs/overlayfs/ovl_entry.h                            |  2 +-
 fs/overlayfs/readdir.c                              |  2 +-
 include/linux/compiler.h                            | 21 +--------------------
 include/linux/rculist.h                             |  4 ++--
 include/linux/rcupdate.h                            |  4 ++--
 kernel/events/core.c                                |  4 ++--
 kernel/seccomp.c                                    |  2 +-
 kernel/task_work.c                                  |  2 +-
 mm/slab.h                                           |  2 +-
 16 files changed, 28 insertions(+), 71 deletions(-)

diff --git a/Documentation/memory-barriers.txt b/Documentation/memory-barriers.txt
index b759a60624fd..470a682f3fa4 100644
--- a/Documentation/memory-barriers.txt
+++ b/Documentation/memory-barriers.txt
@@ -1886,18 +1886,6 @@ There are some more advanced barrier functions:
      See Documentation/atomic_{t,bitops}.txt for more information.
 
 
- (*) lockless_dereference();
-
-     This can be thought of as a pointer-fetch wrapper around the
-     smp_read_barrier_depends() data-dependency barrier.
-
-     This is also similar to rcu_dereference(), but in cases where
-     object lifetime is handled by some mechanism other than RCU, for
-     example, when the objects removed only when the system goes down.
-     In addition, lockless_dereference() is used in some data structures
-     that can be used both with and without RCU.
-
-
  (*) dma_wmb();
  (*) dma_rmb();
 
diff --git a/Documentation/translations/ko_KR/memory-barriers.txt b/Documentation/translations/ko_KR/memory-barriers.txt
index a7a813258013..ec3b46e27b7a 100644
--- a/Documentation/translations/ko_KR/memory-barriers.txt
+++ b/Documentation/translations/ko_KR/memory-barriers.txt
@@ -1858,18 +1858,6 @@ Mandatory 배리어들은 SMP 시스템에서도 UP 시스템에서도 SMP 효
      참고하세요.
 
 
- (*) lockless_dereference();
-
-     이 함수는 smp_read_barrier_depends() 데이터 의존성 배리어를 사용하는
-     포인터 읽어오기 래퍼(wrapper) 함수로 생각될 수 있습니다.
-
-     객체의 라이프타임이 RCU 외의 메커니즘으로 관리된다는 점을 제외하면
-     rcu_dereference() 와도 유사한데, 예를 들면 객체가 시스템이 꺼질 때에만
-     제거되는 경우 등입니다.  또한, lockless_dereference() 은 RCU 와 함께
-     사용될수도, RCU 없이 사용될 수도 있는 일부 데이터 구조에 사용되고
-     있습니다.
-
-
  (*) dma_wmb();
  (*) dma_rmb();
 
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c
index 80534d3c2480..589af1eec7c1 100644
--- a/arch/x86/events/core.c
+++ b/arch/x86/events/core.c
@@ -2371,7 +2371,7 @@ static unsigned long get_segment_base(unsigned int segment)
 		struct ldt_struct *ldt;
 
 		/* IRQs are off, so this synchronizes with smp_store_release */
-		ldt = lockless_dereference(current->active_mm->context.ldt);
+		ldt = READ_ONCE(current->active_mm->context.ldt);
 		if (!ldt || idx >= ldt->nr_entries)
 			return 0;
 
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index c120b5db178a..9037a4e546e8 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -72,8 +72,8 @@ static inline void load_mm_ldt(struct mm_struct *mm)
 #ifdef CONFIG_MODIFY_LDT_SYSCALL
 	struct ldt_struct *ldt;
 
-	/* lockless_dereference synchronizes with smp_store_release */
-	ldt = lockless_dereference(mm->context.ldt);
+	/* READ_ONCE synchronizes with smp_store_release */
+	ldt = READ_ONCE(mm->context.ldt);
 
 	/*
 	 * Any change to mm->context.ldt is followed by an IPI to all
diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c
index f0e64db18ac8..0a21390642c4 100644
--- a/arch/x86/kernel/ldt.c
+++ b/arch/x86/kernel/ldt.c
@@ -101,7 +101,7 @@ static void finalize_ldt_struct(struct ldt_struct *ldt)
 static void install_ldt(struct mm_struct *current_mm,
 			struct ldt_struct *ldt)
 {
-	/* Synchronizes with lockless_dereference in load_mm_ldt. */
+	/* Synchronizes with READ_ONCE in load_mm_ldt. */
 	smp_store_release(&current_mm->context.ldt, ldt);
 
 	/* Activate the LDT for all CPUs using current_mm. */
diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index 11f273d2f018..3f88c9d32f7e 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -366,7 +366,7 @@ static struct pgpath *choose_path_in_pg(struct multipath *m,
 
 	pgpath = path_to_pgpath(path);
 
-	if (unlikely(lockless_dereference(m->current_pg) != pg)) {
+	if (unlikely(READ_ONCE(m->current_pg) != pg)) {
 		/* Only update current_pgpath if pg changed */
 		spin_lock_irqsave(&m->lock, flags);
 		m->current_pgpath = pgpath;
@@ -390,7 +390,7 @@ static struct pgpath *choose_pgpath(struct multipath *m, size_t nr_bytes)
 	}
 
 	/* Were we instructed to switch PG? */
-	if (lockless_dereference(m->next_pg)) {
+	if (READ_ONCE(m->next_pg)) {
 		spin_lock_irqsave(&m->lock, flags);
 		pg = m->next_pg;
 		if (!pg) {
@@ -406,7 +406,7 @@ static struct pgpath *choose_pgpath(struct multipath *m, size_t nr_bytes)
 
 	/* Don't change PG until it has no remaining paths */
 check_current_pg:
-	pg = lockless_dereference(m->current_pg);
+	pg = READ_ONCE(m->current_pg);
 	if (pg) {
 		pgpath = choose_path_in_pg(m, pg, nr_bytes);
 		if (!IS_ERR_OR_NULL(pgpath))
@@ -473,7 +473,7 @@ static int multipath_clone_and_map(struct dm_target *ti, struct request *rq,
 	struct request *clone;
 
 	/* Do we need to select a new pgpath? */
-	pgpath = lockless_dereference(m->current_pgpath);
+	pgpath = READ_ONCE(m->current_pgpath);
 	if (!pgpath || !test_bit(MPATHF_QUEUE_IO, &m->flags))
 		pgpath = choose_pgpath(m, nr_bytes);
 
@@ -535,7 +535,7 @@ static int __multipath_map_bio(struct multipath *m, struct bio *bio, struct dm_m
 	bool queue_io;
 
 	/* Do we need to select a new pgpath? */
-	pgpath = lockless_dereference(m->current_pgpath);
+	pgpath = READ_ONCE(m->current_pgpath);
 	queue_io = test_bit(MPATHF_QUEUE_IO, &m->flags);
 	if (!pgpath || !queue_io)
 		pgpath = choose_pgpath(m, nr_bytes);
@@ -1804,7 +1804,7 @@ static int multipath_prepare_ioctl(struct dm_target *ti,
 	struct pgpath *current_pgpath;
 	int r;
 
-	current_pgpath = lockless_dereference(m->current_pgpath);
+	current_pgpath = READ_ONCE(m->current_pgpath);
 	if (!current_pgpath)
 		current_pgpath = choose_pgpath(m, 0);
 
@@ -1826,7 +1826,7 @@ static int multipath_prepare_ioctl(struct dm_target *ti,
 	}
 
 	if (r == -ENOTCONN) {
-		if (!lockless_dereference(m->current_pg)) {
+		if (!READ_ONCE(m->current_pg)) {
 			/* Path status changed, redo selection */
 			(void) choose_pgpath(m, 0);
 		}
@@ -1895,9 +1895,9 @@ static int multipath_busy(struct dm_target *ti)
 		return (m->queue_mode != DM_TYPE_MQ_REQUEST_BASED);
 
 	/* Guess which priority_group will be used at next mapping time */
-	pg = lockless_dereference(m->current_pg);
-	next_pg = lockless_dereference(m->next_pg);
-	if (unlikely(!lockless_dereference(m->current_pgpath) && next_pg))
+	pg = READ_ONCE(m->current_pg);
+	next_pg = READ_ONCE(m->next_pg);
+	if (unlikely(!READ_ONCE(m->current_pgpath) && next_pg))
 		pg = next_pg;
 
 	if (!pg) {
diff --git a/fs/dcache.c b/fs/dcache.c
index f90141387f01..34c852af215c 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -231,7 +231,7 @@ static inline int dentry_cmp(const struct dentry *dentry, const unsigned char *c
 {
 	/*
 	 * Be careful about RCU walk racing with rename:
-	 * use 'lockless_dereference' to fetch the name pointer.
+	 * use 'READ_ONCE' to fetch the name pointer.
 	 *
 	 * NOTE! Even if a rename will mean that the length
 	 * was not loaded atomically, we don't care. The
@@ -245,7 +245,7 @@ static inline int dentry_cmp(const struct dentry *dentry, const unsigned char *c
 	 * early because the data cannot match (there can
 	 * be no NUL in the ct/tcount data)
 	 */
-	const unsigned char *cs = lockless_dereference(dentry->d_name.name);
+	const unsigned char *cs = READ_ONCE(dentry->d_name.name);
 
 	return dentry_string_cmp(cs, ct, tcount);
 }
diff --git a/fs/overlayfs/ovl_entry.h b/fs/overlayfs/ovl_entry.h
index 25d9b5adcd42..36b49bd09264 100644
--- a/fs/overlayfs/ovl_entry.h
+++ b/fs/overlayfs/ovl_entry.h
@@ -77,5 +77,5 @@ static inline struct ovl_inode *OVL_I(struct inode *inode)
 
 static inline struct dentry *ovl_upperdentry_dereference(struct ovl_inode *oi)
 {
-	return lockless_dereference(oi->__upperdentry);
+	return READ_ONCE(oi->__upperdentry);
 }
diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
index 0f85ee9c3268..c67a7703296b 100644
--- a/fs/overlayfs/readdir.c
+++ b/fs/overlayfs/readdir.c
@@ -754,7 +754,7 @@ static int ovl_dir_fsync(struct file *file, loff_t start, loff_t end,
 	if (!od->is_upper && OVL_TYPE_UPPER(ovl_path_type(dentry))) {
 		struct inode *inode = file_inode(file);
 
-		realfile = lockless_dereference(od->upperfile);
+		realfile = READ_ONCE(od->upperfile);
 		if (!realfile) {
 			struct path upperpath;
 
diff --git a/include/linux/compiler.h b/include/linux/compiler.h
index c714239f453f..4517662f128c 100644
--- a/include/linux/compiler.h
+++ b/include/linux/compiler.h
@@ -242,6 +242,7 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
 		__read_once_size(&(x), __u.__c, sizeof(x));		\
 	else								\
 		__read_once_size_nocheck(&(x), __u.__c, sizeof(x));	\
+	smp_read_barrier_depends(); /* Enforce dependency ordering from x */ \
 	__u.__val;							\
 })
 #define READ_ONCE(x) __READ_ONCE(x, 1)
@@ -345,24 +346,4 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
 	(volatile typeof(x) *)&(x); })
 #define ACCESS_ONCE(x) (*__ACCESS_ONCE(x))
 
-/**
- * lockless_dereference() - safely load a pointer for later dereference
- * @p: The pointer to load
- *
- * Similar to rcu_dereference(), but for situations where the pointed-to
- * object's lifetime is managed by something other than RCU.  That
- * "something other" might be reference counting or simple immortality.
- *
- * The seemingly unused variable ___typecheck_p validates that @p is
- * indeed a pointer type by using a pointer to typeof(*p) as the type.
- * Taking a pointer to typeof(*p) again is needed in case p is void *.
- */
-#define lockless_dereference(p) \
-({ \
-	typeof(p) _________p1 = READ_ONCE(p); \
-	typeof(*(p)) *___typecheck_p __maybe_unused; \
-	smp_read_barrier_depends(); /* Dependency order vs. p above. */ \
-	(_________p1); \
-})
-
 #endif /* __LINUX_COMPILER_H */
diff --git a/include/linux/rculist.h b/include/linux/rculist.h
index b1fd8bf85fdc..3a2bb7d8ed4d 100644
--- a/include/linux/rculist.h
+++ b/include/linux/rculist.h
@@ -274,7 +274,7 @@ static inline void list_splice_tail_init_rcu(struct list_head *list,
  * primitives such as list_add_rcu() as long as it's guarded by rcu_read_lock().
  */
 #define list_entry_rcu(ptr, type, member) \
-	container_of(lockless_dereference(ptr), type, member)
+	container_of(READ_ONCE(ptr), type, member)
 
 /**
  * Where are list_empty_rcu() and list_first_entry_rcu()?
@@ -367,7 +367,7 @@ static inline void list_splice_tail_init_rcu(struct list_head *list,
  * example is when items are added to the list, but never deleted.
  */
 #define list_entry_lockless(ptr, type, member) \
-	container_of((typeof(ptr))lockless_dereference(ptr), type, member)
+	container_of((typeof(ptr))READ_ONCE(ptr), type, member)
 
 /**
  * list_for_each_entry_lockless - iterate over rcu list of given type
diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
index de50d8a4cf41..380a3aeb09d7 100644
--- a/include/linux/rcupdate.h
+++ b/include/linux/rcupdate.h
@@ -346,7 +346,7 @@ static inline void rcu_preempt_sleep_check(void) { }
 #define __rcu_dereference_check(p, c, space) \
 ({ \
 	/* Dependency order vs. p above. */ \
-	typeof(*p) *________p1 = (typeof(*p) *__force)lockless_dereference(p); \
+	typeof(*p) *________p1 = (typeof(*p) *__force)READ_ONCE(p); \
 	RCU_LOCKDEP_WARN(!(c), "suspicious rcu_dereference_check() usage"); \
 	rcu_dereference_sparse(p, space); \
 	((typeof(*p) __force __kernel *)(________p1)); \
@@ -360,7 +360,7 @@ static inline void rcu_preempt_sleep_check(void) { }
 #define rcu_dereference_raw(p) \
 ({ \
 	/* Dependency order vs. p above. */ \
-	typeof(p) ________p1 = lockless_dereference(p); \
+	typeof(p) ________p1 = READ_ONCE(p); \
 	((typeof(*p) __force __kernel *)(________p1)); \
 })
 
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 6bc21e202ae4..417812ce0099 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -4231,7 +4231,7 @@ static void perf_remove_from_owner(struct perf_event *event)
 	 * indeed free this event, otherwise we need to serialize on
 	 * owner->perf_event_mutex.
 	 */
-	owner = lockless_dereference(event->owner);
+	owner = READ_ONCE(event->owner);
 	if (owner) {
 		/*
 		 * Since delayed_put_task_struct() also drops the last
@@ -4328,7 +4328,7 @@ int perf_event_release_kernel(struct perf_event *event)
 		 * Cannot change, child events are not migrated, see the
 		 * comment with perf_event_ctx_lock_nested().
 		 */
-		ctx = lockless_dereference(child->ctx);
+		ctx = READ_ONCE(child->ctx);
 		/*
 		 * Since child_mutex nests inside ctx::mutex, we must jump
 		 * through hoops. We start by grabbing a reference on the ctx.
diff --git a/kernel/seccomp.c b/kernel/seccomp.c
index bb3a38005b9c..1daa8b61a268 100644
--- a/kernel/seccomp.c
+++ b/kernel/seccomp.c
@@ -189,7 +189,7 @@ static u32 seccomp_run_filters(const struct seccomp_data *sd,
 	u32 ret = SECCOMP_RET_ALLOW;
 	/* Make sure cross-thread synced filter points somewhere sane. */
 	struct seccomp_filter *f =
-			lockless_dereference(current->seccomp.filter);
+			READ_ONCE(current->seccomp.filter);
 
 	/* Ensure unexpected behavior doesn't result in failing open. */
 	if (unlikely(WARN_ON(f == NULL)))
diff --git a/kernel/task_work.c b/kernel/task_work.c
index 836a72a66fba..9a9f262fc53d 100644
--- a/kernel/task_work.c
+++ b/kernel/task_work.c
@@ -67,7 +67,7 @@ task_work_cancel(struct task_struct *task, task_work_func_t func)
 	 * we raced with task_work_run(), *pprev == NULL/exited.
 	 */
 	raw_spin_lock_irqsave(&task->pi_lock, flags);
-	while ((work = lockless_dereference(*pprev))) {
+	while ((work = READ_ONCE(*pprev))) {
 		if (work->func != func)
 			pprev = &work->next;
 		else if (cmpxchg(pprev, work, work->next) == work)
diff --git a/mm/slab.h b/mm/slab.h
index 073362816acc..8894f811a89d 100644
--- a/mm/slab.h
+++ b/mm/slab.h
@@ -258,7 +258,7 @@ cache_from_memcg_idx(struct kmem_cache *s, int idx)
 	 * memcg_caches issues a write barrier to match this (see
 	 * memcg_create_kmem_cache()).
 	 */
-	cachep = lockless_dereference(arr->entries[idx]);
+	cachep = READ_ONCE(arr->entries[idx]);
 	rcu_read_unlock();
 
 	return cachep;
-- 
2.5.2

[PATCH 17/19] locking/barriers: Kill lockless_dereference

[Paul E. McKenney]

From: Will Deacon <will.deacon@arm.com>

lockless_dereference is a nice idea, but its gained little traction in
kernel code since it's introduction three years ago. This is partly
because it's a pain to type, but also because using READ_ONCE instead
will work correctly on all architectures apart from Alpha, which is a
fully supported but somewhat niche architecture these days.

This patch moves smp_read_barrier_depends() (a NOP on all architectures
other than Alpha) from lockless_dereference into READ_ONCE, converts
the few actual users over to READ_ONCE and then finally removes
lockless_dereference altogether.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
---
 Documentation/memory-barriers.txt                   | 12 ------------
 .../translations/ko_KR/memory-barriers.txt          | 12 ------------
 arch/x86/events/core.c                              |  2 +-
 arch/x86/include/asm/mmu_context.h                  |  4 ++--
 arch/x86/kernel/ldt.c                               |  2 +-
 drivers/md/dm-mpath.c                               | 20 ++++++++++----------
 fs/dcache.c                                         |  4 ++--
 fs/overlayfs/ovl_entry.h                            |  2 +-
 fs/overlayfs/readdir.c                              |  2 +-
 include/linux/compiler.h                            | 21 +--------------------
 include/linux/rculist.h                             |  4 ++--
 include/linux/rcupdate.h                            |  4 ++--
 kernel/events/core.c                                |  4 ++--
 kernel/seccomp.c                                    |  2 +-
 kernel/task_work.c                                  |  2 +-
 mm/slab.h                                           |  2 +-
 16 files changed, 28 insertions(+), 71 deletions(-)

diff --git a/Documentation/memory-barriers.txt b/Documentation/memory-barriers.txt
index b759a60624fd..470a682f3fa4 100644
--- a/Documentation/memory-barriers.txt
+++ b/Documentation/memory-barriers.txt
@@ -1886,18 +1886,6 @@ There are some more advanced barrier functions:
      See Documentation/atomic_{t,bitops}.txt for more information.
 
 
- (*) lockless_dereference();
-
-     This can be thought of as a pointer-fetch wrapper around the
-     smp_read_barrier_depends() data-dependency barrier.
-
-     This is also similar to rcu_dereference(), but in cases where
-     object lifetime is handled by some mechanism other than RCU, for
-     example, when the objects removed only when the system goes down.
-     In addition, lockless_dereference() is used in some data structures
-     that can be used both with and without RCU.
-
-
  (*) dma_wmb();
  (*) dma_rmb();
 
diff --git a/Documentation/translations/ko_KR/memory-barriers.txt b/Documentation/translations/ko_KR/memory-barriers.txt
index a7a813258013..ec3b46e27b7a 100644
--- a/Documentation/translations/ko_KR/memory-barriers.txt
+++ b/Documentation/translations/ko_KR/memory-barriers.txt
@@ -1858,18 +1858,6 @@ Mandatory 배리어들은 SMP 시스템에서도 UP 시스템에서도 SMP 효
      참고하세요.
 
 
- (*) lockless_dereference();
-
-     이 함수는 smp_read_barrier_depends() 데이터 의존성 배리어를 사용하는
-     포인터 읽어오기 래퍼(wrapper) 함수로 생각될 수 있습니다.
-
-     객체의 라이프타임이 RCU 외의 메커니즘으로 관리된다는 점을 제외하면
-     rcu_dereference() 와도 유사한데, 예를 들면 객체가 시스템이 꺼질 때에만
-     제거되는 경우 등입니다.  또한, lockless_dereference() 은 RCU 와 함께
-     사용될수도, RCU 없이 사용될 수도 있는 일부 데이터 구조에 사용되고
-     있습니다.
-
-
  (*) dma_wmb();
  (*) dma_rmb();
 
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c
index 80534d3c2480..589af1eec7c1 100644
--- a/arch/x86/events/core.c
+++ b/arch/x86/events/core.c
@@ -2371,7 +2371,7 @@ static unsigned long get_segment_base(unsigned int segment)
 		struct ldt_struct *ldt;
 
 		/* IRQs are off, so this synchronizes with smp_store_release */
-		ldt = lockless_dereference(current->active_mm->context.ldt);
+		ldt = READ_ONCE(current->active_mm->context.ldt);
 		if (!ldt || idx >= ldt->nr_entries)
 			return 0;
 
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index c120b5db178a..9037a4e546e8 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -72,8 +72,8 @@ static inline void load_mm_ldt(struct mm_struct *mm)
 #ifdef CONFIG_MODIFY_LDT_SYSCALL
 	struct ldt_struct *ldt;
 
-	/* lockless_dereference synchronizes with smp_store_release */
-	ldt = lockless_dereference(mm->context.ldt);
+	/* READ_ONCE synchronizes with smp_store_release */
+	ldt = READ_ONCE(mm->context.ldt);
 
 	/*
 	 * Any change to mm->context.ldt is followed by an IPI to all
diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c
index f0e64db18ac8..0a21390642c4 100644
--- a/arch/x86/kernel/ldt.c
+++ b/arch/x86/kernel/ldt.c
@@ -101,7 +101,7 @@ static void finalize_ldt_struct(struct ldt_struct *ldt)
 static void install_ldt(struct mm_struct *current_mm,
 			struct ldt_struct *ldt)
 {
-	/* Synchronizes with lockless_dereference in load_mm_ldt. */
+	/* Synchronizes with READ_ONCE in load_mm_ldt. */
 	smp_store_release(&current_mm->context.ldt, ldt);
 
 	/* Activate the LDT for all CPUs using current_mm. */
diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index 11f273d2f018..3f88c9d32f7e 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -366,7 +366,7 @@ static struct pgpath *choose_path_in_pg(struct multipath *m,
 
 	pgpath = path_to_pgpath(path);
 
-	if (unlikely(lockless_dereference(m->current_pg) != pg)) {
+	if (unlikely(READ_ONCE(m->current_pg) != pg)) {
 		/* Only update current_pgpath if pg changed */
 		spin_lock_irqsave(&m->lock, flags);
 		m->current_pgpath = pgpath;
@@ -390,7 +390,7 @@ static struct pgpath *choose_pgpath(struct multipath *m, size_t nr_bytes)
 	}
 
 	/* Were we instructed to switch PG? */
-	if (lockless_dereference(m->next_pg)) {
+	if (READ_ONCE(m->next_pg)) {
 		spin_lock_irqsave(&m->lock, flags);
 		pg = m->next_pg;
 		if (!pg) {
@@ -406,7 +406,7 @@ static struct pgpath *choose_pgpath(struct multipath *m, size_t nr_bytes)
 
 	/* Don't change PG until it has no remaining paths */
 check_current_pg:
-	pg = lockless_dereference(m->current_pg);
+	pg = READ_ONCE(m->current_pg);
 	if (pg) {
 		pgpath = choose_path_in_pg(m, pg, nr_bytes);
 		if (!IS_ERR_OR_NULL(pgpath))
@@ -473,7 +473,7 @@ static int multipath_clone_and_map(struct dm_target *ti, struct request *rq,
 	struct request *clone;
 
 	/* Do we need to select a new pgpath? */
-	pgpath = lockless_dereference(m->current_pgpath);
+	pgpath = READ_ONCE(m->current_pgpath);
 	if (!pgpath || !test_bit(MPATHF_QUEUE_IO, &m->flags))
 		pgpath = choose_pgpath(m, nr_bytes);
 
@@ -535,7 +535,7 @@ static int __multipath_map_bio(struct multipath *m, struct bio *bio, struct dm_m
 	bool queue_io;
 
 	/* Do we need to select a new pgpath? */
-	pgpath = lockless_dereference(m->current_pgpath);
+	pgpath = READ_ONCE(m->current_pgpath);
 	queue_io = test_bit(MPATHF_QUEUE_IO, &m->flags);
 	if (!pgpath || !queue_io)
 		pgpath = choose_pgpath(m, nr_bytes);
@@ -1804,7 +1804,7 @@ static int multipath_prepare_ioctl(struct dm_target *ti,
 	struct pgpath *current_pgpath;
 	int r;
 
-	current_pgpath = lockless_dereference(m->current_pgpath);
+	current_pgpath = READ_ONCE(m->current_pgpath);
 	if (!current_pgpath)
 		current_pgpath = choose_pgpath(m, 0);
 
@@ -1826,7 +1826,7 @@ static int multipath_prepare_ioctl(struct dm_target *ti,
 	}
 
 	if (r == -ENOTCONN) {
-		if (!lockless_dereference(m->current_pg)) {
+		if (!READ_ONCE(m->current_pg)) {
 			/* Path status changed, redo selection */
 			(void) choose_pgpath(m, 0);
 		}
@@ -1895,9 +1895,9 @@ static int multipath_busy(struct dm_target *ti)
 		return (m->queue_mode != DM_TYPE_MQ_REQUEST_BASED);
 
 	/* Guess which priority_group will be used at next mapping time */
-	pg = lockless_dereference(m->current_pg);
-	next_pg = lockless_dereference(m->next_pg);
-	if (unlikely(!lockless_dereference(m->current_pgpath) && next_pg))
+	pg = READ_ONCE(m->current_pg);
+	next_pg = READ_ONCE(m->next_pg);
+	if (unlikely(!READ_ONCE(m->current_pgpath) && next_pg))
 		pg = next_pg;
 
 	if (!pg) {
diff --git a/fs/dcache.c b/fs/dcache.c
index f90141387f01..34c852af215c 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -231,7 +231,7 @@ static inline int dentry_cmp(const struct dentry *dentry, const unsigned char *c
 {
 	/*
 	 * Be careful about RCU walk racing with rename:
-	 * use 'lockless_dereference' to fetch the name pointer.
+	 * use 'READ_ONCE' to fetch the name pointer.
 	 *
 	 * NOTE! Even if a rename will mean that the length
 	 * was not loaded atomically, we don't care. The
@@ -245,7 +245,7 @@ static inline int dentry_cmp(const struct dentry *dentry, const unsigned char *c
 	 * early because the data cannot match (there can
 	 * be no NUL in the ct/tcount data)
 	 */
-	const unsigned char *cs = lockless_dereference(dentry->d_name.name);
+	const unsigned char *cs = READ_ONCE(dentry->d_name.name);
 
 	return dentry_string_cmp(cs, ct, tcount);
 }
diff --git a/fs/overlayfs/ovl_entry.h b/fs/overlayfs/ovl_entry.h
index 25d9b5adcd42..36b49bd09264 100644
--- a/fs/overlayfs/ovl_entry.h
+++ b/fs/overlayfs/ovl_entry.h
@@ -77,5 +77,5 @@ static inline struct ovl_inode *OVL_I(struct inode *inode)
 
 static inline struct dentry *ovl_upperdentry_dereference(struct ovl_inode *oi)
 {
-	return lockless_dereference(oi->__upperdentry);
+	return READ_ONCE(oi->__upperdentry);
 }
diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
index 0f85ee9c3268..c67a7703296b 100644
--- a/fs/overlayfs/readdir.c
+++ b/fs/overlayfs/readdir.c
@@ -754,7 +754,7 @@ static int ovl_dir_fsync(struct file *file, loff_t start, loff_t end,
 	if (!od->is_upper && OVL_TYPE_UPPER(ovl_path_type(dentry))) {
 		struct inode *inode = file_inode(file);
 
-		realfile = lockless_dereference(od->upperfile);
+		realfile = READ_ONCE(od->upperfile);
 		if (!realfile) {
 			struct path upperpath;
 
diff --git a/include/linux/compiler.h b/include/linux/compiler.h
index c714239f453f..4517662f128c 100644
--- a/include/linux/compiler.h
+++ b/include/linux/compiler.h
@@ -242,6 +242,7 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
 		__read_once_size(&(x), __u.__c, sizeof(x));		\
 	else								\
 		__read_once_size_nocheck(&(x), __u.__c, sizeof(x));	\
+	smp_read_barrier_depends(); /* Enforce dependency ordering from x */ \
 	__u.__val;							\
 })
 #define READ_ONCE(x) __READ_ONCE(x, 1)
@@ -345,24 +346,4 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
 	(volatile typeof(x) *)&(x); })
 #define ACCESS_ONCE(x) (*__ACCESS_ONCE(x))
 
-/**
- * lockless_dereference() - safely load a pointer for later dereference
- * @p: The pointer to load
- *
- * Similar to rcu_dereference(), but for situations where the pointed-to
- * object's lifetime is managed by something other than RCU.  That
- * "something other" might be reference counting or simple immortality.
- *
- * The seemingly unused variable ___typecheck_p validates that @p is
- * indeed a pointer type by using a pointer to typeof(*p) as the type.
- * Taking a pointer to typeof(*p) again is needed in case p is void *.
- */
-#define lockless_dereference(p) \
-({ \
-	typeof(p) _________p1 = READ_ONCE(p); \
-	typeof(*(p)) *___typecheck_p __maybe_unused; \
-	smp_read_barrier_depends(); /* Dependency order vs. p above. */ \
-	(_________p1); \
-})
-
 #endif /* __LINUX_COMPILER_H */
diff --git a/include/linux/rculist.h b/include/linux/rculist.h
index b1fd8bf85fdc..3a2bb7d8ed4d 100644
--- a/include/linux/rculist.h
+++ b/include/linux/rculist.h
@@ -274,7 +274,7 @@ static inline void list_splice_tail_init_rcu(struct list_head *list,
  * primitives such as list_add_rcu() as long as it's guarded by rcu_read_lock().
  */
 #define list_entry_rcu(ptr, type, member) \
-	container_of(lockless_dereference(ptr), type, member)
+	container_of(READ_ONCE(ptr), type, member)
 
 /**
  * Where are list_empty_rcu() and list_first_entry_rcu()?
@@ -367,7 +367,7 @@ static inline void list_splice_tail_init_rcu(struct list_head *list,
  * example is when items are added to the list, but never deleted.
  */
 #define list_entry_lockless(ptr, type, member) \
-	container_of((typeof(ptr))lockless_dereference(ptr), type, member)
+	container_of((typeof(ptr))READ_ONCE(ptr), type, member)
 
 /**
  * list_for_each_entry_lockless - iterate over rcu list of given type
diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
index de50d8a4cf41..380a3aeb09d7 100644
--- a/include/linux/rcupdate.h
+++ b/include/linux/rcupdate.h
@@ -346,7 +346,7 @@ static inline void rcu_preempt_sleep_check(void) { }
 #define __rcu_dereference_check(p, c, space) \
 ({ \
 	/* Dependency order vs. p above. */ \
-	typeof(*p) *________p1 = (typeof(*p) *__force)lockless_dereference(p); \
+	typeof(*p) *________p1 = (typeof(*p) *__force)READ_ONCE(p); \
 	RCU_LOCKDEP_WARN(!(c), "suspicious rcu_dereference_check() usage"); \
 	rcu_dereference_sparse(p, space); \
 	((typeof(*p) __force __kernel *)(________p1)); \
@@ -360,7 +360,7 @@ static inline void rcu_preempt_sleep_check(void) { }
 #define rcu_dereference_raw(p) \
 ({ \
 	/* Dependency order vs. p above. */ \
-	typeof(p) ________p1 = lockless_dereference(p); \
+	typeof(p) ________p1 = READ_ONCE(p); \
 	((typeof(*p) __force __kernel *)(________p1)); \
 })
 
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 6bc21e202ae4..417812ce0099 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -4231,7 +4231,7 @@ static void perf_remove_from_owner(struct perf_event *event)
 	 * indeed free this event, otherwise we need to serialize on
 	 * owner->perf_event_mutex.
 	 */
-	owner = lockless_dereference(event->owner);
+	owner = READ_ONCE(event->owner);
 	if (owner) {
 		/*
 		 * Since delayed_put_task_struct() also drops the last
@@ -4328,7 +4328,7 @@ int perf_event_release_kernel(struct perf_event *event)
 		 * Cannot change, child events are not migrated, see the
 		 * comment with perf_event_ctx_lock_nested().
 		 */
-		ctx = lockless_dereference(child->ctx);
+		ctx = READ_ONCE(child->ctx);
 		/*
 		 * Since child_mutex nests inside ctx::mutex, we must jump
 		 * through hoops. We start by grabbing a reference on the ctx.
diff --git a/kernel/seccomp.c b/kernel/seccomp.c
index bb3a38005b9c..1daa8b61a268 100644
--- a/kernel/seccomp.c
+++ b/kernel/seccomp.c
@@ -189,7 +189,7 @@ static u32 seccomp_run_filters(const struct seccomp_data *sd,
 	u32 ret = SECCOMP_RET_ALLOW;
 	/* Make sure cross-thread synced filter points somewhere sane. */
 	struct seccomp_filter *f =
-			lockless_dereference(current->seccomp.filter);
+			READ_ONCE(current->seccomp.filter);
 
 	/* Ensure unexpected behavior doesn't result in failing open. */
 	if (unlikely(WARN_ON(f == NULL)))
diff --git a/kernel/task_work.c b/kernel/task_work.c
index 836a72a66fba..9a9f262fc53d 100644
--- a/kernel/task_work.c
+++ b/kernel/task_work.c
@@ -67,7 +67,7 @@ task_work_cancel(struct task_struct *task, task_work_func_t func)
 	 * we raced with task_work_run(), *pprev == NULL/exited.
 	 */
 	raw_spin_lock_irqsave(&task->pi_lock, flags);
-	while ((work = lockless_dereference(*pprev))) {
+	while ((work = READ_ONCE(*pprev))) {
 		if (work->func != func)
 			pprev = &work->next;
 		else if (cmpxchg(pprev, work, work->next) == work)
diff --git a/mm/slab.h b/mm/slab.h
index 073362816acc..8894f811a89d 100644
--- a/mm/slab.h
+++ b/mm/slab.h
@@ -258,7 +258,7 @@ cache_from_memcg_idx(struct kmem_cache *s, int idx)
 	 * memcg_caches issues a write barrier to match this (see
 	 * memcg_create_kmem_cache()).
 	 */
-	cachep = lockless_dereference(arr->entries[idx]);
+	cachep = READ_ONCE(arr->entries[idx]);
 	rcu_read_unlock();
 
 	return cachep;
-- 
2.5.2

Re: [PATCH 17/19] locking/barriers: Kill lockless_dereference

[Ingo Molnar]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> From: Will Deacon <will.deacon@arm.com>
> 
> lockless_dereference is a nice idea, but its gained little traction in
> kernel code since it's introduction three years ago. This is partly

s/its/it
s/it's/its

> because it's a pain to type, but also because using READ_ONCE instead
> will work correctly on all architectures apart from Alpha, which is a
> fully supported but somewhat niche architecture these days.
> 
> This patch moves smp_read_barrier_depends() (a NOP on all architectures
> other than Alpha) from lockless_dereference into READ_ONCE, converts
> the few actual users over to READ_ONCE and then finally removes
> lockless_dereference altogether.

Nit: if we refer to smp_read_barrier_depends() with parentheses (which is the nice 
thing to do for function-alike symbols), then we should do the same with 
READ_ONCE() and lockless_dereference() as well.

Also, could we please split this into three patches:

 #1: Add smp_read_barrier_depends() to READ_ONCE()
 #2: Convert all lockless_dereference() users to READ_ONCE()
 #3: Remove the now unused lockless_dereference() API

to make it easier to analyze if bisected to, should any problems arise?

Thanks,

	Ingo

Re: [PATCH 17/19] locking/barriers: Kill lockless_dereference

[Ingo Molnar]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> From: Will Deacon <will.deacon@arm.com>
> 
> lockless_dereference is a nice idea, but its gained little traction in
> kernel code since it's introduction three years ago. This is partly

s/its/it
s/it's/its

> because it's a pain to type, but also because using READ_ONCE instead
> will work correctly on all architectures apart from Alpha, which is a
> fully supported but somewhat niche architecture these days.
> 
> This patch moves smp_read_barrier_depends() (a NOP on all architectures
> other than Alpha) from lockless_dereference into READ_ONCE, converts
> the few actual users over to READ_ONCE and then finally removes
> lockless_dereference altogether.

Nit: if we refer to smp_read_barrier_depends() with parentheses (which is the nice 
thing to do for function-alike symbols), then we should do the same with 
READ_ONCE() and lockless_dereference() as well.

Also, could we please split this into three patches:

 #1: Add smp_read_barrier_depends() to READ_ONCE()
 #2: Convert all lockless_dereference() users to READ_ONCE()
 #3: Remove the now unused lockless_dereference() API

to make it easier to analyze if bisected to, should any problems arise?

Thanks,

	Ingo

Re: [PATCH 17/19] locking/barriers: Kill lockless_dereference

[Will Deacon]

On Tue, Oct 24, 2017 at 11:31:04AM +0200, Ingo Molnar wrote:
> 
> * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> 
> > From: Will Deacon <will.deacon@arm.com>
> > 
> > lockless_dereference is a nice idea, but its gained little traction in
> > kernel code since it's introduction three years ago. This is partly
> 
> s/its/it
> s/it's/its

Crikey, no idea what happened there!

> > because it's a pain to type, but also because using READ_ONCE instead
> > will work correctly on all architectures apart from Alpha, which is a
> > fully supported but somewhat niche architecture these days.
> > 
> > This patch moves smp_read_barrier_depends() (a NOP on all architectures
> > other than Alpha) from lockless_dereference into READ_ONCE, converts
> > the few actual users over to READ_ONCE and then finally removes
> > lockless_dereference altogether.
> 
> Nit: if we refer to smp_read_barrier_depends() with parentheses (which is the nice 
> thing to do for function-alike symbols), then we should do the same with 
> READ_ONCE() and lockless_dereference() as well.
> 
> Also, could we please split this into three patches:
> 
>  #1: Add smp_read_barrier_depends() to READ_ONCE()
>  #2: Convert all lockless_dereference() users to READ_ONCE()
>  #3: Remove the now unused lockless_dereference() API
> 
> to make it easier to analyze if bisected to, should any problems arise?

Sure, I'll do that now.

Will

Re: [PATCH 17/19] locking/barriers: Kill lockless_dereference

[Will Deacon]

On Tue, Oct 24, 2017 at 11:31:04AM +0200, Ingo Molnar wrote:
> 
> * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> 
> > From: Will Deacon <will.deacon@arm.com>
> > 
> > lockless_dereference is a nice idea, but its gained little traction in
> > kernel code since it's introduction three years ago. This is partly
> 
> s/its/it
> s/it's/its

Crikey, no idea what happened there!

> > because it's a pain to type, but also because using READ_ONCE instead
> > will work correctly on all architectures apart from Alpha, which is a
> > fully supported but somewhat niche architecture these days.
> > 
> > This patch moves smp_read_barrier_depends() (a NOP on all architectures
> > other than Alpha) from lockless_dereference into READ_ONCE, converts
> > the few actual users over to READ_ONCE and then finally removes
> > lockless_dereference altogether.
> 
> Nit: if we refer to smp_read_barrier_depends() with parentheses (which is the nice 
> thing to do for function-alike symbols), then we should do the same with 
> READ_ONCE() and lockless_dereference() as well.
> 
> Also, could we please split this into three patches:
> 
>  #1: Add smp_read_barrier_depends() to READ_ONCE()
>  #2: Convert all lockless_dereference() users to READ_ONCE()
>  #3: Remove the now unused lockless_dereference() API
> 
> to make it easier to analyze if bisected to, should any problems arise?

Sure, I'll do that now.

Will

Re: [PATCH 17/19] locking/barriers: Kill lockless_dereference

[Ingo Molnar]

* Will Deacon <will.deacon@arm.com> wrote:

> On Tue, Oct 24, 2017 at 11:31:04AM +0200, Ingo Molnar wrote:
> > 
> > * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> > 
> > > From: Will Deacon <will.deacon@arm.com>
> > > 
> > > lockless_dereference is a nice idea, but its gained little traction in
> > > kernel code since it's introduction three years ago. This is partly
> > 
> > s/its/it
> > s/it's/its
> 
> Crikey, no idea what happened there!

NP, just habitual drive-by typo fixing from me, would have fixed them silently 
when applying the patches, had it not been for the split-up suggestion :-)

> > > because it's a pain to type, but also because using READ_ONCE instead
> > > will work correctly on all architectures apart from Alpha, which is a
> > > fully supported but somewhat niche architecture these days.
> > > 
> > > This patch moves smp_read_barrier_depends() (a NOP on all architectures
> > > other than Alpha) from lockless_dereference into READ_ONCE, converts
> > > the few actual users over to READ_ONCE and then finally removes
> > > lockless_dereference altogether.
> > 
> > Nit: if we refer to smp_read_barrier_depends() with parentheses (which is the nice 
> > thing to do for function-alike symbols), then we should do the same with 
> > READ_ONCE() and lockless_dereference() as well.
> > 
> > Also, could we please split this into three patches:
> > 
> >  #1: Add smp_read_barrier_depends() to READ_ONCE()
> >  #2: Convert all lockless_dereference() users to READ_ONCE()
> >  #3: Remove the now unused lockless_dereference() API
> > 
> > to make it easier to analyze if bisected to, should any problems arise?
> 
> Sure, I'll do that now.

Thanks!

	Ingo

Re: [PATCH 17/19] locking/barriers: Kill lockless_dereference

[Ingo Molnar]

* Will Deacon <will.deacon@arm.com> wrote:

> On Tue, Oct 24, 2017 at 11:31:04AM +0200, Ingo Molnar wrote:
> > 
> > * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> > 
> > > From: Will Deacon <will.deacon@arm.com>
> > > 
> > > lockless_dereference is a nice idea, but its gained little traction in
> > > kernel code since it's introduction three years ago. This is partly
> > 
> > s/its/it
> > s/it's/its
> 
> Crikey, no idea what happened there!

NP, just habitual drive-by typo fixing from me, would have fixed them silently 
when applying the patches, had it not been for the split-up suggestion :-)

> > > because it's a pain to type, but also because using READ_ONCE instead
> > > will work correctly on all architectures apart from Alpha, which is a
> > > fully supported but somewhat niche architecture these days.
> > > 
> > > This patch moves smp_read_barrier_depends() (a NOP on all architectures
> > > other than Alpha) from lockless_dereference into READ_ONCE, converts
> > > the few actual users over to READ_ONCE and then finally removes
> > > lockless_dereference altogether.
> > 
> > Nit: if we refer to smp_read_barrier_depends() with parentheses (which is the nice 
> > thing to do for function-alike symbols), then we should do the same with 
> > READ_ONCE() and lockless_dereference() as well.
> > 
> > Also, could we please split this into three patches:
> > 
> >  #1: Add smp_read_barrier_depends() to READ_ONCE()
> >  #2: Convert all lockless_dereference() users to READ_ONCE()
> >  #3: Remove the now unused lockless_dereference() API
> > 
> > to make it easier to analyze if bisected to, should any problems arise?
> 
> Sure, I'll do that now.

Thanks!

	Ingo

Re: [PATCH 17/19] locking/barriers: Kill lockless_dereference

[Will Deacon]

On Tue, Oct 24, 2017 at 11:54:28AM +0200, Ingo Molnar wrote:
> * Will Deacon <will.deacon@arm.com> wrote:
> > On Tue, Oct 24, 2017 at 11:31:04AM +0200, Ingo Molnar wrote:
> > > Also, could we please split this into three patches:
> > > 
> > >  #1: Add smp_read_barrier_depends() to READ_ONCE()
> > >  #2: Convert all lockless_dereference() users to READ_ONCE()
> > >  #3: Remove the now unused lockless_dereference() API
> > > 
> > > to make it easier to analyze if bisected to, should any problems arise?
> > 
> > Sure, I'll do that now.
> 
> Thanks!

https://lkml.org/lkml/2017/10/24/323

Will

[PATCH 18/19] alpha: atomics: Add smp_read_barrier_depends() to release/relaxed atomics

[Paul E. McKenney]

From: Will Deacon <will.deacon@arm.com>

As part of the fight against smp_read_barrier_depends(), we require
dependency ordering to be preserved when a dependency is headed by a load
performed using an atomic operation.

This patch adds smp_read_barrier_depends() to the _release and _relaxed
atomics on alpha, which otherwise lack anything to enforce dependency
ordering.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
---
 arch/alpha/include/asm/atomic.h | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/arch/alpha/include/asm/atomic.h b/arch/alpha/include/asm/atomic.h
index 498933a7df97..16961a3f45ba 100644
--- a/arch/alpha/include/asm/atomic.h
+++ b/arch/alpha/include/asm/atomic.h
@@ -13,6 +13,15 @@
  * than regular operations.
  */
 
+/*
+ * To ensure dependency ordering is preserved for the _relaxed and
+ * _release atomics, an smp_read_barrier_depends() is unconditionally
+ * inserted into the _relaxed variants, which are used to build the
+ * barriered versions. To avoid redundant back-to-back fences, we can
+ * define the _acquire and _fence versions explicitly.
+ */
+#define __atomic_op_acquire(op, args...)	op##_relaxed(args)
+#define __atomic_op_fence			__atomic_op_release
 
 #define ATOMIC_INIT(i)		{ (i) }
 #define ATOMIC64_INIT(i)	{ (i) }
@@ -60,6 +69,7 @@ static inline int atomic_##op##_return_relaxed(int i, atomic_t *v)	\
 	".previous"							\
 	:"=&r" (temp), "=m" (v->counter), "=&r" (result)		\
 	:"Ir" (i), "m" (v->counter) : "memory");			\
+	smp_read_barrier_depends();					\
 	return result;							\
 }
 
@@ -77,6 +87,7 @@ static inline int atomic_fetch_##op##_relaxed(int i, atomic_t *v)	\
 	".previous"							\
 	:"=&r" (temp), "=m" (v->counter), "=&r" (result)		\
 	:"Ir" (i), "m" (v->counter) : "memory");			\
+	smp_read_barrier_depends();					\
 	return result;							\
 }
 
@@ -111,6 +122,7 @@ static __inline__ long atomic64_##op##_return_relaxed(long i, atomic64_t * v)	\
 	".previous"							\
 	:"=&r" (temp), "=m" (v->counter), "=&r" (result)		\
 	:"Ir" (i), "m" (v->counter) : "memory");			\
+	smp_read_barrier_depends();					\
 	return result;							\
 }
 
@@ -128,6 +140,7 @@ static __inline__ long atomic64_fetch_##op##_relaxed(long i, atomic64_t * v)	\
 	".previous"							\
 	:"=&r" (temp), "=m" (v->counter), "=&r" (result)		\
 	:"Ir" (i), "m" (v->counter) : "memory");			\
+	smp_read_barrier_depends();					\
 	return result;							\
 }
 
-- 
2.5.2

[PATCH 19/19] COCCINELLE: treewide: kill off ACCESS_ONCE()

[Paul E. McKenney]

From: Mark Rutland <mark.rutland@arm.com>

Please do not apply this to mainline directly, instead please re-run the
coccinelle script shown below and apply its output.

For several reasons, it is desirable to use {READ,WRITE}_ONCE() in
preference to ACCESS_ONCE(), and new code is expected to use one of the
former. So far, there's been no reason to change most existing uses of
ACCESS_ONCE(), as these aren't harmful, and changing them results in
churn.

However, for some features, the read/write distinction is critical to
correct operation. To distinguish these cases, separate read/write
accessors must be used. This patch migrates (most) remaining
ACCESS_ONCE() instances to {READ,WRITE}_ONCE(), using the following
coccinelle script:

----
// Convert trivial ACCESS_ONCE() uses to equivalent READ_ONCE() and
// WRITE_ONCE()

// $ make coccicheck COCCI=/home/mark/once.cocci SPFLAGS="--include-headers" MODE=patch

virtual patch

@ depends on patch @
expression E1, E2;
@@

- ACCESS_ONCE(E1) = E2
+ WRITE_ONCE(E1, E2)

@ depends on patch @
expression E;
@@

- ACCESS_ONCE(E)
+ READ_ONCE(E)
----

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
---
 arch/arc/kernel/smp.c                              |  2 +-
 arch/arm/include/asm/spinlock.h                    |  6 ++--
 arch/arm/mach-tegra/cpuidle-tegra20.c              |  2 +-
 arch/arm/vdso/vgettimeofday.c                      |  2 +-
 arch/ia64/include/asm/spinlock.h                   |  8 ++---
 arch/mips/include/asm/vdso.h                       |  2 +-
 arch/mips/kernel/pm-cps.c                          |  2 +-
 arch/mn10300/kernel/mn10300-serial.c               |  4 +--
 arch/parisc/include/asm/atomic.h                   |  2 +-
 arch/powerpc/platforms/powernv/opal-msglog.c       |  2 +-
 arch/s390/include/asm/spinlock.h                   |  6 ++--
 arch/s390/lib/spinlock.c                           | 16 +++++-----
 arch/sparc/include/asm/atomic_32.h                 |  2 +-
 arch/tile/gxio/dma_queue.c                         |  4 +--
 arch/tile/include/gxio/dma_queue.h                 |  2 +-
 arch/tile/kernel/ptrace.c                          |  2 +-
 arch/x86/entry/common.c                            |  2 +-
 arch/x86/entry/vdso/vclock_gettime.c               |  2 +-
 arch/x86/events/core.c                             |  2 +-
 arch/x86/include/asm/vgtod.h                       |  2 +-
 arch/x86/kernel/espfix_64.c                        |  6 ++--
 arch/x86/kernel/nmi.c                              |  2 +-
 arch/x86/kvm/mmu.c                                 |  4 +--
 arch/x86/kvm/page_track.c                          |  2 +-
 arch/x86/xen/p2m.c                                 |  2 +-
 arch/xtensa/platforms/xtfpga/lcd.c                 | 14 ++++-----
 block/blk-wbt.c                                    |  2 +-
 drivers/base/core.c                                |  2 +-
 drivers/base/power/runtime.c                       |  4 +--
 drivers/char/random.c                              |  4 +--
 drivers/clocksource/bcm2835_timer.c                |  2 +-
 drivers/crypto/caam/jr.c                           |  4 +--
 drivers/crypto/nx/nx-842-powernv.c                 |  2 +-
 drivers/firewire/ohci.c                            | 10 +++---
 drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c          |  4 +--
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c            |  4 +--
 drivers/gpu/drm/amd/scheduler/gpu_scheduler.c      |  2 +-
 drivers/gpu/drm/radeon/radeon_gem.c                |  4 +--
 drivers/gpu/drm/vmwgfx/vmwgfx_surface.c            |  2 +-
 drivers/infiniband/hw/hfi1/file_ops.c              |  2 +-
 drivers/infiniband/hw/hfi1/pio.c                   |  6 ++--
 drivers/infiniband/hw/hfi1/ruc.c                   |  2 +-
 drivers/infiniband/hw/hfi1/sdma.c                  |  8 ++---
 drivers/infiniband/hw/hfi1/sdma.h                  |  2 +-
 drivers/infiniband/hw/hfi1/uc.c                    |  4 +--
 drivers/infiniband/hw/hfi1/ud.c                    |  4 +--
 drivers/infiniband/hw/hfi1/user_sdma.c             |  8 ++---
 drivers/infiniband/hw/qib/qib_ruc.c                |  2 +-
 drivers/infiniband/hw/qib/qib_uc.c                 |  4 +--
 drivers/infiniband/hw/qib/qib_ud.c                 |  4 +--
 drivers/infiniband/sw/rdmavt/qp.c                  |  6 ++--
 drivers/input/misc/regulator-haptic.c              |  2 +-
 drivers/md/dm-bufio.c                              | 10 +++---
 drivers/md/dm-kcopyd.c                             |  4 +--
 drivers/md/dm-stats.c                              | 36 +++++++++++-----------
 drivers/md/dm-switch.c                             |  2 +-
 drivers/md/dm-thin.c                               |  2 +-
 drivers/md/dm-verity-target.c                      |  2 +-
 drivers/md/dm.c                                    |  4 +--
 drivers/md/md.c                                    |  2 +-
 drivers/md/raid5.c                                 |  2 +-
 drivers/misc/mic/scif/scif_rb.c                    |  8 ++---
 drivers/misc/mic/scif/scif_rma_list.c              |  2 +-
 drivers/net/bonding/bond_alb.c                     |  2 +-
 drivers/net/bonding/bond_main.c                    |  6 ++--
 drivers/net/ethernet/chelsio/cxgb4/sge.c           |  4 +--
 drivers/net/ethernet/emulex/benet/be_main.c        |  2 +-
 drivers/net/ethernet/hisilicon/hip04_eth.c         |  4 +--
 drivers/net/ethernet/intel/i40e/i40e_debugfs.c     |  4 +--
 drivers/net/ethernet/intel/i40e/i40e_ethtool.c     |  2 +-
 drivers/net/ethernet/intel/i40e/i40e_main.c        |  4 +--
 drivers/net/ethernet/intel/i40e/i40e_ptp.c         |  4 +--
 drivers/net/ethernet/intel/igb/e1000_regs.h        |  2 +-
 drivers/net/ethernet/intel/igb/igb_main.c          |  2 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_common.h    |  4 +--
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c      |  8 ++---
 drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c       |  4 +--
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c  |  2 +-
 drivers/net/ethernet/intel/ixgbevf/vf.h            |  2 +-
 drivers/net/ethernet/mellanox/mlx4/en_tx.c         | 12 ++++----
 drivers/net/ethernet/neterion/vxge/vxge-main.c     |  2 +-
 drivers/net/ethernet/sfc/ef10.c                    | 10 +++---
 drivers/net/ethernet/sfc/efx.c                     |  4 +--
 drivers/net/ethernet/sfc/falcon/efx.c              |  4 +--
 drivers/net/ethernet/sfc/falcon/falcon.c           |  4 +--
 drivers/net/ethernet/sfc/falcon/farch.c            |  8 ++---
 drivers/net/ethernet/sfc/falcon/nic.h              |  6 ++--
 drivers/net/ethernet/sfc/falcon/tx.c               |  6 ++--
 drivers/net/ethernet/sfc/farch.c                   |  8 ++---
 drivers/net/ethernet/sfc/nic.h                     |  6 ++--
 drivers/net/ethernet/sfc/ptp.c                     | 10 +++---
 drivers/net/ethernet/sfc/tx.c                      |  6 ++--
 drivers/net/ethernet/sun/niu.c                     |  4 +--
 drivers/net/tap.c                                  |  2 +-
 drivers/net/tun.c                                  |  4 +--
 drivers/net/wireless/ath/ath5k/desc.c              |  8 ++---
 .../wireless/broadcom/brcm80211/brcmfmac/sdio.c    |  2 +-
 drivers/net/wireless/intel/iwlwifi/mvm/ops.c       |  2 +-
 drivers/net/wireless/intel/iwlwifi/mvm/tx.c        |  4 +--
 drivers/net/wireless/intel/iwlwifi/pcie/rx.c       |  2 +-
 drivers/net/wireless/intel/iwlwifi/pcie/trans.c    | 10 +++---
 drivers/net/wireless/mac80211_hwsim.c              |  4 +--
 drivers/scsi/qla2xxx/qla_target.c                  |  2 +-
 drivers/target/target_core_user.c                  |  2 +-
 drivers/usb/class/cdc-wdm.c                        |  2 +-
 drivers/usb/core/devio.c                           |  2 +-
 drivers/usb/core/sysfs.c                           |  4 +--
 drivers/usb/gadget/udc/gr_udc.c                    |  4 +--
 drivers/usb/host/ohci-hcd.c                        |  2 +-
 drivers/usb/host/uhci-hcd.h                        |  4 +--
 drivers/vfio/vfio.c                                |  2 +-
 drivers/vhost/scsi.c                               |  2 +-
 fs/aio.c                                           |  2 +-
 fs/buffer.c                                        |  3 +-
 fs/crypto/keyinfo.c                                |  2 +-
 fs/direct-io.c                                     |  2 +-
 fs/exec.c                                          |  2 +-
 fs/fcntl.c                                         |  2 +-
 fs/fs_pin.c                                        |  4 +--
 fs/fuse/dev.c                                      |  2 +-
 fs/inode.c                                         |  2 +-
 fs/namei.c                                         |  4 +--
 fs/namespace.c                                     |  2 +-
 fs/nfs/dir.c                                       |  8 ++---
 fs/proc/array.c                                    |  2 +-
 fs/proc_namespace.c                                |  2 +-
 fs/splice.c                                        |  2 +-
 fs/userfaultfd.c                                   |  8 ++---
 fs/xfs/xfs_log_priv.h                              |  4 +--
 include/linux/bitops.h                             |  4 +--
 include/linux/dynamic_queue_limits.h               |  2 +-
 include/linux/huge_mm.h                            |  2 +-
 include/linux/if_team.h                            |  2 +-
 include/linux/llist.h                              |  2 +-
 include/linux/pm_runtime.h                         |  2 +-
 include/net/ip_vs.h                                |  6 ++--
 kernel/acct.c                                      |  4 +--
 kernel/events/core.c                               |  6 ++--
 kernel/events/ring_buffer.c                        |  2 +-
 kernel/exit.c                                      |  2 +-
 kernel/trace/ring_buffer.c                         |  2 +-
 kernel/trace/trace.h                               |  2 +-
 kernel/trace/trace_stack.c                         |  2 +-
 kernel/user_namespace.c                            |  2 +-
 lib/assoc_array.c                                  | 20 ++++++------
 lib/dynamic_queue_limits.c                         |  2 +-
 lib/llist.c                                        |  2 +-
 lib/vsprintf.c                                     |  4 +--
 mm/huge_memory.c                                   |  2 +-
 net/core/dev.c                                     |  2 +-
 net/core/pktgen.c                                  |  2 +-
 net/ipv4/inet_fragment.c                           |  2 +-
 net/ipv4/route.c                                   |  2 +-
 net/ipv4/tcp_output.c                              |  2 +-
 net/ipv4/udp.c                                     |  4 +--
 net/ipv6/ip6_tunnel.c                              |  8 ++---
 net/ipv6/udp.c                                     |  4 +--
 net/llc/llc_input.c                                |  4 +--
 net/mac80211/sta_info.c                            |  2 +-
 net/netlabel/netlabel_calipso.c                    |  2 +-
 net/wireless/nl80211.c                             |  2 +-
 sound/firewire/amdtp-am824.c                       |  6 ++--
 sound/firewire/amdtp-stream.c                      | 23 +++++++-------
 sound/firewire/amdtp-stream.h                      |  2 +-
 sound/firewire/digi00x/amdtp-dot.c                 |  6 ++--
 sound/firewire/fireface/amdtp-ff.c                 |  4 +--
 sound/firewire/fireface/ff-midi.c                  | 10 +++---
 sound/firewire/fireface/ff-transaction.c           |  8 ++---
 sound/firewire/isight.c                            | 18 +++++------
 sound/firewire/motu/amdtp-motu.c                   |  4 +--
 sound/firewire/oxfw/oxfw-scs1x.c                   | 12 ++++----
 sound/firewire/tascam/amdtp-tascam.c               |  4 +--
 sound/firewire/tascam/tascam-transaction.c         |  6 ++--
 sound/soc/xtensa/xtfpga-i2s.c                      |  6 ++--
 sound/usb/bcd2000/bcd2000.c                        |  4 +--
 tools/arch/x86/include/asm/atomic.h                |  2 +-
 tools/include/asm-generic/atomic-gcc.h             |  2 +-
 tools/perf/util/auxtrace.h                         |  4 +--
 tools/perf/util/evlist.h                           |  2 +-
 tools/perf/util/session.h                          |  2 +-
 virt/kvm/kvm_main.c                                |  2 +-
 181 files changed, 386 insertions(+), 388 deletions(-)

diff --git a/arch/arc/kernel/smp.c b/arch/arc/kernel/smp.c
index f46267153ec2..94cabe73664b 100644
--- a/arch/arc/kernel/smp.c
+++ b/arch/arc/kernel/smp.c
@@ -245,7 +245,7 @@ static void ipi_send_msg_one(int cpu, enum ipi_msg_type msg)
 	 * and read back old value
 	 */
 	do {
-		new = old = ACCESS_ONCE(*ipi_data_ptr);
+		new = old = READ_ONCE(*ipi_data_ptr);
 		new |= 1U << msg;
 	} while (cmpxchg(ipi_data_ptr, old, new) != old);
 
diff --git a/arch/arm/include/asm/spinlock.h b/arch/arm/include/asm/spinlock.h
index c030143c18c6..060e4ebb4e8a 100644
--- a/arch/arm/include/asm/spinlock.h
+++ b/arch/arm/include/asm/spinlock.h
@@ -73,7 +73,7 @@ static inline void arch_spin_lock(arch_spinlock_t *lock)
 
 	while (lockval.tickets.next != lockval.tickets.owner) {
 		wfe();
-		lockval.tickets.owner = ACCESS_ONCE(lock->tickets.owner);
+		lockval.tickets.owner = READ_ONCE(lock->tickets.owner);
 	}
 
 	smp_mb();
@@ -194,7 +194,7 @@ static inline void arch_write_unlock(arch_rwlock_t *rw)
 }
 
 /* write_can_lock - would write_trylock() succeed? */
-#define arch_write_can_lock(x)		(ACCESS_ONCE((x)->lock) == 0)
+#define arch_write_can_lock(x)		(READ_ONCE((x)->lock) == 0)
 
 /*
  * Read locks are a bit more hairy:
@@ -274,7 +274,7 @@ static inline int arch_read_trylock(arch_rwlock_t *rw)
 }
 
 /* read_can_lock - would read_trylock() succeed? */
-#define arch_read_can_lock(x)		(ACCESS_ONCE((x)->lock) < 0x80000000)
+#define arch_read_can_lock(x)		(READ_ONCE((x)->lock) < 0x80000000)
 
 #define arch_read_lock_flags(lock, flags) arch_read_lock(lock)
 #define arch_write_lock_flags(lock, flags) arch_write_lock(lock)
diff --git a/arch/arm/mach-tegra/cpuidle-tegra20.c b/arch/arm/mach-tegra/cpuidle-tegra20.c
index 76e4c83cd5c8..3f24addd7972 100644
--- a/arch/arm/mach-tegra/cpuidle-tegra20.c
+++ b/arch/arm/mach-tegra/cpuidle-tegra20.c
@@ -179,7 +179,7 @@ static int tegra20_idle_lp2_coupled(struct cpuidle_device *dev,
 	bool entered_lp2 = false;
 
 	if (tegra_pending_sgi())
-		ACCESS_ONCE(abort_flag) = true;
+		WRITE_ONCE(abort_flag, true);
 
 	cpuidle_coupled_parallel_barrier(dev, &abort_barrier);
 
diff --git a/arch/arm/vdso/vgettimeofday.c b/arch/arm/vdso/vgettimeofday.c
index 79214d5ff097..a9dd619c6c29 100644
--- a/arch/arm/vdso/vgettimeofday.c
+++ b/arch/arm/vdso/vgettimeofday.c
@@ -35,7 +35,7 @@ static notrace u32 __vdso_read_begin(const struct vdso_data *vdata)
 {
 	u32 seq;
 repeat:
-	seq = ACCESS_ONCE(vdata->seq_count);
+	seq = READ_ONCE(vdata->seq_count);
 	if (seq & 1) {
 		cpu_relax();
 		goto repeat;
diff --git a/arch/ia64/include/asm/spinlock.h b/arch/ia64/include/asm/spinlock.h
index df2c121164b8..9c1f0b3b7ef5 100644
--- a/arch/ia64/include/asm/spinlock.h
+++ b/arch/ia64/include/asm/spinlock.h
@@ -61,7 +61,7 @@ static __always_inline void __ticket_spin_lock(arch_spinlock_t *lock)
 
 static __always_inline int __ticket_spin_trylock(arch_spinlock_t *lock)
 {
-	int tmp = ACCESS_ONCE(lock->lock);
+	int tmp = READ_ONCE(lock->lock);
 
 	if (!(((tmp >> TICKET_SHIFT) ^ tmp) & TICKET_MASK))
 		return ia64_cmpxchg(acq, &lock->lock, tmp, tmp + 1, sizeof (tmp)) == tmp;
@@ -73,19 +73,19 @@ static __always_inline void __ticket_spin_unlock(arch_spinlock_t *lock)
 	unsigned short	*p = (unsigned short *)&lock->lock + 1, tmp;
 
 	asm volatile ("ld2.bias %0=[%1]" : "=r"(tmp) : "r"(p));
-	ACCESS_ONCE(*p) = (tmp + 2) & ~1;
+	WRITE_ONCE(*p, (tmp + 2) & ~1);
 }
 
 static inline int __ticket_spin_is_locked(arch_spinlock_t *lock)
 {
-	long tmp = ACCESS_ONCE(lock->lock);
+	long tmp = READ_ONCE(lock->lock);
 
 	return !!(((tmp >> TICKET_SHIFT) ^ tmp) & TICKET_MASK);
 }
 
 static inline int __ticket_spin_is_contended(arch_spinlock_t *lock)
 {
-	long tmp = ACCESS_ONCE(lock->lock);
+	long tmp = READ_ONCE(lock->lock);
 
 	return ((tmp - (tmp >> TICKET_SHIFT)) & TICKET_MASK) > 1;
 }
diff --git a/arch/mips/include/asm/vdso.h b/arch/mips/include/asm/vdso.h
index b7cd6cf77b83..91bf0c2c265c 100644
--- a/arch/mips/include/asm/vdso.h
+++ b/arch/mips/include/asm/vdso.h
@@ -99,7 +99,7 @@ static inline u32 vdso_data_read_begin(const union mips_vdso_data *data)
 	u32 seq;
 
 	while (true) {
-		seq = ACCESS_ONCE(data->seq_count);
+		seq = READ_ONCE(data->seq_count);
 		if (likely(!(seq & 1))) {
 			/* Paired with smp_wmb() in vdso_data_write_*(). */
 			smp_rmb();
diff --git a/arch/mips/kernel/pm-cps.c b/arch/mips/kernel/pm-cps.c
index 4655017f2377..1d2996cd58da 100644
--- a/arch/mips/kernel/pm-cps.c
+++ b/arch/mips/kernel/pm-cps.c
@@ -166,7 +166,7 @@ int cps_pm_enter_state(enum cps_pm_state state)
 	nc_core_ready_count = nc_addr;
 
 	/* Ensure ready_count is zero-initialised before the assembly runs */
-	ACCESS_ONCE(*nc_core_ready_count) = 0;
+	WRITE_ONCE(*nc_core_ready_count, 0);
 	coupled_barrier(&per_cpu(pm_barrier, core), online);
 
 	/* Run the generated entry code */
diff --git a/arch/mn10300/kernel/mn10300-serial.c b/arch/mn10300/kernel/mn10300-serial.c
index 7ecf69879e2d..d7ef1232a82a 100644
--- a/arch/mn10300/kernel/mn10300-serial.c
+++ b/arch/mn10300/kernel/mn10300-serial.c
@@ -543,7 +543,7 @@ static void mn10300_serial_receive_interrupt(struct mn10300_serial_port *port)
 
 try_again:
 	/* pull chars out of the hat */
-	ix = ACCESS_ONCE(port->rx_outp);
+	ix = READ_ONCE(port->rx_outp);
 	if (CIRC_CNT(port->rx_inp, ix, MNSC_BUFFER_SIZE) == 0) {
 		if (push && !tport->low_latency)
 			tty_flip_buffer_push(tport);
@@ -1724,7 +1724,7 @@ static int mn10300_serial_poll_get_char(struct uart_port *_port)
 	if (mn10300_serial_int_tbl[port->rx_irq].port != NULL) {
 		do {
 			/* pull chars out of the hat */
-			ix = ACCESS_ONCE(port->rx_outp);
+			ix = READ_ONCE(port->rx_outp);
 			if (CIRC_CNT(port->rx_inp, ix, MNSC_BUFFER_SIZE) == 0)
 				return NO_POLL_CHAR;
 
diff --git a/arch/parisc/include/asm/atomic.h b/arch/parisc/include/asm/atomic.h
index 17b98a87e5e2..c57d4e8307f2 100644
--- a/arch/parisc/include/asm/atomic.h
+++ b/arch/parisc/include/asm/atomic.h
@@ -260,7 +260,7 @@ atomic64_set(atomic64_t *v, s64 i)
 static __inline__ s64
 atomic64_read(const atomic64_t *v)
 {
-	return ACCESS_ONCE((v)->counter);
+	return READ_ONCE((v)->counter);
 }
 
 #define atomic64_inc(v)		(atomic64_add(   1,(v)))
diff --git a/arch/powerpc/platforms/powernv/opal-msglog.c b/arch/powerpc/platforms/powernv/opal-msglog.c
index 7a9cde0cfbd1..acd3206dfae3 100644
--- a/arch/powerpc/platforms/powernv/opal-msglog.c
+++ b/arch/powerpc/platforms/powernv/opal-msglog.c
@@ -43,7 +43,7 @@ ssize_t opal_msglog_copy(char *to, loff_t pos, size_t count)
 	if (!opal_memcons)
 		return -ENODEV;
 
-	out_pos = be32_to_cpu(ACCESS_ONCE(opal_memcons->out_pos));
+	out_pos = be32_to_cpu(READ_ONCE(opal_memcons->out_pos));
 
 	/* Now we've read out_pos, put a barrier in before reading the new
 	 * data it points to in conbuf. */
diff --git a/arch/s390/include/asm/spinlock.h b/arch/s390/include/asm/spinlock.h
index 8182b521c42f..fc03f9dc7cd1 100644
--- a/arch/s390/include/asm/spinlock.h
+++ b/arch/s390/include/asm/spinlock.h
@@ -130,14 +130,14 @@ extern int _raw_write_trylock_retry(arch_rwlock_t *lp);
 
 static inline int arch_read_trylock_once(arch_rwlock_t *rw)
 {
-	int old = ACCESS_ONCE(rw->lock);
+	int old = READ_ONCE(rw->lock);
 	return likely(old >= 0 &&
 		      __atomic_cmpxchg_bool(&rw->lock, old, old + 1));
 }
 
 static inline int arch_write_trylock_once(arch_rwlock_t *rw)
 {
-	int old = ACCESS_ONCE(rw->lock);
+	int old = READ_ONCE(rw->lock);
 	return likely(old == 0 &&
 		      __atomic_cmpxchg_bool(&rw->lock, 0, 0x80000000));
 }
@@ -224,7 +224,7 @@ static inline void arch_read_unlock(arch_rwlock_t *rw)
 	int old;
 
 	do {
-		old = ACCESS_ONCE(rw->lock);
+		old = READ_ONCE(rw->lock);
 	} while (!__atomic_cmpxchg_bool(&rw->lock, old, old - 1));
 }
 
diff --git a/arch/s390/lib/spinlock.c b/arch/s390/lib/spinlock.c
index b12663d653d8..34e30b9ea234 100644
--- a/arch/s390/lib/spinlock.c
+++ b/arch/s390/lib/spinlock.c
@@ -162,8 +162,8 @@ void _raw_read_lock_wait(arch_rwlock_t *rw)
 				smp_yield_cpu(~owner);
 			count = spin_retry;
 		}
-		old = ACCESS_ONCE(rw->lock);
-		owner = ACCESS_ONCE(rw->owner);
+		old = READ_ONCE(rw->lock);
+		owner = READ_ONCE(rw->owner);
 		if (old < 0)
 			continue;
 		if (__atomic_cmpxchg_bool(&rw->lock, old, old + 1))
@@ -178,7 +178,7 @@ int _raw_read_trylock_retry(arch_rwlock_t *rw)
 	int old;
 
 	while (count-- > 0) {
-		old = ACCESS_ONCE(rw->lock);
+		old = READ_ONCE(rw->lock);
 		if (old < 0)
 			continue;
 		if (__atomic_cmpxchg_bool(&rw->lock, old, old + 1))
@@ -202,8 +202,8 @@ void _raw_write_lock_wait(arch_rwlock_t *rw, int prev)
 				smp_yield_cpu(~owner);
 			count = spin_retry;
 		}
-		old = ACCESS_ONCE(rw->lock);
-		owner = ACCESS_ONCE(rw->owner);
+		old = READ_ONCE(rw->lock);
+		owner = READ_ONCE(rw->owner);
 		smp_mb();
 		if (old >= 0) {
 			prev = __RAW_LOCK(&rw->lock, 0x80000000, __RAW_OP_OR);
@@ -230,8 +230,8 @@ void _raw_write_lock_wait(arch_rwlock_t *rw)
 				smp_yield_cpu(~owner);
 			count = spin_retry;
 		}
-		old = ACCESS_ONCE(rw->lock);
-		owner = ACCESS_ONCE(rw->owner);
+		old = READ_ONCE(rw->lock);
+		owner = READ_ONCE(rw->owner);
 		if (old >= 0 &&
 		    __atomic_cmpxchg_bool(&rw->lock, old, old | 0x80000000))
 			prev = old;
@@ -251,7 +251,7 @@ int _raw_write_trylock_retry(arch_rwlock_t *rw)
 	int old;
 
 	while (count-- > 0) {
-		old = ACCESS_ONCE(rw->lock);
+		old = READ_ONCE(rw->lock);
 		if (old)
 			continue;
 		if (__atomic_cmpxchg_bool(&rw->lock, 0, 0x80000000))
diff --git a/arch/sparc/include/asm/atomic_32.h b/arch/sparc/include/asm/atomic_32.h
index 7643e979e333..e2f398e9456c 100644
--- a/arch/sparc/include/asm/atomic_32.h
+++ b/arch/sparc/include/asm/atomic_32.h
@@ -31,7 +31,7 @@ void atomic_set(atomic_t *, int);
 
 #define atomic_set_release(v, i)	atomic_set((v), (i))
 
-#define atomic_read(v)          ACCESS_ONCE((v)->counter)
+#define atomic_read(v)          READ_ONCE((v)->counter)
 
 #define atomic_add(i, v)	((void)atomic_add_return( (int)(i), (v)))
 #define atomic_sub(i, v)	((void)atomic_add_return(-(int)(i), (v)))
diff --git a/arch/tile/gxio/dma_queue.c b/arch/tile/gxio/dma_queue.c
index baa60357f8ba..b7ba577d82ca 100644
--- a/arch/tile/gxio/dma_queue.c
+++ b/arch/tile/gxio/dma_queue.c
@@ -163,14 +163,14 @@ int __gxio_dma_queue_is_complete(__gxio_dma_queue_t *dma_queue,
 				 int64_t completion_slot, int update)
 {
 	if (update) {
-		if (ACCESS_ONCE(dma_queue->hw_complete_count) >
+		if (READ_ONCE(dma_queue->hw_complete_count) >
 		    completion_slot)
 			return 1;
 
 		__gxio_dma_queue_update_credits(dma_queue);
 	}
 
-	return ACCESS_ONCE(dma_queue->hw_complete_count) > completion_slot;
+	return READ_ONCE(dma_queue->hw_complete_count) > completion_slot;
 }
 
 EXPORT_SYMBOL_GPL(__gxio_dma_queue_is_complete);
diff --git a/arch/tile/include/gxio/dma_queue.h b/arch/tile/include/gxio/dma_queue.h
index b9e45e37649e..c8fd47edba30 100644
--- a/arch/tile/include/gxio/dma_queue.h
+++ b/arch/tile/include/gxio/dma_queue.h
@@ -121,7 +121,7 @@ static inline int64_t __gxio_dma_queue_reserve(__gxio_dma_queue_t *dma_queue,
 		 * if the result is LESS than "hw_complete_count".
 		 */
 		uint64_t complete;
-		complete = ACCESS_ONCE(dma_queue->hw_complete_count);
+		complete = READ_ONCE(dma_queue->hw_complete_count);
 		slot |= (complete & 0xffffffffff000000);
 		if (slot < complete)
 			slot += 0x1000000;
diff --git a/arch/tile/kernel/ptrace.c b/arch/tile/kernel/ptrace.c
index e1a078e6828e..d516d61751c2 100644
--- a/arch/tile/kernel/ptrace.c
+++ b/arch/tile/kernel/ptrace.c
@@ -255,7 +255,7 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request,
 
 int do_syscall_trace_enter(struct pt_regs *regs)
 {
-	u32 work = ACCESS_ONCE(current_thread_info()->flags);
+	u32 work = READ_ONCE(current_thread_info()->flags);
 
 	if ((work & _TIF_SYSCALL_TRACE) &&
 	    tracehook_report_syscall_entry(regs)) {
diff --git a/arch/x86/entry/common.c b/arch/x86/entry/common.c
index 03505ffbe1b6..eaa0ba66cf96 100644
--- a/arch/x86/entry/common.c
+++ b/arch/x86/entry/common.c
@@ -75,7 +75,7 @@ static long syscall_trace_enter(struct pt_regs *regs)
 	if (IS_ENABLED(CONFIG_DEBUG_ENTRY))
 		BUG_ON(regs != task_pt_regs(current));
 
-	work = ACCESS_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY;
+	work = READ_ONCE(ti->flags) & _TIF_WORK_SYSCALL_ENTRY;
 
 	if (unlikely(work & _TIF_SYSCALL_EMU))
 		emulated = true;
diff --git a/arch/x86/entry/vdso/vclock_gettime.c b/arch/x86/entry/vdso/vclock_gettime.c
index fa8dbfcf7ed3..11b13c4b43d5 100644
--- a/arch/x86/entry/vdso/vclock_gettime.c
+++ b/arch/x86/entry/vdso/vclock_gettime.c
@@ -318,7 +318,7 @@ int gettimeofday(struct timeval *, struct timezone *)
 notrace time_t __vdso_time(time_t *t)
 {
 	/* This is atomic on x86 so we don't need any locks. */
-	time_t result = ACCESS_ONCE(gtod->wall_time_sec);
+	time_t result = READ_ONCE(gtod->wall_time_sec);
 
 	if (t)
 		*t = result;
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c
index 589af1eec7c1..140d33288e78 100644
--- a/arch/x86/events/core.c
+++ b/arch/x86/events/core.c
@@ -2118,7 +2118,7 @@ static int x86_pmu_event_init(struct perf_event *event)
 			event->destroy(event);
 	}
 
-	if (ACCESS_ONCE(x86_pmu.attr_rdpmc))
+	if (READ_ONCE(x86_pmu.attr_rdpmc))
 		event->hw.flags |= PERF_X86_EVENT_RDPMC_ALLOWED;
 
 	return err;
diff --git a/arch/x86/include/asm/vgtod.h b/arch/x86/include/asm/vgtod.h
index 022e59714562..53dd162576a8 100644
--- a/arch/x86/include/asm/vgtod.h
+++ b/arch/x86/include/asm/vgtod.h
@@ -48,7 +48,7 @@ static inline unsigned gtod_read_begin(const struct vsyscall_gtod_data *s)
 	unsigned ret;
 
 repeat:
-	ret = ACCESS_ONCE(s->seq);
+	ret = READ_ONCE(s->seq);
 	if (unlikely(ret & 1)) {
 		cpu_relax();
 		goto repeat;
diff --git a/arch/x86/kernel/espfix_64.c b/arch/x86/kernel/espfix_64.c
index 9c4e7ba6870c..7d7715dde901 100644
--- a/arch/x86/kernel/espfix_64.c
+++ b/arch/x86/kernel/espfix_64.c
@@ -155,14 +155,14 @@ void init_espfix_ap(int cpu)
 	page = cpu/ESPFIX_STACKS_PER_PAGE;
 
 	/* Did another CPU already set this up? */
-	stack_page = ACCESS_ONCE(espfix_pages[page]);
+	stack_page = READ_ONCE(espfix_pages[page]);
 	if (likely(stack_page))
 		goto done;
 
 	mutex_lock(&espfix_init_mutex);
 
 	/* Did we race on the lock? */
-	stack_page = ACCESS_ONCE(espfix_pages[page]);
+	stack_page = READ_ONCE(espfix_pages[page]);
 	if (stack_page)
 		goto unlock_done;
 
@@ -200,7 +200,7 @@ void init_espfix_ap(int cpu)
 		set_pte(&pte_p[n*PTE_STRIDE], pte);
 
 	/* Job is done for this CPU and any CPU which shares this page */
-	ACCESS_ONCE(espfix_pages[page]) = stack_page;
+	WRITE_ONCE(espfix_pages[page], stack_page);
 
 unlock_done:
 	mutex_unlock(&espfix_init_mutex);
diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c
index 35aafc95e4b8..18bc9b51ac9b 100644
--- a/arch/x86/kernel/nmi.c
+++ b/arch/x86/kernel/nmi.c
@@ -105,7 +105,7 @@ static void nmi_max_handler(struct irq_work *w)
 {
 	struct nmiaction *a = container_of(w, struct nmiaction, irq_work);
 	int remainder_ns, decimal_msecs;
-	u64 whole_msecs = ACCESS_ONCE(a->max_duration);
+	u64 whole_msecs = READ_ONCE(a->max_duration);
 
 	remainder_ns = do_div(whole_msecs, (1000 * 1000));
 	decimal_msecs = remainder_ns / 1000;
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 106d4a029a8a..af19a4291d54 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -443,7 +443,7 @@ static u64 __update_clear_spte_slow(u64 *sptep, u64 spte)
 
 static u64 __get_spte_lockless(u64 *sptep)
 {
-	return ACCESS_ONCE(*sptep);
+	return READ_ONCE(*sptep);
 }
 #else
 union split_spte {
@@ -4818,7 +4818,7 @@ static void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa,
 	 * If we don't have indirect shadow pages, it means no page is
 	 * write-protected, so we can exit simply.
 	 */
-	if (!ACCESS_ONCE(vcpu->kvm->arch.indirect_shadow_pages))
+	if (!READ_ONCE(vcpu->kvm->arch.indirect_shadow_pages))
 		return;
 
 	remote_flush = local_flush = false;
diff --git a/arch/x86/kvm/page_track.c b/arch/x86/kvm/page_track.c
index ea67dc876316..01c1371f39f8 100644
--- a/arch/x86/kvm/page_track.c
+++ b/arch/x86/kvm/page_track.c
@@ -157,7 +157,7 @@ bool kvm_page_track_is_active(struct kvm_vcpu *vcpu, gfn_t gfn,
 		return false;
 
 	index = gfn_to_index(gfn, slot->base_gfn, PT_PAGE_TABLE_LEVEL);
-	return !!ACCESS_ONCE(slot->arch.gfn_track[mode][index]);
+	return !!READ_ONCE(slot->arch.gfn_track[mode][index]);
 }
 
 void kvm_page_track_cleanup(struct kvm *kvm)
diff --git a/arch/x86/xen/p2m.c b/arch/x86/xen/p2m.c
index 6083ba462f35..13b4f19b9131 100644
--- a/arch/x86/xen/p2m.c
+++ b/arch/x86/xen/p2m.c
@@ -547,7 +547,7 @@ int xen_alloc_p2m_entry(unsigned long pfn)
 	if (p2m_top_mfn && pfn < MAX_P2M_PFN) {
 		topidx = p2m_top_index(pfn);
 		top_mfn_p = &p2m_top_mfn[topidx];
-		mid_mfn = ACCESS_ONCE(p2m_top_mfn_p[topidx]);
+		mid_mfn = READ_ONCE(p2m_top_mfn_p[topidx]);
 
 		BUG_ON(virt_to_mfn(mid_mfn) != *top_mfn_p);
 
diff --git a/arch/xtensa/platforms/xtfpga/lcd.c b/arch/xtensa/platforms/xtfpga/lcd.c
index 4dc0c1b43f4b..2f7eb66c23ec 100644
--- a/arch/xtensa/platforms/xtfpga/lcd.c
+++ b/arch/xtensa/platforms/xtfpga/lcd.c
@@ -34,23 +34,23 @@
 static void lcd_put_byte(u8 *addr, u8 data)
 {
 #ifdef CONFIG_XTFPGA_LCD_8BIT_ACCESS
-	ACCESS_ONCE(*addr) = data;
+	WRITE_ONCE(*addr, data);
 #else
-	ACCESS_ONCE(*addr) = data & 0xf0;
-	ACCESS_ONCE(*addr) = (data << 4) & 0xf0;
+	WRITE_ONCE(*addr, data & 0xf0);
+	WRITE_ONCE(*addr, (data << 4) & 0xf0);
 #endif
 }
 
 static int __init lcd_init(void)
 {
-	ACCESS_ONCE(*LCD_INSTR_ADDR) = LCD_DISPLAY_MODE8BIT;
+	WRITE_ONCE(*LCD_INSTR_ADDR, LCD_DISPLAY_MODE8BIT);
 	mdelay(5);
-	ACCESS_ONCE(*LCD_INSTR_ADDR) = LCD_DISPLAY_MODE8BIT;
+	WRITE_ONCE(*LCD_INSTR_ADDR, LCD_DISPLAY_MODE8BIT);
 	udelay(200);
-	ACCESS_ONCE(*LCD_INSTR_ADDR) = LCD_DISPLAY_MODE8BIT;
+	WRITE_ONCE(*LCD_INSTR_ADDR, LCD_DISPLAY_MODE8BIT);
 	udelay(50);
 #ifndef CONFIG_XTFPGA_LCD_8BIT_ACCESS
-	ACCESS_ONCE(*LCD_INSTR_ADDR) = LCD_DISPLAY_MODE4BIT;
+	WRITE_ONCE(*LCD_INSTR_ADDR, LCD_DISPLAY_MODE4BIT);
 	udelay(50);
 	lcd_put_byte(LCD_INSTR_ADDR, LCD_DISPLAY_MODE4BIT);
 	udelay(50);
diff --git a/block/blk-wbt.c b/block/blk-wbt.c
index 6a9a0f03a67b..d822530e6aea 100644
--- a/block/blk-wbt.c
+++ b/block/blk-wbt.c
@@ -261,7 +261,7 @@ static inline bool stat_sample_valid(struct blk_rq_stat *stat)
 
 static u64 rwb_sync_issue_lat(struct rq_wb *rwb)
 {
-	u64 now, issue = ACCESS_ONCE(rwb->sync_issue);
+	u64 now, issue = READ_ONCE(rwb->sync_issue);
 
 	if (!issue || !rwb->sync_cookie)
 		return 0;
diff --git a/drivers/base/core.c b/drivers/base/core.c
index 12ebd055724c..4b8ba2a75a4d 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -668,7 +668,7 @@ const char *dev_driver_string(const struct device *dev)
 	 * so be careful about accessing it.  dev->bus and dev->class should
 	 * never change once they are set, so they don't need special care.
 	 */
-	drv = ACCESS_ONCE(dev->driver);
+	drv = READ_ONCE(dev->driver);
 	return drv ? drv->name :
 			(dev->bus ? dev->bus->name :
 			(dev->class ? dev->class->name : ""));
diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c
index 7bcf80fa9ada..41d7c2b99f69 100644
--- a/drivers/base/power/runtime.c
+++ b/drivers/base/power/runtime.c
@@ -134,11 +134,11 @@ unsigned long pm_runtime_autosuspend_expiration(struct device *dev)
 	if (!dev->power.use_autosuspend)
 		goto out;
 
-	autosuspend_delay = ACCESS_ONCE(dev->power.autosuspend_delay);
+	autosuspend_delay = READ_ONCE(dev->power.autosuspend_delay);
 	if (autosuspend_delay < 0)
 		goto out;
 
-	last_busy = ACCESS_ONCE(dev->power.last_busy);
+	last_busy = READ_ONCE(dev->power.last_busy);
 	elapsed = jiffies - last_busy;
 	if (elapsed < 0)
 		goto out;	/* jiffies has wrapped around. */
diff --git a/drivers/char/random.c b/drivers/char/random.c
index 8ad92707e45f..6c7ccac2679e 100644
--- a/drivers/char/random.c
+++ b/drivers/char/random.c
@@ -641,7 +641,7 @@ static void credit_entropy_bits(struct entropy_store *r, int nbits)
 		return;
 
 retry:
-	entropy_count = orig = ACCESS_ONCE(r->entropy_count);
+	entropy_count = orig = READ_ONCE(r->entropy_count);
 	if (nfrac < 0) {
 		/* Debit */
 		entropy_count += nfrac;
@@ -1265,7 +1265,7 @@ static size_t account(struct entropy_store *r, size_t nbytes, int min,
 
 	/* Can we pull enough? */
 retry:
-	entropy_count = orig = ACCESS_ONCE(r->entropy_count);
+	entropy_count = orig = READ_ONCE(r->entropy_count);
 	ibytes = nbytes;
 	/* never pull more than available */
 	have_bytes = entropy_count >> (ENTROPY_SHIFT + 3);
diff --git a/drivers/clocksource/bcm2835_timer.c b/drivers/clocksource/bcm2835_timer.c
index 39e489a96ad7..60da2537bef9 100644
--- a/drivers/clocksource/bcm2835_timer.c
+++ b/drivers/clocksource/bcm2835_timer.c
@@ -71,7 +71,7 @@ static irqreturn_t bcm2835_time_interrupt(int irq, void *dev_id)
 	if (readl_relaxed(timer->control) & timer->match_mask) {
 		writel_relaxed(timer->match_mask, timer->control);
 
-		event_handler = ACCESS_ONCE(timer->evt.event_handler);
+		event_handler = READ_ONCE(timer->evt.event_handler);
 		if (event_handler)
 			event_handler(&timer->evt);
 		return IRQ_HANDLED;
diff --git a/drivers/crypto/caam/jr.c b/drivers/crypto/caam/jr.c
index d258953ff488..f4f258075b89 100644
--- a/drivers/crypto/caam/jr.c
+++ b/drivers/crypto/caam/jr.c
@@ -172,7 +172,7 @@ static void caam_jr_dequeue(unsigned long devarg)
 
 	while (rd_reg32(&jrp->rregs->outring_used)) {
 
-		head = ACCESS_ONCE(jrp->head);
+		head = READ_ONCE(jrp->head);
 
 		spin_lock(&jrp->outlock);
 
@@ -341,7 +341,7 @@ int caam_jr_enqueue(struct device *dev, u32 *desc,
 	spin_lock_bh(&jrp->inplock);
 
 	head = jrp->head;
-	tail = ACCESS_ONCE(jrp->tail);
+	tail = READ_ONCE(jrp->tail);
 
 	if (!rd_reg32(&jrp->rregs->inpring_avail) ||
 	    CIRC_SPACE(head, tail, JOBR_DEPTH) <= 0) {
diff --git a/drivers/crypto/nx/nx-842-powernv.c b/drivers/crypto/nx/nx-842-powernv.c
index 874ddf5e9087..0f20f5ec9617 100644
--- a/drivers/crypto/nx/nx-842-powernv.c
+++ b/drivers/crypto/nx/nx-842-powernv.c
@@ -193,7 +193,7 @@ static int wait_for_csb(struct nx842_workmem *wmem,
 	ktime_t start = wmem->start, now = ktime_get();
 	ktime_t timeout = ktime_add_ms(start, CSB_WAIT_MAX);
 
-	while (!(ACCESS_ONCE(csb->flags) & CSB_V)) {
+	while (!(READ_ONCE(csb->flags) & CSB_V)) {
 		cpu_relax();
 		now = ktime_get();
 		if (ktime_after(now, timeout))
diff --git a/drivers/firewire/ohci.c b/drivers/firewire/ohci.c
index 8bf89267dc25..ccf52368a073 100644
--- a/drivers/firewire/ohci.c
+++ b/drivers/firewire/ohci.c
@@ -734,7 +734,7 @@ static unsigned int ar_search_last_active_buffer(struct ar_context *ctx,
 	__le16 res_count, next_res_count;
 
 	i = ar_first_buffer_index(ctx);
-	res_count = ACCESS_ONCE(ctx->descriptors[i].res_count);
+	res_count = READ_ONCE(ctx->descriptors[i].res_count);
 
 	/* A buffer that is not yet completely filled must be the last one. */
 	while (i != last && res_count == 0) {
@@ -742,8 +742,7 @@ static unsigned int ar_search_last_active_buffer(struct ar_context *ctx,
 		/* Peek at the next descriptor. */
 		next_i = ar_next_buffer_index(i);
 		rmb(); /* read descriptors in order */
-		next_res_count = ACCESS_ONCE(
-				ctx->descriptors[next_i].res_count);
+		next_res_count = READ_ONCE(ctx->descriptors[next_i].res_count);
 		/*
 		 * If the next descriptor is still empty, we must stop at this
 		 * descriptor.
@@ -759,8 +758,7 @@ static unsigned int ar_search_last_active_buffer(struct ar_context *ctx,
 			if (MAX_AR_PACKET_SIZE > PAGE_SIZE && i != last) {
 				next_i = ar_next_buffer_index(next_i);
 				rmb();
-				next_res_count = ACCESS_ONCE(
-					ctx->descriptors[next_i].res_count);
+				next_res_count = READ_ONCE(ctx->descriptors[next_i].res_count);
 				if (next_res_count != cpu_to_le16(PAGE_SIZE))
 					goto next_buffer_is_active;
 			}
@@ -2812,7 +2810,7 @@ static int handle_ir_buffer_fill(struct context *context,
 	u32 buffer_dma;
 
 	req_count = le16_to_cpu(last->req_count);
-	res_count = le16_to_cpu(ACCESS_ONCE(last->res_count));
+	res_count = le16_to_cpu(READ_ONCE(last->res_count));
 	completed = req_count - res_count;
 	buffer_dma = le32_to_cpu(last->data_address);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
index 333bad749067..303b5e099a98 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
@@ -260,7 +260,7 @@ static void amdgpu_fence_fallback(unsigned long arg)
  */
 int amdgpu_fence_wait_empty(struct amdgpu_ring *ring)
 {
-	uint64_t seq = ACCESS_ONCE(ring->fence_drv.sync_seq);
+	uint64_t seq = READ_ONCE(ring->fence_drv.sync_seq);
 	struct dma_fence *fence, **ptr;
 	int r;
 
@@ -300,7 +300,7 @@ unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring)
 	amdgpu_fence_process(ring);
 	emitted = 0x100000000ull;
 	emitted -= atomic_read(&ring->fence_drv.last_seq);
-	emitted += ACCESS_ONCE(ring->fence_drv.sync_seq);
+	emitted += READ_ONCE(ring->fence_drv.sync_seq);
 	return lower_32_bits(emitted);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 7171968f261e..6149a47fe63d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -788,11 +788,11 @@ static int amdgpu_debugfs_gem_bo_info(int id, void *ptr, void *data)
 	seq_printf(m, "\t0x%08x: %12ld byte %s",
 		   id, amdgpu_bo_size(bo), placement);
 
-	offset = ACCESS_ONCE(bo->tbo.mem.start);
+	offset = READ_ONCE(bo->tbo.mem.start);
 	if (offset != AMDGPU_BO_INVALID_OFFSET)
 		seq_printf(m, " @ 0x%010Lx", offset);
 
-	pin_count = ACCESS_ONCE(bo->pin_count);
+	pin_count = READ_ONCE(bo->pin_count);
 	if (pin_count)
 		seq_printf(m, " pin count %d", pin_count);
 	seq_printf(m, "\n");
diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c
index 97c94f9683fa..46a5ec3eb438 100644
--- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c
+++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c
@@ -187,7 +187,7 @@ static bool amd_sched_entity_is_ready(struct amd_sched_entity *entity)
 	if (kfifo_is_empty(&entity->job_queue))
 		return false;
 
-	if (ACCESS_ONCE(entity->dependency))
+	if (READ_ONCE(entity->dependency))
 		return false;
 
 	return true;
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index 3386452bd2f0..cf3deb283da5 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -451,7 +451,7 @@ int radeon_gem_busy_ioctl(struct drm_device *dev, void *data,
 	else
 		r = 0;
 
-	cur_placement = ACCESS_ONCE(robj->tbo.mem.mem_type);
+	cur_placement = READ_ONCE(robj->tbo.mem.mem_type);
 	args->domain = radeon_mem_type_to_domain(cur_placement);
 	drm_gem_object_put_unlocked(gobj);
 	return r;
@@ -481,7 +481,7 @@ int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 		r = ret;
 
 	/* Flush HDP cache via MMIO if necessary */
-	cur_placement = ACCESS_ONCE(robj->tbo.mem.mem_type);
+	cur_placement = READ_ONCE(robj->tbo.mem.mem_type);
 	if (rdev->asic->mmio_hdp_flush &&
 	    radeon_mem_type_to_domain(cur_placement) == RADEON_GEM_DOMAIN_VRAM)
 		robj->rdev->asic->mmio_hdp_flush(rdev);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c
index a552e4ea5440..6ac094ee8983 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c
@@ -904,7 +904,7 @@ vmw_surface_handle_reference(struct vmw_private *dev_priv,
 		if (unlikely(drm_is_render_client(file_priv)))
 			require_exist = true;
 
-		if (ACCESS_ONCE(vmw_fpriv(file_priv)->locked_master)) {
+		if (READ_ONCE(vmw_fpriv(file_priv)->locked_master)) {
 			DRM_ERROR("Locked master refused legacy "
 				  "surface reference.\n");
 			return -EACCES;
diff --git a/drivers/infiniband/hw/hfi1/file_ops.c b/drivers/infiniband/hw/hfi1/file_ops.c
index d9a1e9893136..97bea2e1aa6a 100644
--- a/drivers/infiniband/hw/hfi1/file_ops.c
+++ b/drivers/infiniband/hw/hfi1/file_ops.c
@@ -380,7 +380,7 @@ static long hfi1_file_ioctl(struct file *fp, unsigned int cmd,
 		if (sc->flags & SCF_FROZEN) {
 			wait_event_interruptible_timeout(
 				dd->event_queue,
-				!(ACCESS_ONCE(dd->flags) & HFI1_FROZEN),
+				!(READ_ONCE(dd->flags) & HFI1_FROZEN),
 				msecs_to_jiffies(SEND_CTXT_HALT_TIMEOUT));
 			if (dd->flags & HFI1_FROZEN)
 				return -ENOLCK;
diff --git a/drivers/infiniband/hw/hfi1/pio.c b/drivers/infiniband/hw/hfi1/pio.c
index 7108a4b5e94c..75e740780285 100644
--- a/drivers/infiniband/hw/hfi1/pio.c
+++ b/drivers/infiniband/hw/hfi1/pio.c
@@ -1423,14 +1423,14 @@ struct pio_buf *sc_buffer_alloc(struct send_context *sc, u32 dw_len,
 			goto done;
 		}
 		/* copy from receiver cache line and recalculate */
-		sc->alloc_free = ACCESS_ONCE(sc->free);
+		sc->alloc_free = READ_ONCE(sc->free);
 		avail =
 			(unsigned long)sc->credits -
 			(sc->fill - sc->alloc_free);
 		if (blocks > avail) {
 			/* still no room, actively update */
 			sc_release_update(sc);
-			sc->alloc_free = ACCESS_ONCE(sc->free);
+			sc->alloc_free = READ_ONCE(sc->free);
 			trycount++;
 			goto retry;
 		}
@@ -1667,7 +1667,7 @@ void sc_release_update(struct send_context *sc)
 
 	/* call sent buffer callbacks */
 	code = -1;				/* code not yet set */
-	head = ACCESS_ONCE(sc->sr_head);	/* snapshot the head */
+	head = READ_ONCE(sc->sr_head);	/* snapshot the head */
 	tail = sc->sr_tail;
 	while (head != tail) {
 		pbuf = &sc->sr[tail].pbuf;
diff --git a/drivers/infiniband/hw/hfi1/ruc.c b/drivers/infiniband/hw/hfi1/ruc.c
index b3291f0fde9a..a7fc664f0d4e 100644
--- a/drivers/infiniband/hw/hfi1/ruc.c
+++ b/drivers/infiniband/hw/hfi1/ruc.c
@@ -363,7 +363,7 @@ static void ruc_loopback(struct rvt_qp *sqp)
 
 again:
 	smp_read_barrier_depends(); /* see post_one_send() */
-	if (sqp->s_last == ACCESS_ONCE(sqp->s_head))
+	if (sqp->s_last == READ_ONCE(sqp->s_head))
 		goto clr_busy;
 	wqe = rvt_get_swqe_ptr(sqp, sqp->s_last);
 
diff --git a/drivers/infiniband/hw/hfi1/sdma.c b/drivers/infiniband/hw/hfi1/sdma.c
index 6781bcdb10b3..08346d25441c 100644
--- a/drivers/infiniband/hw/hfi1/sdma.c
+++ b/drivers/infiniband/hw/hfi1/sdma.c
@@ -1725,7 +1725,7 @@ static inline u16 sdma_gethead(struct sdma_engine *sde)
 
 		swhead = sde->descq_head & sde->sdma_mask;
 		/* this code is really bad for cache line trading */
-		swtail = ACCESS_ONCE(sde->descq_tail) & sde->sdma_mask;
+		swtail = READ_ONCE(sde->descq_tail) & sde->sdma_mask;
 		cnt = sde->descq_cnt;
 
 		if (swhead < swtail)
@@ -1872,7 +1872,7 @@ static void sdma_make_progress(struct sdma_engine *sde, u64 status)
 	if ((status & sde->idle_mask) && !idle_check_done) {
 		u16 swtail;
 
-		swtail = ACCESS_ONCE(sde->descq_tail) & sde->sdma_mask;
+		swtail = READ_ONCE(sde->descq_tail) & sde->sdma_mask;
 		if (swtail != hwhead) {
 			hwhead = (u16)read_sde_csr(sde, SD(HEAD));
 			idle_check_done = 1;
@@ -2222,7 +2222,7 @@ void sdma_seqfile_dump_sde(struct seq_file *s, struct sdma_engine *sde)
 	u16 len;
 
 	head = sde->descq_head & sde->sdma_mask;
-	tail = ACCESS_ONCE(sde->descq_tail) & sde->sdma_mask;
+	tail = READ_ONCE(sde->descq_tail) & sde->sdma_mask;
 	seq_printf(s, SDE_FMT, sde->this_idx,
 		   sde->cpu,
 		   sdma_state_name(sde->state.current_state),
@@ -3305,7 +3305,7 @@ int sdma_ahg_alloc(struct sdma_engine *sde)
 		return -EINVAL;
 	}
 	while (1) {
-		nr = ffz(ACCESS_ONCE(sde->ahg_bits));
+		nr = ffz(READ_ONCE(sde->ahg_bits));
 		if (nr > 31) {
 			trace_hfi1_ahg_allocate(sde, -ENOSPC);
 			return -ENOSPC;
diff --git a/drivers/infiniband/hw/hfi1/sdma.h b/drivers/infiniband/hw/hfi1/sdma.h
index 107011d8613b..374c59784950 100644
--- a/drivers/infiniband/hw/hfi1/sdma.h
+++ b/drivers/infiniband/hw/hfi1/sdma.h
@@ -445,7 +445,7 @@ static inline u16 sdma_descq_freecnt(struct sdma_engine *sde)
 {
 	return sde->descq_cnt -
 		(sde->descq_tail -
-		 ACCESS_ONCE(sde->descq_head)) - 1;
+		 READ_ONCE(sde->descq_head)) - 1;
 }
 
 static inline u16 sdma_descq_inprocess(struct sdma_engine *sde)
diff --git a/drivers/infiniband/hw/hfi1/uc.c b/drivers/infiniband/hw/hfi1/uc.c
index 0b646173ca22..9a31c585427f 100644
--- a/drivers/infiniband/hw/hfi1/uc.c
+++ b/drivers/infiniband/hw/hfi1/uc.c
@@ -80,7 +80,7 @@ int hfi1_make_uc_req(struct rvt_qp *qp, struct hfi1_pkt_state *ps)
 			goto bail;
 		/* We are in the error state, flush the work request. */
 		smp_read_barrier_depends(); /* see post_one_send() */
-		if (qp->s_last == ACCESS_ONCE(qp->s_head))
+		if (qp->s_last == READ_ONCE(qp->s_head))
 			goto bail;
 		/* If DMAs are in progress, we can't flush immediately. */
 		if (iowait_sdma_pending(&priv->s_iowait)) {
@@ -121,7 +121,7 @@ int hfi1_make_uc_req(struct rvt_qp *qp, struct hfi1_pkt_state *ps)
 			goto bail;
 		/* Check if send work queue is empty. */
 		smp_read_barrier_depends(); /* see post_one_send() */
-		if (qp->s_cur == ACCESS_ONCE(qp->s_head)) {
+		if (qp->s_cur == READ_ONCE(qp->s_head)) {
 			clear_ahg(qp);
 			goto bail;
 		}
diff --git a/drivers/infiniband/hw/hfi1/ud.c b/drivers/infiniband/hw/hfi1/ud.c
index 2ba74fdd6f15..7fec6b984e3e 100644
--- a/drivers/infiniband/hw/hfi1/ud.c
+++ b/drivers/infiniband/hw/hfi1/ud.c
@@ -487,7 +487,7 @@ int hfi1_make_ud_req(struct rvt_qp *qp, struct hfi1_pkt_state *ps)
 			goto bail;
 		/* We are in the error state, flush the work request. */
 		smp_read_barrier_depends(); /* see post_one_send */
-		if (qp->s_last == ACCESS_ONCE(qp->s_head))
+		if (qp->s_last == READ_ONCE(qp->s_head))
 			goto bail;
 		/* If DMAs are in progress, we can't flush immediately. */
 		if (iowait_sdma_pending(&priv->s_iowait)) {
@@ -501,7 +501,7 @@ int hfi1_make_ud_req(struct rvt_qp *qp, struct hfi1_pkt_state *ps)
 
 	/* see post_one_send() */
 	smp_read_barrier_depends();
-	if (qp->s_cur == ACCESS_ONCE(qp->s_head))
+	if (qp->s_cur == READ_ONCE(qp->s_head))
 		goto bail;
 
 	wqe = rvt_get_swqe_ptr(qp, qp->s_cur);
diff --git a/drivers/infiniband/hw/hfi1/user_sdma.c b/drivers/infiniband/hw/hfi1/user_sdma.c
index c0c0e0445cbf..8ec6e8a8d6f7 100644
--- a/drivers/infiniband/hw/hfi1/user_sdma.c
+++ b/drivers/infiniband/hw/hfi1/user_sdma.c
@@ -276,7 +276,7 @@ int hfi1_user_sdma_free_queues(struct hfi1_filedata *fd,
 		/* Wait until all requests have been freed. */
 		wait_event_interruptible(
 			pq->wait,
-			(ACCESS_ONCE(pq->state) == SDMA_PKT_Q_INACTIVE));
+			(READ_ONCE(pq->state) == SDMA_PKT_Q_INACTIVE));
 		kfree(pq->reqs);
 		kfree(pq->req_in_use);
 		kmem_cache_destroy(pq->txreq_cache);
@@ -591,7 +591,7 @@ int hfi1_user_sdma_process_request(struct hfi1_filedata *fd,
 			if (ret != -EBUSY) {
 				req->status = ret;
 				WRITE_ONCE(req->has_error, 1);
-				if (ACCESS_ONCE(req->seqcomp) ==
+				if (READ_ONCE(req->seqcomp) ==
 				    req->seqsubmitted - 1)
 					goto free_req;
 				return ret;
@@ -825,7 +825,7 @@ static int user_sdma_send_pkts(struct user_sdma_request *req, unsigned maxpkts)
 		 */
 		if (req->data_len) {
 			iovec = &req->iovs[req->iov_idx];
-			if (ACCESS_ONCE(iovec->offset) == iovec->iov.iov_len) {
+			if (READ_ONCE(iovec->offset) == iovec->iov.iov_len) {
 				if (++req->iov_idx == req->data_iovs) {
 					ret = -EFAULT;
 					goto free_txreq;
@@ -1390,7 +1390,7 @@ static void user_sdma_txreq_cb(struct sdma_txreq *txreq, int status)
 	} else {
 		if (status != SDMA_TXREQ_S_OK)
 			req->status = status;
-		if (req->seqcomp == (ACCESS_ONCE(req->seqsubmitted) - 1) &&
+		if (req->seqcomp == (READ_ONCE(req->seqsubmitted) - 1) &&
 		    (READ_ONCE(req->done) ||
 		     READ_ONCE(req->has_error))) {
 			user_sdma_free_request(req, false);
diff --git a/drivers/infiniband/hw/qib/qib_ruc.c b/drivers/infiniband/hw/qib/qib_ruc.c
index 53efbb0b40c4..9a37e844d4c8 100644
--- a/drivers/infiniband/hw/qib/qib_ruc.c
+++ b/drivers/infiniband/hw/qib/qib_ruc.c
@@ -368,7 +368,7 @@ static void qib_ruc_loopback(struct rvt_qp *sqp)
 
 again:
 	smp_read_barrier_depends(); /* see post_one_send() */
-	if (sqp->s_last == ACCESS_ONCE(sqp->s_head))
+	if (sqp->s_last == READ_ONCE(sqp->s_head))
 		goto clr_busy;
 	wqe = rvt_get_swqe_ptr(sqp, sqp->s_last);
 
diff --git a/drivers/infiniband/hw/qib/qib_uc.c b/drivers/infiniband/hw/qib/qib_uc.c
index 498e2202e72c..bddcc37ace44 100644
--- a/drivers/infiniband/hw/qib/qib_uc.c
+++ b/drivers/infiniband/hw/qib/qib_uc.c
@@ -61,7 +61,7 @@ int qib_make_uc_req(struct rvt_qp *qp, unsigned long *flags)
 			goto bail;
 		/* We are in the error state, flush the work request. */
 		smp_read_barrier_depends(); /* see post_one_send() */
-		if (qp->s_last == ACCESS_ONCE(qp->s_head))
+		if (qp->s_last == READ_ONCE(qp->s_head))
 			goto bail;
 		/* If DMAs are in progress, we can't flush immediately. */
 		if (atomic_read(&priv->s_dma_busy)) {
@@ -91,7 +91,7 @@ int qib_make_uc_req(struct rvt_qp *qp, unsigned long *flags)
 			goto bail;
 		/* Check if send work queue is empty. */
 		smp_read_barrier_depends(); /* see post_one_send() */
-		if (qp->s_cur == ACCESS_ONCE(qp->s_head))
+		if (qp->s_cur == READ_ONCE(qp->s_head))
 			goto bail;
 		/*
 		 * Start a new request.
diff --git a/drivers/infiniband/hw/qib/qib_ud.c b/drivers/infiniband/hw/qib/qib_ud.c
index be4907453ac4..15962ed193ce 100644
--- a/drivers/infiniband/hw/qib/qib_ud.c
+++ b/drivers/infiniband/hw/qib/qib_ud.c
@@ -253,7 +253,7 @@ int qib_make_ud_req(struct rvt_qp *qp, unsigned long *flags)
 			goto bail;
 		/* We are in the error state, flush the work request. */
 		smp_read_barrier_depends(); /* see post_one_send */
-		if (qp->s_last == ACCESS_ONCE(qp->s_head))
+		if (qp->s_last == READ_ONCE(qp->s_head))
 			goto bail;
 		/* If DMAs are in progress, we can't flush immediately. */
 		if (atomic_read(&priv->s_dma_busy)) {
@@ -267,7 +267,7 @@ int qib_make_ud_req(struct rvt_qp *qp, unsigned long *flags)
 
 	/* see post_one_send() */
 	smp_read_barrier_depends();
-	if (qp->s_cur == ACCESS_ONCE(qp->s_head))
+	if (qp->s_cur == READ_ONCE(qp->s_head))
 		goto bail;
 
 	wqe = rvt_get_swqe_ptr(qp, qp->s_cur);
diff --git a/drivers/infiniband/sw/rdmavt/qp.c b/drivers/infiniband/sw/rdmavt/qp.c
index 22df09ae809e..b670cb9d2006 100644
--- a/drivers/infiniband/sw/rdmavt/qp.c
+++ b/drivers/infiniband/sw/rdmavt/qp.c
@@ -1073,7 +1073,7 @@ int rvt_error_qp(struct rvt_qp *qp, enum ib_wc_status err)
 	rdi->driver_f.notify_error_qp(qp);
 
 	/* Schedule the sending tasklet to drain the send work queue. */
-	if (ACCESS_ONCE(qp->s_last) != qp->s_head)
+	if (READ_ONCE(qp->s_last) != qp->s_head)
 		rdi->driver_f.schedule_send(qp);
 
 	rvt_clear_mr_refs(qp, 0);
@@ -1686,7 +1686,7 @@ static inline int rvt_qp_is_avail(
 	if (likely(qp->s_avail))
 		return 0;
 	smp_read_barrier_depends(); /* see rc.c */
-	slast = ACCESS_ONCE(qp->s_last);
+	slast = READ_ONCE(qp->s_last);
 	if (qp->s_head >= slast)
 		avail = qp->s_size - (qp->s_head - slast);
 	else
@@ -1917,7 +1917,7 @@ int rvt_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 	 * ahead and kick the send engine into gear. Otherwise we will always
 	 * just schedule the send to happen later.
 	 */
-	call_send = qp->s_head == ACCESS_ONCE(qp->s_last) && !wr->next;
+	call_send = qp->s_head == READ_ONCE(qp->s_last) && !wr->next;
 
 	for (; wr; wr = wr->next) {
 		err = rvt_post_one_wr(qp, wr, &call_send);
diff --git a/drivers/input/misc/regulator-haptic.c b/drivers/input/misc/regulator-haptic.c
index 2e8f801932be..a1db1e5040dc 100644
--- a/drivers/input/misc/regulator-haptic.c
+++ b/drivers/input/misc/regulator-haptic.c
@@ -233,7 +233,7 @@ static int __maybe_unused regulator_haptic_resume(struct device *dev)
 
 	haptic->suspended = false;
 
-	magnitude = ACCESS_ONCE(haptic->magnitude);
+	magnitude = READ_ONCE(haptic->magnitude);
 	if (magnitude)
 		regulator_haptic_set_voltage(haptic, magnitude);
 
diff --git a/drivers/md/dm-bufio.c b/drivers/md/dm-bufio.c
index d216a8f7bc22..33bb074d6941 100644
--- a/drivers/md/dm-bufio.c
+++ b/drivers/md/dm-bufio.c
@@ -347,7 +347,7 @@ static void __cache_size_refresh(void)
 	BUG_ON(!mutex_is_locked(&dm_bufio_clients_lock));
 	BUG_ON(dm_bufio_client_count < 0);
 
-	dm_bufio_cache_size_latch = ACCESS_ONCE(dm_bufio_cache_size);
+	dm_bufio_cache_size_latch = READ_ONCE(dm_bufio_cache_size);
 
 	/*
 	 * Use default if set to 0 and report the actual cache size used.
@@ -960,7 +960,7 @@ static void __get_memory_limit(struct dm_bufio_client *c,
 {
 	unsigned long buffers;
 
-	if (unlikely(ACCESS_ONCE(dm_bufio_cache_size) != dm_bufio_cache_size_latch)) {
+	if (unlikely(READ_ONCE(dm_bufio_cache_size) != dm_bufio_cache_size_latch)) {
 		if (mutex_trylock(&dm_bufio_clients_lock)) {
 			__cache_size_refresh();
 			mutex_unlock(&dm_bufio_clients_lock);
@@ -1600,7 +1600,7 @@ static bool __try_evict_buffer(struct dm_buffer *b, gfp_t gfp)
 
 static unsigned long get_retain_buffers(struct dm_bufio_client *c)
 {
-        unsigned long retain_bytes = ACCESS_ONCE(dm_bufio_retain_bytes);
+        unsigned long retain_bytes = READ_ONCE(dm_bufio_retain_bytes);
         return retain_bytes >> (c->sectors_per_block_bits + SECTOR_SHIFT);
 }
 
@@ -1647,7 +1647,7 @@ dm_bufio_shrink_count(struct shrinker *shrink, struct shrink_control *sc)
 {
 	struct dm_bufio_client *c = container_of(shrink, struct dm_bufio_client, shrinker);
 
-	return ACCESS_ONCE(c->n_buffers[LIST_CLEAN]) + ACCESS_ONCE(c->n_buffers[LIST_DIRTY]);
+	return READ_ONCE(c->n_buffers[LIST_CLEAN]) + READ_ONCE(c->n_buffers[LIST_DIRTY]);
 }
 
 /*
@@ -1818,7 +1818,7 @@ EXPORT_SYMBOL_GPL(dm_bufio_set_sector_offset);
 
 static unsigned get_max_age_hz(void)
 {
-	unsigned max_age = ACCESS_ONCE(dm_bufio_max_age);
+	unsigned max_age = READ_ONCE(dm_bufio_max_age);
 
 	if (max_age > UINT_MAX / HZ)
 		max_age = UINT_MAX / HZ;
diff --git a/drivers/md/dm-kcopyd.c b/drivers/md/dm-kcopyd.c
index cf2c67e35eaf..eb45cc3df31d 100644
--- a/drivers/md/dm-kcopyd.c
+++ b/drivers/md/dm-kcopyd.c
@@ -107,7 +107,7 @@ static void io_job_start(struct dm_kcopyd_throttle *t)
 try_again:
 	spin_lock_irq(&throttle_spinlock);
 
-	throttle = ACCESS_ONCE(t->throttle);
+	throttle = READ_ONCE(t->throttle);
 
 	if (likely(throttle >= 100))
 		goto skip_limit;
@@ -157,7 +157,7 @@ static void io_job_finish(struct dm_kcopyd_throttle *t)
 
 	t->num_io_jobs--;
 
-	if (likely(ACCESS_ONCE(t->throttle) >= 100))
+	if (likely(READ_ONCE(t->throttle) >= 100))
 		goto skip_limit;
 
 	if (!t->num_io_jobs) {
diff --git a/drivers/md/dm-stats.c b/drivers/md/dm-stats.c
index 6028d8247f58..a1a5eec783cc 100644
--- a/drivers/md/dm-stats.c
+++ b/drivers/md/dm-stats.c
@@ -431,7 +431,7 @@ static int dm_stats_delete(struct dm_stats *stats, int id)
 		synchronize_rcu_expedited();
 		dm_stat_free(&s->rcu_head);
 	} else {
-		ACCESS_ONCE(dm_stat_need_rcu_barrier) = 1;
+		WRITE_ONCE(dm_stat_need_rcu_barrier, 1);
 		call_rcu(&s->rcu_head, dm_stat_free);
 	}
 	return 0;
@@ -639,12 +639,12 @@ void dm_stats_account_io(struct dm_stats *stats, unsigned long bi_rw,
 		 */
 		last = raw_cpu_ptr(stats->last);
 		stats_aux->merged =
-			(bi_sector == (ACCESS_ONCE(last->last_sector) &&
+			(bi_sector == (READ_ONCE(last->last_sector) &&
 				       ((bi_rw == WRITE) ==
-					(ACCESS_ONCE(last->last_rw) == WRITE))
+					(READ_ONCE(last->last_rw) == WRITE))
 				       ));
-		ACCESS_ONCE(last->last_sector) = end_sector;
-		ACCESS_ONCE(last->last_rw) = bi_rw;
+		WRITE_ONCE(last->last_sector, end_sector);
+		WRITE_ONCE(last->last_rw, bi_rw);
 	}
 
 	rcu_read_lock();
@@ -693,22 +693,22 @@ static void __dm_stat_init_temporary_percpu_totals(struct dm_stat_shared *shared
 
 	for_each_possible_cpu(cpu) {
 		p = &s->stat_percpu[cpu][x];
-		shared->tmp.sectors[READ] += ACCESS_ONCE(p->sectors[READ]);
-		shared->tmp.sectors[WRITE] += ACCESS_ONCE(p->sectors[WRITE]);
-		shared->tmp.ios[READ] += ACCESS_ONCE(p->ios[READ]);
-		shared->tmp.ios[WRITE] += ACCESS_ONCE(p->ios[WRITE]);
-		shared->tmp.merges[READ] += ACCESS_ONCE(p->merges[READ]);
-		shared->tmp.merges[WRITE] += ACCESS_ONCE(p->merges[WRITE]);
-		shared->tmp.ticks[READ] += ACCESS_ONCE(p->ticks[READ]);
-		shared->tmp.ticks[WRITE] += ACCESS_ONCE(p->ticks[WRITE]);
-		shared->tmp.io_ticks[READ] += ACCESS_ONCE(p->io_ticks[READ]);
-		shared->tmp.io_ticks[WRITE] += ACCESS_ONCE(p->io_ticks[WRITE]);
-		shared->tmp.io_ticks_total += ACCESS_ONCE(p->io_ticks_total);
-		shared->tmp.time_in_queue += ACCESS_ONCE(p->time_in_queue);
+		shared->tmp.sectors[READ] += READ_ONCE(p->sectors[READ]);
+		shared->tmp.sectors[WRITE] += READ_ONCE(p->sectors[WRITE]);
+		shared->tmp.ios[READ] += READ_ONCE(p->ios[READ]);
+		shared->tmp.ios[WRITE] += READ_ONCE(p->ios[WRITE]);
+		shared->tmp.merges[READ] += READ_ONCE(p->merges[READ]);
+		shared->tmp.merges[WRITE] += READ_ONCE(p->merges[WRITE]);
+		shared->tmp.ticks[READ] += READ_ONCE(p->ticks[READ]);
+		shared->tmp.ticks[WRITE] += READ_ONCE(p->ticks[WRITE]);
+		shared->tmp.io_ticks[READ] += READ_ONCE(p->io_ticks[READ]);
+		shared->tmp.io_ticks[WRITE] += READ_ONCE(p->io_ticks[WRITE]);
+		shared->tmp.io_ticks_total += READ_ONCE(p->io_ticks_total);
+		shared->tmp.time_in_queue += READ_ONCE(p->time_in_queue);
 		if (s->n_histogram_entries) {
 			unsigned i;
 			for (i = 0; i < s->n_histogram_entries + 1; i++)
-				shared->tmp.histogram[i] += ACCESS_ONCE(p->histogram[i]);
+				shared->tmp.histogram[i] += READ_ONCE(p->histogram[i]);
 		}
 	}
 }
diff --git a/drivers/md/dm-switch.c b/drivers/md/dm-switch.c
index 4c8de1ff78ca..8d0ba879777e 100644
--- a/drivers/md/dm-switch.c
+++ b/drivers/md/dm-switch.c
@@ -144,7 +144,7 @@ static unsigned switch_region_table_read(struct switch_ctx *sctx, unsigned long
 
 	switch_get_position(sctx, region_nr, &region_index, &bit);
 
-	return (ACCESS_ONCE(sctx->region_table[region_index]) >> bit) &
+	return (READ_ONCE(sctx->region_table[region_index]) >> bit) &
 		((1 << sctx->region_table_entry_bits) - 1);
 }
 
diff --git a/drivers/md/dm-thin.c b/drivers/md/dm-thin.c
index 1e25705209c2..89e5dff9b4cf 100644
--- a/drivers/md/dm-thin.c
+++ b/drivers/md/dm-thin.c
@@ -2431,7 +2431,7 @@ static void set_pool_mode(struct pool *pool, enum pool_mode new_mode)
 	struct pool_c *pt = pool->ti->private;
 	bool needs_check = dm_pool_metadata_needs_check(pool->pmd);
 	enum pool_mode old_mode = get_pool_mode(pool);
-	unsigned long no_space_timeout = ACCESS_ONCE(no_space_timeout_secs) * HZ;
+	unsigned long no_space_timeout = READ_ONCE(no_space_timeout_secs) * HZ;
 
 	/*
 	 * Never allow the pool to transition to PM_WRITE mode if user
diff --git a/drivers/md/dm-verity-target.c b/drivers/md/dm-verity-target.c
index bda3caca23ca..fba93237a780 100644
--- a/drivers/md/dm-verity-target.c
+++ b/drivers/md/dm-verity-target.c
@@ -589,7 +589,7 @@ static void verity_prefetch_io(struct work_struct *work)
 		verity_hash_at_level(v, pw->block, i, &hash_block_start, NULL);
 		verity_hash_at_level(v, pw->block + pw->n_blocks - 1, i, &hash_block_end, NULL);
 		if (!i) {
-			unsigned cluster = ACCESS_ONCE(dm_verity_prefetch_cluster);
+			unsigned cluster = READ_ONCE(dm_verity_prefetch_cluster);
 
 			cluster >>= v->data_dev_block_bits;
 			if (unlikely(!cluster))
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 4be85324f44d..8aaffa19b29a 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -114,7 +114,7 @@ static unsigned reserved_bio_based_ios = RESERVED_BIO_BASED_IOS;
 
 static int __dm_get_module_param_int(int *module_param, int min, int max)
 {
-	int param = ACCESS_ONCE(*module_param);
+	int param = READ_ONCE(*module_param);
 	int modified_param = 0;
 	bool modified = true;
 
@@ -136,7 +136,7 @@ static int __dm_get_module_param_int(int *module_param, int min, int max)
 unsigned __dm_get_module_param(unsigned *module_param,
 			       unsigned def, unsigned max)
 {
-	unsigned param = ACCESS_ONCE(*module_param);
+	unsigned param = READ_ONCE(*module_param);
 	unsigned modified_param = 0;
 
 	if (!param)
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 0ff1bbf6c90e..447ddcbc9566 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -2651,7 +2651,7 @@ state_show(struct md_rdev *rdev, char *page)
 {
 	char *sep = ",";
 	size_t len = 0;
-	unsigned long flags = ACCESS_ONCE(rdev->flags);
+	unsigned long flags = READ_ONCE(rdev->flags);
 
 	if (test_bit(Faulty, &flags) ||
 	    (!test_bit(ExternalBbl, &flags) &&
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 928e24a07133..7d9a50eed9db 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -6072,7 +6072,7 @@ static inline sector_t raid5_sync_request(struct mddev *mddev, sector_t sector_n
 	 */
 	rcu_read_lock();
 	for (i = 0; i < conf->raid_disks; i++) {
-		struct md_rdev *rdev = ACCESS_ONCE(conf->disks[i].rdev);
+		struct md_rdev *rdev = READ_ONCE(conf->disks[i].rdev);
 
 		if (rdev == NULL || test_bit(Faulty, &rdev->flags))
 			still_degraded = 1;
diff --git a/drivers/misc/mic/scif/scif_rb.c b/drivers/misc/mic/scif/scif_rb.c
index 637cc4686742..b665757ca89a 100644
--- a/drivers/misc/mic/scif/scif_rb.c
+++ b/drivers/misc/mic/scif/scif_rb.c
@@ -138,7 +138,7 @@ void scif_rb_commit(struct scif_rb *rb)
 	 * the read barrier in scif_rb_count(..)
 	 */
 	wmb();
-	ACCESS_ONCE(*rb->write_ptr) = rb->current_write_offset;
+	WRITE_ONCE(*rb->write_ptr, rb->current_write_offset);
 #ifdef CONFIG_INTEL_MIC_CARD
 	/*
 	 * X100 Si bug: For the case where a Core is performing an EXT_WR
@@ -147,7 +147,7 @@ void scif_rb_commit(struct scif_rb *rb)
 	 * This way, if ordering is violated for the Interrupt Message, it will
 	 * fall just behind the first Posted associated with the first EXT_WR.
 	 */
-	ACCESS_ONCE(*rb->write_ptr) = rb->current_write_offset;
+	WRITE_ONCE(*rb->write_ptr, rb->current_write_offset);
 #endif
 }
 
@@ -210,7 +210,7 @@ void scif_rb_update_read_ptr(struct scif_rb *rb)
 	 * scif_rb_space(..)
 	 */
 	mb();
-	ACCESS_ONCE(*rb->read_ptr) = new_offset;
+	WRITE_ONCE(*rb->read_ptr, new_offset);
 #ifdef CONFIG_INTEL_MIC_CARD
 	/*
 	 * X100 Si Bug: For the case where a Core is performing an EXT_WR
@@ -219,7 +219,7 @@ void scif_rb_update_read_ptr(struct scif_rb *rb)
 	 * This way, if ordering is violated for the Interrupt Message, it will
 	 * fall just behind the first Posted associated with the first EXT_WR.
 	 */
-	ACCESS_ONCE(*rb->read_ptr) = new_offset;
+	WRITE_ONCE(*rb->read_ptr, new_offset);
 #endif
 }
 
diff --git a/drivers/misc/mic/scif/scif_rma_list.c b/drivers/misc/mic/scif/scif_rma_list.c
index e1ef8daedd5a..a036dbb4101e 100644
--- a/drivers/misc/mic/scif/scif_rma_list.c
+++ b/drivers/misc/mic/scif/scif_rma_list.c
@@ -277,7 +277,7 @@ int scif_unregister_all_windows(scif_epd_t epd)
 		 * Need to restart list traversal if there has been
 		 * an asynchronous list entry deletion.
 		 */
-		if (ACCESS_ONCE(ep->rma_info.async_list_del))
+		if (READ_ONCE(ep->rma_info.async_list_del))
 			goto retry;
 	}
 	mutex_unlock(&ep->rma_info.rma_lock);
diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
index c02cc817a490..1ed9529e7bd1 100644
--- a/drivers/net/bonding/bond_alb.c
+++ b/drivers/net/bonding/bond_alb.c
@@ -1378,7 +1378,7 @@ int bond_tlb_xmit(struct sk_buff *skb, struct net_device *bond_dev)
 				unsigned int count;
 
 				slaves = rcu_dereference(bond->slave_arr);
-				count = slaves ? ACCESS_ONCE(slaves->count) : 0;
+				count = slaves ? READ_ONCE(slaves->count) : 0;
 				if (likely(count))
 					tx_slave = slaves->arr[hash_index %
 							       count];
diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index c99dc59d729b..af51b90cecbb 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -1167,7 +1167,7 @@ static rx_handler_result_t bond_handle_frame(struct sk_buff **pskb)
 	slave = bond_slave_get_rcu(skb->dev);
 	bond = slave->bond;
 
-	recv_probe = ACCESS_ONCE(bond->recv_probe);
+	recv_probe = READ_ONCE(bond->recv_probe);
 	if (recv_probe) {
 		ret = recv_probe(skb, bond, slave);
 		if (ret == RX_HANDLER_CONSUMED) {
@@ -3810,7 +3810,7 @@ static int bond_xmit_roundrobin(struct sk_buff *skb, struct net_device *bond_dev
 		else
 			bond_xmit_slave_id(bond, skb, 0);
 	} else {
-		int slave_cnt = ACCESS_ONCE(bond->slave_cnt);
+		int slave_cnt = READ_ONCE(bond->slave_cnt);
 
 		if (likely(slave_cnt)) {
 			slave_id = bond_rr_gen_slave_id(bond);
@@ -3972,7 +3972,7 @@ static int bond_3ad_xor_xmit(struct sk_buff *skb, struct net_device *dev)
 	unsigned int count;
 
 	slaves = rcu_dereference(bond->slave_arr);
-	count = slaves ? ACCESS_ONCE(slaves->count) : 0;
+	count = slaves ? READ_ONCE(slaves->count) : 0;
 	if (likely(count)) {
 		slave = slaves->arr[bond_xmit_hash(bond, skb) % count];
 		bond_dev_queue_xmit(bond, skb, slave->dev);
diff --git a/drivers/net/ethernet/chelsio/cxgb4/sge.c b/drivers/net/ethernet/chelsio/cxgb4/sge.c
index 4ef68f69b58c..43f52a8fe708 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/sge.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/sge.c
@@ -405,7 +405,7 @@ void free_tx_desc(struct adapter *adap, struct sge_txq *q,
  */
 static inline int reclaimable(const struct sge_txq *q)
 {
-	int hw_cidx = ntohs(ACCESS_ONCE(q->stat->cidx));
+	int hw_cidx = ntohs(READ_ONCE(q->stat->cidx));
 	hw_cidx -= q->cidx;
 	return hw_cidx < 0 ? hw_cidx + q->size : hw_cidx;
 }
@@ -1375,7 +1375,7 @@ out_free:	dev_kfree_skb_any(skb);
  */
 static inline void reclaim_completed_tx_imm(struct sge_txq *q)
 {
-	int hw_cidx = ntohs(ACCESS_ONCE(q->stat->cidx));
+	int hw_cidx = ntohs(READ_ONCE(q->stat->cidx));
 	int reclaim = hw_cidx - q->cidx;
 
 	if (reclaim < 0)
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 0e3d9f39a807..c6e859a27ee6 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -605,7 +605,7 @@ static void accumulate_16bit_val(u32 *acc, u16 val)
 
 	if (wrapped)
 		newacc += 65536;
-	ACCESS_ONCE(*acc) = newacc;
+	WRITE_ONCE(*acc, newacc);
 }
 
 static void populate_erx_stats(struct be_adapter *adapter,
diff --git a/drivers/net/ethernet/hisilicon/hip04_eth.c b/drivers/net/ethernet/hisilicon/hip04_eth.c
index 0cec06bec63e..340e28211135 100644
--- a/drivers/net/ethernet/hisilicon/hip04_eth.c
+++ b/drivers/net/ethernet/hisilicon/hip04_eth.c
@@ -373,7 +373,7 @@ static int hip04_tx_reclaim(struct net_device *ndev, bool force)
 	unsigned int count;
 
 	smp_rmb();
-	count = tx_count(ACCESS_ONCE(priv->tx_head), tx_tail);
+	count = tx_count(READ_ONCE(priv->tx_head), tx_tail);
 	if (count == 0)
 		goto out;
 
@@ -431,7 +431,7 @@ static int hip04_mac_start_xmit(struct sk_buff *skb, struct net_device *ndev)
 	dma_addr_t phys;
 
 	smp_rmb();
-	count = tx_count(tx_head, ACCESS_ONCE(priv->tx_tail));
+	count = tx_count(tx_head, READ_ONCE(priv->tx_tail));
 	if (count == (TX_DESC_NUM - 1)) {
 		netif_stop_queue(ndev);
 		return NETDEV_TX_BUSY;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
index 8f326f87a815..2cb9539c931e 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
@@ -264,7 +264,7 @@ static void i40e_dbg_dump_vsi_seid(struct i40e_pf *pf, int seid)
 		 vsi->rx_buf_failed, vsi->rx_page_failed);
 	rcu_read_lock();
 	for (i = 0; i < vsi->num_queue_pairs; i++) {
-		struct i40e_ring *rx_ring = ACCESS_ONCE(vsi->rx_rings[i]);
+		struct i40e_ring *rx_ring = READ_ONCE(vsi->rx_rings[i]);
 
 		if (!rx_ring)
 			continue;
@@ -320,7 +320,7 @@ static void i40e_dbg_dump_vsi_seid(struct i40e_pf *pf, int seid)
 			 ITR_IS_DYNAMIC(rx_ring->rx_itr_setting) ? "dynamic" : "fixed");
 	}
 	for (i = 0; i < vsi->num_queue_pairs; i++) {
-		struct i40e_ring *tx_ring = ACCESS_ONCE(vsi->tx_rings[i]);
+		struct i40e_ring *tx_ring = READ_ONCE(vsi->tx_rings[i]);
 
 		if (!tx_ring)
 			continue;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index 05e89864f781..e9e04a485e0a 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -1570,7 +1570,7 @@ static void i40e_get_ethtool_stats(struct net_device *netdev,
 	}
 	rcu_read_lock();
 	for (j = 0; j < vsi->num_queue_pairs; j++) {
-		tx_ring = ACCESS_ONCE(vsi->tx_rings[j]);
+		tx_ring = READ_ONCE(vsi->tx_rings[j]);
 
 		if (!tx_ring)
 			continue;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 6498da8806cb..de1fcac7834d 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -455,7 +455,7 @@ static void i40e_get_netdev_stats_struct(struct net_device *netdev,
 		u64 bytes, packets;
 		unsigned int start;
 
-		tx_ring = ACCESS_ONCE(vsi->tx_rings[i]);
+		tx_ring = READ_ONCE(vsi->tx_rings[i]);
 		if (!tx_ring)
 			continue;
 		i40e_get_netdev_stats_struct_tx(tx_ring, stats);
@@ -791,7 +791,7 @@ static void i40e_update_vsi_stats(struct i40e_vsi *vsi)
 	rcu_read_lock();
 	for (q = 0; q < vsi->num_queue_pairs; q++) {
 		/* locate Tx ring */
-		p = ACCESS_ONCE(vsi->tx_rings[q]);
+		p = READ_ONCE(vsi->tx_rings[q]);
 
 		do {
 			start = u64_stats_fetch_begin_irq(&p->syncp);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ptp.c b/drivers/net/ethernet/intel/i40e/i40e_ptp.c
index d8456c381c99..97381238eb7c 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ptp.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ptp.c
@@ -130,7 +130,7 @@ static int i40e_ptp_adjfreq(struct ptp_clock_info *ptp, s32 ppb)
 	}
 
 	smp_mb(); /* Force any pending update before accessing. */
-	adj = ACCESS_ONCE(pf->ptp_base_adj);
+	adj = READ_ONCE(pf->ptp_base_adj);
 
 	freq = adj;
 	freq *= ppb;
@@ -499,7 +499,7 @@ void i40e_ptp_set_increment(struct i40e_pf *pf)
 	wr32(hw, I40E_PRTTSYN_INC_H, incval >> 32);
 
 	/* Update the base adjustement value. */
-	ACCESS_ONCE(pf->ptp_base_adj) = incval;
+	WRITE_ONCE(pf->ptp_base_adj, incval);
 	smp_mb(); /* Force the above update. */
 }
 
diff --git a/drivers/net/ethernet/intel/igb/e1000_regs.h b/drivers/net/ethernet/intel/igb/e1000_regs.h
index 58adbf234e07..31a3f09df9f7 100644
--- a/drivers/net/ethernet/intel/igb/e1000_regs.h
+++ b/drivers/net/ethernet/intel/igb/e1000_regs.h
@@ -375,7 +375,7 @@ u32 igb_rd32(struct e1000_hw *hw, u32 reg);
 /* write operations, indexed using DWORDS */
 #define wr32(reg, val) \
 do { \
-	u8 __iomem *hw_addr = ACCESS_ONCE((hw)->hw_addr); \
+	u8 __iomem *hw_addr = READ_ONCE((hw)->hw_addr); \
 	if (!E1000_REMOVED(hw_addr)) \
 		writel((val), &hw_addr[(reg)]); \
 } while (0)
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index fd4a46b03cc8..6bccc2be2b91 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -750,7 +750,7 @@ static void igb_cache_ring_register(struct igb_adapter *adapter)
 u32 igb_rd32(struct e1000_hw *hw, u32 reg)
 {
 	struct igb_adapter *igb = container_of(hw, struct igb_adapter, hw);
-	u8 __iomem *hw_addr = ACCESS_ONCE(hw->hw_addr);
+	u8 __iomem *hw_addr = READ_ONCE(hw->hw_addr);
 	u32 value = 0;
 
 	if (E1000_REMOVED(hw_addr))
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.h
index e083732adf64..a01409e2e06c 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.h
@@ -161,7 +161,7 @@ static inline bool ixgbe_removed(void __iomem *addr)
 
 static inline void ixgbe_write_reg(struct ixgbe_hw *hw, u32 reg, u32 value)
 {
-	u8 __iomem *reg_addr = ACCESS_ONCE(hw->hw_addr);
+	u8 __iomem *reg_addr = READ_ONCE(hw->hw_addr);
 
 	if (ixgbe_removed(reg_addr))
 		return;
@@ -180,7 +180,7 @@ static inline void writeq(u64 val, void __iomem *addr)
 
 static inline void ixgbe_write_reg64(struct ixgbe_hw *hw, u32 reg, u64 value)
 {
-	u8 __iomem *reg_addr = ACCESS_ONCE(hw->hw_addr);
+	u8 __iomem *reg_addr = READ_ONCE(hw->hw_addr);
 
 	if (ixgbe_removed(reg_addr))
 		return;
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index d962368d08d0..df2c8ecafcf4 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -380,7 +380,7 @@ static void ixgbe_check_remove(struct ixgbe_hw *hw, u32 reg)
  */
 u32 ixgbe_read_reg(struct ixgbe_hw *hw, u32 reg)
 {
-	u8 __iomem *reg_addr = ACCESS_ONCE(hw->hw_addr);
+	u8 __iomem *reg_addr = READ_ONCE(hw->hw_addr);
 	u32 value;
 
 	if (ixgbe_removed(reg_addr))
@@ -8626,7 +8626,7 @@ static void ixgbe_get_stats64(struct net_device *netdev,
 
 	rcu_read_lock();
 	for (i = 0; i < adapter->num_rx_queues; i++) {
-		struct ixgbe_ring *ring = ACCESS_ONCE(adapter->rx_ring[i]);
+		struct ixgbe_ring *ring = READ_ONCE(adapter->rx_ring[i]);
 		u64 bytes, packets;
 		unsigned int start;
 
@@ -8642,12 +8642,12 @@ static void ixgbe_get_stats64(struct net_device *netdev,
 	}
 
 	for (i = 0; i < adapter->num_tx_queues; i++) {
-		struct ixgbe_ring *ring = ACCESS_ONCE(adapter->tx_ring[i]);
+		struct ixgbe_ring *ring = READ_ONCE(adapter->tx_ring[i]);
 
 		ixgbe_get_ring_stats64(stats, ring);
 	}
 	for (i = 0; i < adapter->num_xdp_queues; i++) {
-		struct ixgbe_ring *ring = ACCESS_ONCE(adapter->xdp_ring[i]);
+		struct ixgbe_ring *ring = READ_ONCE(adapter->xdp_ring[i]);
 
 		ixgbe_get_ring_stats64(stats, ring);
 	}
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
index 86d6924a2b71..ae312c45696a 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
@@ -378,7 +378,7 @@ static int ixgbe_ptp_adjfreq_82599(struct ptp_clock_info *ptp, s32 ppb)
 	}
 
 	smp_mb();
-	incval = ACCESS_ONCE(adapter->base_incval);
+	incval = READ_ONCE(adapter->base_incval);
 
 	freq = incval;
 	freq *= ppb;
@@ -1159,7 +1159,7 @@ void ixgbe_ptp_start_cyclecounter(struct ixgbe_adapter *adapter)
 	}
 
 	/* update the base incval used to calculate frequency adjustment */
-	ACCESS_ONCE(adapter->base_incval) = incval;
+	WRITE_ONCE(adapter->base_incval, incval);
 	smp_mb();
 
 	/* need lock to prevent incorrect read while modifying cyclecounter */
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 032f8ac06357..cacb30682434 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -164,7 +164,7 @@ static void ixgbevf_check_remove(struct ixgbe_hw *hw, u32 reg)
 
 u32 ixgbevf_read_reg(struct ixgbe_hw *hw, u32 reg)
 {
-	u8 __iomem *reg_addr = ACCESS_ONCE(hw->hw_addr);
+	u8 __iomem *reg_addr = READ_ONCE(hw->hw_addr);
 	u32 value;
 
 	if (IXGBE_REMOVED(reg_addr))
diff --git a/drivers/net/ethernet/intel/ixgbevf/vf.h b/drivers/net/ethernet/intel/ixgbevf/vf.h
index 04d8d4ee4f04..c651fefcc3d2 100644
--- a/drivers/net/ethernet/intel/ixgbevf/vf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/vf.h
@@ -182,7 +182,7 @@ struct ixgbevf_info {
 
 static inline void ixgbe_write_reg(struct ixgbe_hw *hw, u32 reg, u32 value)
 {
-	u8 __iomem *reg_addr = ACCESS_ONCE(hw->hw_addr);
+	u8 __iomem *reg_addr = READ_ONCE(hw->hw_addr);
 
 	if (IXGBE_REMOVED(reg_addr))
 		return;
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index 8a32a8f7f9c0..3541a7f9d12e 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -414,8 +414,8 @@ bool mlx4_en_process_tx_cq(struct net_device *dev,
 
 	index = cons_index & size_mask;
 	cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor;
-	last_nr_txbb = ACCESS_ONCE(ring->last_nr_txbb);
-	ring_cons = ACCESS_ONCE(ring->cons);
+	last_nr_txbb = READ_ONCE(ring->last_nr_txbb);
+	ring_cons = READ_ONCE(ring->cons);
 	ring_index = ring_cons & size_mask;
 	stamp_index = ring_index;
 
@@ -479,8 +479,8 @@ bool mlx4_en_process_tx_cq(struct net_device *dev,
 	wmb();
 
 	/* we want to dirty this cache line once */
-	ACCESS_ONCE(ring->last_nr_txbb) = last_nr_txbb;
-	ACCESS_ONCE(ring->cons) = ring_cons + txbbs_skipped;
+	WRITE_ONCE(ring->last_nr_txbb, last_nr_txbb);
+	WRITE_ONCE(ring->cons, ring_cons + txbbs_skipped);
 
 	if (cq->type == TX_XDP)
 		return done < budget;
@@ -858,7 +858,7 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
 		goto tx_drop;
 
 	/* fetch ring->cons far ahead before needing it to avoid stall */
-	ring_cons = ACCESS_ONCE(ring->cons);
+	ring_cons = READ_ONCE(ring->cons);
 
 	real_size = get_real_size(skb, shinfo, dev, &lso_header_size,
 				  &inline_ok, &fragptr);
@@ -1066,7 +1066,7 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
 		 */
 		smp_rmb();
 
-		ring_cons = ACCESS_ONCE(ring->cons);
+		ring_cons = READ_ONCE(ring->cons);
 		if (unlikely(!mlx4_en_is_tx_ring_full(ring))) {
 			netif_tx_wake_queue(ring->tx_queue);
 			ring->wake_queue++;
diff --git a/drivers/net/ethernet/neterion/vxge/vxge-main.c b/drivers/net/ethernet/neterion/vxge/vxge-main.c
index 50ea69d88480..5dd5f61e1114 100644
--- a/drivers/net/ethernet/neterion/vxge/vxge-main.c
+++ b/drivers/net/ethernet/neterion/vxge/vxge-main.c
@@ -2629,7 +2629,7 @@ static void vxge_poll_vp_lockup(unsigned long data)
 		ring = &vdev->vpaths[i].ring;
 
 		/* Truncated to machine word size number of frames */
-		rx_frms = ACCESS_ONCE(ring->stats.rx_frms);
+		rx_frms = READ_ONCE(ring->stats.rx_frms);
 
 		/* Did this vpath received any packets */
 		if (ring->stats.prev_rx_frms == rx_frms) {
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 13f72f5b18d2..a95a46bcd339 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -2073,7 +2073,7 @@ static irqreturn_t efx_ef10_msi_interrupt(int irq, void *dev_id)
 	netif_vdbg(efx, intr, efx->net_dev,
 		   "IRQ %d on CPU %d\n", irq, raw_smp_processor_id());
 
-	if (likely(ACCESS_ONCE(efx->irq_soft_enabled))) {
+	if (likely(READ_ONCE(efx->irq_soft_enabled))) {
 		/* Note test interrupts */
 		if (context->index == efx->irq_level)
 			efx->last_irq_cpu = raw_smp_processor_id();
@@ -2088,7 +2088,7 @@ static irqreturn_t efx_ef10_msi_interrupt(int irq, void *dev_id)
 static irqreturn_t efx_ef10_legacy_interrupt(int irq, void *dev_id)
 {
 	struct efx_nic *efx = dev_id;
-	bool soft_enabled = ACCESS_ONCE(efx->irq_soft_enabled);
+	bool soft_enabled = READ_ONCE(efx->irq_soft_enabled);
 	struct efx_channel *channel;
 	efx_dword_t reg;
 	u32 queues;
@@ -3291,7 +3291,7 @@ static int efx_ef10_handle_rx_event(struct efx_channel *channel,
 	bool rx_cont;
 	u16 flags = 0;
 
-	if (unlikely(ACCESS_ONCE(efx->reset_pending)))
+	if (unlikely(READ_ONCE(efx->reset_pending)))
 		return 0;
 
 	/* Basic packet information */
@@ -3428,7 +3428,7 @@ efx_ef10_handle_tx_event(struct efx_channel *channel, efx_qword_t *event)
 	unsigned int tx_ev_q_label;
 	int tx_descs = 0;
 
-	if (unlikely(ACCESS_ONCE(efx->reset_pending)))
+	if (unlikely(READ_ONCE(efx->reset_pending)))
 		return 0;
 
 	if (unlikely(EFX_QWORD_FIELD(*event, ESF_DZ_TX_DROP_EVENT)))
@@ -5316,7 +5316,7 @@ static void efx_ef10_filter_remove_old(struct efx_nic *efx)
 	int i;
 
 	for (i = 0; i < HUNT_FILTER_TBL_ROWS; i++) {
-		if (ACCESS_ONCE(table->entry[i].spec) &
+		if (READ_ONCE(table->entry[i].spec) &
 		    EFX_EF10_FILTER_FLAG_AUTO_OLD) {
 			rc = efx_ef10_filter_remove_internal(efx,
 					1U << EFX_FILTER_PRI_AUTO, i, true);
diff --git a/drivers/net/ethernet/sfc/efx.c b/drivers/net/ethernet/sfc/efx.c
index b9cb697b2818..016616a63880 100644
--- a/drivers/net/ethernet/sfc/efx.c
+++ b/drivers/net/ethernet/sfc/efx.c
@@ -2809,7 +2809,7 @@ static void efx_reset_work(struct work_struct *data)
 	unsigned long pending;
 	enum reset_type method;
 
-	pending = ACCESS_ONCE(efx->reset_pending);
+	pending = READ_ONCE(efx->reset_pending);
 	method = fls(pending) - 1;
 
 	if (method == RESET_TYPE_MC_BIST)
@@ -2874,7 +2874,7 @@ void efx_schedule_reset(struct efx_nic *efx, enum reset_type type)
 	/* If we're not READY then just leave the flags set as the cue
 	 * to abort probing or reschedule the reset later.
 	 */
-	if (ACCESS_ONCE(efx->state) != STATE_READY)
+	if (READ_ONCE(efx->state) != STATE_READY)
 		return;
 
 	/* efx_process_channel() will no longer read events once a
diff --git a/drivers/net/ethernet/sfc/falcon/efx.c b/drivers/net/ethernet/sfc/falcon/efx.c
index 29614da91cbf..7263275fde4a 100644
--- a/drivers/net/ethernet/sfc/falcon/efx.c
+++ b/drivers/net/ethernet/sfc/falcon/efx.c
@@ -2545,7 +2545,7 @@ static void ef4_reset_work(struct work_struct *data)
 	unsigned long pending;
 	enum reset_type method;
 
-	pending = ACCESS_ONCE(efx->reset_pending);
+	pending = READ_ONCE(efx->reset_pending);
 	method = fls(pending) - 1;
 
 	if ((method == RESET_TYPE_RECOVER_OR_DISABLE ||
@@ -2605,7 +2605,7 @@ void ef4_schedule_reset(struct ef4_nic *efx, enum reset_type type)
 	/* If we're not READY then just leave the flags set as the cue
 	 * to abort probing or reschedule the reset later.
 	 */
-	if (ACCESS_ONCE(efx->state) != STATE_READY)
+	if (READ_ONCE(efx->state) != STATE_READY)
 		return;
 
 	queue_work(reset_workqueue, &efx->reset_work);
diff --git a/drivers/net/ethernet/sfc/falcon/falcon.c b/drivers/net/ethernet/sfc/falcon/falcon.c
index 93c713c1f627..cd8bb472d758 100644
--- a/drivers/net/ethernet/sfc/falcon/falcon.c
+++ b/drivers/net/ethernet/sfc/falcon/falcon.c
@@ -452,7 +452,7 @@ static irqreturn_t falcon_legacy_interrupt_a1(int irq, void *dev_id)
 		   "IRQ %d on CPU %d status " EF4_OWORD_FMT "\n",
 		   irq, raw_smp_processor_id(), EF4_OWORD_VAL(*int_ker));
 
-	if (!likely(ACCESS_ONCE(efx->irq_soft_enabled)))
+	if (!likely(READ_ONCE(efx->irq_soft_enabled)))
 		return IRQ_HANDLED;
 
 	/* Check to see if we have a serious error condition */
@@ -1372,7 +1372,7 @@ static void falcon_reconfigure_mac_wrapper(struct ef4_nic *efx)
 	ef4_oword_t reg;
 	int link_speed, isolate;
 
-	isolate = !!ACCESS_ONCE(efx->reset_pending);
+	isolate = !!READ_ONCE(efx->reset_pending);
 
 	switch (link_state->speed) {
 	case 10000: link_speed = 3; break;
diff --git a/drivers/net/ethernet/sfc/falcon/farch.c b/drivers/net/ethernet/sfc/falcon/farch.c
index 05916c710d8c..494884f6af4a 100644
--- a/drivers/net/ethernet/sfc/falcon/farch.c
+++ b/drivers/net/ethernet/sfc/falcon/farch.c
@@ -834,7 +834,7 @@ ef4_farch_handle_tx_event(struct ef4_channel *channel, ef4_qword_t *event)
 	struct ef4_nic *efx = channel->efx;
 	int tx_packets = 0;
 
-	if (unlikely(ACCESS_ONCE(efx->reset_pending)))
+	if (unlikely(READ_ONCE(efx->reset_pending)))
 		return 0;
 
 	if (likely(EF4_QWORD_FIELD(*event, FSF_AZ_TX_EV_COMP))) {
@@ -990,7 +990,7 @@ ef4_farch_handle_rx_event(struct ef4_channel *channel, const ef4_qword_t *event)
 	struct ef4_rx_queue *rx_queue;
 	struct ef4_nic *efx = channel->efx;
 
-	if (unlikely(ACCESS_ONCE(efx->reset_pending)))
+	if (unlikely(READ_ONCE(efx->reset_pending)))
 		return;
 
 	rx_ev_cont = EF4_QWORD_FIELD(*event, FSF_AZ_RX_EV_JUMBO_CONT);
@@ -1504,7 +1504,7 @@ irqreturn_t ef4_farch_fatal_interrupt(struct ef4_nic *efx)
 irqreturn_t ef4_farch_legacy_interrupt(int irq, void *dev_id)
 {
 	struct ef4_nic *efx = dev_id;
-	bool soft_enabled = ACCESS_ONCE(efx->irq_soft_enabled);
+	bool soft_enabled = READ_ONCE(efx->irq_soft_enabled);
 	ef4_oword_t *int_ker = efx->irq_status.addr;
 	irqreturn_t result = IRQ_NONE;
 	struct ef4_channel *channel;
@@ -1596,7 +1596,7 @@ irqreturn_t ef4_farch_msi_interrupt(int irq, void *dev_id)
 		   "IRQ %d on CPU %d status " EF4_OWORD_FMT "\n",
 		   irq, raw_smp_processor_id(), EF4_OWORD_VAL(*int_ker));
 
-	if (!likely(ACCESS_ONCE(efx->irq_soft_enabled)))
+	if (!likely(READ_ONCE(efx->irq_soft_enabled)))
 		return IRQ_HANDLED;
 
 	/* Handle non-event-queue sources */
diff --git a/drivers/net/ethernet/sfc/falcon/nic.h b/drivers/net/ethernet/sfc/falcon/nic.h
index a4c4592f6023..54ca457cdb15 100644
--- a/drivers/net/ethernet/sfc/falcon/nic.h
+++ b/drivers/net/ethernet/sfc/falcon/nic.h
@@ -83,7 +83,7 @@ static inline struct ef4_tx_queue *ef4_tx_queue_partner(struct ef4_tx_queue *tx_
 static inline bool __ef4_nic_tx_is_empty(struct ef4_tx_queue *tx_queue,
 					 unsigned int write_count)
 {
-	unsigned int empty_read_count = ACCESS_ONCE(tx_queue->empty_read_count);
+	unsigned int empty_read_count = READ_ONCE(tx_queue->empty_read_count);
 
 	if (empty_read_count == 0)
 		return false;
@@ -464,11 +464,11 @@ irqreturn_t ef4_farch_fatal_interrupt(struct ef4_nic *efx);
 
 static inline int ef4_nic_event_test_irq_cpu(struct ef4_channel *channel)
 {
-	return ACCESS_ONCE(channel->event_test_cpu);
+	return READ_ONCE(channel->event_test_cpu);
 }
 static inline int ef4_nic_irq_test_irq_cpu(struct ef4_nic *efx)
 {
-	return ACCESS_ONCE(efx->last_irq_cpu);
+	return READ_ONCE(efx->last_irq_cpu);
 }
 
 /* Global Resources */
diff --git a/drivers/net/ethernet/sfc/falcon/tx.c b/drivers/net/ethernet/sfc/falcon/tx.c
index 6a75f4140a4b..6486814e97dc 100644
--- a/drivers/net/ethernet/sfc/falcon/tx.c
+++ b/drivers/net/ethernet/sfc/falcon/tx.c
@@ -134,8 +134,8 @@ static void ef4_tx_maybe_stop_queue(struct ef4_tx_queue *txq1)
 	 */
 	netif_tx_stop_queue(txq1->core_txq);
 	smp_mb();
-	txq1->old_read_count = ACCESS_ONCE(txq1->read_count);
-	txq2->old_read_count = ACCESS_ONCE(txq2->read_count);
+	txq1->old_read_count = READ_ONCE(txq1->read_count);
+	txq2->old_read_count = READ_ONCE(txq2->read_count);
 
 	fill_level = max(txq1->insert_count - txq1->old_read_count,
 			 txq2->insert_count - txq2->old_read_count);
@@ -524,7 +524,7 @@ void ef4_xmit_done(struct ef4_tx_queue *tx_queue, unsigned int index)
 
 	/* Check whether the hardware queue is now empty */
 	if ((int)(tx_queue->read_count - tx_queue->old_write_count) >= 0) {
-		tx_queue->old_write_count = ACCESS_ONCE(tx_queue->write_count);
+		tx_queue->old_write_count = READ_ONCE(tx_queue->write_count);
 		if (tx_queue->read_count == tx_queue->old_write_count) {
 			smp_mb();
 			tx_queue->empty_read_count =
diff --git a/drivers/net/ethernet/sfc/farch.c b/drivers/net/ethernet/sfc/farch.c
index ba45150f53c7..86454d25a405 100644
--- a/drivers/net/ethernet/sfc/farch.c
+++ b/drivers/net/ethernet/sfc/farch.c
@@ -827,7 +827,7 @@ efx_farch_handle_tx_event(struct efx_channel *channel, efx_qword_t *event)
 	struct efx_nic *efx = channel->efx;
 	int tx_packets = 0;
 
-	if (unlikely(ACCESS_ONCE(efx->reset_pending)))
+	if (unlikely(READ_ONCE(efx->reset_pending)))
 		return 0;
 
 	if (likely(EFX_QWORD_FIELD(*event, FSF_AZ_TX_EV_COMP))) {
@@ -979,7 +979,7 @@ efx_farch_handle_rx_event(struct efx_channel *channel, const efx_qword_t *event)
 	struct efx_rx_queue *rx_queue;
 	struct efx_nic *efx = channel->efx;
 
-	if (unlikely(ACCESS_ONCE(efx->reset_pending)))
+	if (unlikely(READ_ONCE(efx->reset_pending)))
 		return;
 
 	rx_ev_cont = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_JUMBO_CONT);
@@ -1520,7 +1520,7 @@ irqreturn_t efx_farch_fatal_interrupt(struct efx_nic *efx)
 irqreturn_t efx_farch_legacy_interrupt(int irq, void *dev_id)
 {
 	struct efx_nic *efx = dev_id;
-	bool soft_enabled = ACCESS_ONCE(efx->irq_soft_enabled);
+	bool soft_enabled = READ_ONCE(efx->irq_soft_enabled);
 	efx_oword_t *int_ker = efx->irq_status.addr;
 	irqreturn_t result = IRQ_NONE;
 	struct efx_channel *channel;
@@ -1612,7 +1612,7 @@ irqreturn_t efx_farch_msi_interrupt(int irq, void *dev_id)
 		   "IRQ %d on CPU %d status " EFX_OWORD_FMT "\n",
 		   irq, raw_smp_processor_id(), EFX_OWORD_VAL(*int_ker));
 
-	if (!likely(ACCESS_ONCE(efx->irq_soft_enabled)))
+	if (!likely(READ_ONCE(efx->irq_soft_enabled)))
 		return IRQ_HANDLED;
 
 	/* Handle non-event-queue sources */
diff --git a/drivers/net/ethernet/sfc/nic.h b/drivers/net/ethernet/sfc/nic.h
index 4d7fb8af880d..7b51b6371724 100644
--- a/drivers/net/ethernet/sfc/nic.h
+++ b/drivers/net/ethernet/sfc/nic.h
@@ -81,7 +81,7 @@ static struct efx_tx_queue *efx_tx_queue_partner(struct efx_tx_queue *tx_queue)
 static inline bool __efx_nic_tx_is_empty(struct efx_tx_queue *tx_queue,
 					 unsigned int write_count)
 {
-	unsigned int empty_read_count = ACCESS_ONCE(tx_queue->empty_read_count);
+	unsigned int empty_read_count = READ_ONCE(tx_queue->empty_read_count);
 
 	if (empty_read_count == 0)
 		return false;
@@ -617,11 +617,11 @@ irqreturn_t efx_farch_fatal_interrupt(struct efx_nic *efx);
 
 static inline int efx_nic_event_test_irq_cpu(struct efx_channel *channel)
 {
-	return ACCESS_ONCE(channel->event_test_cpu);
+	return READ_ONCE(channel->event_test_cpu);
 }
 static inline int efx_nic_irq_test_irq_cpu(struct efx_nic *efx)
 {
-	return ACCESS_ONCE(efx->last_irq_cpu);
+	return READ_ONCE(efx->last_irq_cpu);
 }
 
 /* Global Resources */
diff --git a/drivers/net/ethernet/sfc/ptp.c b/drivers/net/ethernet/sfc/ptp.c
index 60cdb97f58e2..56c2db398def 100644
--- a/drivers/net/ethernet/sfc/ptp.c
+++ b/drivers/net/ethernet/sfc/ptp.c
@@ -658,7 +658,7 @@ static void efx_ptp_send_times(struct efx_nic *efx,
 
 	/* Write host time for specified period or until MC is done */
 	while ((timespec64_compare(&now.ts_real, &limit) < 0) &&
-	       ACCESS_ONCE(*mc_running)) {
+	       READ_ONCE(*mc_running)) {
 		struct timespec64 update_time;
 		unsigned int host_time;
 
@@ -668,7 +668,7 @@ static void efx_ptp_send_times(struct efx_nic *efx,
 		do {
 			pps_get_ts(&now);
 		} while ((timespec64_compare(&now.ts_real, &update_time) < 0) &&
-			 ACCESS_ONCE(*mc_running));
+			 READ_ONCE(*mc_running));
 
 		/* Synchronise NIC with single word of time only */
 		host_time = (now.ts_real.tv_sec << MC_NANOSECOND_BITS |
@@ -832,14 +832,14 @@ static int efx_ptp_synchronize(struct efx_nic *efx, unsigned int num_readings)
 		       ptp->start.dma_addr);
 
 	/* Clear flag that signals MC ready */
-	ACCESS_ONCE(*start) = 0;
+	WRITE_ONCE(*start, 0);
 	rc = efx_mcdi_rpc_start(efx, MC_CMD_PTP, synch_buf,
 				MC_CMD_PTP_IN_SYNCHRONIZE_LEN);
 	EFX_WARN_ON_ONCE_PARANOID(rc);
 
 	/* Wait for start from MCDI (or timeout) */
 	timeout = jiffies + msecs_to_jiffies(MAX_SYNCHRONISE_WAIT_MS);
-	while (!ACCESS_ONCE(*start) && (time_before(jiffies, timeout))) {
+	while (!READ_ONCE(*start) && (time_before(jiffies, timeout))) {
 		udelay(20);	/* Usually start MCDI execution quickly */
 		loops++;
 	}
@@ -849,7 +849,7 @@ static int efx_ptp_synchronize(struct efx_nic *efx, unsigned int num_readings)
 	if (!time_before(jiffies, timeout))
 		++ptp->sync_timeouts;
 
-	if (ACCESS_ONCE(*start))
+	if (READ_ONCE(*start))
 		efx_ptp_send_times(efx, &last_time);
 
 	/* Collect results */
diff --git a/drivers/net/ethernet/sfc/tx.c b/drivers/net/ethernet/sfc/tx.c
index 32bf1fecf864..efb66ea21f27 100644
--- a/drivers/net/ethernet/sfc/tx.c
+++ b/drivers/net/ethernet/sfc/tx.c
@@ -136,8 +136,8 @@ static void efx_tx_maybe_stop_queue(struct efx_tx_queue *txq1)
 	 */
 	netif_tx_stop_queue(txq1->core_txq);
 	smp_mb();
-	txq1->old_read_count = ACCESS_ONCE(txq1->read_count);
-	txq2->old_read_count = ACCESS_ONCE(txq2->read_count);
+	txq1->old_read_count = READ_ONCE(txq1->read_count);
+	txq2->old_read_count = READ_ONCE(txq2->read_count);
 
 	fill_level = max(txq1->insert_count - txq1->old_read_count,
 			 txq2->insert_count - txq2->old_read_count);
@@ -752,7 +752,7 @@ void efx_xmit_done(struct efx_tx_queue *tx_queue, unsigned int index)
 
 	/* Check whether the hardware queue is now empty */
 	if ((int)(tx_queue->read_count - tx_queue->old_write_count) >= 0) {
-		tx_queue->old_write_count = ACCESS_ONCE(tx_queue->write_count);
+		tx_queue->old_write_count = READ_ONCE(tx_queue->write_count);
 		if (tx_queue->read_count == tx_queue->old_write_count) {
 			smp_mb();
 			tx_queue->empty_read_count =
diff --git a/drivers/net/ethernet/sun/niu.c b/drivers/net/ethernet/sun/niu.c
index 6a4e8e1bbd90..8ab0fb6892d5 100644
--- a/drivers/net/ethernet/sun/niu.c
+++ b/drivers/net/ethernet/sun/niu.c
@@ -6245,7 +6245,7 @@ static void niu_get_rx_stats(struct niu *np,
 
 	pkts = dropped = errors = bytes = 0;
 
-	rx_rings = ACCESS_ONCE(np->rx_rings);
+	rx_rings = READ_ONCE(np->rx_rings);
 	if (!rx_rings)
 		goto no_rings;
 
@@ -6276,7 +6276,7 @@ static void niu_get_tx_stats(struct niu *np,
 
 	pkts = errors = bytes = 0;
 
-	tx_rings = ACCESS_ONCE(np->tx_rings);
+	tx_rings = READ_ONCE(np->tx_rings);
 	if (!tx_rings)
 		goto no_rings;
 
diff --git a/drivers/net/tap.c b/drivers/net/tap.c
index 21b71ae947fd..b55b29b90b88 100644
--- a/drivers/net/tap.c
+++ b/drivers/net/tap.c
@@ -257,7 +257,7 @@ static struct tap_queue *tap_get_queue(struct tap_dev *tap,
 	 * and validate that the result isn't NULL - in case we are
 	 * racing against queue removal.
 	 */
-	int numvtaps = ACCESS_ONCE(tap->numvtaps);
+	int numvtaps = READ_ONCE(tap->numvtaps);
 	__u32 rxq;
 
 	if (!numvtaps)
diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index 5ce580f413b9..6aee2afb1f69 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -469,7 +469,7 @@ static u16 tun_select_queue(struct net_device *dev, struct sk_buff *skb,
 	u32 numqueues = 0;
 
 	rcu_read_lock();
-	numqueues = ACCESS_ONCE(tun->numqueues);
+	numqueues = READ_ONCE(tun->numqueues);
 
 	txq = __skb_get_hash_symmetric(skb);
 	if (txq) {
@@ -864,7 +864,7 @@ static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	rcu_read_lock();
 	tfile = rcu_dereference(tun->tfiles[txq]);
-	numqueues = ACCESS_ONCE(tun->numqueues);
+	numqueues = READ_ONCE(tun->numqueues);
 
 	/* Drop packet if interface is not attached */
 	if (txq >= numqueues)
diff --git a/drivers/net/wireless/ath/ath5k/desc.c b/drivers/net/wireless/ath/ath5k/desc.c
index bd8d4392d68b..80f75139495f 100644
--- a/drivers/net/wireless/ath/ath5k/desc.c
+++ b/drivers/net/wireless/ath/ath5k/desc.c
@@ -500,13 +500,13 @@ ath5k_hw_proc_4word_tx_status(struct ath5k_hw *ah,
 
 	tx_status = &desc->ud.ds_tx5212.tx_stat;
 
-	txstat1 = ACCESS_ONCE(tx_status->tx_status_1);
+	txstat1 = READ_ONCE(tx_status->tx_status_1);
 
 	/* No frame has been send or error */
 	if (unlikely(!(txstat1 & AR5K_DESC_TX_STATUS1_DONE)))
 		return -EINPROGRESS;
 
-	txstat0 = ACCESS_ONCE(tx_status->tx_status_0);
+	txstat0 = READ_ONCE(tx_status->tx_status_0);
 
 	/*
 	 * Get descriptor status
@@ -700,14 +700,14 @@ ath5k_hw_proc_5212_rx_status(struct ath5k_hw *ah,
 	u32 rxstat0, rxstat1;
 
 	rx_status = &desc->ud.ds_rx.rx_stat;
-	rxstat1 = ACCESS_ONCE(rx_status->rx_status_1);
+	rxstat1 = READ_ONCE(rx_status->rx_status_1);
 
 	/* No frame received / not ready */
 	if (unlikely(!(rxstat1 & AR5K_5212_RX_DESC_STATUS1_DONE)))
 		return -EINPROGRESS;
 
 	memset(rs, 0, sizeof(struct ath5k_rx_status));
-	rxstat0 = ACCESS_ONCE(rx_status->rx_status_0);
+	rxstat0 = READ_ONCE(rx_status->rx_status_0);
 
 	/*
 	 * Frame receive status
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
index 613caca7dc02..785a0f33b7e6 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
@@ -3628,7 +3628,7 @@ static void brcmf_sdio_dataworker(struct work_struct *work)
 
 	bus->dpc_running = true;
 	wmb();
-	while (ACCESS_ONCE(bus->dpc_triggered)) {
+	while (READ_ONCE(bus->dpc_triggered)) {
 		bus->dpc_triggered = false;
 		brcmf_sdio_dpc(bus);
 		bus->idlecount = 0;
diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/ops.c b/drivers/net/wireless/intel/iwlwifi/mvm/ops.c
index 231878969332..0f45f34e39d3 100644
--- a/drivers/net/wireless/intel/iwlwifi/mvm/ops.c
+++ b/drivers/net/wireless/intel/iwlwifi/mvm/ops.c
@@ -1118,7 +1118,7 @@ void iwl_mvm_set_hw_ctkill_state(struct iwl_mvm *mvm, bool state)
 static bool iwl_mvm_set_hw_rfkill_state(struct iwl_op_mode *op_mode, bool state)
 {
 	struct iwl_mvm *mvm = IWL_OP_MODE_GET_MVM(op_mode);
-	bool calibrating = ACCESS_ONCE(mvm->calibrating);
+	bool calibrating = READ_ONCE(mvm->calibrating);
 
 	if (state)
 		set_bit(IWL_MVM_STATUS_HW_RFKILL, &mvm->status);
diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/tx.c b/drivers/net/wireless/intel/iwlwifi/mvm/tx.c
index 6f2e2af23219..6e9d3289b9d0 100644
--- a/drivers/net/wireless/intel/iwlwifi/mvm/tx.c
+++ b/drivers/net/wireless/intel/iwlwifi/mvm/tx.c
@@ -652,7 +652,7 @@ int iwl_mvm_tx_skb_non_sta(struct iwl_mvm *mvm, struct sk_buff *skb)
 				return -1;
 		} else if (info.control.vif->type == NL80211_IFTYPE_STATION &&
 			   is_multicast_ether_addr(hdr->addr1)) {
-			u8 ap_sta_id = ACCESS_ONCE(mvmvif->ap_sta_id);
+			u8 ap_sta_id = READ_ONCE(mvmvif->ap_sta_id);
 
 			if (ap_sta_id != IWL_MVM_INVALID_STA)
 				sta_id = ap_sta_id;
@@ -700,7 +700,7 @@ static int iwl_mvm_tx_tso(struct iwl_mvm *mvm, struct sk_buff *skb,
 	snap_ip_tcp = 8 + skb_transport_header(skb) - skb_network_header(skb) +
 		tcp_hdrlen(skb);
 
-	dbg_max_amsdu_len = ACCESS_ONCE(mvm->max_amsdu_len);
+	dbg_max_amsdu_len = READ_ONCE(mvm->max_amsdu_len);
 
 	if (!sta->max_amsdu_len ||
 	    !ieee80211_is_data_qos(hdr->frame_control) ||
diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/rx.c b/drivers/net/wireless/intel/iwlwifi/pcie/rx.c
index a06b6612b658..f25ce3a1ea50 100644
--- a/drivers/net/wireless/intel/iwlwifi/pcie/rx.c
+++ b/drivers/net/wireless/intel/iwlwifi/pcie/rx.c
@@ -1247,7 +1247,7 @@ static void iwl_pcie_rx_handle(struct iwl_trans *trans, int queue)
 	spin_lock(&rxq->lock);
 	/* uCode's read index (stored in shared DRAM) indicates the last Rx
 	 * buffer that the driver may process (last buffer filled by ucode). */
-	r = le16_to_cpu(ACCESS_ONCE(rxq->rb_stts->closed_rb_num)) & 0x0FFF;
+	r = le16_to_cpu(READ_ONCE(rxq->rb_stts->closed_rb_num)) & 0x0FFF;
 	i = rxq->read;
 
 	/* W/A 9000 device step A0 wrap-around bug */
diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/trans.c b/drivers/net/wireless/intel/iwlwifi/pcie/trans.c
index 2e3e013ec95a..9ad3f4fe5894 100644
--- a/drivers/net/wireless/intel/iwlwifi/pcie/trans.c
+++ b/drivers/net/wireless/intel/iwlwifi/pcie/trans.c
@@ -2076,12 +2076,12 @@ static int iwl_trans_pcie_wait_txq_empty(struct iwl_trans *trans, int txq_idx)
 
 	IWL_DEBUG_TX_QUEUES(trans, "Emptying queue %d...\n", txq_idx);
 	txq = trans_pcie->txq[txq_idx];
-	wr_ptr = ACCESS_ONCE(txq->write_ptr);
+	wr_ptr = READ_ONCE(txq->write_ptr);
 
-	while (txq->read_ptr != ACCESS_ONCE(txq->write_ptr) &&
+	while (txq->read_ptr != READ_ONCE(txq->write_ptr) &&
 	       !time_after(jiffies,
 			   now + msecs_to_jiffies(IWL_FLUSH_WAIT_MS))) {
-		u8 write_ptr = ACCESS_ONCE(txq->write_ptr);
+		u8 write_ptr = READ_ONCE(txq->write_ptr);
 
 		if (WARN_ONCE(wr_ptr != write_ptr,
 			      "WR pointer moved while flushing %d -> %d\n",
@@ -2553,7 +2553,7 @@ static u32 iwl_trans_pcie_dump_rbs(struct iwl_trans *trans,
 
 	spin_lock(&rxq->lock);
 
-	r = le16_to_cpu(ACCESS_ONCE(rxq->rb_stts->closed_rb_num)) & 0x0FFF;
+	r = le16_to_cpu(READ_ONCE(rxq->rb_stts->closed_rb_num)) & 0x0FFF;
 
 	for (i = rxq->read, j = 0;
 	     i != r && j < allocated_rb_nums;
@@ -2814,7 +2814,7 @@ static struct iwl_trans_dump_data
 		/* Dump RBs is supported only for pre-9000 devices (1 queue) */
 		struct iwl_rxq *rxq = &trans_pcie->rxq[0];
 		/* RBs */
-		num_rbs = le16_to_cpu(ACCESS_ONCE(rxq->rb_stts->closed_rb_num))
+		num_rbs = le16_to_cpu(READ_ONCE(rxq->rb_stts->closed_rb_num))
 				      & 0x0FFF;
 		num_rbs = (num_rbs - rxq->read) & RX_QUEUE_MASK;
 		len += num_rbs * (sizeof(*data) +
diff --git a/drivers/net/wireless/mac80211_hwsim.c b/drivers/net/wireless/mac80211_hwsim.c
index 6467ffac9811..d2b3d6177a55 100644
--- a/drivers/net/wireless/mac80211_hwsim.c
+++ b/drivers/net/wireless/mac80211_hwsim.c
@@ -1380,7 +1380,7 @@ static void mac80211_hwsim_tx(struct ieee80211_hw *hw,
 	mac80211_hwsim_monitor_rx(hw, skb, channel);
 
 	/* wmediumd mode check */
-	_portid = ACCESS_ONCE(data->wmediumd);
+	_portid = READ_ONCE(data->wmediumd);
 
 	if (_portid)
 		return mac80211_hwsim_tx_frame_nl(hw, skb, _portid);
@@ -1477,7 +1477,7 @@ static void mac80211_hwsim_tx_frame(struct ieee80211_hw *hw,
 				    struct ieee80211_channel *chan)
 {
 	struct mac80211_hwsim_data *data = hw->priv;
-	u32 _pid = ACCESS_ONCE(data->wmediumd);
+	u32 _pid = READ_ONCE(data->wmediumd);
 
 	if (ieee80211_hw_check(hw, SUPPORTS_RC_TABLE)) {
 		struct ieee80211_tx_info *txi = IEEE80211_SKB_CB(skb);
diff --git a/drivers/scsi/qla2xxx/qla_target.c b/drivers/scsi/qla2xxx/qla_target.c
index f05cfc83c9c8..f946bf889015 100644
--- a/drivers/scsi/qla2xxx/qla_target.c
+++ b/drivers/scsi/qla2xxx/qla_target.c
@@ -996,7 +996,7 @@ static void qlt_free_session_done(struct work_struct *work)
 	if (logout_started) {
 		bool traced = false;
 
-		while (!ACCESS_ONCE(sess->logout_completed)) {
+		while (!READ_ONCE(sess->logout_completed)) {
 			if (!traced) {
 				ql_dbg(ql_dbg_tgt_mgt, vha, 0xf086,
 					"%s: waiting for sess %p logout\n",
diff --git a/drivers/target/target_core_user.c b/drivers/target/target_core_user.c
index 942d094269fb..9469695f5871 100644
--- a/drivers/target/target_core_user.c
+++ b/drivers/target/target_core_user.c
@@ -985,7 +985,7 @@ static unsigned int tcmu_handle_completions(struct tcmu_dev *udev)
 	mb = udev->mb_addr;
 	tcmu_flush_dcache_range(mb, sizeof(*mb));
 
-	while (udev->cmdr_last_cleaned != ACCESS_ONCE(mb->cmd_tail)) {
+	while (udev->cmdr_last_cleaned != READ_ONCE(mb->cmd_tail)) {
 
 		struct tcmu_cmd_entry *entry = (void *) mb + CMDR_OFF + udev->cmdr_last_cleaned;
 		struct tcmu_cmd *cmd;
diff --git a/drivers/usb/class/cdc-wdm.c b/drivers/usb/class/cdc-wdm.c
index 3e865dbf878c..fbaa2a90d25d 100644
--- a/drivers/usb/class/cdc-wdm.c
+++ b/drivers/usb/class/cdc-wdm.c
@@ -483,7 +483,7 @@ static ssize_t wdm_read
 	if (rv < 0)
 		return -ERESTARTSYS;
 
-	cntr = ACCESS_ONCE(desc->length);
+	cntr = READ_ONCE(desc->length);
 	if (cntr == 0) {
 		desc->read = 0;
 retry:
diff --git a/drivers/usb/core/devio.c b/drivers/usb/core/devio.c
index 4664e543cf2f..07cd2d245e65 100644
--- a/drivers/usb/core/devio.c
+++ b/drivers/usb/core/devio.c
@@ -150,7 +150,7 @@ static int usbfs_increase_memory_usage(u64 amount)
 {
 	u64 lim;
 
-	lim = ACCESS_ONCE(usbfs_memory_mb);
+	lim = READ_ONCE(usbfs_memory_mb);
 	lim <<= 20;
 
 	atomic64_add(amount, &usbfs_memory_usage);
diff --git a/drivers/usb/core/sysfs.c b/drivers/usb/core/sysfs.c
index d930bfda4010..58d59c5f8592 100644
--- a/drivers/usb/core/sysfs.c
+++ b/drivers/usb/core/sysfs.c
@@ -973,7 +973,7 @@ static ssize_t interface_show(struct device *dev, struct device_attribute *attr,
 	char *string;
 
 	intf = to_usb_interface(dev);
-	string = ACCESS_ONCE(intf->cur_altsetting->string);
+	string = READ_ONCE(intf->cur_altsetting->string);
 	if (!string)
 		return 0;
 	return sprintf(buf, "%s\n", string);
@@ -989,7 +989,7 @@ static ssize_t modalias_show(struct device *dev, struct device_attribute *attr,
 
 	intf = to_usb_interface(dev);
 	udev = interface_to_usbdev(intf);
-	alt = ACCESS_ONCE(intf->cur_altsetting);
+	alt = READ_ONCE(intf->cur_altsetting);
 
 	return sprintf(buf, "usb:v%04Xp%04Xd%04Xdc%02Xdsc%02Xdp%02X"
 			"ic%02Xisc%02Xip%02Xin%02X\n",
diff --git a/drivers/usb/gadget/udc/gr_udc.c b/drivers/usb/gadget/udc/gr_udc.c
index 1f9941145746..0b59fa50aa30 100644
--- a/drivers/usb/gadget/udc/gr_udc.c
+++ b/drivers/usb/gadget/udc/gr_udc.c
@@ -1261,7 +1261,7 @@ static int gr_handle_in_ep(struct gr_ep *ep)
 	if (!req->last_desc)
 		return 0;
 
-	if (ACCESS_ONCE(req->last_desc->ctrl) & GR_DESC_IN_CTRL_EN)
+	if (READ_ONCE(req->last_desc->ctrl) & GR_DESC_IN_CTRL_EN)
 		return 0; /* Not put in hardware buffers yet */
 
 	if (gr_read32(&ep->regs->epstat) & (GR_EPSTAT_B1 | GR_EPSTAT_B0))
@@ -1290,7 +1290,7 @@ static int gr_handle_out_ep(struct gr_ep *ep)
 	if (!req->curr_desc)
 		return 0;
 
-	ctrl = ACCESS_ONCE(req->curr_desc->ctrl);
+	ctrl = READ_ONCE(req->curr_desc->ctrl);
 	if (ctrl & GR_DESC_OUT_CTRL_EN)
 		return 0; /* Not received yet */
 
diff --git a/drivers/usb/host/ohci-hcd.c b/drivers/usb/host/ohci-hcd.c
index 44924824fa41..c86f89babd57 100644
--- a/drivers/usb/host/ohci-hcd.c
+++ b/drivers/usb/host/ohci-hcd.c
@@ -785,7 +785,7 @@ static void io_watchdog_func(unsigned long _ohci)
 		}
 
 		/* find the last TD processed by the controller. */
-		head = hc32_to_cpu(ohci, ACCESS_ONCE(ed->hwHeadP)) & TD_MASK;
+		head = hc32_to_cpu(ohci, READ_ONCE(ed->hwHeadP)) & TD_MASK;
 		td_start = td;
 		td_next = list_prepare_entry(td, &ed->td_list, td_list);
 		list_for_each_entry_continue(td_next, &ed->td_list, td_list) {
diff --git a/drivers/usb/host/uhci-hcd.h b/drivers/usb/host/uhci-hcd.h
index 91b22b2ea3aa..09a2a259941b 100644
--- a/drivers/usb/host/uhci-hcd.h
+++ b/drivers/usb/host/uhci-hcd.h
@@ -186,7 +186,7 @@ struct uhci_qh {
  * We need a special accessor for the element pointer because it is
  * subject to asynchronous updates by the controller.
  */
-#define qh_element(qh)		ACCESS_ONCE((qh)->element)
+#define qh_element(qh)		READ_ONCE((qh)->element)
 
 #define LINK_TO_QH(uhci, qh)	(UHCI_PTR_QH((uhci)) | \
 				cpu_to_hc32((uhci), (qh)->dma_handle))
@@ -274,7 +274,7 @@ struct uhci_td {
  * subject to asynchronous updates by the controller.
  */
 #define td_status(uhci, td)		hc32_to_cpu((uhci), \
-						ACCESS_ONCE((td)->status))
+						READ_ONCE((td)->status))
 
 #define LINK_TO_TD(uhci, td)		(cpu_to_hc32((uhci), (td)->dma_handle))
 
diff --git a/drivers/vfio/vfio.c b/drivers/vfio/vfio.c
index f5a86f651f38..2bc3705a99bd 100644
--- a/drivers/vfio/vfio.c
+++ b/drivers/vfio/vfio.c
@@ -665,7 +665,7 @@ static int vfio_dev_viable(struct device *dev, void *data)
 {
 	struct vfio_group *group = data;
 	struct vfio_device *device;
-	struct device_driver *drv = ACCESS_ONCE(dev->driver);
+	struct device_driver *drv = READ_ONCE(dev->driver);
 	struct vfio_unbound_dev *unbound;
 	int ret = -EINVAL;
 
diff --git a/drivers/vhost/scsi.c b/drivers/vhost/scsi.c
index 046f6d280af5..35e929f132e8 100644
--- a/drivers/vhost/scsi.c
+++ b/drivers/vhost/scsi.c
@@ -929,7 +929,7 @@ vhost_scsi_handle_vq(struct vhost_scsi *vs, struct vhost_virtqueue *vq)
 			continue;
 		}
 
-		tpg = ACCESS_ONCE(vs_tpg[*target]);
+		tpg = READ_ONCE(vs_tpg[*target]);
 		if (unlikely(!tpg)) {
 			/* Target does not exist, fail the request */
 			vhost_scsi_send_bad_target(vs, vq, head, out);
diff --git a/fs/aio.c b/fs/aio.c
index 5a2487217072..e6de7715228c 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -576,7 +576,7 @@ static int kiocb_cancel(struct aio_kiocb *kiocb)
 	 * actually has a cancel function, hence the cmpxchg()
 	 */
 
-	cancel = ACCESS_ONCE(kiocb->ki_cancel);
+	cancel = READ_ONCE(kiocb->ki_cancel);
 	do {
 		if (!cancel || cancel == KIOCB_CANCELLED)
 			return -EINVAL;
diff --git a/fs/buffer.c b/fs/buffer.c
index 170df856bdb9..32ce01f0f95f 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -1692,7 +1692,8 @@ static struct buffer_head *create_page_buffers(struct page *page, struct inode *
 	BUG_ON(!PageLocked(page));
 
 	if (!page_has_buffers(page))
-		create_empty_buffers(page, 1 << ACCESS_ONCE(inode->i_blkbits), b_state);
+		create_empty_buffers(page, 1 << READ_ONCE(inode->i_blkbits),
+				     b_state);
 	return page_buffers(page);
 }
 
diff --git a/fs/crypto/keyinfo.c b/fs/crypto/keyinfo.c
index 018c588c7ac3..768263c2d077 100644
--- a/fs/crypto/keyinfo.c
+++ b/fs/crypto/keyinfo.c
@@ -368,7 +368,7 @@ void fscrypt_put_encryption_info(struct inode *inode, struct fscrypt_info *ci)
 	struct fscrypt_info *prev;
 
 	if (ci == NULL)
-		ci = ACCESS_ONCE(inode->i_crypt_info);
+		ci = READ_ONCE(inode->i_crypt_info);
 	if (ci == NULL)
 		return;
 
diff --git a/fs/direct-io.c b/fs/direct-io.c
index 62cf812ed0e5..066cf268694f 100644
--- a/fs/direct-io.c
+++ b/fs/direct-io.c
@@ -1140,7 +1140,7 @@ do_blockdev_direct_IO(struct kiocb *iocb, struct inode *inode,
 		      get_block_t get_block, dio_iodone_t end_io,
 		      dio_submit_t submit_io, int flags)
 {
-	unsigned i_blkbits = ACCESS_ONCE(inode->i_blkbits);
+	unsigned i_blkbits = READ_ONCE(inode->i_blkbits);
 	unsigned blkbits = i_blkbits;
 	unsigned blocksize_mask = (1 << blkbits) - 1;
 	ssize_t retval = -EINVAL;
diff --git a/fs/exec.c b/fs/exec.c
index 5470d3c1892a..13331bc42454 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -1910,7 +1910,7 @@ void set_dumpable(struct mm_struct *mm, int value)
 		return;
 
 	do {
-		old = ACCESS_ONCE(mm->flags);
+		old = READ_ONCE(mm->flags);
 		new = (old & ~MMF_DUMPABLE_MASK) | value;
 	} while (cmpxchg(&mm->flags, old, new) != old);
 }
diff --git a/fs/fcntl.c b/fs/fcntl.c
index 448a1119f0be..57bf2964bb83 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -724,7 +724,7 @@ static void send_sigio_to_task(struct task_struct *p,
 	 * F_SETSIG can change ->signum lockless in parallel, make
 	 * sure we read it once and use the same value throughout.
 	 */
-	int signum = ACCESS_ONCE(fown->signum);
+	int signum = READ_ONCE(fown->signum);
 
 	if (!sigio_perm(p, fown, signum))
 		return;
diff --git a/fs/fs_pin.c b/fs/fs_pin.c
index e747b3d720ee..2d07f292b625 100644
--- a/fs/fs_pin.c
+++ b/fs/fs_pin.c
@@ -78,7 +78,7 @@ void mnt_pin_kill(struct mount *m)
 	while (1) {
 		struct hlist_node *p;
 		rcu_read_lock();
-		p = ACCESS_ONCE(m->mnt_pins.first);
+		p = READ_ONCE(m->mnt_pins.first);
 		if (!p) {
 			rcu_read_unlock();
 			break;
@@ -92,7 +92,7 @@ void group_pin_kill(struct hlist_head *p)
 	while (1) {
 		struct hlist_node *q;
 		rcu_read_lock();
-		q = ACCESS_ONCE(p->first);
+		q = READ_ONCE(p->first);
 		if (!q) {
 			rcu_read_unlock();
 			break;
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index 13c65dd2d37d..a42d89371748 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -33,7 +33,7 @@ static struct fuse_dev *fuse_get_dev(struct file *file)
 	 * Lockless access is OK, because file->private data is set
 	 * once during mount and is valid until the file is released.
 	 */
-	return ACCESS_ONCE(file->private_data);
+	return READ_ONCE(file->private_data);
 }
 
 static void fuse_request_init(struct fuse_req *req, struct page **pages,
diff --git a/fs/inode.c b/fs/inode.c
index d1e35b53bb23..fd401028a309 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -2090,7 +2090,7 @@ void inode_set_flags(struct inode *inode, unsigned int flags,
 
 	WARN_ON_ONCE(flags & ~mask);
 	do {
-		old_flags = ACCESS_ONCE(inode->i_flags);
+		old_flags = READ_ONCE(inode->i_flags);
 		new_flags = (old_flags & ~mask) | flags;
 	} while (unlikely(cmpxchg(&inode->i_flags, old_flags,
 				  new_flags) != old_flags));
diff --git a/fs/namei.c b/fs/namei.c
index c75ea03ca147..40a0f34bf990 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -1209,7 +1209,7 @@ static int follow_managed(struct path *path, struct nameidata *nd)
 	/* Given that we're not holding a lock here, we retain the value in a
 	 * local variable for each dentry as we look at it so that we don't see
 	 * the components of that value change under us */
-	while (managed = ACCESS_ONCE(path->dentry->d_flags),
+	while (managed = READ_ONCE(path->dentry->d_flags),
 	       managed &= DCACHE_MANAGED_DENTRY,
 	       unlikely(managed != 0)) {
 		/* Allow the filesystem to manage the transit without i_mutex
@@ -1394,7 +1394,7 @@ int follow_down(struct path *path)
 	unsigned managed;
 	int ret;
 
-	while (managed = ACCESS_ONCE(path->dentry->d_flags),
+	while (managed = READ_ONCE(path->dentry->d_flags),
 	       unlikely(managed & DCACHE_MANAGED_DENTRY)) {
 		/* Allow the filesystem to manage the transit without i_mutex
 		 * being held.
diff --git a/fs/namespace.c b/fs/namespace.c
index 3b601f115b6c..f65f2dc786cf 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -353,7 +353,7 @@ int __mnt_want_write(struct vfsmount *m)
 	 * incremented count after it has set MNT_WRITE_HOLD.
 	 */
 	smp_mb();
-	while (ACCESS_ONCE(mnt->mnt.mnt_flags) & MNT_WRITE_HOLD)
+	while (READ_ONCE(mnt->mnt.mnt_flags) & MNT_WRITE_HOLD)
 		cpu_relax();
 	/*
 	 * After the slowpath clears MNT_WRITE_HOLD, mnt_is_readonly will
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 5ceaeb1f6fb6..f439f1c45008 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1081,7 +1081,7 @@ static int nfs_lookup_revalidate(struct dentry *dentry, unsigned int flags)
 	int error;
 
 	if (flags & LOOKUP_RCU) {
-		parent = ACCESS_ONCE(dentry->d_parent);
+		parent = READ_ONCE(dentry->d_parent);
 		dir = d_inode_rcu(parent);
 		if (!dir)
 			return -ECHILD;
@@ -1168,7 +1168,7 @@ static int nfs_lookup_revalidate(struct dentry *dentry, unsigned int flags)
 	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
  out_valid:
 	if (flags & LOOKUP_RCU) {
-		if (parent != ACCESS_ONCE(dentry->d_parent))
+		if (parent != READ_ONCE(dentry->d_parent))
 			return -ECHILD;
 	} else
 		dput(parent);
@@ -1582,7 +1582,7 @@ static int nfs4_lookup_revalidate(struct dentry *dentry, unsigned int flags)
 		struct inode *dir;
 
 		if (flags & LOOKUP_RCU) {
-			parent = ACCESS_ONCE(dentry->d_parent);
+			parent = READ_ONCE(dentry->d_parent);
 			dir = d_inode_rcu(parent);
 			if (!dir)
 				return -ECHILD;
@@ -1596,7 +1596,7 @@ static int nfs4_lookup_revalidate(struct dentry *dentry, unsigned int flags)
 			ret = -ECHILD;
 		if (!(flags & LOOKUP_RCU))
 			dput(parent);
-		else if (parent != ACCESS_ONCE(dentry->d_parent))
+		else if (parent != READ_ONCE(dentry->d_parent))
 			return -ECHILD;
 		goto out;
 	}
diff --git a/fs/proc/array.c b/fs/proc/array.c
index 77a8eacbe032..375e8bf0dd24 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -453,7 +453,7 @@ static int do_task_stat(struct seq_file *m, struct pid_namespace *ns,
 		cutime = sig->cutime;
 		cstime = sig->cstime;
 		cgtime = sig->cgtime;
-		rsslim = ACCESS_ONCE(sig->rlim[RLIMIT_RSS].rlim_cur);
+		rsslim = READ_ONCE(sig->rlim[RLIMIT_RSS].rlim_cur);
 
 		/* add up live thread stats at the group level */
 		if (whole) {
diff --git a/fs/proc_namespace.c b/fs/proc_namespace.c
index 99dff222fe67..03afd5150916 100644
--- a/fs/proc_namespace.c
+++ b/fs/proc_namespace.c
@@ -27,7 +27,7 @@ static unsigned mounts_poll(struct file *file, poll_table *wait)
 
 	poll_wait(file, &p->ns->poll, wait);
 
-	event = ACCESS_ONCE(ns->event);
+	event = READ_ONCE(ns->event);
 	if (m->poll_event != event) {
 		m->poll_event = event;
 		res |= POLLERR | POLLPRI;
diff --git a/fs/splice.c b/fs/splice.c
index f3084cce0ea6..39e2dc01ac12 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -253,7 +253,7 @@ EXPORT_SYMBOL(add_to_pipe);
  */
 int splice_grow_spd(const struct pipe_inode_info *pipe, struct splice_pipe_desc *spd)
 {
-	unsigned int buffers = ACCESS_ONCE(pipe->buffers);
+	unsigned int buffers = READ_ONCE(pipe->buffers);
 
 	spd->nr_pages_max = buffers;
 	if (buffers <= PIPE_DEF_BUFFERS)
diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
index 1c713fd5b3e6..f46d133c0949 100644
--- a/fs/userfaultfd.c
+++ b/fs/userfaultfd.c
@@ -381,7 +381,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason)
 	 * in __get_user_pages if userfaultfd_release waits on the
 	 * caller of handle_userfault to release the mmap_sem.
 	 */
-	if (unlikely(ACCESS_ONCE(ctx->released))) {
+	if (unlikely(READ_ONCE(ctx->released))) {
 		/*
 		 * Don't return VM_FAULT_SIGBUS in this case, so a non
 		 * cooperative manager can close the uffd after the
@@ -477,7 +477,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason)
 						       vmf->flags, reason);
 	up_read(&mm->mmap_sem);
 
-	if (likely(must_wait && !ACCESS_ONCE(ctx->released) &&
+	if (likely(must_wait && !READ_ONCE(ctx->released) &&
 		   (return_to_userland ? !signal_pending(current) :
 		    !fatal_signal_pending(current)))) {
 		wake_up_poll(&ctx->fd_wqh, POLLIN);
@@ -586,7 +586,7 @@ static void userfaultfd_event_wait_completion(struct userfaultfd_ctx *ctx,
 		set_current_state(TASK_KILLABLE);
 		if (ewq->msg.event == 0)
 			break;
-		if (ACCESS_ONCE(ctx->released) ||
+		if (READ_ONCE(ctx->released) ||
 		    fatal_signal_pending(current)) {
 			/*
 			 * &ewq->wq may be queued in fork_event, but
@@ -833,7 +833,7 @@ static int userfaultfd_release(struct inode *inode, struct file *file)
 	struct userfaultfd_wake_range range = { .len = 0, };
 	unsigned long new_flags;
 
-	ACCESS_ONCE(ctx->released) = true;
+	WRITE_ONCE(ctx->released, true);
 
 	if (!mmget_not_zero(mm))
 		goto wakeup;
diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
index 51bf7b827387..129975970d99 100644
--- a/fs/xfs/xfs_log_priv.h
+++ b/fs/xfs/xfs_log_priv.h
@@ -592,9 +592,9 @@ xlog_valid_lsn(
 	 * a transiently forward state. Instead, we can see the LSN in a
 	 * transiently behind state if we happen to race with a cycle wrap.
 	 */
-	cur_cycle = ACCESS_ONCE(log->l_curr_cycle);
+	cur_cycle = READ_ONCE(log->l_curr_cycle);
 	smp_rmb();
-	cur_block = ACCESS_ONCE(log->l_curr_block);
+	cur_block = READ_ONCE(log->l_curr_block);
 
 	if ((CYCLE_LSN(lsn) > cur_cycle) ||
 	    (CYCLE_LSN(lsn) == cur_cycle && BLOCK_LSN(lsn) > cur_block)) {
diff --git a/include/linux/bitops.h b/include/linux/bitops.h
index 8fbe259b197c..0a7ce668f8e0 100644
--- a/include/linux/bitops.h
+++ b/include/linux/bitops.h
@@ -236,7 +236,7 @@ static inline unsigned long __ffs64(u64 word)
 	typeof(*ptr) old, new;					\
 								\
 	do {							\
-		old = ACCESS_ONCE(*ptr);			\
+		old = READ_ONCE(*ptr);			\
 		new = (old & ~mask) | bits;			\
 	} while (cmpxchg(ptr, old, new) != old);		\
 								\
@@ -251,7 +251,7 @@ static inline unsigned long __ffs64(u64 word)
 	typeof(*ptr) old, new;					\
 								\
 	do {							\
-		old = ACCESS_ONCE(*ptr);			\
+		old = READ_ONCE(*ptr);			\
 		new = old & ~clear;				\
 	} while (!(old & test) &&				\
 		 cmpxchg(ptr, old, new) != old);		\
diff --git a/include/linux/dynamic_queue_limits.h b/include/linux/dynamic_queue_limits.h
index a4be70398ce1..36dd4ffb5715 100644
--- a/include/linux/dynamic_queue_limits.h
+++ b/include/linux/dynamic_queue_limits.h
@@ -88,7 +88,7 @@ static inline void dql_queued(struct dql *dql, unsigned int count)
 /* Returns how many objects can be queued, < 0 indicates over limit. */
 static inline int dql_avail(const struct dql *dql)
 {
-	return ACCESS_ONCE(dql->adj_limit) - ACCESS_ONCE(dql->num_queued);
+	return READ_ONCE(dql->adj_limit) - READ_ONCE(dql->num_queued);
 }
 
 /* Record number of completed objects and recalculate the limit. */
diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index 14bc21c2ee7f..785a00ca4628 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -221,7 +221,7 @@ extern struct page *huge_zero_page;
 
 static inline bool is_huge_zero_page(struct page *page)
 {
-	return ACCESS_ONCE(huge_zero_page) == page;
+	return READ_ONCE(huge_zero_page) == page;
 }
 
 static inline bool is_huge_zero_pmd(pmd_t pmd)
diff --git a/include/linux/if_team.h b/include/linux/if_team.h
index 30294603526f..d95cae09dea0 100644
--- a/include/linux/if_team.h
+++ b/include/linux/if_team.h
@@ -247,7 +247,7 @@ static inline struct team_port *team_get_port_by_index(struct team *team,
 
 static inline int team_num_to_port_index(struct team *team, unsigned int num)
 {
-	int en_port_count = ACCESS_ONCE(team->en_port_count);
+	int en_port_count = READ_ONCE(team->en_port_count);
 
 	if (unlikely(!en_port_count))
 		return 0;
diff --git a/include/linux/llist.h b/include/linux/llist.h
index 1957635e6d5f..85abc2915e8d 100644
--- a/include/linux/llist.h
+++ b/include/linux/llist.h
@@ -198,7 +198,7 @@ static inline void init_llist_head(struct llist_head *list)
  */
 static inline bool llist_empty(const struct llist_head *head)
 {
-	return ACCESS_ONCE(head->first) == NULL;
+	return READ_ONCE(head->first) == NULL;
 }
 
 static inline struct llist_node *llist_next(struct llist_node *node)
diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h
index 2efb08a60e63..f0fc4700b6ff 100644
--- a/include/linux/pm_runtime.h
+++ b/include/linux/pm_runtime.h
@@ -105,7 +105,7 @@ static inline bool pm_runtime_callbacks_present(struct device *dev)
 
 static inline void pm_runtime_mark_last_busy(struct device *dev)
 {
-	ACCESS_ONCE(dev->power.last_busy) = jiffies;
+	WRITE_ONCE(dev->power.last_busy, jiffies);
 }
 
 static inline bool pm_runtime_is_irq_safe(struct device *dev)
diff --git a/include/net/ip_vs.h b/include/net/ip_vs.h
index 4f4f786255ef..3fadb6f9982b 100644
--- a/include/net/ip_vs.h
+++ b/include/net/ip_vs.h
@@ -983,12 +983,12 @@ static inline int sysctl_sync_threshold(struct netns_ipvs *ipvs)
 
 static inline int sysctl_sync_period(struct netns_ipvs *ipvs)
 {
-	return ACCESS_ONCE(ipvs->sysctl_sync_threshold[1]);
+	return READ_ONCE(ipvs->sysctl_sync_threshold[1]);
 }
 
 static inline unsigned int sysctl_sync_refresh_period(struct netns_ipvs *ipvs)
 {
-	return ACCESS_ONCE(ipvs->sysctl_sync_refresh_period);
+	return READ_ONCE(ipvs->sysctl_sync_refresh_period);
 }
 
 static inline int sysctl_sync_retries(struct netns_ipvs *ipvs)
@@ -1013,7 +1013,7 @@ static inline int sysctl_sloppy_sctp(struct netns_ipvs *ipvs)
 
 static inline int sysctl_sync_ports(struct netns_ipvs *ipvs)
 {
-	return ACCESS_ONCE(ipvs->sysctl_sync_ports);
+	return READ_ONCE(ipvs->sysctl_sync_ports);
 }
 
 static inline int sysctl_sync_persist_mode(struct netns_ipvs *ipvs)
diff --git a/kernel/acct.c b/kernel/acct.c
index 5e72af29ab73..21eedd0dd81a 100644
--- a/kernel/acct.c
+++ b/kernel/acct.c
@@ -146,7 +146,7 @@ static struct bsd_acct_struct *acct_get(struct pid_namespace *ns)
 again:
 	smp_rmb();
 	rcu_read_lock();
-	res = to_acct(ACCESS_ONCE(ns->bacct));
+	res = to_acct(READ_ONCE(ns->bacct));
 	if (!res) {
 		rcu_read_unlock();
 		return NULL;
@@ -158,7 +158,7 @@ static struct bsd_acct_struct *acct_get(struct pid_namespace *ns)
 	}
 	rcu_read_unlock();
 	mutex_lock(&res->lock);
-	if (res != to_acct(ACCESS_ONCE(ns->bacct))) {
+	if (res != to_acct(READ_ONCE(ns->bacct))) {
 		mutex_unlock(&res->lock);
 		acct_put(res);
 		goto again;
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 417812ce0099..a1322e855e9f 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -1200,7 +1200,7 @@ perf_event_ctx_lock_nested(struct perf_event *event, int nesting)
 
 again:
 	rcu_read_lock();
-	ctx = ACCESS_ONCE(event->ctx);
+	ctx = READ_ONCE(event->ctx);
 	if (!atomic_inc_not_zero(&ctx->refcount)) {
 		rcu_read_unlock();
 		goto again;
@@ -5302,8 +5302,8 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
 		if (!rb)
 			goto aux_unlock;
 
-		aux_offset = ACCESS_ONCE(rb->user_page->aux_offset);
-		aux_size = ACCESS_ONCE(rb->user_page->aux_size);
+		aux_offset = READ_ONCE(rb->user_page->aux_offset);
+		aux_size = READ_ONCE(rb->user_page->aux_size);
 
 		if (aux_offset < perf_data_size(rb) + PAGE_SIZE)
 			goto aux_unlock;
diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c
index f684d8e5fa2b..f3e37971c842 100644
--- a/kernel/events/ring_buffer.c
+++ b/kernel/events/ring_buffer.c
@@ -381,7 +381,7 @@ void *perf_aux_output_begin(struct perf_output_handle *handle,
 	 * (B) <-> (C) ordering is still observed by the pmu driver.
 	 */
 	if (!rb->aux_overwrite) {
-		aux_tail = ACCESS_ONCE(rb->user_page->aux_tail);
+		aux_tail = READ_ONCE(rb->user_page->aux_tail);
 		handle->wakeup = rb->aux_wakeup + rb->aux_watermark;
 		if (aux_head - aux_tail < perf_aux_size(rb))
 			handle->size = CIRC_SPACE(aux_head, aux_tail, perf_aux_size(rb));
diff --git a/kernel/exit.c b/kernel/exit.c
index f2cd53e92147..e4fee98803b3 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -1339,7 +1339,7 @@ static int wait_consider_task(struct wait_opts *wo, int ptrace,
 	 * Ensure that EXIT_ZOMBIE -> EXIT_DEAD/EXIT_TRACE transition
 	 * can't confuse the checks below.
 	 */
-	int exit_state = ACCESS_ONCE(p->exit_state);
+	int exit_state = READ_ONCE(p->exit_state);
 	int ret;
 
 	if (unlikely(exit_state == EXIT_DEAD))
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index 81279c6602ff..845f3805c73d 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -2724,7 +2724,7 @@ rb_reserve_next_event(struct ring_buffer *buffer,
 	 * if it happened, we have to fail the write.
 	 */
 	barrier();
-	if (unlikely(ACCESS_ONCE(cpu_buffer->buffer) != buffer)) {
+	if (unlikely(READ_ONCE(cpu_buffer->buffer) != buffer)) {
 		local_dec(&cpu_buffer->committing);
 		local_dec(&cpu_buffer->commits);
 		return NULL;
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 652c682707cd..9050c8b3ccde 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -1459,7 +1459,7 @@ extern struct trace_event_file *find_event_file(struct trace_array *tr,
 
 static inline void *event_file_data(struct file *filp)
 {
-	return ACCESS_ONCE(file_inode(filp)->i_private);
+	return READ_ONCE(file_inode(filp)->i_private);
 }
 
 extern struct mutex event_mutex;
diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c
index 49cb41412eec..780262210c9a 100644
--- a/kernel/trace/trace_stack.c
+++ b/kernel/trace/trace_stack.c
@@ -77,7 +77,7 @@ check_stack(unsigned long ip, unsigned long *stack)
 {
 	unsigned long this_size, flags; unsigned long *p, *top, *start;
 	static int tracer_frame;
-	int frame_size = ACCESS_ONCE(tracer_frame);
+	int frame_size = READ_ONCE(tracer_frame);
 	int i, x;
 
 	this_size = ((unsigned long)stack) & (THREAD_SIZE-1);
diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c
index c490f1e4313b..d32b45662fb6 100644
--- a/kernel/user_namespace.c
+++ b/kernel/user_namespace.c
@@ -894,7 +894,7 @@ static bool new_idmap_permitted(const struct file *file,
 int proc_setgroups_show(struct seq_file *seq, void *v)
 {
 	struct user_namespace *ns = seq->private;
-	unsigned long userns_flags = ACCESS_ONCE(ns->flags);
+	unsigned long userns_flags = READ_ONCE(ns->flags);
 
 	seq_printf(seq, "%s\n",
 		   (userns_flags & USERNS_SETGROUPS_ALLOWED) ?
diff --git a/lib/assoc_array.c b/lib/assoc_array.c
index 155c55d8db5f..fe7953aead82 100644
--- a/lib/assoc_array.c
+++ b/lib/assoc_array.c
@@ -39,7 +39,7 @@ static int assoc_array_subtree_iterate(const struct assoc_array_ptr *root,
 		/* Descend through a shortcut */
 		shortcut = assoc_array_ptr_to_shortcut(cursor);
 		smp_read_barrier_depends();
-		cursor = ACCESS_ONCE(shortcut->next_node);
+		cursor = READ_ONCE(shortcut->next_node);
 	}
 
 	node = assoc_array_ptr_to_node(cursor);
@@ -55,7 +55,7 @@ static int assoc_array_subtree_iterate(const struct assoc_array_ptr *root,
 	 */
 	has_meta = 0;
 	for (; slot < ASSOC_ARRAY_FAN_OUT; slot++) {
-		ptr = ACCESS_ONCE(node->slots[slot]);
+		ptr = READ_ONCE(node->slots[slot]);
 		has_meta |= (unsigned long)ptr;
 		if (ptr && assoc_array_ptr_is_leaf(ptr)) {
 			/* We need a barrier between the read of the pointer
@@ -89,7 +89,7 @@ static int assoc_array_subtree_iterate(const struct assoc_array_ptr *root,
 	smp_read_barrier_depends();
 
 	for (; slot < ASSOC_ARRAY_FAN_OUT; slot++) {
-		ptr = ACCESS_ONCE(node->slots[slot]);
+		ptr = READ_ONCE(node->slots[slot]);
 		if (assoc_array_ptr_is_meta(ptr)) {
 			cursor = ptr;
 			goto begin_node;
@@ -98,7 +98,7 @@ static int assoc_array_subtree_iterate(const struct assoc_array_ptr *root,
 
 finished_node:
 	/* Move up to the parent (may need to skip back over a shortcut) */
-	parent = ACCESS_ONCE(node->back_pointer);
+	parent = READ_ONCE(node->back_pointer);
 	slot = node->parent_slot;
 	if (parent == stop)
 		return 0;
@@ -107,7 +107,7 @@ static int assoc_array_subtree_iterate(const struct assoc_array_ptr *root,
 		shortcut = assoc_array_ptr_to_shortcut(parent);
 		smp_read_barrier_depends();
 		cursor = parent;
-		parent = ACCESS_ONCE(shortcut->back_pointer);
+		parent = READ_ONCE(shortcut->back_pointer);
 		slot = shortcut->parent_slot;
 		if (parent == stop)
 			return 0;
@@ -147,7 +147,7 @@ int assoc_array_iterate(const struct assoc_array *array,
 					void *iterator_data),
 			void *iterator_data)
 {
-	struct assoc_array_ptr *root = ACCESS_ONCE(array->root);
+	struct assoc_array_ptr *root = READ_ONCE(array->root);
 
 	if (!root)
 		return 0;
@@ -194,7 +194,7 @@ assoc_array_walk(const struct assoc_array *array,
 
 	pr_devel("-->%s()\n", __func__);
 
-	cursor = ACCESS_ONCE(array->root);
+	cursor = READ_ONCE(array->root);
 	if (!cursor)
 		return assoc_array_walk_tree_empty;
 
@@ -220,7 +220,7 @@ assoc_array_walk(const struct assoc_array *array,
 
 	slot = segments >> (level & ASSOC_ARRAY_KEY_CHUNK_MASK);
 	slot &= ASSOC_ARRAY_FAN_MASK;
-	ptr = ACCESS_ONCE(node->slots[slot]);
+	ptr = READ_ONCE(node->slots[slot]);
 
 	pr_devel("consider slot %x [ix=%d type=%lu]\n",
 		 slot, level, (unsigned long)ptr & 3);
@@ -294,7 +294,7 @@ assoc_array_walk(const struct assoc_array *array,
 	} while (sc_level < shortcut->skip_to_level);
 
 	/* The shortcut matches the leaf's index to this point. */
-	cursor = ACCESS_ONCE(shortcut->next_node);
+	cursor = READ_ONCE(shortcut->next_node);
 	if (((level ^ sc_level) & ~ASSOC_ARRAY_KEY_CHUNK_MASK) != 0) {
 		level = sc_level;
 		goto jumped;
@@ -337,7 +337,7 @@ void *assoc_array_find(const struct assoc_array *array,
 	 * the terminal node.
 	 */
 	for (slot = 0; slot < ASSOC_ARRAY_FAN_OUT; slot++) {
-		ptr = ACCESS_ONCE(node->slots[slot]);
+		ptr = READ_ONCE(node->slots[slot]);
 		if (ptr && assoc_array_ptr_is_leaf(ptr)) {
 			/* We need a barrier between the read of the pointer
 			 * and dereferencing the pointer - but only if we are
diff --git a/lib/dynamic_queue_limits.c b/lib/dynamic_queue_limits.c
index f346715e2255..81770a55cb16 100644
--- a/lib/dynamic_queue_limits.c
+++ b/lib/dynamic_queue_limits.c
@@ -20,7 +20,7 @@ void dql_completed(struct dql *dql, unsigned int count)
 	unsigned int ovlimit, completed, num_queued;
 	bool all_prev_completed;
 
-	num_queued = ACCESS_ONCE(dql->num_queued);
+	num_queued = READ_ONCE(dql->num_queued);
 
 	/* Can't complete more than what's in queue */
 	BUG_ON(count > num_queued - dql->num_completed);
diff --git a/lib/llist.c b/lib/llist.c
index ae5872b1df0c..7062e931a7bb 100644
--- a/lib/llist.c
+++ b/lib/llist.c
@@ -41,7 +41,7 @@ bool llist_add_batch(struct llist_node *new_first, struct llist_node *new_last,
 	struct llist_node *first;
 
 	do {
-		new_last->next = first = ACCESS_ONCE(head->first);
+		new_last->next = first = READ_ONCE(head->first);
 	} while (cmpxchg(&head->first, first, new_first) != first);
 
 	return !first;
diff --git a/lib/vsprintf.c b/lib/vsprintf.c
index 86c3385b9eb3..1746bae94d41 100644
--- a/lib/vsprintf.c
+++ b/lib/vsprintf.c
@@ -620,8 +620,8 @@ char *dentry_name(char *buf, char *end, const struct dentry *d, struct printf_sp
 
 	rcu_read_lock();
 	for (i = 0; i < depth; i++, d = p) {
-		p = ACCESS_ONCE(d->d_parent);
-		array[i] = ACCESS_ONCE(d->d_name.name);
+		p = READ_ONCE(d->d_parent);
+		array[i] = READ_ONCE(d->d_name.name);
 		if (p == d) {
 			if (i)
 				array[i] = "";
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 269b5df58543..c3bf907a03ee 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2715,7 +2715,7 @@ static unsigned long deferred_split_count(struct shrinker *shrink,
 		struct shrink_control *sc)
 {
 	struct pglist_data *pgdata = NODE_DATA(sc->nid);
-	return ACCESS_ONCE(pgdata->split_queue_len);
+	return READ_ONCE(pgdata->split_queue_len);
 }
 
 static unsigned long deferred_split_scan(struct shrinker *shrink,
diff --git a/net/core/dev.c b/net/core/dev.c
index 588b473194a8..648be63520c6 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -3725,7 +3725,7 @@ bool rps_may_expire_flow(struct net_device *dev, u16 rxq_index,
 	flow_table = rcu_dereference(rxqueue->rps_flow_table);
 	if (flow_table && flow_id <= flow_table->mask) {
 		rflow = &flow_table->flows[flow_id];
-		cpu = ACCESS_ONCE(rflow->cpu);
+		cpu = READ_ONCE(rflow->cpu);
 		if (rflow->filter == filter_id && cpu < nr_cpu_ids &&
 		    ((int)(per_cpu(softnet_data, cpu).input_queue_head -
 			   rflow->last_qtail) <
diff --git a/net/core/pktgen.c b/net/core/pktgen.c
index 6e1e10ff433a..3b2034f6d49d 100644
--- a/net/core/pktgen.c
+++ b/net/core/pktgen.c
@@ -3377,7 +3377,7 @@ static void pktgen_wait_for_skb(struct pktgen_dev *pkt_dev)
 
 static void pktgen_xmit(struct pktgen_dev *pkt_dev)
 {
-	unsigned int burst = ACCESS_ONCE(pkt_dev->burst);
+	unsigned int burst = READ_ONCE(pkt_dev->burst);
 	struct net_device *odev = pkt_dev->odev;
 	struct netdev_queue *txq;
 	struct sk_buff *skb;
diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c
index af74d0433453..f9597ba26599 100644
--- a/net/ipv4/inet_fragment.c
+++ b/net/ipv4/inet_fragment.c
@@ -164,7 +164,7 @@ static void inet_frag_worker(struct work_struct *work)
 
 	local_bh_disable();
 
-	for (i = ACCESS_ONCE(f->next_bucket); budget; --budget) {
+	for (i = READ_ONCE(f->next_bucket); budget; --budget) {
 		evicted += inet_evict_bucket(f, &f->hash[i]);
 		i = (i + 1) & (INETFRAGS_HASHSZ - 1);
 		if (evicted > INETFRAGS_EVICT_MAX)
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index ac6fde5d45f1..ea67e330be9e 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -495,7 +495,7 @@ u32 ip_idents_reserve(u32 hash, int segs)
 {
 	u32 *p_tstamp = ip_tstamps + hash % IP_IDENTS_SZ;
 	atomic_t *p_id = ip_idents + hash % IP_IDENTS_SZ;
-	u32 old = ACCESS_ONCE(*p_tstamp);
+	u32 old = READ_ONCE(*p_tstamp);
 	u32 now = (u32)jiffies;
 	u32 new, delta = 0;
 
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 0bc9e46a5369..48531da1aba6 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1908,7 +1908,7 @@ static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb,
 	if ((skb != tcp_write_queue_tail(sk)) && (limit >= skb->len))
 		goto send_now;
 
-	win_divisor = ACCESS_ONCE(sysctl_tcp_tso_win_divisor);
+	win_divisor = READ_ONCE(sysctl_tcp_tso_win_divisor);
 	if (win_divisor) {
 		u32 chunk = min(tp->snd_wnd, tp->snd_cwnd * tp->mss_cache);
 
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 5676237d2b0f..1049347f2fbc 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -1856,7 +1856,7 @@ static int udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
 		 */
 
 		/* if we're overly short, let UDP handle it */
-		encap_rcv = ACCESS_ONCE(up->encap_rcv);
+		encap_rcv = READ_ONCE(up->encap_rcv);
 		if (encap_rcv) {
 			int ret;
 
@@ -2305,7 +2305,7 @@ void udp_destroy_sock(struct sock *sk)
 	unlock_sock_fast(sk, slow);
 	if (static_key_false(&udp_encap_needed) && up->encap_type) {
 		void (*encap_destroy)(struct sock *sk);
-		encap_destroy = ACCESS_ONCE(up->encap_destroy);
+		encap_destroy = READ_ONCE(up->encap_destroy);
 		if (encap_destroy)
 			encap_destroy(sk);
 	}
diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index a1c24443cd9e..dab946554157 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -490,7 +490,7 @@ ip6_tnl_err(struct sk_buff *skb, __u8 ipproto, struct inet6_skb_parm *opt,
 	if (!t)
 		goto out;
 
-	tproto = ACCESS_ONCE(t->parms.proto);
+	tproto = READ_ONCE(t->parms.proto);
 	if (tproto != ipproto && tproto != 0)
 		goto out;
 
@@ -899,7 +899,7 @@ static int ipxip6_rcv(struct sk_buff *skb, u8 ipproto,
 	t = ip6_tnl_lookup(dev_net(skb->dev), &ipv6h->saddr, &ipv6h->daddr);
 
 	if (t) {
-		u8 tproto = ACCESS_ONCE(t->parms.proto);
+		u8 tproto = READ_ONCE(t->parms.proto);
 
 		if (tproto != ipproto && tproto != 0)
 			goto drop;
@@ -1233,7 +1233,7 @@ ip4ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
 
-	tproto = ACCESS_ONCE(t->parms.proto);
+	tproto = READ_ONCE(t->parms.proto);
 	if (tproto != IPPROTO_IPIP && tproto != 0)
 		return -1;
 
@@ -1303,7 +1303,7 @@ ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
 	u8 tproto;
 	int err;
 
-	tproto = ACCESS_ONCE(t->parms.proto);
+	tproto = READ_ONCE(t->parms.proto);
 	if ((tproto != IPPROTO_IPV6 && tproto != 0) ||
 	    ip6_tnl_addr_conflict(t, ipv6h))
 		return -1;
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index 40d7234c27b9..3f30fa313bf2 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -606,7 +606,7 @@ static int udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
 		 */
 
 		/* if we're overly short, let UDP handle it */
-		encap_rcv = ACCESS_ONCE(up->encap_rcv);
+		encap_rcv = READ_ONCE(up->encap_rcv);
 		if (encap_rcv) {
 			int ret;
 
@@ -1432,7 +1432,7 @@ void udpv6_destroy_sock(struct sock *sk)
 
 	if (static_key_false(&udpv6_encap_needed) && up->encap_type) {
 		void (*encap_destroy)(struct sock *sk);
-		encap_destroy = ACCESS_ONCE(up->encap_destroy);
+		encap_destroy = READ_ONCE(up->encap_destroy);
 		if (encap_destroy)
 			encap_destroy(sk);
 	}
diff --git a/net/llc/llc_input.c b/net/llc/llc_input.c
index dd3e83328ad5..82cb93f66b9b 100644
--- a/net/llc/llc_input.c
+++ b/net/llc/llc_input.c
@@ -193,7 +193,7 @@ int llc_rcv(struct sk_buff *skb, struct net_device *dev,
 	 */
 	rcv = rcu_dereference(sap->rcv_func);
 	dest = llc_pdu_type(skb);
-	sap_handler = dest ? ACCESS_ONCE(llc_type_handlers[dest - 1]) : NULL;
+	sap_handler = dest ? READ_ONCE(llc_type_handlers[dest - 1]) : NULL;
 	if (unlikely(!sap_handler)) {
 		if (rcv)
 			rcv(skb, dev, pt, orig_dev);
@@ -214,7 +214,7 @@ int llc_rcv(struct sk_buff *skb, struct net_device *dev,
 	kfree_skb(skb);
 	goto out;
 handle_station:
-	sta_handler = ACCESS_ONCE(llc_station_handler);
+	sta_handler = READ_ONCE(llc_station_handler);
 	if (!sta_handler)
 		goto drop;
 	sta_handler(skb);
diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c
index 69615016d5bf..214d2ba02877 100644
--- a/net/mac80211/sta_info.c
+++ b/net/mac80211/sta_info.c
@@ -2008,7 +2008,7 @@ static void sta_stats_decode_rate(struct ieee80211_local *local, u16 rate,
 
 static int sta_set_rate_info_rx(struct sta_info *sta, struct rate_info *rinfo)
 {
-	u16 rate = ACCESS_ONCE(sta_get_last_rx_stats(sta)->last_rate);
+	u16 rate = READ_ONCE(sta_get_last_rx_stats(sta)->last_rate);
 
 	if (rate == STA_STATS_RATE_INVALID)
 		return -EINVAL;
diff --git a/net/netlabel/netlabel_calipso.c b/net/netlabel/netlabel_calipso.c
index d177dd066504..4d748975117d 100644
--- a/net/netlabel/netlabel_calipso.c
+++ b/net/netlabel/netlabel_calipso.c
@@ -393,7 +393,7 @@ EXPORT_SYMBOL(netlbl_calipso_ops_register);
 
 static const struct netlbl_calipso_ops *netlbl_calipso_ops_get(void)
 {
-	return ACCESS_ONCE(calipso_ops);
+	return READ_ONCE(calipso_ops);
 }
 
 /**
diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c
index 690874293cfc..fe63060e8c1b 100644
--- a/net/wireless/nl80211.c
+++ b/net/wireless/nl80211.c
@@ -14191,7 +14191,7 @@ static bool __nl80211_unexpected_frame(struct net_device *dev, u8 cmd,
 	struct cfg80211_registered_device *rdev = wiphy_to_rdev(wdev->wiphy);
 	struct sk_buff *msg;
 	void *hdr;
-	u32 nlportid = ACCESS_ONCE(wdev->ap_unexpected_nlportid);
+	u32 nlportid = READ_ONCE(wdev->ap_unexpected_nlportid);
 
 	if (!nlportid)
 		return false;
diff --git a/sound/firewire/amdtp-am824.c b/sound/firewire/amdtp-am824.c
index 23ccddb20de1..4210e5c6262e 100644
--- a/sound/firewire/amdtp-am824.c
+++ b/sound/firewire/amdtp-am824.c
@@ -247,7 +247,7 @@ void amdtp_am824_midi_trigger(struct amdtp_stream *s, unsigned int port,
 	struct amdtp_am824 *p = s->protocol;
 
 	if (port < p->midi_ports)
-		ACCESS_ONCE(p->midi[port]) = midi;
+		WRITE_ONCE(p->midi[port], midi);
 }
 EXPORT_SYMBOL_GPL(amdtp_am824_midi_trigger);
 
@@ -336,7 +336,7 @@ static unsigned int process_rx_data_blocks(struct amdtp_stream *s, __be32 *buffe
 					   unsigned int data_blocks, unsigned int *syt)
 {
 	struct amdtp_am824 *p = s->protocol;
-	struct snd_pcm_substream *pcm = ACCESS_ONCE(s->pcm);
+	struct snd_pcm_substream *pcm = READ_ONCE(s->pcm);
 	unsigned int pcm_frames;
 
 	if (pcm) {
@@ -357,7 +357,7 @@ static unsigned int process_tx_data_blocks(struct amdtp_stream *s, __be32 *buffe
 					   unsigned int data_blocks, unsigned int *syt)
 {
 	struct amdtp_am824 *p = s->protocol;
-	struct snd_pcm_substream *pcm = ACCESS_ONCE(s->pcm);
+	struct snd_pcm_substream *pcm = READ_ONCE(s->pcm);
 	unsigned int pcm_frames;
 
 	if (pcm) {
diff --git a/sound/firewire/amdtp-stream.c b/sound/firewire/amdtp-stream.c
index 3fc581a5ad62..4a1dc145327b 100644
--- a/sound/firewire/amdtp-stream.c
+++ b/sound/firewire/amdtp-stream.c
@@ -376,7 +376,7 @@ static void update_pcm_pointers(struct amdtp_stream *s,
 	ptr = s->pcm_buffer_pointer + frames;
 	if (ptr >= pcm->runtime->buffer_size)
 		ptr -= pcm->runtime->buffer_size;
-	ACCESS_ONCE(s->pcm_buffer_pointer) = ptr;
+	WRITE_ONCE(s->pcm_buffer_pointer, ptr);
 
 	s->pcm_period_pointer += frames;
 	if (s->pcm_period_pointer >= pcm->runtime->period_size) {
@@ -388,7 +388,7 @@ static void update_pcm_pointers(struct amdtp_stream *s,
 static void pcm_period_tasklet(unsigned long data)
 {
 	struct amdtp_stream *s = (void *)data;
-	struct snd_pcm_substream *pcm = ACCESS_ONCE(s->pcm);
+	struct snd_pcm_substream *pcm = READ_ONCE(s->pcm);
 
 	if (pcm)
 		snd_pcm_period_elapsed(pcm);
@@ -453,7 +453,7 @@ static int handle_out_packet(struct amdtp_stream *s,
 		s->data_block_counter =
 				(s->data_block_counter + data_blocks) & 0xff;
 
-	buffer[0] = cpu_to_be32(ACCESS_ONCE(s->source_node_id_field) |
+	buffer[0] = cpu_to_be32(READ_ONCE(s->source_node_id_field) |
 				(s->data_block_quadlets << CIP_DBS_SHIFT) |
 				((s->sph << CIP_SPH_SHIFT) & CIP_SPH_MASK) |
 				s->data_block_counter);
@@ -472,7 +472,7 @@ static int handle_out_packet(struct amdtp_stream *s,
 	if (queue_out_packet(s, payload_length) < 0)
 		return -EIO;
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (pcm && pcm_frames > 0)
 		update_pcm_pointers(s, pcm, pcm_frames);
 
@@ -504,7 +504,7 @@ static int handle_out_packet_without_header(struct amdtp_stream *s,
 	if (queue_out_packet(s, payload_length) < 0)
 		return -EIO;
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (pcm && pcm_frames > 0)
 		update_pcm_pointers(s, pcm, pcm_frames);
 
@@ -621,7 +621,7 @@ static int handle_in_packet(struct amdtp_stream *s,
 	if (queue_in_packet(s) < 0)
 		return -EIO;
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (pcm && pcm_frames > 0)
 		update_pcm_pointers(s, pcm, pcm_frames);
 
@@ -649,7 +649,7 @@ static int handle_in_packet_without_header(struct amdtp_stream *s,
 	if (queue_in_packet(s) < 0)
 		return -EIO;
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (pcm && pcm_frames > 0)
 		update_pcm_pointers(s, pcm, pcm_frames);
 
@@ -947,7 +947,7 @@ unsigned long amdtp_stream_pcm_pointer(struct amdtp_stream *s)
 	if (!in_interrupt() && amdtp_stream_running(s))
 		fw_iso_context_flush_completions(s->context);
 
-	return ACCESS_ONCE(s->pcm_buffer_pointer);
+	return READ_ONCE(s->pcm_buffer_pointer);
 }
 EXPORT_SYMBOL(amdtp_stream_pcm_pointer);
 
@@ -977,9 +977,8 @@ EXPORT_SYMBOL(amdtp_stream_pcm_ack);
 void amdtp_stream_update(struct amdtp_stream *s)
 {
 	/* Precomputing. */
-	ACCESS_ONCE(s->source_node_id_field) =
-		(fw_parent_device(s->unit)->card->node_id << CIP_SID_SHIFT) &
-								CIP_SID_MASK;
+	WRITE_ONCE(s->source_node_id_field,
+                   (fw_parent_device(s->unit)->card->node_id << CIP_SID_SHIFT) & CIP_SID_MASK);
 }
 EXPORT_SYMBOL(amdtp_stream_update);
 
@@ -1022,7 +1021,7 @@ void amdtp_stream_pcm_abort(struct amdtp_stream *s)
 {
 	struct snd_pcm_substream *pcm;
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (pcm)
 		snd_pcm_stop_xrun(pcm);
 }
diff --git a/sound/firewire/amdtp-stream.h b/sound/firewire/amdtp-stream.h
index ed6eafd10992..f9abd8b07ce6 100644
--- a/sound/firewire/amdtp-stream.h
+++ b/sound/firewire/amdtp-stream.h
@@ -220,7 +220,7 @@ static inline bool amdtp_stream_pcm_running(struct amdtp_stream *s)
 static inline void amdtp_stream_pcm_trigger(struct amdtp_stream *s,
 					    struct snd_pcm_substream *pcm)
 {
-	ACCESS_ONCE(s->pcm) = pcm;
+	WRITE_ONCE(s->pcm, pcm);
 }
 
 static inline bool cip_sfc_is_base_44100(enum cip_sfc sfc)
diff --git a/sound/firewire/digi00x/amdtp-dot.c b/sound/firewire/digi00x/amdtp-dot.c
index 1453c34ce99f..4a884a335248 100644
--- a/sound/firewire/digi00x/amdtp-dot.c
+++ b/sound/firewire/digi00x/amdtp-dot.c
@@ -327,7 +327,7 @@ void amdtp_dot_midi_trigger(struct amdtp_stream *s, unsigned int port,
 	struct amdtp_dot *p = s->protocol;
 
 	if (port < MAX_MIDI_PORTS)
-		ACCESS_ONCE(p->midi[port]) = midi;
+		WRITE_ONCE(p->midi[port], midi);
 }
 
 static unsigned int process_tx_data_blocks(struct amdtp_stream *s,
@@ -338,7 +338,7 @@ static unsigned int process_tx_data_blocks(struct amdtp_stream *s,
 	struct snd_pcm_substream *pcm;
 	unsigned int pcm_frames;
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (pcm) {
 		read_pcm_s32(s, pcm, buffer, data_blocks);
 		pcm_frames = data_blocks;
@@ -359,7 +359,7 @@ static unsigned int process_rx_data_blocks(struct amdtp_stream *s,
 	struct snd_pcm_substream *pcm;
 	unsigned int pcm_frames;
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (pcm) {
 		write_pcm_s32(s, pcm, buffer, data_blocks);
 		pcm_frames = data_blocks;
diff --git a/sound/firewire/fireface/amdtp-ff.c b/sound/firewire/fireface/amdtp-ff.c
index 780da9deb2f0..77c7598b61ab 100644
--- a/sound/firewire/fireface/amdtp-ff.c
+++ b/sound/firewire/fireface/amdtp-ff.c
@@ -108,7 +108,7 @@ static unsigned int process_rx_data_blocks(struct amdtp_stream *s,
 					   unsigned int data_blocks,
 					   unsigned int *syt)
 {
-	struct snd_pcm_substream *pcm = ACCESS_ONCE(s->pcm);
+	struct snd_pcm_substream *pcm = READ_ONCE(s->pcm);
 	unsigned int pcm_frames;
 
 	if (pcm) {
@@ -127,7 +127,7 @@ static unsigned int process_tx_data_blocks(struct amdtp_stream *s,
 					   unsigned int data_blocks,
 					   unsigned int *syt)
 {
-	struct snd_pcm_substream *pcm = ACCESS_ONCE(s->pcm);
+	struct snd_pcm_substream *pcm = READ_ONCE(s->pcm);
 	unsigned int pcm_frames;
 
 	if (pcm) {
diff --git a/sound/firewire/fireface/ff-midi.c b/sound/firewire/fireface/ff-midi.c
index 949ee56b4e0e..6a49611ee462 100644
--- a/sound/firewire/fireface/ff-midi.c
+++ b/sound/firewire/fireface/ff-midi.c
@@ -22,7 +22,7 @@ static int midi_playback_open(struct snd_rawmidi_substream *substream)
 	ff->running_status[substream->number] = 0;
 	ff->rx_midi_error[substream->number] = false;
 
-	ACCESS_ONCE(ff->rx_midi_substreams[substream->number]) = substream;
+	WRITE_ONCE(ff->rx_midi_substreams[substream->number], substream);
 
 	return 0;
 }
@@ -38,7 +38,7 @@ static int midi_playback_close(struct snd_rawmidi_substream *substream)
 	struct snd_ff *ff = substream->rmidi->private_data;
 
 	cancel_work_sync(&ff->rx_midi_work[substream->number]);
-	ACCESS_ONCE(ff->rx_midi_substreams[substream->number]) = NULL;
+	WRITE_ONCE(ff->rx_midi_substreams[substream->number], NULL);
 
 	return 0;
 }
@@ -52,10 +52,10 @@ static void midi_capture_trigger(struct snd_rawmidi_substream *substream,
 	spin_lock_irqsave(&ff->lock, flags);
 
 	if (up)
-		ACCESS_ONCE(ff->tx_midi_substreams[substream->number]) =
-								substream;
+		WRITE_ONCE(ff->tx_midi_substreams[substream->number],
+			   substream);
 	else
-		ACCESS_ONCE(ff->tx_midi_substreams[substream->number]) = NULL;
+		WRITE_ONCE(ff->tx_midi_substreams[substream->number], NULL);
 
 	spin_unlock_irqrestore(&ff->lock, flags);
 }
diff --git a/sound/firewire/fireface/ff-transaction.c b/sound/firewire/fireface/ff-transaction.c
index dd6c8e839647..332b29f8ed75 100644
--- a/sound/firewire/fireface/ff-transaction.c
+++ b/sound/firewire/fireface/ff-transaction.c
@@ -12,7 +12,7 @@ static void finish_transmit_midi_msg(struct snd_ff *ff, unsigned int port,
 				     int rcode)
 {
 	struct snd_rawmidi_substream *substream =
-				ACCESS_ONCE(ff->rx_midi_substreams[port]);
+				READ_ONCE(ff->rx_midi_substreams[port]);
 
 	if (rcode_is_permanent_error(rcode)) {
 		ff->rx_midi_error[port] = true;
@@ -60,7 +60,7 @@ static inline void fill_midi_buf(struct snd_ff *ff, unsigned int port,
 static void transmit_midi_msg(struct snd_ff *ff, unsigned int port)
 {
 	struct snd_rawmidi_substream *substream =
-			ACCESS_ONCE(ff->rx_midi_substreams[port]);
+			READ_ONCE(ff->rx_midi_substreams[port]);
 	u8 *buf = (u8 *)ff->msg_buf[port];
 	int i, len;
 
@@ -159,7 +159,7 @@ static void handle_midi_msg(struct fw_card *card, struct fw_request *request,
 		 */
 		index = (quad >> 8) & 0xff;
 		if (index > 0) {
-			substream = ACCESS_ONCE(ff->tx_midi_substreams[0]);
+			substream = READ_ONCE(ff->tx_midi_substreams[0]);
 			if (substream != NULL) {
 				byte = quad & 0xff;
 				snd_rawmidi_receive(substream, &byte, 1);
@@ -169,7 +169,7 @@ static void handle_midi_msg(struct fw_card *card, struct fw_request *request,
 		/* Message in second port. */
 		index = (quad >> 24) & 0xff;
 		if (index > 0) {
-			substream = ACCESS_ONCE(ff->tx_midi_substreams[1]);
+			substream = READ_ONCE(ff->tx_midi_substreams[1]);
 			if (substream != NULL) {
 				byte = (quad >> 16) & 0xff;
 				snd_rawmidi_receive(substream, &byte, 1);
diff --git a/sound/firewire/isight.c b/sound/firewire/isight.c
index 5826aa8362f1..46092fa3ff9b 100644
--- a/sound/firewire/isight.c
+++ b/sound/firewire/isight.c
@@ -96,7 +96,7 @@ static void isight_update_pointers(struct isight *isight, unsigned int count)
 	ptr += count;
 	if (ptr >= runtime->buffer_size)
 		ptr -= runtime->buffer_size;
-	ACCESS_ONCE(isight->buffer_pointer) = ptr;
+	WRITE_ONCE(isight->buffer_pointer, ptr);
 
 	isight->period_counter += count;
 	if (isight->period_counter >= runtime->period_size) {
@@ -111,7 +111,7 @@ static void isight_samples(struct isight *isight,
 	struct snd_pcm_runtime *runtime;
 	unsigned int count1;
 
-	if (!ACCESS_ONCE(isight->pcm_running))
+	if (!READ_ONCE(isight->pcm_running))
 		return;
 
 	runtime = isight->pcm->runtime;
@@ -131,7 +131,7 @@ static void isight_samples(struct isight *isight,
 
 static void isight_pcm_abort(struct isight *isight)
 {
-	if (ACCESS_ONCE(isight->pcm_active))
+	if (READ_ONCE(isight->pcm_active))
 		snd_pcm_stop_xrun(isight->pcm);
 }
 
@@ -141,7 +141,7 @@ static void isight_dropped_samples(struct isight *isight, unsigned int total)
 	u32 dropped;
 	unsigned int count1;
 
-	if (!ACCESS_ONCE(isight->pcm_running))
+	if (!READ_ONCE(isight->pcm_running))
 		return;
 
 	runtime = isight->pcm->runtime;
@@ -293,7 +293,7 @@ static int isight_hw_params(struct snd_pcm_substream *substream,
 	if (err < 0)
 		return err;
 
-	ACCESS_ONCE(isight->pcm_active) = true;
+	WRITE_ONCE(isight->pcm_active, true);
 
 	return 0;
 }
@@ -331,7 +331,7 @@ static int isight_hw_free(struct snd_pcm_substream *substream)
 {
 	struct isight *isight = substream->private_data;
 
-	ACCESS_ONCE(isight->pcm_active) = false;
+	WRITE_ONCE(isight->pcm_active, false);
 
 	mutex_lock(&isight->mutex);
 	isight_stop_streaming(isight);
@@ -424,10 +424,10 @@ static int isight_trigger(struct snd_pcm_substream *substream, int cmd)
 
 	switch (cmd) {
 	case SNDRV_PCM_TRIGGER_START:
-		ACCESS_ONCE(isight->pcm_running) = true;
+		WRITE_ONCE(isight->pcm_running, true);
 		break;
 	case SNDRV_PCM_TRIGGER_STOP:
-		ACCESS_ONCE(isight->pcm_running) = false;
+		WRITE_ONCE(isight->pcm_running, false);
 		break;
 	default:
 		return -EINVAL;
@@ -439,7 +439,7 @@ static snd_pcm_uframes_t isight_pointer(struct snd_pcm_substream *substream)
 {
 	struct isight *isight = substream->private_data;
 
-	return ACCESS_ONCE(isight->buffer_pointer);
+	return READ_ONCE(isight->buffer_pointer);
 }
 
 static int isight_create_pcm(struct isight *isight)
diff --git a/sound/firewire/motu/amdtp-motu.c b/sound/firewire/motu/amdtp-motu.c
index 96f0091144bb..f0555a24d90e 100644
--- a/sound/firewire/motu/amdtp-motu.c
+++ b/sound/firewire/motu/amdtp-motu.c
@@ -310,7 +310,7 @@ static unsigned int process_tx_data_blocks(struct amdtp_stream *s,
 	if (p->midi_ports)
 		read_midi_messages(s, buffer, data_blocks);
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (data_blocks > 0 && pcm)
 		read_pcm_s32(s, pcm->runtime, buffer, data_blocks);
 
@@ -374,7 +374,7 @@ static unsigned int process_rx_data_blocks(struct amdtp_stream *s,
 	if (p->midi_ports)
 		write_midi_messages(s, buffer, data_blocks);
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (pcm)
 		write_pcm_s32(s, pcm->runtime, buffer, data_blocks);
 	else
diff --git a/sound/firewire/oxfw/oxfw-scs1x.c b/sound/firewire/oxfw/oxfw-scs1x.c
index 02d595665898..f33497cdc706 100644
--- a/sound/firewire/oxfw/oxfw-scs1x.c
+++ b/sound/firewire/oxfw/oxfw-scs1x.c
@@ -112,7 +112,7 @@ static void handle_hss(struct fw_card *card, struct fw_request *request,
 	}
 
 	if (length >= 1) {
-		stream = ACCESS_ONCE(scs->input);
+		stream = READ_ONCE(scs->input);
 		if (stream)
 			midi_input_packet(scs, stream, data, length);
 	}
@@ -183,7 +183,7 @@ static void scs_output_work(struct work_struct *work)
 	if (scs->transaction_running)
 		return;
 
-	stream = ACCESS_ONCE(scs->output);
+	stream = READ_ONCE(scs->output);
 	if (!stream || scs->error) {
 		scs->output_idle = true;
 		wake_up(&scs->idle_wait);
@@ -291,9 +291,9 @@ static void midi_capture_trigger(struct snd_rawmidi_substream *stream, int up)
 
 	if (up) {
 		scs->input_escape_count = 0;
-		ACCESS_ONCE(scs->input) = stream;
+		WRITE_ONCE(scs->input, stream);
 	} else {
-		ACCESS_ONCE(scs->input) = NULL;
+		WRITE_ONCE(scs->input, NULL);
 	}
 }
 
@@ -319,10 +319,10 @@ static void midi_playback_trigger(struct snd_rawmidi_substream *stream, int up)
 		scs->transaction_bytes = 0;
 		scs->error = false;
 
-		ACCESS_ONCE(scs->output) = stream;
+		WRITE_ONCE(scs->output, stream);
 		schedule_work(&scs->work);
 	} else {
-		ACCESS_ONCE(scs->output) = NULL;
+		WRITE_ONCE(scs->output, NULL);
 	}
 }
 static void midi_playback_drain(struct snd_rawmidi_substream *stream)
diff --git a/sound/firewire/tascam/amdtp-tascam.c b/sound/firewire/tascam/amdtp-tascam.c
index 6aff1fc1c72d..ab482423c165 100644
--- a/sound/firewire/tascam/amdtp-tascam.c
+++ b/sound/firewire/tascam/amdtp-tascam.c
@@ -124,7 +124,7 @@ static unsigned int process_tx_data_blocks(struct amdtp_stream *s,
 {
 	struct snd_pcm_substream *pcm;
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (data_blocks > 0 && pcm)
 		read_pcm_s32(s, pcm, buffer, data_blocks);
 
@@ -143,7 +143,7 @@ static unsigned int process_rx_data_blocks(struct amdtp_stream *s,
 	/* This field is not used. */
 	*syt = 0x0000;
 
-	pcm = ACCESS_ONCE(s->pcm);
+	pcm = READ_ONCE(s->pcm);
 	if (pcm)
 		write_pcm_s32(s, pcm, buffer, data_blocks);
 	else
diff --git a/sound/firewire/tascam/tascam-transaction.c b/sound/firewire/tascam/tascam-transaction.c
index 8967c52f5032..2ad692dd4b13 100644
--- a/sound/firewire/tascam/tascam-transaction.c
+++ b/sound/firewire/tascam/tascam-transaction.c
@@ -148,7 +148,7 @@ static void async_midi_port_callback(struct fw_card *card, int rcode,
 				     void *callback_data)
 {
 	struct snd_fw_async_midi_port *port = callback_data;
-	struct snd_rawmidi_substream *substream = ACCESS_ONCE(port->substream);
+	struct snd_rawmidi_substream *substream = READ_ONCE(port->substream);
 
 	/* This port is closed. */
 	if (substream == NULL)
@@ -173,7 +173,7 @@ static void midi_port_work(struct work_struct *work)
 {
 	struct snd_fw_async_midi_port *port =
 			container_of(work, struct snd_fw_async_midi_port, work);
-	struct snd_rawmidi_substream *substream = ACCESS_ONCE(port->substream);
+	struct snd_rawmidi_substream *substream = READ_ONCE(port->substream);
 	int generation;
 
 	/* Under transacting or error state. */
@@ -282,7 +282,7 @@ static void handle_midi_tx(struct fw_card *card, struct fw_request *request,
 				bytes = 3;
 		}
 
-		substream = ACCESS_ONCE(tscm->tx_midi_substreams[port]);
+		substream = READ_ONCE(tscm->tx_midi_substreams[port]);
 		if (substream != NULL)
 			snd_rawmidi_receive(substream, b + 1, bytes);
 	}
diff --git a/sound/soc/xtensa/xtfpga-i2s.c b/sound/soc/xtensa/xtfpga-i2s.c
index 8382ffa3bcaf..2472144b329e 100644
--- a/sound/soc/xtensa/xtfpga-i2s.c
+++ b/sound/soc/xtensa/xtfpga-i2s.c
@@ -165,7 +165,7 @@ static bool xtfpga_pcm_push_tx(struct xtfpga_i2s *i2s)
 	tx_substream = rcu_dereference(i2s->tx_substream);
 	tx_active = tx_substream && snd_pcm_running(tx_substream);
 	if (tx_active) {
-		unsigned tx_ptr = ACCESS_ONCE(i2s->tx_ptr);
+		unsigned tx_ptr = READ_ONCE(i2s->tx_ptr);
 		unsigned new_tx_ptr = i2s->tx_fn(i2s, tx_substream->runtime,
 						 tx_ptr);
 
@@ -437,7 +437,7 @@ static int xtfpga_pcm_trigger(struct snd_pcm_substream *substream, int cmd)
 	case SNDRV_PCM_TRIGGER_START:
 	case SNDRV_PCM_TRIGGER_RESUME:
 	case SNDRV_PCM_TRIGGER_PAUSE_RELEASE:
-		ACCESS_ONCE(i2s->tx_ptr) = 0;
+		WRITE_ONCE(i2s->tx_ptr, 0);
 		rcu_assign_pointer(i2s->tx_substream, substream);
 		xtfpga_pcm_refill_fifo(i2s);
 		break;
@@ -459,7 +459,7 @@ static snd_pcm_uframes_t xtfpga_pcm_pointer(struct snd_pcm_substream *substream)
 {
 	struct snd_pcm_runtime *runtime = substream->runtime;
 	struct xtfpga_i2s *i2s = runtime->private_data;
-	snd_pcm_uframes_t pos = ACCESS_ONCE(i2s->tx_ptr);
+	snd_pcm_uframes_t pos = READ_ONCE(i2s->tx_ptr);
 
 	return pos < runtime->buffer_size ? pos : 0;
 }
diff --git a/sound/usb/bcd2000/bcd2000.c b/sound/usb/bcd2000/bcd2000.c
index 7371e5b06035..fc579f330601 100644
--- a/sound/usb/bcd2000/bcd2000.c
+++ b/sound/usb/bcd2000/bcd2000.c
@@ -108,7 +108,7 @@ static void bcd2000_midi_handle_input(struct bcd2000 *bcd2k,
 	unsigned int payload_length, tocopy;
 	struct snd_rawmidi_substream *midi_receive_substream;
 
-	midi_receive_substream = ACCESS_ONCE(bcd2k->midi_receive_substream);
+	midi_receive_substream = READ_ONCE(bcd2k->midi_receive_substream);
 	if (!midi_receive_substream)
 		return;
 
@@ -139,7 +139,7 @@ static void bcd2000_midi_send(struct bcd2000 *bcd2k)
 
 	BUILD_BUG_ON(sizeof(device_cmd_prefix) >= BUFSIZE);
 
-	midi_out_substream = ACCESS_ONCE(bcd2k->midi_out_substream);
+	midi_out_substream = READ_ONCE(bcd2k->midi_out_substream);
 	if (!midi_out_substream)
 		return;
 
diff --git a/tools/arch/x86/include/asm/atomic.h b/tools/arch/x86/include/asm/atomic.h
index 328eeceec709..96e2d06cb031 100644
--- a/tools/arch/x86/include/asm/atomic.h
+++ b/tools/arch/x86/include/asm/atomic.h
@@ -24,7 +24,7 @@
  */
 static inline int atomic_read(const atomic_t *v)
 {
-	return ACCESS_ONCE((v)->counter);
+	return READ_ONCE((v)->counter);
 }
 
 /**
diff --git a/tools/include/asm-generic/atomic-gcc.h b/tools/include/asm-generic/atomic-gcc.h
index 5e9738f97bf3..97427e700e3b 100644
--- a/tools/include/asm-generic/atomic-gcc.h
+++ b/tools/include/asm-generic/atomic-gcc.h
@@ -21,7 +21,7 @@
  */
 static inline int atomic_read(const atomic_t *v)
 {
-	return ACCESS_ONCE((v)->counter);
+	return READ_ONCE((v)->counter);
 }
 
 /**
diff --git a/tools/perf/util/auxtrace.h b/tools/perf/util/auxtrace.h
index 33b5e6cdf38c..d19e11b68de7 100644
--- a/tools/perf/util/auxtrace.h
+++ b/tools/perf/util/auxtrace.h
@@ -378,7 +378,7 @@ struct addr_filters {
 static inline u64 auxtrace_mmap__read_snapshot_head(struct auxtrace_mmap *mm)
 {
 	struct perf_event_mmap_page *pc = mm->userpg;
-	u64 head = ACCESS_ONCE(pc->aux_head);
+	u64 head = READ_ONCE(pc->aux_head);
 
 	/* Ensure all reads are done after we read the head */
 	rmb();
@@ -389,7 +389,7 @@ static inline u64 auxtrace_mmap__read_head(struct auxtrace_mmap *mm)
 {
 	struct perf_event_mmap_page *pc = mm->userpg;
 #if BITS_PER_LONG == 64 || !defined(HAVE_SYNC_COMPARE_AND_SWAP_SUPPORT)
-	u64 head = ACCESS_ONCE(pc->aux_head);
+	u64 head = READ_ONCE(pc->aux_head);
 #else
 	u64 head = __sync_val_compare_and_swap(&pc->aux_head, 0, 0);
 #endif
diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
index bf2c4936e35f..b6767855d916 100644
--- a/tools/perf/util/evlist.h
+++ b/tools/perf/util/evlist.h
@@ -288,7 +288,7 @@ int perf_evlist__strerror_mmap(struct perf_evlist *evlist, int err, char *buf, s
 static inline u64 perf_mmap__read_head(struct perf_mmap *mm)
 {
 	struct perf_event_mmap_page *pc = mm->base;
-	u64 head = ACCESS_ONCE(pc->data_head);
+	u64 head = READ_ONCE(pc->data_head);
 	rmb();
 	return head;
 }
diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h
index 47b5e7dbcb18..aae9645c7122 100644
--- a/tools/perf/util/session.h
+++ b/tools/perf/util/session.h
@@ -113,7 +113,7 @@ int __perf_session__set_tracepoints_handlers(struct perf_session *session,
 
 extern volatile int session_done;
 
-#define session_done()	ACCESS_ONCE(session_done)
+#define session_done()	READ_ONCE(session_done)
 
 int perf_session__deliver_synth_event(struct perf_session *session,
 				      union perf_event *event,
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 9deb5a245b83..ce507ae1d4f5 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -2302,7 +2302,7 @@ void kvm_vcpu_on_spin(struct kvm_vcpu *me, bool yield_to_kernel_mode)
 				continue;
 			} else if (pass && i > last_boosted_vcpu)
 				break;
-			if (!ACCESS_ONCE(vcpu->preempted))
+			if (!READ_ONCE(vcpu->preempted))
 				continue;
 			if (vcpu == me)
 				continue;
-- 
2.5.2

Re: [GIT PULL de-alpha] Make core code not need to know about Alpha for v4.15

[Ingo Molnar]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> Mark Rutland (14):
>       dm integrity: Kill off ACCESS_ONCE()
>       EDAC, altera: Kill off ACCESS_ONCE()
>       firmware/ivc: Kill off ACCESS_ONCE()
>       fs: dcache: Kill off ACCESS_ONCE()
>       fs: ncpfs: Kill off ACCESS_ONCE()
>       media: dvb_ringbuffer: Kill off ACCESS_ONCE()
>       net: netlink/netfilter: Kill off ACCESS_ONCE()
>       net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()
>       net: average: Kill off ACCESS_ONCE()
>       samples: mic/mpssd/mpssd.c: Kill off ACCESS_ONCE()
>       selftests/powerpc: Kill off ACCESS_ONCE()
>       workqueue: Kill off ACCESS_ONCE()
>       rcutorture: formal: Prepare for ACCESS_ONCE() removal
>       COCCINELLE: treewide: kill off ACCESS_ONCE()
> 
> Paul E. McKenney (3):
>       doc: filesystems: Kill off ACCESS_ONCE()
>       mm: Kill off ACCESS_ONCE()

Quick question: are there any ACCESS_ONCE() users left after this? If not then we 
should add a separate patch that either removes it, or at least adds a build time 
warning which should trigger with new users.

Thanks,

	Ingo

Re: [GIT PULL de-alpha] Make core code not need to know about Alpha for v4.15

[Ingo Molnar]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> Mark Rutland (14):
>       dm integrity: Kill off ACCESS_ONCE()
>       EDAC, altera: Kill off ACCESS_ONCE()
>       firmware/ivc: Kill off ACCESS_ONCE()
>       fs: dcache: Kill off ACCESS_ONCE()
>       fs: ncpfs: Kill off ACCESS_ONCE()
>       media: dvb_ringbuffer: Kill off ACCESS_ONCE()
>       net: netlink/netfilter: Kill off ACCESS_ONCE()
>       net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()
>       net: average: Kill off ACCESS_ONCE()
>       samples: mic/mpssd/mpssd.c: Kill off ACCESS_ONCE()
>       selftests/powerpc: Kill off ACCESS_ONCE()
>       workqueue: Kill off ACCESS_ONCE()
>       rcutorture: formal: Prepare for ACCESS_ONCE() removal
>       COCCINELLE: treewide: kill off ACCESS_ONCE()
> 
> Paul E. McKenney (3):
>       doc: filesystems: Kill off ACCESS_ONCE()
>       mm: Kill off ACCESS_ONCE()

Quick question: are there any ACCESS_ONCE() users left after this? If not then we 
should add a separate patch that either removes it, or at least adds a build time 
warning which should trigger with new users.

Thanks,

	Ingo

Re: [GIT PULL de-alpha] Make core code not need to know about Alpha for v4.15

[Ingo Molnar]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> Hello, Ingo,
> 
> This series is a first step towards making the core kernel no longer
> need to consider DEC Alpha as a special case.  This is accomplished
> by two sets of patches, followed by a Coccinelle script:
> 
> 1.	Patches 1/19 through 15/19 in the following patches, which
> 	change non-Coccinelle-susceptible instances of ACCESS_ONCE()
> 	to either READ_ONCE() or WRITE_ONCE(), as appropriate.	Most of
> 	these patches are courtesy of Mark Rutland.
> 
> 2.	Patches 16/19 through 18/19 in the following patches, which
> 	add smp_read_barrier_depends() to READ_ONCE() and to Alpha's
> 	value-returning _release and _relaxed atomic read-modify-write
> 	operations, while also removing lockless_dereference() in favor
> 	of the modified READ_ONCE().
> 
> 3.	A patch generated by Coccinelle (script in commit log) that
> 	converts ACCESS_ONCE() to either READ_ONCE() or WRITE_ONCE(),
> 	as appropriate.  Of course, this patch should not be applied
> 	as-is to mainline, instead, the Coccinelle script should be
> 	re-run in order to account for any changes between now and
> 	the time of merging into mainline.
> 
> A later series (most likely targeted to v4.16) will remove instances of
> smp_read_barrier_depends() that are made redundant by this series, that
> is to say, almost all of them.
> 
> These changes are available in the git repository at:
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git rcu/alpha-cocci
> 
> for you to fetch changes up to d19fa5187a69bd7685e9c7452ce9e339b1b079fc:
> 
>   COCCINELLE: treewide: kill off ACCESS_ONCE() (2017-10-21 12:52:09 -0700)
> 
> ----------------------------------------------------------------
> Mark Rutland (14):
>       dm integrity: Kill off ACCESS_ONCE()
>       EDAC, altera: Kill off ACCESS_ONCE()
>       firmware/ivc: Kill off ACCESS_ONCE()
>       fs: dcache: Kill off ACCESS_ONCE()
>       fs: ncpfs: Kill off ACCESS_ONCE()
>       media: dvb_ringbuffer: Kill off ACCESS_ONCE()
>       net: netlink/netfilter: Kill off ACCESS_ONCE()
>       net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()
>       net: average: Kill off ACCESS_ONCE()
>       samples: mic/mpssd/mpssd.c: Kill off ACCESS_ONCE()
>       selftests/powerpc: Kill off ACCESS_ONCE()
>       workqueue: Kill off ACCESS_ONCE()
>       rcutorture: formal: Prepare for ACCESS_ONCE() removal
>       COCCINELLE: treewide: kill off ACCESS_ONCE()

None of the patches from Mark include your Signed-off-by - I'll add them, let me 
know if that's not OK. I suppose this happened because you rebased his tree?

Thanks,

	Ingo

Re: [GIT PULL de-alpha] Make core code not need to know about Alpha for v4.15

[Paul E. McKenney]

On Wed, Oct 25, 2017 at 10:45:48AM +0200, Ingo Molnar wrote:
> 
> * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> 
> > Hello, Ingo,
> > 
> > This series is a first step towards making the core kernel no longer
> > need to consider DEC Alpha as a special case.  This is accomplished
> > by two sets of patches, followed by a Coccinelle script:
> > 
> > 1.	Patches 1/19 through 15/19 in the following patches, which
> > 	change non-Coccinelle-susceptible instances of ACCESS_ONCE()
> > 	to either READ_ONCE() or WRITE_ONCE(), as appropriate.	Most of
> > 	these patches are courtesy of Mark Rutland.
> > 
> > 2.	Patches 16/19 through 18/19 in the following patches, which
> > 	add smp_read_barrier_depends() to READ_ONCE() and to Alpha's
> > 	value-returning _release and _relaxed atomic read-modify-write
> > 	operations, while also removing lockless_dereference() in favor
> > 	of the modified READ_ONCE().
> > 
> > 3.	A patch generated by Coccinelle (script in commit log) that
> > 	converts ACCESS_ONCE() to either READ_ONCE() or WRITE_ONCE(),
> > 	as appropriate.  Of course, this patch should not be applied
> > 	as-is to mainline, instead, the Coccinelle script should be
> > 	re-run in order to account for any changes between now and
> > 	the time of merging into mainline.
> > 
> > A later series (most likely targeted to v4.16) will remove instances of
> > smp_read_barrier_depends() that are made redundant by this series, that
> > is to say, almost all of them.
> > 
> > These changes are available in the git repository at:
> > 
> >   git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git rcu/alpha-cocci
> > 
> > for you to fetch changes up to d19fa5187a69bd7685e9c7452ce9e339b1b079fc:
> > 
> >   COCCINELLE: treewide: kill off ACCESS_ONCE() (2017-10-21 12:52:09 -0700)
> > 
> > ----------------------------------------------------------------
> > Mark Rutland (14):
> >       dm integrity: Kill off ACCESS_ONCE()
> >       EDAC, altera: Kill off ACCESS_ONCE()
> >       firmware/ivc: Kill off ACCESS_ONCE()
> >       fs: dcache: Kill off ACCESS_ONCE()
> >       fs: ncpfs: Kill off ACCESS_ONCE()
> >       media: dvb_ringbuffer: Kill off ACCESS_ONCE()
> >       net: netlink/netfilter: Kill off ACCESS_ONCE()
> >       net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()
> >       net: average: Kill off ACCESS_ONCE()
> >       samples: mic/mpssd/mpssd.c: Kill off ACCESS_ONCE()
> >       selftests/powerpc: Kill off ACCESS_ONCE()
> >       workqueue: Kill off ACCESS_ONCE()
> >       rcutorture: formal: Prepare for ACCESS_ONCE() removal
> >       COCCINELLE: treewide: kill off ACCESS_ONCE()
> 
> None of the patches from Mark include your Signed-off-by - I'll add them, let me 
> know if that's not OK. I suppose this happened because you rebased his tree?

I did pull his tree rather than applying from email, and no problem
adding my Signed-off-by.

								Thanx, Paul

Re: [GIT PULL de-alpha] Make core code not need to know about Alpha for v4.15

[Ingo Molnar]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> On Wed, Oct 25, 2017 at 10:45:48AM +0200, Ingo Molnar wrote:
> > 
> > * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> > 
> > > Hello, Ingo,
> > > 
> > > This series is a first step towards making the core kernel no longer
> > > need to consider DEC Alpha as a special case.  This is accomplished
> > > by two sets of patches, followed by a Coccinelle script:
> > > 
> > > 1.	Patches 1/19 through 15/19 in the following patches, which
> > > 	change non-Coccinelle-susceptible instances of ACCESS_ONCE()
> > > 	to either READ_ONCE() or WRITE_ONCE(), as appropriate.	Most of
> > > 	these patches are courtesy of Mark Rutland.
> > > 
> > > 2.	Patches 16/19 through 18/19 in the following patches, which
> > > 	add smp_read_barrier_depends() to READ_ONCE() and to Alpha's
> > > 	value-returning _release and _relaxed atomic read-modify-write
> > > 	operations, while also removing lockless_dereference() in favor
> > > 	of the modified READ_ONCE().
> > > 
> > > 3.	A patch generated by Coccinelle (script in commit log) that
> > > 	converts ACCESS_ONCE() to either READ_ONCE() or WRITE_ONCE(),
> > > 	as appropriate.  Of course, this patch should not be applied
> > > 	as-is to mainline, instead, the Coccinelle script should be
> > > 	re-run in order to account for any changes between now and
> > > 	the time of merging into mainline.
> > > 
> > > A later series (most likely targeted to v4.16) will remove instances of
> > > smp_read_barrier_depends() that are made redundant by this series, that
> > > is to say, almost all of them.
> > > 
> > > These changes are available in the git repository at:
> > > 
> > >   git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git rcu/alpha-cocci
> > > 
> > > for you to fetch changes up to d19fa5187a69bd7685e9c7452ce9e339b1b079fc:
> > > 
> > >   COCCINELLE: treewide: kill off ACCESS_ONCE() (2017-10-21 12:52:09 -0700)
> > > 
> > > ----------------------------------------------------------------
> > > Mark Rutland (14):
> > >       dm integrity: Kill off ACCESS_ONCE()
> > >       EDAC, altera: Kill off ACCESS_ONCE()
> > >       firmware/ivc: Kill off ACCESS_ONCE()
> > >       fs: dcache: Kill off ACCESS_ONCE()
> > >       fs: ncpfs: Kill off ACCESS_ONCE()
> > >       media: dvb_ringbuffer: Kill off ACCESS_ONCE()
> > >       net: netlink/netfilter: Kill off ACCESS_ONCE()
> > >       net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()
> > >       net: average: Kill off ACCESS_ONCE()
> > >       samples: mic/mpssd/mpssd.c: Kill off ACCESS_ONCE()
> > >       selftests/powerpc: Kill off ACCESS_ONCE()
> > >       workqueue: Kill off ACCESS_ONCE()
> > >       rcutorture: formal: Prepare for ACCESS_ONCE() removal
> > >       COCCINELLE: treewide: kill off ACCESS_ONCE()
> > 
> > None of the patches from Mark include your Signed-off-by - I'll add them, let me 
> > know if that's not OK. I suppose this happened because you rebased his tree?
> 
> I did pull his tree rather than applying from email, [...]

You might have pulled it, but the commits in rcu/alpha-cocci suggest that the tree 
was also rebased:

 commit d19fa5187a69bd7685e9c7452ce9e339b1b079fc
 Author:     Mark Rutland <mark.rutland@arm.com>
 AuthorDate: Thu Oct 19 10:42:03 2017 -0700
 Commit:     Paul E. McKenney <paulmck@linux.vnet.ibm.com>
 CommitDate: Sat Oct 21 12:52:09 2017 -0700

    COCCINELLE: treewide: kill off ACCESS_ONCE()

Note the "Commit:" field which says the commit was created by you, not by Mark.

Thanks,

	Ingo

Re: [GIT PULL de-alpha] Make core code not need to know about Alpha for v4.15

[Ingo Molnar]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> On Wed, Oct 25, 2017 at 10:45:48AM +0200, Ingo Molnar wrote:
> > 
> > * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> > 
> > > Hello, Ingo,
> > > 
> > > This series is a first step towards making the core kernel no longer
> > > need to consider DEC Alpha as a special case.  This is accomplished
> > > by two sets of patches, followed by a Coccinelle script:
> > > 
> > > 1.	Patches 1/19 through 15/19 in the following patches, which
> > > 	change non-Coccinelle-susceptible instances of ACCESS_ONCE()
> > > 	to either READ_ONCE() or WRITE_ONCE(), as appropriate.	Most of
> > > 	these patches are courtesy of Mark Rutland.
> > > 
> > > 2.	Patches 16/19 through 18/19 in the following patches, which
> > > 	add smp_read_barrier_depends() to READ_ONCE() and to Alpha's
> > > 	value-returning _release and _relaxed atomic read-modify-write
> > > 	operations, while also removing lockless_dereference() in favor
> > > 	of the modified READ_ONCE().
> > > 
> > > 3.	A patch generated by Coccinelle (script in commit log) that
> > > 	converts ACCESS_ONCE() to either READ_ONCE() or WRITE_ONCE(),
> > > 	as appropriate.  Of course, this patch should not be applied
> > > 	as-is to mainline, instead, the Coccinelle script should be
> > > 	re-run in order to account for any changes between now and
> > > 	the time of merging into mainline.
> > > 
> > > A later series (most likely targeted to v4.16) will remove instances of
> > > smp_read_barrier_depends() that are made redundant by this series, that
> > > is to say, almost all of them.
> > > 
> > > These changes are available in the git repository at:
> > > 
> > >   git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git rcu/alpha-cocci
> > > 
> > > for you to fetch changes up to d19fa5187a69bd7685e9c7452ce9e339b1b079fc:
> > > 
> > >   COCCINELLE: treewide: kill off ACCESS_ONCE() (2017-10-21 12:52:09 -0700)
> > > 
> > > ----------------------------------------------------------------
> > > Mark Rutland (14):
> > >       dm integrity: Kill off ACCESS_ONCE()
> > >       EDAC, altera: Kill off ACCESS_ONCE()
> > >       firmware/ivc: Kill off ACCESS_ONCE()
> > >       fs: dcache: Kill off ACCESS_ONCE()
> > >       fs: ncpfs: Kill off ACCESS_ONCE()
> > >       media: dvb_ringbuffer: Kill off ACCESS_ONCE()
> > >       net: netlink/netfilter: Kill off ACCESS_ONCE()
> > >       net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()
> > >       net: average: Kill off ACCESS_ONCE()
> > >       samples: mic/mpssd/mpssd.c: Kill off ACCESS_ONCE()
> > >       selftests/powerpc: Kill off ACCESS_ONCE()
> > >       workqueue: Kill off ACCESS_ONCE()
> > >       rcutorture: formal: Prepare for ACCESS_ONCE() removal
> > >       COCCINELLE: treewide: kill off ACCESS_ONCE()
> > 
> > None of the patches from Mark include your Signed-off-by - I'll add them, let me 
> > know if that's not OK. I suppose this happened because you rebased his tree?
> 
> I did pull his tree rather than applying from email, [...]

You might have pulled it, but the commits in rcu/alpha-cocci suggest that the tree 
was also rebased:

 commit d19fa5187a69bd7685e9c7452ce9e339b1b079fc
 Author:     Mark Rutland <mark.rutland@arm.com>
 AuthorDate: Thu Oct 19 10:42:03 2017 -0700
 Commit:     Paul E. McKenney <paulmck@linux.vnet.ibm.com>
 CommitDate: Sat Oct 21 12:52:09 2017 -0700

    COCCINELLE: treewide: kill off ACCESS_ONCE()

Note the "Commit:" field which says the commit was created by you, not by Mark.

Thanks,

	Ingo

Re: [GIT PULL de-alpha] Make core code not need to know about Alpha for v4.15

[Paul E. McKenney]

On Wed, Oct 25, 2017 at 11:04:43AM +0200, Ingo Molnar wrote:
> 
> * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> 
> > On Wed, Oct 25, 2017 at 10:45:48AM +0200, Ingo Molnar wrote:
> > > 
> > > * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> > > 
> > > > Hello, Ingo,
> > > > 
> > > > This series is a first step towards making the core kernel no longer
> > > > need to consider DEC Alpha as a special case.  This is accomplished
> > > > by two sets of patches, followed by a Coccinelle script:
> > > > 
> > > > 1.	Patches 1/19 through 15/19 in the following patches, which
> > > > 	change non-Coccinelle-susceptible instances of ACCESS_ONCE()
> > > > 	to either READ_ONCE() or WRITE_ONCE(), as appropriate.	Most of
> > > > 	these patches are courtesy of Mark Rutland.
> > > > 
> > > > 2.	Patches 16/19 through 18/19 in the following patches, which
> > > > 	add smp_read_barrier_depends() to READ_ONCE() and to Alpha's
> > > > 	value-returning _release and _relaxed atomic read-modify-write
> > > > 	operations, while also removing lockless_dereference() in favor
> > > > 	of the modified READ_ONCE().
> > > > 
> > > > 3.	A patch generated by Coccinelle (script in commit log) that
> > > > 	converts ACCESS_ONCE() to either READ_ONCE() or WRITE_ONCE(),
> > > > 	as appropriate.  Of course, this patch should not be applied
> > > > 	as-is to mainline, instead, the Coccinelle script should be
> > > > 	re-run in order to account for any changes between now and
> > > > 	the time of merging into mainline.
> > > > 
> > > > A later series (most likely targeted to v4.16) will remove instances of
> > > > smp_read_barrier_depends() that are made redundant by this series, that
> > > > is to say, almost all of them.
> > > > 
> > > > These changes are available in the git repository at:
> > > > 
> > > >   git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git rcu/alpha-cocci
> > > > 
> > > > for you to fetch changes up to d19fa5187a69bd7685e9c7452ce9e339b1b079fc:
> > > > 
> > > >   COCCINELLE: treewide: kill off ACCESS_ONCE() (2017-10-21 12:52:09 -0700)
> > > > 
> > > > ----------------------------------------------------------------
> > > > Mark Rutland (14):
> > > >       dm integrity: Kill off ACCESS_ONCE()
> > > >       EDAC, altera: Kill off ACCESS_ONCE()
> > > >       firmware/ivc: Kill off ACCESS_ONCE()
> > > >       fs: dcache: Kill off ACCESS_ONCE()
> > > >       fs: ncpfs: Kill off ACCESS_ONCE()
> > > >       media: dvb_ringbuffer: Kill off ACCESS_ONCE()
> > > >       net: netlink/netfilter: Kill off ACCESS_ONCE()
> > > >       net/ipv4/tcp_input.c: Kill off ACCESS_ONCE()
> > > >       net: average: Kill off ACCESS_ONCE()
> > > >       samples: mic/mpssd/mpssd.c: Kill off ACCESS_ONCE()
> > > >       selftests/powerpc: Kill off ACCESS_ONCE()
> > > >       workqueue: Kill off ACCESS_ONCE()
> > > >       rcutorture: formal: Prepare for ACCESS_ONCE() removal
> > > >       COCCINELLE: treewide: kill off ACCESS_ONCE()
> > > 
> > > None of the patches from Mark include your Signed-off-by - I'll add them, let me 
> > > know if that's not OK. I suppose this happened because you rebased his tree?
> > 
> > I did pull his tree rather than applying from email, [...]
> 
> You might have pulled it, but the commits in rcu/alpha-cocci suggest that the tree 
> was also rebased:
> 
>  commit d19fa5187a69bd7685e9c7452ce9e339b1b079fc
>  Author:     Mark Rutland <mark.rutland@arm.com>
>  AuthorDate: Thu Oct 19 10:42:03 2017 -0700
>  Commit:     Paul E. McKenney <paulmck@linux.vnet.ibm.com>
>  CommitDate: Sat Oct 21 12:52:09 2017 -0700
> 
>     COCCINELLE: treewide: kill off ACCESS_ONCE()
> 
> Note the "Commit:" field which says the commit was created by you, not by Mark.

Hmmm...  Looking back through my old versions, it looks like my original
pull from Mark's git tree did the rebase.  Not sure how that happened,
given that both are based on v4.14-rc4.  I am going to have to proceed
more carefully next time!

							Thanx, Paul