qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Michael Roth <mdroth@linux.vnet.ibm.com>
To: qemu-devel@nongnu.org
Cc: qemu-stable@nongnu.org, Nicholas Piggin <npiggin@gmail.com>,
	David Gibson <david@gibson.dropbear.id.au>
Subject: [PATCH 68/78] target/ppc: Fix mtmsr(d) L=1 variant that loses interrupts
Date: Tue, 16 Jun 2020 09:15:37 -0500	[thread overview]
Message-ID: <20200616141547.24664-69-mdroth@linux.vnet.ibm.com> (raw)
In-Reply-To: <20200616141547.24664-1-mdroth@linux.vnet.ibm.com>

From: Nicholas Piggin <npiggin@gmail.com>

If mtmsr L=1 sets MSR[EE] while there is a maskable exception pending,
it does not cause an interrupt. This causes the test case to hang:

https://lists.gnu.org/archive/html/qemu-ppc/2019-10/msg00826.html

More recently, Linux reduced the occurance of operations (e.g., rfi)
which stop translation and allow pending interrupts to be processed.
This started causing hangs in Linux boot in long-running kernel tests,
running with '-d int' shows the decrementer stops firing despite DEC
wrapping and MSR[EE]=1.

https://lists.ozlabs.org/pipermail/linuxppc-dev/2020-April/208301.html

The cause is the broken mtmsr L=1 behaviour, which is contrary to the
architecture. From Power ISA v3.0B, p.977, Move To Machine State Register,
Programming Note states:

    If MSR[EE]=0 and an External, Decrementer, or Performance Monitor
    exception is pending, executing an mtmsrd instruction that sets
    MSR[EE] to 1 will cause the interrupt to occur before the next
    instruction is executed, if no higher priority exception exists

Fix this by handling L=1 exactly the same way as L=0, modulo the MSR
bits altered.

The confusion arises from L=0 being "context synchronizing" whereas L=1
is "execution synchronizing", which is a weaker semantic. However this
is not a relaxation of the requirement that these exceptions cause
interrupts when MSR[EE]=1 (e.g., when mtmsr executes to completion as
TCG is doing here), rather it specifies how a pipelined processor can
have multiple instructions in flight where one may influence how another
behaves.

Cc: qemu-stable@nongnu.org
Reported-by: Anton Blanchard <anton@ozlabs.org>
Reported-by: Nathan Chancellor <natechancellor@gmail.com>
Tested-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-Id: <20200414111131.465560-1-npiggin@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Tested-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
(cherry picked from commit 5ed195065cc6895f61b9d59bfa0a0536ed5ed51e)
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
---
 target/ppc/translate.c | 46 +++++++++++++++++++++++++-----------------
 1 file changed, 27 insertions(+), 19 deletions(-)

diff --git a/target/ppc/translate.c b/target/ppc/translate.c
index f87f6eeaf7..4f5008ed6f 100644
--- a/target/ppc/translate.c
+++ b/target/ppc/translate.c
@@ -4361,30 +4361,34 @@ static void gen_mtmsrd(DisasContext *ctx)
     CHK_SV;
 
 #if !defined(CONFIG_USER_ONLY)
+    if (tb_cflags(ctx->base.tb) & CF_USE_ICOUNT) {
+        gen_io_start();
+    }
     if (ctx->opcode & 0x00010000) {
-        /* Special form that does not need any synchronisation */
+        /* L=1 form only updates EE and RI */
         TCGv t0 = tcg_temp_new();
+        TCGv t1 = tcg_temp_new();
         tcg_gen_andi_tl(t0, cpu_gpr[rS(ctx->opcode)],
                         (1 << MSR_RI) | (1 << MSR_EE));
-        tcg_gen_andi_tl(cpu_msr, cpu_msr,
+        tcg_gen_andi_tl(t1, cpu_msr,
                         ~(target_ulong)((1 << MSR_RI) | (1 << MSR_EE)));
-        tcg_gen_or_tl(cpu_msr, cpu_msr, t0);
+        tcg_gen_or_tl(t1, t1, t0);
+
+        gen_helper_store_msr(cpu_env, t1);
         tcg_temp_free(t0);
+        tcg_temp_free(t1);
+
     } else {
         /*
          * XXX: we need to update nip before the store if we enter
          *      power saving mode, we will exit the loop directly from
          *      ppc_store_msr
          */
-        if (tb_cflags(ctx->base.tb) & CF_USE_ICOUNT) {
-            gen_io_start();
-        }
         gen_update_nip(ctx, ctx->base.pc_next);
         gen_helper_store_msr(cpu_env, cpu_gpr[rS(ctx->opcode)]);
-        /* Must stop the translation as machine state (may have) changed */
-        /* Note that mtmsr is not always defined as context-synchronizing */
-        gen_stop_exception(ctx);
     }
+    /* Must stop the translation as machine state (may have) changed */
+    gen_stop_exception(ctx);
 #endif /* !defined(CONFIG_USER_ONLY) */
 }
 #endif /* defined(TARGET_PPC64) */
@@ -4394,15 +4398,23 @@ static void gen_mtmsr(DisasContext *ctx)
     CHK_SV;
 
 #if !defined(CONFIG_USER_ONLY)
-   if (ctx->opcode & 0x00010000) {
-        /* Special form that does not need any synchronisation */
+    if (tb_cflags(ctx->base.tb) & CF_USE_ICOUNT) {
+        gen_io_start();
+    }
+    if (ctx->opcode & 0x00010000) {
+        /* L=1 form only updates EE and RI */
         TCGv t0 = tcg_temp_new();
+        TCGv t1 = tcg_temp_new();
         tcg_gen_andi_tl(t0, cpu_gpr[rS(ctx->opcode)],
                         (1 << MSR_RI) | (1 << MSR_EE));
-        tcg_gen_andi_tl(cpu_msr, cpu_msr,
+        tcg_gen_andi_tl(t1, cpu_msr,
                         ~(target_ulong)((1 << MSR_RI) | (1 << MSR_EE)));
-        tcg_gen_or_tl(cpu_msr, cpu_msr, t0);
+        tcg_gen_or_tl(t1, t1, t0);
+
+        gen_helper_store_msr(cpu_env, t1);
         tcg_temp_free(t0);
+        tcg_temp_free(t1);
+
     } else {
         TCGv msr = tcg_temp_new();
 
@@ -4411,9 +4423,6 @@ static void gen_mtmsr(DisasContext *ctx)
          *      power saving mode, we will exit the loop directly from
          *      ppc_store_msr
          */
-        if (tb_cflags(ctx->base.tb) & CF_USE_ICOUNT) {
-            gen_io_start();
-        }
         gen_update_nip(ctx, ctx->base.pc_next);
 #if defined(TARGET_PPC64)
         tcg_gen_deposit_tl(msr, cpu_msr, cpu_gpr[rS(ctx->opcode)], 0, 32);
@@ -4422,10 +4431,9 @@ static void gen_mtmsr(DisasContext *ctx)
 #endif
         gen_helper_store_msr(cpu_env, msr);
         tcg_temp_free(msr);
-        /* Must stop the translation as machine state (may have) changed */
-        /* Note that mtmsr is not always defined as context-synchronizing */
-        gen_stop_exception(ctx);
     }
+    /* Must stop the translation as machine state (may have) changed */
+    gen_stop_exception(ctx);
 #endif
 }
 
-- 
2.17.1



  parent reply	other threads:[~2020-06-16 15:05 UTC|newest]

Thread overview: 92+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-16 14:14 [PATCH 00/78] Patch Round-up for stable 4.2.1, freeze on 2020-06-22 Michael Roth
2020-06-16 14:14 ` [PATCH 01/78] block/nbd: extract the common cleanup code Michael Roth
2020-06-16 14:14 ` [PATCH 02/78] block/nbd: fix memory leak in nbd_open() Michael Roth
2020-06-16 14:14 ` [PATCH 03/78] i386: Resolve CPU models to v1 by default Michael Roth
2020-06-16 14:14 ` [PATCH 04/78] qapi: better document NVMe blockdev @device parameter Michael Roth
2020-06-16 14:14 ` [PATCH 05/78] target/arm: ensure we use current exception state after SCR update Michael Roth
2020-06-16 14:14 ` [PATCH 06/78] block: Activate recursively even for already active nodes Michael Roth
2020-06-16 14:14 ` [PATCH 07/78] virtio-blk: fix out-of-bounds access to bitmap in notify_guest_bh Michael Roth
2020-06-16 14:14 ` [PATCH 08/78] numa: remove not needed check Michael Roth
2020-06-16 14:14 ` [PATCH 09/78] numa: properly check if numa is supported Michael Roth
2020-06-16 14:14 ` [PATCH 10/78] backup-top: Begin drain earlier Michael Roth
2020-06-16 14:14 ` [PATCH 11/78] arm/arm-powerctl: set NSACR.{CP11, CP10} bits in arm_set_cpu_on() Michael Roth
2020-06-16 14:14 ` [PATCH 12/78] arm/arm-powerctl: rebuild hflags after setting CP15 " Michael Roth
2020-06-16 14:14 ` [PATCH 13/78] hw/i386/pc: fix regression in parsing vga cmdline parameter Michael Roth
2020-06-16 14:14 ` [PATCH 14/78] tests/ide-test: Create a single unit-test covering more PRDT cases Michael Roth
2020-06-16 14:14 ` [PATCH 15/78] ide: Fix incorrect handling of some PRDTs in ide_dma_cb() Michael Roth
2020-06-16 14:14 ` [PATCH 16/78] target/arm: Set ISSIs16Bit in make_issinfo Michael Roth
2020-06-16 14:14 ` [PATCH 17/78] virtio: update queue size on guest write Michael Roth
2020-06-16 14:14 ` [PATCH 18/78] virtio-mmio: " Michael Roth
2020-06-16 14:14 ` [PATCH 19/78] virtio: add ability to delete vq through a pointer Michael Roth
2020-06-16 14:14 ` [PATCH 20/78] virtio: make virtio_delete_queue idempotent Michael Roth
2020-06-16 14:14 ` [PATCH 21/78] virtio: reset region cache when on queue deletion Michael Roth
2020-06-16 14:14 ` [PATCH 22/78] virtio-net: delete also control queue when TX/RX deleted Michael Roth
2020-06-16 14:14 ` [PATCH 23/78] intel_iommu: a fix to vtd_find_as_from_bus_num() Michael Roth
2020-06-16 14:14 ` [PATCH 24/78] intel_iommu: add present bit check for pasid table entries Michael Roth
2020-06-16 14:14 ` [PATCH 25/78] vfio/pci: Don't remove irqchip notifier if not registered Michael Roth
2020-06-16 14:14 ` [PATCH 26/78] qcow2-bitmaps: fix qcow2_can_store_new_dirty_bitmap Michael Roth
2020-06-16 14:14 ` [PATCH 27/78] dp8393x: Mask EOL bit from descriptor addresses Michael Roth
2020-06-16 14:14 ` [PATCH 28/78] dp8393x: Always use 32-bit accesses Michael Roth
2020-06-16 14:14 ` [PATCH 29/78] dp8393x: Clean up endianness hacks Michael Roth
2020-06-16 14:14 ` [PATCH 30/78] dp8393x: Have dp8393x_receive() return the packet size Michael Roth
2020-06-16 14:15 ` [PATCH 31/78] dp8393x: Update LLFA and CRDA registers from rx descriptor Michael Roth
2020-06-16 14:15 ` [PATCH 32/78] dp8393x: Clear RRRA command register bit only when appropriate Michael Roth
2020-06-16 14:15 ` [PATCH 33/78] dp8393x: Implement packet size limit and RBAE interrupt Michael Roth
2020-06-16 14:15 ` [PATCH 34/78] dp8393x: Don't clobber packet checksum Michael Roth
2020-06-16 14:15 ` [PATCH 35/78] dp8393x: Use long-word-aligned RRA pointers in 32-bit mode Michael Roth
2020-06-16 14:15 ` [PATCH 36/78] dp8393x: Pad frames to word or long word boundary Michael Roth
2020-06-16 14:15 ` [PATCH 37/78] dp8393x: Clear descriptor in_use field to release packet Michael Roth
2020-06-16 14:15 ` [PATCH 38/78] dp8393x: Always update RRA pointers and sequence numbers Michael Roth
2020-06-16 14:15 ` [PATCH 39/78] dp8393x: Don't reset Silicon Revision register Michael Roth
2020-06-16 14:15 ` [PATCH 40/78] dp8393x: Don't stop reception upon RBE interrupt assertion Michael Roth
2020-06-16 14:15 ` [PATCH 41/78] s390/sclp: improve special wait psw logic Michael Roth
2020-06-16 14:15 ` [PATCH 42/78] plugins/core: add missing break in cb_to_tcg_flags Michael Roth
2020-06-16 14:15 ` [PATCH 43/78] tcg: save vaddr temp for plugin usage Michael Roth
2020-06-16 14:15 ` [PATCH 44/78] qcow2: update_refcount(): Reset old_table_index after qcow2_cache_put() Michael Roth
2020-06-16 14:15 ` [PATCH 45/78] qcow2: Fix qcow2_alloc_cluster_abort() for external data file Michael Roth
2020-06-16 14:15 ` [PATCH 46/78] iotests: Test copy offloading with " Michael Roth
2020-06-16 14:15 ` [PATCH 47/78] qcow2: Fix alloc_cluster_abort() for pre-existing clusters Michael Roth
2020-06-16 14:15 ` [PATCH 48/78] iotests/026: Test EIO on preallocated zero cluster Michael Roth
2020-06-16 14:15 ` [PATCH 49/78] iotests/026: Test EIO on allocation in a data-file Michael Roth
2020-06-16 14:15 ` [PATCH 50/78] virtio: gracefully handle invalid region caches Michael Roth
2020-06-16 14:15 ` [PATCH 51/78] scsi/qemu-pr-helper: Fix out-of-bounds access to trnptid_list[] Michael Roth
2020-06-16 14:15 ` [PATCH 52/78] block/qcow2-threads: fix qcow2_decompress Michael Roth
2020-06-16 14:15 ` [PATCH 53/78] job: refactor progress to separate object Michael Roth
2020-06-16 14:15 ` [PATCH 54/78] block/block-copy: fix progress calculation Michael Roth
2020-06-16 14:15 ` [PATCH 55/78] target/ppc: Fix rlwinm on ppc64 Michael Roth
2020-06-16 14:15 ` [PATCH 56/78] block/io: fix bdrv_co_do_copy_on_readv Michael Roth
2020-06-16 14:15 ` [PATCH 57/78] compat: disable edid on correct virtio-gpu device Michael Roth
2020-06-16 14:15 ` [PATCH 58/78] qga: Installer: Wait for installation to finish Michael Roth
2020-06-16 14:15 ` [PATCH 59/78] qga-win: Handle VSS_E_PROVIDER_ALREADY_REGISTERED error Michael Roth
2020-06-16 14:15 ` [PATCH 60/78] qga-win: prevent crash when executing guest-file-read with large count Michael Roth
2020-06-16 14:15 ` [PATCH 61/78] qga: Fix undefined C behavior Michael Roth
2020-06-16 14:15 ` [PATCH 62/78] qemu-ga: document vsock-listen in the man page Michael Roth
2020-06-16 14:15 ` [PATCH 63/78] hw/i386/amd_iommu.c: Fix corruption of log events passed to guest Michael Roth
2020-06-16 14:15 ` [PATCH 64/78] tcg/i386: Fix INDEX_op_dup2_vec Michael Roth
2020-06-16 14:15 ` [PATCH 65/78] dump: Fix writing of ELF section Michael Roth
2020-06-16 14:15 ` [PATCH 66/78] xen-block: Fix double qlist remove and request leak Michael Roth
2020-06-16 14:15 ` [PATCH 67/78] vhost-user-gpu: Release memory returned by vu_queue_pop() with free() Michael Roth
2020-06-16 14:15 ` Michael Roth [this message]
2020-06-16 14:15 ` [PATCH 69/78] hostmem: don't use mbind() if host-nodes is empty Michael Roth
2020-06-16 14:15 ` [PATCH 70/78] target/arm: Clear tail in gvec_fmul_idx_*, gvec_fmla_idx_* Michael Roth
2020-06-16 14:15 ` [PATCH 71/78] qemu-nbd: Close inherited stderr Michael Roth
2020-06-16 14:15 ` [PATCH 72/78] 9p: Lock directory streams with a CoMutex Michael Roth
2020-06-16 15:14   ` Greg Kurz
2020-06-16 16:09     ` Christian Schoenebeck
2020-06-16 16:41       ` Greg Kurz
2020-06-16 22:46         ` Michael Roth
2020-06-18 13:47           ` Christian Schoenebeck
2020-06-16 14:15 ` [PATCH 73/78] net: Do not include a newline in the id of -nic devices Michael Roth
2020-06-16 14:15 ` [PATCH 74/78] nbd/server: Avoid long error message assertions CVE-2020-10761 Michael Roth
2020-06-16 14:15 ` [PATCH 75/78] virtio-balloon: fix free page hinting without an iothread Michael Roth
2020-06-16 14:15 ` [PATCH 76/78] virtio-balloon: fix free page hinting check on unrealize Michael Roth
2020-06-16 14:15 ` [PATCH 77/78] virtio-balloon: unref the iothread when unrealizing Michael Roth
2020-06-16 14:15 ` [PATCH 78/78] block: Call attention to truncation of long NBD exports Michael Roth
2020-06-17 14:39 ` [PATCH 00/78] Patch Round-up for stable 4.2.1, freeze on 2020-06-22 Cole Robinson
2020-06-17 15:54 ` Liam Merwick
2020-06-17 20:02 ` Karl Heubaum
2020-06-20  0:14 ` Finn Thain
2020-06-20  3:39   ` Finn Thain
2020-06-22 20:31     ` Michael Roth
2020-06-20 21:44 ` Bruce Rogers
2020-06-22 20:26 ` Michael Roth

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200616141547.24664-69-mdroth@linux.vnet.ibm.com \
    --to=mdroth@linux.vnet.ibm.com \
    --cc=david@gibson.dropbear.id.au \
    --cc=npiggin@gmail.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-stable@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).