stable.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Alex Elder <elder@linaro.org>,
	Ilya Dryomov <idryomov@gmail.com>
Subject: [PATCH 4.1 55/84] rbd: fix copyup completion race
Date: Fri, 14 Aug 2015 10:42:23 -0700	[thread overview]
Message-ID: <20150814174211.886050670@linuxfoundation.org> (raw)
In-Reply-To: <20150814174210.214822912@linuxfoundation.org>

4.1-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Ilya Dryomov <idryomov@gmail.com>

commit 2761713d35e370fd640b5781109f753066b746c4 upstream.

For write/discard obj_requests that involved a copyup method call, the
opcode of the first op is CEPH_OSD_OP_CALL and the ->callback is
rbd_img_obj_copyup_callback().  The latter frees copyup pages, sets
->xferred and delegates to rbd_img_obj_callback(), the "normal" image
object callback, for reporting to block layer and putting refs.

rbd_osd_req_callback() however treats CEPH_OSD_OP_CALL as a trivial op,
which means obj_request is marked done in rbd_osd_trivial_callback(),
*before* ->callback is invoked and rbd_img_obj_copyup_callback() has
a chance to run.  Marking obj_request done essentially means giving
rbd_img_obj_callback() a license to end it at any moment, so if another
obj_request from the same img_request is being completed concurrently,
rbd_img_obj_end_request() may very well be called on such prematurally
marked done request:

<obj_request-1/2 reply>
handle_reply()
  rbd_osd_req_callback()
    rbd_osd_trivial_callback()
    rbd_obj_request_complete()
    rbd_img_obj_copyup_callback()
    rbd_img_obj_callback()
                                    <obj_request-2/2 reply>
                                    handle_reply()
                                      rbd_osd_req_callback()
                                        rbd_osd_trivial_callback()
      for_each_obj_request(obj_request->img_request) {
        rbd_img_obj_end_request(obj_request-1/2)
        rbd_img_obj_end_request(obj_request-2/2) <--
      }

Calling rbd_img_obj_end_request() on such a request leads to trouble,
in particular because its ->xfferred is 0.  We report 0 to the block
layer with blk_update_request(), get back 1 for "this request has more
data in flight" and then trip on

    rbd_assert(more ^ (which == img_request->obj_request_count));

with rhs (which == ...) being 1 because rbd_img_obj_end_request() has
been called for both requests and lhs (more) being 1 because we haven't
got a chance to set ->xfferred in rbd_img_obj_copyup_callback() yet.

To fix this, leverage that rbd wants to call class methods in only two
cases: one is a generic method call wrapper (obj_request is standalone)
and the other is a copyup (obj_request is part of an img_request).  So
make a dedicated handler for CEPH_OSD_OP_CALL and directly invoke
rbd_img_obj_copyup_callback() from it if obj_request is part of an
img_request, similar to how CEPH_OSD_OP_READ handler invokes
rbd_img_obj_request_read_callback().

Since rbd_img_obj_copyup_callback() is now being called from the OSD
request callback (only), it is renamed to rbd_osd_copyup_callback().

Cc: Alex Elder <elder@linaro.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/block/rbd.c |   22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)

--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -522,6 +522,7 @@ void rbd_warn(struct rbd_device *rbd_dev
 #  define rbd_assert(expr)	((void) 0)
 #endif /* !RBD_DEBUG */
 
+static void rbd_osd_copyup_callback(struct rbd_obj_request *obj_request);
 static int rbd_img_obj_request_submit(struct rbd_obj_request *obj_request);
 static void rbd_img_parent_read(struct rbd_obj_request *obj_request);
 static void rbd_dev_remove_parent(struct rbd_device *rbd_dev);
@@ -1797,6 +1798,16 @@ static void rbd_osd_stat_callback(struct
 	obj_request_done_set(obj_request);
 }
 
+static void rbd_osd_call_callback(struct rbd_obj_request *obj_request)
+{
+	dout("%s: obj %p\n", __func__, obj_request);
+
+	if (obj_request_img_data_test(obj_request))
+		rbd_osd_copyup_callback(obj_request);
+	else
+		obj_request_done_set(obj_request);
+}
+
 static void rbd_osd_req_callback(struct ceph_osd_request *osd_req,
 				struct ceph_msg *msg)
 {
@@ -1845,6 +1856,8 @@ static void rbd_osd_req_callback(struct
 		rbd_osd_discard_callback(obj_request);
 		break;
 	case CEPH_OSD_OP_CALL:
+		rbd_osd_call_callback(obj_request);
+		break;
 	case CEPH_OSD_OP_NOTIFY_ACK:
 	case CEPH_OSD_OP_WATCH:
 		rbd_osd_trivial_callback(obj_request);
@@ -2509,13 +2522,15 @@ out_unwind:
 }
 
 static void
-rbd_img_obj_copyup_callback(struct rbd_obj_request *obj_request)
+rbd_osd_copyup_callback(struct rbd_obj_request *obj_request)
 {
 	struct rbd_img_request *img_request;
 	struct rbd_device *rbd_dev;
 	struct page **pages;
 	u32 page_count;
 
+	dout("%s: obj %p\n", __func__, obj_request);
+
 	rbd_assert(obj_request->type == OBJ_REQUEST_BIO ||
 		obj_request->type == OBJ_REQUEST_NODATA);
 	rbd_assert(obj_request_img_data_test(obj_request));
@@ -2542,9 +2557,7 @@ rbd_img_obj_copyup_callback(struct rbd_o
 	if (!obj_request->result)
 		obj_request->xferred = obj_request->length;
 
-	/* Finish up with the normal image object callback */
-
-	rbd_img_obj_callback(obj_request);
+	obj_request_done_set(obj_request);
 }
 
 static void
@@ -2629,7 +2642,6 @@ rbd_img_obj_parent_read_full_callback(st
 
 	/* All set, send it off. */
 
-	orig_request->callback = rbd_img_obj_copyup_callback;
 	osdc = &rbd_dev->rbd_client->client->osdc;
 	img_result = rbd_obj_request_submit(osdc, orig_request);
 	if (!img_result)



  parent reply	other threads:[~2015-08-14 17:43 UTC|newest]

Thread overview: 83+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-08-14 17:41 [PATCH 4.1 00/84] 4.1.6-stable review Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 01/84] MIPS: unaligned: Fix build error on big endian R6 kernels Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 02/84] MIPS: Replace add and sub instructions in relocate_kernel.S with addiu Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 03/84] MIPS: Malta: Dont reinitialise RTC Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 04/84] MIPS: Fix sched_getaffinity with MT FPAFF enabled Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 05/84] MIPS: Export get_c0_perfcount_int() Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 06/84] MIPS: do_mcheck: Fix kernel code dump with EVA Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 07/84] MIPS: show_stack: Fix stack trace " Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 09/84] MIPS: Flush RPS on kernel entry " Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 10/84] MIPS: Make set_pte() SMP safe Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 11/84] fsnotify: fix oops in fsnotify_clear_marks_by_group_flags() Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 12/84] drm/i915: Declare the swizzling unknown for L-shaped configurations Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 14/84] drm/radeon: rework audio detect (v4) Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 15/84] drm/radeon/combios: add some validation of lvds values Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 16/84] drm/dp-mst: Remove debug WARN_ON Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 17/84] ipr: Fix locking for unit attention handling Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 18/84] ipr: Fix incorrect trace indexing Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 19/84] ipr: Fix invalid array indexing for HRRQ Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 20/84] Bluetooth: Fix NULL pointer dereference in smp_conn_security Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 21/84] dmaengine: pl330: Fix overflow when reporting residue in memcpy Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 22/84] dmaengine: pl330: Really fix choppy sound because of wrong residue calculation Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 24/84] drivers/usb: Delete XHCI command timer if necessary Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 25/84] USB: sierra: add 1199:68AB device ID Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 26/84] usb: udc: core: add device_del() call to error pathway Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 27/84] usb: chipidea: ehci_init_driver is intended to call one time Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 28/84] phy: twl4030-usb: make runtime pm more reliable Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 29/84] md: use kzalloc() when bitmap is disabled Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 30/84] ath10k: fix qca61x4 hw2.1 support Greg Kroah-Hartman
2015-08-14 17:41 ` [PATCH 4.1 31/84] x86/asm/entry/64: Remove pointless jump to irq_return Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 32/84] x86/nmi: Enable nested do_nmi() handling for 64-bit kernels Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 33/84] x86/nmi/64: Remove asm code that saves CR2 Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 34/84] x86/nmi/64: Switch stacks on userspace NMI entry Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 35/84] x86/nmi/64: Improve nested NMI comments Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 36/84] x86/nmi/64: Reorder nested NMI checks Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 37/84] x86/nmi/64: Use DF to avoid userspace RSP confusing nested NMI detection Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 38/84] dmaengine: at_xdmac: fix transfer data width in at_xdmac_prep_slave_sg() Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 39/84] crypto: nx - Fixing NX data alignment with nx_sg list Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 40/84] crypto: nx - Fixing SHA update bug Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 41/84] crypto: nx - Fix reentrancy bugs Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 42/84] [PATCH] sparc64: Fix userspace FPU register corruptions Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 43/84] clk: keystone: add support for post divider register for main pll Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 44/84] ARM: dts: keystone: fix dt bindings to use post div register for mainpll Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 45/84] ASoC: Intel: Get correct usage_count value to load firmware Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 46/84] ASoC: ssm4567: Keep TDM_BCLKS in ssm4567_set_dai_fmt Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 47/84] ASoC: pcm1681: Fix setting de-emphasis sampling rate selection Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 48/84] ASoC: dapm: Lock during userspace access Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 49/84] ASoC: dapm: Dont add prefix to widget stream name Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 50/84] x86/xen: Probe target addresses in set_aliased_prot() before the hypercall Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 52/84] hwrng: core - correct error check of kthread_run call Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 53/84] crypto: qat - Fix invalid synchronization between register/unregister sym algs Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 54/84] crypto: ixp4xx - Remove bogus BUG_ON on scattered dst buffer Greg Kroah-Hartman
2015-08-14 17:42 ` Greg Kroah-Hartman [this message]
2015-08-14 17:42 ` [PATCH 4.1 56/84] ARM: dts: i.MX35: Fix can support Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 57/84] ARM: OMAP2+: hwmod: Fix _wait_target_ready() for hwmods without sysc Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 58/84] ALSA: hda - fix cs4210_spdif_automute() Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 59/84] ALSA: hda - one Dell machine needs the headphone white noise fixup Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 60/84] ALSA: fireworks/firewire-lib: add support for recent firmware quirk Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 61/84] hwmon: (nct7904) Export I2C module alias information Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 63/84] ipc: modify message queue accounting to not take kernel data structures into account Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 64/84] ocfs2: fix BUG in ocfs2_downconvert_thread_do_work() Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 65/84] ocfs2: fix shift left overflow Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 66/84] nfsd: Drop BUG_ON and ignore SECLABEL on absent filesystem Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 67/84] PCI: Restore PCI_MSIX_FLAGS_BIRMASK definition Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 68/84] md/raid1: extend spinlock to protect raid1_end_read_request against inconsistencies Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 69/84] dm: fix dm_merge_bvec regression on 32 bit systems Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 70/84] staging: vt6655: vnt_bss_info_changed check conf->beacon_rate is not NULL Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 71/84] staging: lustre: Include unaligned.h instead of access_ok.h Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 72/84] usb: gadget: f_uac2: fix calculation of uac2->p_interval Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 73/84] USB: qcserial/option: make AT URCs work for Sierra Wireless MC7305/MC7355 Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 75/84] mtd: nand: Fix NAND_USE_BOUNCE_BUFFER flag conflict Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 76/84] Input: alps - only Dell laptops have separate button bits for v2 dualpoint sticks Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 77/84] thermal: exynos: Disable the regulator on probe failure Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 78/84] mm, vmscan: Do not wait for page writeback for GFP_NOFS allocations Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 79/84] signalfd: fix information leak in signalfd_copyinfo Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 80/84] signal: fix information leak in copy_siginfo_to_user Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 81/84] signal: fix information leak in copy_siginfo_from_user32 Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 82/84] kvm: x86: fix kvm_apic_has_events to check for NULL pointer Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 83/84] nfsd: refactor nfs4_preprocess_stateid_op Greg Kroah-Hartman
2015-08-14 17:42 ` [PATCH 4.1 84/84] nfsd: do nfs4_check_fh in nfs4_check_file instead of nfs4_check_olstateid Greg Kroah-Hartman
2015-08-15  0:10 ` [PATCH 4.1 00/84] 4.1.6-stable review Shuah Khan
2015-08-15  0:46   ` Greg Kroah-Hartman
2015-08-15 15:21 ` Guenter Roeck
2015-08-15 16:40   ` Greg Kroah-Hartman

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=20150814174211.886050670@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=elder@linaro.org \
    --cc=idryomov@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stable@vger.kernel.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).