[5.10,09/86] dm: fix mempool NULL pointer race when completing IO

From: Jiazi Li <jqqlijiazi@gmail.com>

From: Jiazi Li <jqqlijiazi@gmail.com>

commit d208b89401e073de986dc891037c5a668f5d5d95 upstream.

dm_io_dec_pending() calls end_io_acct() first and will then dec md
in-flight pending count. But if a task is swapping DM table at same
time this can result in a crash due to mempool->elements being NULL:

task1                             task2
do_resume
 ->do_suspend
  ->dm_wait_for_completion
                                  bio_endio
				   ->clone_endio
				    ->dm_io_dec_pending
				     ->end_io_acct
				      ->wakeup task1
 ->dm_swap_table
  ->__bind
   ->__bind_mempools
    ->bioset_exit
     ->mempool_exit
                                     ->free_io

[ 67.330330] Unable to handle kernel NULL pointer dereference at
virtual address 0000000000000000
......
[ 67.330494] pstate: 80400085 (Nzcv daIf +PAN -UAO)
[ 67.330510] pc : mempool_free+0x70/0xa0
[ 67.330515] lr : mempool_free+0x4c/0xa0
[ 67.330520] sp : ffffff8008013b20
[ 67.330524] x29: ffffff8008013b20 x28: 0000000000000004
[ 67.330530] x27: ffffffa8c2ff40a0 x26: 00000000ffff1cc8
[ 67.330535] x25: 0000000000000000 x24: ffffffdada34c800
[ 67.330541] x23: 0000000000000000 x22: ffffffdada34c800
[ 67.330547] x21: 00000000ffff1cc8 x20: ffffffd9a1304d80
[ 67.330552] x19: ffffffdada34c970 x18: 000000b312625d9c
[ 67.330558] x17: 00000000002dcfbf x16: 00000000000006dd
[ 67.330563] x15: 000000000093b41e x14: 0000000000000010
[ 67.330569] x13: 0000000000007f7a x12: 0000000034155555
[ 67.330574] x11: 0000000000000001 x10: 0000000000000001
[ 67.330579] x9 : 0000000000000000 x8 : 0000000000000000
[ 67.330585] x7 : 0000000000000000 x6 : ffffff80148b5c1a
[ 67.330590] x5 : ffffff8008013ae0 x4 : 0000000000000001
[ 67.330596] x3 : ffffff80080139c8 x2 : ffffff801083bab8
[ 67.330601] x1 : 0000000000000000 x0 : ffffffdada34c970
[ 67.330609] Call trace:
[ 67.330616] mempool_free+0x70/0xa0
[ 67.330627] bio_put+0xf8/0x110
[ 67.330638] dec_pending+0x13c/0x230
[ 67.330644] clone_endio+0x90/0x180
[ 67.330649] bio_endio+0x198/0x1b8
[ 67.330655] dec_pending+0x190/0x230
[ 67.330660] clone_endio+0x90/0x180
[ 67.330665] bio_endio+0x198/0x1b8
[ 67.330673] blk_update_request+0x214/0x428
[ 67.330683] scsi_end_request+0x2c/0x300
[ 67.330688] scsi_io_completion+0xa0/0x710
[ 67.330695] scsi_finish_command+0xd8/0x110
[ 67.330700] scsi_softirq_done+0x114/0x148
[ 67.330708] blk_done_softirq+0x74/0xd0
[ 67.330716] __do_softirq+0x18c/0x374
[ 67.330724] irq_exit+0xb4/0xb8
[ 67.330732] __handle_domain_irq+0x84/0xc0
[ 67.330737] gic_handle_irq+0x148/0x1b0
[ 67.330744] el1_irq+0xe8/0x190
[ 67.330753] lpm_cpuidle_enter+0x4f8/0x538
[ 67.330759] cpuidle_enter_state+0x1fc/0x398
[ 67.330764] cpuidle_enter+0x18/0x20
[ 67.330772] do_idle+0x1b4/0x290
[ 67.330778] cpu_startup_entry+0x20/0x28
[ 67.330786] secondary_start_kernel+0x160/0x170

Fix this by:
1) Establishing pointers to 'struct dm_io' members in
dm_io_dec_pending() so that they may be passed into end_io_acct()
_after_ free_io() is called.
2) Moving end_io_acct() after free_io().

Cc: stable@vger.kernel.org
Signed-off-by: Jiazi Li <lijiazi@xiaomi.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Akilesh Kailash <akailash@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 drivers/md/dm.c |   17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

Message ID	20220426081741.477201436@linuxfoundation.org
State	New
Headers	show Return-Path: <stable-owner@kernel.org> From: Greg Kroah-Hartman <gregkh@linuxfoundation.org> To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>, stable@vger.kernel.org, Jiazi Li <lijiazi@xiaomi.com>, Mike Snitzer <snitzer@redhat.com>, Akilesh Kailash <akailash@google.com> Subject: [PATCH 5.10 09/86] dm: fix mempool NULL pointer race when completing IO Date: Tue, 26 Apr 2022 10:20:37 +0200 Message-Id: <20220426081741.477201436@linuxfoundation.org> In-Reply-To: <20220426081741.202366502@linuxfoundation.org> References: <20220426081741.202366502@linuxfoundation.org> User-Agent: quilt/0.66 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Precedence: bulk
Series	None \| expand [5.10,02/86] mm: page_alloc: fix building error on -Werror=array-compare [5.10,03/86] tracing: Dump stacktrace trigger to the corresponding instance [5.10,04/86] perf tools: Fix segfault accessing sample_id xyarray [5.10,05/86] gfs2: assign rgrp glock before compute_bitstructs [5.10,06/86] net/sched: cls_u32: fix netns refcount changes in u32_change() [5.10,07/86] ALSA: usb-audio: Clear MIDI port active flag after draining [5.10,08/86] ALSA: hda/realtek: Add quirk for Clevo NP70PNP [5.10,09/86] dm: fix mempool NULL pointer race when completing IO [5.10,10/86] ASoC: atmel: Remove system clock tree configuration for at91sam9g20ek [5.10,11/86] ASoC: msm8916-wcd-digital: Check failure for devm_snd_soc_register_component [5.10,12/86] ASoC: codecs: wcd934x: do not switch off SIDO Buck when codec is in use [5.10,13/86] dmaengine: imx-sdma: Fix error checking in sdma_event_remap [5.10,14/86] dmaengine: mediatek:Fix PM usage reference leak of mtk_uart_apdma_alloc_chan_resources [5.10,15/86] spi: spi-mtk-nor: initialize spi controller after resume [5.10,16/86] esp: limit skb_page_frag_refill use to a single page [5.10,17/86] igc: Fix infinite loop in release_swfw_sync [5.10,18/86] igc: Fix BUG: scheduling while atomic [5.10,19/86] rxrpc: Restore removed timer deletion [5.10,20/86] net/smc: Fix sock leak when release after smc_shutdown() [5.10,21/86] net/packet: fix packet_sock xmit return value checking [5.10,22/86] ip6_gre: Avoid updating tunnel->tun_hlen in __gre6_xmit() [5.10,23/86] ip6_gre: Fix skb_under_panic in __gre6_xmit() [5.10,24/86] net/sched: cls_u32: fix possible leak in u32_init_knode() [5.10,25/86] l3mdev: l3mdev_master_upper_ifindex_by_index_rcu should be using netdev_master_upper... [5.10,26/86] ipv6: make ip6_rt_gc_expire an atomic_t [5.10,27/86] netlink: reset network and mac headers in netlink_dump() [5.10,28/86] net: stmmac: Use readl_poll_timeout_atomic() in atomic state [5.10,29/86] dmaengine: idxd: add RO check for wq max_batch_size write [5.10,30/86] dmaengine: idxd: add RO check for wq max_transfer_size write [5.10,31/86] selftests: mlxsw: vxlan_flooding: Prevent flooding of unwanted packets [5.10,32/86] arm64/mm: Remove [PUD\|PMD]_TABLE_BIT from [pud\|pmd]_bad() [5.10,33/86] arm64: mm: fix p?d_leaf() [5.10,34/86] ARM: vexpress/spc: Avoid negative array index when !SMP [5.10,35/86] reset: tegra-bpmp: Restore Handle errors in BPMP response [5.10,36/86] platform/x86: samsung-laptop: Fix an unsigned comparison which can never be negative [5.10,37/86] ALSA: usb-audio: Fix undefined behavior due to shift overflowing the constant [5.10,38/86] arm64: dts: imx: Fix imx8*-var-som touchscreen property sizes [5.10,39/86] vxlan: fix error return code in vxlan_fdb_append [5.10,40/86] cifs: Check the IOCB_DIRECT flag, not O_DIRECT [5.10,41/86] net: atlantic: Avoid out-of-bounds indexing [5.10,42/86] mt76: Fix undefined behavior due to shift overflowing the constant [5.10,43/86] brcmfmac: sdio: Fix undefined behavior due to shift overflowing the constant [5.10,44/86] dpaa_eth: Fix missing of_node_put in dpaa_get_ts_info() [5.10,45/86] drm/msm/mdp5: check the return of kzalloc() [5.10,46/86] net: macb: Restart tx only if queue pointer is lagging [5.10,47/86] scsi: qedi: Fix failed disconnect handling [5.10,48/86] stat: fix inconsistency between struct stat and struct compat_stat [5.10,49/86] nvme: add a quirk to disable namespace identifiers [5.10,50/86] nvme-pci: disable namespace identifiers for Qemu controllers [5.10,51/86] EDAC/synopsys: Read the error count from the correct register [5.10,52/86] mm, hugetlb: allow for "high" userspace addresses [5.10,53/86] oom_kill.c: futex: delay the OOM reaper to allow time for proper futex cleanup [5.10,54/86] mm/mmu_notifier.c: fix race in mmu_interval_notifier_remove() [5.10,55/86] ata: pata_marvell: Check the bmdma_addr beforing reading [5.10,56/86] dma: at_xdmac: fix a missing check on list iterator [5.10,57/86] net: atlantic: invert deep par in pm functions, preventing null derefs [5.10,58/86] xtensa: patch_text: Fixup last cpu should be master [5.10,59/86] xtensa: fix a7 clobbering in coprocessor context load/store [5.10,60/86] openvswitch: fix OOB access in reserve_sfa_size() [5.10,61/86] gpio: Request interrupts after IRQ is initialized [5.10,62/86] ASoC: soc-dapm: fix two incorrect uses of list iterator [5.10,63/86] e1000e: Fix possible overflow in LTR decoding [5.10,64/86] ARC: entry: fix syscall_trace_exit argument [5.10,65/86] arm_pmu: Validate single/group leader events [5.10,66/86] sched/pelt: Fix attach_entity_load_avg() corner case [5.10,67/86] perf/core: Fix perf_mmap fail when CONFIG_PERF_USE_VMALLOC enabled [5.10,68/86] drm/panel/raspberrypi-touchscreen: Avoid NULL deref if not initialised [5.10,69/86] drm/panel/raspberrypi-touchscreen: Initialise the bridge in prepare [5.10,70/86] KVM: PPC: Fix TCE handling for VFIO [5.10,71/86] drm/vc4: Use pm_runtime_resume_and_get to fix pm_runtime_get_sync() usage [5.10,72/86] powerpc/perf: Fix power9 event alternatives [5.10,73/86] perf report: Set PERF_SAMPLE_DATA_SRC bit for Arm SPE event [5.10,74/86] ext4: fix fallocate to use file_modified to update permissions consistently [5.10,75/86] ext4: fix symlink file size not match to file content [5.10,76/86] ext4: fix use-after-free in ext4_search_dir [5.10,77/86] ext4: limit length to bitmap_maxbytes - blocksize in punch_hole [5.10,78/86] ext4, doc: fix incorrect h_reserved size [5.10,79/86] ext4: fix overhead calculation to account for the reserved gdt blocks [5.10,80/86] ext4: force overhead calculation if the s_overhead_cluster makes no sense [5.10,81/86] can: isotp: stop timeout monitoring when no first frame was sent [5.10,82/86] jbd2: fix a potential race while discarding reserved buffers after an abort [5.10,83/86] spi: atmel-quadspi: Fix the buswidth adjustment between spi-mem and controller [5.10,84/86] staging: ion: Prevent incorrect reference counting behavour [5.10,85/86] block/compat_ioctl: fix range check in BLKGETSIZE [5.10,86/86] Revert "net: micrel: fix KS8851_MLL Kconfig"

[5.10,09/86] dm: fix mempool NULL pointer race when completing IO

Commit Message

Patch