[BUG] RDMA/rtrs-clt: race between close_work submission and synchronous close

[BUG] RDMA/rtrs-clt: race between close_work submission and synchronous close

[Shuangpeng Bai]

Hi Kernel Maintainers,

I hit the following report while testing current upstream kernel:

KASAN: slab-use-after-free in rtrs_clt_stop_and_destroy_conns

on commit: 93f51579e7df248780214094418f205253383cc5 (2026-09-20)

To help trigger the bug more reliably, we applied a minimal diagnostic patch
that only adds delays and print statements.

The reproducer and .config files are here.
https://gist.github.com/shuangpengbai/1b0782d283051b34df6548d387be90a6

I'm happy to test debug patches or provide additional information.

Reported-by: Shuangpeng Bai <shuangpeng.kernel@gmail.com>

[   86.996512] ==================================================================
[   86.996866] BUG: KASAN: slab-use-after-free in rtrs_clt_stop_and_destroy_conns (./include/linux/instrumented.h:82 ./include/linux/atomic/atomic-instrumented.h:32 drivers/infiniband/ulp/rtrs/rtrs-clt.c:2237)
[   86.996866] Read of size 4 at addr ffff8881038672fc by task kworker/0:0/9
[   86.996866]
[   86.999871] Hardware name: QEMU Ubuntu 24.04 PC v2 (i440FX + PIIX, arch_caps fix, 1996), BIOS 1.16.3-debian-1.16.3-2 04/01/2014
[   86.999871] Workqueue: rtrs_client_wq rtrs_clt_close_work
[   86.999871] Call Trace:
[   87.001599]  <TASK>
[   87.001599]  dump_stack_lvl (lib/dump_stack.c:94 lib/dump_stack.c:120)
[   87.001599]  print_report (mm/kasan/report.c:378 mm/kasan/report.c:482)
[   87.001599]  kasan_report (mm/kasan/report.c:595)
[   87.001599]  kasan_check_range (mm/kasan/generic.c:? mm/kasan/generic.c:200)
[   87.001599]  rtrs_clt_stop_and_destroy_conns (./include/linux/instrumented.h:82 ./include/linux/atomic/atomic-instrumented.h:32 drivers/infiniband/ulp/rtrs/rtrs-clt.c:2237)
[   87.001599]  rtrs_clt_close_work (drivers/infiniband/ulp/rtrs/rtrs-clt.c:2364)
[   87.001599]  process_scheduled_works (kernel/workqueue.c:3396 kernel/workqueue.c:3479)
[   87.001599]  worker_thread (kernel/workqueue.c:3560)
[   87.001599]  kthread (kernel/kthread.c:436)
[   87.001599]  ret_from_fork (arch/x86/kernel/process.c:158)
[   87.001599]  ret_from_fork_asm (arch/x86/entry/entry_64.S:245)
[   87.001599]  </TASK>
[   87.001599]
[   87.001599] Allocated by task 1 on cpu 1 at 34.840929s:
[   87.001599]  kasan_save_track (mm/kasan/common.c:57 mm/kasan/common.c:78)
[   87.001599]  __kasan_kmalloc (mm/kasan/common.c:398 mm/kasan/common.c:415)
[   87.001599]  alloc_path (./include/linux/slab.h:991 ./include/linux/slab.h:1312 drivers/infiniband/ulp/rtrs/rtrs-clt.c:1544)
[   87.001599]  rtrs_clt_open (drivers/infiniband/ulp/rtrs/rtrs-clt.c:2870)
[   87.001599]  rnbd_clt_map_device (drivers/block/rnbd/rnbd-clt.c:1268 drivers/block/rnbd/rnbd-clt.c:1551)
[   87.001599]  rnbd_clt_map_device_store (drivers/block/rnbd/rnbd-clt-sysfs.c:612)
[   87.001599]  kernfs_fop_write_iter (fs/kernfs/file.c:345)
[   87.001599]  vfs_write (fs/read_write.c:595 fs/read_write.c:687)
[   87.001599]  ksys_write (fs/read_write.c:739)
[   87.001599]  do_syscall_64 (arch/x86/entry/syscall_64.c:61 arch/x86/entry/syscall_64.c:84)
[   87.001599]  entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)
[   87.001599]
[   87.001599] Freed by task 1 on cpu 1 at 56.591404s:
[   87.001599]  kasan_save_track (mm/kasan/common.c:57 mm/kasan/common.c:78)
[   87.001599]  kasan_save_free_info (mm/kasan/generic.c:584)
[   87.001599]  __kasan_slab_free (mm/kasan/common.c:253 mm/kasan/common.c:285)
[   87.001599]  kfree (./include/linux/kasan.h:235 mm/slub.c:2748 mm/slub.c:6508 mm/slub.c:6801)
[   87.001599]  kobject_put (lib/kobject.c:689 lib/kobject.c:720 ./include/linux/kref.h:65 lib/kobject.c:737)
[   87.001599]  rtrs_clt_close (drivers/infiniband/ulp/rtrs/rtrs-clt.c:2933)
[   87.001599]  free_sess (drivers/block/rnbd/rnbd-clt.c:754 drivers/block/rnbd/rnbd-clt.c:766)
[   87.001599]  rnbd_clt_put_dev (drivers/block/rnbd/rnbd-clt.c:48 drivers/block/rnbd/rnbd-clt.c:61)
[   87.001599]  rnbd_clt_unmap_device (drivers/block/rnbd/rnbd-clt.c:1703)
[   87.001599]  rnbd_clt_unmap_dev_store (drivers/block/rnbd/rnbd-clt-sysfs.c:334)
[   87.001599]  kernfs_fop_write_iter (fs/kernfs/file.c:345)
[   87.001599]  vfs_write (fs/read_write.c:595 fs/read_write.c:687)
[   87.001599]  ksys_write (fs/read_write.c:739)
[   87.001599]
[   87.001599] Last potentially related work creation:
[   87.001599]  kasan_save_stack (mm/kasan/common.c:57)
[   87.001599]  kasan_record_aux_stack (mm/kasan/generic.c:556)
[   87.001599]  insert_work (kernel/workqueue.c:2272)
[   87.001599]  __queue_work (kernel/workqueue.c:2448)
[   87.001599]  queue_work_on (kernel/workqueue.c:2501)
[   87.001599]  rtrs_clt_close_conns (./include/linux/workqueue.h:700 drivers/infiniband/ulp/rtrs/rtrs-clt.c:1967)
[   87.001599]  rtrs_clt_rdma_cm_handler (drivers/infiniband/ulp/rtrs/rtrs-clt.c:2063)
[   87.001599]  cma_cm_event_handler (drivers/infiniband/core/cma.c:2181)
[   87.001599]  cma_process_remove (drivers/infiniband/core/cma.c:5313 drivers/infiniband/core/cma.c:5346)
[   87.001599]  cma_remove_one (drivers/infiniband/core/cma.c:5453)
[   87.001599]  remove_client_context (drivers/infiniband/core/device.c:810)
[   87.001599]  disable_device (drivers/infiniband/core/device.c:1325)
[   87.001599]  __ib_unregister_device (drivers/infiniband/core/device.c:1571)
[   87.001599]  ib_unregister_device_and_put (drivers/infiniband/core/device.c:1636)
[   87.001599]  nldev_dellink (drivers/infiniband/core/nldev.c:1902)
[   87.001599]  rdma_nl_rcv (drivers/infiniband/core/netlink.c:? drivers/infiniband/core/netlink.c:239 drivers/infiniband/core/netlink.c:259)
[   87.001599]  netlink_unicast (net/netlink/af_netlink.c:1338 net/netlink/af_netlink.c:1364)
[   87.001599]  netlink_sendmsg (net/netlink/af_netlink.c:1919)
[   87.001599]  __sock_sendmsg (net/socket.c:800 net/socket.c:815)
[   87.001599]  __sys_sendto (net/socket.c:2281)
[   87.001599]  __x64_sys_sendto (net/socket.c:2288 net/socket.c:2284 net/socket.c:2284)
[   87.001599]  do_syscall_64 (arch/x86/entry/syscall_64.c:61 arch/x86/entry/syscall_64.c:84)
[   87.001599]  entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)
[   87.001599]
[   87.001599] The buggy address belongs to the object at ffff888103867000
[   87.001599]  which belongs to the cache kmalloc-2k of size 2048
[   87.001599] The buggy address is located 764 bytes inside of
[   87.001599]  freed 2048-byte region [ffff888103867000, ffff888103867800)
[   87.001599] ==================================================================


Best,
Shuangpeng