Re: [BUG] RDMA/siw: KASAN: slab-use-after-free in siw_accept
From: Bernard Metzler
Date: Mon Aug 24 2026 - 11:32:37 EST
On 24.08.2026 10:58, Guoqing Jiang wrote:
Hi,Thanks Guoqing, that should fix it. But even better if we
On 8/22/26 03:46, Shuangpeng Bai wrote:
Hi Kernel Maintainers,
I hit the following report while testing current upstream kernel:
KASAN: slab-use-after-free in siw_accept
on commit: bd5f485f3f026225b86573e559af0b7254ef4184 (2026-08-18)
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/28addfcbf79470a62d5b4cd784e95cd2
The reproducer triggers a concurrent userspace QP transition to ERROR while
siw_accept() is unwinding after a QP initialization allocation failure. The
QP ERROR path drops qp->cep and frees the same struct siw_cep before
siw_accept() stores cep->qp = NULL.
This report is separate from the withdrawn follow-up patch:
https://lore.kernel.org/linux-rdma/20260816061353.4059242-1-shuangpeng.kernel@xxxxxxxxx/
That patch concerned a concurrent QP-destruction path. This report instead
involves a normal userspace ibv_modify_qp() transition to IBV_QPS_ERR on a
still-live QP.
We also reproduced this issue on a tree containing
a93949718259 ("RDMA/siw: Fix use-after-free in siw_accept()"). That commit
covers the qp->cep != cep case; this report exercises the distinct
qp->cep == cep race during a concurrent userspace ibv_modify_qp(...,
IBV_QPS_ERR).
The QP initialization failure was induced through the built-in fail_page_alloc
fault injector, restricted to the siw_qp_modify() call stack. The failing
vcalloc() path is a real -ENOMEM path.
The diagnostic patch shifts source line numbers in the decoded trace; in the
unmodified tree, the corresponding access is `siw_cm.c:1758` and the freeing
put is `siw_qp.c:663`.
I'm happy to test debug patches or provide additional information.
Reported-by: Shuangpeng Bai <shuangpeng.kernel@xxxxxxxxx>
[ 409.854273] ==================================================================
[ 409.856424] BUG: KASAN: slab-use-after-free in siw_accept (drivers/infiniband/sw/siw/siw_cm.c:1766)
[ 409.858419] Write of size 8 at addr ffff888176620068 by task siw_accept_foll/8801
[ 409.860281] 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
[ 409.860284] Call Trace:
[ 409.860288] <TASK>
[ 409.860291] dump_stack_lvl (lib/dump_stack.c:94 lib/dump_stack.c:120)
[ 409.860298] print_report (mm/kasan/report.c:378 mm/kasan/report.c:482)
[ 409.860315] kasan_report (mm/kasan/report.c:595)
[ 409.860325] siw_accept (drivers/infiniband/sw/siw/siw_cm.c:1766)
[ 409.860382] iw_cm_accept (drivers/infiniband/core/iwcm.c:670)
[ 409.860388] rdma_accept (drivers/infiniband/core/cma.c:4643 drivers/infiniband/core/cma.c:4727)
[ 409.860484] ucma_accept (drivers/infiniband/core/ucma.c:1208)
[ 409.860504] ucma_write (drivers/infiniband/core/ucma.c:1856)
[ 409.860528] vfs_write (fs/read_write.c:685)
[ 409.860554] ksys_write (fs/read_write.c:739)
[ 409.860574] do_syscall_64 (arch/x86/entry/syscall_64.c:61 arch/x86/entry/syscall_64.c:84)
[ 409.860579] entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)
[ 409.860583] RIP: 0033:0x7f179cdf6fef
[ 409.860588] Code: 89 54 24 18 48 89 74 24 10 89 7c 24 08 e8 29 fd ff ff 48 8b 54 24 18 48 8b 74 24 10 41 89 c0 8b 7c 24 08 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 31 44 89 c7 48 89 44 24 08 e8 5c fd ff ff 48
All code
========
0: 89 54 24 18 mov %edx,0x18(%rsp)
4: 48 89 74 24 10 mov %rsi,0x10(%rsp)
9: 89 7c 24 08 mov %edi,0x8(%rsp)
d: e8 29 fd ff ff call 0xfffffffffffffd3b
12: 48 8b 54 24 18 mov 0x18(%rsp),%rdx
17: 48 8b 74 24 10 mov 0x10(%rsp),%rsi
1c: 41 89 c0 mov %eax,%r8d
1f: 8b 7c 24 08 mov 0x8(%rsp),%edi
23: b8 01 00 00 00 mov $0x1,%eax
28: 0f 05 syscall
2a:* 48 3d 00 f0 ff ff cmp $0xfffffffffffff000,%rax <-- trapping instruction
30: 77 31 ja 0x63
32: 44 89 c7 mov %r8d,%edi
35: 48 89 44 24 08 mov %rax,0x8(%rsp)
3a: e8 5c fd ff ff call 0xfffffffffffffd9b
3f: 48 rex.W
Code starting with the faulting instruction
===========================================
0: 48 3d 00 f0 ff ff cmp $0xfffffffffffff000,%rax
6: 77 31 ja 0x39
8: 44 89 c7 mov %r8d,%edi
b: 48 89 44 24 08 mov %rax,0x8(%rsp)
10: e8 5c fd ff ff call 0xfffffffffffffd71
15: 48 rex.W
[ 409.860592] RSP: 002b:00007ffe992e7720 EFLAGS: 00000293 ORIG_RAX: 0000000000000001
[ 409.860599] RAX: ffffffffffffffda RBX: 00007ffe992e79c0 RCX: 00007f179cdf6fef
[ 409.860602] RDX: 0000000000000130 RSI: 00007ffe992e7800 RDI: 0000000000000003
[ 409.860605] RBP: 000055f0d1977070 R08: 0000000000000000 R09: 00007f179ca7e700
[ 409.860608] R10: fffffffffffffaa5 R11: 0000000000000293 R12: 0000000000000000
[ 409.860610] R13: 0000000000000000 R14: 0000000000000001 R15: 00007ffe992e7800
[ 409.860615] </TASK>
[ 409.897413] Allocated by task 8115 on cpu 1 at 408.747286s:
[ 409.897984] kasan_save_track (mm/kasan/common.c:57 mm/kasan/common.c:78)
[ 409.898409] __kasan_kmalloc (mm/kasan/common.c:398 mm/kasan/common.c:415)
[ 409.898822] __kmalloc_cache_noprof (./include/linux/kasan.h:263 mm/slub.c:5489)
[ 409.899305] siw_cep_alloc (./include/linux/slab.h:988 ./include/linux/slab.h:1309 drivers/infiniband/sw/siw/siw_cm.c:210)
[ 409.899712] siw_cm_work_handler (drivers/infiniband/sw/siw/siw_cm.c:982 drivers/infiniband/sw/siw/siw_cm.c:1077)
[ 409.900176] process_scheduled_works (kernel/workqueue.c:3322 kernel/workqueue.c:3405)
[ 409.900666] worker_thread (kernel/workqueue.c:3486)
[ 409.901080] kthread (kernel/kthread.c:436)
[ 409.901451] ret_from_fork (archkernel/process.c:158)
[ 409.901865] ret_from_fork_asm (arch/x86/entry/entry_64.S:245)
[ 409.902511] Freed by task 8817 on cpu 0 at 409.051801s:
[ 409.903051] kasan_save_track (mm/kasan/common.c:57 mm/kasan/common.c:78)
[ 409.903477] kasan_save_free_info (mm/kasan/generic.c:584)
[ 409.903930] __kasan_slab_free (mm/kasan/common.c:253 mm/kasan/common.c:285)
[ 409.904366] kfree (./include/linux/kasan.h:235 mm/slub.c:2677 mm/slub.c:6377 mm/slub.c:6692)
[ 409.904718] siw_qp_modify (drivers/infiniband/sw/siw/siw_qp.c:669 drivers/infiniband/sw/siw/siw_qp.c:823)
[ 409.905144] siw_verbs_modify_qp (drivers/infiniband/sw/siw/siw_verbs.c:600)
[ 409.905609] ib_security_modify_qp (drivers/infiniband/core/security.c:625)
[ 409.906082] _ib_modify_qp (drivers/infiniband/core/verbs.c:1912)
[ 409.906502] modify_qp (drivers/infiniband/core/uverbs_cmd.c:1925)
[ 409.906890] ib_uverbs_modify_qp (drivers/infiniband/core/uverbs_cmd.c:1951)
[ 409.907339] ib_uverbs_handler_UVERBS_METHOD_INVOKE_WRITE (drivers/infiniband/core/uverbs_std_types_device.c:44)
[ 409.907989] ib_uverbs_cmd_verbs (drivers/infiniband/core/uverbs_ioctl.c:387 drivers/infiniband/core/uverbs_ioctl.c:526)
[ 409.908469] ib_uverbs_ioctl (drivers/infiniband/core/uverbs_ioctl.c:555)
[ 409.908898] __se_sys_ioctl (fs/ioctl.c:51 fs/ioctl.c:597 fs/ioctl.c:583)
[ 409.909319] do_syscall_64 (arch/x86/entry/syscall_64.c:61 arch/x86/entry/syscall_64.c:84)
[ 409.909729] entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)
[ 409.910480] The buggy address belongs to the object at ffff888176620000
which belongs to the cache kmalloc-256 of size 256
[ 409.911716] The buggy address is located 104 bytes inside of
freed 256-byte region [ffff888176620000, ffff888176620100)
I suppose it is necessary to clear qp and cep before release state_lock since both siw_qp_llp_close and
siw_qp_modify->siw_qp_llp_close hold the lock during close cep, something like
@@ -1719,6 +1719,16 @@ int siw_accept(struct iw_cm_id *id, struct iw_cm_conn_param *params)
SIW_QP_ATTR_STATE | SIW_QP_ATTR_LLP_HANDLE |
SIW_QP_ATTR_ORD | SIW_QP_ATTR_IRD |
SIW_QP_ATTR_MPA);
+ if (rv) {
+ cep->qp = NULL;
+ qp->cep = NULL;
+ siw_cep_put(cep);
+ }
up_write(&qp->state_lock);
goto error_unlock directly and remove the now obsolete
'if (rv != 0) goto error' below. Can you pls provide
a new patch?
Shuangpeng - could you pls test this with your
test environment?
Thank you both!
Bernard.
Thanks,
Guoqing