[syzbot] [rdma?] kernel BUG in ib_device_get_by_index

[syzbot] [rdma?] kernel BUG in ib_device_get_by_index

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit:    779cae956c83 Add linux-next specific files for 20260223
git tree:       linux-next
console output: https://syzkaller.appspot.com/x/log.txt?x=172ebfb2580000
kernel config:  https://syzkaller.appspot.com/x/.config?x=ee920513e4deca5f
dashboard link: https://syzkaller.appspot.com/bug?extid=53cf317e7803e4ef2f33
compiler:       Debian clang version 21.1.8 (++20251221033036+2078da43e25a-1~exp1~20251221153213.50), Debian LLD 21.1.8

Unfortunately, I don't have any reproducer for this issue yet.

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/0110c968351a/disk-779cae95.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/582a1d69266e/vmlinux-779cae95.xz
kernel image: https://storage.googleapis.com/syzbot-assets/13619a146558/bzImage-779cae95.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+53cf317e7803e4ef2f33@syzkaller.appspotmail.com

------------[ cut here ]------------
kernel BUG at ./include/rdma/ib_verbs.h:4611!
Oops: invalid opcode: 0000 [#1] SMP KASAN PTI
CPU: 0 UID: 0 PID: 5945 Comm: syz.0.12 Not tainted syzkaller #0 PREEMPT(full) 
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 02/12/2026
RIP: 0010:ib_device_try_get include/rdma/ib_verbs.h:4611 [inline]
RIP: 0010:ib_device_get_by_index+0x2ae/0x2b0 drivers/infiniband/core/device.c:331
Code: f9 e9 76 fe ff ff 44 89 f1 80 e1 07 80 c1 03 38 c1 0f 8c 9d fe ff ff 4c 89 f7 e8 6d 6c 88 f9 e9 90 fe ff ff e8 23 46 1e f9 90 <0f> 0b 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 0f 1e fa
RSP: 0018:ffffc900043e6fe0 EFLAGS: 00010287
RAX: ffffffff88a79c3d RBX: ffff8880277bc000 RCX: 0000000000080000
RDX: ffffc9000d882000 RSI: 0000000000000110 RDI: 0000000000000111
RBP: ffffc900043e7090 R08: ffff88801e7ebc80 R09: 0000000000000002
R10: 0000000000000406 R11: 0000000000000002 R12: 1ffff9200087ce00
R13: dffffc0000000000 R14: ffff8880277bd104 R15: dffffc0000000000
FS:  00007f2f67eba6c0(0000) GS:ffff888125009000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000200000001000 CR3: 0000000076cc8000 CR4: 00000000003526f0
Call Trace:
 <TASK>
 nldev_set_doit+0x23a/0x4c0 drivers/infiniband/core/nldev.c:1134
 rdma_nl_rcv_msg drivers/infiniband/core/netlink.c:-1 [inline]
 rdma_nl_rcv_skb drivers/infiniband/core/netlink.c:239 [inline]
 rdma_nl_rcv+0x6d7/0xa10 drivers/infiniband/core/netlink.c:259
 netlink_unicast_kernel net/netlink/af_netlink.c:1318 [inline]
 netlink_unicast+0x80f/0x9b0 net/netlink/af_netlink.c:1344
 netlink_sendmsg+0x813/0xb40 net/netlink/af_netlink.c:1894
 sock_sendmsg_nosec+0x18f/0x1d0 net/socket.c:737
 __sock_sendmsg net/socket.c:752 [inline]
 ____sys_sendmsg+0x589/0x8c0 net/socket.c:2610
 ___sys_sendmsg+0x2a5/0x360 net/socket.c:2664
 __sys_sendmsg net/socket.c:2696 [inline]
 __do_sys_sendmsg net/socket.c:2701 [inline]
 __se_sys_sendmsg net/socket.c:2699 [inline]
 __x64_sys_sendmsg+0x1bd/0x2a0 net/socket.c:2699
 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
 do_syscall_64+0x14d/0xf80 arch/x86/entry/syscall_64.c:94
 entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x7f2f66f9c629
Code: ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 e8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007f2f67eba028 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 00007f2f67216090 RCX: 00007f2f66f9c629
RDX: 0000000000040810 RSI: 0000200000000140 RDI: 0000000000000004
RBP: 00007f2f67032b39 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 00007f2f67216128 R14: 00007f2f67216090 R15: 00007fff55890bd8
 </TASK>
Modules linked in:
---[ end trace 0000000000000000 ]---
RIP: 0010:ib_device_try_get include/rdma/ib_verbs.h:4611 [inline]
RIP: 0010:ib_device_get_by_index+0x2ae/0x2b0 drivers/infiniband/core/device.c:331
Code: f9 e9 76 fe ff ff 44 89 f1 80 e1 07 80 c1 03 38 c1 0f 8c 9d fe ff ff 4c 89 f7 e8 6d 6c 88 f9 e9 90 fe ff ff e8 23 46 1e f9 90 <0f> 0b 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 0f 1e fa
RSP: 0018:ffffc900043e6fe0 EFLAGS: 00010287
RAX: ffffffff88a79c3d RBX: ffff8880277bc000 RCX: 0000000000080000
RDX: ffffc9000d882000 RSI: 0000000000000110 RDI: 0000000000000111
RBP: ffffc900043e7090 R08: ffff88801e7ebc80 R09: 0000000000000002
R10: 0000000000000406 R11: 0000000000000002 R12: 1ffff9200087ce00
R13: dffffc0000000000 R14: ffff8880277bd104 R15: dffffc0000000000
FS:  00007f2f67eba6c0(0000) GS:ffff888125109000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f37a17e82f8 CR3: 0000000076cc8000 CR4: 00000000003526f0


---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzkaller@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

If the report is already addressed, let syzbot know by replying with:
#syz fix: exact-commit-title

If you want to overwrite report's subsystems, reply with:
#syz set subsystems: new-subsystem
(See the list of subsystem names on the web dashboard)

If the report is a duplicate of another one, reply with:
#syz dup: exact-subject-of-another-report

If you want to undo deduplication, reply with:
#syz undup

Re: [syzbot] [rdma?] kernel BUG in ib_device_get_by_index

[Tetsuo Handa]

Hmm, this assertion was wrong because ib_device_get_by_index()
might be called before enable_device_and_get() is called.

#syz invalid

Re: [syzbot] [rdma?] kernel BUG in ib_device_get_by_index

[Leon Romanovsky]

On Sat, Feb 28, 2026 at 02:07:46PM +0900, Tetsuo Handa wrote:
> Hmm, this assertion was wrong because ib_device_get_by_index()
> might be called before enable_device_and_get() is called.
> 
> #syz invalid

I think this is a valid syzkaller report. As you correctly noted, the device
was inserted into the xarray database in assign_name(), but its refcount was
only set later in enable_device_and_get().

The proper fix can be something like that:

diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index c7b227e2e657..5fc2604ec482 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -321,7 +321,7 @@ struct ib_device *ib_device_get_by_index(const struct net *net, u32 index)

        down_read(&devices_rwsem);
        device = xa_load(&devices, index);
-       if (device) {
+       if (device && xa_get_mark(&devices, index, DEVICE_REGISTERED)) {
                if (!rdma_dev_access_netns(device, net)) {
                        device = NULL;
                        goto out;

Thanks

Re: [syzbot] [rdma?] kernel BUG in ib_device_get_by_index

[Tetsuo Handa]

On 2026/03/03 4:17, Leon Romanovsky wrote:
> On Sat, Feb 28, 2026 at 02:07:46PM +0900, Tetsuo Handa wrote:
>> Hmm, this assertion was wrong because ib_device_get_by_index()
>> might be called before enable_device_and_get() is called.
>>
>> #syz invalid
> 
> I think this is a valid syzkaller report. As you correctly noted, the device
> was inserted into the xarray database in assign_name(), but its refcount was
> only set later in enable_device_and_get().

I was wondering why enable_device_and_get() is using not refcount_add()
but refcount_set(), and I tried
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit?id=510cd4b7d46753b4bf0f57004aa7b53b91b2b25a
in case commit 9af0feae8016 ("RDMA/core: Fix stale RoCE GIDs during netdev
events at registration") unexpectedly triggered modification of ->refcount
before refcount_set(&device->refcount, 2) is called.

But I concluded from this syzbot report that the reason enable_device_and_get() is
using refcount_set() is that we cannot use refcount_add() because ->refcount == 0.

Therefore, it is safe to call ib_device_try_get() before enable_device_and_get()
calls refcount_set().

> 
> The proper fix can be something like that:
> 
          down_read(&devices_rwsem);
          device = xa_load(&devices, index);
  -       if (device) {
  +       if (device && xa_get_mark(&devices, index, DEVICE_REGISTERED)) {
                  if (!rdma_dev_access_netns(device, net)) {
                          device = NULL;
                          goto out;
                  }
   
                  if (!ib_device_try_get(device))
                          device = NULL;
          }

Why do you want to make this change? Unless it is unsafe to call
rdma_dev_access_netns() when DEVICE_REGISTERED is not set,
refcount_inc_not_zero() from ib_device_try_get() makes the final
result same (i.e. device == NULL).

Since enable_device_and_get() sets ->refcount immediately before
xa_set_mark() is called, adding xa_get_mark() check does not change
effective behavior.

What I rather worry is that refcount_set() is called too early if
there is an ib_device_try_get() user who expects that
device->ops.enable_driver()/add_client_context()/add_compat_devs()
have already completed when ib_device_try_get() succeeded.

Re: [syzbot] [rdma?] kernel BUG in ib_device_get_by_index

[Leon Romanovsky]

On Tue, Mar 03, 2026 at 10:38:17PM +0900, Tetsuo Handa wrote:
> On 2026/03/03 4:17, Leon Romanovsky wrote:
> > On Sat, Feb 28, 2026 at 02:07:46PM +0900, Tetsuo Handa wrote:
> >> Hmm, this assertion was wrong because ib_device_get_by_index()
> >> might be called before enable_device_and_get() is called.
> >>
> >> #syz invalid
> > 
> > I think this is a valid syzkaller report. As you correctly noted, the device
> > was inserted into the xarray database in assign_name(), but its refcount was
> > only set later in enable_device_and_get().
> 
> I was wondering why enable_device_and_get() is using not refcount_add()
> but refcount_set(), and I tried
> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit?id=510cd4b7d46753b4bf0f57004aa7b53b91b2b25a
> in case commit 9af0feae8016 ("RDMA/core: Fix stale RoCE GIDs during netdev
> events at registration") unexpectedly triggered modification of ->refcount
> before refcount_set(&device->refcount, 2) is called.
> 
> But I concluded from this syzbot report that the reason enable_device_and_get() is
> using refcount_set() is that we cannot use refcount_add() because ->refcount == 0.
> 
> Therefore, it is safe to call ib_device_try_get() before enable_device_and_get()
> calls refcount_set().
> 
> > 
> > The proper fix can be something like that:
> > 
>           down_read(&devices_rwsem);
>           device = xa_load(&devices, index);
>   -       if (device) {
>   +       if (device && xa_get_mark(&devices, index, DEVICE_REGISTERED)) {
>                   if (!rdma_dev_access_netns(device, net)) {
>                           device = NULL;
>                           goto out;
>                   }
>    
>                   if (!ib_device_try_get(device))
>                           device = NULL;
>           }
> 
> Why do you want to make this change? Unless it is unsafe to call
> rdma_dev_access_netns() when DEVICE_REGISTERED is not set,
> refcount_inc_not_zero() from ib_device_try_get() makes the final
> result same (i.e. device == NULL).
> 
> Since enable_device_and_get() sets ->refcount immediately before
> xa_set_mark() is called, adding xa_get_mark() check does not change
> effective behavior.

xa_set_mark() is performed under down_write(&devices_rwsem) and it
ensures that xa_load(...) will return fully initialized device.

But yes, you are right, ib_device_try_get() should return 0 if this
device isn't set yet.

Thanks

> 
> What I rather worry is that refcount_set() is called too early if
> there is an ib_device_try_get() user who expects that
> device->ops.enable_driver()/add_client_context()/add_compat_devs()
> have already completed when ib_device_try_get() succeeded.
>