Skip to content

Commit

Permalink
nouveau: fix instmem race condition around ptr stores
Browse files Browse the repository at this point in the history
Running a lot of VK CTS in parallel against nouveau, once every
few hours you might see something like this crash.

BUG: kernel NULL pointer dereference, address: 0000000000000008
PGD 8000000114e6e067 P4D 8000000114e6e067 PUD 109046067 PMD 0
Oops: 0000 [#1] PREEMPT SMP PTI
CPU: 7 PID: 53891 Comm: deqp-vk Not tainted 6.8.0-rc6+ #27
Hardware name: Gigabyte Technology Co., Ltd. Z390 I AORUS PRO WIFI/Z390 I AORUS PRO WIFI-CF, BIOS F8 11/05/2021
RIP: 0010:gp100_vmm_pgt_mem+0xe3/0x180 [nouveau]
Code: c7 48 01 c8 49 89 45 58 85 d2 0f 84 95 00 00 00 41 0f b7 46 12 49 8b 7e 08 89 da 42 8d 2c f8 48 8b 47 08 41 83 c7 01 48 89 ee <48> 8b 40 08 ff d0 0f 1f 00 49 8b 7e 08 48 89 d9 48 8d 75 04 48 c1
RSP: 0000:ffffac20c5857838 EFLAGS: 00010202
RAX: 0000000000000000 RBX: 00000000004d8001 RCX: 0000000000000001
RDX: 00000000004d8001 RSI: 00000000000006d8 RDI: ffffa07afe332180
RBP: 00000000000006d8 R08: ffffac20c5857ad0 R09: 0000000000ffff10
R10: 0000000000000001 R11: ffffa07af27e2de0 R12: 000000000000001c
R13: ffffac20c5857ad0 R14: ffffa07a96fe9040 R15: 000000000000001c
FS:  00007fe395eed7c0(0000) GS:ffffa07e2c980000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000008 CR3: 000000011febe001 CR4: 00000000003706f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:

...

 ? gp100_vmm_pgt_mem+0xe3/0x180 [nouveau]
 ? gp100_vmm_pgt_mem+0x37/0x180 [nouveau]
 nvkm_vmm_iter+0x351/0xa20 [nouveau]
 ? __pfx_nvkm_vmm_ref_ptes+0x10/0x10 [nouveau]
 ? __pfx_gp100_vmm_pgt_mem+0x10/0x10 [nouveau]
 ? __pfx_gp100_vmm_pgt_mem+0x10/0x10 [nouveau]
 ? __lock_acquire+0x3ed/0x2170
 ? __pfx_gp100_vmm_pgt_mem+0x10/0x10 [nouveau]
 nvkm_vmm_ptes_get_map+0xc2/0x100 [nouveau]
 ? __pfx_nvkm_vmm_ref_ptes+0x10/0x10 [nouveau]
 ? __pfx_gp100_vmm_pgt_mem+0x10/0x10 [nouveau]
 nvkm_vmm_map_locked+0x224/0x3a0 [nouveau]

Adding any sort of useful debug usually makes it go away, so I hand
wrote the function in a line, and debugged the asm.

Every so often pt->memory->ptrs is NULL. This ptrs ptr is set in
the nv50_instobj_acquire called from nvkm_kmap.

If Thread A and Thread B both get to nv50_instobj_acquire around
the same time, and Thread A hits the refcount_set line, and in
lockstep thread B succeeds at refcount_inc_not_zero, there is a
chance the ptrs value won't have been stored since refcount_set
is unordered. Force a memory barrier here, I picked smp_mb, since
we want it on all CPUs and it's write followed by a read.

v2: use paired smp_rmb/smp_wmb.

Cc: <stable@vger.kernel.org>
Fixes: be55287 ("drm/nouveau/imem/nv50: embed nvkm_instobj directly into nv04_instobj")
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Danilo Krummrich <dakr@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240411011510.2546857-1-airlied@gmail.com
  • Loading branch information
airlied authored and dakr committed Apr 15, 2024
1 parent 4c08f01 commit fff1386
Showing 1 changed file with 6 additions and 1 deletion.
7 changes: 6 additions & 1 deletion drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c
Original file line number Diff line number Diff line change
Expand Up @@ -222,8 +222,11 @@ nv50_instobj_acquire(struct nvkm_memory *memory)
void __iomem *map = NULL;

/* Already mapped? */
if (refcount_inc_not_zero(&iobj->maps))
if (refcount_inc_not_zero(&iobj->maps)) {
/* read barrier match the wmb on refcount set */
smp_rmb();
return iobj->map;
}

/* Take the lock, and re-check that another thread hasn't
* already mapped the object in the meantime.
Expand All @@ -250,6 +253,8 @@ nv50_instobj_acquire(struct nvkm_memory *memory)
iobj->base.memory.ptrs = &nv50_instobj_fast;
else
iobj->base.memory.ptrs = &nv50_instobj_slow;
/* barrier to ensure the ptrs are written before refcount is set */
smp_wmb();
refcount_set(&iobj->maps, 1);
}

Expand Down

0 comments on commit fff1386

Please sign in to comment.