drm/vmwgfx: Keep a gem reference to user bos in surfaces
authorZack Rusin <zackr@vmware.com>
Thu, 28 Sep 2023 04:13:55 +0000 (00:13 -0400)
committerZack Rusin <zackr@vmware.com>
Mon, 9 Oct 2023 19:53:30 +0000 (15:53 -0400)
commit91398b413d03660fd5828f7b4abc64e884b98069
tree2f5adc456f62a8f6ee8f52cbd5ac570605f31ada
parent39465cac283702a7d4a507a558db81898029c6d3
drm/vmwgfx: Keep a gem reference to user bos in surfaces

Surfaces can be backed (i.e. stored in) memory objects (mob's) which
are created and managed by the userspace as GEM buffers. Surfaces
grab only a ttm reference which means that the gem object can
be deleted underneath us, especially in cases where prime buffer
export is used.

Make sure that all userspace surfaces which are backed by gem objects
hold a gem reference to make sure they're not deleted before vmw
surfaces are done with them, which fixes:
------------[ cut here ]------------
refcount_t: underflow; use-after-free.
WARNING: CPU: 2 PID: 2632 at lib/refcount.c:28 refcount_warn_saturate+0xfb/0x150
Modules linked in: overlay vsock_loopback vmw_vsock_virtio_transport_common vmw_vsock_vmci_transport vsock snd_ens1371 snd_ac97_codec ac97_bus snd_pcm gameport>
CPU: 2 PID: 2632 Comm: vmw_ref_count Not tainted 6.5.0-rc2-vmwgfx #1
Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 11/12/2020
RIP: 0010:refcount_warn_saturate+0xfb/0x150
Code: eb 9e 0f b6 1d 8b 5b a6 01 80 fb 01 0f 87 ba e4 80 00 83 e3 01 75 89 48 c7 c7 c0 3c f9 a3 c6 05 6f 5b a6 01 01 e8 15 81 98 ff <0f> 0b e9 6f ff ff ff 0f b>
RSP: 0018:ffffbdc34344bba0 EFLAGS: 00010286
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000027
RDX: ffff960475ea1548 RSI: 0000000000000001 RDI: ffff960475ea1540
RBP: ffffbdc34344bba8 R08: 0000000000000003 R09: 65646e75203a745f
R10: ffffffffa5b32b20 R11: 72657466612d6573 R12: ffff96037d6a6400
R13: ffff9603484805b0 R14: 000000000000000b R15: ffff9603bed06060
FS:  00007f5fd8520c40(0000) GS:ffff960475e80000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f5fda755000 CR3: 000000010d012005 CR4: 00000000003706e0
Call Trace:
 <TASK>
 ? show_regs+0x6e/0x80
 ? refcount_warn_saturate+0xfb/0x150
 ? __warn+0x91/0x150
 ? refcount_warn_saturate+0xfb/0x150
 ? report_bug+0x19d/0x1b0
 ? handle_bug+0x46/0x80
 ? exc_invalid_op+0x1d/0x80
 ? asm_exc_invalid_op+0x1f/0x30
 ? refcount_warn_saturate+0xfb/0x150
 drm_gem_object_handle_put_unlocked+0xba/0x110 [drm]
 drm_gem_object_release_handle+0x6e/0x80 [drm]
 drm_gem_handle_delete+0x6a/0xc0 [drm]
 ? __pfx_vmw_bo_unref_ioctl+0x10/0x10 [vmwgfx]
 vmw_bo_unref_ioctl+0x33/0x40 [vmwgfx]
 drm_ioctl_kernel+0xbc/0x160 [drm]
 drm_ioctl+0x2d2/0x580 [drm]
 ? __pfx_vmw_bo_unref_ioctl+0x10/0x10 [vmwgfx]
 ? do_vmi_munmap+0xee/0x180
 vmw_generic_ioctl+0xbd/0x180 [vmwgfx]
 vmw_unlocked_ioctl+0x19/0x20 [vmwgfx]
 __x64_sys_ioctl+0x99/0xd0
 do_syscall_64+0x5d/0x90
 ? syscall_exit_to_user_mode+0x2a/0x50
 ? do_syscall_64+0x6d/0x90
 ? handle_mm_fault+0x16e/0x2f0
 ? exit_to_user_mode_prepare+0x34/0x170
 ? irqentry_exit_to_user_mode+0xd/0x20
 ? irqentry_exit+0x3f/0x50
 ? exc_page_fault+0x8e/0x190
 entry_SYSCALL_64_after_hwframe+0x6e/0xd8
RIP: 0033:0x7f5fda51aaff
Code: 00 48 89 44 24 18 31 c0 48 8d 44 24 60 c7 04 24 10 00 00 00 48 89 44 24 08 48 8d 44 24 20 48 89 44 24 10 b8 10 00 00 00 0f 05 <41> 89 c0 3d 00 f0 ff ff 7>
RSP: 002b:00007ffd536a4d30 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 00007ffd536a4de0 RCX: 00007f5fda51aaff
RDX: 00007ffd536a4de0 RSI: 0000000040086442 RDI: 0000000000000003
RBP: 0000000040086442 R08: 000055fa603ada50 R09: 0000000000000000
R10: 0000000000000001 R11: 0000000000000246 R12: 00007ffd536a51b8
R13: 0000000000000003 R14: 000055fa5ebb4c80 R15: 00007f5fda90f040
 </TASK>
---[ end trace 0000000000000000 ]---

A lot of the analyis on the bug was done by Murray McAllister and
Ian Forbes.

Reported-by: Murray McAllister <murray.mcallister@gmail.com>
Cc: Ian Forbes <iforbes@vmware.com>
Signed-off-by: Zack Rusin <zackr@vmware.com>
Fixes: a950b989ea29 ("drm/vmwgfx: Do not drop the reference to the handle too soon")
Cc: <stable@vger.kernel.org> # v6.2+
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230928041355.737635-1-zack@kde.org
drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
drivers/gpu/drm/vmwgfx/vmwgfx_bo.h
drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c
drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
drivers/gpu/drm/vmwgfx/vmwgfx_gem.c
drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c
drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
drivers/gpu/drm/vmwgfx/vmwgfx_surface.c