[Intel-gfx] [PATCH 1/2] drm/i915/ppgtt: Pin page directories before allocation

[Matthew Auld]

On 31 January 2018 at 21:44, Chris Wilson <chris at chris-wilson.co.uk> wrote:
> Commit e2b763caa6eb ("drm/i915: Remove bitmap tracking for used-pdpes")
> believed that because it did not insert its freshly allocated page
> directory into the pd tree, it was safe from the shrinker. I failed to
> heed the lesson learnt from commit dd19674bacba ("drm/i915: Remove bitmap
> tracking for used-ptes") that we need to pin all the levels in the tree
> before hitting the shrinker or else the shrinker may free an upper layer
> as we proceed to allocate the tree. Thus leaving dangling pointers
> everywhere and a GPF should we hit direct reclaim at just the wrong
> moment.
>
> Jan 24 10:32:48 eric-macbookpro kernel: CPU: 0 PID: 7374 Comm: chromium Tainted: P           O    4.14.13-1-ARCH #1
> Jan 24 10:32:48 eric-macbookpro kernel: Hardware name: Apple Inc. MacBookPro12,1/Mac-E43C1C25D4880AD6, BIOS MBP121.88Z.0167.B33.1706181928 06/18/2017
> Jan 24 10:32:48 eric-macbookpro kernel: task: ffff994f696c2c40 task.stack: ffffb1a789d4c000
> Jan 24 10:32:48 eric-macbookpro kernel: RIP: 0010:gen8_ppgtt_set_pde.isra.40+0x48/0x70 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel: RSP: 0018:ffffb1a789d4f940 EFLAGS: 00010206
> Jan 24 10:32:48 eric-macbookpro kernel: RAX: 81c1788cc4f68138 RBX: ffff994f54db8000 RCX: ffff994f696c2c40
> Jan 24 10:32:48 eric-macbookpro kernel: RDX: 000000023bc73003 RSI: ffff994d598b6b80 RDI: ffff994f54db8000
> Jan 24 10:32:48 eric-macbookpro kernel: RBP: ffff994d598b6b80 R08: 0000000000000000 R09: 0000000000000000
> Jan 24 10:32:48 eric-macbookpro kernel: R10: ffffb1a789d4f550 R11: ffff994eaf3c3208 R12: 0000000000000027
> Jan 24 10:32:48 eric-macbookpro kernel: R13: 0000000000005000 R14: 0000000004e8f000 R15: ffff994f54dba000
> Jan 24 10:32:48 eric-macbookpro kernel: FS:  00007f585886aa00(0000) GS:ffff994faec00000(0000) knlGS:0000000000000000
> Jan 24 10:32:48 eric-macbookpro kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> Jan 24 10:32:48 eric-macbookpro kernel: CR2: 00000000004ac8e8 CR3: 00000002552c8004 CR4: 00000000003606f0
> Jan 24 10:32:48 eric-macbookpro kernel: Call Trace:
> Jan 24 10:32:48 eric-macbookpro kernel:  gen8_ppgtt_alloc_pdp+0x178/0x320 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  gen8_ppgtt_alloc_4lvl+0x5f/0x150 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  ppgtt_bind_vma+0x30/0x70 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  i915_vma_bind+0x68/0xd0 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  __i915_vma_do_pin+0x2d6/0x3a0 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  eb_lookup_vmas+0x7a2/0xb50 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  i915_gem_do_execbuffer+0x4d7/0x10e0 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  ? sock_wfree+0x34/0x60
> Jan 24 10:32:48 eric-macbookpro kernel:  ? unix_stream_read_generic+0x1f9/0x7e0
> Jan 24 10:32:48 eric-macbookpro kernel:  ? import_iovec+0x37/0xd0
> Jan 24 10:32:48 eric-macbookpro kernel:  ? i915_gem_execbuffer2+0x5d/0x390 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  i915_gem_execbuffer2+0x1b7/0x390 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  ? i915_gem_execbuffer+0x2d0/0x2d0 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  drm_ioctl_kernel+0x59/0xb0 [drm]
> Jan 24 10:32:48 eric-macbookpro kernel:  drm_ioctl+0x2d5/0x370 [drm]
> Jan 24 10:32:48 eric-macbookpro kernel:  ? i915_gem_execbuffer+0x2d0/0x2d0 [i915]
> Jan 24 10:32:48 eric-macbookpro kernel:  ? __seccomp_filter+0x3b/0x260
> Jan 24 10:32:48 eric-macbookpro kernel:  do_vfs_ioctl+0xa1/0x610
> Jan 24 10:32:48 eric-macbookpro kernel:  ? syscall_trace_enter+0xdb/0x2b0
> Jan 24 10:32:48 eric-macbookpro kernel:  SyS_ioctl+0x74/0x80
> Jan 24 10:32:48 eric-macbookpro kernel:  do_syscall_64+0x55/0x110
> Jan 24 10:32:48 eric-macbookpro kernel:  entry_SYSCALL64_slow_path+0x25/0x25
> Jan 24 10:32:48 eric-macbookpro kernel: RIP: 0033:0x7f584fa82d27
> Jan 24 10:32:48 eric-macbookpro kernel: RSP: 002b:00007ffee14a7828 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
> Jan 24 10:32:48 eric-macbookpro kernel: RAX: ffffffffffffffda RBX: 000003b0126a1030 RCX: 00007f584fa82d27
> Jan 24 10:32:48 eric-macbookpro kernel: RDX: 00007ffee14a7870 RSI: 0000000040406469 RDI: 0000000000000080
> Jan 24 10:32:48 eric-macbookpro kernel: RBP: 00007ffee14a7870 R08: 0000000000000002 R09: 0000000000000077
> Jan 24 10:32:48 eric-macbookpro kernel: R10: 00007f5839f2b780 R11: 0000000000000246 R12: 0000000040406469
> Jan 24 10:32:48 eric-macbookpro kernel: R13: 0000000000000080 R14: 00007f5842b00040 R15: 0000000000000000
> Jan 24 10:32:48 eric-macbookpro kernel: Code: 01 00 83 81 58 0a 00 00 01 48 2b 05 13 9d fd c9 48 c1 f8 06 48 c1 e0 0c 48 8d 04 d0 48 8b 56 08 48 03 05 0c 9d fd c9 48 83 ca 03 <48> 89 10 83 a9 58 0a 00 00 01 65 ff 0d 37 03 fb 3e 74 02 f3 c3
> Jan 24 10:32:48 eric-macbookpro kernel: RIP: gen8_ppgtt_set_pde.isra.40+0x48/0x70 [i915] RSP: ffffb1a789d4f940
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104773
> Fixes: e2b763caa6eb ("drm/i915: Remove bitmap tracking for used-pdpes")
> References: dd19674bacba ("drm/i915: Remove bitmap tracking for used-ptes")
> Testcase: igt/drv_selftest/live_gtt (igt_ppgtt_shrink_boom)
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Matthew Auld <matthew.auld at intel.com>
Reviewed-by: Matthew Auld <matthew.auld at intel.com>