[Intel-gfx] QM57/i7: i915 DRI kernel driver causes crash

[Chris Wilson]

On Wed, 19 Jan 2011 14:51:07 -0800, "John J. Stimson-III" <john at idsfa.net> wrote:
> The kernel messages from running "modprobe i915" did get logged:
> 
> Jan 19 12:05:08 harlie kernel: mtrr: type mismatch for c0000000,10000000 old: write-back new: write-combining
> Jan 19 12:05:08 harlie kernel: BUG: unable to handle kernel NULL pointer dereference at   (null)
> Jan 19 12:05:08 harlie kernel: IP: [<  (null)>]   (null)
> Jan 19 12:05:08 harlie kernel: *pdpt = 000000002f183001 *pde = 0000000000000000
> Jan 19 12:05:08 harlie kernel: Oops: 0010 [#1] SMP
> Jan 19 12:05:08 harlie kernel: last sysfs file: /sys/devices/pci0000:00/0000:00:02.0/boot_vga
> Jan 19 12:05:08 harlie kernel: Modules linked in: i915(+) drm_kms_helper cfbcopyarea video backlight cfbimgblt cfbfillrect
> Jan 19 12:05:08 harlie kernel:
> Jan 19 12:05:08 harlie kernel: Pid: 4166, comm: modprobe Not tainted 2.6.37Harlie #8 To be filled by O.E.M./To be filled by O.E.M.
> Jan 19 12:05:08 harlie kernel: EIP: 0060:[<00000000>] EFLAGS: 00010246 CPU: 3
> Jan 19 12:05:08 harlie kernel: EIP is at 0x0
> Jan 19 12:05:08 harlie kernel: EAX: 00000000 EBX: f77ed000 ECX: f3383100 EDX: f77ed000
> Jan 19 12:05:08 harlie kernel: ESI: f2c8a118 EDI: 00000000 EBP: f2c93c98 ESP: f2c93c7c
> Jan 19 12:05:08 harlie kernel:  DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
> Jan 19 12:05:08 harlie kernel: Process modprobe (pid: 4166, ti=f2c92000 task=f2eb72a0 task.ti=f2c92000)
> Jan 19 12:05:08 harlie kernel: Stack:
> Jan 19 12:05:08 harlie kernel:  c1071821 000000d0 ecf67110 00000000 000a13d2 f26cb600 00000001 f2c93cac
> Jan 19 12:05:08 harlie kernel:  c107192f 00000000 000a13d2 00000000 f2c93cd4 f86f40ed f2d3ddc0 00001000
> Jan 19 12:05:08 harlie kernel:  00000000 000a13d2 f2c8a118 f26cb600 ed700000 00001000 f2c93d04 f86f57ab
> Jan 19 12:05:08 harlie kernel: Call Trace:
> Jan 19 12:05:08 harlie kernel:  [<c1071821>] ? do_read_cache_page+0x71/0x160
> Jan 19 12:05:08 harlie kernel:  [<c107192f>] ? read_cache_page_gfp+0x1f/0x30
> Jan 19 12:05:08 harlie kernel:  [<f86f40ed>] ? i915_gem_object_get_pages+0xad/0x1d0 [i915]
> Jan 19 12:05:08 harlie kernel:  [<f86f57ab>] ? i915_gem_object_bind_to_gtt+0xeb/0x2d0 [i915]
> Jan 19 12:05:08 harlie kernel:  [<f86f7961>] ? i915_gem_object_pin+0x151/0x190 [i915]
> Jan 19 12:05:08 harlie kernel:  [<c11dedcd>] ? drm_gem_object_init+0x3d/0x60
> Jan 19 12:05:08 harlie kernel:  [<f86f7aa5>] ? i915_gem_init_ringbuffer+0x105/0x1e0 [i915]

I've never seen that one before. Looks suspiciously like that the backing
shmem file was not initialised. It would be useful to find out just which
line do_read_cache_page+0x71 corresponds to, i.e.:

  gdb vmlinux
  list *do_read_cache_page+0x71

and double check that CONFIG_SHMEM is indeed builtin.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre