On Tue, Jan 24, 2012 at 7:12 PM, Martin Nyhus <martin.nyhus at gmx.com> wrote:
> On Tue, 24 Jan 2012 17:33:19 -0500 Jerome Glisse <j.glisse at gmail.com>
> wrote:
>> Can you please both test if attached patch fix it for you ?
>
> Thanks. It looks good too me, but it crashes a little later due to vma->node
> being invalid:
>
> Jan 25 00:54:21 callisto kernel: [ ?119.038357] [drm] nouveau_vm_unmap vma
> ffff880057502f50
> Jan 25 00:54:21 callisto kernel: [ ?119.038360] [drm] nouveau_vm_unmap
> vma->node ffff8800576b87a8
> Jan 25 00:54:21 callisto kernel: [ ?119.038363] [drm] nouveau_vm_unmap
> vma->node->length 58
> Jan 25 00:54:21 callisto kernel: [ ?119.038477] [drm] nouveau_vm_unmap vma
> ffff8800577beab8
> Jan 25 00:54:21 callisto kernel: [ ?119.038479] [drm] nouveau_vm_unmap
> vma->node ffff8800577bf880
> Jan 25 00:54:21 callisto kernel: [ ?119.038482] [drm] nouveau_vm_unmap
> vma->node->length 1
> Jan 25 00:54:21 callisto kernel: [ ?119.078025] [drm] nouveau_vm_unmap vma
> ffffffff8148df45
> Jan 25 00:54:21 callisto kernel: [ ?119.078029] [drm] nouveau_vm_unmap
> vma->node 8b48084b8b480000
> Jan 25 00:54:21 callisto kernel: [ ?119.078040] general protection fault:
> 0000 [#1] SMP
> Jan 25 00:54:21 callisto kernel: [ ?119.078133] CPU 0
> Jan 25 00:54:21 callisto kernel: [ ?119.078138] Modules linked in: tun
> iwl4965 iwlegacy mac80211 cfg80211 tg3 psmouse rtc_cmos evdev ehci_hcd
> uhci_hcd usbcore usb_common [last unloaded: scsi_wait_scan]
> Jan 25 00:54:21 callisto kernel: [ ?119.078542]
> Jan 25 00:54:21 callisto kernel: [ ?119.078914] Pid: 3220, comm: Xorg
> Tainted: G ? ? ? ?W ? ?3.3.0-rc1-00076-g44d4826-dirty #75 Dell Inc. XPS M1330
> /0PU073
> Jan 25 00:54:21 callisto kernel: [ ?119.079331] RIP:
> 0010:[<ffffffff814b2f7f>] ?[<ffffffff814b2f7f>] nouveau_vm_unmap+0x4f/0x80
> Jan 25 00:54:21 callisto kernel: [ ?119.079778] RSP: 0018:ffff88005c167868
> ?EFLAGS: 00010292
> Jan 25 00:54:21 callisto kernel: [ ?119.080266] RAX: 8b48084b8b480000 RBX:
> ffffffff8148df45 RCX: 0000000000000006
> Jan 25 00:54:21 callisto kernel: [ ?119.080712] RDX: 0000000000000000 RSI:
> ffffffff81868740 RDI: ffffffff81a6e040
> Jan 25 00:54:21 callisto kernel: [ ?119.081218] RBP: ffff88005c167878 R08:
> 0000000000000001 R09: 0000000000000000
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] R10: 0000000000000000 R11:
> 0000000000000001 R12: 0000000000000000
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] R13: ffff88006c309c80 R14:
> ffff88006c309a40 R15: ffff880037180590
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] FS: ?00007f141232f880(0000)
> GS:ffff88007fc00000(0000) knlGS:0000000000000000
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] CS: ?0010 DS: 0000 ES: 0000
> CR0: 0000000080050033
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] CR2: 00007fb09c1de000 CR3:
> 000000005ce28000 CR4: 00000000000006f0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] DR0: 0000000000000000 DR1:
> 0000000000000000 DR2: 0000000000000000
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] DR3: 0000000000000000 DR6:
> 00000000ffff0ff0 DR7: 0000000000000400
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] Process Xorg (pid: 3220,
> threadinfo ffff88005c166000, task ffff88005f502180)
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] Stack:
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?ffff88005f502180
> ffffffff8148df45 ffff88005c1678a8 ffffffff8148c0e8
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?ffff88006c309a40
> 0000000000000002 ffff880037180b00 ffff880079ff5e68
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?ffff88005c1678c8
> ffffffff814792b1 ffff880079ff5e68 ffff88006c309a40
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] Call Trace:
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8148df45>] ?
> nouveau_bo_move+0xb5/0x270
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8148c0e8>]
> nouveau_bo_move_ntfy+0x38/0xc0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff814792b1>]
> ttm_bo_cleanup_memtype_use+0x21/0xa0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8147a5b5>]
> ttm_bo_cleanup_refs_or_queue+0x165/0x190
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8147a675>]
> ttm_bo_release+0x95/0xd0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8147a6ef>]
> ttm_bo_unref+0x3f/0x60
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8147cae3>]
> ttm_bo_move_accel_cleanup+0x213/0x240
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8148db28>]
> nouveau_bo_move_m2mf+0x148/0x1b0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff817bfd49>] ?
> mutex_unlock+0x9/0x10
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8148df45>]
> nouveau_bo_move+0xb5/0x270
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8147ab66>]
> ttm_bo_handle_move_mem+0x1e6/0x3d0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8147bcba>]
> ttm_bo_move_buffer+0x14a/0x160
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8147bdb7>]
> ttm_bo_validate+0xe7/0xf0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8148cbdd>]
> nouveau_bo_validate+0x1d/0x20
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8148f2a0>]
> validate_list+0xc0/0x360
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8148fafa>]
> nouveau_gem_pushbuf_validate+0x9a/0x210
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8149064d>]
> nouveau_gem_ioctl_pushbuf+0x1bd/0x8d0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff810960c1>] ?
> __lock_release+0xc1/0xe0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff8145f994>]
> drm_ioctl+0x444/0x510
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff81490490>] ?
> nouveau_gem_ioctl_new+0x170/0x170
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff81152147>]
> do_vfs_ioctl+0x87/0x330
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff81344e78>] ?
> selinux_file_ioctl+0x68/0x140
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff81152481>]
> sys_ioctl+0x91/0xa0
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?[<ffffffff817cade2>]
> system_call_fastpath+0x16/0x1b
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] Code: 48 8b 53 20 48 c7 c6 40
> 87 86 81 48 c7 c7 17 3a a5 81 31 c0 e8 05 77 2f 00 48 8b 43 20 48 c7 c6 40 87
> 86 81 48 c7 c7 40 e0 a6 81 <8b> 50 38 31 c0 e8 e9 76 2f 00 48 8b 43 20 48 89
> df 31 f6 8b 50
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] RIP ?[<ffffffff814b2f7f>]
> nouveau_vm_unmap+0x4f/0x80
> Jan 25 00:54:21 callisto kernel: [ ?119.081320] ?RSP <ffff88005c167868>
> Jan 25 00:54:21 callisto kernel: [ ?119.128824] ---[ end trace
> a7919e7f17c0a727 ]---
>
> The taint is because of a failing self test (debug_objects_selftest) and the
> -dirty and extra lines at the start of the log are from this patch:
>
> diff --git a/drivers/gpu/drm/nouveau/nouveau_vm.c
> b/drivers/gpu/drm/nouveau/nouveau_vm.c
> index 2bf6c03..2b788c3 100644
> --- a/drivers/gpu/drm/nouveau/nouveau_vm.c
> +++ b/drivers/gpu/drm/nouveau/nouveau_vm.c
> @@ -150,6 +150,9 @@ nouveau_vm_unmap_at(struct nouveau_vma *vma, u64 delta,
> u64 length)
> ?void
> ?nouveau_vm_unmap(struct nouveau_vma *vma)
> ?{
> + ? ? ? DRM_INFO("%s vma %p\n", __func__, vma);
> + ? ? ? DRM_INFO("%s vma->node %p\n", __func__, vma->node);
> + ? ? ? DRM_INFO("%s vma->node->length %u\n", __func__, vma->node->length);
> ? ? ? ?nouveau_vm_unmap_at(vma, 0, (u64)vma->node->length << 12);
> ?}
>
> To reproduce I do exactly the same as before, it just takes a little longer
> before it crashes.
>
> ? ? ? ?Martin
Ben posted a proper patch on dri-devel.
Cheers,
Jerome
