Re: 2.6.18-rc4-mm1 BUG, drm relatedy

[Frederik Deweerdt <deweerdt@free.fr>]

On Mon, Aug 21, 2006 at 12:24:51PM +0100, Dave Airlie wrote:
> 
> >>[   40.276000] [drm:drm_unlock] *ERROR* Process 8914 using kernel context 0
> 
> Can you send me an lspci -v (or have you already??)
The lspci output is available at http://fdeweerdt.free.fr/drm_bug,
along with the .config, and dmesg.
> 
> I've been busy lately so having trouble following this stuff in a timely manner, I think this is an error path which the 
> userpsace code doesn't clean up properly, your patch is most definitely not correct..
I see. So this is most likely X's having trouble with the EFAULT on
unlock? I've gone through the 2.6.18-rc4-mm1 patches, and I couldn't
relate the updates and that new drm_unlock() error message. Any idea?
> 
> if I had to guess I'd say you have an AGP machine + card but no AGP support driver loaded...
I've got CONFIG_AGP=y and CONFIG_AGP_INTEL=y, do I need some more
options? What's strange is that this worked perfectly with
2.6.18-rc3-mm2...

Thanks,
Frederik

> 
> Dave.
> 
> 
> >>[   41.024000] BUG: unable to handle kernel paging request at virtual address 6e756f73
> >>[   41.024000]  printing eip:
> >>[   41.024000] c01b5771
> >>[   41.024000] *pde = 00000000
> >>[   41.024000] Oops: 0000 [#1]
> >>[   41.024000] 8K_STACKS PREEMPT
> >>[   41.024000] last sysfs file: /devices/pci0000:00/0000:00:1d.7/usb5/5-0:1.0/bInterfaceProtocol
> >>[   41.024000] Modules linked in: snd_seq snd_seq_device ohci_hcd parport_pc parport pcspkr ipw2200 yenta_socket 
> >>rsrc_nonstatic pcmcia_core snd_intel8x0 snd_ac97_codec snd_ac97_bus snd_pcm snd_timer snd soundcore snd_page_alloc ehci_hcd 
> >>uhci_hcd usbcore cpufreq_stats cpufreq_powersave cpufreq_ondemand cpufreq_conservative speedstep_centrino freq_table 
> >>processor ac battery i915 drm tg3 joydev tsdev
> >>[   41.024000] CPU:    0
> >>[   41.024000] EIP:    0060:[<c01b5771>]    Not tainted VLI
> >>[   41.024000] EFLAGS: 00210246   (2.6.18-rc4-mm1-def01 #1)
> >>[   41.024000] EIP is at sysfs_lookup+0x65/0xb0
> >>[   41.024000] eax: f3161e40   ebx: f316842c   ecx: f73f6280   edx: f316842c
> >>[   41.024000] esi: 6e756f73   edi: f3161ec4   ebp: f6c35dfc   esp: f6c35de0
> >>[   41.024000] ds: 007b   es: 007b   ss: 0068
> >>[   41.024000] Process modprobe (pid: 8952, ti=f6c34000 task=f7d17550 task.ti=f6c34000)
> >>[   41.024000] Stack: f316842c f3161e94 f31684bc 00000000 fffffff4 f73f72ec f3161e40 f6c35e1c
> >>[   41.024000]        c0184d42 f73f72ec f3161e40 00000000 ffffffff c03a6656 12fd28db f6c35e4c
> >>[   41.024000]        c0184e02 f6c35e30 f3161ee8 00000000 12fd28db 00000005 c03a6651 c038230e
> >>[   41.024000] Call Trace:
> >>[   41.024000]  [<c0184d42>] __lookup_hash+0x9d/0xcc
> >>[   41.024000]  [<c0184e02>] lookup_one_len+0x71/0x86
> >>[   41.024000]  [<c01b51da>] create_dir+0x43/0x23f
> >>[   41.024000]  [<c01b53fc>] sysfs_create_subdir+0x26/0x28
> >>[   41.024000]  [<c01b6c56>] sysfs_create_group+0x77/0x97
> >>[   41.024000]  [<c02903af>] dpm_sysfs_add+0x1e/0x20
> >>[   41.024000]  [<c028f6b3>] device_pm_add+0x64/0x89
> >>[   41.024000]  [<c028930a>] device_add+0x1d9/0x380
> >>[   41.024000]  [<c02894cb>] device_register+0x1a/0x20
> >>[   41.024000]  [<c0289811>] device_create+0xaa/0xc4
> >>[   41.024000]  [<f8a0c472>] snd_register_device+0xcf/0x104 [snd]
> >>[   41.024000]  [<f8abd0c2>] snd_sequencer_device_init+0x4e/0x7c [snd_seq]
> >>[   41.024000]  [<f8abd02f>] alsa_seq_init+0x2f/0x51 [snd_seq]
> >>[   41.024000]  [<c014186c>] sys_init_module+0x163/0x221
> >>[   41.024000]  [<c0103135>] sysenter_past_esp+0x56/0x8d
> >>[   41.024000]  [<b7fb0410>] 0xb7fb0410
> >>[   41.024000]  [<c0104017>] show_trace_log_lvl+0x2f/0x45
> >>[   41.024000]  [<c01040ee>] show_stack_log_lvl+0x98/0xb2
> >>[   41.024000]  [<c0104351>] show_registers+0x1eb/0x289
> >>[   41.024000]  [<c0104587>] die+0x134/0x241
> >>[   41.024000]  [<c0385e80>] do_page_fault+0x395/0x620
> >>[   41.024000]  [<c0384401>] error_code+0x39/0x40
> >>[   41.024000]  [<c0184d42>] __lookup_hash+0x9d/0xcc
> >>[   41.024000]  [<c0184e02>] lookup_one_len+0x71/0x86
> >>[   41.024000]  [<c01b51da>] create_dir+0x43/0x23f
> >>[   41.024000]  [<c01b53fc>] sysfs_create_subdir+0x26/0x28
> >>[   41.024000]  [<c01b6c56>] sysfs_create_group+0x77/0x97
> >>[   41.024000]  [<c02903af>] dpm_sysfs_add+0x1e/0x20
> >>[   41.024000]  [<c028f6b3>] device_pm_add+0x64/0x89
> >>[   41.024000]  [<c028930a>] device_add+0x1d9/0x380
> >>[   41.024000]  [<c02894cb>] device_register+0x1a/0x20
> >>[   41.024000]  [<c0289811>] device_create+0xaa/0xc4
> >>[   41.024000]  [<f8a0c472>] snd_register_device+0xcf/0x104 [snd]
> >>[   41.024000]  [<f8abd0c2>] snd_sequencer_device_init+0x4e/0x7c [snd_seq]
> >>[   41.024000]  [<f8abd02f>] alsa_seq_init+0x2f/0x51 [snd_seq]
> >>[   41.024000]  [<c014186c>] sys_init_module+0x163/0x221
> >>[   41.024000]  [<c0103135>] sysenter_past_esp+0x56/0x8d
> >>[   41.024000]  =======================
> >>[   41.024000] Code: 42 fc 89 c3 8b 40 04 0f 18 00 90 3b 55 ec 75 e6 8b 45 f0 83 c4 10 5b 5e 5f 5d c3 89 1c 24 e8 27 e9 ff ff 
> >>89 c6 8b 45 0c 8b 78 48 <ac> ae 75 08 84 c0 75 f8 31 c0 eb 04 19 c0 0c 01 85 c0 75 bd f6
> >>[   41.024000] EIP: [<c01b5771>] sysfs_lookup+0x65/0xb0 SS:ESP 0068:f6c35de0
> >>[   41.024000]
> >>
> >>
> >Hi Andrew,
> >
> >I think that this definitely qualifies as a drm bug. Here are the tests
> >that I did and that make me think so:
> >- Removing the line 'Load  "dri"' from xorg.conf makes the oops
> > disappear.
> >- The bisection I performed indicates that the problem comes from the
> > following set of patches:
> > git-drm.patch
> > drm-build-fix.patch
> > drm-build-fixes-2.patch
> > allow-drm-detection-of-new-via-chipsets.patch
> > git-drm-build-fix.patch
> >- Appliying the attached patch makes the oops disapear too.
> >
> >Now, I don't know the drm code, and I'm not able to explain why the fix
> >works... Any ideas?
> >
> >Regards,
> >Frederik
> >
> >Signed-off-by: Frederik Deweerdt <frederik.deweerdt@gmail.com>
> >diff --git a/drivers/char/drm/drm_lock.c b/drivers/char/drm/drm_lock.c
> >index f9e4530..a9e01b7 100644
> >--- a/drivers/char/drm/drm_lock.c
> >+++ b/drivers/char/drm/drm_lock.c
> >@@ -65,12 +65,6 @@ int drm_lock(struct inode *inode, struct
> >	if (copy_from_user(&lock, (drm_lock_t __user *) arg, sizeof(lock)))
> >		return -EFAULT;
> >
> >-	if (lock.context == DRM_KERNEL_CONTEXT) {
> >-		DRM_ERROR("Process %d using kernel context %d\n",
> >-			  current->pid, lock.context);
> >-		return -EINVAL;
> >-	}
> >-
> >	DRM_DEBUG("%d (pid %d) requests lock (0x%08x), flags = 0x%08x\n",
> >		  lock.context, current->pid,
> >		  dev->lock.hw_lock->lock, lock.flags);
> >
> 
> -- 
> David Airlie, Software Engineer
> http://www.skynet.ie/~airlied / airlied at skynet.ie
> Linux kernel - DRI, VAX / pam_smb / ILUG
> 
>