[Openchrome-users] Almost always frozen, with dri + glx

Mon Apr 13 14:47:21 PDT 2009

Sorry for the bunch of letters... I hope some developers read this
list, and can make anything useful from my details.

I got back to the TTM branch because I learnt I can mplayer with "-vo
gl2". If I try with "-vo gl", I get only a white content in the
window. That may be related to the problem I see in google earth,
where all places where there should be text in the GL viewport, there
are white (sometimes blue) rectangles.

I've noticed that the 'xv' mplayer-ing fails if I run my window
manager (dwm). If I don't run any window manager, all works. I notice
these differences in the Xorg logs:
Without windowmanager, so xv working:
Raw rSrc  X (0,720) Y (0,404)
Raw rDest  X (152,872) Y (182,586)
videoflag=0x4000020
// Upd_Video:
Modified rSrc  X (0,720) Y (0,404)
Modified rDest  X (152,872) Y (182,586)
===srcWidth= 720
===srcHeight= 404
===dwOffset= 0x0
===fetch= 0x5b

With windowmanager 'dwm', so xv crashing X and often also the kernel module:
Raw rDest  X (2,1026) Y (99,673)
videoflag=0x4000020
// Upd_Video:
Modified rSrc  X (0,719) Y (0,404)
Modified rDest  X (2,1024) Y (99,673)
===srcWidth= 719
===srcHeight= 404
===dwOffset= 0x0
===fetch= 0x5b
SetVideoWindow: X (2,1023) Y (99,672)
    First HQV

Backtrace:
0: /nix/store/48pi7im9w6plr68qv7nyqy67wrwxk4q5-xorg-server-1.5.3/bin/X(xorg_backtrace+0x3b)
[0x8123b9b]
1: /nix/store/48pi7im9w6plr68qv7nyqy67wrwxk4q5-xorg-server-1.5.3/bin/X(xf86SigHandler+0x51)
[0x80c04a1]
2: [0xffffe400]
3: /nix/store/sal6rzaykm3xa953k0vvakjgb6q666zx-openchrome-ttm/lib/xorg/modules/drivers//openchrome_drv.so(VIAVidUpdateOverlay+0xfdd)
[0xb7a8397d]
...

2009/4/12 Llu?s Batlle <viriketo at gmail.com>:
> I've stopped using the TTM branch by now, as mplayer-ing a movie
> through xv kills my X server. Do you think it should work?
>
> 2009/4/12 Llu?s Batlle <viriketo at gmail.com>:
>> The TTM_branch works quite well for me! I could play neverball.
>> I noticed glxgears triggers the next complain, although I have the
>> DisableIRQ option in xorg.conf:
>> ?do_wait: drmWaitVBlank returned -1, IRQs don't seem to be working correctly.
>>
>> I saw an OOPS, I think that was triggered after just launching firefox
>> (which had many web pages opened). It's a 2.6.28.9, with xorg server
>> 1.5.3. After reboot, now I'm using firefox normally.
>>
>> Here is what remain in /var/log/messages:
>> Apr 12 13:32:17 mandarina kernel: Freeing 0xf6cf9ec0
>> Apr 12 13:32:22 mandarina kernel: Freeing 0xf6cf9ec0
>> Apr 12 13:33:03 mandarina kernel: BUG: unable to handle kernel paging
>> request at 00b56c56
>> Apr 12 13:33:03 mandarina kernel: IP: [free_block+0x86/0x120]
>> free_block+0x86/0x120
>> Apr 12 13:33:03 mandarina kernel: IP: [<c01a0f96>] free_block+0x86/0x120
>> Apr 12 13:33:03 mandarina kernel: *pdpt = 0000000036c77001 *pde =
>> 0000000000000000
>> Apr 12 13:33:03 mandarina kernel: Oops: 0002 [#1] SMP
>> Apr 12 13:33:03 mandarina kernel: last sysfs file:
>> /sys/devices/pci0000:00/0000:00:18.3/resource
>> Apr 12 13:33:03 mandarina kernel: Modules linked in: tun bridge stp
>> openchrome drm ipv6 usblp raid456 async_xor async_memcpy async_tx xor
>> raid1 raid0 af_packet snd_pcm_oss snd_mixer_oss i2c_viapro i2c_core
>> rtc_cmos rtc_core rtc_lib snd_via82xx gameport snd_via82xx_modem
>> snd_ac97_codec ac97_bus snd_mpu401_uart k8temp snd_rawmidi
>> snd_seq_device serio_raw via_rhine joydev snd_pcm snd_timer hwmon
>> via_ircc snd_page_alloc mii snd soundcore ipw2200 irda sg shpchp
>> uhci_hcd crc_ccitt ieee80211 ieee80211_crypt pci_hotplug yenta_socket
>> rsrc_nonstatic video output battery ac thermal button amd64_agp
>> agpgart kqemu powernow_k8 processor cpufreq_powersave
>> cpufreq_userspace cpufreq_ondemand freq_table fuse dm_mod usbhid
>> ohci_hcd ssb pcmcia pcmcia_core ehci_hcd usbcore ext3 jbd ide_generic
>> ide_gd_mod ide_cd_mod ide_core sr_mod cdrom sd_mod crc_t10dif
>> pata_marvell ata_piix sata_uli sata_sis pata_sis sata_via sata_nv ahci
>> reiserfs pata_via libata scsi_mod
>> Apr 12 13:33:03 mandarina kernel:
>> Apr 12 13:33:03 mandarina kernel: Pid: 6, comm: events/0 Not tainted
>> (2.6.28.9-default #1)
>> Apr 12 13:33:03 mandarina kernel: EIP: 0060:[free_block+0x86/0x120]
>> EFLAGS: 00010082 CPU: 0
>> Apr 12 13:33:03 mandarina kernel: EIP: 0060:[<c01a0f96>] EFLAGS: 00010082 CPU: 0
>> Apr 12 13:33:03 mandarina kernel: EIP is at free_block+0x86/0x120
>> Apr 12 13:33:03 mandarina kernel: EAX: 00bd755a EBX: f75038c0 ECX:
>> ce849040 EDX: 00b56c52
>> Apr 12 13:33:03 mandarina kernel: ESI: ce849000 EDI: f74d3ac0 EBP:
>> 00000000 ESP: f7489f28
>> Apr 12 13:33:03 mandarina kernel: ?DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
>> Apr 12 13:33:03 mandarina kernel: Process events/0 (pid: 6,
>> ti=f7488000 task=f747c450 task.ti=f7488000)
>> Apr 12 13:33:03 mandarina kernel: Stack:
>> Apr 12 13:33:03 mandarina kernel: ?0000000c 00000000 00000004 f75062d4
>> f75038c0 f75062d4 f75062c0 f74d3ae4
>> Apr 12 13:33:03 mandarina kernel: ?00000004 c01a10a9 00000000 f75038c0
>> f74d3ac0 f75038c0 c1788f60 f7463f40
>> Apr 12 13:33:03 mandarina kernel: ?c01a20a9 00000000 00000000 c1788f64
>> f7463f40 c1788f60 c014a29a c0125805
>> Apr 12 13:33:03 mandarina kernel: Call Trace:
>> Apr 12 13:33:03 mandarina kernel: ?[drain_array+0x79/0xd0] drain_array+0x79/0xd0
>> Apr 12 13:33:03 mandarina kernel: ?[<c01a10a9>] drain_array+0x79/0xd0
>> Apr 12 13:33:03 mandarina kernel: ?[cache_reap+0x89/0x110] cache_reap+0x89/0x110
>> Apr 12 13:33:04 mandarina kernel: ?[<c01a20a9>] cache_reap+0x89/0x110
>> Apr 12 13:33:04 mandarina kernel: ?[run_workqueue+0xca/0x170]
>> run_workqueue+0xca/0x170
>> Apr 12 13:33:04 mandarina kernel: ?[<c014a29a>] run_workqueue+0xca/0x170
>> Apr 12 13:33:04 mandarina kernel: ?[default_spin_lock_flags+0x5/0x10]
>> default_spin_lock_flags+0x5/0x10
>> Apr 12 13:33:04 mandarina kernel: ?[<c0125805>] default_spin_lock_flags+0x5/0x10
>> Apr 12 13:33:04 mandarina kernel:
>> [rtc_cmos:_spin_lock_irqsave+0x33/0x50] _spin_lock_irqsave+0x33/0x50
>> Apr 12 13:33:04 mandarina kernel: ?[<c03cfa13>] _spin_lock_irqsave+0x33/0x50
>> Apr 12 13:33:04 mandarina kernel: ?[cache_reap+0x0/0x110] cache_reap+0x0/0x110
>> Apr 12 13:33:04 mandarina kernel: ?[<c01a2020>] cache_reap+0x0/0x110
>> Apr 12 13:33:04 mandarina kernel: ?[worker_thread+0x0/0xe0]
>> worker_thread+0x0/0xe0
>> Apr 12 13:33:04 mandarina kernel: ?[<c014abd0>] worker_thread+0x0/0xe0
>> Apr 12 13:33:04 mandarina kernel: ?[worker_thread+0x80/0xe0]
>> worker_thread+0x80/0xe0
>> Apr 12 13:33:04 mandarina kernel: ?[<c014ac50>] worker_thread+0x80/0xe0
>> Apr 12 13:33:04 mandarina kernel: ?[<c014d8c0>]
>> autoremove_wake_function+0x0/0x40
>> Apr 12 13:33:04 mandarina kernel: ?[worker_thread+0x0/0xe0]
>> worker_thread+0x0/0xe0
>> Apr 12 13:33:04 mandarina kernel: ?[<c014abd0>] worker_thread+0x0/0xe0
>> Apr 12 13:33:04 mandarina kernel: ?[kthread+0x42/0x70] kthread+0x42/0x70
>> Apr 12 13:33:05 mandarina kernel: ?[<c014d5a2>] kthread+0x42/0x70
>> Apr 12 13:33:05 mandarina kernel: ?[kthread+0x0/0x70] kthread+0x0/0x70
>> Apr 12 13:33:05 mandarina kernel: ?[<c014d560>] kthread+0x0/0x70
>> Apr 12 13:33:05 mandarina kernel: ?[kernel_thread_helper+0x7/0x10]
>> kernel_thread_helper+0x7/0x10
>> Apr 12 13:33:05 mandarina kernel: ?[<c010ac7f>] kernel_thread_helper+0x7/0x10
>> Apr 12 13:33:05 mandarina kernel: Code: 80 30 5e c0 8b 10 f6 c6 40 0f
>> 85 a0 00 00 00 84 d2 0f 89 a5 00 00 00 8b 70 1c 8b 5c 24 10 8b 54 24
>> 28 8b 46 04 8b 7c 93 68 8b 16 <89> 42 04 89 10 2b 4e 0c c7 06 00 01 10
>> 00 c7 46 04 00 02 20 00
>> Apr 12 13:33:05 mandarina kernel: EIP: [free_block+0x86/0x120]
>> free_block+0x86/0x120 SS:ESP 0068:f7489f28
>> Apr 12 13:33:05 mandarina kernel: EIP: [<c01a0f96>]
>> free_block+0x86/0x120 SS:ESP 0068:f7489f28
>> Apr 12 13:33:05 mandarina kernel: ---[ end trace 9bea13080343841c ]---
>> Apr 12 13:33:25 mandarina exiting on signal 15
>>
>>
>> 2009/4/12 Llu?s Batlle <viriketo at gmail.com>:
>>> Tomorrow I'll try the TTM_branch, if noone here advises not to do so.
>>> schlobinux advised that on #unichrome.
>>>
>>> 2009/4/11 Llu?s Batlle <viriketo at gmail.com>:
>>>> Here I attach the big letter I sent yesterday, which has some logs.
>>>> There I mention an X config with 'Clone On' and DisableIRQ true, but I
>>>> get exactly the same results without those lines in the Device
>>>> section.
>>>
>>
>