Random computer lockups

Message

Post by **NeddySeagoon** » Wed Oct 06, 2010 7:01 pm

panderiz,

Do a few things you can do for free ...

Run memtest from a CD, instead of the kernel.
emerge smartmontools and look at the disk drives internal error log.

Lastly remove the RAM and replace each stick in same socket it was removed from. This wipes the contacts, which breaks any oxide coating that may have formed.
There shouldn't be any, with gold on both surfaces, but its harmless.

If its still faulty remove all but one stick of RAM. If that seems OK, add another ... and so on.

If you have spare RAM slots, move each stick round one slot, then again if it fails ... and so on.

If you have any unused plug in cards, remove them, that helps the PSU. Likewise any connected but unused disks.

It will take a while to work your way through the above.

panderiz · Post by **panderiz** » Fri Oct 08, 2010 3:52 am

NeddySeagoon wrote:panderiz,

Do a few things you can do for free ...

Run memtest from a CD, instead of the kernel.
emerge smartmontools and look at the disk drives internal error log.

Lastly remove the RAM and replace each stick in same socket it was removed from. This wipes the contacts, which breaks any oxide coating that may have formed.
There shouldn't be any, with gold on both surfaces, but its harmless.

If its still faulty remove all but one stick of RAM. If that seems OK, add another ... and so on.

If you have spare RAM slots, move each stick round one slot, then again if it fails ... and so on.

If you have any unused plug in cards, remove them, that helps the PSU. Likewise any connected but unused disks.

It will take a while to work your way through the above.

CD?! What is this CD you speak of?! =P I'll run a memtest some time tomorrow when I got time from my live usb (should be the same), I'm installing smartmontools right now and will give that a look in a short while. I'll mess with the ram tomorrow if I got time to boot into a liveusb and test stuff out as well. Will post back when I do your above list with results as well as follow up any future posts that're made here.

Post by **NeddySeagoon** » Fri Oct 08, 2010 9:57 pm

panderiz,

As long as you do not run memtest inside Linux as that is not very useful, since it can't beat the virtual memory system.
It must be run in place of kernel so it can have the entire system to itself.

panderiz · Post by **panderiz** » Fri Oct 08, 2010 11:22 pm

NeddySeagoon wrote:panderiz,

As long as you do not run memtest inside Linux as that is not very useful, since it can't beat the virtual memory system.
It must be run in place of kernel so it can have the entire system to itself.

I actually meant I'll be using my live USB of some linux distribution to run memtest, should be the same. Will run test in about an hour.

panderiz · Post by **panderiz** » Tue Nov 09, 2010 3:54 pm

Double posting cause I got something that might actually be of use now...

Code: Select all

Nov  9 11:20:53 dm400 kernel: [   45.448782] [fglrx] GART Table is not in FRAME_BUFFER range
Nov  9 11:20:53 dm400 kernel: [   45.448937] [fglrx] Could not enable MSI; System prevented initialization
Nov  9 11:20:53 dm400 kernel: [   45.449397] [fglrx] Firegl kernel thread PID: 5159
Nov  9 11:20:53 dm400 kernel: [   45.449673] [fglrx] IRQ 18 Enabled
Nov  9 11:20:54 dm400 kernel: [   46.022903] [fglrx] Gart USWC size:1160 M.
Nov  9 11:20:54 dm400 kernel: [   46.022906] [fglrx] Gart cacheable size:459 M.
Nov  9 11:20:54 dm400 kernel: [   46.022911] [fglrx] Reserved FB block: Shared offset:0, size:1000000
Nov  9 11:20:54 dm400 kernel: [   46.022914] [fglrx] Reserved FB block: Unshared offset:17ffb000, size:5000
Nov  9 11:21:25 dm400 polkitd[5376]: started daemon version 0.96 using authority implementation `local' version `0.96'
Nov  9 11:26:54 dm400 sshd[5427]: Accepted keyboard-interactive/pam for root from 192.168.2.15 port 1128 ssh2
Nov  9 11:26:54 dm400 sshd[5427]: pam_unix(sshd:session): session opened for user root by (uid=0)
Nov  9 11:36:01 dm400 kernel: [  953.197233] [fglrx] ASIC hang happened
Nov  9 11:36:01 dm400 kernel: [  953.197237] Pid: 5414, comm: ExeFile.exe Tainted: P            2.6.35-gentoo-r5 #6
Nov  9 11:36:01 dm400 kernel: [  953.197240] Call Trace:
Nov  9 11:36:01 dm400 kernel: [  953.197289]  [<ffffffffa0005edb>] KCL_DEBUG_OsDump+0x9/0xb [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197327]  [<ffffffffa001300c>] firegl_hardwareHangRecovery+0x1c/0x50 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197377]  [<ffffffffa008b999>] ? _ZN4Asic9WaitUntil15ResetASICIfHungEv+0x9/0x10 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197426]  [<ffffffffa008b94c>] ? _ZN4Asic9WaitUntil15WaitForCompleteEv+0x6c/0xb0 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197473]  [<ffffffffa008aad0>] ? _ZN4Asic19PM4ElapsedTimeStampEj14_LARGE_INTEGER+0x90/0x130 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197518]  [<ffffffffa0070000>] ? _ZN10QS_PRIVATED0Ev+0x20/0x90 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197565]  [<ffffffffa008375e>] ? _ZN15QS_PRIVATE_CORE27multiVpuPM4ElapsedTimeStampEj14_LARGE_INTEGER+0x2e/0x50 [fglrx]
Nov  9 11:36:01 dm400 sshd[5427]: pam_unix(sshd:session): session closed for user root
Nov  9 11:36:01 dm400 kernel: [  953.197611]  [<ffffffffa007dcaf>] ? _Z19uQSTimeStampRetiredmjj14_LARGE_INTEGER+0x6f/0x80 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197658]  [<ffffffffa0079f2e>] ? _Z8uCWDDEQCmjjPvjS_+0x36e/0xf00 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197704]  [<ffffffffa007ab73>] ? _Z13uCWDDEQCThunkmjjPvjS_+0xb3/0x14c0 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197709]  [<ffffffff81033d7b>] ? select_task_rq_fair+0x778/0x780
Nov  9 11:36:01 dm400 kernel: [  953.197712]  [<ffffffff810330dc>] ? enqueue_entity+0x27b/0x288
Nov  9 11:36:01 dm400 kernel: [  953.197715]  [<ffffffff8101d91b>] ? flat_send_IPI_mask+0xc/0xe
Nov  9 11:36:01 dm400 kernel: [  953.197718]  [<ffffffff81019918>] ? native_smp_send_reschedule+0x51/0x53
Nov  9 11:36:01 dm400 kernel: [  953.197722]  [<ffffffff8102dbe8>] ? resched_task+0x60/0x62
Nov  9 11:36:01 dm400 kernel: [  953.197725]  [<ffffffff8102dda7>] ? check_preempt_curr_idle+0x10/0x12
Nov  9 11:36:01 dm400 kernel: [  953.197727]  [<ffffffff8102db47>] ? check_preempt_curr+0x1b/0x33
Nov  9 11:36:01 dm400 kernel: [  953.197730]  [<ffffffff810381a2>] ? try_to_wake_up+0x2a1/0x2b3
Nov  9 11:36:01 dm400 kernel: [  953.197768]  [<ffffffffa002e1c2>] ? firegl_trace+0x72/0x1e0 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197805]  [<ffffffffa002e1c2>] ? firegl_trace+0x72/0x1e0 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197808]  [<ffffffff810381c1>] ? default_wake_function+0xd/0xf
Nov  9 11:36:01 dm400 kernel: [  953.197845]  [<ffffffffa002e1c2>] ? firegl_trace+0x72/0x1e0 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197847]  [<ffffffff810381b4>] ? default_wake_function+0x0/0xf
Nov  9 11:36:01 dm400 kernel: [  953.197884]  [<ffffffffa002e1c2>] ? firegl_trace+0x72/0x1e0 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197918]  [<ffffffffa0001718>] ? KCL_MEM_SmallBufferAlloc+0xe/0x10 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197922]  [<ffffffff810d4b2b>] ? __kmalloc+0x11b/0x12d
Nov  9 11:36:01 dm400 kernel: [  953.197955]  [<ffffffffa0001718>] ? KCL_MEM_SmallBufferAlloc+0xe/0x10 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.197992]  [<ffffffffa003067e>] ? firegl_cmmqs_CWDDE_32+0x35e/0x440 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.198030]  [<ffffffffa002f0e0>] ? firegl_cmmqs_CWDDE32+0x70/0x100 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.198067]  [<ffffffffa002f070>] ? firegl_cmmqs_CWDDE32+0x0/0x100 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.198102]  [<ffffffffa000ec9a>] ? firegl_ioctl+0x1ea/0xeb0 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.198125]  [<ffffffffa003620e>] ? firegl_cmmqs_CWDDE32_32on64+0xde/0x180 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.198183]  [<ffffffffa0036402>] ? firegl_compat_ioctl+0x72/0xb0 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.198217]  [<ffffffffa00051ac>] ? ip_firegl_compat_ioctl+0x45/0x77 [fglrx]
Nov  9 11:36:01 dm400 kernel: [  953.198220]  [<ffffffff8111031d>] ? compat_sys_ioctl+0x215/0x151c
Nov  9 11:36:01 dm400 kernel: [  953.198224]  [<ffffffff8104bb76>] ? sigprocmask+0xa1/0xc3
Nov  9 11:36:01 dm400 kernel: [  953.198227]  [<ffffffff8105d016>] ? getnstimeofday+0x5f/0xbe
Nov  9 11:36:01 dm400 kernel: [  953.198230]  [<ffffffff8105d0cc>] ? do_gettimeofday+0x15/0x35
Nov  9 11:36:01 dm400 kernel: [  953.198233]  [<ffffffff8102bf68>] ? cstar_dispatch+0x7/0x27
Nov  9 11:36:01 dm400 kernel: [  953.198237] pubdev:0xffffffffa020ccb0, num of device:1 , name:fglrx, major 8, minor 77.
Nov  9 11:36:01 dm400 kernel: [  953.198240] device 0 : 0xffff88011bf10000 .
Nov  9 11:36:01 dm400 kernel: [  953.198243] Asic ID:0x9614, revision:0x13, MMIOReg:0xffffc90011120000.
Nov  9 11:36:01 dm400 kernel: [  953.198245] FB phys addr: 0xd0000000, MC :0xc0000000, Total FB size :0x18000000.
Nov  9 11:36:01 dm400 kernel: [  953.198248] gart table MC:0x0, Physical:0x0, size:0x0.
Nov  9 11:36:01 dm400 kernel: [  953.198250] mc_node :FB, total 1 zones
Nov  9 11:36:01 dm400 kernel: [  953.198253]     MC start:0xc0000000, Physical:0xd0000000, size:0x10000000.
Nov  9 11:36:01 dm400 kernel: [  953.198256]     Mapped heap -- Offset:0x0, size:0x10000000, reference count:8, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198259]     Mapped heap -- Offset:0x0, size:0x1000000, reference count:1, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198261] mc_node :INV_FB, total 1 zones
Nov  9 11:36:01 dm400 kernel: [  953.198263]     MC start:0xd0000000, Physical:0x128000000, size:0x8000000.
Nov  9 11:36:01 dm400 kernel: [  953.198266]     Mapped heap -- Offset:0x0, size:0x7ffb000, reference count:2, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198269]     Mapped heap -- Offset:0x7ffb000, size:0x5000, reference count:1, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198272] mc_node :GART_USWC, total 2 zones
Nov  9 11:36:01 dm400 kernel: [  953.198274]     MC start:0x3b8d0000, Physical:0x0, size:0x48800000.
Nov  9 11:36:01 dm400 kernel: [  953.198277]     Mapped heap -- Offset:0x3830000, size:0x800000, reference count:1, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198280]     Mapped heap -- Offset:0x3030000, size:0x800000, reference count:1, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198283]     Mapped heap -- Offset:0x2830000, size:0x800000, reference count:6, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198286]     Mapped heap -- Offset:0x2030000, size:0x800000, reference count:6, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198289]     Mapped heap -- Offset:0x30000, size:0x2000000, reference count:31, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198292] mc_node :GART_CACHEABLE, total 3 zones
Nov  9 11:36:01 dm400 kernel: [  953.198294]     MC start:0x10400000, Physical:0x0, size:0x2b4d0000.
Nov  9 11:36:01 dm400 kernel: [  953.198297]     Mapped heap -- Offset:0x0, size:0x200000, reference count:3, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198300]     Mapped heap -- Offset:0xef000, size:0x11000, reference count:1, mapping count:0,
Nov  9 11:36:01 dm400 kernel: [  953.198302] Dump the trace queue.
Nov  9 11:36:01 dm400 kernel: [  953.198303] End of dump

This has happened several times while I play the game EVE online, as well as a few times while playing minecraft. Don't play many other games but I'm fairly certain though this would happen with most other games as well...

Any insight about this regarding what I could do? And yes, I am considering now buying an nvidea graphics card.

Post by **NeddySeagoon** » Tue Nov 09, 2010 6:41 pm

panderiz,

Code: Select all

Nov  9 11:36:01 dm400 kernel: [  953.197233] [fglrx] ASIC hang happened

looks like a debug message from the ATI binary blob.
Try upgrading the firmware on your graphics card and/or a newer version of the binary blob.
Its clear that ATI know about the issue, or that debug message wouldn't be there.

After this incident, can you still ssh into the system ?
Its unusual but not unknown, for graphics issues to kill the system stone dead.

An ASIC, (Application Specific Integrated Circuit) is a chip on your graphics card. The best you can hope for is a work around in the card firmware or driver.

panderiz · Post by **panderiz** » Tue Nov 09, 2010 11:45 pm

Well normally I reboot instantly, but when I just walked away for a bit and came back I was able to ssh that one time. Currently running the wine app windowed and haven't had the crash happen... yet. (It just happened... Wasn't patient enough to wait to see if I had access after some time)

I'm not entirely sure what you mean upgrade the binary blob and my firmware... If you mean upgrading ati-drivers I have always used to most up to date.

Post by **NeddySeagoon** » Wed Nov 10, 2010 8:56 pm

panderiz,

Binary-blob is a disparaging way to refer to closed source code, in this case, ati-drivers. When you say you use the latest driver, you mean the Gentoo testing one?

Your video card also has firmware, which is run to set the card up so its useful to any operating sustem. Its like the BIOS in your PC, which does the same thing for your motherboard. Most video card have updateable firmware. The video card firmware version is normally displayed on boot but you may need to boot several times to get it all as its only there for a second or two.

Check the card vendors site for later firmware for your card and try to determine what it fixes before you update.
As with BIOS updates, firmware updates are not risk free, in that a loss of power mid process may leave you with a perfectly servicable card but with no firmware loaded and no means of loading the firmware either. This state is known as 'bricked' because the card won't work with no firmware.

If you know how, have a JTAG programmer and know where to connect it, its recoverable but you really don't want to go there.

panderiz · Post by **panderiz** » Thu Nov 11, 2010 6:39 pm

I'm currently using ati-drivers-10.9-r1 from portage.

My video card is onboard so I'm presuming that I'd have to do an actual BIOS update from the motherboards website. Doesn't seem any BIOS update addresses anything related to that ASIC or video card. Might be missing something though. http://www.ecs.com.tw/ECSWebSite/Produc ... 20&LanID=9 Is what I'm looking at. Assuming ecs is the right site for my ECS motherboard.

Post by **NeddySeagoon** » Thu Nov 11, 2010 10:12 pm

panderiz,

I can't comment on your BIOS.

Your ati-drivers is the latest in portage.

panderiz · Post by **panderiz** » Fri Nov 12, 2010 1:38 am

Alright... Guess I can check to see for any work around's exist. I'm currently sitting here after it locked up again while I was playing Quake Live (Flash based). Seems like any kind of game will cause a lock up eventually.

Well after waiting ten minutes I decided I wasn't going to get access back so I rebooted. /var/log/message shows nothing even related to the lock up like it did the previous time. Would I be right for assuming the lock up happens before anything could be written to the log?

Just tried out magic sysrq and that wouldn't work when it locked up like normal, but when I caused X to lockup it works, so I know it does in fact work normally but this fglrx related lockup is causing a full desktop lockup... Lame

denysonique · Post by **denysonique** » Fri Nov 12, 2010 5:24 am

I have a similar lockup problem that is related to my graphic card. For instance when I run Compiz it hapens, when I play full screen YouTube videos it may happen too. My problem is a Xorg/driver (Intel GM45) issue.
If you think it's a xorg problem too you can try to debug: http://www.x.org/wiki/Development/Docum ... rDebugging

panderiz · Post by **panderiz** » Fri Nov 12, 2010 7:07 am

denysonique wrote:I have a similar lockup problem that is related to my graphic card. For instance when I run Compiz it hapens, when I play full screen YouTube videos it may happen too. My problem is a Xorg/driver (Intel GM45) issue.
If you think it's a xorg problem too you can try to debug: http://www.x.org/wiki/Development/Docum ... rDebugging

Well I've caused Xorg to lockup on me for testing purposes, and I'm still able to use magic sysrq, as well as being able to ssh into my machine.

When my lockups that I'm posting about happen my whole computer locks up I can't even ping it from my laptop, magic sysrq doesn't work and I (so far) have to do an unsafe reboot which a few things I have had running in the past have become corrupted..

denysonique · Post by **denysonique** » Fri Nov 12, 2010 10:34 am

you still may get some useful info in GDB. does magic+b or k work. or magic just doesn't work at all?