https://bugzilla.novell.com/show_bug.cgi?id=376165 User drankinatty@suddenlinkmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=376165#c1 --- Comment #1 from David Rankin 2008-04-01 23:37:23 MST --- Created an attachment (id=205559) --> (https://bugzilla.novell.com/attachment.cgi?id=205559) machine check events log machine check events log with nvidia kernel module unload indicated at bottom of file (there are no further mce after unloading the nvidia module) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=376165 User drankinatty@suddenlinkmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=376165#c3 --- Comment #3 from David Rankin 2008-04-01 23:40:19 MST --- Created an attachment (id=205561) --> (https://bugzilla.novell.com/attachment.cgi?id=205561) dmesg without acpi_use_timer_override kernel parameter dmesg file without passing acpi_use_timer_override to the kernel at boot -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=376165 User drankinatty@suddenlinkmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=376165#c5 --- Comment #5 from David Rankin 2008-04-01 23:42:33 MST --- Created an attachment (id=205563) --> (https://bugzilla.novell.com/attachment.cgi?id=205563) hwinfo for system hwinfo for the system -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=376165 User drankinatty@suddenlinkmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=376165#c7 --- Comment #7 from David Rankin 2008-04-03 10:37:47 MST --- Guys, Here are additional mce(s) captured in syslog but absent in /var/log/mcelog. The nvidia driver was *not* loaded when these occurred and apparently were only written to syslog due to the crash/reboot occurring before the next cron run of /usr/sbin/mcelog. (currently set to run at 1 min. intervals). The following of the mce(s) caught in /var/log/messages: Apr 1 01:35:01 nirvana /usr/sbin/cron[3761]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:35:50 nirvana kernel: [ 198.079706] Machine check events logged Apr 1 01:36:01 nirvana /usr/sbin/cron[3766]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:36:01 nirvana mcelog: HARDWARE ERROR. This is *NOT* a software problem! Apr 1 01:36:01 nirvana mcelog: Please contact your hardware vendor Apr 1 01:36:01 nirvana mcelog: CPU 0 1 instruction cache Apr 1 01:36:01 nirvana mcelog: TSC 6f61bca83a Apr 1 01:36:01 nirvana mcelog: ADDR 2b66e64040f0 Apr 1 01:36:01 nirvana mcelog: Apr 1 01:36:01 nirvana mcelog: memory/cache error 'instruction fetch mem transaction, instruction transaction, level 1' Apr 1 01:36:01 nirvana mcelog: STATUS 9400000000000151 MCGSTATUS 0 Apr 1 01:36:01 nirvana mcelog: HARDWARE ERROR. This is *NOT* a software problem! Apr 1 01:36:01 nirvana mcelog: Please contact your hardware vendor Apr 1 01:36:01 nirvana mcelog: CPU 1 1 instruction cache Apr 1 01:36:01 nirvana mcelog: TSC 6f61bd690e Apr 1 01:36:01 nirvana mcelog: ADDR ffff804454f0 Apr 1 01:36:01 nirvana mcelog: Apr 1 01:36:01 nirvana mcelog: bit62 = error overflow (multiple errors) Apr 1 01:36:01 nirvana mcelog: memory/cache error 'instruction fetch mem transaction, instruction transaction, level 1' Apr 1 01:36:01 nirvana mcelog: STATUS d400000000000151 MCGSTATUS 0 Apr 1 01:37:01 nirvana /usr/sbin/cron[3805]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:38:01 nirvana /usr/sbin/cron[3812]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:39:01 nirvana /usr/sbin/cron[3820]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:39:01 nirvana mcelog: HARDWARE ERROR. This is *NOT* a software problem! Apr 1 01:39:01 nirvana mcelog: Please contact your hardware vendor Apr 1 01:39:01 nirvana mcelog: CPU 0 1 instruction cache Apr 1 01:39:01 nirvana mcelog: TSC 9c52fbb5cd Apr 1 01:39:01 nirvana mcelog: ADDR 77a8d270 Apr 1 01:39:01 nirvana mcelog: Apr 1 01:39:01 nirvana mcelog: bit62 = error overflow (multiple errors) Apr 1 01:39:01 nirvana mcelog: memory/cache error 'instruction fetch mem transaction, instruction transaction, level 1' Apr 1 01:39:01 nirvana mcelog: STATUS d400000000000151 MCGSTATUS 0 The complete collection of the mce(s) from /var/log/messages are contained in the attachment "mce_syslog" provided along with this post. Let me know what else I can provide and I will respond as soon as I can. Also, if an account on the box would be helpful, we can arrange that as well. Thanks! -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=376165 User drankinatty@suddenlinkmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=376165#c9 --- Comment #9 from David Rankin 2008-04-03 11:25:09 MST --- (In reply to comment #7 from David Rankin)

Guys,

Here are additional mce(s) captured in syslog but absent in /var/log/mcelog. The nvidia driver was *not* loaded when these occurred and apparently were only written to syslog due to the crash/reboot occurring before the next cron run of /usr/sbin/mcelog. (currently set to run at 1 min. intervals).

The following of the mce(s) caught in /var/log/messages:

Apr 1 01:35:01 nirvana /usr/sbin/cron[3761]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:35:50 nirvana kernel: [ 198.079706] Machine check events logged Apr 1 01:36:01 nirvana /usr/sbin/cron[3766]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:36:01 nirvana mcelog: HARDWARE ERROR. This is *NOT* a software problem! Apr 1 01:36:01 nirvana mcelog: Please contact your hardware vendor Apr 1 01:36:01 nirvana mcelog: CPU 0 1 instruction cache Apr 1 01:36:01 nirvana mcelog: TSC 6f61bca83a Apr 1 01:36:01 nirvana mcelog: ADDR 2b66e64040f0 Apr 1 01:36:01 nirvana mcelog: Apr 1 01:36:01 nirvana mcelog: memory/cache error 'instruction fetch mem transaction, instruction transaction, level 1' Apr 1 01:36:01 nirvana mcelog: STATUS 9400000000000151 MCGSTATUS 0 Apr 1 01:36:01 nirvana mcelog: HARDWARE ERROR. This is *NOT* a software problem! Apr 1 01:36:01 nirvana mcelog: Please contact your hardware vendor Apr 1 01:36:01 nirvana mcelog: CPU 1 1 instruction cache Apr 1 01:36:01 nirvana mcelog: TSC 6f61bd690e Apr 1 01:36:01 nirvana mcelog: ADDR ffff804454f0 Apr 1 01:36:01 nirvana mcelog: Apr 1 01:36:01 nirvana mcelog: bit62 = error overflow (multiple errors) Apr 1 01:36:01 nirvana mcelog: memory/cache error 'instruction fetch mem transaction, instruction transaction, level 1' Apr 1 01:36:01 nirvana mcelog: STATUS d400000000000151 MCGSTATUS 0 Apr 1 01:37:01 nirvana /usr/sbin/cron[3805]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:38:01 nirvana /usr/sbin/cron[3812]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:39:01 nirvana /usr/sbin/cron[3820]: (root) CMD (/usr/sbin/mcelog --k8 --syslog) Apr 1 01:39:01 nirvana mcelog: HARDWARE ERROR. This is *NOT* a software problem! Apr 1 01:39:01 nirvana mcelog: Please contact your hardware vendor Apr 1 01:39:01 nirvana mcelog: CPU 0 1 instruction cache Apr 1 01:39:01 nirvana mcelog: TSC 9c52fbb5cd Apr 1 01:39:01 nirvana mcelog: ADDR 77a8d270 Apr 1 01:39:01 nirvana mcelog: Apr 1 01:39:01 nirvana mcelog: bit62 = error overflow (multiple errors) Apr 1 01:39:01 nirvana mcelog: memory/cache error 'instruction fetch mem transaction, instruction transaction, level 1' Apr 1 01:39:01 nirvana mcelog: STATUS d400000000000151 MCGSTATUS 0

The complete collection of the mce(s) from /var/log/messages are contained in the attachment "mce_syslog" provided along with this post.

Let me know what else I can provide and I will respond as soon as I can. Also, if an account on the box would be helpful, we can arrange that as well.

Thanks!

NOTE: For testing purposes the "acpi_use_timer_override" kernel parameter was removed after the 4/3 reboot at 0400. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=376165 User drankinatty@suddenlinkmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=376165#c11 David Rankin changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|INVALID | --- Comment #11 from David Rankin 2008-04-03 23:40:12 MST --- WAIT, WAIT a minute. I don't mind buying a new board, but I don't think this is a HARDWARE error. That is the point of the bug report. Just because the log says it is, that doesn't make it so. Since when did we start dismissing potential problems because the "log says it's hardware". God knows log error messages are NOT infallible. Did you even read the discussion above?? The problem goes AWAY when the *software* nvidia driver isn't loaded. I've been running mprime for the past 10 hours with NO errors what so ever. This certainly merits more attention than an arbitrary invalidation of the bug "because the log says so?" 10 hours no error no *software* driver loaded: ps axf | grep mprime 788 pts/1 S+ 0:00 \_ grep mprime 19956 ? RN 628:02 ./mprime I want help you validate that there isn't a generic problem with some of the newer x86_64 chipsets. If you have already diagnosed the problem from your look at the thorough information I provided, then "what is it?" Are you telling the world that you are absolutely sure a software error could not create the crash problem that is being experienced when the nvidia kernel module is loaded? What triggered the errors with the nvidia software module loaded at addresses: ADDR 2aaaad165cf0 ADDR 2aaf6b4262a0 ADDR 2aafbc3102b0 ADDR 2ab5d9f3d5f0 ADDR 2acbc6b2dbf0 ADDR 2ad0a10d93e0 ADDR 2ad2c6d787a0 ADDR 2b063ff544f0 ADDR 2b351ba197a0 ADDR 2b351ba197a0 ADDR 2b351ba197a0 ADDR 2b351ba197a0 ADDR 2b47840ed7b0 ADDR 2b7505fe1fa0 ADDR 2b76e1f8d3c0 ADDR 2b81c17b6ad0 ADDR 2b850b97d1f0 ADDR 2b8b094a6e20 ADDR 41df70 ADDR 43d930 ADDR eb4ceaa0 ADDR f0ce3900 ADDR f0cfb940 ADDR f0cfde20 ADDR f0d150f0 ADDR f688caf0 ADDR f688cc30 ADDR f6894530 ADDR f68adb70 ADDR f68ba940 ADDR f68d50f0 ADDR f69553f0 ADDR f74cda70 ADDR f74cfb20 ADDR f78cdbb0 ADDR f78e3900 ADDR ffff80220a20 ADDR ffff802603e0 ADDR ffff802962c0 ADDR ffff802a88a0 ADDR ffff8034d760 ADDR ffff803586a0 ADDR ffff803f3d80 ADDR ffff803f55f0 ADDR ffff803f55f0 ADDR ffff803f56b0 ADDR ffff880b1020 ADDR ffff88658f20 Why do the errors disappear when the *software* module is unloaded? If you're right, I'm happy to send the board back, burn it, whatever. However, if I'm right, and you miss this opportunity to better your product and this problem snowballs, a lot of people will be getting screwed because we couldn't take the time necessary to either rule in or rule out a software problem. If you know what the problem is, just post it and I'll happily send the system back. If you don't know what it is and are just trying to punt this bug as invalid, then it is time we get down to work and figure out what the problem is. Thanks. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=376165 User drankinatty@suddenlinkmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=376165#c13 --- Comment #13 from David Rankin 2008-04-05 00:33:16 MST --- Guys, I apologize for my lack of understanding regarding the mce being solely hardware related. My concern was a possible software bug masquerading as a hardware error. Anders J. and Carlos E.R. shed a little more light on the situation. I will refocus on the hardware aspects of the issue. Thanks. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=376165 User drankinatty@suddenlinkmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=376165#c15 --- Comment #15 from David Rankin 2008-05-06 00:20:26 MST --- Guys, I thought I would pass along the fact that you were correct. After replacing virtually everything in the box, I rma'ed the motherboard back to Tyan and was given a replacement. All works as it should and without a single mce. Thanks for your help. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.