Date: Wed, 12 Dec 2001 16:50:02 -0800 (PST) From: Goran Dokic <goran@euro.net> To: freebsd-bugs@FreeBSD.org Subject: Re: kern/25788: microuptime() weny backwards, again Message-ID: <200112130050.fBD0o2N14289@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR kern/25788; it has been noted by GNATS. From: Goran Dokic <goran@euro.net> To: freebsd-gnats-submit@FreeBSD.org, ming@unixbox.com Cc: Subject: Re: kern/25788: microuptime() weny backwards, again Date: Thu, 13 Dec 2001 01:49:25 +0100 Hi, I just suffered the 'microuptime() went backwards' bug in FreeBSD fire.8hz.com 4.4-RELEASE FreeBSD 4.4-RELEASE #1: i386 And XFree 4.1. I was diffing a CD I just burned (cd was in my 32* UltraPlex CD-ROM on an Adaptec 2940 controller) with the original file when the machine locked up: Dec 13 00:35:51 fire /kernel: cd9660: RockRidge Extension Dec 13 00:37:46 fire /kernel: microuptime() went backwards (520036.285909 -> 520 036.-695120903) That was the last log entry and almost exactly at that moment I lost the machine. The only thing doing anything was the audio card, it kept playing the last byte it received. It was running diff quite actively at the moment it chocked. It might not have anything to do with each other. But as it came pretty much at the same time as the crash and I never before saw a message stating anything about the microuptime(), I figure it might be related. For the interested: Dec 13 00:41:49 fire /kernel: goran@fire.8hz.com:/usr/src/sys/compile/FIRE Dec 13 00:41:49 fire /kernel: Timecounter "i8254" frequency 1193182 Hz Dec 13 00:41:49 fire /kernel: CPU: Pentium III/Pentium III Xeon/Celeron (930.24- MHz 686-class CPU) Dec 13 00:41:49 fire /kernel: Origin = "GenuineIntel" Id = 0x686 Stepping = 6 Dec 13 00:41:49 fire /kernel: Features=0x383f9ff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE, CX8,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,MMX,FXSR,SSE> Dec 13 00:41:49 fire /kernel: real memory = 133955584 (130816K bytes) Dec 13 00:41:49 fire /kernel: avail memory = 126476288 (123512K bytes) I have an Intel OEM board (gateway 2k) which are not the fastest, but experience shows they tend to be quite stable under varied conditions... I just swapped a crocky ABIT BE6-II board in the hope of parting from the numerous lockups (those boards do suck, btw). But, not a week after the swap I get this. (The messages file saved from the other board didn't speak of 'microuptime() went backwards', the hardware just sucked). Well, I hope someone finds the cure soon... I'm cvsupping to 4.4STABLE now, to see if it'll happen again. If it does, i'll try setting the kern.timecounter.method to 1, as has been suggested to ming@unixbox.com. These are my current settings: fire% sysctl kern.timecounter kern.timecounter.method: 0 kern.timecounter.hardware: i8254 goran To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200112130050.fBD0o2N14289>