I have some machines running 1.4_rc2. After kernel upgrade to gentoo-sources 2.4.20-r5, some of them became very unstable, hangs with no message shortly after(or during) boot process. 2 boards with problems: Intel L440GX+(Dual P3 Coppermine) Tyan 2462UNG(Dual Athlon MP) Other matchines work fine. Lowering optimization flags does not help. I tried to tweak/disable many kernel options, does not help either. Change to vanilla- sources helped me out. Older ebuild files are not there now, can't test them. Reproducible: Always Steps to Reproduce: 1. 2. 3.
please create 2 new attachments with your kernel configs for the respective systems. Thanks, Jay
Created attachment 12186 [details] Kernel config Kernel config for Tyan 2462UNG Dual Athlon-MP The config for L440GX+ is almost the same(except for the CPU type).
I also have problems with athlon MP and gentoo-sources 2.4.20-r4 and r5. The machine doesn't crash but seems to lose interrupts and doesn't respond correctly. ie to get through "init" I have to keep hitting the keys. 2.4.20-r2 is fine. I've done a little problem hunting and the problem seems to be related to (200) Timer frequency (HZ) (200). I see the same problem when I tried the ck-sources which has the same patch. I don't believe these kernels where ever tested on a MP system which is ok, but it should not be in stable or it needs to have some other method in portage so it flags as a untested in MP systems.
I have this problem as well. Let me know if you need anything tested. cheers. Jonathan.
well, i have some dual systems here that do not reproduce the problem, so i have some ideas and would like some testing since the few of you here seem to be having the issue. i think this issue may stem from the new adaptec code (01May2003). ie . CONFIG_SCSI_AIC7XXX=y let me get some changes together, and i'll post the patches here or on my website. Thanks, Jay
I just tried a kernel with no SCSI support at all and the problem is still there. ac-sources have been ok if that helps. My money is still on something to do with this patch. │ (200) Timer frequency (HZ) (200) I did a quick check on gentoo-users asking if there was anyone with an MP-Athlon machine that had problems. Got a few confirmations of problems there, nobody said they had a working version of gentoo-sources-r5.
Same here. 2.4.20-r5 brings system into unstable mode. switching to 2.4.19-r10 solves problem.
forgot to mention: System is: Dual Athlon XP, Tyan Tiger motherboard.
is this still an issue ?
dunno, I've been running 2.6-test for a while now.
If have problems like this too. I'm using gentoo-sources 2.4.20-r7 and from time to time the whole pc freezes. No input, neither mouse nor keyboard is possible, remote login times out. There is no coredump and no entry in one of the logs that looks like a problem. I can't reproduce this behaviour as it has nothing to do with the utilization of my PCs resources. Sometime I'm compiling some gentoo stuff, sometimes I'm only playing KMines, sometimes all is fine. I haven't tried with another kernel (older gentoo oder vanilla) yet. My system seems to work just fine under RH 7.3, at least I do not experience this freezing nor any other crashes. My hardware is a 1 GHz Athlon T-Bird on a DFI Board with VIA KT133 chipset, 512 MB SD-RAM, NVIDIA GForce2 MX, Tekram DC390F SCSI-Controller. Friend of mine experiences similar behaviour with his gentoo-box. regards, Maik
Bumping priority... Please do the following: remove ACPI and APM support, and enable or disable the Preemptive scheduling support [i.e. on->off, off->on]. Maik: Also please attach your kernel .config
Created attachment 19620 [details] Maiks Kernel-Configuration This is my current config (before changing settings for Preemt Scheduler and APM/ACPI). I'm currently compiling a kernel with switched-off Low-Latency, Preemptive Kernel, Powermanagement. regards, Maik
Hi there, I still got the problems even without PM and Preemptive Scheduling. Still there are no entrys in the logs. Any suggstions? rgds, MaDMaik
Okay. Can I suggest trying vanilla-2.4.20 [[ NOT 2.4.21+ as we can't reproduce this otherwise ]] or trying the latest gentoo-sources. I'd be grateful if you can confirm this or confirm that you don't get this on vanilla-2.4.20. Thanks.
Please comment if this still holds true for more recent versions of the gentoo-sources since we are up to 2.4.22-r1.
Hi there, I'm now at gentoo-sources 2.4.22-r2 for some days and at the moment I have no more problems. BTW 2.6.0-gentoo-r1 works excellent for me, as I test this for quite a while now. But I will stay at 2.4.22-r2 for some time now to check whether my problem will happen again. As I said before, for some days now it seems quite ok. rgds, Maik
Good morning, I have some news. First neither 2.4.22-gentoo-r2 nor 2.6.1-gentoo prevented me from getting a frozen system. Today it was a bit different. First part of my kde locked up but I could still use the application currently open (it was konqueror) then my keyboard locked with my mouse still working. A look into the log-files: <<<< Jan 22 19:36:18 [kernel] Unable to handle kernel paging request at virtual address 736d3a74 >>>> nothing more but I guess this is a start. By the way I have the feeling that my problem is somewhere related to X. Any suggestions? rgds, MaDMaik
did you try memtest86 and cpuburn ?
Yes, I just did memtest and cpuburn. Memtest ran for 8 hours doing 12 tests with no errors. cpuburn ran for 1,5 hours with no problems at all, System Temperature was afterwards 35 Celsius, CPU Temperature 41 Celsius.
Maik: Bug #44391. Anybody else getting this for 2.4.20: Can you set CONFIG_HZ to 100 rather than 200?
Do the lockups occur with a newer 2.6 kernel or a newer 2.4 kernel?
2.4.20 is ancient. This was fixed in newer kernels.