Hello gentlemans, I'm back to this issue, hopefully I'm not bothering too much :( To summarize first: we're experiencing problems with 2.6.32.x KVM guests crashing when running tcpdump. The problem has been fixed since then, in newer kernels, but I'd like to find the fix and get it to 2.6.32.x stable series. I've been bisecting this lately, and narrowed this to 2.6.34 - the problem has been fixed there. The problem is, I've somehow got cornered there: I've got ~3200 commits left, but got into state, where I can't test commits further - there was regression in 2.6.34-rc1 (I think), which caused tcpdump to stop working (see http://lkml.org/lkml/2010/3/9/439) Regression was introduced by 914c8ad2d18b62ad1420f518c0cab0b0b90ab308 and fixed by 1162563f82b434e3099c9e6c1bbdba846d792f0d. I can still reproduce the problem with 1162563f82b434e3099c9e6c1bbdba846d792f0d, and can't with 4961e02f1999e1c3468c09b2669c94d7c3ae82a8. When I try to continue to bisect, git gets me to fa0d976298b25d090fafc3460c63fee1c8eea854, the commit which is somewhere between 1162563f82b434e3099c9e6c1bbdba846d792f0d and 4961e02f1999e1c3468c09b2669c94d7c3ae82a8. Strange thing is, that I'm again hitting the other problem with tcpdump, which should have been fixed by 1162563f82b434e3099c9e6c1bbdba846d792f0d! When I check the net/packet/af_packet.c, the fix really doesn't seem to be there although that commit is a lot later then 1162563f82b434e3099c9e6c1bbdba846d792f0d! I'm not sure about what I'm doing wrong, I guess it has to do something with merging branches... Can somebody more knowledgeable about git hint me about how to continue? I'd really like to get this fixed for 2.6.32... Thanks a lot to anyone willing to help! with best regards nik On Sun, Apr 17, 2011 at 05:15:24PM +0200, Nikola Ciprich wrote: > OK, just wanted to let You know I were testing it quite a lot, and I'm not able to reproduce this with 2.6.38.3-rc1. > So the bug must have been fixed. > I'll bisect it to find proper fix so it can be posted to stable... > n. > > > On Sat, Apr 02, 2011 at 09:42:26PM +0200, Nikola Ciprich wrote: > > Hello Stefan! > > > > > It looks like your guests are SMP. How many vcpus are you running? > > > How many physical cpus does /proc/cpuinfo list on the host? > > one of guests is SMP (8cpus), one is UP, host has 2x4 cores. > > > > > > Is the host overloaded when this occurs? > > nope > > > > > > > > Are there any clues in host dmesg? > > nothing :( > > I guess I shall try 2.6.38 or maybe latest git to check if the problem > > is still present... > > > > > > > > Stefan > > > -- > > > To unsubscribe from this list: send the line "unsubscribe kvm" in > > > the body of a message to majordomo@xxxxxxxxxxxxxxx > > > More majordomo info at http://vger.kernel.org/majordomo-info.html > > > > > > > -- > > ------------------------------------- > > Ing. Nikola CIPRICH > > LinuxBox.cz, s.r.o. > > 28. rijna 168, 709 01 Ostrava > > > > tel.: +420 596 603 142 > > fax: +420 596 621 273 > > mobil: +420 777 093 799 > > > > www.linuxbox.cz > > > > mobil servis: +420 737 238 656 > > email servis: servis@xxxxxxxxxxx > > ------------------------------------- > > -- > > To unsubscribe from this list: send the line "unsubscribe kvm" in > > the body of a message to majordomo@xxxxxxxxxxxxxxx > > More majordomo info at http://vger.kernel.org/majordomo-info.html > > > > -- > ------------------------------------- > Ing. Nikola CIPRICH > LinuxBox.cz, s.r.o. > 28. rijna 168, 709 01 Ostrava > > tel.: +420 596 603 142 > fax: +420 596 621 273 > mobil: +420 777 093 799 > > www.linuxbox.cz > > mobil servis: +420 737 238 656 > email servis: servis@xxxxxxxxxxx > ------------------------------------- -- ------------------------------------- Ing. Nikola CIPRICH LinuxBox.cz, s.r.o. 28. rijna 168, 709 01 Ostrava tel.: +420 596 603 142 fax: +420 596 621 273 mobil: +420 777 093 799 www.linuxbox.cz mobil servis: +420 737 238 656 email servis: servis@xxxxxxxxxxx -------------------------------------
Attachment:
pgp2vT5RmTPaK.pgp
Description: PGP signature
