[00:15:20] --- Jeffrey Altman has left [01:46:39] --- Simon Wilkinson has left [03:40:28] --- mdionne has become available [03:41:48] only tried more recent kernels so far - 2.6.35 and current git (pre 2.6.36 rc). 2.6.34 won't boot with my current userspace [03:42:42] if I can find a kernel where it works, I might be able to do a bisection on the kernel side [03:43:11] --- mdionne has left [04:06:53] --- Simon Wilkinson has become available [05:49:54] --- Simon Wilkinson has left [05:50:08] it may be that i didn't get the same messages as you, and so it didn't work in the same cases here either, marc [06:36:33] --- Simon Wilkinson has become available [07:00:07] --- deason has become available [07:07:04] --- reuteras has left [07:16:52] --- jaltman has left: Disconnected [07:17:04] --- jaltman has become available [07:19:06] --- Simon Wilkinson has left [08:02:49] --- meffie has become available [08:48:58] --- meffie has left [09:06:32] --- rra has become available [09:23:21] --- jaltman has left: Disconnected [09:23:25] --- jaltman has become available [09:49:14] --- jaltman has left: Disconnected [10:10:23] --- jaltman has become available [11:31:33] --- pod has become available [11:55:37] --- matt has become available [13:04:07] --- mdionne has become available [13:09:51] --- mdionne has left [13:20:52] --- mdionne has become available [13:22:30] --- shadow has left [13:24:17] the oops itself is simple enough - we specify a length of 256 to kernel_recvmsg and give it a NULL iovec pointer. if length > 0, the kernel tries to test iovec->iov_len [13:26:55] but the oops is not the real problem here. after fixing the above, I now see several of the afs background daemons go to 100% CPU, and the system eventually freezes solid [13:49:11] well, it's disabled as of last night anyway [13:50:20] yeah I saw that, thanks. but I still wanted to see where the oops was coming from. [13:51:51] oh, i'd like to know what the othe problem is too, but at least no one else need be saddled with it for now [13:53:40] change 2546 for the oops [14:00:44] you're faster than buildbot [14:01:30] it's easy to rebuild one file. [14:01:51] buildbot can't assume things about it's build sandbox that i can about mine [14:01:56] (in this case) [14:06:42] BTW I didn't test those 2.6.36 changes against older kernels (rhel 5, etc.). My VMs are not functional at the moment [14:07:10] they seem to be working here. i'll reconfirm in a moment. [14:10:43] does 1.4.12.1 work with linux 2.6.35? or was 2.6.34 the last one? [14:22:59] don't remember, but the only fix that looks related to 2.6.35 is not in the 1.4 branch (commit 4d89ce0a) [14:25:08] so I think 1.4.12.1 supports up to 2.6.34 [14:57:58] --- jaltman has left: Disconnected [14:58:01] --- jaltman has become available [15:04:58] When it says "called with avc write locked" I guess you're supposed to. [15:05:36] ptobably helps [15:07:33] --- jaltman has left: Replaced by new connection [15:07:35] --- jaltman has become available [15:19:19] --- jaltman has left: Disconnected [15:50:58] --- deason has left [16:35:13] --- jaltman has become available [16:51:33] --- matt has left [16:52:14] --- mdionne has left [17:31:38] --- mdionne has become available [18:45:42] --- mdionne has left [18:51:55] --- rra has left: Disconnected [19:08:59] --- Russ has become available [21:03:26] --- deason has become available [21:59:16] --- deason has left [22:23:28] --- kaj has become available [22:41:38] --- reuteras has become available [22:59:58] --- Russ has left: Disconnected [23:01:30] --- kaj has left [23:54:35] --- kaj has become available