[00:57:27] --- lars.malinowsky has left [01:35:04] --- lars.malinowsky has become available [03:42:53] --- reuteras has left [04:04:12] --- reuteras has become available [04:04:17] --- reuteras has left [05:49:40] --- steven.jenkins has become available [06:12:32] --- abo has left [06:26:57] --- abo has become available [06:31:12] --- reuteras has become available [07:12:50] --- reuteras has left [07:12:51] --- reuteras has become available [07:32:11] --- deason has become available [07:32:23] --- reuteras has left [07:32:23] --- reuteras has become available [07:37:58] --- reuteras has left [07:37:59] --- reuteras has become available [07:42:06] --- reuteras has left [07:42:06] --- reuteras has become available [07:44:49] --- reuteras has left [07:44:49] --- reuteras has become available [07:48:33] --- lars.malinowsky has left [07:53:18] --- reuteras has left [07:53:18] --- reuteras has become available [08:11:07] --- meffie has left [08:18:05] --- jaltman/FrogsLeap has left: Replaced by new connection [08:18:06] --- jaltman/FrogsLeap has become available [08:34:42] --- Russ has become available [08:47:59] --- reuteras has left [09:11:00] --- Russ has left: Disconnected [09:23:33] --- Simon Wilkinson has left [09:32:55] --- sxw has left [09:44:43] --- jaltman/FrogsLeap has left: Disconnected [09:44:51] --- jaltman/FrogsLeap has become available [09:58:17] --- sxw has become available [12:04:19] --- ksumner has left [12:04:28] --- jaltman/FrogsLeap has left: Disconnected [12:08:01] --- jaltman/FrogsLeap has become available [13:21:06] --- mfelliott has left [14:30:35] The openafs-security queue would be much more effective if RT actually sent emails for bugs in it. [14:33:38] what are the "substantial changes" in 2.6.39, anyway? [14:33:57] 2.6.39 does away with the big kernel lock entirely. [14:34:26] ah right; I remember seeing that discussion [14:34:27] 1.4 isn't safe without the BKL, and it needs a substantial number of changes to make it safe. [14:35:17] glock doesn't help us there? [14:35:35] No, because the GLOCK doesn't stop other Linux processes from doing things behind our back. [14:36:54] so it's adapating to the correct linux-level locks, not anything really to do with our internal locking mechanisms [14:36:57] The problem is with things like dcache accesses, where we aren't necessarily the only ones who are allowed to access the dcache. Because (in 1.4) we don't use the proper accessor functions, we aren't thread safe against other bits of the kernel which are working there. [14:37:28] Yes - our internal locking mechanism remains as wretched as ever :) [15:17:33] --- mfelliott has become available [15:35:06] --- Simon Wilkinson has become available [15:43:45] --- deason has left [15:57:34] --- Russ has become available [17:29:26] --- Russ has left: Disconnected [19:16:37] --- Simon Wilkinson has left [19:38:44] --- Russ has become available [20:26:49] --- Russ has left: Disconnected [20:27:11] --- Russ has become available [23:20:05] --- Russ has left: Disconnected