[00:15:30] --- dev-zero@jabber.org has become available [00:20:27] --- sxw mobile has become available [00:25:00] --- sxw mobile has left [00:33:10] --- dev-zero@jabber.org has left [01:00:11] --- sxw mobile has become available [01:00:27] --- sxw mobile has left [01:53:33] --- Russ has left: Disconnected [02:45:11] --- Jeffrey Altman has left: Replaced by new connection [03:16:32] --- jaltman has left: Lost connection [07:30:53] --- dev-zero@jabber.org has become available [07:31:18] --- dev-zero@jabber.org has left [07:31:24] --- dev-zero@jabber.org has become available [07:53:06] --- asedeno has left [07:53:17] --- asedeno has become available [07:59:17] --- asedeno has left [07:59:26] --- asedeno has become available [08:05:27] --- asedeno has left [08:05:37] --- asedeno has become available [08:11:37] --- asedeno has left [08:11:47] --- asedeno has become available [08:14:52] --- dev-zero@jabber.org has left [08:17:47] --- asedeno has left [08:17:57] --- asedeno has become available [08:23:57] --- asedeno has left [08:24:06] --- asedeno has become available [08:30:07] --- asedeno has left [08:30:16] --- asedeno has become available [08:36:17] --- asedeno has left [08:36:27] --- asedeno has become available [08:42:28] --- asedeno has left [08:42:36] --- asedeno has become available [08:48:37] --- asedeno has left [08:48:46] --- asedeno has become available [08:50:55] --- dev-zero@jabber.org has become available [08:54:47] --- asedeno has left [08:54:57] --- asedeno has become available [09:00:58] --- asedeno has left [09:01:06] --- asedeno has become available [09:05:23] --- dev-zero@jabber.org has left [09:05:23] --- dev-zero@jabber.org has become available [09:07:06] --- asedeno has left [09:07:16] --- asedeno has become available [09:13:16] --- asedeno has left [09:13:26] --- asedeno has become available [09:19:27] --- asedeno has left [09:19:36] --- asedeno has become available [09:25:37] --- asedeno has left [09:25:46] --- asedeno has become available [09:31:47] --- asedeno has left [09:31:56] --- asedeno has become available [09:38:11] --- asedeno has left [09:38:37] --- asedeno has become available [09:59:53] --- dev-zero@jabber.org has left [10:17:44] --- jaltman has become available [10:18:06] --- Russ has become available [11:31:00] Is it a bug that commit 35d5b6c “Remove unused configuration tests” removed HAVE_IN_SYSTM_H, which is still tested in src/rx/rx_kcommon.h? [11:32:13] probably [11:32:33] Er, no, because it was previously defined twice. Never mind. [11:33:31] ok then [11:34:56] I’m working on getting 1.4.x to compile on 2.6.33. I just pushed a series for master, but for 1.4.x, one of the tests deleted in that commit needs to be updated. Should I backport 35d5b6c, or fix the test, or do nothing? [11:36:43] "one of". which? [11:37:57] LINUX_KERNEL_SELINUX has an #include that shouldn’t be there. [11:38:35] ok. looking. [11:39:40] a minimal backport which just removes the LINUX_KERNEL_SELINUX test might be best, as it has the least possibility of side effect [11:50:03] --- Russ has left: Disconnected [11:54:24] --- haba has become available [13:00:27] --- Russ has become available [13:23:32] --- Claudio Bisegni has become available [13:23:46] derrick are you here? [13:32:02] --- Claudio Bisegni has left [14:07:53] --- jaltman has left: Replaced by new connection [14:07:54] --- jaltman has become available [14:44:29] --- haba has left [15:08:37] So, fcntl locks aren’t working correctly for me on a 2.6.33 client. [15:09:10] I get the message “afs: byte-range locks only enforced for processes on this machine.”, and indeed the locks are not being enforced on other machines. [15:09:59] I’m using Python’s fcntl.lockf, which straces to fcntl(3, F_SETLKW, {type=F_WRLCK, whence=SEEK_SET, start=0, len=0}) = 0 [15:11:23] Some debugging code I added to afs_lockctl: afs_warn("type=%hx whence=%hx start=%llx len=%llx pid=%x\n", af->l_type, af->l_whence, af->l_start, af->l_len, af->l_pid); shows type=1 whence=0 start=0 len=8000000000000000 pid=56c [15:45:58] Oh. I bet this broke in commit 5979cb6 (linux-byte-range-locks-sensibility-20090526): - flock.l_len = flp->fl_end - flp->fl_start; + flock.l_len = flp->fl_end - flp->fl_start + 1; That can overflow if fl_end == OFFSET_MAX, fl_start == 0. [16:03:19] Yeah, the problem goes away after reverting 5979cb6. [16:43:54] --- jaltman has left: Disconnected [16:44:09] --- jaltman has become available [19:13:59] --- Kevin Sumner has become available [21:14:24] --- deason has become available [23:48:04] --- deason has left