[00:57:27] --- Simon Wilkinson has become available [01:01:08] If we need to pullup changes for 2.6.38, then either you could do that in the packaging, or we could do 1.4.14.1 with them in [01:17:06] It’s getting more annoying to cherry-pick these because of 02f2c7cb “Linux: Add autoconf macro for structure checks”. [01:19:26] You could pull in the macro definition, and only use it for new checks. [02:42:49] Okay, well, now I have a 1.4.x that seems to work on natty. In other news, I feel like I want to go take a shower. [02:43:49] Which patches did you need to pullup? [02:44:15] We should make sure those are on openafs-stable-1_4_x, and then take a view as to whether we do 1.4.14.1, or look again at 1.4.15 [02:45:27] “Pullup” is a stretch, but a97b201 Linux: Add autoconf macro for structure checks 7a77ad8 Linux: Add general autoconf macro for Linux kernel e735919 linux: 2.6.38: New d_op handling 5a6d2c2 linux: 2.6.38: Make d_revalidate RCU-walk aware 8d44de7 Linux: allow compile flags to be passed to AC_CHECK_LINUX_BUILD 0b1ccb1 Linux: 2.6.38: Adjust for permission inode operation changes ba86aaf Linux: 2.6.38: deal with dcache_lock removal 28e7412 Linux: 2.6.38: dentry->d_count is not an atomic [02:46:53] Which is just everything mentioning “2.6.38” and three dependencies for autoconf macros. [02:47:13] If you fancy pushing the work that you've done to gerrit (refs/for/openafs-stable-1_4_x) that would be great [02:47:33] Sure. [02:47:42] Even if the patches aren't direct cherry-picks, include the "cherry picked from commit" line, as that's what our tools use to work out what's on which branches. [02:48:42] Right, except I wasn’t sure what to do about the autoconf patches, which I only picked a small part of. [02:49:58] We're never going to pull up the full autoconf patches, so I think it's fine to include the cherry picked line, and note below it that it's a small subset of the master patch. [02:50:11] If I'd thought, I would have pushed those changes in two parts, to avoid exactly this problem. [02:52:24] The other nasty bit was that LINUX_IOP_I_PERMISSION_TAKES_NAMEIDATA was modified both by the autoconf patches and by “Adjust for permission inode operation changes”. It looked like a bugfix to the test, so I left that all in. [02:59:25] Pushed to gerrit (3992–3999). [03:01:01] Yeah, I think that was a bugfix, from memory. [03:01:06] Cool. I shall take a look. [03:02:37] I think some bits of that has gone horribly wrong. [03:03:39] You don't seem to have modified the ChangeId on "Linux: Add autoconf macro for structure checks", so your patch has been added to the master change (1853), rather than created as a new change. [03:04:09] Oh no, tell a lie. I'm just not properly awake yet. [03:23:28] --- Simon Wilkinson has left [06:33:17] --- jaltman/FrogsLeap has left: Disconnected [07:01:31] --- mdionne has become available [07:13:13] > Yeah, I think that was a bugfix, from memory. It's a bugfix, but a bug specific to 2.6.38. The test was OK with older kernels. [07:14:05] The kernel change that makes it a bug is the same one for which the new configure test is added, hence the single commit. [08:50:41] --- mdionne has left [09:51:53] --- Russ has become available [11:10:55] --- deason has become available [15:15:20] --- deason has left [15:50:46] --- deason has become available [16:31:35] --- jaltman/FrogsLeap has become available [17:40:27] --- jaltman/FrogsLeap has left: Replaced by new connection [17:40:28] --- jaltman/FrogsLeap has become available [17:41:26] --- jaltman/FrogsLeap has left: Replaced by new connection [17:41:26] --- jaltman/FrogsLeap has become available [17:42:37] --- jaltman/FrogsLeap has left: Replaced by new connection [17:42:38] --- jaltman/FrogsLeap has become available [17:43:53] --- jaltman/FrogsLeap has left: Replaced by new connection [17:43:53] --- jaltman/FrogsLeap has become available [17:50:24] --- jaltman/FrogsLeap has left: Replaced by new connection [17:50:25] --- jaltman/FrogsLeap has become available [17:57:32] --- jaltman/FrogsLeap has left: Replaced by new connection [17:57:33] --- jaltman/FrogsLeap has become available [17:58:19] --- jaltman/FrogsLeap has left: Replaced by new connection [17:58:20] --- jaltman/FrogsLeap has become available [18:28:09] --- jaltman/FrogsLeap has left: Disconnected [18:28:17] --- jaltman/FrogsLeap has become available [19:52:49] Note that gerrit/4020 means the amd64_fbsd_90 package has a kernel module that won't load. [19:55:50] Hmm, probably the fbsd_82 one as well. [19:59:46] /afs/sipb.mit.edu/project/freebsd/openafs/amd64_fbsd_{82,90}/ updated. [20:24:18] Also, -j seems broken on master for me, yet. [20:24:50] /afs/sipb.mit.edu/project/freebsd/openafs/build.log [20:34:43] --- asedeno has left [20:36:05] > cd: can't cd to src > *** Error code 2 [20:36:20] can't say I've seen that before [20:36:39] --- asedeno has become available [20:37:21] (which is probably from the "cd src && cd venus/test && make all", which is probably after the completed src/venus submake) [20:45:17] Also, 'make clean' complained in src/JAVA/libjafs (I think) about needing an operator in several places. [20:58:11] I presume because it contains "ifeq"s and stuff; JAVA/libjafs doesn't get built on fbsd, but the 'clean' target still tries to go through it [20:58:38] Yeah, that seems right. [22:12:26] --- deason has left