[00:35:07] --- dev-zero@jabber.org has left [00:37:14] --- dev-zero@jabber.org has become available [00:42:05] --- dev-zero@jabber.org has left [00:45:30] --- dev-zero@jabber.org has become available [00:49:09] --- dev-zero@jabber.org has left [00:53:20] --- dev-zero@jabber.org has become available [01:01:49] --- dev-zero@jabber.org has left [01:05:23] --- dev-zero@jabber.org has become available [01:50:14] --- squinney has become available [02:37:29] --- squinney has left [02:40:00] --- stephen.quinney has become available [02:55:07] --- ballbery has left [02:55:16] --- ballbery has become available [02:57:42] --- stephen.quinney has left [03:00:12] --- stephen.quinney has become available [03:08:48] --- Simon Wilkinson has become available [03:45:27] --- Simon Wilkinson has left [04:00:49] --- Simon Wilkinson has become available [04:22:36] --- stephen.quinney has left [04:22:58] --- stephen.quinney has become available [04:46:26] --- kula has left [04:46:49] --- Simon Wilkinson has left [04:47:08] --- kula has become available [05:08:09] --- kula has left: Lost connection [05:08:09] --- dev-zero@jabber.org has left: Lost connection [05:13:29] --- dev-zero@jabber.org has become available [05:52:21] --- Simon Wilkinson has become available [06:11:41] --- kula has become available [06:23:15] --- kula has left [06:24:27] --- kula has become available [06:29:46] --- kula has left [06:31:48] --- kula has become available [06:34:10] --- mvitale has become available [06:39:41] --- stephen.quinney has left [06:43:05] --- dev-zero@jabber.org has left [07:26:13] --- dev-zero@jabber.org has become available [07:53:07] --- stephen.quinney has become available [07:53:27] --- stephen.quinney has left [07:58:42] --- deason has become available [10:31:56] --- dev-zero@jabber.org has left [12:19:20] --- Simon Wilkinson has left [13:06:09] --- dev-zero@jabber.org has become available [13:07:01] --- dev-zero@jabber.org has left [13:07:03] --- dev-zero@jabber.org has become available [13:07:36] --- dev-zero@jabber.org has left: offline [13:59:08] bnode_Register() has a very curious use of opr_queue_Scan: for (opr_queue_Scan(&allTypes, cursor), tt = NULL) { That's the only place in the tree where the opr_queue_Scan() macro is added to as part of the loop body, which seems quite fishy. I'm looking at the history now, but do others agree that it seems worth rewriting the code to not use this construct? [14:05:51] --- Simon Wilkinson has become available [14:29:18] --- Ken Dreyer has become available [14:44:35] Am I missing something here? It looks like bnode_Delete() calls bnode_Release() which recurses into bnode_Delete() and I don't see how the chain would stop. [14:56:29] It only recurses if BNODE_DELETE is set [14:58:12] But bnode_Delete ... only sets BNODE_DELETE if the refcount is positive, gotcha. [14:58:48] Tangentially related, I think that the refcounting situation is pretty bogus right now, but I'm not sure that I want to fix it in the same patchset that converts to pthreads. [15:00:09] Like, FindFOO should return a held reference that the caller must release, etc. [15:03:22] Pretty much all of the reference counting in OpenAFS is bogus when you dig into it [15:04:51] "Oh, goody." [15:05:13] In any case, this is just the bosserver, it's hardly loaded and even my current scheme is probably way overkill. [15:09:23] I take it you've seen Chas's pthreaded bos server work? [15:09:33] I ... apparently not. [15:10:42] http://gerrit.openafs.org/#dashboard,1000049 [15:11:35] Looks like it stalled around Christmas [15:11:40] Indeed. [15:12:03] I'm not sure there's reason to keep a LWP version around; I think we could just wholesale convert to pthreads. [15:12:19] That's certainly my view [15:12:53] But Chas is adding a tbozo. [15:13:11] --- mvitale has left [15:13:36] I guess I can fetch the whole patch chain and look at it a bit locally, to see whether I want to adopt it or continue with what I've started. [15:13:42] Yeah, I think I moaned about that in an earlier review comment [15:14:06] The big issue with bozo is getting softsig working properly. What's in OpenAFS currently doesn't - you can intermittently lose segfaults [15:15:00] That sounds poor. I have not encountered softsig yet, so I may be missing some background. [15:15:17] pthreads and signals, it's a whole world of pain [15:15:30] Yup. [15:17:13] 6947 was my attempt at fixing it for OpenAFS, but it doesn't work with LinuxThreads [15:19:15] That just gives us an excuse to drop support for LINUX24, right? ;) [15:26:46] --- shadow@gmail.com/barnowl5ABA57FD has left [15:27:14] --- shadow@gmail.com/barnowl5ABA57FD has become available [15:48:28] Chas' thing kind of feels overkill to me at this point, but maybe I just haven't gotten to the hart parts yet. [16:16:00] --- deason has left [17:16:47] --- Simon Wilkinson has left [17:27:17] --- Ken Dreyer has left [17:27:52] --- Ken Dreyer has become available [19:25:51] --- cudave has left: Disconnected [19:26:37] --- cudave has become available [22:37:22] --- dev-zero@jabber.org has become available [23:01:04] --- shadow@gmail.com/barnowl5ABA57FD has left [23:01:26] --- shadow@gmail.com/barnowl5ABA57FD has become available [23:02:51] --- dev-zero@jabber.org has left [23:34:09] --- Simon Wilkinson has become available [23:34:28] --- Simon Wilkinson has left [23:35:16] --- Simon Wilkinson has become available