Re: [Openais] [PATCH] corosync-keygen.c: diagnose a few more failures

Steven Dake wrote: looks good On Fri, 2009-04-10 at 09:03 +0200, Jim Meyering wrote: Date: Fri, 10 Apr 2009 09:02:36 +0200 Subject: [PATCH] corosync-keygen.c: diagnose a few more failures * tools/corosync-keygen.c (main): Diagnose short reads, failed mkdir and fchmod; detect write

Re: [Openais] two small patches

Steven Dake wrote: looks good regards -steve On Fri, 2009-04-10 at 18:58 +0200, Jim Meyering wrote: From 01aaca1325f4dedd3f0be0bab7fd753a83f3e0b1 Mon Sep 17 00:00:00 2001 From: Jim Meyering meyer...@redhat.com Date: Fri, 10 Apr 2009 18:00:01 +0200 Subject: [PATCH 1/2] change a few int

Re: [Openais] totemrrp.h, totemsrp.h, totemnet.h: const+size_t

Steven Dake wrote: Try using this patch with your patchset. It makes a copy of the constant void *msg into a new merge_detect struct. If it looks good go ahead and merge in with your changes on wednesday. That looks fine and confirms that msg should remain const. I've just committed your

[Openais] [RFC] corosync/trunk: logsys v3 implementation take 3

Hi guys, changes compared to take2: - move SUBSYS_MAX and SUBSYS_MAXLEN in logsys.h. The former is required to assert within DECLARE_* and the latter is good to have visibile as it's part of the public API limit for subsystem name. - make DECLARE_* macros assert on error. - fix error handling

[Openais] [PATCH] totemsrp.c: more const propagation

This started by my removing useless (and risky) casts like this: - struct orf_token *token = (struct orf_token *)msg; + const struct orf_token *token = msg; Those changes propagated and exposed one problem that could not be resolved with simple const addition, so I mentioned it in

[Openais] Patch for WhiteTank to prevent shared memory leak

Hi, Here is a patch to fix a shared memory leak in WhiteTank when used on FreeBSD systems. The problem was that the connection ref counter wasn't decremented as it should when a client disconnected (see patch) and so the shared memory segment was never destroyed by

[Openais] openais unicast?

I am using qpid to handle message queuing and it uses openais to handle clustering (via libcpg). I was wondering if that clustering could be done via unicast? Thanks, Adam ___ Openais mailing list Openais@lists.linux-foundation.org

[Openais] [PATCH 06/15] don't shadow file-scoped global, handle

From: Jim Meyering meyer...@redhat.com * test/testquorum.c: Rename: s/handle/g_handle/. --- test/testquorum.c | 10 +- 1 files changed, 5 insertions(+), 5 deletions(-) diff --git a/test/testquorum.c b/test/testquorum.c index 9139f3f..74b97ec 100644 --- a/test/testquorum.c +++

[Openais] [PATCH 08/15] coroipcs.c: don't shadow functions read and write

From: Jim Meyering meyer...@redhat.com * exec/coroipcs.c (req_setup_recv): rename locals: s/read/n_read/; s/write/n_write. Also declare param to be const. --- exec/coroipcs.c | 18 +- 1 files changed, 9 insertions(+), 9 deletions(-) diff --git a/exec/coroipcs.c

[Openais] [PATCH 15/15] cpg.c: rename file-scoped global to avoid shadowing warnings

From: Jim Meyering meyer...@redhat.com * services/cpg.c: s/req_exec_cpg_downlist/g_req_exec_cpg_downlist/ --- services/cpg.c | 18 +- 1 files changed, 9 insertions(+), 9 deletions(-) diff --git a/services/cpg.c b/services/cpg.c index f60427a..2be795e 100644 ---

[Openais] [PATCH 05/15] wthread.c: Rename file-scoped type to avoid shadows.

From: Jim Meyering meyer...@redhat.com * exec/wthread.c (struct worker_thread_t): Rename from struct worker_thread. (start_worker_thread): Rename from function worker_thread. * exec/wthread.h (struct worker_thread_group) [threads]: Update member type: s/struct worker_thread/struct

[Openais] [PATCH 09/15] configure.ac: enable gcc's -Wshadow warning

From: Jim Meyering meyer...@redhat.com * configure.ac (WARNLIST): Add -Wshadow to the list. --- configure.ac |1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/configure.ac b/configure.ac index 4b034f8..3d5b15e 100644 --- a/configure.ac +++ b/configure.ac @@ -262,6 +262,7

[Openais] [PATCH 10/15] corosync-fplay.c: avoid shadowin: s/index/idx/

From: Jim Meyering meyer...@redhat.com * tools/corosync-fplay.c (printer_totempg_mcast_fits): --- tools/corosync-fplay.c |4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/corosync-fplay.c b/tools/corosync-fplay.c index 9eed9f7..2c53383 100644 ---

[Openais] [PATCH 03/15] list.h: avoid shadowing warning

From: Jim Meyering meyer...@redhat.com * include/corosync/list.h: Don't use remove as param name. --- include/corosync/list.h | 10 +- 1 files changed, 5 insertions(+), 5 deletions(-) diff --git a/include/corosync/list.h b/include/corosync/list.h index e9df38b..1790a69 100644 ---

[Openais] [PATCH 11/15] testvotequorum1.c: don't shadow file-scoped global, handle

From: Jim Meyering meyer...@redhat.com * test/testvotequorum1.c (main): Rename: s/handle/g_handle/. --- test/testvotequorum1.c | 16 1 files changed, 8 insertions(+), 8 deletions(-) diff --git a/test/testvotequorum1.c b/test/testvotequorum1.c index b8f0648..5050ba7 100644 ---

[Openais] [PATCH 02/15] coroipcc.c: avoid shadowing warning

From: Jim Meyering meyer...@redhat.com * lib/coroipcc.c: Don't use read as param name. --- lib/coroipcc.c | 12 ++-- 1 files changed, 6 insertions(+), 6 deletions(-) diff --git a/lib/coroipcc.c b/lib/coroipcc.c index 038ded8..cc3a540 100644 --- a/lib/coroipcc.c +++ b/lib/coroipcc.c @@

[Openais] [PATCH 01/15] totemsrp.c: more const propagation

From: Jim Meyering meyer...@redhat.com * exec/totemsrp.c (message_handler_memb_commit_token): Don't cast away const on memb_commit_token. This exposed a const violation. Fix that with minor rearrangement and an added memcpy. --- exec/totemsrp.c | 64

[Openais] [PATCH 04/15] don't shadow the global type name, timer_handle

From: Jim Meyering meyer...@vv.meyering.net.meyering.net * exec/coropoll.c (poll_timer_delete): Rename locals and/or params. * exec/timer.c (corosync_timer_delete): Likewise. (corosync_timer_expire_time_get): Likewise. * exec/tlist.h (timerlist_del, timerlist_expire_time): Likewise.

Re: [Openais] [PATCH 01/15] totemsrp.c: more const propagation

Jim Meyering wrote: From: Jim Meyering meyer...@redhat.com * exec/totemsrp.c (message_handler_memb_commit_token): Don't cast away const on memb_commit_token. This exposed a const violation. Fix that with minor rearrangement and an added memcpy. --- exec/totemsrp.c | 64

Re: [Openais] [PATCH 14/15] vsf_ykd.c: Rename param to avoid shadowing global ydk_state

this patchset looks fine regards -steve On Wed, 2009-04-15 at 18:07 +0200, Jim Meyering wrote: From: Jim Meyering meyer...@redhat.com * exec/vsf_ykd.c (ykd_state_endian_convert): Rename param: s/ydk_state/state/ --- exec/vsf_ykd.c | 18 +- 1 files changed, 9

Re: [Openais] [PATCH 14/15] vsf_ykd.c: Rename param to avoid shadowing global ydk_state

Steven Dake wrote: this patchset looks fine Thanks for the review. Committing all 15 of them now. ___ Openais mailing list Openais@lists.linux-foundation.org https://lists.linux-foundation.org/mailman/listinfo/openais

[Openais] delayed shutdown

If I run 'cman_tool leave' on four nodes in parallel, node1 will leave right away, but the other three nodes don't leave until the token timeout expires for node1 causing a confchg for it, after which the other three all leave right away. This has only been annoying me recently, so I think it

Re: [Openais] Patch for WhiteTank to prevent shared memory leak

I don't think this is a proper fix as it would break the proper refcounting on Linux systems. I have tested conn_info_destroy pretty extensively on Linux and it works well. Could you explain your test case? This may be a bsd-ism that we need to work around. Regards -steve On Wed, 2009-04-15 at

Re: [Openais] delayed shutdown

On Wed, 2009-04-15 at 14:35 -0500, David Teigland wrote: If I run 'cman_tool leave' on four nodes in parallel, node1 will leave right away, but the other three nodes don't leave until the token timeout expires for node1 causing a confchg for it, after which the other three all leave right