merged
thanks
-steve
On Wed, 2009-06-17 at 11:25 +0200, Jérôme Flesch wrote:
> This patch adds a call to madvise(..., MADV_NOSYNC) after each call to
> mmap() . MADV_NOSYNC is an option specific to BSD systems specifying to
> the system to not write data on the disk unless it's really needed.
Committed revision 2252.
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> logsys: remove leftover files from running tests
___
Openais mailing list
Openais@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/open
Committed revision 2251.
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 77e8bd7... 90efdf8... Mexec/logsys.c
> exec/logsys.c | 14 +++---
> 1 files changed, 7 insertions(+), 7 deletions(-)
>
> diff --git a
Committed revision 2250.
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 4f97dfc... 77e8bd7... Mexec/logsys.c
> :100644 100644 9901788... 3f4e425... Mexec/main.c
> :100644 100644 b1f19c4... e7ead32... M
Committed revision 2249.
and add () as requested. If I missed any, we can just add them.
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> rec_ident should contain 3 info: log level, subsystem id and
> message type.
>
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 ea53cac
Committed as part of revision 2249.
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 5629879... 7a7d78b... Minclude/corosync/engine/logsys.h
> include/corosync/engine/logsys.h |2 +-
> 1 files changed, 1 insertions(
Committed revision 2248.
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> most of the values in logsys.h are very useful for non logsys library
> API users.
>
> Allow to import them without sucking the whole lib.
>
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 59c8a3e.
Committed revision 2247.
On Wed, 2009-06-17 at 08:42 -0700, Steven Dake wrote:
> looks good
>
> On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> > Signed-off-by: Fabio M. Di Nitto
> > ---
> > :100644 100644 e39837a... e19836e... M man/logsys_overview.8
> > man/logsys_overview
Committed revision 2246.
On Wed, 2009-06-17 at 08:42 -0700, Steven Dake wrote:
> Looks good
>
> On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> > Signed-off-by: Fabio M. Di Nitto
> > ---
> > :100644 100644 682d7dd... 4f97dfc... M exec/logsys.c
> > :100644 100644 2e9a337... 99
Committed revision 2245
On Wed, 2009-06-17 at 08:39 -0700, Steven Dake wrote:
> looks good
>
> On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> > LOGSYS_LEVEL_SECURITY is specific to corosync/openais and it
> > is used only in the totem configuration.
> >
> > Drop the special case f
+1
Fabio
On Wed, 2009-06-17 at 16:31 -0700, Steven Dake wrote:
> I have given some thought to versioning and what it means for us. This
> is what I suggest:
>
> X.Y.Z
>
> X indicates a major release. X is maintained as a branch in the branch
> directory of the repository. During a major rele
third party service engines may want to shutdown the system cleanly.
Currently they may exit which does not release semaphores or shared
memory. Also other parts of openais may shutdown which results in these
same leaks.
This patch formalizes the shutdown method to call the api
openais_shutdown (
I have given some thought to versioning and what it means for us. This
is what I suggest:
X.Y.Z
X indicates a major release. X is maintained as a branch in the branch
directory of the repository. During a major release no changes to
existing APIs may be done including coroapi.h and the header
On Wed, 2009-06-17 at 14:04 -0700, Steven Dake wrote:
> On Wed, 2009-06-17 at 22:56 +0200, Fabio M. Di Nitto wrote:
> > On Wed, 2009-06-17 at 08:55 -0700, Steven Dake wrote:
> > > looks good
> > >
> > > Did you try benchmarks prior/before? Curious if faster/slower.
> >
> > Yes I did and the pac
On Wed, 2009-06-17 at 22:56 +0200, Fabio M. Di Nitto wrote:
> On Wed, 2009-06-17 at 08:55 -0700, Steven Dake wrote:
> > looks good
> >
> > Did you try benchmarks prior/before? Curious if faster/slower.
>
> Yes I did and the packing/unpacking is slower but barely noticeable.
>
> before it was a
On Wed, 2009-06-17 at 08:55 -0700, Steven Dake wrote:
> looks good
>
> Did you try benchmarks prior/before? Curious if faster/slower.
Yes I did and the packing/unpacking is slower but barely noticeable.
before it was approx: 670 ops/sec
after it was approx: 668 ops/sec
Fabio
___
On Wed, 2009-06-17 at 08:48 -0700, Steven Dake wrote:
> All defines operations should always be in () for example:
> #define LOGSYS_LEVEL_END 3 should be
> #define LOGSYS_LEVEL_END (3)
>
> While in this particular case it isn't necessary it is in any time an
> operation is done and is good practic
On Wed, 2009-06-17 at 08:49 -0700, Steven Dake wrote:
> I noticed in this patch LOGSYS_RECID_MAX - 1 without parens. Must have
> missed seeing that in previous patch.
>
> This patch has same issue as 5/9 though but looks fine after that fix
Yeps.. gather that.. i'll just pass it all over.
Fab
thanks that was annoying
pls merge
On Wed, 2009-06-17 at 11:12 +0200, Jim Meyering wrote:
> Without this patch, running automake provokes this:
>
> exec/Makefile.am:66: SHARED_LIBS:%.so.$(SONAME: non-POSIX variable name
> exec/Makefile.am:67: SHARED_LIBS:%.so.$(SONAME: non-POSIX variable nam
thanks looks good
On Wed, 2009-06-17 at 12:41 +0200, Jim Meyering wrote:
> Jan Friesse wrote:
>
> > Jim,
> > OpenAIS need some love too.
> >
> > Jim Meyering wrote:
> >> Without this patch, running automake provokes this:
> >>
> >> exec/Makefile.am:66: SHARED_LIBS:%.so.$(SONAME: non-POSIX varia
thanks looks good
On Wed, 2009-06-17 at 11:14 +0200, Jim Meyering wrote:
> Jan Friesse mentioned that the existing use of "which" in autogen.sh
> was causing trouble on a fresh F11 install (it's not installed).
>
> I propose to replace that file with a one-liner:
>
> From d0c4fa62fadebaa801997a6
What can I say other then pure brilliance!
Regards
-steve
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Hi Steven,
>
> as we discussed on IRC:
>
> - removal of tags
> - packing of rec_ident
> - misc cleanup around (whitespaces, defines, Makefiles)
>
> I'l
I noticed in this patch LOGSYS_RECID_MAX - 1 without parens. Must have
missed seeing that in previous patch.
This patch has same issue as 5/9 though but looks fine after that fix
regards
-steve
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Signed-off-by: Fabio M. Di Nitto
> -
looks good
Did you try benchmarks prior/before? Curious if faster/slower.
regards
-steve
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 4f97dfc... 77e8bd7... Mexec/logsys.c
> :100644 100644 9901788... 3f4e425... M
good for merge
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 a0ebbdd... d426fa6... Mtest/Makefile.am
> test/Makefile.am |3 +++
> 1 files changed, 3 insertions(+), 0 deletions(-)
>
> diff --git a/test/Makefile.a
Looks good
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 77e8bd7... 90efdf8... Mexec/logsys.c
> exec/logsys.c | 14 +++---
> 1 files changed, 7 insertions(+), 7 deletions(-)
>
> diff --git a/exec/logsys.c
All defines operations should always be in () for example:
#define LOGSYS_LEVEL_END 3 should be
#define LOGSYS_LEVEL_END (3)
While in this particular case it isn't necessary it is in any time an
operation is done and is good practice to follow.
The bitmask generation also looks complicated (I pre
looks good
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> LOGSYS_LEVEL_SECURITY is specific to corosync/openais and it
> is used only in the totem configuration.
>
> Drop the special case from logsys that's meant to be a generic
> logging library and specify the correct equivalent
looks good
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> most of the values in logsys.h are very useful for non logsys library
> API users.
>
> Allow to import them without sucking the whole lib.
>
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 59c8a3e... b1f19c4... M
Looks good
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 682d7dd... 4f97dfc... Mexec/logsys.c
> :100644 100644 2e9a337... 9901788... Mexec/main.c
> :100644 100644 5db1fba... 847194a... Mexec/mainconfig
looks good
On Wed, 2009-06-17 at 12:53 +0200, Fabio M. Di Nitto wrote:
> Signed-off-by: Fabio M. Di Nitto
> ---
> :100644 100644 e39837a... e19836e... Mman/logsys_overview.8
> man/logsys_overview.8 | 32 +++-
> 1 files changed, 7 insertions(+), 25 deletions(
Signed-off-by: Fabio M. Di Nitto
---
:100644 100644 a0ebbdd... d426fa6... M test/Makefile.am
test/Makefile.am |3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/test/Makefile.am b/test/Makefile.am
index a0ebbdd..d426fa6 100644
--- a/test/Makefile.am
+++ b/test/Makefile.am
Signed-off-by: Fabio M. Di Nitto
---
:100644 100644 682d7dd... 4f97dfc... M exec/logsys.c
:100644 100644 2e9a337... 9901788... M exec/main.c
:100644 100644 5db1fba... 847194a... M exec/mainconfig.c
:100644 100644 6705311... 59c8a3e... M exec/totemnet.c
:100644 100644 2ea45ff... bb34451... M e
Signed-off-by: Fabio M. Di Nitto
---
:100644 100644 4f97dfc... 77e8bd7... M exec/logsys.c
:100644 100644 9901788... 3f4e425... M exec/main.c
:100644 100644 b1f19c4... e7ead32... M exec/totemnet.c
:100644 100644 7565805... 7d24ae1... M exec/totempg.c
:100644 100644 e0d19ca... c0eb2e7... M exec
most of the values in logsys.h are very useful for non logsys library
API users.
Allow to import them without sucking the whole lib.
Signed-off-by: Fabio M. Di Nitto
---
:100644 100644 59c8a3e... b1f19c4... M exec/totemnet.c
:100644 100644 bb34451... 7565805... M exec/totempg.c
:100644 100644
Signed-off-by: Fabio M. Di Nitto
---
:100644 100644 5629879... 7a7d78b... M include/corosync/engine/logsys.h
include/corosync/engine/logsys.h |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/include/corosync/engine/logsys.h b/include/corosync/engine/logsys.h
index 562987
Signed-off-by: Fabio M. Di Nitto
---
:100644 100644 77e8bd7... 90efdf8... M exec/logsys.c
exec/logsys.c | 14 +++---
1 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/exec/logsys.c b/exec/logsys.c
index 77e8bd7..90efdf8 100644
--- a/exec/logsys.c
+++ b/exec/logsys.c
@@ -18
rec_ident should contain 3 info: log level, subsystem id and
message type.
Signed-off-by: Fabio M. Di Nitto
---
:100644 100644 ea53cac... 5629879... M include/corosync/engine/logsys.h
include/corosync/engine/logsys.h | 71 --
1 files changed, 60 insertions(
Signed-off-by: Fabio M. Di Nitto
---
:100644 100644 e39837a... e19836e... M man/logsys_overview.8
man/logsys_overview.8 | 32 +++-
1 files changed, 7 insertions(+), 25 deletions(-)
diff --git a/man/logsys_overview.8 b/man/logsys_overview.8
index e39837a..e19836e 10
LOGSYS_LEVEL_SECURITY is specific to corosync/openais and it
is used only in the totem configuration.
Drop the special case from logsys that's meant to be a generic
logging library and specify the correct equivalent for totem config.
Signed-off-by: Fabio M. Di Nitto
---
:100644 100644 7b62270...
Hi Steven,
as we discussed on IRC:
- removal of tags
- packing of rec_ident
- misc cleanup around (whitespaces, defines, Makefiles)
I'll take care to port openais/cman as soon as this patch set (or
reviewed/acked equivalent) will hit the tree.
The last beast that n
Jan Friesse wrote:
> Jim,
> OpenAIS need some love too.
>
> Jim Meyering wrote:
>> Without this patch, running automake provokes this:
>>
>> exec/Makefile.am:66: SHARED_LIBS:%.so.$(SONAME: non-POSIX variable name
>> exec/Makefile.am:67: SHARED_LIBS:%.so.$(SONAME: non-POSIX variable name
;-)
H
Jim,
OpenAIS need some love too.
Jim Meyering wrote:
> Jan Friesse mentioned that the existing use of "which" in autogen.sh
> was causing trouble on a fresh F11 install (it's not installed).
>
> I propose to replace that file with a one-liner:
>
> From d0c4fa62fadebaa801997a619e7b2bc2cd8c14d9 Mo
Jim,
OpenAIS need some love too.
Jim Meyering wrote:
> Without this patch, running automake provokes this:
>
> exec/Makefile.am:66: SHARED_LIBS:%.so.$(SONAME: non-POSIX variable name
> exec/Makefile.am:67: SHARED_LIBS:%.so.$(SONAME: non-POSIX variable name
> exec/Makefile.am:77: `%'-style p
This patch adds a call to madvise(..., MADV_NOSYNC) after each call to
mmap() . MADV_NOSYNC is an option specific to BSD systems specifying to
the system to not write data on the disk unless it's really needed.
It doesn't seem really required to make Corosync works on FreeBSD
(performances loo
Jan Friesse mentioned that the existing use of "which" in autogen.sh
was causing trouble on a fresh F11 install (it's not installed).
I propose to replace that file with a one-liner:
>From d0c4fa62fadebaa801997a619e7b2bc2cd8c14d9 Mon Sep 17 00:00:00 2001
From: Jim Meyering
Date: Wed, 17 Jun 2009
Without this patch, running automake provokes this:
exec/Makefile.am:66: SHARED_LIBS:%.so.$(SONAME: non-POSIX variable name
exec/Makefile.am:67: SHARED_LIBS:%.so.$(SONAME: non-POSIX variable name
exec/Makefile.am:77: `%'-style pattern rules are a GNU make extension
exec/Makefile.am:97: `%'
47 matches
Mail list logo