rpms/rsyslog
5
0
Fork 0
Code Issues Pull Requests Releases Activity

Add patches to resolve rhbz971471

Browse Source
This commit is contained in:
Tomas Heinrich 2013-06-11 10:27:42 +02:00
parent e32e6e538e
commit 5b94587fc1
4 changed files with 115 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
rsyslog-7.4.0-imjournal-segv.rhbz971471.patch Normal file
View File

@ -0,0 +1,35 @@
From 0082647fa5e6849b83ab4c6b9ab2e8803245db14 Mon Sep 17 00:00:00 2001
From: Tomas Heinrich <theinric@redhat.com>
Date: Fri, 7 Jun 2013 01:15:10 +0200
Subject: [PATCH] bugfix: be more tolerant to malformed journal fields
This prevents a segfault when a malformed journal entry field doesn't
contain an equal sign. Should not ever happen but was actually
triggered by a real bug in systemd journal.
---
plugins/imjournal/imjournal.c | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/plugins/imjournal/imjournal.c b/plugins/imjournal/imjournal.c
index ae29154..cce45b9 100755
--- a/plugins/imjournal/imjournal.c
+++ b/plugins/imjournal/imjournal.c
@@ -244,7 +244,14 @@ readjournal() {
SD_JOURNAL_FOREACH_DATA(j, get, l) {
/* locate equal sign, this is always present */
equal_sign = memchr(get, '=', l);
- assert (equal_sign != NULL);
+
+ /* ... but we know better than to trust the specs */
+ if (equal_sign == NULL) {
+ errmsg.LogError(0, RS_RET_ERR,"SD_JOURNAL_FOREACH_DATA()"
+ " returned a malformed field (has no '='): '%s'",
+ get);
+ continue; /* skip the entry */
+ }
/* get length of journal data prefix */
prefixlen = ((char *)equal_sign - (char *)get);
--
1.7.10.4

38
rsyslog-7.4.0-ratelimiter-loop.rhbz971471.patch Normal file
View File

@ -0,0 +1,38 @@
From f11a873dc4e258c346765af9d5d23a1180493ee8 Mon Sep 17 00:00:00 2001
From: Tomas Heinrich <theinric@redhat.com>
Date: Sat, 8 Jun 2013 23:27:48 +0200
Subject: [PATCH 2/2] bugfix: prevent an endless loop in the ratelimiter
If messages are being dropped because of ratelimiting, an internal
message is generated to inform about this fact. This should happen
only uppon the firs occurance but the counter that tracks the number
of dropped messages was incremented only after sending the message. If
the message itself gets ratelimited, an endless loop spins out of
control. Thanks to Jerry James for notifying about this.
---
runtime/ratelimit.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/runtime/ratelimit.c b/runtime/ratelimit.c
index d83da2d..ec24855 100644
--- a/runtime/ratelimit.c
+++ b/runtime/ratelimit.c
@@ -167,13 +167,13 @@ withinRatelimit(ratelimit_t *ratelimit, time_t tt)
ratelimit->done++;
ret = 1;
} else {
- if(ratelimit->missed == 0) {
+ ratelimit->missed++;
+ if(ratelimit->missed == 1) {
snprintf((char*)msgbuf, sizeof(msgbuf),
"%s: begin to drop messages due to rate-limiting",
ratelimit->name);
logmsgInternal(RS_RET_RATE_LIMITED, LOG_SYSLOG|LOG_INFO, msgbuf, 0);
}
- ratelimit->missed++;
ret = 0;
}
--
1.7.10.4

29
rsyslog-7.4.0-ratelimiter-loop2.rhbz971471.patch Normal file
View File

@ -0,0 +1,29 @@
From 82db8e6fbba89bf486cc7e642e4f8daaa43852e8 Mon Sep 17 00:00:00 2001
From: Tomas Heinrich <theinric@redhat.com>
Date: Mon, 10 Jun 2013 23:09:38 +0200
Subject: [PATCH] bugfix: prevent another endless loop in the ratelimiter
The message that reports how many messages were lost due to
ratelimiting was sent before reseting the state that led to it. If it
itself got ratelimited, this could lead to an endless loop.
---
runtime/ratelimit.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/runtime/ratelimit.c b/runtime/ratelimit.c
index ec24855..443a5fa 100644
--- a/runtime/ratelimit.c
+++ b/runtime/ratelimit.c
@@ -128,8 +128,8 @@ tellLostCnt(ratelimit_t *ratelimit)
snprintf((char*)msgbuf, sizeof(msgbuf),
"%s: %u messages lost due to rate-limiting",
ratelimit->name, ratelimit->missed);
- logmsgInternal(RS_RET_RATE_LIMITED, LOG_SYSLOG|LOG_INFO, msgbuf, 0);
ratelimit->missed = 0;
+ logmsgInternal(RS_RET_RATE_LIMITED, LOG_SYSLOG|LOG_INFO, msgbuf, 0);
}
}
--
1.7.10.4

13
rsyslog.spec
View File

@ -26,7 +26,13 @@ Patch0: rsyslog-7.2.2-systemd.patch
Patch1: rsyslog-7.2.2-manpage-dbg-mode.patch
# prevent modification of trusted properties (proposed upstream)
Patch2: rsyslog-7.2.1-msg_c_nonoverwrite_merge.patch
# merged upstream
Patch3: rsyslog-7.3.15-imuxsock-warning.patch
# merged upstream
Patch4: rsyslog-7.4.0-imjournal-segv.rhbz971471.patch
# merged upstream
Patch5: rsyslog-7.4.0-ratelimiter-loop.rhbz971471.patch
Patch6: rsyslog-7.4.0-ratelimiter-loop2.rhbz971471.patch
BuildRequires: bison
BuildRequires: flex
@ -244,6 +250,9 @@ of source ports.
%patch1 -p1
%patch2 -p1
%patch3 -p1
%patch4 -p1
%patch5 -p1
%patch6 -p1
%build
%ifarch sparc64
@ -469,6 +478,10 @@ done
- rebase to 7.4.0
- drop autoconf automake libtool from BuildRequires
- depends on systemd >= 201 because of the sd_journal_get_events() api
- add a patch to prevent a segfault in imjournal caused by a bug in
systemd journal
- add a patch to prevent an endless loop in the ratelimiter
- add a patch to prevent another endless loop in the ratelimiter
* Tue Jun 04 2013 Tomas Heinrich <theinric@redhat.com> 7.3.15-1.20130604git6e72fa6
- rebase to an upstream snapshot, effectively version 7.3.15