opcontrol relies on 'objdump' which is part of binutils, but not tracked
as dependancy
** Changed in: oprofile (Ubuntu)
Status: New => Confirmed
--
Oprofile fails to load vmlinux image if binutils is not installed
https://bugs.launchpad.net/bugs/251290
You received this bug notification b
It fixed at least for my VM tests (though they were mostly same as initial
testcase).
Thanks Steve.
--
locks on unlinked files leak memory in apparmor
https://bugs.launchpad.net/bugs/329489
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu
and, of course:
Linux srv47 2.6.24-19-server #1 SMP Wed Aug 20 18:43:06 UTC 2008 x86_64
GNU/Linux
hardy with all updates
--
frequent fcntl()s leak memory in apparmor
https://bugs.launchpad.net/bugs/329489
You received this bug notification because you are a member of Ubuntu
Bugs, which is subsc
** Attachment added: "graph of leaked memory"
http://launchpadlibrarian.net/22615274/graph.php.gif
--
frequent fcntl()s leak memory in apparmor
https://bugs.launchpad.net/bugs/329489
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
--
Public bug reported:
after switching locking in apache/PHP/APC from shared-memory based to
fcntl() we observed that apparmor (running in audit mode) leaks memory
inside kernel like crazy:
MemTotal: 2062736 kB
MemFree: 16160 kB
Buffers: 496 kB
Cached: 7120 kB
SwapC
Testcase immediately leaks memory, as well as kernel message buffer gets
lots of smashed buffer crap:
http://p.defau.lt/?HxHScO_HJyFcgLVE2Fp8qw
--
frequent fcntl()s leak memory in apparmor
https://bugs.launchpad.net/bugs/329489
You received this bug notification because you are a member of Ubunt
testcase:
#include
#include
main() {
int fd = open("/tmp/.lockfile", O_RDWR|O_CREAT);
unlink("/tmp/.lockfile");
fork();
fork();
fork();
fork();
while(1) {
struct flock lock;
lock.l_type=F_WRLCK;
lock.l_start=0;
lock.l_whence=SEEK_END;
lock.l_len=0;
One more detail, this is excerpt from /proc/slabinfo:
kmalloc-1024 806816 102441 : tunables000 :
slabdata204204 0
kmalloc-512 51752851281 : tunables000 :
slabdata 66 66 0
kmalloc-256 907300 907
Steve,
it does same in all modes:
aa-complain
aa-audit
aa-enforce
--
frequent fcntl()s leak memory in apparmor
https://bugs.launchpad.net/bugs/329489
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
--
ubuntu-bugs mailing list
ubuntu-b
** Summary changed:
- frequent fcntl()s leak memory in apparmor
+ fcntl() locks on unlinked files leak memory in apparmor
--
operations on unlinked files leak memory in apparmor
https://bugs.launchpad.net/bugs/329489
You received this bug notification because you are a member of Ubuntu
Bugs, whi
OK, sorry folks for having too difficult start, I ended up with
extremely simple testcase, leaking 100MB/s:
#include
#include
main() {
int fd = open("/tmp/.lockfile", O_RDWR | O_CREAT);
unlink("/tmp/.lockfile");
while (1) write(fd,"a",1);
}
I'm changing the issue summary yet again.
How did you make the loop run only 1000 times?!!!? :)
Thanks for 'slabtop' hint - nice tool for tourists from userland.
Now, let me put my MySQL hat on:
- MySQL ships by default with apparmor profile
- MySQL uses unlinked files quite a bit (implicit temptables, sort buffer
overflows, etc)
So, t
damn, after all, it may be just fcntl() - I can't reproduce the write() case
anymore, and I wonder if I ever could.
teaches me that C is a compiled language!
** Summary changed:
- operations on unlinked files leak memory in apparmor
+ locks on unlinked files leak memory in apparmor
--
locks o
resetting back to new, maybe I failed something
** Changed in: mysql-dfsg-5.0 (Ubuntu)
Status: Confirmed => New
--
mysql user has home directory writable by mysqld
https://bugs.launchpad.net/bugs/293258
You received this bug notification because you are a member of Ubuntu
Bugs, which is a
intrepid, seems to be same in lenny, hardy, etc.
--
mysql user has home directory writable by mysqld
https://bugs.launchpad.net/bugs/293258
You received this bug notification because you are a member of Ubuntu
Bugs, which is a direct subscriber.
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubu
Public bug reported:
Binary package hint: findutils
Various troubles can lead to mount entries not being written in /etc/mtab
(say, if single NFS mount fails, when others not in 'mount -a -t nfs').
/etc/mtab information is used for find to determine remote file systems (to
skip them), but if i
Public bug reported:
Imagemagick signal handlers, if compiled with pthread support, may enter
recursive mutex locks.
This may lead to deadlock in single-threaded programs such as 'convert'.
Generally, either error checking mutexes have to be used (and crash
horribly), or signal handlers simplifi
Excuse me, what do you want to have clarified?
Doesn't stack trace show you that there's a deadlock condition inside
ImageMagick?
My fix is having it compiled without pthreads support.
--
imagemagick tools should be built without threading (or recursive mutexes fixed)
https://bugs.launchpad.ne
Dear sirs, jaunty (and intrepid, it seems) does not have linux-image-debug
packages, and ddebs.ubuntu.com don't carry anything pre-karmic at the moment.
As systemtap, oprofile and plethora of other tools depend on this, this is
quite a show stopper.
--
Missing linux-image-debug packages and met
http://bugs.mysql.com/bug.php?id=35477
** Bug watch added: MySQL Bug System #35477
http://bugs.mysql.com/bug.php?id=35477
--
mysql-server-5.0 on 08.04 LTS leaks filedescriptors
https://bugs.launchpad.net/bugs/644288
You received this bug notification because you are a member of Ubuntu
Bugs, w
20 matches
Mail list logo
