Bugzilla – Bug 720882
fuser creates random output
Last modified: 2011-10-05 10:00:07 UTC
This is not a problem on x86_64, but on 32bit fuser is completely broken Either fuser -v /dev/tty1 (just an example, it works for all devices) outputs all processes or none. This very much looks like an uninitialized variable and valgrind blames timeout.c:182 Backing out the timeout patch fixes the problem for now.
This is an autogenerated message for OBS integration: This bug (720882) was mentioned in https://build.opensuse.org/request/show/85293 Factory / psmisc
What exactly does valgrind blame ... as the vraible handle is initialitzed and the pipes array should be also initialized after start() has been executed (which in fact is done as constructor or due restart of the forked subprocess)
I have no time to debug this for you, sorry.
On which system and with which gcc do you see this? As I do *not* see this behaviour ...
I've seen it on _all_ i586 systems running factory. fuser /dev/snd/* returned all processes.
The message of valgrind is nonsense ... the real reason is the internal type of structure stat used by glibc on i386 which in fact is struct stat64 !
psmisc/psmisc-22.14> sudo src/fuser -v /dev/tty1 USER PID ACCESS COMMAND /dev/tty1: root 17685 F.... mingetty psmisc/psmisc-22.14> uname -i i386 that's it
Fixed and submitted
This is an autogenerated message for OBS integration: This bug (720882) was mentioned in https://build.opensuse.org/request/show/86648 Factory / psmisc