View previous topic :: View next topic |
Author |
Message |
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Thu Sep 30, 2004 2:15 am Post subject: Segmentation Fault - VIM |
|
|
Hi all,
Since I use vim who knows how many times a day, this is pretty annoying...
I've tried just about everything, from emerge -e vim to emerging a previous vim ebuild. No luck. I can get into vim if I just type 'vim', otherwise.. 'vim <filename>' will always seg fault.
The last couple lines of an strace vim /etc/X11/xorg.conf :
Code: |
open("/dev/urandom", O_RDONLY) = 3
read(3, "pF}\333", 4) = 4
close(3) = 0
getrlimit(RLIMIT_STACK, {rlim_cur=8192*1024, rlim_max=RLIM_INFINITY}) = 0
brk(0) = 0x822d000
brk(0x824e000) = 0x824e000
brk(0) = 0x824e000
sigaltstack({ss_sp=0x822d050, ss_flags=0, ss_size=8192}, NULL) = 0
stat64("/usr/share/vim/vim63", {st_mode=S_IFDIR|0777, st_size=872, ...}) = 0
ioctl(1, SNDCTL_TMR_TIMEBASE or TCGETS, {B38400 opost isig icanon echo ...}) = 0getcwd("/home/bhogg", 1024) = 12
chdir("/home/bhogg") = 0
chdir("/home/bhogg") = 0
getcwd("/home/bhogg", 1025) = 12
chdir("/home/bhogg") = 0
sysinfo({uptime=28506, loads=[3968, 12416, 28672] totalram=518340608, freeram=187871232, sharedram=0, bufferram=12742656} totalswap=518180864, freeswap=458702848, procs=109}) = 0
getrlimit(RLIMIT_DATA, {rlim_cur=RLIM_INFINITY, rlim_max=RLIM_INFINITY}) = 0
stat64("/usr/share/vim", {st_mode=S_IFDIR|0777, st_size=72, ...}) = 0
--- SIGSEGV (Segmentation fault) @ 0 (0) ---
+++ killed by SIGSEGV +++
|
Any ideas?
Thanks,
Brian |
|
Back to top |
|
|
hardcore l33t
Joined: 01 Nov 2003 Posts: 626 Location: MSU, MI
|
Posted: Thu Sep 30, 2004 4:02 am Post subject: |
|
|
Try removing .viminfo in your home directory. _________________ Nothing can stop me now, cuz I just don't care. |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Thu Sep 30, 2004 9:14 pm Post subject: |
|
|
I deleted the file and I'm now getting segmentation fault at all times. The last lines of strace still look similar:
Code: |
sigaltstack({ss_sp=0x822d050, ss_flags=0, ss_size=8192}, NULL) = 0
stat64("/usr/share/vim/vim63", {st_mode=S_IFDIR|0777, st_size=872, ...}) = 0
ioctl(1, SNDCTL_TMR_TIMEBASE or TCGETS, {B38400 opost isig icanon echo ...}) = 0
getcwd("/home/bhogg", 1024) = 12
chdir("/home/bhogg") = 0
chdir("/home/bhogg") = 0
getcwd("/home/bhogg", 1025) = 12
chdir("/home/bhogg") = 0
sysinfo({uptime=5364, loads=[102080, 84352, 76576] totalram=518340608, freeram=6291456, sharedram=0, bufferram=22298624} totalswap=518180864, freeswap=487882752, procs=123}) = 0
getrlimit(RLIMIT_DATA, {rlim_cur=RLIM_INFINITY, rlim_max=RLIM_INFINITY}) = 0
stat64("/usr/share/vim", {st_mode=S_IFDIR|0777, st_size=72, ...}) = 0
--- SIGSEGV (Segmentation fault) @ 0 (0) ---
+++ killed by SIGSEGV +++
|
Any other suggestions..?
Thanks,
Brian |
|
Back to top |
|
|
didl Retired Dev
Joined: 09 Sep 2003 Posts: 1106 Location: Pittsburgh, PA
|
Posted: Thu Sep 30, 2004 10:06 pm Post subject: |
|
|
Did you try moving ~/.vim ?? |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Fri Oct 01, 2004 12:35 am Post subject: |
|
|
I don't have a .vim file... just .vimrc, which contains just basic options (syntax on, set nowrap...). Same error occurs if I move that also...
Brian |
|
Back to top |
|
|
hardcore l33t
Joined: 01 Nov 2003 Posts: 626 Location: MSU, MI
|
Posted: Fri Oct 01, 2004 4:32 am Post subject: |
|
|
Have you tried compiling with generic cflags? _________________ Nothing can stop me now, cuz I just don't care. |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Fri Oct 01, 2004 11:29 am Post subject: |
|
|
I thought you were on to something, since I did change the CFLAGS slightly trying to get evolution fixed. But still no luck. When I emerge vim-core, then vim, I get this error just before the emerge is done:
Code: |
* Updating documentation tags in /usr/share/vim/vim63
/usr/sbin/ebuild.sh: line 1436: 3035 Segmentation fault DISPLAY= $vim -u NONE -U NONE -T xterm -X -n -f '+set nobackup nomore' "+helptags $d/doc" '+qa!' </dev/null >&/dev/null
>>> original instance of package unmerged safely.
* Updating documentation tags in /usr/share/vim/vim63
/usr/sbin/ebuild.sh: line 1436: 3078 Segmentation fault DISPLAY= $vim -u NONE -U NONE -T xterm -X -n -f '+set nobackup nomore' "+helptags $d/doc" '+qa!' </dev/null >&/dev/null
|
If I see that I pretty much know it's not going to work.
Brian |
|
Back to top |
|
|
hardcore l33t
Joined: 01 Nov 2003 Posts: 626 Location: MSU, MI
|
Posted: Fri Oct 01, 2004 8:22 pm Post subject: |
|
|
Hmm does only vim segfault or is it everything you try to compile/run? It might be a hardware overheating issue, possibly bad ram. Run memtest86 if you haven't yet. _________________ Nothing can stop me now, cuz I just don't care. |
|
Back to top |
|
|
didl Retired Dev
Joined: 09 Sep 2003 Posts: 1106 Location: Pittsburgh, PA
|
Posted: Fri Oct 01, 2004 10:45 pm Post subject: |
|
|
Did you by any chance upgrade glibc/gcc or any of the libraries
that vim links to (check via ldd /usr/bin/vim)? |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Fri Oct 01, 2004 11:10 pm Post subject: |
|
|
It could definitely have happened... emerge -u world upgraded quite a bit. Here's the output to ldd:
Code: |
linux-gate.so.1 => (0xffffe000)
libncurses.so.5 => /lib/libncurses.so.5 (0xb7f9b000)
libgpm.so.1 => /usr/lib/libgpm.so.1 (0xb7f95000)
libperl.so.1 => /usr/lib/libperl.so.1 (0xb7e9c000)
libutil.so.1 => /lib/libutil.so.1 (0xb7e99000)
libc.so.6 => /lib/libc.so.6 (0xb7d91000)
libpthread.so.0 => /lib/libpthread.so.0 (0xb7d41000)
libm.so.6 => /lib/libm.so.6 (0xb7d1f000)
libdl.so.2 => /lib/libdl.so.2 (0xb7d1c000)
libnsl.so.1 => /lib/libnsl.so.1 (0xb7d08000)
libcrypt.so.1 => /lib/libcrypt.so.1 (0xb7cdc000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0xb7fee000)
|
Quite annoying... would've been nice if I set the options to roll back before hand.
Brian |
|
Back to top |
|
|
ciaranm Retired Dev
Joined: 19 Jul 2003 Posts: 1719 Location: In Hiding
|
Posted: Fri Oct 01, 2004 11:23 pm Post subject: |
|
|
I bet you used some broken CFLAGS for something on your system at some point. Try rebuilding your entire system with known safe flags, see if that fixes it. |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Sat Oct 02, 2004 4:52 am Post subject: |
|
|
Since I haven't rebuilt the entire system other than installing from stage 1 and the occasional program reemerge, what's the best way to do so?
The flags I had for most of the system were:
-O2 -fomit-frame-pointer -march=pentium4
They should be fairly safe for then ASUS laptop, centrino. If I do rebuild the system I'll just so an -O2 -march=pentium4. Any other options in make.conf I should enable for the 'safest' rebuild possible?
Thanks again,
Brian |
|
Back to top |
|
|
hardcore l33t
Joined: 01 Nov 2003 Posts: 626 Location: MSU, MI
|
Posted: Sat Oct 02, 2004 5:05 am Post subject: |
|
|
Those are pretty safe flags, I've compiled my system with much more agressive flags, so i don't think your cflags are the problem. _________________ Nothing can stop me now, cuz I just don't care. |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Sat Oct 02, 2004 5:11 am Post subject: |
|
|
Good news, I think... lol.
They were changed a bit due to the evolution problem, and I can't remember (or be absolutely sure) of what they were for the majority of the system. Don't think it can hurt to rebuild the system so everything is done the same.
emerge -e vim redid everything for vim and it's dependencies, would emerge -e world do the system? Not everything's in the world list so there must be a better way to get the libraries and the rest of the system...
Brian |
|
Back to top |
|
|
hardcore l33t
Joined: 01 Nov 2003 Posts: 626 Location: MSU, MI
|
Posted: Sat Oct 02, 2004 5:15 am Post subject: |
|
|
Correct me if I'm wrong but, I believe emerge -eD world will get everything in the world file and all of their dependencies. _________________ Nothing can stop me now, cuz I just don't care. |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Sat Oct 02, 2004 6:10 pm Post subject: |
|
|
72 of 436 is emerging... keep your fingers crossed.
Brian |
|
Back to top |
|
|
ciaranm Retired Dev
Joined: 19 Jul 2003 Posts: 1719 Location: In Hiding
|
Posted: Sat Oct 02, 2004 6:25 pm Post subject: |
|
|
Oooh, I just thought of something... What kernel? |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Sun Oct 03, 2004 9:32 pm Post subject: |
|
|
2.6, mm sources.
If I copy over the binary from my P4 desktop to the laptop, vim works perfectly. It's just getting it to recompile.
I was about 70 programs away from being done emerge -eD world, and got an error (alsa-driver is still loaded, but it's internal to the kernel now). Is there any way to skip that package when I do emerge --resume?
At least the binary gives me a working solution...
Thanks,
Brian |
|
Back to top |
|
|
psyqil Advocate
Joined: 26 May 2003 Posts: 2767
|
Posted: Sun Oct 03, 2004 10:39 pm Post subject: |
|
|
emerge --skipfirst, but that won't help with emerge -e...
Where did you get your kernel from? Does portage know about it? |
|
Back to top |
|
|
bhogg Tux's lil' helper
Joined: 29 Jun 2003 Posts: 96
|
Posted: Sun Oct 03, 2004 11:32 pm Post subject: |
|
|
That was what I was looking for... --skipfirst --resume to skip over the one package it wouldn't emerge correctly.
It's just the mm-sources, emerged normally... vim has compiled with these sources before, so I'm not sure what it could be. As I mentioned before, there was some experimentation trying to get evolution to work, but the emerge -eD world should help restore stuff...
Brian |
|
Back to top |
|
|
|