Bug 242612 - glibc "Segmentation Fault" issues after update to f7-ga
glibc "Segmentation Fault" issues after update to f7-ga
Product: Fedora
Classification: Fedora
Component: kernel (Show other bugs)
i386 Linux
low Severity low
: ---
: ---
Assigned To: Kernel Maintainer List
Brian Brock
: 242749 (view as bug list)
Depends On:
  Show dependency treegraph
Reported: 2007-06-04 23:01 EDT by Steve Ebersole
Modified: 2008-08-02 19:40 EDT (History)
3 users (show)

See Also:
Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2007-08-29 14:34:38 EDT
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)
requested strace output (1.06 KB, text/plain)
2007-06-05 07:14 EDT, Steve Ebersole
no flags Details

  None (edit)
Description Steve Ebersole 2007-06-04 23:01:48 EDT
This relates to running a java application (IntelliJ java IDE) on a system
originally built using fedora7-test4.  Oddly enough I had no problems in the
original installs.  However, either during the update to fedora7-ga or since
then, I am no longer able to run this application.  This is using the sun jdk,
not gcj.  I am pretty new to Linux and Fedora, so based on the advice of someone
more knowledgeable I ran the following commands, and based on the results he
suggested I open a case as it seemed "that glibc is screwed up".  Below are the

[steve@localhost bin]$ gdb --args /usr/java/jdk1.5.0_11/bin/java -Xms64m
-Xmx256m -XX:MaxPermSize=99m -ea -Xbootclasspath/p:../lib/boot.jar
GNU gdb Red Hat Linux (6.6-8.fc7rh)
Copyright (C) 2006 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you
welcome to change it and/or distribute copies of it under certain
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for
This GDB was configured as "i386-redhat-linux-gnu"...
(no debugging symbols found)
Using host libthread_db library "/lib/libthread_db.so.1".
(gdb) run
Starting program: /usr/java/jdk1.5.0_11/bin/java -Xms64m -Xmx256m
-XX:MaxPermSize=99m -ea -Xbootclasspath/p:../lib/boot.jar
(no debugging symbols found)
(no debugging symbols found)

Program received signal SIGSEGV, Segmentation fault.
0x4d16ac07 in memset () from /lib/ld-linux.so.2
(gdb) bt
#0  0x4d16ac07 in memset () from /lib/ld-linux.so.2
#1  0x4d15b6df in _dl_map_object_from_fd () from /lib/ld-linux.so.2
#2  0x4d15d33c in _dl_map_object () from /lib/ld-linux.so.2
#3  0x4d1611c6 in openaux () from /lib/ld-linux.so.2
#4  0x4d162ec6 in _dl_catch_error () from /lib/ld-linux.so.2
#5  0x4d161785 in _dl_map_object_deps () from /lib/ld-linux.so.2
#6  0x4d15840d in dl_main () from /lib/ld-linux.so.2
#7  0x4d168f6b in _dl_sysdep_start () from /lib/ld-linux.so.2
#8  0x4d1562b8 in _dl_start () from /lib/ld-linux.so.2
#9  0x4d155817 in _start () from /lib/ld-linux.so.2

Further, he suggested running `dmesg` and including its output here. 
Unfortunately, that just gives me the following over and over and over again:
bcm43xx: Error: Microcode "bcm43xx_microcode5.fw" not available or load failed.
bcm43xx: core_up for active 802.11 core failed (-2)

This is a laptop currently with a broadcom card, which I've obviously had wfi
issues with.
Comment 1 Jakub Jelinek 2007-06-05 04:42:30 EDT
When you see SIGSEGV in gdb, in addition to bt please issue also
info regs
and in a different shell run
cat /proc/`pidof java`/maps
Also, running the program under strace instead of gdb and attaching strace
output would be helpful.
Comment 2 Steve Ebersole 2007-06-05 07:07:49 EDT
(gdb) info reqs
Undefined info command: "reqs".  Try "help info".
(gdb) info regs
Undefined info command: "regs".  Try "help info".
(gdb) regs
Undefined command: "regs".  Try "help".
(gdb) info
"info" must be followed by the name of an info command.
List of info subcommands:

info address -- Describe where symbol SYM is stored
info all-registers -- List of all registers and their contents
info args -- Argument variables of current stack frame
info auxv -- Display the inferior's auxiliary vector
info breakpoints -- Status of user-settable breakpoints
info catch -- Exceptions that can be caught in the current stack frame
info checkpoints -- IDs of currently known forks/checkpoints
info classes -- All Objective-C classes
info common -- Print out the values contained in a Fortran COMMON block
info copying -- Conditions for redistributing copies of GDB
info dcache -- Print information on the dcache performance
info display -- Expressions to display when program stops
info extensions -- All filename extensions associated with a source language
info files -- Names of targets and files being debugged
info float -- Print the status of the floating point unit
info forks -- IDs of currently known forks/checkpoints
info frame -- All about selected stack frame
info functions -- All function names
info handle -- What debugger does when program gets various signals
info line -- Core addresses of the code for a source line
info linkmap -- Display the inferior's linkmap
---Type <return> to continue, or q <return> to quit---

Are you sure it should be `info regs`?
Comment 3 Steve Ebersole 2007-06-05 07:09:51 EDT
[steve@localhost ~]$ cat /proc/`pidof java`/maps
bash: pidof: command not found
cat: /proc//maps: No such file or directory
Comment 4 Steve Ebersole 2007-06-05 07:14:19 EDT
Created attachment 156202 [details]
requested strace output

Attaching the output of strace.  As I mentioned I am a Linux/Fedora noob, so I
"believe" I got the command right.  The command I ran is:
strace -o /home/steve/Desktop/strace.txt /usr/java/jdk1.5.0_11/bin/java -Xms64m
-Xmx256m -XX:MaxPermSize=99m -ea -Xbootclasspath/p:../lib/boot.jar
Comment 5 Jakub Jelinek 2007-06-05 07:27:08 EDT
Sorry, typo,
info reg
Use /sbin/pidof if you don't have /sbin in your PATH.

Anyway, from strace the following looks wrong
mmap2(0x4d540000, 8192, PROT_READ|PROT_WRITE,
With MAP_FIXED, kernel must honor the address or the syscall must fail, but in
this case succeeded with a different address.  That would be a kernel bug.
What kernel are you using?
Comment 6 Steve Ebersole 2007-06-05 07:31:21 EDT
Whatever "system update" told me to use ;)

Actually I just got a new update last night as well with no improvement.

As for the specific version, how do I tell?
Comment 7 Jakub Jelinek 2007-06-05 07:36:05 EDT
uname -a
Comment 8 Steve Ebersole 2007-06-05 07:39:45 EDT
[steve@localhost ~]$ uname -a
Linux localhost.localdomain 2.6.21-1.3200.fc8 #1 SMP Sat Jun 2 05:42:16 EDT 2007
i686 i686 i386 GNU/Linux
Comment 9 Dave Jones 2007-06-05 17:33:38 EDT
ok, I think I see what's wrong with the MAP_FIXED code. I've got a patch, I just
want Ingo to take a look at it before I commit it, as it touches execshield.
Comment 10 Jakub Jelinek 2007-06-05 17:35:48 EDT
*** Bug 242749 has been marked as a duplicate of this bug. ***
Comment 11 Dave Jones 2007-06-07 13:52:37 EDT
Can you try the 3218 kernel from http://people.redhat.com/davej/kernels/Fedora/fc7

That should have fixed MAP_FIXED behaviour.
Comment 12 Tom London 2007-06-07 14:28:04 EDT
Works for me:

execve("/usr/bin/vmware-ping", ["vmware-ping", "-q", ""], [/* 20 vars
*/]) = 0
brk(0)                                  = 0x9a79000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY)      = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=78336, ...}) = 0
mmap2(NULL, 78336, PROT_READ, MAP_PRIVATE, 3, 0) = 0x111000
close(3)                                = 0
open("/lib/libc.so.6", O_RDONLY)        = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\320\20"..., 512) = 512
fstat64(3, {st_mode=S_IFREG|0755, st_size=1673760, ...}) = 0
mmap2(0x45efb000, 1390032, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0)
= 0x45efb000
mmap2(0x46049000, 12288, PROT_READ|PROT_WRITE,
mmap2(0x4604c000, 9680, PROT_READ|PROT_WRITE,
close(3)                                = 0

Fails with 2.6.21-1.3209.fc8PAE
Comment 13 Tom London 2007-06-07 20:57:16 EDT
Works in 2.6.21-1.3213.fc8PAE too.

Note You need to log in before you can comment on or make changes to this bug.