Dual opteron various segfaults with 2.6.14.2 and earlier kernels
Hi all,
I'm seeing several segfaults on a couple of HP DL585 Dual Opterons, 8Gb ram each.
The segfaults are like this:
factorial[17031]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffe287e0 error 4
factorial[17034]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffc6f450 error 4
factorial[17038]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffdbd060 error 4
factorial[17044]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffb48fa0 error 4
factorial[17046]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffc2a7f0 error 4
ld[3997]: segfault at 0000000000000020 rip 00002aaaaad1a525 rsp 00007fffffa8e960 error 4
ld[4234]: segfault at 0000000000000020 rip 00002aaaaad1a525 rsp 00007fffffc3a1e0 error 4
This is only an example; often during some "make", also sed segfaults (!). I've seen this with 2.6.12, 2.6.13.4, 2.6.14.2
distro: gentoo
gcc (GCC) 3.4.4 (Gentoo 3.4.4-r1, ssp-3.4.4-1.0, pie-8.7.8)
I've attached config.gz; I use Discontiguous Memory model, since Sparse Memory simply won't boot.
I've googled a bit to find some help, but while I've find only problems like mine, I've found no hints unfortunately.
Can someone give me some help?
Of course, more details are available.
Signature · 4 lines
Fabio "Cova" Coatti http://members.ferrara.linux.it/cova Ferrara Linux Users Group http://ferrara.linux.it GnuPG fp:9765 A5B6 6843 17BC A646 BE8C FA56 373A 5374 C703 Old SysOps never die... they simply forget their password.
Quoted 3 lines · Fabio Coatti
Hi all,
I'm seeing several segfaults on a couple of HP DL585 Dual Opterons, 8Gb ram each.
are you using the gentoo buildstuff for this? eg libjail or whatever it's called?
Alle 15:41, mercoledì 23 novembre 2005, Arjan van de Ven ha scritto:
Quoted 7 lines · Fabio Coatti
Quoted 3 lines
Hi all,
I'm seeing several segfaults on a couple of HP DL585 Dual Opterons, 8Gb ram each.are you using the gentoo buildstuff for this? eg libjail or whatever it's called?
ldd /usr/bin/sed
libc.so.6 => /lib/libc.so.6 (0x00002aaaaabc1000)
/lib64/ld-linux-x86-64.so.2 (0x00002aaaaaaab000)The kernel is compiled in usual way, with vanilla sources, no distro patches.
Signature · 4 lines
Fabio "Cova" Coatti http://members.ferrara.linux.it/cova Ferrara Linux Users Group http://ferrara.linux.it GnuPG fp:9765 A5B6 6843 17BC A646 BE8C FA56 373A 5374 C703 Old SysOps never die... they simply forget their password.
Quoted 20 lines · Fabio Coatti
Hi all,
I'm seeing several segfaults on a couple of HP DL585 Dual Opterons, 8Gb ram each.The segfaults are like this:
factorial[17031]: segfault at0000000000020f31rip00000000004035aersp00007fffffe287e0error 4
factorial[17034]: segfault at0000000000020f31rip00000000004035aersp00007fffffc6f450error 4
factorial[17038]: segfault at0000000000020f31rip00000000004035aersp00007fffffdbd060error 4
factorial[17044]: segfault at0000000000020f31rip00000000004035aersp00007fffffb48fa0error 4
factorial[17046]: segfault at0000000000020f31rip00000000004035aersp00007fffffc2a7f0error 4
ld[3997]: segfault at 0000000000000020 rip00002aaaaad1a525rsp00007fffffa8e960error 4
ld[4234]: segfault at 0000000000000020 rip00002aaaaad1a525rsp00007fffffc3a1e0error 4
Quoted 3 lines
This is only an example; often during some "make", also sed segfaults (!). I've seen this with 2.6.12, 2.6.13.4, 2.6.14.2
The symtoms look just like the TLB flush filter errata which affected SMP x86_64 kernels upto (at least) 2.6.13.4. IIRC it was fixed for 2.6.14 (at least I stopped using the patch after 2.6.13.4).
Are you sure you saw this with 2.6.14+ ?
Andrew Walrond
Alle 23:55, mercoledì 23 novembre 2005, Andrew Walrond ha scritto:
Quoted 21 lines · Fabio Coatti
Quoted 20 lines
Hi all,
I'm seeing several segfaults on a couple of HP DL585 Dual Opterons, 8Gb ram each.The segfaults are like this:
factorial[17031]: segfault at0000000000020f31rip00000000004035aersp00007fffffe287e0error 4
factorial[17034]: segfault at0000000000020f31rip00000000004035aersp00007fffffc6f450error 4
factorial[17038]: segfault at0000000000020f31rip00000000004035aersp00007fffffdbd060error 4
factorial[17044]: segfault at0000000000020f31rip00000000004035aersp00007fffffb48fa0error 4
factorial[17046]: segfault at0000000000020f31rip00000000004035aersp00007fffffc2a7f0error 4
ld[3997]: segfault at 0000000000000020 rip00002aaaaad1a525rsp00007fffffa8e960error 4
ld[4234]: segfault at 0000000000000020 rip00002aaaaad1a525rsp00007fffffc3a1e0error 4
Quoted 8 lines
Quoted 2 lines
This is only an example; often during some "make", also sed segfaults (!). I've seen this with 2.6.12, 2.6.13.4, 2.6.14.2
The symtoms look just like the TLB flush filter errata which affected SMP
x86_64kernels upto (at least) 2.6.13.4. IIRC it was fixed for 2.6.14 (at least I stopped using the patch after 2.6.13.4).Are you sure you saw this with 2.6.14+ ?
yes, uname says 2.6.14.2; on a second identical machine, I've just seen this:
factorial[2352]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffbfaf60 error 4
factorial[2354]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffe3fc70 error 4
factorial[2361]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffb07c50 error 4
factorial[2358]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffb07c50 error 4
factorial[2363]: segfault at 0000000000020f31 rip 00000000004035ae rsp 00007fffffe6d270 error 4
the kernel and HW are the same.
Hello Fabio,
Quoted 14 lines · Fabio Coatti
yes, uname says 2.6.14.2; on a second identical machine, I've just seen this:
factorial[2352]: segfault at
0000000000020f31rip00000000004035aersp00007fffffbfaf60error 4
factorial[2354]: segfault at0000000000020f31rip00000000004035aersp00007fffffe3fc70error 4
factorial[2361]: segfault at0000000000020f31rip00000000004035aersp00007fffffb07c50error 4
factorial[2358]: segfault at0000000000020f31rip00000000004035aersp00007fffffb07c50error 4
factorial[2363]: segfault at0000000000020f31rip00000000004035aersp00007fffffe6d270error 4
Quoted 1 line
the kernel and HW are the same.
Error 4 means a read in userland on a not mapped area.
The above isn't necessairly a kernel or hardware problem, it looks like an userland bug if it segfaults at such a low address (20f31). Nothig is mapped below "0x400000" exactly to catch these kind of bugs.
You should debug the program and check what's the code at address 0x4035ae? You can check it with gdb or objdump -d. Probably there's a 64bit bug in the program that doesn't trigger on x86 32bit (or you may not be noticing the segfault on 32bits because it wouldn't be logged in the syslog).
Hope this helps ;)
Quoted 18 lines · Andrea Arcangeli
Hello Fabio,
Quoted 15 lines · Fabio Coatti
yes, uname says 2.6.14.2; on a second identical machine, I've just seen this:
factorial[2352]: segfault at
0000000000020f31rip00000000004035aersp00007fffffbfaf60error 4
factorial[2354]: segfault at0000000000020f31rip00000000004035aersp00007fffffe3fc70error 4
factorial[2361]: segfault at0000000000020f31rip00000000004035aersp00007fffffb07c50error 4
factorial[2358]: segfault at0000000000020f31rip00000000004035aersp00007fffffb07c50error 4
factorial[2363]: segfault at0000000000020f31rip00000000004035aersp00007fffffe6d270error 4
Quoted 7 lines
Quoted 1 line
the kernel and HW are the same.
Error 4 means a read in userland on a not mapped area.
The above isn't necessairly a kernel or hardware problem, it looks like an userland bug if it segfaults at such a low address (20f31). Nothig is mapped below "
0x400000" exactly to catch these kind of bugs.
Which makes sense; the sed failures seen during 'make' runs were probably a result of the TLB flush filter errata on kernels prior to 2.6.14 , whereas the above is a userland bug which occurs with all kernels.
Andrew Walrond