Kernel 4.9 branch causing issue with several BOINC projects

Kernel 4.9 branch causing issue with several BOINC projects

[Pierre B]

[-- Attachment #1.1: Type: text/plain, Size: 345 bytes --]

When running BOINC under any of the 4.9 kernel, several projects tasks 
terminate immediately in error with messages like this:

process got signal 11
SIGSEGV: segmentation violation

(example above is from WorldCommunityGrid)

Returning to 4.8 kernel allows the sames tasks to run fine.

Any idea what changes 4.9 introduces that breaks this ?

[-- Attachment #1.2: Type: text/html, Size: 1477 bytes --]

Re: Kernel 4.9 branch causing issue with several BOINC projects

[Christian Neukirchen]

Pierre B <pierr...@gmail.com> writes:

> When running BOINC under any of the 4.9 kernel, several projects tasks 
> terminate immediately in error with messages like this:
>
> process got signal 11
> SIGSEGV: segmentation violation
>
> (example above is from WorldCommunityGrid)
>
> Returning to 4.8 kernel allows the sames tasks to run fine.
>
> Any idea what changes 4.9 introduces that breaks this ?

Show a backtrace, either from running in gdb or enable coredumps and
print them from one.

-- 
Christian Neukirchen  <chneuk...@gmail.com>  http://chneukirchen.org

Re: Kernel 4.9 branch causing issue with several BOINC projects

[Pierre B]

[-- Attachment #1.1: Type: text/plain, Size: 3716 bytes --]

Not sure if I'm doing this right but here's what I managed to get:

[root@boinc-zfsnas ~]# uname -a
Linux boinc-zfsnas 4.9.9_1 #1 SMP PREEMPT Thu Feb 9 14:09:14 UTC 2017 
x86_64 GNU/Linux
[root@boinc-zfsnas 0]# gdb wcgrid_scc1_vina_7.08_x86_64-pc-linux-gnu
GNU gdb (GDB) 7.12.1
Copyright (C) 2017 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http:
//gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-unknown-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
<http://www.gnu.org/software/gdb/documentation/>.
For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from wcgrid_scc1_vina_7.08_x86_64-pc-linux-gnu...(no 
debugging symbols found)...done.
(gdb) run -jobfile SCC1_0000083_Bct-C_10945.job -inputfile SCC1_0000083_Bct-
C_10945.zip -seed 218789685
Starting program: /root/0/wcgrid_scc1_vina_7.08_x86_64-pc-linux-gnu -jobfile 
SCC1_0000083_Bct-C_10945.job -inputfile SCC1_0000083_Bct-C_10945.zip -seed 
218789685


Program received signal SIGSEGV, Segmentation fault.
0xffffffffff600000 in ?? ()
(gdb)
(gdb) info program
        Using the running image of child process 970.
Program stopped at 0xffffffffff600000.
It stopped with signal SIGSEGV, Segmentation fault.
Type "info stack" or "info registers" for more information.
(gdb) info stack
#0  0xffffffffff600000 in ?? ()
#1  0x00000000005afeed in ?? ()
#2  0x00000000004d65be in ?? ()
#3  0x00000000004d16af in ?? ()
#4  0x00000000004c6688 in ?? ()
#5  0x00000000004c31ad in ?? ()
#6  0x00000000004c33be in ?? ()
#7  0x00000000004c3c7d in ?? ()
#8  0x00000000004b5b95 in ?? ()
#9  0x000000000047abdc in ?? ()
#10 0x000000000055ad9b in ?? ()
#11 0x0000000000400449 in ?? ()
#12 0x00007fffffffe738 in ?? ()
#13 0x0000000000000000 in ?? ()
(gdb) info registers
rax            0xffffffffff600000       -10485760
rbx            0x7ffffffe2a90   140737488235152
rcx            0x0      0
rdx            0x9      9
rsi            0x0      0
rdi            0x7ffffffe2530   140737488233776
rbp            0x4      0x4
rsp            0x7ffffffe2518   0x7ffffffe2518
r8             0x8d6880 9267328
r9             0x0      0
r10            0x2      2
r11            0x7ffffffe20c8   140737488232648
r12            0xff8    4088
r13            0x7ffffffe2a98   140737488235160
r14            0x7ffffffe2690   140737488234128
r15            0x100    256
rip            0xffffffffff600000       0xffffffffff600000
eflags         0x10202  [ IF RF ]
cs             0x33     51
ss             0x2b     43
ds             0x0      0
es             0x0      0
fs             0x0      0
gs             0x0      0






On Friday, February 10, 2017 at 6:20:46 PM UTC+1, Christian Neukirchen 
wrote:
>
> Pierre B <pier...@gmail.com <javascript:>> writes: 
>
> > When running BOINC under any of the 4.9 kernel, several projects tasks 
> > terminate immediately in error with messages like this: 
> > 
> > process got signal 11 
> > SIGSEGV: segmentation violation 
> > 
> > (example above is from WorldCommunityGrid) 
> > 
> > Returning to 4.8 kernel allows the sames tasks to run fine. 
> > 
> > Any idea what changes 4.9 introduces that breaks this ? 
>
> Show a backtrace, either from running in gdb or enable coredumps and 
> print them from one. 
>
> -- 
> Christian Neukirchen  <chne...@gmail.com <javascript:>>  
> http://chneukirchen.org 
>

[-- Attachment #1.2: Type: text/html, Size: 30751 bytes --]

Re: Kernel 4.9 branch causing issue with several BOINC projects

[Christian Neukirchen]

Pierre B <pierr...@gmail.com> writes:

> Program received signal SIGSEGV, Segmentation fault.
> 0xffffffffff600000 in ?? ()

This is the vdso address.

According to this commit, it probably broke due to
CONFIG_LEGACY_VSYSCALL_NONE=y:

https://gitweb.gentoo.org/repo/gentoo.git/commit/sci-misc?id=38460857021466100646b9a8b110798cc7661abf

-- 
Christian Neukirchen  <chneuk...@gmail.com>  http://chneukirchen.org

Re: Kernel 4.9 branch causing issue with several BOINC projects

[Pierre B]

[-- Attachment #1.1: Type: text/plain, Size: 759 bytes --]

Nice find!

if I understand correctly:
- either programs need to be update/recompiled to no longer use the 
old vsyscall 
- either kernel has to be compile with 
option CONFIG_LEGACY_VSYSCALL_EMULATE=y


On Friday, February 10, 2017 at 9:08:54 PM UTC+1, Christian Neukirchen 
wrote:
>
> Pierre B <pier...@gmail.com <javascript:>> writes: 
>
> > Program received signal SIGSEGV, Segmentation fault. 
> > 0xffffffffff600000 in ?? () 
>
> This is the vdso address. 
>
> According to this commit, it probably broke due to 
> CONFIG_LEGACY_VSYSCALL_NONE=y: 
>
>
> https://gitweb.gentoo.org/repo/gentoo.git/commit/sci-misc?id=38460857021466100646b9a8b110798cc7661abf 
>
> -- 
> Christian Neukirchen  <chne...@gmail.com <javascript:>>  
> http://chneukirchen.org 
>

[-- Attachment #1.2: Type: text/html, Size: 2524 bytes --]