Skip to contentSkip to conversation

some gij-4.3/2.6.24-rc8 crashes

DISCUSSION Helge Deller 3 participants 3 messages Jan 19, 2008 – Jan 22, 2008

lore
Share & tools
lore
#0 lore raw

I noticed the following crashes yesterday with a plain 2.6.24-rc8 kernel. Maybe someone has an idea?
The system still worked afterwards.

      _______________________________
     < Your System ate a SPARC! Gah! >
      -------------------------------
             \   ^__^
              \  (xx)\_______
                 (__)\       )\/\
                  U  ||----w |
                     ||     ||

gij-4.3.bin (pid 23032): Protection id trap (code 27)

YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
PSW: 00000000000001000000000000001111 Not tainted

r00-03  0004000f 10280000 1018d72c 60314388
r04-07  00000000 000197a8 00000001 60314388
r08-11  00000000 fffffff2 603140c8 00000006
r12-15  00176fc0 0001bb40 00000014 00001000
r16-19  402c3554 00019840 0001bb54 00000001
r20-23  000197ae 00000000 6031439e 00000000
r24-27  00000001 1050eaac 7e6c2970 1059f010
r28-31  0000000f 00000001 60314500 1018d4e4

sr00-03 00000000 00000000 00000000 000000ca sr04-07 00000000 00000000 00000000 00000000

IASQ: 00000000 00000000 IAOQ: 1018d76c 1018d770
IIR: 0e95d240 ISR: 000000ca IOR: 000197ae
CPU: 0 CR30: 60314000 CR31: 11111111
ORIG_R28: 1017d598
IAOQ[0]: do_sys_poll+0x138/0x1b0 IAOQ[1]: do_sys_poll+0x13c/0x1b0 RP(r2): do_sys_poll+0xf8/0x1b0

Backtrace:
 [<10106cf0>] die_if_kernel+0x130/0x1a0
 [<1010755c>] handle_interruption+0x3f4/0x58c
 [<1010a078>] intr_check_sig+0x0/0x34
      _______________________________
     < Your System ate a SPARC! Gah! >
      -------------------------------
             \   ^__^
              \  (xx)\_______
                 (__)\       )\/\
                  U  ||----w |
                     ||     ||

gij-4.3.bin (pid 23166): Protection id trap (code 27)

YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
PSW: 00000000000001001111111100001111 Tainted: G D

r00-03  0004ff0f 10280000 1010f864 600b8148
r04-07  445e5400 54aec974 54aec778 445e5424
r08-11  00000000 00000001 54aec974 600b8148
r12-15  00000012 0001bb40 00000014 00001000
r16-19  402c3554 00019840 0001bb54 00000001
r20-23  00000012 fffffff8 37c6ea54 600b80c8
r24-27  445e5424 600b80c8 445e5424 1059f010
r28-31  00000012 00040001 600b82c0 10179e38

sr00-03 00000000 00000000 00000000 000000ca sr04-07 00000000 00000000 00000000 00000000

IASQ: 00000000 00000000 IAOQ: 10138ae4 10138ae8
IIR: 0f54d280 ISR: 000000ca IOR: 445e5424
CPU: 0 CR30: 600b8000 CR31: 11111111
ORIG_R28: 1017d598
IAOQ[0]: copy_siginfo_to_user+0x54/0x458 IAOQ[1]: copy_siginfo_to_user+0x58/0x458 RP(r2): setup_rt_frame+0xb8/0x388

Backtrace:
 [<10106cf0>] die_if_kernel+0x130/0x1a0
 [<1010755c>] handle_interruption+0x3f4/0x58c
 [<1010a078>] intr_check_sig+0x0/0x34
#1 lore raw
Quoted 2 lines · Helge Deller

I noticed the following crashes yesterday with a plain 2.6.24-rc8 kernel. Maybe someone has an idea?

Please see the following URL and my inline comments.
http://www.parisc-linux.org/faq/kernelbug-howto.html

Quoted 68 lines

The system still worked afterwards.

      _______________________________
     < Your System ate a SPARC! Gah! >
      -------------------------------
             \   ^__^
              \  (xx)\_______
                 (__)\       )\/\
                  U  ||----w |
                     ||     ||

gij-4.3.bin (pid 23032): Protection id trap (code 27)

YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
PSW: 00000000000001000000000000001111 Not tainted

r00-03  0004000f 10280000 1018d72c 60314388
r04-07  00000000 000197a8 00000001 60314388
r08-11  00000000 fffffff2 603140c8 00000006
r12-15  00176fc0 0001bb40 00000014 00001000
r16-19  402c3554 00019840 0001bb54 00000001
r20-23  000197ae 00000000 6031439e 00000000
r24-27  00000001 1050eaac 7e6c2970 1059f010
r28-31  0000000f 00000001 60314500 1018d4e4

sr00-03 00000000 00000000 00000000 000000ca sr04-07 00000000 00000000 00000000 00000000

IASQ: 00000000 00000000 IAOQ: 1018d76c 1018d770
IIR: 0e95d240 ISR: 000000ca IOR: 000197ae
CPU: 0 CR30: 60314000 CR31: 11111111
ORIG_R28: 1017d598
IAOQ[0]: do_sys_poll+0x138/0x1b0 IAOQ[1]: do_sys_poll+0x13c/0x1b0 RP(r2): do_sys_poll+0xf8/0x1b0

Backtrace:
 [<10106cf0>] die_if_kernel+0x130/0x1a0
 [<1010755c>] handle_interruption+0x3f4/0x58c
 [<1010a078>] intr_check_sig+0x0/0x34
      _______________________________
     < Your System ate a SPARC! Gah! >
      -------------------------------
             \   ^__^
              \  (xx)\_______
                 (__)\       )\/\
                  U  ||----w |
                     ||     ||

gij-4.3.bin (pid 23166): Protection id trap (code 27)

YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
PSW: 00000000000001001111111100001111 Tainted: G D

r00-03  0004ff0f 10280000 1010f864 600b8148
r04-07  445e5400 54aec974 54aec778 445e5424
r08-11  00000000 00000001 54aec974 600b8148
r12-15  00000012 0001bb40 00000014 00001000
r16-19  402c3554 00019840 0001bb54 00000001
r20-23  00000012 fffffff8 37c6ea54 600b80c8
r24-27  445e5424 600b80c8 445e5424 1059f010
r28-31  00000012 00040001 600b82c0 10179e38

sr00-03 00000000 00000000 00000000 000000ca sr04-07 00000000 00000000 00000000 00000000

IASQ: 00000000 00000000 IAOQ: 10138ae4 10138ae8
IIR: 0f54d280 ISR: 000000ca IOR: 445e5424
CPU: 0 CR30: 600b8000 CR31: 11111111
ORIG_R28: 1017d598
IAOQ[0]: copy_siginfo_to_user+0x54/0x458 IAOQ[1]: copy_siginfo_to_user+0x58/0x458 RP(r2): setup_rt_frame+0xb8/0x388

Quoted 4 lines
Backtrace:
 [<10106cf0>] die_if_kernel+0x130/0x1a0
 [<1010755c>] handle_interruption+0x3f4/0x58c
 [<1010a078>] intr_check_sig+0x0/0x34

- The userspace process received a signal.
- We setup the signal frame on the users stack.
- We start copying the siginfo struct into userspace and fail.

This isn't fatal because we are writing into a userspace address, it *is* fatal to the userspace process.

Did you run out of memory?

c.

#2 lore raw
Quoted 4 lines

This isn't fatal because we are writing into a userspace address, it *is* fatal to the userspace process.

Did you run out of memory?

Don't really have a solution. I used to see this all the time but recent kernels have been much better in handling out of memory situations. There is a little program in the gas testsuite that allocates 1 GB of memory and it used to crash my c3k all the time.

The jc1 compiler is something of a memory hog. I've stopped testing java on my c200 because it takes forever to compile the library. Although the compiler adapts somewhat to the memory available in a machine, I have the sense that the current optimizations run best on machines with 1 GB or more of memory for data. Indeed, I think the relative performance of the 64-bit compiler is better than it used to be because of the larger address space and available memory.

Dave

Signature · 2 lines
J. David Anglin                                  [email protected]
National Research Council of Canada              (613) 990-0752 (FAX: 952-6602)