[parisc-linux] Kernel crash while running glibc's make check

[parisc-linux] Kernel crash while running glibc's make check

[Jeff Bailey]

Hi!  As an attempt to help the glibc/NPTL efforts, I've upgraded to a
recent kernel in order to make sure problems I'm seeing that are
triggered by kernel bugs haven't already been fixed.  However, I can no
longer make it through a glibc make check pass.

I have reproduced this twice now, and both times it happens while
running the tst-gettext4.sh script.  It doesn't seem to matter whether
the system has a long or short uptime, or the amount of work done.

Bug report particulars follow:

$ uname -a
Linux test148.lamont 2.6.18-rc4-pa4 #2 Sat Aug 26 08:52:30 EDT 2006 parisc GNU/Linux

This is using the current gcc-4.1 in edgy with a binutils snapshot from
mid-July.

The machine is a j6700.

I am using a serial console.

jbailey@test148:/var/lib/dpkg/info$ cat /proc/cmdline 
root=/dev/sda5 HOME=/ console=ttyS0 TERM=vt102 palo_kernel=2/vmlinuz-2.6.18-rc4-pa4

Noise from the console:

test148.lamont login: kernel BUG at mm/mmap.c:1970!
Backtrace:
 [<1015cd58>] exit_mmap+0x17c/0x19c
 [<10123a70>] mmput+0x48/0xbc
 [<10128f18>] do_exit+0x114/0x8ac
 [<10129728>] do_group_exit+0x78/0xbc
 [<101348c0>] get_signal_to_deliver+0x2c8/0x4cc
 [<1010fab0>] do_signal+0x68/0x814
 [<1010c384>] syscall_do_signal+0x54/0xa0
Kernel panic - not syncing: BUG!
 <3>BUG: scheduling while atomic: ld.so.1/0x00000100/10062
Backtrace:
 [<101023a0>] schedule+0x538/0x724
 [<10138e48>] flush_cpu_workqueue+0xa0/0xf0
 [<10138ee4>] cancel_rearming_delayed_workqueue+0x28/0x60
 [<10231750>] lcd_print+0x58/0x178
 [<10232c44>] process_shutdown+0x3c/0x6c
 [<1012c33c>] tasklet_action+0x6c/0xf4
 [<1012c198>] __do_softirq+0xd8/0xec
 [<10104670>] __lock_text_end+0x48/0x4c
 [<1010b068>] intr_return+0x0/0xc

System map is at http://people.ubuntu.com/~jbailey/System.map-2.6.18-rc4-pa4

The kernel was built with the c3000_defconfig.

-- 
I do not agree with a word you say, but I will defend to the death your
right to say it. 
 - Voltaire
_______________________________________________
parisc-linux mailing list
parisc-linux@lists.parisc-linux.org
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

Re: [parisc-linux] Kernel crash while running glibc's make check

[Thibaut VARENE]

On 8/27/06, Jeff Bailey <jbailey@raspberryginger.com> wrote:
> Hi!  As an attempt to help the glibc/NPTL efforts, I've upgraded to a
> recent kernel in order to make sure problems I'm seeing that are
> triggered by kernel bugs haven't already been fixed.  However, I can no
> longer make it through a glibc make check pass.
>
> I have reproduced this twice now, and both times it happens while
> running the tst-gettext4.sh script.  It doesn't seem to matter whether
> the system has a long or short uptime, or the amount of work done.
>
> Bug report particulars follow:
>
> $ uname -a
> Linux test148.lamont 2.6.18-rc4-pa4 #2 Sat Aug 26 08:52:30 EDT 2006 parisc GNU/Linux
>
> This is using the current gcc-4.1 in edgy with a binutils snapshot from
> mid-July.

Totally out of curiosity, if you have a chance to rebuild your kernel
with gcc-3.X and see if it dies the same way. I've been spotting some
strange variations between kernels built with gcc-3.X and gcc-4.X...

> I am using a serial console.
>
> jbailey@test148:/var/lib/dpkg/info$ cat /proc/cmdline
> root=/dev/sda5 HOME=/ console=ttyS0 TERM=vt102 palo_kernel=2/vmlinuz-2.6.18-rc4-pa4
>
> Noise from the console:
>
> test148.lamont login: kernel BUG at mm/mmap.c:1970!
> Backtrace:
>  [<1015cd58>] exit_mmap+0x17c/0x19c
>  [<10123a70>] mmput+0x48/0xbc
>  [<10128f18>] do_exit+0x114/0x8ac
>  [<10129728>] do_group_exit+0x78/0xbc
>  [<101348c0>] get_signal_to_deliver+0x2c8/0x4cc
>  [<1010fab0>] do_signal+0x68/0x814
>  [<1010c384>] syscall_do_signal+0x54/0xa0
> Kernel panic - not syncing: BUG!

Died in signal code, why am I not surprised? ;)

>  <3>BUG: scheduling while atomic: ld.so.1/0x00000100/10062
> Backtrace:
>  [<101023a0>] schedule+0x538/0x724
>  [<10138e48>] flush_cpu_workqueue+0xa0/0xf0
>  [<10138ee4>] cancel_rearming_delayed_workqueue+0x28/0x60
>  [<10231750>] lcd_print+0x58/0x178
>  [<10232c44>] process_shutdown+0x3c/0x6c
>  [<1012c33c>] tasklet_action+0x6c/0xf4
>  [<1012c198>] __do_softirq+0xd8/0xec
>  [<10104670>] __lock_text_end+0x48/0x4c
>  [<1010b068>] intr_return+0x0/0xc

This is unrelated. For some reason the led/lcd tasklet doesn't
'shutdown' properly on panic. I wonder if it's hooked to the panic
handler correctly (if it should)...

-- 
Thibaut VARENE
http://www.parisc-linux.org/~varenet/

<jbailey> Insert tab A into slot B doesn't take too long to figure
out.  Leave a couple horny kids in a room for an hour, they don't need
instructions.
_______________________________________________
parisc-linux mailing list
parisc-linux@lists.parisc-linux.org
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

Re: [parisc-linux] Kernel crash while running glibc's make check

[Jeff Bailey]

[-- Attachment #1: Type: text/plain, Size: 1516 bytes --]

On Sun, Aug 27, 2006 at 11:05:34AM -0700, Jeff Bailey wrote:
> Hi!  As an attempt to help the glibc/NPTL efforts, I've upgraded to a
> recent kernel in order to make sure problems I'm seeing that are
> triggered by kernel bugs haven't already been fixed.  However, I can no
> longer make it through a glibc make check pass.

Because reduced test cases are cuddly, I have produced one (attached).
=)

I found I could produce the same failure with the ltp testsuite, so I
reduced that instead.

And the failure message again:
root@test148:~/foo# ./adjtimex02
kernel BUG at mm/mmap.c:1970!
Backtrace:
 [<10160978>] exit_mmap+0x16c/0x18c
 [<10124160>] mmput+0x4c/0xc0
 [<101290c0>] do_exit+0x130/0x840
 [<10129860>] do_group_exit+0x6c/0xcc
 [<1010c134>] syscall_exit+0x0/0x14
Kernel panic - not syncing: BUG!
 <3>BUG: scheduling while atomic: adjtimex02/0x00000100/2627
Backtrace:
 [<1010262c>] schedule+0x6d0/0x6f4
 [<10139c54>] flush_cpu_workqueue+0xb8/0x110
 [<1013a164>] cancel_rearming_delayed_workqueue+0x58/0x5c
 [<10238ce4>] lcd_print+0x54/0x1b0
 [<1023a270>] process_shutdown+0x3c/0x6c
 [<1012cf78>] tasklet_action+0x64/0xec
 [<1012cb68>] __do_softirq+0xd8/0x114
 [<10104918>] __lock_text_end+0x48/0x54
 [<1010b068>] intr_return+0x0/0xc

I have updated the system map at
http://people.ubuntu.com/~jbailey/System.map-2.6.18-rc4-pa4
with the one for the kernel recompiled with gcc-3.4.6.

Tks,
Jeff Bailey

-- 
I do not agree with a word you say, but I will defend to the death your
right to say it. 
 - Voltaire

[-- Attachment #2: mmaptstcase.c --]
[-- Type: text/x-csrc, Size: 476 bytes --]

/*
 * Testcase reduced from ltp, adjtimex02
 *
 * 1) adjtimex(2) fails with errno set to EFAULT if buf does
 *	   not point to writable memory
 */

#include <stdio.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/timex.h>

int
main()
{
	struct timex *buffp = (struct timex *) -1;

	/* Check whether we are root */
	if (geteuid() != 0) {
		fprintf(stderr, "This needs to run as root.\n");
		return(-1);
	}
	
	/* Call adjtimex(2) */
	adjtimex(buffp);

	return 0;
}



[-- Attachment #3: Type: text/plain, Size: 169 bytes --]

_______________________________________________
parisc-linux mailing list
parisc-linux@lists.parisc-linux.org
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux