Got hold of one of these machines in this "everything is exploding"
state.  Used the below test program to dump out the static variables and
obtain the alignment of the corruption.  (This program does not
manipulate this data which eliminates a bug in dpkg as cause.)  Note
that the corruption is at the start of the page (and although most
elided here repeats on each page thereafter):

===
#include <stdio.h>

static char b[65536 * 16];

main(int argc, char *argv[])
{
        int p;

        printf("%08lx\n", (long)b);
        for (p = 0; p < sizeof(b); p++) {
                if (b[p]) {
                        printf("%d != 0 @ %d [%08lx]\n", b[p], p, (long)&b[p]);
                }
        }
}
===
10011068
68 != 0 @ 61336 [10020000]
20 != 0 @ 61340 [10020004]
2 != 0 @ 61342 [10020006]
1 != 0 @ 61344 [10020008]
75 != 0 @ 61348 [1002000c]
3 != 0 @ 61349 [1002000d]
2 != 0 @ 61352 [10020010]
8 != 0 @ 61353 [10020011]
[...]
68 != 0 @ 126872 [10030000]
20 != 0 @ 126876 [10030004]
2 != 0 @ 126878 [10030006]
1 != 0 @ 126880 [10030008]
75 != 0 @ 126884 [1003000c]
3 != 0 @ 126885 [1003000d]
2 != 0 @ 126888 [10030010]
8 != 0 @ 126889 [10030011]
[...]
===

I also dumped the corruption in full in a more readable form, I would
note that this seems to contain 'lo' and 'eth0' as if it were networking
related:

===
000000    0044    0000    0014    0002    0001    0000    034b    0000
         D  \0  \0  \0 024  \0 002  \0 001  \0  \0  \0   K 003  \0  \0
000010    0802    fe80    0001    0000    0008    0001    007f    0100
       002  \b 200 376 001  \0  \0  \0  \b  \0 001  \0 177  \0  \0 001
000020    0008    0002    007f    0100    0007    0003    6f6c    0000
        \b  \0 002  \0 177  \0  \0 001  \a  \0 003  \0   l   o  \0  \0
000030    0014    0006    ffff    ffff    ffff    ffff    bb72    0054
       024  \0 006  \0 377 377 377 377 377 377 377 377   r 273   T  \0
000040    bb72    0054    0050    0000    0014    0002    0001    0000
         r 273   T  \0   P  \0  \0  \0 024  \0 002  \0 001  \0  \0  \0
000050    034b    0000    1802    0080    000c    0000    0008    0001
         K 003  \0  \0 002 030 200  \0  \f  \0  \0  \0  \b  \0 001  \0
000060    000a    8203    0008    0002    000a    8203    0008    0004
        \n  \0 003 202  \b  \0 002  \0  \n  \0 003 202  \b  \0 004  \0
000070    000a    ff03    0009    0003    7465    3068    0000    0000
        \n  \0 003 377  \t  \0 003  \0   e   t   h   0  \0  \0  \0  \0
000080    0014    0006    ffff    ffff    ffff    ffff    bcc8    0054
       024  \0 006  \0 377 377 377 377 377 377 377 377 310 274   T  \0
000090    bcc8    0054    0000    0000    0000    0000    0000    0000
       310 274   T  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0
0000a0    0000    0000    0000    0000    0000    0000    0000    0000
        \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0  \0
*
010000    0044    0000    0014    0002    0001    0000    034b    0000
===

I should note at this point that this differs from the corruption as
seen by @vorlon which showed a single bit change in each page.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1301496

Title:
  kernel crash: Unable to handle kernel paging request for data

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1301496/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to