[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: JFFS2:0 powerfailtesting-checkfs:4



Sorry about the delayed reply. I've been out sick with a bad cold (Aaachu!
;).

>vmalik@xxxxxxx.com said:
>>  UhOh! and it was going so well. After 1692 power cycles, I got the
>> following:

>Hmmm. All looks normal till the panic. Is it repeatable?

Yes (mostly, see below), though the panic only takes place when I try to
boot from this
partition. Whe I boot from my hard disk and try to mount this partition
under /mnt/jffs2, this is what I get:

***CONSOLE LOG****
<...snip of some (similar) lines....>
Found empty flash at 0x4f370c
ofs 0x004f370c has already been seen. Skipping
Found empty flash at 0x4f3714
Found empty flash at 0x4f3718
ofs 0x004f3718 has already been seen. Skipping
Found empty flash at 0x4f3724
Found empty flash at 0x4f372c
Found empty flash at 0x4f373c
Found empty flash at 0x4f375c
Found empty flash at 0x4f3774
Found empty flash at 0x4f3778
ofs 0x004f3778 has already been seen. Skipping
Found empty flash at 0x4f377c
ofs 0x004f377c has already been seen. Skipping
Found empty flash at 0x4f3780
Found empty flash at 0x4f3784
ofs 0x004f3784 has already been seen. Skipping
Found empty flash at 0x4f3788
ofs 0x004f3788 has already been seen. Skipping
Found empty flash at 0x4f3790
Found empty flash at 0x4f37b0
ofs 0x004f37b0 has already been seen. Skipping
Found empty flash at 0x4f37b4
Found empty flash at 0x4f37c4
ofs 0x004f37c4 has already been seen. Skipping
Found empty flash at 0x4f37c8
Found empty flash at 0x4f37d8
***END CONSOLE LOG****

That's it! KERN_DEBUG was enabled. The system is not
completely dead at this point. I can still ping it,
and the console echo's back a CR on the terminal.

However, it behaves as if there are (almost) no user CPU 
resources available. I cannot telnet to it (no login prompt ever
comes up), and all telnet sessions are frozen.

I've left the system from more than 4-5 days with no progress
in the situation.

I then rebooted the system and tried booting again from the
JFFS2 partition (that I was manually trying to mount in above)
and I got the kernel panic again, but at a different location.
This time the (last of the)console messages were:

***CONSOLE LOG****
<...snip...>
ofs 0x004f576c has already been seen. Skipping
ofs 0x004f5778 has already been seen. Skipping
ofs 0x004f577c has already been seen. Skipping
ofs 0x004f57e4 has already been seen. Skipping
ofs 0x004f5818 has already been seen. Skipping
ofs 0x004f582c has already been seen. Skipping
ofs 0x004f58c0 has already been seen. Skipping
ofs 0x004f5968 has already been seen. Skipping
ofs 0x004f5984 has already been seen. Skipping
ofs 0x004f59c8 has already been seen. Skipping
ofs 0x004f59cc has already been seen. Skipping
Kernel panic: Out of memory and no killable processes...
****END CONSOLE LOG****

Notice the different address on the (second) last line
of the log (the first (panic) time this address was
0x004f639c).

> Can you make
>the 
>offending panic code print the number of objects/pages used in each of
the 
>jffs2_* slab caches? And also BUG() so we can see how much memory was
being 
>asked for and from where.

Please explain in a bit more detail. This is greek to me :)


>We shouldn't be allocating any memory in the scanning of the 
>partially-erased block.

Another interesting point is that the kernel panic happens
really quickly. Only 3127 "ofs 0x00xxxxxx has already been seen. Skipping"
lines are printed before the kernel panic, and I have 
64megs in the machine. We must be leaking a lot of memory
somewhere.


Vipin

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo@xxxxxxx.com