Samba dies suddenly

[Zigster]

Hi forum,

just suddenly my samba on my Wl500gP V2 dies suddenly when accessing any data via samba with:


Mar 16 20:28:17 kernel: Assertion failure in do_get_write_access() at transaction.c:716: "handle->h_buffer_credits > 0"
Mar 16 20:28:17 kernel: kernel BUG at transaction.c:716!
Mar 16 20:28:17 kernel: Unable to handle kernel paging request at virtual address 00000000, epc == 8006da2c, ra == 8006da2c
Mar 16 20:28:17 kernel: Oops in fault.c::do_page_fault, line 192:
Mar 16 20:28:17 kernel: $0 : 00000000 1000dc00 00000021 00000001 81b6e000 00000000 00000001 00000000
Mar 16 20:28:17 kernel: $8 : 0000215e 8020f31f 00000000 00000000 fffffff9 ffffffff 0000000a 00000002
Mar 16 20:28:17 kernel: $16: 801ae6d0 00000000 81f71580 80eb99d0 81e1f1c0 00000000 8156ed40 00000196
Mar 16 20:28:17 kernel: $24: 81593b42 00000002 81592000 81593c78 00000000 8006da2c
Mar 16 20:28:17 kernel: Hi : 00000000
Mar 16 20:28:17 kernel: Lo : 00000100
Mar 16 20:28:17 kernel: epc : 8006da2c Tainted: P
Mar 16 20:28:17 kernel: Status: 1000dc03
Mar 16 20:28:17 kernel: Cause : 0000000c
Mar 16 20:28:17 kernel: Process smbd (pid: 317, stackpage=81592000)
Mar 16 20:28:17 kernel: Stack: 801ae738 801ae6d0 000002cc 0000001f 801ae9c8 81f71580 00000000
Mar 16 20:28:17 kernel: ffffffd6 81f71580 8156ed40 80eb99d0 81f715f8 8156ed40 81a90248 801ad28c
Mar 16 20:28:17 kernel: 00000196 7fff6df8 8006dc64 81f71580 ffffffd6 81f715f8 8006cab0 00000000
Mar 16 20:28:17 kernel: 81a90200 80a98400 80a98578 80068310 81ee0320 81f71580 8156ed40 80a98400
Mar 16 20:28:17 kernel: 80a98518 00000000 00000400 8006ccb0 801228b8 81b25008 00000001 00000001
Mar 16 20:28:17 kernel: 801e918c ...
Mar 16 20:28:17 kernel: Call Trace: [<801ae738>] [<801ae6d0>] [<801ae9c8>] [<801ad28c>] [<8006dc64>]
Mar 16 20:28:17 kernel: [<8006cab0>] [<80068310>] [<8006ccb0>] [<801228b8>] [<801007a0>] [<c000c1e0>]
Mar 16 20:28:17 kernel: [<80065520>] [<801327d0>] [<80100984>] [<80021cb8>] [<80050b90>] [<80022374>]
Mar 16 20:28:17 kernel: [<80022300>] [<80066238>] [<800661d4>] [<80020228>] [<80050d78>] [<80034eec>]
Mar 16 20:28:17 kernel: [<80023b5c>] [<800364cc>] [<80034fc8>] [<800066dc>] [<800085c4>] [<8000c2b8>]
Mar 16 20:28:17 kernel:
Mar 16 20:28:17 kernel: Code: 02002821 0c0043f5 240602cc <ac000000> 0801b65a 8ec20004 26500078 24020001 c244007c


Web browsing is still possible, but access via all other means is not possibly and I need to unplug power and replug to get access to the box.

Is there anyone familiar with that issue ?

Thanx for any help or advice

Zigster

[wpte]

Unable to handle kernel paging request at virtual address 00000000, epc == 8006da2c, ra == 8006da2c

I think there is something wrong with the swap partition...
what kinda storage device do you use for swap?
because it might be broken at some places


short explenation:
the router has little memory, only 32MB ram, wich it can use to run programs.
A trick of modern kernels is to place data that doesnt fit on the RAM to something called swap space. A special mechanism will activate as soon as the CPU wants data from the RAM memory, however the data is not on the RAM but on swap. The mechanism will look up where it is placed on swap when this so called "hard error" (this is not a bad error, no worries) occurs, and sends it to the ram memory so the cpu can read it.
However, now this piece that is on your swap space isn't there or corrupted, so the CPU expects something to be there, but it isnt, so the process involved crashes.

clear?:D

[Zigster]

Hi,

yep seems to be clear.

My suspicion was a faulty memory, but since I have both wl500gp V1 and wl500gp V2 I was very in doubt, that both were damaged ... so I was looking further, to no prevail yet.

Your hint with the swap is quite a trace...!

Do you think deleting the swap partition and creating a new one would help? Or is there another way to clean that partition ?

Thanx in advance

Zigster

[Zigster]

hmm,

tried:


swapoff
mkswap -v1 <dev>

reboot


... to no prevail.

Will try to completely clean that partition and setup a new swap partition.

I'll keep you posted...and awaiting some enlightment :-)

Zigster

[Zigster]

Hi again,

here's what I did (no success though):

- deleted the swap partition with fdisk.
- created a new partition.
- made a linux swap partition type
- called mkswap on that device
- rebooted the wl500gp
- after complete bootup, called samba init.d script

Result: same crash right after smbd/nmbd were trying to run

I'm somewhat running out of ideas ...

Any hint is much appreciated.

Zigster


PS: Here's what the interactive mode of smbd is outputting:
smbd -i
smbd version 3.2.8 started.
Copyright Andrew Tridgell and the Samba Team 1992-2009
Unable to open printcap file /etc/printcap for read!
Unable to open printcap file /etc/printcap for read!
Segmentation fault

[wpte]

Hi again,

here's what I did (no success though):

- deleted the swap partition with fdisk.
- created a new partition.
- made a linux swap partition type
- called mkswap on that device
- rebooted the wl500gp
- after complete bootup, called samba init.d script

Result: same crash right after smbd/nmbd were trying to run

I'm somewhat running out of ideas ...

Any hint is much appreciated.

Zigster


PS: Here's what the interactive mode of smbd is outputting:
smbd -i
smbd version 3.2.8 started.
Copyright Andrew Tridgell and the Samba Team 1992-2009
Unable to open printcap file /etc/printcap for read!
Unable to open printcap file /etc/printcap for read!
Segmentation fault

hmm.. version 3, I never had that version working.
This error might also reffer to a broken binary, you might need to redownload the IPKG package with:

ipkg update
ipkg upgrade
ipkg install [your package] -force-reinstall -force-overwrite

otherwise you might need to try samba 2:confused:

[Zigster]

Well, Samba 3 seemed to work ok for a while... I switched back to Oleg's Samba ... works OK :-)

Thx wpte ;-)

Zigster