Skip to contentSkip to conversation

http_load causes 405GP kernel panic w/ linux_2_4_devel

DISCUSSION Bryan Rittmeyer 2 participants 3 messages Oct 25, 2001 – Oct 29, 2001

lore
Share & tools
lore
#0 lore raw

Hi,

Looks like linux_2_4_devel is very unstable under high TCP/IP loads and low amounts of memory on the Walnut.

Here's how to get a *double* kernel panic:

1) download http_load from http://www.acme.com/software/http_load/

2) cross compile for ppc405 using MontaVista JE 2.0 (very easy to do, just change CC in the Makefile)

3) setup fast webserver on same network as Walnut, with one large page availible (I used the linux kernel tarball and khttpd on a x86 box).

4) boot Walnut with "mem=8M" or "mem=4M" -- it's probably possible to make this panic happen with larger amounts of physical memory, but it is much faster and easier if you restrict mem via cmdline.

5) run "http_load -parallel 200 -fetches 1000 urls" where "urls" is a file in cwd containing something like "http://192.168.1.1:8080/linux.tar.gz"

6) sit back and wait a few seconds. you should see something like this:

http://192.168.3.126:8080/foo.htm: check-connect SUCCEEDED, ignoring

http://192.168.3.126:8080/foo.htm: check-connect SUCCEEDED, ignoring

...
http://192.168.3.126:8080/foo.htm: check-connect SUCCEEDED, ignoring

http:/Oops: kernel access of bad area, sig: 11

NIP: C009CB00 XER: 00000000 LR: C009CAD8 SP: C0E25DC0 REGS: c0e25d00
TRAP: 0800 Not tainted

MSR: 00009030 EE: 1 PR: 0 FP: 0 ME: 1 IR/DR: 11

TASK = c0e24000[47] 'http_load' Last syscall: 4

last math 00000000 last altivec 00000000

PLB0: bear= 0x00aff17c acr= 0x00000000 besr= 0x00000000

PLB0 to OPB: bear= 0x00000000 besr0= 0x00000000 besr1= 0x00000000

GPR00: C009CAD8 C0E25DC0 C0E24000 00000000 00001030 C012A514 C0222F38 45442C20

GPR08: C01D6140 00000001 00000010 C0187380 00000018 1001FB04 00000000 00000000

GPR16: 00000000 00000000 10010000 FFFFFFFF 00001032 00E25E40 00000000 C0002860

GPR24: C0003964 00000180 C01E6200 C0FFC028 000005EE C0FF1010 C0FFC000 00000000

Call backtrace:

C009CAD8 C00038C0 C00039F4 C0002860 C003365C C000263C 946ECA52

0FF3C1EC 0FF3CE20 0FF2E280 0FF29558 0FF31C58 10002468 100017F8

0FF06298 00000000

ppc405_eth_serr: Receive channel 0

----- latched error -----

DE: descriptor error

----- cumulative errors -----

DEI: descriptor error interrupt

PPC405 Ethernet driver received a receive descriptor error

EMAC DEBUG **********

EMAC_M0 ==> 0x18000000

EMAC_M1 ==> 0x8788000

EMAC_TXM0==> 0x0

EMAC_TXM1==> 0x380f0000

EMAC_RXM ==> 0x580000

EMAC_ISR ==> 0x42

EMAC_IER ==> 0x1af0000

EMAC_IAH ==> 0x4

EMAC_IAL ==> 0xace310b1

EMAC_VLAN_TPID_REG ==> 0x8808

MAL DEBUG **********

MCR ==> 0x7c082

ESR ==> 0x40100000

IER ==> 0x1f

DBR ==> 0x0

TXCASR ==> 0x80000000

TXCARR ==> 0x80000000

TXEOBISR ==> 0x0

TXDEIR ==> 0x0

RXCASR ==> 0x0

RXCARR ==> 0x0

RXEOBISR ==> 0x80000000

RXDEIR ==> 0x80000000

TXCTP0R ==> 0xff6000

TXCTP1R ==> 0xf80000

RXCTP0R ==> 0xff1000

RCBS0 ==> 0x80

dumping the receive descriptors: current slot is 2

Desc 00: status 0x1c00, length 70, addr 0x224010

Desc 01: status 0x1c00, length 70, addr 0x223010

Desc 02: status 0x1c00, length 1518, addr 0x780010

Desc 03: status 0x1c00, length 1518, addr 0x25f010

Desc 04: status 0x1c00, length 1518, addr 0x25e010

Desc 05: status 0x1c00, length 1518, addr 0x25d010

Desc 06: status 0x1c00, length 1518, addr 0x25c010

Desc 07: status 0x1c00, length 70, addr 0x25b010

Desc 08: status 0x1c00, length 70, addr 0x259010

Desc 09: status 0x1c00, length 70, addr 0x258010

Desc 10: status 0x1c00, length 70, addr 0x257010

Desc 11: status 0x1c00, length 70, addr 0x256010

Desc 12: status 0x1c00, length 70, addr 0x255010

Desc 13: status 0x1c00, length 70, addr 0x254010

Desc 14: status 0x1c00, length 70, addr 0x253010

Desc 15: status 0x1c00, length 70, addr 0x252010

Desc 16: status 0x1c00, length 70, addr 0x251010

Desc 17: status 0x1c00, length 70, addr 0x250010

Desc 18: status 0x1c00, length 70, addr 0x24f010

Desc 19: status 0x1c00, length 70, addr 0x74e010

Desc 20: status 0x1c00, length 70, addr 0x2a3010

Desc 21: status 0x1c00, length 70, addr 0x24e010

Desc 22: status 0x1c00, length 70, addr 0x24d010

Desc 23: status 0x1c00, length 70, addr 0x24c010

Desc 24: status 0x1c00, length 70, addr 0x24b010

Desc 25: status 0x1c00, length 70, addr 0x24a010

Desc 26: status 0x1c00, length 70, addr 0x249010

Desc 27: status 0x1c00, length 70, addr 0x248010

Desc 28: status 0x1c00, length 70, addr 0x247010

Desc 29: status 0x1c00, length 70, addr 0x246010

Desc 30: status 0x1c00, length 70, addr 0x245010

Desc 31: status 0x1c00, length 70, addr 0x244010

Desc 32: status 0x1c00, length 70, addr 0x243010

Desc 33: status 0x1c00, length 70, addr 0x242010

Desc 34: status 0x1c00, length 70, addr 0x241010

Desc 35: status 0x1c00, length 70, addr 0x240010

Desc 36: status 0x1c00, length 70, addr 0x602010

Desc 37: status 0x1c00, length 70, addr 0x23f010

Desc 38: status 0x1c00, length 70, addr 0x23e010

Desc 39: status 0x1c00, length 70, addr 0x23c010

Desc 40: status 0x1c00, length 70, addr 0x23b010

Desc 41: status 0x1c00, length 70, addr 0x23a010

Desc 42: status 0x1c00, length 70, addr 0x239010

Desc 43: status 0x1c00, length 70, addr 0x238010

Desc 44: status 0x1c00, length 247, addr 0x237010

Desc 45: status 0x1c00, length 70, addr 0x236010

Desc 46: status 0x1c00, length 70, addr 0x358010

Desc 47: status 0x1c00, length 70, addr 0x235010

Desc 48: status 0x1c00, length 70, addr 0x234010

Desc 49: status 0x1c00, length 70, addr 0x233010

Desc 50: status 0x1c00, length 70, addr 0x232010

Desc 51: status 0x1c00, length 70, addr 0x231010

Desc 52: status 0x1c00, length 70, addr 0x230010

Desc 53: status 0x1c00, length 70, addr 0x22f010

Desc 54: status 0x1c00, length 70, addr 0x22e010

Desc 55: status 0x1c00, length 70, addr 0x22d010

Desc 56: status 0x1c00, length 70, addr 0x22c010

Desc 57: status 0x1c00, length 70, addr 0x22b010

Desc 58: status 0x1c00, length 70, addr 0x22a010

Desc 59: status 0x1c00, length 70, addr 0x229010

Desc 60: status 0x1c00, length 70, addr 0x228010

Desc 61: status 0x1c00, length 70, addr 0x227010

Desc 62: status 0x1c00, length 70, addr 0x226010

Desc 63: status 0x5c00, length 70, addr 0x225010

Oops: kernel access of bad area, sig: 11

NIP: C009CD8C XER: 20000000 LR: C009CD38 SP: C0E25AB0 REGS: c0e259f0
TRAP: 0800 Not tainted

MSR: 00009030 EE: 1 PR: 0 FP: 0 ME: 1 IR/DR: 11

TASK = c0e24000[47] 'http_load' Last syscall: 4

last math 00000000 last altivec 00000000

PLB0: bear= 0x00aff17c acr= 0x00000000 besr= 0x00000000

PLB0 to OPB: bear= 0x00000000 besr0= 0x00000000 besr1= 0x00000000

GPR00: 00001C00 C0E25AB0 C0E24000 00000035 00001030 00000001 00000020 C0160000

GPR08: 00001E7A 00000000 000005EE 00000010 00000018 1001FB04 00000000 00000000

GPR16: 00000000 00000000 10010000 FFFFFFFF 00001032 00E25B30 00000000 C0002860

GPR24: C0003964 00000002 C01E6200 00000010 C0FFC000 C0FFC020 00000002 00000008

Call backtrace:

C009CD38 C00038C0 C00039F4 C0002860 C00029F0 C000B020 C000AE98

C0002868 C009CAD8 C00038C0 C00039F4 C0002860 C003365C C000263C

946ECA52 0FF3C1EC 0FF3CE20 0FF2E280 0FF29558 0FF31C58 10002468

100017F8 0FF06298 00000000

Kernel panic: Aiee, killing interrupt handler!

In interrupt handler - not syncing

<0>Rebooting in 180 seconds..

---

This panic looks like a PPC 405GP ethernet driver issue, perhaps with some VM interaction thrown in. I have access to a BDI2000 and may be able to help debug.

Feel free to e-mail me privately if you'd like my kernel configuration or any additional info... should be fairly easy to reproduce with the above recipe, though.

Regards,

-Bryan

--
Bryan Rittmeyer
mailto:[email protected]

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

#1 lore raw
Quoted 2 lines · Bryan Rittmeyer

Looks like linux_2_4_devel is very unstable under high TCP/IP loads and low amounts of memory on the Walnut.

Yeah, that's why it's called a _development_ tree. If you want to play around here, the purpose is for you to help us fix the bugs.

When you find a solution, post a patch and we'll all take a look at it. If you are willing to help out, that's great, but we all have plenty to do so just expecting someone else to look at it and fix it for you isn't going to be very satisfying :-).

Thanks.

	-- Dan

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

#2 lore raw
Quoted 5 lines · Dan Malek
Quoted 2 lines

Looks like linux_2_4_devel is very unstable under high TCP/IP loads and low amounts of memory on the Walnut.

Yeah, that's why it's called a _development_ tree. If you want to play around here, the purpose is for you to help us fix the bugs.

yep, fair enough.

Quoted 4 lines

When you find a solution, post a patch and we'll all take a look at it. If you are willing to help out, that's great, but we all have plenty to do so just expecting someone else to look at it and fix it for you isn't going to be very satisfying :-).

I am certainly going to investigate it myself, but I thought posting a public bug report might be useful for other users/developers, even before I have a fix.

Anyway, I'll send the patch your way as soon as I have one.

Thanks,

-Bryan

--
Bryan Rittmeyer
mailto:[email protected]

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/