Skip to contentSkip to conversation

2.6.6 kernel oops on ifconfig eth0 mtu 1300

DISCUSSION Soeren Sonnenburg 2 participants 4 messages May 15, 2004 – May 16, 2004

lore
Share & tools
lore
#0 lore raw

Hi.

When I have some ethernet connection and then do
ifconfig eth0 mtu 1300 I get an immediate kernel panic (kernel 2.6.6) powerbook g4 15" 1ghz.

(xmon trace (jpeg) on request)

Ideas ?
Soeren

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

#1 lore raw
Quoted 7 lines · Soeren Sonnenburg

Hi.

When I have some ethernet connection and then do
ifconfig eth0 mtu 1300 I get an immediate kernel panic (kernel 2.6.6) powerbook g4 15" 1ghz.

(xmon trace (jpeg) on request)

Yes, please, the trace

Ben.

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

#2 lore raw
Quoted 10 lines · Benjamin Herrenschmidt
Quoted 7 lines · Soeren Sonnenburg

Hi.

When I have some ethernet connection and then do
ifconfig eth0 mtu 1300 I get an immediate kernel panic (kernel 2.6.6) powerbook g4 15" 1ghz.

(xmon trace (jpeg) on request)

Yes, please, the trace

doesn't it happen when you do ifconfig eth0 mtu 1300 ? anyway here it
is: use a web-browser to view it (it is a redirect):
http://www.nn7.de/kernel/mtu1300.jpg

Soeren

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

#3 lore raw
Quoted 8 lines · Soeren Sonnenburg
Quoted 7 lines · Benjamin Herrenschmidt
Quoted 6 lines · Soeren Sonnenburg

Hi.

When I have some ethernet connection and then do
ifconfig eth0 mtu 1300 I get an immediate kernel panic (kernel 2.6.6) powerbook g4 15" 1ghz.

Quoted 7 lines
Quoted 3 lines
Quoted 1 line

(xmon trace (jpeg) on request)

Yes, please, the trace

doesn't it happen when you do ifconfig eth0 mtu 1300 ? anyway here it
is: use a web-browser to view it (it is a redirect):
http://www.nn7.de/kernel/mtu1300.jpg

Heh, didn't want to crash the laptop today. I CC'ed David Miller on that as he is the main author of the driver.

David: the crash is apparently us doing skb_put of ETH_FRAME_LEN + RX_LEN, so we tune the skb allocator to the MTU ? It's commented out in my bk snapshot but then, I may be missing something. Should we change sungem to use the mtu instead of ETH_FRAME_LEN ?

Ben.

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