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

Re: system hang + kernel oops because of high usb load?



Orjan Friberg wrote on 16.12.2003 16:11 MET:
> Friedrich Lobenstock wrote:
> 
>>
>> After some time (approx. 1 hour IIRC) I see the following
>> messages appear on the console:
>> Found enabled EP for epid 0, starting bulk channel.
>> Found enabled EP for epid 0, starting bulk channel.
> 
> This message, while printed, is harmless.  There is a timer that 
> periodically checks if the bulk DMA channel needs to be started, because 
> when we insert traffic we might do it just as the DMA controller is 
> about to stop the channel.

Ok. Those messages happend just some minutes before the lookup
therefore I thought that they might be in a connection to it.

> Which USB host driver version is this?  (I assume 1.19, since you've 
> asked related earlier.)

Sorry, I forgot to tell that I'm using the usb host driver 1.19 on
SDK R1_91 with:
   kernel 2.4.22
   + axis-linux-2_4_22.diff  (applied via running ./install in the SDK dir)
   + ebtables-brnf-3_vs_2.4.22.diff

I've just removed the ebtables-brnf patch, compiled again and
I'm currently re-running the test. If it take again about 2,5 hours
to reach the lookup we've got another 1,5 hours to go...

BTW if it makes a difference here the differences between
my kernelconfig and the axis default:

diff -u kernelconfig.axis kernelconfig.latest | grep "^[\+\-]" | grep -v "^[\+\-]#"
-CONFIG_ETRAX_NETWORK_LED_ON_WHEN_ACTIVITY=y
+CONFIG_ETRAX_NETWORK_LED_ON_WHEN_LINK=y
+CONFIG_IP_ADVANCED_ROUTER=y
+CONFIG_IP_NF_FTP=y
+CONFIG_IP_NF_MATCH_HELPER=y
+CONFIG_IP_NF_MATCH_STATE=y
+CONFIG_IP_NF_MATCH_CONNTRACK=y
+CONFIG_IP_NF_MATCH_UNCLEAN=y
+CONFIG_IP_NF_NAT_FTP=y
+CONFIG_IP_NF_MANGLE=y
+CONFIG_IP_NF_TARGET_TOS=y
+CONFIG_IP_NF_TARGET_ECN=y
+CONFIG_IP_NF_TARGET_LOG=y
+CONFIG_IP_NF_TARGET_TCPMSS=y
+CONFIG_BRIDGE=y
-CONFIG_NFS_FS=y
-CONFIG_NFS_V3=y
-CONFIG_SUNRPC=y
-CONFIG_LOCKD=y
-CONFIG_LOCKD_V4=y

-- 
MfG / Regards
Friedrich Lobenstock