Re: ksoftirqd uses 99% CPU triggered by network traffic (maybe RLT-8139 related)

2 messages, 2 authors, 2004-08-03 · open the first message on its own page

Re: ksoftirqd uses 99% CPU triggered by network traffic (maybe RLT-8139 related)

From: Francois Romieu <romieu@fr.zoreil.com>
Date: 2004-08-02 22:38:04

Pasi Sjoholm [off-list ref] :
I forgot to mention that it was quite hard to crash the driver with that 
/* Clear out errors and receive interrupts */-patch. Took about 15minutes 
everytime, when normally it takes about 2mins.
I have made a few changes. Please enable the DEBUG option and set msglvl
to its maximal value via ethtool. You may test the patches separately if
you find some time but the log once both r8139-10.patch and r8139-20.patch
are applied would be enough.

If the log fills too fast, you may comment out any message which does
not belong to rtl8139_rx().

--
Ueimor

Re: ksoftirqd uses 99% CPU triggered by network traffic (maybe RLT-8139 related)

From: Pasi Sjoholm <hidden>
Date: 2004-08-03 12:34:41

On Tue, 3 Aug 2004, Francois Romieu wrote:
I have made a few changes. Please enable the DEBUG option and set msglvl
to its maximal value via ethtool. You may test the patches separately if
you find some time but the log once both r8139-10.patch and r8139-20.patch
are applied would be enough.
The full logfiles can be downloaded from:

http://www.cc.jyu.fi/~ptsjohol/syslog1.gz 
http://www.cc.jyu.fi/~ptsjohol/syslog2.gz 

The first log file is with both patchs applied and the second one with one 
little change to rx8139_rx() to show if it even goes to through 

"        while (netif_running(dev) && received < budget
               && (RTL_R8 (ChipCmd) & RxBufEmpty) == 0) {"-section.

This was the change which I made.. so you can see in the second log file 
that there won't be any of these messages after the driver has crashed. 

/*              if (netif_msg_rx_status(tp))*/
        printk(KERN_DEBUG "%s:  rtl8139_rx() status %4.4x, size %4.4x,"
        " cur %4.4x.\n", dev->name, rx_status, rx_size, cur_rx);

For the first logfile the exact crash time is "13:02:22" and the for the 
second one it is "14:54:49".
If the log fills too fast, you may comment out any message which does
not belong to rtl8139_rx().
I took out those "exiting with interrupt"-messages.

--
Pasi Sjöholm
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help