Re: 2.6.3 RT8139too NIC problems [NOT resolved]

Re: 2.6.3 RT8139too NIC problems [NOT resolved]

[Nick Warne]

 > yes, that is the pre-NAPI driver. You seem to be the only one having
> problems with the NAPI driver, so please help getting it fixed.
> 

OK, please what do I need to do?  I did build with debug, but dmesg 
gets filled up so fast I have lost the logs.  Also, when I do use the 
new 8139too.c driver, the box (my gateway) is unusable for any type 
of network traffic load.

So I will need to do it in one hit to grab the info and switch back 
again.

Thanks for reply - I want to help and reolve this too...

Nick

-- 
"When you're chewing on life's gristle,
Don't grumble,
Give a whistle
And this'll help things turn out for the best."

Re: 2.6.3 RT8139too NIC problems [NOT resolved]

[Nick Warne]

> > > > > Linux233 kernel: NETDEV WATCHDOG: eth1: transmit timed out
> > > > > Linux233 kernel: eth1: link up, 10Mbps, half-duplex, lpa 0x0000
> > > > > Linux233 kernel: nfs: server 486Linux not responding, still trying
> > > > > Linux233 kernel: nfs: server 486Linux not responding, still trying
> > > > > Linux233 kernel: NETDEV WATCHDOG: eth0: transmit timed out
> > > > > Linux233 kernel: nfs: server 486Linux OK
> > > > > Linux233 kernel: nfs: server 486Linux OK
> > > > > Linux233 kernel: nfs: server 486Linux not responding, still trying
> > > > > Linux233 kernel: NETDEV WATCHDOG: eth0: transmit timed out
> > > > > Linux233 kernel: nfs: server 486Linux OK
> > 
> > Well, I am at a loss now or any idea what to do next.  I have tried 
> > everything this morning to build this an eliminate the problem.  
> > Whatever I do, kernel builds nice, boots nice and no problems... 
> > except for these NIC timeouts - it makes 2.6.3 totally unusable for 
> > me.
> > 
> > I state again, these _very_same_ cards work perfectly under any other 
> > kernel I have ever used over the last 3 years (like I am back on 
> > 2.6.2 right now).
> 
> "This is usually irq routing related...  Try booting with 'noapic' or 
> similar. Jeff"
> 
> OK, this was the solution.  I am right bloody idiot.  Good call Jeff.
> 
> I believed I had APIC turned off (I didn't).  So I then compiled 
> kernel with debugging set in 8139too.c, and thought I would be real 
> cool and catch the problem and submit a kernel patch (Yea, right!).  
> I also looked at what I done.
> 
> I was booting with append=noapic... then read a bit... it needed to 
> be a string, append="noapic".
> 
> It all works great now on 2.6.3.
> 
> Sorry to bother you guys when nothing is wrong, and I apologise for 
> me being a dipstick twice over :/

I spoke too bloody soon.  The problem still exists, but on a lesser 
scale:

Feb 23 19:25:39 Linux233 kernel: nfs: server 486Linux not responding, 
still trying
Feb 23 19:25:41 Linux233 kernel: NETDEV WATCHDOG: eth0: transmit 
timed out
Feb 23 19:25:41 Linux233 kernel: nfs: server 486Linux OK
Feb 23 19:28:27 Linux233 kernel: nfs: server 486Linux not responding, 
still trying
Feb 23 19:28:31 Linux233 kernel: NETDEV WATCHDOG: eth1: transmit 
timed out
Feb 23 19:28:31 Linux233 kernel: eth1: link up, 10Mbps, half-duplex, 
lpa 0x0000
Feb 23 19:28:35 Linux233 kernel: nfs: server 486Linux not responding, 
still trying
Feb 23 19:28:35 Linux233 kernel: NETDEV WATCHDOG: eth0: transmit 
timed out
Feb 23 19:28:35 Linux233 kernel: nfs: server 486Linux OK

I lose 10 seconds on network when this happens, and eth0 going AWOL 
then seems to make eth1 go awol straight after.

The cards all work FINE on 2.6.2/1/0 (2.4.1.->.24)...  what has 
changed so much to cause this, so at least I can supply some info.

:/

Nick

-- 
"I am not Spock", said Leonard Nimoy.
"And it is highly illogical of humans to assume so."

Re: 2.6.3 RT8139too NIC problems [NOT resolved]

[Stephen Hemminger]

The current 2.6.3 driver uses NAPI, and maybe that is related to your troubles.

What exactly is the hardware config (lspci output)?

There are debug messages produced, something like:
	eth0: Tx queue start entry 0  dirty entry 13
do you see these in dmesg output? or change your syslog.conf to capture
debug kernel messages to a different file.

Also, rebuild with RTL8139_DEBUG defined to pick up more info.

Seems like interrupts are getting disabled somehow.

Re: 2.6.3 RT8139too NIC problems [NOT resolved]

[Nick Warne]

> The current 2.6.3 driver uses NAPI, and maybe that is related to your troubles.
> 
> What exactly is the hardware config (lspci output)?
> 
> There are debug messages produced, something like:
> 	eth0: Tx queue start entry 0  dirty entry 13
> do you see these in dmesg output? or change your syslog.conf to capture
> debug kernel messages to a different file.
> 
> Also, rebuild with RTL8139_DEBUG defined to pick up more info.
> 
> Seems like interrupts are getting disabled somehow.

Hi Stephen,

I have resolved it short-term re this post:

http://marc.theaimsgroup.com/?l=linux-kernel&m=107765547029938&w=2

regards,

Nick 

-- 
"When you're chewing on life's gristle,
Don't grumble,
Give a whistle
And this'll help things turn out for the best."