lkml.org 
[lkml]   [2000]   [Mar]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [2.3.99-pre3] via-rhine.o died again!
On Fri, 24 Mar 2000, Jeff Garzik wrote:

> Jean-Luc Pedneault wrote:
> >
> > 2.3.50 up to 2.3.99-pre1 killed the driver, 2.3.99-pre2 worked fine, but
> > 2.3.99-pre3 made it die again.

2.3.50 had a broken non-alignment of the descriptor rings, that was fixed
(in or before pre2).

There are a few changes between pre2 and pre3, if you take the pre2 tree
and apply the testing/pre3-7.gz patch does that work?
(the driver should identify itself as v1.03a-LK1.1.2)

The 2.3.99-pre3 contains the changes made to use the new dynamic DMA
mapping interface (Documentation/DMA-mapping.txt).

What card do you have (+ dmesg output for the card)?
x86 system?


> > The fact is, the driver loads, but the very first packet I send is going
> > to hang the via-rhine.o driver. This means I can see the stats in
> > /sbin/ifconfig eth0.
> >
> > In 2.2.12-20, I can clearly see that the IRQ and base addresses are the
> > same - they should be anyway since they're assigned by the BIOS.
> >
> > After shutting down the eth0 interface many times, removed the driver,
> > reloaded the driver, reactivated the eth0 interface, the very same error
> > keeps getting back every time I send the first packet:
> >
> > ---<snip>---
> > eth0: Oversized Ethernet frame spanned multiple buffers, entry 0x1 length
> > 0 status 00000000!
> > eth0: Oversized Ethernet frame c73ec010 vs c73ec010.
> > ---<snip>---
> >
> > What does it means, I have no clue. The values in the second error change
> > everytime (I suppose it's a frame dynamically allocated in memory). Every
> > other packet I send after displays nothing, they just add up in the
> > Dropped counter in /sbin/ifconfig eth0.
> >
> > Also, it's worth noting I already got one of the two above errors before
> > (I think it's the second one) in the 2.2.14 kernel (not sure), when the
> > network was in heavy load, especially when two systems used it at the same
> > time.
>
> There were some incorrect endian changes which made it into the pre
> patch...

... but the "final" 2.3.99-pre3 has David Millers (by him untested) fixes
for this.

I added Justin to the Cc since he can also see problems with this version
of the driver and he is much more likely than me to solve this. For me
things are fine, unfortunately.

/Urban


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:57    [W:0.216 / U:0.096 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site