net: skge breakage on 2.6.24-rc1

net: skge breakage on 2.6.24-rc1

[Heikki Orsila]

After some bisecting, I found that net skge driver broke on

commit 7fb7ac241162dc51ec0f7644d4a97b2855213c32

My network card is:

0000:00:0e.0 Ethernet controller: 3Com Corporation 3c940 10/100/1000Base-T [Marvell] (rev 12)

Linux cheradenine 2.6.24-rc1-dirty #15 Wed Nov 7 23:39:28 EET 2007 x86_64 GNU/Linux

-- 
Heikki Orsila			Barbie's law:
heikki.orsila@iki.fi		"Math is hard, let's go shopping!"
http://www.iki.fi/shd

Re: net: skge breakage on 2.6.24-rc1

[Heikki Orsila]

On Wed, Nov 07, 2007 at 11:46:21PM +0200, Heikki Orsila wrote:
> After some bisecting, I found that net skge driver broke on
> 
> commit 7fb7ac241162dc51ec0f7644d4a97b2855213c32
> 
> My network card is:
> 
> 0000:00:0e.0 Ethernet controller: 3Com Corporation 3c940 10/100/1000Base-T [Marvell] (rev 12)
> 
> Linux cheradenine 2.6.24-rc1-dirty #15 Wed Nov 7 23:39:28 EET 2007 x86_64 GNU/Linux

Sorry, forgot to say what the problem is. Doing

	nc host port < /dev/zero

on a sending machine (not skge) to an skge machine that is receiving:

	nc -l -p port >/dev/null

with ~60 MiB/s speed, causes the interface go malfunct. A slow 
transfer doesn't cause a problem.

Also, after some fiddling, I noticed that not changing the register 
write order on patch:

+       skge_write32(hw, RB_ADDR(q, RB_END), end);
        skge_write32(hw, RB_ADDR(q, RB_WP), start);
        skge_write32(hw, RB_ADDR(q, RB_RP), start);
-       skge_write32(hw, RB_ADDR(q, RB_END), end);

fixes the visible effect.. Possibly not the root cause of the problem, 
but changing the order back fixes networking here.

-- 
Heikki Orsila			Barbie's law:
heikki.orsila@iki.fi		"Math is hard, let's go shopping!"
http://www.iki.fi/shd

Re: net: skge breakage on 2.6.24-rc1

[Rafael J. Wysocki]

On Wednesday, 7 of November 2007, Heikki Orsila wrote:
> On Wed, Nov 07, 2007 at 11:46:21PM +0200, Heikki Orsila wrote:
> > After some bisecting, I found that net skge driver broke on
> > 
> > commit 7fb7ac241162dc51ec0f7644d4a97b2855213c32
> > 
> > My network card is:
> > 
> > 0000:00:0e.0 Ethernet controller: 3Com Corporation 3c940 10/100/1000Base-T [Marvell] (rev 12)
> > 
> > Linux cheradenine 2.6.24-rc1-dirty #15 Wed Nov 7 23:39:28 EET 2007 x86_64 GNU/Linux
> 
> Sorry, forgot to say what the problem is. Doing
> 
> 	nc host port < /dev/zero
> 
> on a sending machine (not skge) to an skge machine that is receiving:
> 
> 	nc -l -p port >/dev/null
> 
> with ~60 MiB/s speed, causes the interface go malfunct. A slow 
> transfer doesn't cause a problem.
> 
> Also, after some fiddling, I noticed that not changing the register 
> write order on patch:
> 
> +       skge_write32(hw, RB_ADDR(q, RB_END), end);
>         skge_write32(hw, RB_ADDR(q, RB_WP), start);
>         skge_write32(hw, RB_ADDR(q, RB_RP), start);
> -       skge_write32(hw, RB_ADDR(q, RB_END), end);
> 
> fixes the visible effect.. Possibly not the root cause of the problem, 
> but changing the order back fixes networking here.

Please put all of the relevant information into the bugzilla entry at
http://bugzilla.kernel.org/show_bug.cgi?id=9321

Thanks,
Rafael

Re: net: skge breakage on 2.6.24-rc1

[Andrew Morton]

> On Wed, 7 Nov 2007 23:50:30 +0100 "Rafael J. Wysocki" <rjw@sisk.pl> wrote:
> On Wednesday, 7 of November 2007, Heikki Orsila wrote:
> > On Wed, Nov 07, 2007 at 11:46:21PM +0200, Heikki Orsila wrote:
> > > After some bisecting, I found that net skge driver broke on
> > > 
> > > commit 7fb7ac241162dc51ec0f7644d4a97b2855213c32


Thanks for doing the bisection.  It's a good idea to cc the author of the
offending patch after having done this.

> > > My network card is:
> > > 
> > > 0000:00:0e.0 Ethernet controller: 3Com Corporation 3c940 10/100/1000Base-T [Marvell] (rev 12)
> > > 
> > > Linux cheradenine 2.6.24-rc1-dirty #15 Wed Nov 7 23:39:28 EET 2007 x86_64 GNU/Linux
> > 
> > Sorry, forgot to say what the problem is. Doing
> > 
> > 	nc host port < /dev/zero
> > 
> > on a sending machine (not skge) to an skge machine that is receiving:
> > 
> > 	nc -l -p port >/dev/null
> > 
> > with ~60 MiB/s speed, causes the interface go malfunct. A slow 
> > transfer doesn't cause a problem.
> > 
> > Also, after some fiddling, I noticed that not changing the register 
> > write order on patch:
> > 
> > +       skge_write32(hw, RB_ADDR(q, RB_END), end);
> >         skge_write32(hw, RB_ADDR(q, RB_WP), start);
> >         skge_write32(hw, RB_ADDR(q, RB_RP), start);
> > -       skge_write32(hw, RB_ADDR(q, RB_END), end);
> > 
> > fixes the visible effect.. Possibly not the root cause of the problem, 
> > but changing the order back fixes networking here.
> 
> Please put all of the relevant information into the bugzilla entry at
> http://bugzilla.kernel.org/show_bug.cgi?id=9321

Please cc netdev@vger.kernel.org on networking-related discussions.

Re: net: skge breakage on 2.6.24-rc1

[Heikki Orsila]

On Wed, Nov 07, 2007 at 03:06:21PM -0800, Andrew Morton wrote:
> > On Wed, 7 Nov 2007 23:50:30 +0100 "Rafael J. Wysocki" <rjw@sisk.pl> wrote:
> > On Wednesday, 7 of November 2007, Heikki Orsila wrote:
> > > On Wed, Nov 07, 2007 at 11:46:21PM +0200, Heikki Orsila wrote:
> > > > After some bisecting, I found that net skge driver broke on
> > > > 
> > > > commit 7fb7ac241162dc51ec0f7644d4a97b2855213c32
> 
> 
> Thanks for doing the bisection.  It's a good idea to cc the author of the
> offending patch after having done this.

Sorry, I just forgot this time. Just after the bug report I did mail 
the author in private (out of the list).

-- 
Heikki Orsila			Barbie's law:
heikki.orsila@iki.fi		"Math is hard, let's go shopping!"
http://www.iki.fi/shd