Re: r8169 patches

Re: r8169 patches

[Francois Romieu]

(netdev Cced. People, please, please Cc: netdev)

Soren Hansen <soren@ubuntu.com> :
[...]
> I have an Asrock ConRoe945G-DVI motherboard with a RTL8111/8168B in it.
> The stock r8159 driver locks up from time to time. I've described it
> here: https://launchpad.net/bugs/121815 . The patchset at
> http://www.fr.zoreil.com/linux/kernel/2.6.x/2.6.22-rc3/ seems to fix it.
> Can you explain why the patchset is not in the mainline kernel?

1. thanks for the report
2. increase net.core.wmem_max, ping -f -q -l xx -s yy with xx and yy
   big enough and the driver will _still_ stop receiving/transmitting :o(
3. if I take the time to write a nice message to tell you why, it will
   delay me:
   - publishing an updated patchset against -rc6 that can handle davem's
     git-net with minor pain for akpm
   - preparing a serie for Jeff
   - answering the rot^Wpending r8169 reports

-- 
Ueimor

Re: r8169 patches

[Soren Hansen]

[-- Attachment #1: Type: text/plain, Size: 1308 bytes --]

On Thu, Jun 28, 2007 at 11:10:37PM +0200, Francois Romieu wrote:
> > I have an Asrock ConRoe945G-DVI motherboard with a RTL8111/8168B in
> > it.  The stock r8159 driver locks up from time to time. I've
> > described it here: https://launchpad.net/bugs/121815 . The patchset
> > at http://www.fr.zoreil.com/linux/kernel/2.6.x/2.6.22-rc3/ seems to
> > fix it.  Can you explain why the patchset is not in the mainline
> > kernel?
> 1. thanks for the report

Thanks for providing the patch!

> 2. increase net.core.wmem_max, ping -f -q -l xx -s yy with xx and yy
>    big enough and the driver will _still_ stop receiving/transmitting :o(

Oh, dear.

> 3. if I take the time to write a nice message to tell you why, it will
>    delay me:
>    - publishing an updated patchset against -rc6 that can handle davem's
>      git-net with minor pain for akpm
>    - preparing a serie for Jeff
>    - answering the rot^Wpending r8169 reports

Thanks for taking time to responding anyway!

The reason I asked was that I was hoping the response would be that the
patchset had no know regressions, but still lacked testing. In that
case, I would have discussed its inclusion in our (Ubuntu's) alpha
releases with our kernel team, but if you feel you're on track already,
I'll stop annoying you. :)

Thanks for your time.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]