freebsd stable - Apr 2013 - 9.1-REL Supermicro H8DCL-iF kernel panic

This has been brought up before for a different board, none specifically
mentioned this one nor the if_em driver (it was if_igb before on a X8DTU-6+
board).

References:
http://lists.freebsd.org/pipermail/freebsd-stable/2011-September/063958.html
http://www.freebsd.org/cgi/query-pr.cgi?pr=kern/172113

Panic image from H8DCl-iF:
http://nitemail.net/img/crash91-h8dcl-if.png

Original image from X8DTU-6+:
http://www.grosbein.net/img/crash-91rc.png

While I'd love to try out some fixes myself however I don't have a ton
of
time with this system in my hands to be developed on. Jack (as referenced
in the kernel pr) felt that it was too few system specific to go further
with the issue as per the last response.

I would be more than happy to assist or even lend remote access to this
machine to figure out just what's causing the problem if anyone is up for
the task, but I will only have about 3-4 days with it. I have confirmed
dumping msix on the boot loader will permit the system to boot up and
function, however horridly slow (6gbps drives pushing 8mbyte/sec isnt even
usable). Maybe there's a grander problem here.. with supermicro? Let me
know, I'll happily test whatever if it benefits the community.

Ryan McIntosh
e: rmcintosh at nitemare.net

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 4/1/13 12:04 AM, Ryan McIntosh wrote:
> This has been brought up before for a different board, none
> specifically mentioned this one nor the if_em driver (it was if_igb
> before on a X8DTU-6+ board).
> 
> References: 
>
http://lists.freebsd.org/pipermail/freebsd-stable/2011-September/063958.html
>
> 
http://www.freebsd.org/cgi/query-pr.cgi?pr=kern/172113
> 
> Panic image from H8DCl-iF: 
> http://nitemail.net/img/crash91-h8dcl-if.png
> 
> Original image from X8DTU-6+: 
> http://www.grosbein.net/img/crash-91rc.png
> 
> While I'd love to try out some fixes myself however I don't have a
> ton of time with this system in my hands to be developed on. Jack
> (as referenced in the kernel pr) felt that it was too few system
> specific to go further with the issue as per the last response.
> 
> I would be more than happy to assist or even lend remote access to
> this machine to figure out just what's causing the problem if
> anyone is up for the task, but I will only have about 3-4 days with
> it. I have confirmed dumping msix on the boot loader will permit
> the system to boot up and function, however horridly slow (6gbps
> drives pushing 8mbyte/sec isnt even usable). Maybe there's a
> grander problem here.. with supermicro? Let me know, I'll happily
> test whatever if it benefits the community.
I tend to agree with John's patch (on Feb 21, 2013 on kern/172113),
will you have a chance to test it?

(My thought is that we should probably just initialize
adapter->rx_mbuf_sz = MCLBYTES; in _attach() right after adapter->dev
assignment?)

Cheers,
-----BEGIN PGP SIGNATURE-----

iQEcBAEBCAAGBQJRWTe/AAoJEG80Jeu8UPuzMf8IALPmOclwXEwLHSzVX8BEDbq7
YOgCsyNAa9Yi3aeSjDkH3Hvqi3XZzc5FtIeXODUMoU9+vTtI9KQSh4As3UsIYJG5
rGAS9dyT8hJ/VNVAzDNAPRmOTaeSRyXCughfCd5MCJXMTG/6KtVkJ2z/js9VpP4r
sqA3qD21p9q88wfPJIhCj7hHFRLa5emv5Ir76pnZHrQ7ORerwtHEbTosWBWQdG7+
9gREoMl0VB28Zoh8Ai1+TvB79MsUklB4F/XZ63sM2ccJ0Uk1Egn6shI9VMBqbh75
tkMSsjenxbM6/BgK86cyNS+NA8Nyh9hGrpNfab5qj79usJKFskSxUpP/iszHCJc=xUCH
-----END PGP SIGNATURE-----