Messages in this thread | | | From | "Pan, Jacob jun" <> | Date | Wed, 14 Jul 2010 11:19:45 -0700 | Subject | RE: Regression: 2.6.34 boot fails on E5405 system, bisected: de08e2c26 |
| |
>-----Original Message----- >From: Ben Greear [mailto:greearb@candelatech.com] >Sent: Wednesday, July 14, 2010 10:07 AM >To: Pan, Jacob jun >Cc: Robert Hancock; linux-kernel; jbarnes@virtuousgeek.org >Subject: Re: Regression: 2.6.34 boot fails on E5405 system, bisected: >de08e2c26 > >On 07/14/2010 08:36 AM, Pan, Jacob jun wrote: >> what is the config size of 10.1? >> ls -l /sys/bus/pci/devices/0000:00:10.1/config >> >> if that is 256, it might be related to this patch. > >That patch is already in 2.6.34.y (with slight white-space >change it seems: space before <). > >I just posted a patch to lkml that fixes the problem for me, >based on a suggestion by Robert Hancock. > >I think this or something similar should to go 2.6.34.y stable >as well. >
I have not seen the patch yet, but there is no guarantee that capabilities are always laid out in ascending address. So I think we cannot bail out when pcie_cap >> 20 <= pos
If that is some bug in the config space, can we fix it with some quirks?
Thanks,
Jacob
| |