Re: Kernel fix for ICH7M controller

[Tejun Heo <htejun@gmail.com>]

Jeff Garzik wrote:
> Tejun Heo wrote:
>> Jonathan Dieter wrote:
>>> The ata_piix module doesn't detect that my laptop's disk controller 
>>> (Intel Corporation 82801GBM/GHM - ICH7 Family) is configured to have 
>>> PATA drives in the first two channels.  I looked at Intel's 
>>> documentation and they say that my controller isn't supposed to 
>>> support this configuration, but it's the default setup on this 
>>> laptop, there's no way to change it in the BIOS, and it works in 
>>> Windows, so I'm assuming this is a documentation error.
>>>
>>> This patch changes one line in ata_piix.c so that it sees this 
>>> configuration as valid (and therefore recognizes my drives).  I've 
>>> tested it on my laptop and the module is now working perfectly 
>>> (whereas the kernel was using the old IDE driver previously with no 
>>> DMA enabled).
>>
>> Hello,
>>
>> Interesting.  Can you post the result of 'lspci -n -vvv -xxx'?  I 
>> don't see why this support shouldn't go into mainline but wanna see if 
>> narrowing down the effect a bit (if PCI ID allows us to) would be a 
>> good idea.
> 
> FWIW several PCI devices use the ich6m info entry.  We may want to split 
> out ICH7M rather than assume other chips behave similarly.

Hello, Jeff.

The controller is reporting subsystem id 1025(Acer):0107.  Either we can 
enable that entry for all ICH7Ms or only for the Acer.  I'm more 
inclined to the former.  What do you think?

Thanks.

-- 
tejun

-- 
VGER BF report: H 6.45778e-11