libata and legacy ide pcmcia failure

[Tejun Heo]

Robert de Rooy wrote:
> Alan Cox wrote:
>>>>   http://thread.gmane.org/gmane.linux.kernel/530099
>>>>
>>>> It seems we're losing interrupts from the CFA device.  Any ideas?
>>>>       
>>> Alan probably knows more, but ISTR some CFA PCMCIA devices that
>>> needed polling...
>>>     
>>
>> Not that I know of. Not devices anyway - there are embedded boxes with no
>> IRQ configuration for the CF slot but we've never supported those (indeed
>> there are even people who bitbang ATA on GPIO pins...)
>>
>>   
> I was able to test the same PCMCIA adapter in a different ThinkPad, an X40.
> The X40 also fails although it has a totally different Cardbus
> controller (TI on the T40 vs Ricoh on the X40)
> 
> Here is the log from the X40 running the same kernel as before
> (2.6.22rc1-git5 + timing-debug.patch +
> disable-dev_init_param-and-setxfermode-for-CFA.patch)

Can you test the attached patch?

-- 
tejun
-------------- next part --------------
A non-text attachment was scrubbed...
Name: libata-dont-test-slave-register-readiness-after-srst.patch
Type: text/x-patch
Size: 1161 bytes
Desc: not available
Url : http://lists.infradead.org/pipermail/linux-pcmcia/attachments/20070607/48c7b0af/attachment-0001.bin