Re: PROBLEM: ata_piix.c for the ICH5 SATA Controller.

[Mark Lord <liml@rtr.ca>]

Mark Lord wrote:
> Johny Mail list wrote:
>> 2007/6/28, Mark Lord <liml@rtr.ca>:
>>> I have an ugly (but working) hack for the ICH5 ata_piix driver
>>> to support hot insertion/removal of drives, but I don't know if/when
>>> I'll be pushing it upstream.
>>
>> Yes it hang permanently there, after this messages i generally reboot
>> the server.
>> Yes it not support SATA drive hot insertion/removal, but i have make
>> the same test on windows. I unplug one disk when i'm logged and the
>> system don't stop. The drive is removed from the devices list.
>>
>> If you can give me the patch for testing it... I would give you my
>> returns about the good/bad functioning in my case.
> 
> Okay, Here is a working patch for a very specific variant of ICH5.
> If your PCI IDs don't match what the patch is looking for,
> then it should have no effect -- you may need to patch the patch
> to contain the correct PCI IDs (from lspci -n).
> * * *
> 
> Implement ICH5 chipset handling for drive hot insertion/removal.
> This cannot go upstream, as it conflicts with a more generic
> polled-hotplug framework that is currently in development.
> 
> Hot-inserted drives are automatically detected within a second or two,
> and are ready-to-use within 30 seconds or so.  This could be even faster,
> but the 2.6.18.8 libata implementation of error-handling is what slows
> us down here.
...

This patch was for 2.6.18.8 -- it *might* apply to newer kernels,
but I haven't ported it forward yet.

Cheers