Re: [RFC] ata/core: don't enable the interrupt while activating the host

[Tejun Heo <tj@kernel.org>]

Hello,

On 03/12/2010 07:04 AM, Sebastian Andrzej Siewior wrote:
> I just migrated from ide_platform to pata_platform on my Swarm/Mips
> board. This resulted in "nobody cared" during request_irq().
> The only difference I noticed between ATA and IDE before registering the
> interrupt handler is that IDE does not enable the interrupt.
> After removing the flag it looks like the controller is working with the
> ATA layer.
> 
> If I flip the polarity of the interrupt, I see the "nobody cared"
> message shortly after seeing ata_sff_softreset() during boot. So I tried
> to perform the software reset with ATA_SRST before requesting the
> interrupt hoping that it makes interrupt go away. It did not so.
> 
> Does anyone have an idea what might go wrong here?

ata_host_start() calls ata_eh_freeze_port() on each port which is
supposed to block the IRQs.  For pata_platform, the callback is
libata-sff.c::ata_sff_freeze() which does all that's necessary.  Can
you please play with ata_sff_freeze() and find out why it's not
working?

Thanks.

-- 
tejun