Re: [PATCH] libahci: ahci interrupt check for disabled port since private_data may be NULL

From: Alexander Gordeev <agordeev@redhat.com>
To: Tejun Heo <tj@kernel.org>
Cc: David Milburn <dmilburn@redhat.com>, linux-ide@vger.kernel.org
Subject: Re: [PATCH] libahci: ahci interrupt check for disabled port since private_data may be NULL
Date: Thu, 17 Apr 2014 15:07:32 +0200	[thread overview]
Message-ID: <20140417130732.GB2293@dhcp-26-207.brq.redhat.com> (raw)
In-Reply-To: <20140417134543.GD15326@htj.dyndns.org>

On Thu, Apr 17, 2014 at 09:45:43AM -0400, Tejun Heo wrote:
> > 	if (pp)
> > 		rc = devm_request_threaded_irq(..., irq + i, ...)
> > 	else
> > 		disable_irq(irq + i);
> 
> Ah, was looking at the old code.  But again, we didn't even request
> the irq, why would we disable it?

The driver does not request IRQs for dummy ports, but it still requests
MSIs for them. The MSI framework initializes the resources for all MSIs
(IRQ #s, CPU interrupt vectors). Hence, although we do not request a
IRQ #, it still exists - just with no handler associated with the driver.
IOW - with a default handler, which would probably complain loudly once
an interrupt from a dummy port comes (I suppose David could check it).

So if we just disable dummy port's IRQ #s we should never hear from them.

> Thanks.
> 
> -- 
> tejun

-- 
Regards,
Alexander Gordeev
agordeev@redhat.com