From mboxrd@z Thu Jan  1 00:00:00 1970
From: thomas.petazzoni@free-electrons.com (Thomas Petazzoni)
Date: Sat, 5 Apr 2014 19:34:35 +0200
Subject: Intel I350 mini-PCIe card (igb) on Mirabox (mvebu / Armada 370)
In-Reply-To: <alpine.DEB.2.10.1404041408280.8617@vroombuntu>
References: <20140325202249.GA10378@obsidianresearch.com>
 <20140325213638.5aba54b6@skate>
 <CAAfodu3RmoN9MYNahwv2kLz5=GvZvK6G9fJorsyVgU8C0J-NEg@mail.gmail.com>
 <20140325222404.GC14718@obsidianresearch.com>
 <20140325223510.GD14718@obsidianresearch.com>
 <CAAfodu3HuzBCAc3r=PPufjrFhEKUQK2b_WRow=GJuC6hSqiuKw@mail.gmail.com>
 <20140326201243.GA1536@obsidianresearch.com>
 <CAAfodu0xZ0V8ikJMT1Z+0JSLeD+nezWAcGNF2_ryvad5tsVDWA@mail.gmail.com>
 <20140326214259.GA12330@obsidianresearch.com>
 <alpine.DEB.2.10.1403270025160.1545@vroombuntu>
 <20140327044054.GA22681@obsidianresearch.com>
 <alpine.DEB.2.10.1403280056250.4521@vroombuntu>
 <alpine.DEB.2.10.1404041408280.8617@vroombuntu>
Message-ID: <20140405193435.50d8dd81@skate>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Dear Neil Greatorex,

On Fri, 4 Apr 2014 14:19:44 +0100 (BST), Neil Greatorex wrote:

> From 50aa11018059704229dd43ca1016defdda04f90c Mon Sep 17 00:00:00 2001
> From: Neil Greatorex <neil@fatboyfat.co.uk>
> Date: Fri, 4 Apr 2014 13:47:09 +0100
> Subject: [PATCH] irqchip: armada-370-xp: Fix releasing of MSIs
> 
> This patch moves the call to irq_dispose_mapping() to after the call to
> armada_370_xp_free_msi(). Without this patch, the armada_370_xp_free_msi
> function would always free MSI#0, no matter what was passed to it.
> 
> Signed-off-by: <neil@fatboyfat.co.uk>
> ---
>   drivers/irqchip/irq-armada-370-xp.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/irqchip/irq-armada-370-xp.c b/drivers/irqchip/irq-armada-370-xp.c
> index 5409564..f5e129e 100644
> --- a/drivers/irqchip/irq-armada-370-xp.c
> +++ b/drivers/irqchip/irq-armada-370-xp.c
> @@ -157,8 +157,8 @@ static void armada_370_xp_teardown_msi_irq(struct msi_chip *chip,
>   					   unsigned int irq)
>   {
>   	struct irq_data *d = irq_get_irq_data(irq);
> -	irq_dispose_mapping(irq);
>   	armada_370_xp_free_msi(d->hwirq);
> +	irq_dispose_mapping(irq);
>   }
> 
>   static struct irq_chip armada_370_xp_msi_irq_chip = {

I want to give it some test, but as it is, I'd prefer to have the
irq_dispose_mapping() done before, and use a local variable to store
d->hwirq to that armada_370_xp_free_msi() can be called after
irq_dispose_mapping(). This is to ensure the sequence is symmetrical
with the MSI setup sequence.

Also, can you detail how you tested with just one port?

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com