From: Ingo Molnar <mingo@elte.hu>
Subject: Re: [PATCH 36 of 38] xen: route hardware irqs via Xen
Date: Thu, 20 Nov 2008 10:31:59 +0100
Message-ID: <20081120093159.GA6885@elte.hu>
References: <patchbomb.1226603398@abulafia.goop.org> <00079eee3818aa0205e2.1226603434@abulafia.goop.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Return-path: <linux-kernel-owner+glk-linux-kernel-3=40m.gmane.org-S1754581AbYKTJc1@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <00079eee3818aa0205e2.1226603434@abulafia.goop.org>
Sender: linux-kernel-owner@vger.kernel.org
To: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: linux-kernel@vger.kernel.org, Xen-devel <xen-devel@lists.xensource.com>, the arch/x86 maintainers <x86@kernel.org>, Ian Campbell <ian.campbell@citrix.com>
List-Id: xen-devel@lists.xenproject.org


> +#ifdef CONFIG_XEN_PCI
> +	irq = xen_register_gsi(gsi, triggering, polarity);
> +	if ((int)irq >= 0)
> +		return irq;
> +#endif

why not change irq to 'int' and avoid the cast?

also, please eliminate the #ifdef by turning xen_register_gsi() into a 
'return -1' inline on !CONFIG_XEN_PCI.

> +#ifdef CONFIG_XEN_PCI
> +	xen_pci_init();
> +#endif

hide the #ifdef in a header please. (like you already properly do for
xen_setup_pirqs())

> +	if (rc != 0) {
>  		if (!probing_irq(irq))
>  			printk(KERN_INFO "Failed to obtain physical IRQ %d\n",
>  			       irq);
> +		dump_stack();

generally it's better to use WARN() or WARN_ONCE() to get good debug 
feedback and stackdumps. (they also document the reason for the dump)

> @@ -523,8 +526,6 @@
>  	} else
>  		irq = find_unbound_irq();
>  
> -	spin_unlock(&irq_mapping_update_lock);
> -
>  	set_irq_chip_and_handler_name(irq, &xen_pirq_chip,
>  				      handle_level_irq, "pirq");

hm, looks like a stray bugfix?

	Ingo