From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Cooper Subject: Re: 4.0/4.1 requests - IO-APIC EOI [RFC] Date: Thu, 8 Sep 2011 14:56:57 +0100 Message-ID: <4E68C9A9.8090707@citrix.com> References: <4E689D9A.1020405@citrix.com> <4E68C6C10200007800055485@nat28.tlf.novell.com> <4E68C09B.8050409@citrix.com> <4E68E1D30200007800055501@nat28.tlf.novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <4E68E1D30200007800055501@nat28.tlf.novell.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Jan Beulich Cc: Keir Fraser , "xen-devel@lists.xensource.com" List-Id: xen-devel@lists.xenproject.org On 08/09/11 14:40, Jan Beulich wrote: >>>> On 08.09.11 at 15:18, Andrew Cooper wrote: >> --- SNIP --- >> >> Attached is my first attempt at expanding the EOI to work on older >> IO-APICs. It has not undergone any significant testing yet. >> >> Does this look suitable? > It looks correct, but I'd prefer not having two io_apic_eoi_*() functions: > Namely in the "normal" __eoi_IO_APIC_irq() case you have pin *and* > vector readily available, and hence there is no need to look up anything. > So perhaps a better option would be to pass both pin and vector to > io_apic_eoi() (using e.g. -1 to identify the "unknown-needs-lookup" > case). > > Jan Ok - I will refactor with that suggestion. Also, the code in end_level_ioapic_irq() suggests masking the entry while playing with it, so I will include that as well. -- Andrew Cooper - Dom0 Kernel Engineer, Citrix XenServer T: +44 (0)1223 225 900, http://www.citrix.com