From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Cooper Subject: Re: [PATCH v2] AMD/IO-APIC: Use old IO-APIC ack method if AMD 813{1, 2} PCI-X tunnel is present Date: Wed, 5 Jun 2013 10:43:03 +0100 Message-ID: <51AF0827.7060507@citrix.com> References: <115069a50a65f219a8b4.1370340835@andrewcoop.uk.xensource.com> <51AE31FE02000078000DB300@nat28.tlf.novell.com> <51AE1A78.5030805@citrix.com> <51AF110B02000078000DB586@nat28.tlf.novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <51AF110B02000078000DB586@nat28.tlf.novell.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Jan Beulich Cc: "Keir (Xen.org)" , Jacob Shin , SuraveeSuthikulpanit , "xen-devel@lists.xen.org" List-Id: xen-devel@lists.xenproject.org On 05/06/13 09:20, Jan Beulich wrote: >>>> On 04.06.13 at 18:48, Andrew Cooper wrote: >> On 04/06/13 17:29, Jan Beulich wrote: >>>> + "Use ioapic_ack_mode=old\n"); >>> "Don't use \"ioapic_ack=new\"\n" would seem better. >>> >>> Jan >> No - this is more informative to someone who doesn't know the Xen >> command line arguments inside/out. > At least the spelling of the option needs to be fixed anyway. And > the reason I prefer the inverted statement is that if this someone > adds the suggested option to the beginning of the command line, > but leaves the bad option in somewhere towards the tail, the bad > behavior (and the message) will still occur. I completely missed the spelling - that does need fixing. While I can appreciate your point of view with the statement, it is only rare cases where someone would set io_apic_mode anyway and when given instructions to use =new, can purge instances of =old. > > But anyway - I continue to be unconvinced that this case can't be > easily enough dealt with the admin adding "ioapic_ack=old" to the > command line. > > Jan > That describes half the errata workarounds we do. Why does this deserve different treatment?. It can certainly can be worked around by using io_apci_ack=old on the command line, and that is how we verified the 'fix'. But in the meantime it took a normally technically-savvy customer 2 months of time in highest level support (i.e. my colleagues and I) before we got to the bottom of the issue. ~Andrew