From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Jan Beulich" Subject: Re: Re: [PATCH] x86: fix the potential of encounteringpanic "IO-APIC + timer doesn't work! ..." Date: Fri, 12 Dec 2008 09:54:17 +0000 Message-ID: <494242D9.76E4.0078.0@novell.com> References: <20081212092441.GJ25106@elte.hu> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: Content-Disposition: inline List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Keir Fraser Cc: Ingo Molnar , xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org >>> Keir Fraser 12.12.08 10:37 >>> >On 12/12/2008 09:24, "Ingo Molnar" wrote: > >>>> note that we had to revert that eventually. The real solution is an >>>> array of fixes from Maciej spread out over many months. >>>=20 >>> Does it do more harm than good on some systems, or is it simply not a >>> comprehensive fix? >>=20 >> more harm than good - fixes some, breaks others. Have a look at the >> latest upstream io_apic.c, that has it all settled down reasonably = well. > >Thanks, Ingo. > >Jan: Do you agree we should revert the io_apic.c changes, but keep your >addition of an explicit local_irq_enable() to __start_xen()? No, please don't until Ingo clarified his indication of this having got reverted (see my other response to him). I really can't see why properly enabling/disabling interrupts in that code could do any harm. Jan