From: Keir Fraser Subject: Re: [PATCH v2] xen: handle paged gfn in wrmsr_hypervisor_regs Date: Fri, 03 May 2013 15:26:57 +0100 Message-ID: References: <5183DEBE02000078000D31BB@nat28.tlf.novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <5183DEBE02000078000D31BB@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 , Olaf Hering Cc: xen-devel@lists.xen.org List-Id: xen-devel@lists.xenproject.org On 03/05/2013 14:58, "Jan Beulich" wrote: >>>> On 03.05.13 at 14:57, Olaf Hering wrote: >> @@ -1682,14 +1682,25 @@ static int svm_msr_write_intercept(unsig >> if ( wrmsr_viridian_regs(msr, msr_content) ) >> break; >> >> - wrmsr_hypervisor_regs(msr, msr_content); >> + ret = wrmsr_hypervisor_regs(msr, msr_content); >> + switch ( ret ) >> + { >> + case -EAGAIN: >> + result = X86EMUL_RETRY; >> + break; >> + case 0: >> + result = X86EMUL_UNHANDLEABLE; >> + break; >> + default: >> + break; > > As you had already noticed the hard way - case 0 and default of > course need to be switched (0 -> okay, anything else -> > unhandleable). No! Actually anything other than -EAGAIN should be handled here as 'okay'. In fact the return codes from wrmsr_hypervisor_regs are going to be a bit of a mess if we're not careful. I suggest the following return codes: 0: not handled 1: handled -EINVAL: error during handling -EAGAIN: retry The HVM callers should then handle as follows: -EAGAIN: rc = X86EMUL_RETRY -EINVAL: goto gp_fault 0: try other msr handlers (if any) 1: we're done, return X86EMUL_OKAY Does that make sense? -- Keir > Jan > > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xen.org > http://lists.xen.org/xen-devel