From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paolo Bonzini Date: Wed, 02 Mar 2016 11:37:59 +0000 Subject: Re: [PATCH v3] powerpc: Add tests for sPAPR h-calls Message-Id: <56D6D097.5010109@redhat.com> List-Id: References: <1456908043-11184-1-git-send-email-thuth@redhat.com> <56D6BB7B.70401@redhat.com> <56D6C6C2.3000207@redhat.com> In-Reply-To: <56D6C6C2.3000207@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Thomas Huth , Laurent Vivier , kvm@vger.kernel.org, kvm-ppc@vger.kernel.org, drjones@redhat.com Cc: dgibson@redhat.com On 02/03/2016 11:56, Thomas Huth wrote: > > > + > > > + asm volatile (" sc 1 " : "+r"(r3), "=r"(r4) : "r"(r3)); > > > > as you use "+r" with r3 in the output operands field, I think you don't > > have to declare it in the input operands field. > > Ah, ok, ... gcc inline assembly is always confusing... > Paolo, Andrew, shall I sent a new version for fixing this nit, or is it > ok to keep it in the current, slightly redundant shape? I guess the current one is fine because of the asm register in the declarations. The "really correct" one would use "=r" in the r3 output and "0" in the r3 input, I think. Paolo