On Wed, Sep 21, 2016 at 08:35:41AM -0500, Wei Huang wrote:
> On 09/21/2016 02:53 AM, Andrew Jones wrote:
> > On Tue, Sep 20, 2016 at 10:33:53PM -0400, Wei Huang wrote:
[snip]
> >> + if (cpu->has_pmu == ON_OFF_AUTO_ON && !kvm_enabled()) {
> >> + cpu->has_pmu = ON_OFF_AUTO_OFF;
> >> + if (!pmu_warned && !qtest_enabled()) {
> >
> > Why do we need this qtest_enabled thing?
>
> Without this, "make check" will print out the warning message (using a
> qtest machine type). This seems to be a common practice across QEMU code.
I see. Too bad it's so ugly. Maybe we need a new error_report that
hides that ugliness, i.e. an error_report that only reports when
!qtest_enabled(). CC'ing Markus to get his opinion, but I'm off topic
now and wouldn't expect anything like that to be worked out before
accepting this patch.
>
> >
> >> + error_report("warning: pmu can't be enabled without KVM
> >> acceleration");
> >> + pmu_warned = true;
> >> + }
> >> + }
> >> + if (cpu->has_pmu == ON_OFF_AUTO_OFF) {
> >> + unset_feature(env, ARM_FEATURE_PMU);
> >> + }
> >> +
[snip]
Thanks,
drew
