On Mon, Sep 03, 2018 at 03:14:31PM +0200, Christoffer Dall wrote: > On Mon, Sep 03, 2018 at 02:11:30PM +0200, Andrew Jones wrote: > > On Mon, Sep 03, 2018 at 01:12:36PM +0200, Christoffer Dall wrote: > > > On Thu, Aug 30, 2018 at 04:17:33PM +0200, Andrew Jones wrote: > > > > From: Shih-Wei Li <shihwei@xxxxxxxxxxxxxxx> > > > > > > > > Here we provide the support for measuring various micro level > > > > operations on arm64. Measurements are currently obtained using > > > > timer counters. Further modifications in KVM will be required > > > > to support timestamping using cycle counters, as KVM now disables > > > > accesses to the PMU counters from the VM. > > > > > > > > Signed-off-by: Shih-Wei Li <shihwei@xxxxxxxxxxxxxxx> > > > > Signed-off-by: Christoffer Dall <cdall@xxxxxxxxxxxxxxx> > > > > Signed-off-by: Andrew Jones <drjones@xxxxxxxxxx> > > > > > > Acked-by: Christoffer Dall <christoffer.dall@xxxxxxx> > > > > Thanks Christoffer! While running this I just noticed a bug though. > > It seemed my hardware was super-duper fast, because I was missing... > > > > > > +#define PS_PER_SEC (1000 * 1000 * 1000 * 1000U) > > ...an 'L' here ^ > > > > I'll send a v2 with that fix now and also add your acked-by. > > > > I didn't verify the numeric aspects of the test this time around, but > looked at what the tests do and ran this on a couple of platforms. > > I figured we should merge this and we can fix small issues later if they > appear. Yeah, I agree. We should get it merged and then as people use it, we'll be able to better see what's wrong or how it can be improved. I think v2 is in decent shape now. Hopefully those that run these types of benchmarks will try it out and suggest/post improvements. Thanks, drew
