Re: Does perf-annotate work correctly?

From: Du, Changbin
Date: Mon Oct 16 2017 - 05:41:34 EST


On Mon, Oct 16, 2017 at 11:28:53AM +0200, Jiri Olsa wrote:
> On Fri, Oct 13, 2017 at 06:15:00PM +0800, Du, Changbin wrote:
> > Hi Jiri,
> > Sorry, missed you (but get_maintainer.pl doesn't list you). Here is ealier email.
> > https://lkml.org/lkml/2017/9/12/158
> >
> > Do you think if this is a real issue?
>
> SNIP
>
> > > 0.02 â test %esi,%esi â
> > > â â js 25 â
> > > 99.98 â â retq â
> > > â25: push %rbp â
> > > â mov $0x440a,%ecx â
> > > â mov $0x440c,%edx â
> > > â vmx_complete_interrupts(): â
> > > â break; â
> > > â } â
> > > â } â
> > > â â
> > > â static void vmx_complete_interrupts(struct vcpu_vmx *vmx) â
> > > â { â
> > > â mov %rsp,%rbp â
> > > â â callq __vmx_complete_interrupts.part.64 â
> > > â __vmx_complete_interrupts(&vmx->vcpu, vmx->idt_vectoring_info, â
> > > â pop %rbp â
> > > â â retq â
>
> hi,
> there's 'o' key to togle the instruction address or you
> can use the perf annotate --stdio to get it.. should be
> easier to tell if that's the same instruction
>

Thanks for replying. I know the reason now, the instructions are shown in pc
address order, and some C statments are split into chunks. It gives me a
illusion. Thanks.

> jirka
>

--
Thanks,
Changbin Du

Attachment: signature.asc
Description: PGP signature