Re: [PATCH V4 01/13] perf/core, x86: Add PERF_SAMPLE_DATA_PAGE_SIZE

[Liang, Kan]

On 2/1/2019 4:22 AM, Peter Zijlstra wrote:
> On Thu, Jan 31, 2019 at 12:27:54PM -0800, kan.liang@xxxxxxxxxxxxxxx wrote:
> > diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c
> > index 374a197..229a73b 100644
> > --- a/arch/x86/events/core.c
> > +++ b/arch/x86/events/core.c
> > @@ -2578,3 +2578,34 @@ void perf_get_x86_pmu_capability(struct x86_pmu_capability *cap)
> >   	cap->events_mask_len	= x86_pmu.events_mask_len;
> >   }
> >   EXPORT_SYMBOL_GPL(perf_get_x86_pmu_capability);
> > +
> > +u64 perf_get_page_size(u64 virt)
> > +{
> > +	unsigned long flags;
> > +	unsigned int level;
> > +	pte_t *pte;
> > +
> > +	if (!virt)
> > +		return 0;
> > +
> > +	/*
> > +	 * Interrupts are disabled, so it prevents any tear down
> > +	 * of the page tables.
> > +	 * See the comment near struct mmu_table_batch.
> > +	 */
> > +	local_irq_save(flags);
> > +	if (virt >= TASK_SIZE)
> > +		pte = lookup_address(virt, &level);
> > +	else {
> > +		if (current->mm) {
> > +			pte = lookup_address_in_pgd(pgd_offset(current->mm, virt),
> > +						    virt, &level);
> > +		} else
> > +			level = PG_LEVEL_NUM;
> > +	}
> > +	local_irq_restore(flags);
> > +	if (level >= PG_LEVEL_NUM)
> > +		return 0;
> > +
> > +	return (u64)page_level_size(level);
> > +}
>
> *sigh* there really isn't anything x86 specific there.

OK. I will split the patch and move the common code to a dedicated patch 
in V5. I will try the proposed code and do some tests on X86.

> >> diff --git a/kernel/events/core.c b/kernel/events/core.c
> > index 236bb8d..d233f45 100644
> > --- a/kernel/events/core.c
> > +++ b/kernel/events/core.c
> > @@ -6352,6 +6358,12 @@ static u64 perf_virt_to_phys(u64 virt)
> >   	return phys_addr;
> >   }
> >   
> > +/* Return page size of given virtual address. IRQ-safe required. */
> > +u64 __weak perf_get_page_size(u64 virt)
> > +{
> > +	return 0;
> > +}
> > +
> >   static struct perf_callchain_entry __empty_callchain = { .nr = 0, };
> >   
> >   struct perf_callchain_entry *
>
> How about something like so instead?
>
> (completely untested, will likely make your grandma eat puppies
That's not funny!


Kan