On Fri, Oct 04, 2019 at 01:42:36PM +0200, Peter Zijlstra wrote:
> If you can find anything there that isn't right, please explain that in
> detail and we'll need to look hard at fixing _that_.
Also, I can't imagine Nick is happy with 128 CPUs banging on that atomic
counter, esp. since atomics are
On Thu, Oct 03, 2019 at 05:36:31PM -0300, Leonardo Bras wrote:
> > Also, I'm not sure I understand things properly.
> >
> > So serialize_against_pte_lookup() wants to wait for all currently
> > out-standing __find_linux_pte() instances (which are very similar to
> > gup_fast).
> >
> > It seems
On Thu, 2019-10-03 at 13:49 -0700, John Hubbard wrote:
> Yes. And to clarify, I was assuming that the changes to mm/gup.c were
> required in order to accomplish your goals. Given that assumption, I
> wanted the generic code to be "proper", and that's what that feedback
> is about.
You assumed
On 10/3/19 1:36 PM, Leonardo Bras wrote:
> On Thu, 2019-10-03 at 09:29 +0200, Peter Zijlstra wrote:
>> On Wed, Oct 02, 2019 at 10:33:14PM -0300, Leonardo Bras wrote:
...
>> This is something entirely specific to Power, you shouldn't be touching
>> generic code at all.
>
> Up to v4, I was
Hello Peter, thanks for the feedback!
On Thu, 2019-10-03 at 09:29 +0200, Peter Zijlstra wrote:
> On Wed, Oct 02, 2019 at 10:33:14PM -0300, Leonardo Bras wrote:
> > If a process (qemu) with a lot of CPUs (128) try to munmap() a large
> > chunk of memory (496GB) mapped with THP, it takes an average
On Wed, Oct 02, 2019 at 10:33:14PM -0300, Leonardo Bras wrote:
> If a process (qemu) with a lot of CPUs (128) try to munmap() a large
> chunk of memory (496GB) mapped with THP, it takes an average of 275
> seconds, which can cause a lot of problems to the load (in qemu case,
> the guest will lock