[v10,0/4] randomize memory access of dirty_log_perf_test

Message

Colton Lewis Nov. 7, 2022, 6:22 p.m. UTC

Add the ability to randomize parts of dirty_log_perf_test,
specifically the order pages are accessed and whether pages are read
or written.

v10:

Move setting default random seed to argument parsing code.

Colton Lewis (4):
  KVM: selftests: implement random number generator for guest code
  KVM: selftests: create -r argument to specify random seed
  KVM: selftests: randomize which pages are written vs read
  KVM: selftests: randomize page access order

 .../selftests/kvm/access_tracking_perf_test.c |  2 +-
 .../selftests/kvm/dirty_log_perf_test.c       | 59 ++++++++++++++-----
 .../selftests/kvm/include/perf_test_util.h    |  8 ++-
 .../testing/selftests/kvm/include/test_util.h |  7 +++
 .../selftests/kvm/lib/perf_test_util.c        | 33 +++++++++--
 tools/testing/selftests/kvm/lib/test_util.c   | 17 ++++++
 6 files changed, 102 insertions(+), 24 deletions(-)

--
2.38.1.431.g37b22c650d-goog

Comments

Oliver Upton Nov. 30, 2022, 5:55 p.m. UTC | #1

On Mon, Nov 07, 2022 at 06:22:04PM +0000, Colton Lewis wrote:
> Add the ability to randomize parts of dirty_log_perf_test,
> specifically the order pages are accessed and whether pages are read
> or written.
> 
> v10:
> 
> Move setting default random seed to argument parsing code.
> 
> Colton Lewis (4):
>   KVM: selftests: implement random number generator for guest code
>   KVM: selftests: create -r argument to specify random seed
>   KVM: selftests: randomize which pages are written vs read
>   KVM: selftests: randomize page access order

Does someone want to pick this up for 6.2? Also, what tree are we
routing these architecture-generic selftests changes through, Paolo's?

--
Thanks,
Oliver

Marc Zyngier Nov. 30, 2022, 6:11 p.m. UTC | #2

On Wed, 30 Nov 2022 17:55:09 +0000,
Oliver Upton <[e-mail]> wrote:
> 
> On Mon, Nov 07, 2022 at 06:22:04PM +0000, Colton Lewis wrote:
> > Add the ability to randomize parts of dirty_log_perf_test,
> > specifically the order pages are accessed and whether pages are read
> > or written.
> > 
> > v10:
> > 
> > Move setting default random seed to argument parsing code.
> > 
> > Colton Lewis (4):
> >   KVM: selftests: implement random number generator for guest code
> >   KVM: selftests: create -r argument to specify random seed
> >   KVM: selftests: randomize which pages are written vs read
> >   KVM: selftests: randomize page access order
> 
> Does someone want to pick this up for 6.2? Also, what tree are we
> routing these architecture-generic selftests changes through, Paolo's?

That's the usual route, but I can also take them if that makes
someone's life easier. Just let me know.

	M.

Sean Christopherson Nov. 30, 2022, 8:11 p.m. UTC | #3

On Wed, Nov 30, 2022, Marc Zyngier wrote:
> On Wed, 30 Nov 2022 17:55:09 +0000,
> Oliver Upton <[e-mail]> wrote:
> > 
> > On Mon, Nov 07, 2022 at 06:22:04PM +0000, Colton Lewis wrote:
> > > Add the ability to randomize parts of dirty_log_perf_test,
> > > specifically the order pages are accessed and whether pages are read
> > > or written.
> > > 
> > > v10:
> > > 
> > > Move setting default random seed to argument parsing code.
> > > 
> > > Colton Lewis (4):
> > >   KVM: selftests: implement random number generator for guest code
> > >   KVM: selftests: create -r argument to specify random seed
> > >   KVM: selftests: randomize which pages are written vs read
> > >   KVM: selftests: randomize page access order
> > 
> > Does someone want to pick this up for 6.2? Also, what tree are we
> > routing these architecture-generic selftests changes through, Paolo's?
> 
> That's the usual route, but I can also take them if that makes
> someone's life easier. Just let me know.

Doh, sorry.  This is already in kvm/queue, I sent Paolo a pull request[*] for 6.2
for a bunch of the generic selftests updates a while back, but forgot to go back and
respond to the individual series to spread the news.

[*] https://lore.kernel.org/all/[e-mail]