<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"> <head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> <meta name="Generator" content="Microsoft Word 15 (filtered medium)"> <style></style> </head> <body lang="EN-US" link="blue" vlink="purple" style="word-wrap:break-word"> <div class="WordSection1"> <div> <div style="border:none;border-top:solid #E1E1E1 1.0pt;padding:3.0pt 0in 0in 0in"> From: Randhawa, Jagmeet <jagmeet.randhawa@intel.com> Sent: Wednesday, October 23, 2024 1:29 PM To: Cavitt, Jonathan <jonathan.cavitt@intel.com> Cc: igt-dev@lists.freedesktop.org Subject: Re: [PATCH] [i-g-t] tests/intel/xe_vm:Reduce n_execs for bind-array-enobufs<o:p></o:p> </div> </div> > > > <o:p></o:p> > > > <o:p></o:p> <div> > > > On 10/23/2024 7:12 AM, Cavitt, Jonathan wrote:<o:p></o:p> > > > -----Original Message-----<o:p></o:p> > > > From: Randhawa, Jagmeet <a href="mailto:jagmeet.randhawa@intel.com"> <jagmeet.randhawa@intel.com></a> <o:p></o:p> > > > Sent: Tuesday, October 22, 2024 1:12 PM<o:p></o:p> > > > Cc: <a href="mailto:igt-dev@lists.freedesktop.org">igt-dev@lists.freedesktop.org</a>; Cavitt, Jonathan <a href="mailto:jonathan.cavitt@intel.com"><jonathan.cavitt@intel.com></a>; Zuo, Alex <a href="mailto:alex.zuo@intel.com"><alex.zuo@intel.com></a>; Konieczny, Kamil <a href="mailto:kamil.konieczny@intel.com"><kamil.konieczny@intel.com></a>; Randhawa, Jagmeet <a href="mailto:jagmeet.randhawa@intel.com"><jagmeet.randhawa@intel.com></a><o:p></o:p> > > > Subject: [PATCH] [i-g-t] tests/intel/xe_vm:Reduce n_execs for bind-array-enobufs<o:p></o:p> > > > <o:p></o:p> > > > The bind-array-enobufs test in xe_vm.c is intended to trigger<o:p></o:p> > > > an -ENOBUFS error by submitting an oversized bind array.<o:p></o:p> > > > After encountering and handling the error, the test reduces<o:p></o:p> > > > n_execs by half and retries the operation. On some environments<o:p></o:p> > > > with stricter resource limits (e.g. simulator),<o:p></o:p> > > > <o:p></o:p> > > Is it "just" simulator this is an issue on? Because if so, we could probably<o:p></o:p> > > use igt_run_in_simulation() to narrow the scope:<o:p></o:p> > > <o:p></o:p> > > """<o:p></o:p> > > xe_cork_fini(&cork);<o:p></o:p> > > - n_execs = n_execs / 2;<o:p></o:p> > > + n_execs /= igt_run_in_simulation() ? 4 : 2;<o:p></o:p> > > }<o:p></o:p> > > """<o:p></o:p> > > <o:p></o:p> > > If not, then<o:p></o:p> > > Reviewed-by: Jonathan Cavitt <a href="mailto:jonathan.cavitt@intel.com"> <jonathan.cavitt@intel.com></a><o:p></o:p> > > -Jonathan Cavitt<o:p></o:p> > <o:p></o:p> </div> > Thank you for your comment and review Jonathan. Oddly when I tested the following line<o:p></o:p> > "n_execs /= igt_run_in_simulation() ? 4 : 2;" The test fails as it was before the original patch was sent<o:p></o:p> > so I suspect igt_run_in_simulation() is returning false and we are taking the "2" integer in this case.<o:p></o:p> > I believe it is just simulator we are facing this issue on, however I think leaving the code as <o:p></o:p> > "n_execs = n_execs / 4;" doesn't do any harm if we run it on any other platform such as hardware<o:p></o:p> > , as it does not affect the basis of the test which is to trigger enobufs error, handle it, and then <o:p></o:p> > run with less n_execs to pass on the next iterations. This patch will just accelerate the test on hardware.<o:p></o:p> > If this works with you, may I ask if I still have your reviewed-by? I can change it and investigate<o:p></o:p> > if you feel strongly about it. Thanks again.<o:p></o:p> <o:p> </o:p> That’s certainly strange, but I’ll accept that as an explanation. My Reviewed-by stands.<o:p></o:p> -Jonathan Cavitt<o:p></o:p> <o:p> </o:p> > <o:p></o:p> > Jagmeet<o:p></o:p> > <o:p></o:p> > > > <o:p></o:p> > > > <o:p></o:p> > > > halving n_execs isn't sufficient to prevent the ENOBUFS error on the retry.<o:p></o:p> > > > Reducing n_execs further to n_execs / 4 allows the test to pass<o:p></o:p> > > > successfully. This patch modifies the test to reduce n_execs<o:p></o:p> > > > to n_execs / 4 after an ENOBUFS error is handled. This ensures<o:p></o:p> > > > compatibility with environments that have tighter resource<o:p></o:p> > > > constraints while maintaining the test's integrity.<o:p></o:p> > > > > Cc: Jonathan Cavitt <a href="mailto:jonathan.cavitt@intel.com"> <jonathan.cavitt@intel.com></a><o:p></o:p> > > > Signed-off-by: Jagmeet Randhawa <a href="mailto:jagmeet.randhawa@intel.com"> <jagmeet.randhawa@intel.com></a><o:p></o:p> > > > ---<o:p></o:p> > > > tests/intel/xe_vm.c | 2 +-<o:p></o:p> > > > 1 file changed, 1 insertion(+), 1 deletion(-)<o:p></o:p> > > > > > diff --git a/tests/intel/xe_vm.c b/tests/intel/xe_vm.c<o:p></o:p> > > > index 7a8740b69..e78ddd0e5 100644<o:p></o:p> > > > --- a/tests/intel/xe_vm.c<o:p></o:p> > > > +++ b/tests/intel/xe_vm.c<o:p></o:p> > > > @@ -957,7 +957,7 @@ test_bind_array(int fd, struct drm_xe_engine_class_instance *eci, int n_execs,<o:p></o:p> > > > xe_cork_end(&cork);<o:p></o:p> > > > xe_cork_wait_done(&cork);<o:p></o:p> > > > xe_cork_fini(&cork);<o:p></o:p> > > > - n_execs = n_execs / 2;<o:p></o:p> > > > + n_execs = n_execs / 4;<o:p></o:p> > > > }<o:p></o:p> > > > <o:p></o:p> > > > xe_vm_bind_array(fd, vm, bind_exec_queue, bind_ops, n_execs, sync, 1);<o:p></o:p> > > > -- <o:p></o:p> > > > 2.34.1 <o:p></o:p> > <o:p></o:p> </div> </body> </html>