From: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
To: Krzysztof Niemiec <krzysztof.niemiec@intel.com>,
igt-dev@lists.freedesktop.org
Cc: Andi Shyti <andi.shyti@linux.intel.com>,
Krzysztof Karas <krzysztof.karas@intel.com>,
Sebastian Brzezinka <sebastian.brzezinka@intel.com>
Subject: Re: [PATCH i-g-t v3] tests/intel/gem_exec_endless: Use the per-gt interface to set frequencies
Date: Fri, 10 Apr 2026 14:16:33 +0200 [thread overview]
Message-ID: <80fb52072aec29387944507afc63ceda0fc72210.camel@linux.intel.com> (raw)
In-Reply-To: <20260409140640.48827-2-krzysztof.niemiec@intel.com>
On Thu, 2026-04-09 at 16:06 +0200, Krzysztof Niemiec wrote:
> The test alters the environment, specifically the gt frequencies, for
> its duration. Using the legacy interface however makes it impossible for
> the pre-test state to be restored due to its limitations in multi-gt
> systems; by using the legacy interface only a single value can be set,
> but each gt might accept values from a different range and default to
> a different value that we might want to restore.
>
> Use the per-gt interface instead of the legacy one to properly clean up.
>
> Signed-off-by: Krzysztof Niemiec <krzysztof.niemiec@intel.com>
Reviewed-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
NIT: Please note we have a habit, in IGT and i915 at least, of including
change logs into commit descriptions.
Thanks,
Janusz
> ---
>
> v3:
> - Expand the commit description for extra clarity (Janusz)
>
> v2:
> - Remove the redundant fd open from the i915_for_each_gt loop, as one
> is already created as part of the loop (Sebastian)
> - Remove extra newline (Krzysztof K)
>
> tests/intel/gem_exec_endless.c | 33 +++++++++++++++++----------------
> 1 file changed, 17 insertions(+), 16 deletions(-)
>
> diff --git a/tests/intel/gem_exec_endless.c b/tests/intel/gem_exec_endless.c
> index 7f35e985c..f3ed8556c 100644
> --- a/tests/intel/gem_exec_endless.c
> +++ b/tests/intel/gem_exec_endless.c
> @@ -316,28 +316,28 @@ static void endless_dispatch(int i915, const struct intel_execution_engine2 *e)
> for_each_if(gem_class_can_store_dword(i915, (e)->class)) \
> igt_dynamic_f("%s", (e)->name)
>
> -static void pin_rps(int sysfs)
> +static void pin_rps(int sysfs_gt)
> {
> unsigned int max;
>
> - if (igt_sysfs_scanf(sysfs, "gt_RP0_freq_mhz", "%u", &max) != 1)
> + if (igt_sysfs_scanf(sysfs_gt, "rps_RP0_freq_mhz", "%u", &max) != 1)
> return;
>
> - igt_sysfs_printf(sysfs, "gt_min_freq_mhz", "%u", max);
> - igt_sysfs_printf(sysfs, "gt_max_freq_mhz", "%u", max);
> - igt_sysfs_printf(sysfs, "gt_boost_freq_mhz", "%u", max);
> + igt_sysfs_printf(sysfs_gt, "rps_min_freq_mhz", "%u", max);
> + igt_sysfs_printf(sysfs_gt, "rps_max_freq_mhz", "%u", max);
> + igt_sysfs_printf(sysfs_gt, "rps_boost_freq_mhz", "%u", max);
> }
>
> -static void unpin_rps(int sysfs)
> +static void unpin_rps(int sysfs_gt)
> {
> unsigned int v;
>
> - if (igt_sysfs_scanf(sysfs, "gt_RPn_freq_mhz", "%u", &v) == 1)
> - igt_sysfs_printf(sysfs, "gt_min_freq_mhz", "%u", v);
> + if (igt_sysfs_scanf(sysfs_gt, "rps_RPn_freq_mhz", "%u", &v) == 1)
> + igt_sysfs_printf(sysfs_gt, "rps_min_freq_mhz", "%u", v);
>
> - if (igt_sysfs_scanf(sysfs, "gt_RP0_freq_mhz", "%u", &v) == 1) {
> - igt_sysfs_printf(sysfs, "gt_max_freq_mhz", "%u", v);
> - igt_sysfs_printf(sysfs, "gt_boost_freq_mhz", "%u", v);
> + if (igt_sysfs_scanf(sysfs_gt, "rps_RP0_freq_mhz", "%u", &v) == 1) {
> + igt_sysfs_printf(sysfs_gt, "rps_max_freq_mhz", "%u", v);
> + igt_sysfs_printf(sysfs_gt, "rps_boost_freq_mhz", "%u", v);
> }
> }
>
> @@ -355,7 +355,7 @@ int igt_main()
>
> igt_subtest_group() {
> struct intel_mmio_data mmio;
> - int sysfs;
> + int sysfs_gt, gt;
>
> igt_fixture() {
> igt_require(gem_scheduler_enabled(i915));
> @@ -365,16 +365,17 @@ int igt_main()
> igt_device_get_pci_device(i915),
> false);
>
> - sysfs = igt_sysfs_open(i915);
> - pin_rps(sysfs);
> + i915_for_each_gt(i915, sysfs_gt, gt)
> + pin_rps(sysfs_gt);
> }
>
> test_each_engine("dispatch", i915, e)
> endless_dispatch(i915, e);
>
> igt_fixture() {
> - unpin_rps(sysfs);
> - close(sysfs);
> + i915_for_each_gt(i915, sysfs_gt, gt)
> + unpin_rps(sysfs_gt);
> +
> intel_register_access_fini(&mmio);
> }
> }
next prev parent reply other threads:[~2026-04-10 12:16 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-09 14:06 [PATCH i-g-t v3] tests/intel/gem_exec_endless: Use the per-gt interface to set frequencies Krzysztof Niemiec
2026-04-10 3:03 ` ✓ Xe.CI.BAT: success for tests/intel/gem_exec_endless: Use the per-gt interface to set frequencies (rev3) Patchwork
2026-04-10 3:19 ` ✓ i915.CI.BAT: " Patchwork
2026-04-10 7:15 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-04-10 12:16 ` Janusz Krzysztofik [this message]
2026-04-10 22:39 ` ✗ i915.CI.Full: " Patchwork
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=80fb52072aec29387944507afc63ceda0fc72210.camel@linux.intel.com \
--to=janusz.krzysztofik@linux.intel.com \
--cc=andi.shyti@linux.intel.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=krzysztof.karas@intel.com \
--cc=krzysztof.niemiec@intel.com \
--cc=sebastian.brzezinka@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox