From: "Wang, X" <x.wang@intel.com>
To: "Zbigniew Kempczyński" <zbigniew.kempczynski@intel.com>,
igt-dev@lists.freedesktop.org
Cc: Matthew Auld <matthew.auld@intel.com>,
Matt Roper <matthew.d.roper@intel.com>
Subject: Re: [PATCH i-g-t v7 4/4] tests/xe_pat: Verify PAT entries contain expected values
Date: Wed, 1 Apr 2026 10:54:14 -0700 [thread overview]
Message-ID: <b19b1853-6f32-4b45-b807-072e8514ed42@intel.com> (raw)
In-Reply-To: <20260401135343.2874239-10-zbigniew.kempczynski@intel.com>
On 4/1/2026 06:53, Zbigniew Kempczyński wrote:
> Debugfs exposes PAT entries for different GTs so verify they
> contain expected values. Verify they are correct also in reset
> and suspend scenarios.
>
> Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
> Cc: Matthew Auld <matthew.auld@intel.com>
> Cc: Matt Roper <matthew.d.roper@intel.com>
> Cc: X Wang <x.wang@intel.com>
> ---
> v4: Test is rewritten to support PAT + PTA + ATS entries.
> v6: Skip on !pf and assert in case of problem with access PAT (Wang)
> v7: Skip on vf
> ---
> tests/intel/xe_pat.c | 85 ++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 85 insertions(+)
>
> diff --git a/tests/intel/xe_pat.c b/tests/intel/xe_pat.c
> index 2e2d7e01e7..81c28677ac 100644
> --- a/tests/intel/xe_pat.c
> +++ b/tests/intel/xe_pat.c
> @@ -19,6 +19,7 @@
> #include "igt_fs.h"
> #include "igt_kmod.h"
> #include "igt_map.h"
> +#include "igt_sriov_device.h"
> #include "igt_syncobj.h"
> #include "igt_sysfs.h"
> #include "igt_vgem.h"
> @@ -238,6 +239,81 @@ static void pat_sanity(int fd)
> }
> }
>
> +enum pat_test_opts {
> + PAT_CHECK = 1,
> + PAT_RESET_GT = 2,
> + PAT_SUSPEND = 3,
> +};
> +
> +/**
> + * SUBTEST: pat-sw-hw-compare
> + * Description: verify debugfs 'pat' reflects 'pat_sw_config'
> + *
> + * SUBTEST: pat-sw-hw-reset-compare
> + * Description: verify debugfs 'pat' reflects 'pat_sw_config' after gt reset
> + *
> + * SUBTEST: pat-sw-hw-suspend
> + * Description: verify debugfs 'pat' reflects 'pat_sw_config' after suspend
> + */
> +static void pat_sw_hw_compare(int fd, enum pat_test_opts opts)
> +{
> + bool matches = true;
> + int gt;
> +
> + igt_skip_on(intel_is_vf_device(fd));
> +
> + xe_for_each_gt(fd, gt) {
> + struct intel_pat_cache pat_hw_config = {};
> + struct intel_pat_cache pat_sw_config = {};
> + int hw_entries, sw_entries;
> +
> + if (opts == PAT_RESET_GT)
> + xe_force_gt_reset_sync(fd, gt);
> + else if (opts == PAT_SUSPEND)
> + igt_system_suspend_autoresume(SUSPEND_STATE_STANDBY, SUSPEND_TEST_NONE);
> +
> + hw_entries = xe_get_pat_hw_config(fd, &pat_hw_config, gt);
> + sw_entries = xe_get_pat_sw_config(fd, &pat_sw_config, gt);
> +
> + igt_debug("[GT%d] hw_entries: %d, sw_entries: %d\n", gt, hw_entries, sw_entries);
> +
> + igt_assert_eq(hw_entries, sw_entries);
> + igt_assert_lt(0, hw_entries);
> + igt_assert_lt(0, sw_entries);
> +
> + for (int i = 0; i < hw_entries; i++) {
> + uint32_t hw_pat, sw_pat;
> +
> + hw_pat = pat_hw_config.entries[i].pat;
> + sw_pat = pat_sw_config.entries[i].pat;
> + if (hw_pat != sw_pat) {
> + igt_debug("[GT%d] Mismatch of pat register vs sw config "
> + "- index: %i, entries: %08x <> %08x\n",
> + gt, i, hw_pat, sw_pat);
> + matches = false;
> + }
> + }
> +
> + /* Check PTA if was explicitly programmed */
> + if (pat_sw_config.pta_mode != UINT32_MAX &&
> + pat_hw_config.pta_mode != pat_sw_config.pta_mode) {
> + igt_debug("[GT%d] Mismatch of PTA_MODE - pta: %x, pta expected: %x\n",
> + gt, pat_hw_config.pta_mode, pat_sw_config.pta_mode);
> + matches = false;
> + }
> +
> + /* Check ATS if was explicitly programmed */
> + if (pat_sw_config.pat_ats != UINT32_MAX &&
> + pat_hw_config.pat_ats != pat_sw_config.pat_ats) {
> + igt_debug("[GT%d] Mismatch of PAT_ATS - ats: %x, ats expected: %x\n",
> + gt, pat_hw_config.pat_ats, pat_sw_config.pat_ats);
> + matches = false;
> + }
> + }
> +
> + igt_assert_eq(matches, true);
> +}
> +
> /**
> * SUBTEST: pat-index-all
> * Test category: functionality test
> @@ -2152,6 +2228,15 @@ int igt_main_args("V", NULL, help_str, opt_handler, NULL)
> igt_subtest("pat-sanity")
> pat_sanity(fd);
>
> + igt_subtest("pat-sw-hw-compare")
> + pat_sw_hw_compare(fd, 0);
> +
> + igt_subtest("pat-sw-hw-reset-compare")
> + pat_sw_hw_compare(fd, PAT_RESET_GT);
> +
> + igt_subtest("pat-sw-hw-suspend")
> + pat_sw_hw_compare(fd, PAT_SUSPEND);
> +
> igt_subtest("pat-index-all")
> pat_index_all(fd);
>
Reviewed-by: Xin Wang <x.wang@intel.com>
next prev parent reply other threads:[~2026-04-01 17:54 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-01 13:53 [PATCH i-g-t v7 0/4] Verify PTA entries contains expected values Zbigniew Kempczyński
2026-04-01 13:53 ` [PATCH i-g-t v7 1/4] lib/intel_pat: Support other than gt0 during parsing debugfs PAT Zbigniew Kempczyński
2026-04-01 13:53 ` [PATCH i-g-t v7 2/4] lib/intel_pat: Expose getter of PAT registers Zbigniew Kempczyński
2026-04-01 13:53 ` [PATCH i-g-t v7 3/4] lib/intel_pat: Set PTA + ATS PAT entries to invalid value Zbigniew Kempczyński
2026-04-01 13:53 ` [PATCH i-g-t v7 4/4] tests/xe_pat: Verify PAT entries contain expected values Zbigniew Kempczyński
2026-04-01 17:54 ` Wang, X [this message]
2026-04-01 19:44 ` ✓ Xe.CI.BAT: success for Verify PTA entries contains expected values (rev7) Patchwork
2026-04-01 19:59 ` ✓ i915.CI.BAT: " Patchwork
2026-04-02 1:31 ` ✓ Xe.CI.FULL: " Patchwork
2026-04-02 14:14 ` ✓ 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=b19b1853-6f32-4b45-b807-072e8514ed42@intel.com \
--to=x.wang@intel.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=matthew.auld@intel.com \
--cc=matthew.d.roper@intel.com \
--cc=zbigniew.kempczynski@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