From: "Dixit, Ashutosh" <ashutosh.dixit@intel.com>
To: Shekhar Chauhan <shekhar.chauhan@intel.com>
Cc: <igt-dev@lists.freedesktop.org>
Subject: Re: [PATCH i-g-t v3] tools/xe-perf-recorder: expose some OA stream open properties
Date: Thu, 06 Aug 2026 11:16:55 -0700 [thread overview]
Message-ID: <87fr0rf71k.wl-ashutosh.dixit@intel.com> (raw)
In-Reply-To: <20260806063402.4003061-1-shekhar.chauhan@intel.com>
On Wed, 05 Aug 2026 23:34:02 -0700, Shekhar Chauhan wrote:
>
> xe-perf-recorder only passes 5 of the OA stream open properties to
> DRM_XE_OBSERVATION_OP_STREAM_OPEN, leaving the rest without an open
> source consumer. Add few of the remaining ones.
>
> v2: Remove properties which don't exercise anything in xe-perf-recorder.
> v3: Minor nitpicks.
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
>
> Signed-off-by: Shekhar Chauhan <shekhar.chauhan@intel.com>
> ---
> tools/xe-perf/xe_perf_recorder.c | 32 ++++++++++++++++++++++++++++++--
> 1 file changed, 30 insertions(+), 2 deletions(-)
>
> diff --git a/tools/xe-perf/xe_perf_recorder.c b/tools/xe-perf/xe_perf_recorder.c
> index c69050b43..e13eef519 100644
> --- a/tools/xe-perf/xe_perf_recorder.c
> +++ b/tools/xe-perf/xe_perf_recorder.c
> @@ -364,6 +364,9 @@ struct recording_context {
> struct drm_xe_oa_unit *oa_unit;
> struct drm_xe_engine_class_instance *hwe;
>
> + int oa_buffer_size;
> + int wait_num_reports;
> +
> uint32_t vm;
> uint32_t exec_queue;
> struct intel_bb *ibb;
> @@ -487,6 +490,12 @@ perf_open(struct recording_context *ctx)
> DRM_XE_OA_PROPERTY_OA_METRIC_SET, ctx->metric_set->perf_oa_metrics_set,
> DRM_XE_OA_PROPERTY_OA_FORMAT, __ff(ctx->metric_set->perf_oa_format),
> DRM_XE_OA_PROPERTY_OA_PERIOD_EXPONENT, ctx->oa_exponent,
> +
> + DRM_XE_OA_PROPERTY_OA_BUFFER_SIZE, ctx->oa_buffer_size,
> + DRM_XE_OA_PROPERTY_WAIT_NUM_REPORTS, ctx->wait_num_reports,
> +
> + /* Open disabled; the stream is enabled right after open */
> + DRM_XE_OA_PROPERTY_OA_DISABLED, true,
> };
> struct intel_xe_oa_open_prop param = {
> .num_properties = ARRAY_SIZE(properties) / 2,
> @@ -924,7 +933,10 @@ usage(const char *name)
> " --output, -o <path> Output file (default = xe_perf.record)\n"
> " --cpu-clock, -k <path> Cpu clock to use for correlations\n"
> " Values: boot, mono, mono_raw (default = mono)\n"
> - " --oa-unit-id -u <value> OA unit id for the capture.\n",
> + " --oa-unit-id -u <value> OA unit id for the capture.\n"
> + " --oa-buffer-size -b <value> OA buffer size in bytes (default = 64M)\n"
> + " --wait-num-reports -w <value> Num reports before unblocking poll()/read()\n"
> + " (default = 1)\n",
> name);
> }
>
> @@ -1038,6 +1050,8 @@ main(int argc, char *argv[])
> {"command-fifo", required_argument, 0, 'f'},
> {"cpu-clock", required_argument, 0, 'k'},
> {"oa-unit-id", required_argument, 0, 'u'},
> + {"oa-buffer-size", required_argument, 0, 'b'},
> + {"wait-num-reports", required_argument, 0, 'w'},
> {0, 0, 0, 0}
> };
> const struct {
> @@ -1066,9 +1080,11 @@ main(int argc, char *argv[])
> .command_fifo_fd = -1,
>
> .oa_unit_id = 0,
> + .oa_buffer_size = 64 * 1024 * 1024,
> + .wait_num_reports = 1,
> };
>
> - while ((opt = getopt_long(argc, argv, "hc:d:p:m:Co:s:f:k:P:u:", long_options, NULL)) != -1) {
> + while ((opt = getopt_long(argc, argv, "hc:d:p:m:Co:s:f:k:P:u:b:w:", long_options, NULL)) != -1) {
> switch (opt) {
> case 'h':
> usage(argv[0]);
> @@ -1119,6 +1135,12 @@ main(int argc, char *argv[])
> case 'u':
> ctx.oa_unit_id = atoi(optarg);
> break;
> + case 'b':
> + ctx.oa_buffer_size = atoi(optarg);
> + break;
> + case 'w':
> + ctx.wait_num_reports = atoi(optarg);
> + break;
> default:
> fprintf(stderr, "Internal error: "
> "unexpected getopt value: %d\n", opt);
> @@ -1291,6 +1313,12 @@ main(int argc, char *argv[])
> goto fail;
> }
>
> + if (perf_ioctl(ctx.perf_fd, DRM_XE_OBSERVATION_IOCTL_ENABLE, 0) < 0) {
> + fprintf(stderr, "Unable to enable xe oa stream: %s\n",
> + strerror(errno));
> + goto fail;
> + }
> +
> init_mmio_trigger_ctx(&ctx);
> emit_oa_trigger(&ctx, 0xc0ffee01);
>
> --
> 2.53.0
>
next prev parent reply other threads:[~2026-08-06 18:17 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-06 6:34 [PATCH i-g-t v3] tools/xe-perf-recorder: expose some OA stream open properties Shekhar Chauhan
2026-08-06 18:16 ` Dixit, Ashutosh [this message]
2026-08-06 20:49 ` ✓ Xe.CI.BAT: success for tools/xe-perf-recorder: expose some OA stream open properties (rev3) Patchwork
2026-08-06 21:06 ` ✓ i915.CI.BAT: " Patchwork
2026-08-07 1:59 ` ✓ i915.CI.Full: " Patchwork
2026-08-07 10:11 ` ✓ Xe.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=87fr0rf71k.wl-ashutosh.dixit@intel.com \
--to=ashutosh.dixit@intel.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=shekhar.chauhan@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