From: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
To: Riana Tauro <riana.tauro@intel.com>
Cc: <intel-xe@lists.freedesktop.org>, <anshuman.gupta@intel.com>,
<rodrigo.vivi@intel.com>, <vinay.belgaumkar@intel.com>,
<aravind.iddamsetty@intel.com>, <john.c.harrison@intel.com>,
<ashutosh.dixit@intel.com>, <soham.purkait@intel.com>
Subject: Re: [PATCH 6/7] drm/xe/guc: Expose engine busyness only for supported GuC version
Date: Mon, 18 Nov 2024 15:32:31 -0800 [thread overview]
Message-ID: <ZzvOj3lFfaGnL0/K@orsosgc001> (raw)
In-Reply-To: <8ceea4a6-96d2-432d-8375-73d968d068c4@intel.com>
On Mon, Nov 18, 2024 at 01:07:32PM +0530, Riana Tauro wrote:
>Hi Umesh
>
>On 11/15/2024 5:42 AM, Umesh Nerlige Ramappa wrote:
>>On Wed, Nov 13, 2024 at 10:25:48AM +0530, Riana Tauro wrote:
>>>Engine busyness is supported only on GuC versions >= 70.32.0
>>>Allow enabling/reading engine busyness only on supported
>>>GuC versions. Warn once if not supported.
>>>
>>>Signed-off-by: Riana Tauro <riana.tauro@intel.com>
>>>---
>>>drivers/gpu/drm/xe/xe_engine_activity.c | 22 ++++++++++++++++++++++
>>>1 file changed, 22 insertions(+)
>>>
>>>diff --git a/drivers/gpu/drm/xe/xe_engine_activity.c
>>>b/drivers/gpu/ drm/xe/xe_engine_activity.c
>>>index 7a2ae09d23de..fc34dcf3057a 100644
>>>--- a/drivers/gpu/drm/xe/xe_engine_activity.c
>>>+++ b/drivers/gpu/drm/xe/xe_engine_activity.c
>>>@@ -236,6 +236,16 @@ static u32 gpm_timestamp_shift(struct xe_gt *gt)
>>> return 3 - REG_FIELD_GET(RPM_CONFIG0_CTC_SHIFT_PARAMETER_MASK, reg);
>>>}
>>>
>>>+static bool guc_engine_activity_supported(struct xe_guc *guc)
>>>+{
>>>+ if (GUC_FIRMWARE_VER(guc) >= MAKE_GUC_VER(70, 32, 0))
>>>+ return true;
>>>+
>>>+ drm_WARN_ON_ONCE(&guc_to_xe(guc)->drm,
>>>+ "Engine busyness not supported in this GuC version\n");
>>>+ return false;
>>>+}
>>>+
>>>/**
>>> * xe_engine_activity_get_active_ticks - Get engine active ticks
>>> * @hwe: The hw_engine object
>>>@@ -246,6 +256,9 @@ u64 xe_engine_activity_get_active_ticks(struct
>>>xe_hw_engine *hwe)
>>>{
>>> struct xe_guc *guc = &hwe->gt->uc.guc;
>>>
>>>+ if (!guc_engine_activity_supported(guc))
>>>+ return 0;
>>>+
>>> return get_engine_active_ticks(guc, hwe);
>>>}
>>>
>>>@@ -259,6 +272,9 @@ u64 xe_engine_activity_get_total_ticks(struct
>>>xe_hw_engine *hwe)
>>>{
>>> struct xe_guc *guc = &hwe->gt->uc.guc;
>>>
>>>+ if (!guc_engine_activity_supported(guc))
>>
>>Hmm, not sure my response to this landed on the ML. Sending again..
>>
>>If GuC is an older version and UMDs have implemented the new
>>interface, we could return CPU time in ticks here so that UMDs will
>>still be able to use something in the denominator instead of a
>>divide by zero. Busy value will still be 0. Thoughts?
>We could do this. But shouldn't UMD only read total ticks if they see
>an actual busyness? Busyness % will anyway be 0 even if we return the
>cpu time in ticks
Well, I think we shouldn't have to impose that condition on the UMDs -
read total ticks only if engine busy ticks is non-zero. At the same time
they should not use a zero delta in the denominator anyways, so you can
ignore this for now. We can add it if there's an issue later.
Thanks,
Umesh
>
>Thanks
>Riana Tauro
>>
>>Umesh
>>
>>>+ return 0;
>>>+
>>> return get_engine_total_ticks(guc, hwe);
>>>}
>
next prev parent reply other threads:[~2024-11-18 23:32 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-11-13 4:55 [PATCH 0/7] Add PMU support for single engine busyness Riana Tauro
2024-11-13 4:55 ` [PATCH 1/7] [DO NOT REVIEW] drm/xe/pmu: Enable PMU interface Riana Tauro
2024-11-13 4:55 ` [PATCH 2/7] [DO NOT REVIEW] drm/xe/pmu: Add GT C6 events Riana Tauro
2024-11-13 4:55 ` [PATCH 3/7] [DO NOT REVIEW] drm/xe/pmu: Add GT frequency events Riana Tauro
2024-11-13 4:55 ` [PATCH 4/7] drm/xe: add function to convert xe hw engine class to user class Riana Tauro
2024-11-14 16:14 ` Umesh Nerlige Ramappa
2024-11-15 7:18 ` Riana Tauro
2024-11-14 17:46 ` Rodrigo Vivi
2024-11-15 7:21 ` Riana Tauro
2024-11-13 4:55 ` [PATCH 5/7] drm/xe: Add single engine busyness support Riana Tauro
2024-11-15 0:08 ` Umesh Nerlige Ramappa
2024-11-18 7:33 ` Riana Tauro
2024-11-13 4:55 ` [PATCH 6/7] drm/xe/guc: Expose engine busyness only for supported GuC version Riana Tauro
2024-11-14 21:12 ` Umesh Nerlige Ramappa
2024-11-15 0:12 ` Umesh Nerlige Ramappa
2024-11-18 7:37 ` Riana Tauro
2024-11-18 23:32 ` Umesh Nerlige Ramappa [this message]
2024-11-13 4:55 ` [PATCH 7/7] drm/xe/pmu: Add PMU support for engine busyness Riana Tauro
2024-11-13 14:41 ` ✓ CI.Patch_applied: success for Add PMU support for single " Patchwork
2024-11-13 14:42 ` ✗ CI.checkpatch: warning " Patchwork
2024-11-13 14:43 ` ✓ CI.KUnit: success " Patchwork
2024-11-13 14:55 ` ✓ CI.Build: " Patchwork
2024-11-13 14:57 ` ✗ CI.Hooks: failure " Patchwork
2024-11-13 14:59 ` ✓ CI.checksparse: success " Patchwork
2024-11-13 15:19 ` ✗ CI.BAT: failure " Patchwork
2024-11-13 21:22 ` ✓ CI.FULL: success " 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=ZzvOj3lFfaGnL0/K@orsosgc001 \
--to=umesh.nerlige.ramappa@intel.com \
--cc=anshuman.gupta@intel.com \
--cc=aravind.iddamsetty@intel.com \
--cc=ashutosh.dixit@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=john.c.harrison@intel.com \
--cc=riana.tauro@intel.com \
--cc=rodrigo.vivi@intel.com \
--cc=soham.purkait@intel.com \
--cc=vinay.belgaumkar@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