From: "Ghimiray, Himal Prasad" <himal.prasad.ghimiray@intel.com>
To: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>,
<intel-xe@lists.freedesktop.org>
Subject: Re: [Intel-xe] [PATCH v3 2/2] drm/xe: Add CONFIG_DRM_XE_PREEMPT_TIMEOUT
Date: Wed, 16 Aug 2023 21:18:12 +0530 [thread overview]
Message-ID: <439234fc-4dcc-4f60-b71e-32e1be96d1b5@intel.com> (raw)
In-Reply-To: <20230816115238.2340-3-niranjana.vishwanathapura@intel.com>
[-- Attachment #1: Type: text/plain, Size: 4252 bytes --]
Hi Niranjana,
On 16-08-2023 17:22, Niranjana Vishwanathapura wrote:
> Allow preemption timeout to be specified as a config option.
>
> v2: Change unit to microseconds (Tejas)
> v3: Remove get_default_preempt_timeout()
>
> Signed-off-by: Niranjana Vishwanathapura<niranjana.vishwanathapura@intel.com>
> ---
> drivers/gpu/drm/xe/Kconfig.profile | 11 +++++++++++
> drivers/gpu/drm/xe/xe_hw_engine.c | 5 ++++-
> drivers/gpu/drm/xe/xe_hw_engine.h | 5 +++++
> 3 files changed, 20 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/xe/Kconfig.profile b/drivers/gpu/drm/xe/Kconfig.profile
> index 9ad487647ca1..a3655ee6978e 100644
> --- a/drivers/gpu/drm/xe/Kconfig.profile
> +++ b/drivers/gpu/drm/xe/Kconfig.profile
> @@ -22,6 +22,17 @@ config DRM_XE_TIMESLICE_MIN
> help
> Configures the default min timeslice duration between multiple
> contexts by guc scheduling.
> +config DRM_XE_PREEMPT_TIMEOUT
> + int "Preempt timeout (us, jiffy granularity)"
> + default 640000 # microseconds
> + help
> + How long to wait (in microseconds) for a preemption event to occur
> + when submitting a new context. If the current context does not hit
> + an arbitration point and yield to HW before the timer expires, the
> + HW will be reset to allow the more important context to execute.
> +
> + Value should be within the DRM_XE_PREEMPT_TIMEOUT_MIN/MAX limits.
Is there plan to make this config dependent on
CONFIG_DRM_XE_PREEMPT_TIMEOUT_MIN/CONFIG_DRM_XE_PREEMPT_TIMEOUT_MAX ?
If not, value should be within
XE_HW_ENGINE_PREEMPT_TIMEOUT_MIN/XE_HW_ENGINE_PREEMPT_TIMEOUT_MAX.
> + This is adjustable via XE_ENGINE_SET_PROPERTY_PREEMPTION_TIMEOUT.
If IIRC ioctl calls names are changed from XE_ENGINE_* to
XE_EXEC_QUUEUE_*. Please confirm.
And this statement seems misleading, ioctl callwill set the properties
for specific exec_queue and config is meant to be for all exec_queues
unless changed by ioctl.
IMO better description will be "preemption property for exec_queue can
be adjusted by XE_EXEC_QUEUE_SET_PROPERTY_PREEMPTION_TIMEOUT" or we can
skip this
info altogether in kconfig because ioctl description might be already
covering it.
BR
Himal Ghimiray
> config DRM_XE_PREEMPT_TIMEOUT_MAX
> int "Default max preempt timeout (us)"
> default 10000000 # microseconds
> diff --git a/drivers/gpu/drm/xe/xe_hw_engine.c b/drivers/gpu/drm/xe/xe_hw_engine.c
> index c44540684462..4c812d04e182 100644
> --- a/drivers/gpu/drm/xe/xe_hw_engine.c
> +++ b/drivers/gpu/drm/xe/xe_hw_engine.c
> @@ -370,7 +370,7 @@ static void hw_engine_init_early(struct xe_gt *gt, struct xe_hw_engine *hwe,
> hwe->eclass->sched_props.timeslice_us = 1 * 1000;
> hwe->eclass->sched_props.timeslice_min = XE_HW_ENGINE_TIMESLICE_MIN;
> hwe->eclass->sched_props.timeslice_max = XE_HW_ENGINE_TIMESLICE_MAX;
> - hwe->eclass->sched_props.preempt_timeout_us = 640 * 1000;
> + hwe->eclass->sched_props.preempt_timeout_us = XE_HW_ENGINE_PREEMPT_TIMEOUT;
> hwe->eclass->sched_props.preempt_timeout_min = XE_HW_ENGINE_PREEMPT_TIMEOUT_MIN;
> hwe->eclass->sched_props.preempt_timeout_max = XE_HW_ENGINE_PREEMPT_TIMEOUT_MAX;
> /* Record default props */
> @@ -562,6 +562,9 @@ int xe_hw_engines_init_early(struct xe_gt *gt)
> read_copy_fuses(gt);
> read_compute_fuses(gt);
>
> + BUILD_BUG_ON(XE_HW_ENGINE_PREEMPT_TIMEOUT < XE_HW_ENGINE_PREEMPT_TIMEOUT_MIN);
> + BUILD_BUG_ON(XE_HW_ENGINE_PREEMPT_TIMEOUT > XE_HW_ENGINE_PREEMPT_TIMEOUT_MAX);
> +
> for (i = 0; i < ARRAY_SIZE(gt->hw_engines); i++)
> hw_engine_init_early(gt, >->hw_engines[i], i);
>
> diff --git a/drivers/gpu/drm/xe/xe_hw_engine.h b/drivers/gpu/drm/xe/xe_hw_engine.h
> index 3d37d6d44261..71968ee2f600 100644
> --- a/drivers/gpu/drm/xe/xe_hw_engine.h
> +++ b/drivers/gpu/drm/xe/xe_hw_engine.h
> @@ -30,6 +30,11 @@ struct drm_printer;
> #else
> #define XE_HW_ENGINE_TIMESLICE_MAX (10 * 1000 * 1000)
> #endif
> +#ifdef CONFIG_DRM_XE_PREEMPT_TIMEOUT
> +#define XE_HW_ENGINE_PREEMPT_TIMEOUT CONFIG_DRM_XE_PREEMPT_TIMEOUT
> +#else
> +#define XE_HW_ENGINE_PREEMPT_TIMEOUT (640 * 1000)
> +#endif
> #ifdef CONFIG_DRM_XE_PREEMPT_TIMEOUT_MIN
> #define XE_HW_ENGINE_PREEMPT_TIMEOUT_MIN CONFIG_DRM_XE_PREEMPT_TIMEOUT_MIN
> #else
[-- Attachment #2: Type: text/html, Size: 5384 bytes --]
next prev parent reply other threads:[~2023-08-16 15:48 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-16 11:52 [Intel-xe] [PATCH v3 0/2] drm/xe: Add preemption timeout config options Niranjana Vishwanathapura
2023-08-16 11:52 ` [Intel-xe] [PATCH v3 1/2] drm/xe: Simplify engine class sched_props setting Niranjana Vishwanathapura
2023-08-16 11:52 ` [Intel-xe] [PATCH v3 2/2] drm/xe: Add CONFIG_DRM_XE_PREEMPT_TIMEOUT Niranjana Vishwanathapura
2023-08-16 13:57 ` Upadhyay, Tejas
2023-08-16 15:48 ` Ghimiray, Himal Prasad [this message]
2023-08-16 15:57 ` Niranjana Vishwanathapura
2023-08-16 16:30 ` Ghimiray, Himal Prasad
2023-08-16 16:41 ` Niranjana Vishwanathapura
2023-08-16 13:05 ` [Intel-xe] ✓ CI.Patch_applied: success for drm/xe: Add preemption timeout config options Patchwork
2023-08-16 13:05 ` [Intel-xe] ✗ CI.checkpatch: warning " Patchwork
2023-08-16 13:06 ` [Intel-xe] ✓ CI.KUnit: success " Patchwork
2023-08-16 13:10 ` [Intel-xe] ✓ CI.Build: " Patchwork
2023-08-16 13:10 ` [Intel-xe] ✓ CI.Hooks: " Patchwork
2023-08-16 13:11 ` [Intel-xe] ✗ CI.checksparse: warning " 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=439234fc-4dcc-4f60-b71e-32e1be96d1b5@intel.com \
--to=himal.prasad.ghimiray@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=niranjana.vishwanathapura@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