Date:   Tue, 7 Jun 2022 09:02:00 -0700
From:   Rob Clark <robdclark@...il.com>
To:     Tvrtko Ursulin <tvrtko.ursulin@...ux.intel.com>
Cc:     dri-devel <dri-devel@...ts.freedesktop.org>,
        freedreno <freedreno@...ts.freedesktop.org>,
        linux-arm-msm <linux-arm-msm@...r.kernel.org>,
        Rob Clark <robdclark@...omium.org>,
        Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
        Maxime Ripard <mripard@...nel.org>,
        Thomas Zimmermann <tzimmermann@...e.de>,
        David Airlie <airlied@...ux.ie>,
        Daniel Vetter <daniel@...ll.ch>,
        Jonathan Corbet <corbet@....net>,
        Abhinav Kumar <quic_abhinavk@...cinc.com>,
        Dmitry Baryshkov <dmitry.baryshkov@...aro.org>,
        Sean Paul <sean@...rly.run>,
        "open list:DOCUMENTATION" <linux-doc@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2 2/2] drm/msm: Expose client engine utilization via fdinfo

On Tue, Jun 7, 2022 at 1:56 AM Tvrtko Ursulin
<tvrtko.ursulin@...ux.intel.com> wrote:
>
>
> On 06/06/2022 20:54, Rob Clark wrote:
> > From: Rob Clark <robdclark@...omium.org>
> >
> > Similar to AMD commit
> > 874442541133 ("drm/amdgpu: Add show_fdinfo() interface"), using the
> > infrastructure added in previous patches, we add basic client info
> > and GPU engine utilisation for msm.
> >
> > Example output:
> >
> >       # cat /proc/`pgrep glmark2`/fdinfo/6
> >       pos:    0
> >       flags:  02400002
> >       mnt_id: 21
> >       ino:    162
> >       drm-driver:     msm
> >       drm-client-id:  7
> >       drm-engine-gpu: 1734371319 ns
> >       drm-cycles-gpu: 1153645024
> >       drm-maxfreq-gpu:        800000000 Hz
> >
> > See also: https://patchwork.freedesktop.org/patch/468505/
> >
> > Signed-off-by: Rob Clark <robdclark@...omium.org>
> > ---
> >   Documentation/gpu/drm-usage-stats.rst | 21 +++++++++++++++++++++
> >   drivers/gpu/drm/msm/msm_drv.c         | 19 ++++++++++++++++++-
> >   drivers/gpu/drm/msm/msm_gpu.c         | 21 +++++++++++++++++++--
> >   drivers/gpu/drm/msm/msm_gpu.h         | 19 +++++++++++++++++++
> >   4 files changed, 77 insertions(+), 3 deletions(-)
> >
> > diff --git a/Documentation/gpu/drm-usage-stats.rst b/Documentation/gpu/drm-usage-stats.rst
> > index 6c9f166a8d6f..60e5cc9c13ad 100644
> > --- a/Documentation/gpu/drm-usage-stats.rst
> > +++ b/Documentation/gpu/drm-usage-stats.rst
> > @@ -105,6 +105,27 @@ object belong to this client, in the respective memory region.
> >   Default unit shall be bytes with optional unit specifiers of 'KiB' or 'MiB'
> >   indicating kibi- or mebi-bytes.
> >
> > +- drm-cycles-<str> <uint>
> > +
> > +Engine identifier string must be the same as the one specified in the
> > +drm-engine-<str> tag and shall contain the number of busy cycles for the given
> > +engine.
> > +
> > +Values are not required to be constantly monotonic if it makes the driver
> > +implementation easier, but are required to catch up with the previously reported
> > +larger value within a reasonable period. Upon observing a value lower than what
> > +was previously read, userspace is expected to stay with that larger previous
> > +value until a monotonic update is seen.
> > +
> > +- drm-maxfreq-<str> <uint> [Hz|MHz|KHz]
> > +
> > +Engine identifier string must be the same as the one specified in the
> > +drm-engine-<str> tag and shall contain the maxium frequence for the given
>
> maximum frequency
>
> > +engine.  Taken together with drm-cycles-<str>, this can be used to calculate
> > +percentage utilization of the engine, whereas drm-engine-<str> only refects
>
> reflects
>
> > +time active without considering what frequency the engine is operating as a
> > +percentage of it's maximum frequency.
>
> Cycles vs max freq sounds very useful. My reservations is that how come
> the idea hasn't happened in the CPU world. Or maybe it has and I am
> un-informed?

I do often pay attention to both where tasks get scheduled, and the
individual CPU freq when I'm profiling CPU side stuff (eg. in
perfetto)

I could also report "always-count" cycles, I think, which could be
used by gputop to derive freq.  I'd have to think about that a bit,
since keeping the result monotinic(ish) might be a bit tricky (the hw
counter loses state across runtime suspend)

> In any case, if going with this I think we need to clarify the text that
> the value should reflect the current soft limit, where the driver
> supports that, in case it has been set to lower than the maximum
> frequency hardware can support. I am thinking about avoiding "my gpu
> cannot hit 100%" support incidents in cases when user/admin lowered the
> soft limit for some reason. Possibly does not apply to msm but can apply
> to i915, if we decided to export the same data.

Yes, with pm-qos thermal or userspace could limit the max freq.. but
we also internally use a pm-qos constraint to reduce freq when the GPU
is idle, and I don't think there is a good way to differentiate
*which* constraint is which.  I'll add something involving the word
"recommended" ;-)

BR,
-R

>
> No other gotchas come to mind at the moment.
>
> Regards,
>
> Tvrtko
>
> > +
> >   ===============================
> >   Driver specific implementations
> >   ===============================
> > diff --git a/drivers/gpu/drm/msm/msm_drv.c b/drivers/gpu/drm/msm/msm_drv.c
> > index 14ab9a627d8b..57a66093e671 100644
> > --- a/drivers/gpu/drm/msm/msm_drv.c
> > +++ b/drivers/gpu/drm/msm/msm_drv.c
> > @@ -948,7 +948,24 @@ static const struct drm_ioctl_desc msm_ioctls[] = {
> >       DRM_IOCTL_DEF_DRV(MSM_SUBMITQUEUE_QUERY, msm_ioctl_submitqueue_query, DRM_RENDER_ALLOW),
> >   };
> >
> > -DEFINE_DRM_GEM_FOPS(fops);
> > +static void msm_fop_show_fdinfo(struct seq_file *m, struct file *f)
> > +{
> > +     struct drm_file *file = f->private_data;
> > +     struct drm_device *dev = file->minor->dev;
> > +     struct msm_drm_private *priv = dev->dev_private;
> > +     struct drm_printer p = drm_seq_file_printer(m);
> > +
> > +     if (!priv->gpu)
> > +             return;
> > +
> > +     msm_gpu_show_fdinfo(priv->gpu, file->driver_priv, &p);
> > +}
> > +
> > +static const struct file_operations fops = {
> > +     .owner = THIS_MODULE,
> > +     DRM_GEM_FOPS,
> > +     .show_fdinfo = msm_fop_show_fdinfo,
> > +};
> >
> >   static const struct drm_driver msm_driver = {
> >       .driver_features    = DRIVER_GEM |
> > diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c
> > index eb8a6663f309..333a9a299b41 100644
> > --- a/drivers/gpu/drm/msm/msm_gpu.c
> > +++ b/drivers/gpu/drm/msm/msm_gpu.c
> > @@ -4,6 +4,8 @@
> >    * Author: Rob Clark <robdclark@...il.com>
> >    */
> >
> > +#include "drm/drm_drv.h"
> > +
> >   #include "msm_gpu.h"
> >   #include "msm_gem.h"
> >   #include "msm_mmu.h"
> > @@ -146,6 +148,16 @@ int msm_gpu_pm_suspend(struct msm_gpu *gpu)
> >       return 0;
> >   }
> >
> > +void msm_gpu_show_fdinfo(struct msm_gpu *gpu, struct msm_file_private *ctx,
> > +                      struct drm_printer *p)
> > +{
> > +     drm_printf(p, "drm-driver:\t%s\n", gpu->dev->driver->name);
> > +     drm_printf(p, "drm-client-id:\t%u\n", ctx->seqno);
> > +     drm_printf(p, "drm-engine-gpu:\t%llu ns\n", ctx->elapsed_ns);
> > +     drm_printf(p, "drm-cycles-gpu:\t%llu\n", ctx->cycles);
> > +     drm_printf(p, "drm-maxfreq-gpu:\t%lu Hz\n", gpu->fast_rate);
> > +}
> > +
> >   int msm_gpu_hw_init(struct msm_gpu *gpu)
> >   {
> >       int ret;
> > @@ -652,7 +664,7 @@ static void retire_submit(struct msm_gpu *gpu, struct msm_ringbuffer *ring,
> >   {
> >       int index = submit->seqno % MSM_GPU_SUBMIT_STATS_COUNT;
> >       volatile struct msm_gpu_submit_stats *stats;
> > -     u64 elapsed, clock = 0;
> > +     u64 elapsed, clock = 0, cycles;
> >       unsigned long flags;
> >
> >       stats = &ring->memptrs->stats[index];
> > @@ -660,12 +672,17 @@ static void retire_submit(struct msm_gpu *gpu, struct msm_ringbuffer *ring,
> >       elapsed = (stats->alwayson_end - stats->alwayson_start) * 10000;
> >       do_div(elapsed, 192);
> >
> > +     cycles = stats->cpcycles_end - stats->cpcycles_start;
> > +
> >       /* Calculate the clock frequency from the number of CP cycles */
> >       if (elapsed) {
> > -             clock = (stats->cpcycles_end - stats->cpcycles_start) * 1000;
> > +             clock = cycles * 1000;
> >               do_div(clock, elapsed);
> >       }
> >
> > +     submit->queue->ctx->elapsed_ns += elapsed;
> > +     submit->queue->ctx->cycles     += cycles;
> > +
> >       trace_msm_gpu_submit_retired(submit, elapsed, clock,
> >               stats->alwayson_start, stats->alwayson_end);
> >
> > diff --git a/drivers/gpu/drm/msm/msm_gpu.h b/drivers/gpu/drm/msm/msm_gpu.h
> > index 6def00883046..4911943ba53b 100644
> > --- a/drivers/gpu/drm/msm/msm_gpu.h
> > +++ b/drivers/gpu/drm/msm/msm_gpu.h
> > @@ -361,6 +361,22 @@ struct msm_file_private {
> >       /** cmdline: Overridden task cmdline, see MSM_PARAM_CMDLINE */
> >       char *cmdline;
> >
> > +     /**
> > +      * elapsed:
> > +      *
> > +      * The total (cumulative) elapsed time GPU was busy with rendering
> > +      * from this context in ns.
> > +      */
> > +     uint64_t elapsed_ns;
> > +
> > +     /**
> > +      * cycles:
> > +      *
> > +      * The total (cumulative) GPU cycles elapsed attributed to this
> > +      * context.
> > +      */
> > +     uint64_t cycles;
> > +
> >       /**
> >        * entities:
> >        *
> > @@ -544,6 +560,9 @@ static inline void gpu_write64(struct msm_gpu *gpu, u32 lo, u32 hi, u64 val)
> >   int msm_gpu_pm_suspend(struct msm_gpu *gpu);
> >   int msm_gpu_pm_resume(struct msm_gpu *gpu);
> >
> > +void msm_gpu_show_fdinfo(struct msm_gpu *gpu, struct msm_file_private *ctx,
> > +                      struct drm_printer *p);
> > +
> >   int msm_submitqueue_init(struct drm_device *drm, struct msm_file_private *ctx);
> >   struct msm_gpu_submitqueue *msm_submitqueue_get(struct msm_file_private *ctx,
> >               u32 id);

Powered by blists - more mailing lists