Re: [PATCH 5/9] drm/i915: Enable i915 perf stream for Haswell OA unit

[Robert Bragg <robert@xxxxxxxxxxxxx>]

On Wed, Apr 20, 2016 at 10:11 PM, Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> wrote:

On Wed, Apr 20, 2016 at 03:23:10PM +0100, Robert Bragg wrote:
> +static void gen7_update_oacontrol_locked(struct drm_i915_private *dev_priv)
> +{
> +     assert_spin_locked(&dev_priv->perf.hook_lock);
> +
> +     if (dev_priv->perf.oa.exclusive_stream->enabled) {
> +             unsigned long ctx_id = 0;
> +
> +             if (dev_priv->perf.oa.exclusive_stream->ctx)
> +                     ctx_id = dev_priv->perf.oa.specific_ctx_id;
> +
> +             if (dev_priv->perf.oa.exclusive_stream->ctx == NULL || ctx_id) {
> +                     bool periodic = dev_priv->perf.oa.periodic;
> +                     u32 period_exponent = dev_priv->perf.oa.period_exponent;
> +                     u32 report_format = dev_priv->perf.oa.oa_buffer.format;
> +
> +                     I915_WRITE(GEN7_OACONTROL,
> +                                (ctx_id & GEN7_OACONTROL_CTX_MASK) |
> +                                (period_exponent <<
> +                                 GEN7_OACONTROL_TIMER_PERIOD_SHIFT) |
> +                                (periodic ?
> +                                 GEN7_OACONTROL_TIMER_ENABLE : 0) |
> +                                (report_format <<
> +                                 GEN7_OACONTROL_FORMAT_SHIFT) |
> +                                (ctx_id ?
> +                                 GEN7_OACONTROL_PER_CTX_ENABLE : 0) |
> +                                GEN7_OACONTROL_ENABLE);

So this works by only recording when the OACONTROL context address
matches the CCID.

Rather than hooking into switch context and checking every batch whether
you have the exclusive context in case it changed address, you could
just pin the exclusive context when told by the user to bind perf to
that context. Then it will also have the same address until oa is
finished (and releases it vma pin).

Yeah, this was the approach I first went with when the driver was perf based, though we ended up deciding to got with hooking into pinning and updating the OA state in the end.

E.g. for reference: https://lists.freedesktop.org/archives/intel-gfx/2014-November/055385.html (wow, sad face after seeing how long I've been kicking this stuff)

I'd prefer to stick with this approach now, unless you see a big problem with it.

Regards,

- Robert

 

-Chris

--
Chris Wilson, Intel Open Source Technology Centre

_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx