[v2] drm/i915/selftest/gsc: Ensure GSC Proxy init completes before selftests

[Dong, Zhanjun]

> -----Original Message-----
> From: Teres Alexis, Alan Previn <alan.previn.teres.alexis at intel.com>
> Sent: June 8, 2023 2:31 PM
> To: Dong, Zhanjun <zhanjun.dong at intel.com>; intel-
> gfx at lists.freedesktop.org
> Cc: dri-devel at lists.freedesktop.org
> Subject: Re: [v2] drm/i915/selftest/gsc: Ensure GSC Proxy init completes
> before selftests
> 
> On Thu, 2023-06-08 at 18:14 +0000, Dong, Zhanjun wrote:
> > See my comments below.
> >
> > > -----Original Message-----
> > > From: Alan Previn <alan.previn.teres.alexis at intel.com>
> alan:snip
> 
> > > +static int
> > > +__wait_gsc_proxy_completed(struct drm_i915_private *i915,
> > > +			   unsigned long timeout_ms)
> > > +{
> > > +	bool need_to_wait = (IS_ENABLED(CONFIG_INTEL_MEI_GSC_PROXY)
> > > &&
> > > +			     i915->media_gt &&
> > > +			     HAS_ENGINE(i915->media_gt, GSC0) &&
> > > +			     intel_uc_fw_is_loadable(&i915->media_gt-
> > > > uc.gsc.fw));
> > > +
> > > +	/*
> > > +	 * For gsc proxy component loading + init, we need a much longer
> > > timeout
> > > +	 * than what CI selftest infrastrucutre currently uses. This longer wait
> > > +	 * period depends on the kernel config and component driver load
> > > ordering
> > > +	 */
> > > +	if (timeout_ms < 8000)
> > > +		timeout_ms = 8000;
> >
> >
> > Lgtm, just an concern about the fixed number here, shall we set the
> minimal here, or let i915_selftest.timeout_ms take control? Thus no longer
> need coding change here in the future.
> >
> > Reviewed-by: Zhanjun Dong <zhanjun.dong at intel.com>
> 
> Thanks Zhanjun, unfortunately, based on internal testing,
> i915_selftest.timeout_ms default is too
> low that it does occasionally timeout for CI. From experience, with a lean
> ubuntu config, it typically
> takes about ~1 seconds for the mei-gsc-sw-proxy component driver to load
> after i915 loads.
> Since CI regular unloads and reloads i915, the timeout observed ends up
> being reported as issue.
> 
> 8 seconds was based on internal testing of the worst case scenario - which
> hardly ever happens.
> We've only seen the 8 second happen when the kernel config has configs
> enabled for very many SOC IP
> drivers and component driver (seen one at least one customer config) or if
> the MTL board IFWI was only
> just reflashed (this would be a one-off 8 seconds, we suspect due to the
> firmware doing additional steps)
> 

Thanks for detailed info. The i915_selftest.timeout_ms is too low for this test, so it need an special minimal for itself, valid reason.
The concern I raised is at minor level. Looks good to me.

Regards,
Zhanjun