Re: [igt-dev] [PATCH i-g-t] tests/kms_dp_dsc: Force a full modeset when we force dsc enable

From: Imre Deak <imre.deak@intel.com>
To: Manasi Navare <manasi.d.navare@intel.com>
Cc: igt-dev@lists.freedesktop.org,
	Anusha Srivatsa <anusha.srivatsa@intel.com>,
	Petri Latvala <petri.latvala@intel.com>
Subject: Re: [igt-dev] [PATCH i-g-t] tests/kms_dp_dsc: Force a full modeset when we force dsc enable
Date: Thu, 4 Apr 2019 02:02:42 +0300	[thread overview]
Message-ID: <20190403230242.GA22054@ideak-desk.fi.intel.com> (raw)
In-Reply-To: <20190403224823.GG17236@intel.com>

On Wed, Apr 03, 2019 at 03:48:24PM -0700, Manasi Navare wrote:
> On Thu, Apr 04, 2019 at 12:42:33AM +0300, Imre Deak wrote:
> > On Wed, Apr 03, 2019 at 01:49:44PM -0700, Manasi Navare wrote:
> > > [...]
> > > Currently we have inh i915_dsc_fec_support_show:
> > > DSC_enabled:
> > > DSC_sink_support:
> > > FEC_sink_support:
> > 
> > Yes, and after the kernel change you'll have an additional
> >   
> >   Force DSC enable: yes/no
> > 
> > line here.
> > 
> > > And this node we open as a read only first and then as write node and
> > > force_dsc_enable() from IGT just writes a 1 to this, so you are saying
> > > that in the kernel _write() node, I accept this and write to
> > > force_dsc_enable entry in this file instead of rewriting the whole
> > > file?
> > 
> > I'm not sure what you mean. Currently "1" or "0" is written to the
> > debugfs file which will set intel_dp->force_dsc_enable. This doesn't
> > need to be changed at all, no need to rewrite the whole file.
> > 
> > > So in IGT also I will have to parse the file and write 1 for
> > > force_dsc_enable entry?
> > 
> > You can get the original value of it similarly to how it's done already
> > for 'DSC_Sink_Support', by reading the file and doing
> > 
> > 	strstr(buf, "Force DSC enable: yes");
> > 
> > on the content.
> > 
> > > I still cant figure out how the restore() function work, we write the
> > > stored_values and write for each of the entries in the file? And then
> > > in the kernel debugfs_write, get these and write back to each entry?
> > 
> > Only force_dsc_enable need to be restored, the only value that you
> > change by writing to the debugfs file. All the rest are calculated
> > values that you don't need to care about.
> 
> But now that I rething then, force_dsc_enable is always 0, it s a variable
> that we use to force DSC only if IGT asks to, so for test we force it to 1
> and then after the test always clear it since the original value is always 0.
> So why do we need a separate restore() for just force_dsc_enable if the value
> is always 0.

It's not necessarily 0. Before the test starts someone could've written
1 to that file.

> 
> Manasi
> 
> > 
> > --Imre
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev