Re: [PATCH 02/12] soc: qcom: aoss: Drop power domain support

From: Sibi Sankar <sibis@codeaurora.org>
To: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Stephen Boyd <swboyd@chromium.org>,
	mathieu.poirier@linaro.org, robh+dt@kernel.org,
	ulf.hansson@linaro.org, rjw@rjwysocki.net, agross@kernel.org,
	ohad@wizery.com, linux-arm-msm@vger.kernel.org,
	linux-remoteproc@vger.kernel.org, devicetree@vger.kernel.org,
	linux-kernel@vger.kernel.org, dianders@chromium.org,
	rishabhb@codeaurora.org, sidgup@codeaurora.org
Subject: Re: [PATCH 02/12] soc: qcom: aoss: Drop power domain support
Date: Wed, 02 Jun 2021 09:26:59 +0530	[thread overview]
Message-ID: <7d371c8ac3ad4dbc45739481ec19c62f@codeaurora.org> (raw)
In-Reply-To: <YLBsCLNLBlWwoPQj@builder.lan>

On 2021-05-28 09:35, Bjorn Andersson wrote:
> On Tue 27 Apr 01:25 CDT 2021, Sibi Sankar wrote:
> 
>> On 2021-04-18 07:31, Stephen Boyd wrote:
>> > Quoting Sibi Sankar (2021-04-16 05:03:48)
>> > > The load state resources are expected to follow the life cycle of the
>> > > remote processor it tracks. However, modeling load state resources as
>> > > power-domains result in them getting turned off during system suspend
>> > > and thereby falling out of sync with the remote processors that are
>> > > still
>> > > on. Fix this by replacing load state resource control through the
>> > > generic
>> > > qmp message send interface instead.
>> > >
>> > > Signed-off-by: Sibi Sankar <sibis@codeaurora.org>
>> > > ---
>> >
>> > Is it possible to keep this code around for a cycle so that there isn't
>> > the chance that someone is using the deprecated DT bindings with a new
>> > kernel? I worry that ripping the code out will cause them angst.
>> 
>> deprecated bindings with a newer kernel
>> shouldn't cause any problems since it is
>> the driver changes that make AOSS PD
>> mandatory or not. So the newer kernel will
>> just use qmp_send and leave the PD unused.
>> 
> 
> Maybe I'm missing something in your argument here, but I see two 
> issues:
> * The changes here requires that the new qcom,qmp property is defined,
>   or the qcom_qmp_get() will be unable to find the qmp instance.
> * Between patch 2 and 5 there's no load_state handling.
> 
> Perhaps we can carry the power-domain handling as a fallback i
> qcom_qmp_get() fails, for a few releases?

The load_state implementation is currently
broken i.e. it currently sends that the
remoteproc is down during suspend. AFAIK it
can be safely dropped without side-effects.
I'll respin the series fixing Rob's comments.

> 
> 
> Other than the ordering and backwards compatibility issue I think this
> looks good. So can you please respin this based on the later revision 
> of
> the qmp patch? (And fix Rob's request on the commit message)
> 
> https://lore.kernel.org/linux-arm-msm/1620320818-2206-2-git-send-email-deesin@codeaurora.org/
> 
> Regards,
> Bjorn
> 
>> > Certainly we have to keep the code in place until DT is updated, so this
>> > patch should come last?
>> 
>> sure I don't mind, as long as it simplifies
>> the merge process.
>> 
>> --
>> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
>> a Linux Foundation Collaborative Project.

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project.