Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 3 Oct 2019 13:52:16 +0100
From:   Daniel Thompson <daniel.thompson@linaro.org>
To:     Suzuki K Poulose <suzuki.poulose@arm.com>
Cc:     mathieu.poirier@linaro.org, saiprakash.ranjan@codeaurora.org,
        jeffrey.l.hugo@gmail.com, mark.rutland@arm.com,
        rnayak@codeaurora.org, alexander.shishkin@linux.intel.com,
        linux-arm-msm@vger.kernel.org, marc.w.gonzalez@free.fr,
        linux-kernel@vger.kernel.org, bjorn.andersson@linaro.org,
        david.brown@linaro.org, agross@kernel.org, sibis@codeaurora.org,
        leo.yan@linaro.org, linux-arm-kernel@lists.infradead.org,
        linux-arm-msm-owner@vger.kernel.org
Subject: Re: [PATCHv9 2/3] arm64: dts: qcom: msm8998: Add Coresight support
Message-ID: <20191003125216.f5sjl6yhhb2t7uqq@holly.lan>
References: <90114e06825e537c3aafd3de5c78743a9de6fadc.1564550873.git.saiprakash.ranjan@codeaurora.org>
 <CAOCk7NrK+wY8jfHdS8781NOQtyLm2RRe1NW2Rm3_zeaot0Q99Q@mail.gmail.com>
 <16212a577339204e901cf4eefa5e82f1@codeaurora.org>
 <CAOCk7NohO67qeYCnrjy4P0KN9nLUiamphHRvj-bFP++K7khPOw@mail.gmail.com>
 <fa5a35f0e993f2b604b60d5cead3cf28@codeaurora.org>
 <CAOCk7NodWtC__W3=AQfXcjF-W9Az_NNUN0r8w5WmqJMziCcvig@mail.gmail.com>
 <5b8835905a704fb813714694a792df54@codeaurora.org>
 <CANLsYkxPOOorqcnPrbhZLzGV9Y7EGWUUyxvi-Cm5xxnzhx=Ecg@mail.gmail.com>
 <20191003102023.qk6ik5vmatheaofs@holly.lan>
 <57349bda-0e86-5fe0-3be0-55b12748c346@arm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <57349bda-0e86-5fe0-3be0-55b12748c346@arm.com>
User-Agent: NeoMutt/20180716
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Oct 03, 2019 at 11:52:36AM +0100, Suzuki K Poulose wrote:
> On 10/03/2019 11:20 AM, Daniel Thompson wrote:
> > On Wed, Oct 02, 2019 at 09:03:59AM -0600, Mathieu Poirier wrote:
> > > On Tue, 1 Oct 2019 at 12:05, Sai Prakash Ranjan
> > > <saiprakash.ranjan@codeaurora.org> wrote:
> > > > 
> > > > On 2019-10-01 11:01, Jeffrey Hugo wrote:
> > > > > On Tue, Oct 1, 2019 at 11:52 AM Sai Prakash Ranjan
> > > > > <saiprakash.ranjan@codeaurora.org> wrote:
> > > > > > 
> > > > > > 
> > > > > > Haan then likely it's the firmware issue.
> > > > > > We should probably disable coresight in soc dtsi and enable only for
> > > > > > MTP. For now you can add a status=disabled for all coresight nodes in
> > > > > > msm8998.dtsi and I will send the patch doing the same in a day or
> > > > > > two(sorry I am travelling currently).
> > > > > 
> > > > > This sounds sane to me (and is what I did while bisecting the issue).
> > > > > When you do create the patch, feel free to add the following tags as
> > > > > you see fit.
> > > > > 
> > > > > Reported-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com>
> > > > > Tested-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com>
> > > > 
> > > > Thanks Jeffrey, I will add them.
> > > > Hope Mathieu and Suzuki are OK with this.
> > > 
> > > The problem here is that a debug and production device are using the
> > > same device tree, i.e msm8998.dtsi.  Disabling coresight devices in
> > > the DTS file will allow the laptop to boot but completely disabled
> > > coresight blocks on the MTP board.  Leaving things as is breaks the
> > > laptop but allows coresight to be used on the MTP board.  One of three
> > > things can happen:
> > > 
> > > 1) Nothing gets done and production board can't boot without DTS modifications.
> > > 2) Disable tags are added to the DTS file and the debug board can't
> > > use coresight without modifications.
> > > 2) The handling of the debug power domain is done properly on the
> > > MSM8998 rather than relying on the bootloader to enable it.
> > > 3) The DTS file is split or reorganised to account for debug/production devices.
> > 
> > msm8998.dtsi is a SoC include file. Can't whatever default it adopts be
> > reversed in the board include files such as msm8998-mtp.dtsi or
> > msm8998-clamshell.dtsi ?
> 
> Or like Mathieu said, all the Coresight specific nodes could be moved in
> to say, msm8998-coresight.dtsi and could be included into the platforms
> where it actually works.

Sure, that works too.

Maybe it depends in you view the mtp as including the feature or as the
laptops as taking it away ;-) .

Treating it as a feature a board can disable also works nicely on systems
where the board include file should be setting secure-status a board
(although that's probably not the case for these boards since the
firmware is proprietary).


Daniel.