Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.37 as permitted sender) client-ip=23.128.96.37;
Date:   Fri, 22 Sep 2023 09:35:00 +0200
From:   Stephan Gerhold <stephan@gerhold.net>
To:     Bjorn Andersson <andersson@kernel.org>
Cc:     Bjorn Andersson <quic_bjorande@quicinc.com>,
        Andy Gross <agross@kernel.org>,
        Konrad Dybcio <konrad.dybcio@linaro.org>,
        Rob Herring <robh+dt@kernel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
        Conor Dooley <conor+dt@kernel.org>,
        linux-arm-msm@vger.kernel.org, devicetree@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3 2/3] soc: qcom: rmtfs: Support discarding guard pages
Message-ID: <ZQ1DmSRIe56WQmnK@gerhold.net>
References: <20230920-rmtfs-mem-guard-pages-v3-0-305b37219b78@quicinc.com>
 <20230920-rmtfs-mem-guard-pages-v3-2-305b37219b78@quicinc.com>
 <ZQyFliFYV7dUwGJg@gerhold.net>
 <j5ivvbjprgdcekujzjiobk7gpoaz3jxpnda5dsgoh6f4hsn5id@3ug5i74ubkqf>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <j5ivvbjprgdcekujzjiobk7gpoaz3jxpnda5dsgoh6f4hsn5id@3ug5i74ubkqf>
Content-Transfer-Encoding: 7bit
Precedence: bulk

eOn Thu, Sep 21, 2023 at 07:51:42PM -0700, Bjorn Andersson wrote:
> On Thu, Sep 21, 2023 at 08:04:06PM +0200, Stephan Gerhold wrote:
> > On Wed, Sep 20, 2023 at 07:37:31PM -0700, Bjorn Andersson wrote:
> > > In some configurations, the exact placement of the rmtfs shared memory
> > > region isn't so strict. The DeviceTree author can then choose to use the
> > > "size" property and rely on the OS for placement (in combination with
> > > "alloc-ranges", if desired).
> > > 
> > > But on some platforms the rmtfs memory region may not be allocated
> > > adjacent to regions allocated by other clients. Add support for
> > > discarding the first and last 4k block in the region, if
> > > qcom,use-guard-pages is specified in DeviceTree.
> > > 
> > > Signed-off-by: Bjorn Andersson <quic_bjorande@quicinc.com>
> > > ---
> > >  drivers/soc/qcom/rmtfs_mem.c | 9 +++++++++
> > >  1 file changed, 9 insertions(+)
> > > 
> > > diff --git a/drivers/soc/qcom/rmtfs_mem.c b/drivers/soc/qcom/rmtfs_mem.c
> > > index f83811f51175..83bba9321e72 100644
> > > --- a/drivers/soc/qcom/rmtfs_mem.c
> > > +++ b/drivers/soc/qcom/rmtfs_mem.c
> > > @@ -200,6 +200,15 @@ static int qcom_rmtfs_mem_probe(struct platform_device *pdev)
> > >  	rmtfs_mem->client_id = client_id;
> > >  	rmtfs_mem->size = rmem->size;
> > >  
> > > +	/*
> > > +	 * If requested, discard the first and last 4k block in order to ensure
> > > +	 * that the rmtfs region isn't adjacent to other protected regions.
> > > +	 */
> > > +	if (of_property_present(node, "qcom,use-guard-pages")) {
> > 
> > I think of_property_read_bool() would be more fitting here. Right now
> > of_property_present() is just a wrapper around of_property_read_bool().
> > Semantically reading a bool fits better here though. :-)
> > 
> 
> Are you saying that you would prefer this to be a bool, so hat you can
> give it a "false" value? Or you are simply saying "it walks like a
> boolean, quacks like a boolean, let's use the boolean accessor"?
> 

The latter. I would expect that of_property_present() is used for
properties which usually have a value, while of_property_read_bool()
is used for pure bool values which can be present or not but must not
have a value. I think a "bool" in terms of DT is simply a present or
not-present property without any value?

For example consider

  regulator-min-microvolts = <4200000000>;
  regulator-always-on;

Then I would expect

  - of_property_present(..., "regulator-min-microvolts"), but
  - of_property_read_bool(..., "regulator-always-on")

Does that make sense? :D

> > Feel free to fix that up while applying.
> > 
> > FWIW I don't really have an opinion if "qcom,use-guard-pages" is a good
> > way to describe this in the DT. For the implementation side feel free to
> > add my
> > 
> 
> Right, I don't think I commented on your suggestion to make the size of
> the guard page configurable. I am not aware of any current or upcoming
> reasons for adding such complexity, so I'd simply prefer to stick with a
> boolean. Should that need arise, I think this model would allow
> extension to express that.
> 

I must admit I forgot that I suggested this until now. :')
I don't see a use case for a different "guard size" either so I think
it's fine to have it as a bool.

Thanks,
Stephan