linux-kernel - Re: [PATCH v2 2/8] remoteproc: qcom: Introduce driver to store pil info in IMEM

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20200123053811.GU1511@yoga>
Date:   Wed, 22 Jan 2020 21:38:11 -0800
From:   Bjorn Andersson <bjorn.andersson@...aro.org>
To:     rishabhb@...eaurora.org
Cc:     Rob Herring <robh+dt@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Ohad Ben-Cohen <ohad@...ery.com>,
        linux-arm-msm@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-remoteproc@...r.kernel.org,
        Sibi Sankar <sibis@...eaurora.org>
Subject: Re: [PATCH v2 2/8] remoteproc: qcom: Introduce driver to store pil
 info in IMEM

On Wed 22 Jan 15:58 PST 2020, rishabhb@...eaurora.org wrote:

> On 2020-01-22 15:08, Bjorn Andersson wrote:
> > On Wed 22 Jan 14:56 PST 2020, rishabhb@...eaurora.org wrote:
> > > On 2019-12-26 21:32, Bjorn Andersson wrote:
> > > > diff --git a/drivers/remoteproc/qcom_pil_info.c
> > [..]
> > > > +static int pil_reloc_probe(struct platform_device *pdev)
> > > > +{
> > > > +	struct pil_reloc *reloc;
> > > > +
> > > > +	reloc = devm_kzalloc(&pdev->dev, sizeof(*reloc), GFP_KERNEL);
> > > > +	if (!reloc)
> > > > +		return -ENOMEM;
> > > > +
> > > > +	reloc->dev = &pdev->dev;
> > > > +	reloc->map = syscon_node_to_regmap(pdev->dev.parent->of_node);
> > > If there are multiple entries like "pil-reloc" in the imem node
> > > mapping the entire imem multiple times may not work. Is there a way
> > > we can somehow just iomap the required region for pil?
> > 
> > With the entire imem being represented as a syscon this will be
> > ioremapped once and all callers of syscon_node_to_regmap() (or one of
> > the other syscon getters) will get a regmap back that reference this one
> > mapping.
> > 
> > So doing it this way allow us to "map" sections of imem that is smaller
> > than PAGE_SIZE.
> > 
> > 
> > That said, it means that all imem users/clients should access imem
> > through this syscon regmap.
> > 
> > Regards,
> > Bjorn
> Yes, the clients are spread around in different drivers currently.
> So accessing same regmap is not possible.

The few examples upstream are children of the imem simple-mfd/syscon and
will thereby naturally request the regmap of the parent syscon.

For driver that doesn't fit this model (I don't find one right now), or
if you have downstream drivers that are designed differently you could
use syscon_regmap_lookup_by_phandle() to acquire the imem regmap from
any device in the system.

Regards,
Bjorn.