linux-kernel - Re: [Patch v3 3/9] dt-bindings: arm: tegra: Add NVIDIA Tegra194 axi2apb binding

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <YhY1Hhgz/O724oYL@orome>
Date:   Wed, 23 Feb 2022 14:22:38 +0100
From:   Thierry Reding <thierry.reding@...il.com>
To:     Rob Herring <robh@...nel.org>
Cc:     Sumit Gupta <sumitg@...dia.com>,
        linux-tegra <linux-tegra@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        devicetree@...r.kernel.org, Jon Hunter <jonathanh@...dia.com>,
        kbuild-all@...ts.01.org, bbasu@...dia.com, vsethi@...dia.com,
        jsequeira@...dia.com, Thierry Reding <treding@...dia.com>
Subject: Re: [Patch v3 3/9] dt-bindings: arm: tegra: Add NVIDIA Tegra194
 axi2apb binding

On Mon, Dec 27, 2021 at 11:41:10AM -0400, Rob Herring wrote:
> On Thu, Dec 23, 2021 at 4:24 AM Sumit Gupta <sumitg@...dia.com> wrote:
> > > On Tue, Dec 21, 2021 at 06:21:11PM +0530, Sumit Gupta wrote:
> > >> Add device-tree binding documentation to represent the axi2apb bridges
> > >> used by Control Backbone (CBB) 1.0 in Tegra194 SOC. All errors for APB
> > >> slaves are reported as slave error because APB bas single bit to report
> > >> error. So, CBB driver needs to further check error status registers of
> > >> all the axi2apb bridges to find error type.
> > >>
> > >> Signed-off-by: Sumit Gupta <sumitg@...dia.com>
> > >> Signed-off-by: Thierry Reding <treding@...dia.com>
> > >> ---
> > >>   .../arm/tegra/nvidia,tegra194-axi2apb.yaml    | 40 +++++++++++++++++++
> > >>   1 file changed, 40 insertions(+)
> > >>   create mode 100644 Documentation/devicetree/bindings/arm/tegra/nvidia,tegra194-axi2apb.yaml
> > >>
> > >> diff --git a/Documentation/devicetree/bindings/arm/tegra/nvidia,tegra194-axi2apb.yaml b/Documentation/devicetree/bindings/arm/tegra/nvidia,tegra194-axi2apb.yaml
> > >> new file mode 100644
> > >> index 000000000000..788a13f8aa93
> > >> --- /dev/null
> > >> +++ b/Documentation/devicetree/bindings/arm/tegra/nvidia,tegra194-axi2apb.yaml
> > >> @@ -0,0 +1,40 @@
> > >> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> > >> +%YAML 1.2
> > >> +---
> > >> +$id: "http://devicetree.org/schemas/arm/tegra/nvidia,tegra194-axi2apb.yaml#"
> > >> +$schema: "http://devicetree.org/meta-schemas/core.yaml#"
> > >> +
> > >> +title: NVIDIA Tegra194 AXI2APB bridge
> > >> +
> > >> +maintainers:
> > >> +  - Sumit Gupta <sumitg@...dia.com>
> > >> +
> > >> +properties:
> > >> +  $nodename:
> > >> +    pattern: "^axi2apb@([0-9a-f]+)$"
> > >> +
> > >> +  compatible:
> > >> +    enum:
> > >> +      - nvidia,tegra194-axi2apb
> > >> +
> > >> +  reg:
> > >> +    maxItems: 6
> > >> +    description: Physical base address and length of registers for all bridges
> > >> +
> > >> +additionalProperties: false
> > >> +
> > >> +required:
> > >> +  - compatible
> > >> +  - reg
> > >> +
> > >> +examples:
> > >> +  - |
> > >> +    axi2apb: axi2apb@...0000 {
> > >
> > > As axi2apb appears to be a bus, then all the child nodes (APB devices)
> > > should be under this node.
> >
> > axi2apb is a bridge which coverts an AXI to APB interface and not a bus.
> 
> A bus and bridge node are pretty much one and the same in DT
> representation. A PCI host bridge has a PCI bus beneath it for
> example.

Sorry for taking so long to reply, this fell through the cracks.

These aren't really bridges as such. CBB (which we call /bus@0 in DT) is
a sort of large container for all IP. Within that there are various shim
layers that connect these "legacy" interfaces to CBB. I suppose you
could call them bridges, but it's a bit of a stretch. From a software
point of view there is no observable translation happening. The only
reason why we need this is for improved error reporting.

The TRM also doesn't make a distinction between the various bridges. The
devices are all just mapped into a single address space via the CBB.

My understanding is that this is also gone in newer chips, so matters
become a bit simpler there.

Reorganizing /bus@0 into multiple bridges and busses would be a lot of
churn and likely confuse people that want to correlate what's in the TRM
to what's in DT, so I don't think it's worth it.

For newer chips we may want to keep this in mind so we structure the DT
more accurately from the beginning, though as I said, things have been
simplified a bit, so this may not be an issue anymore.

Thierry

Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)