[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID:
<IA0PR12MB7699EDFA3753D25C8126D901DC762@IA0PR12MB7699.namprd12.prod.outlook.com>
Date: Mon, 30 Sep 2024 15:44:47 +0000
From: "Mahapatra, Amit Kumar" <amit.kumar-mahapatra@....com>
To: Conor Dooley <conor@...nel.org>, Krzysztof Kozlowski <krzk@...nel.org>
CC: "broonie@...nel.org" <broonie@...nel.org>, "robh@...nel.org"
<robh@...nel.org>, "krzk+dt@...nel.org" <krzk+dt@...nel.org>,
"conor+dt@...nel.org" <conor+dt@...nel.org>, "Simek, Michal"
<michal.simek@....com>, "linux-spi@...r.kernel.org"
<linux-spi@...r.kernel.org>, "devicetree@...r.kernel.org"
<devicetree@...r.kernel.org>, "linux-arm-kernel@...ts.infradead.org"
<linux-arm-kernel@...ts.infradead.org>, "linux-kernel@...r.kernel.org"
<linux-kernel@...r.kernel.org>, "git (AMD-Xilinx)" <git@....com>,
"amitrkcian2002@...il.com" <amitrkcian2002@...il.com>
Subject: RE: [PATCH] dt-bindings: spi: xilinx: Add clocks & clock-names
properties
Hello Conor,
> > >>>> Subject: Re: [PATCH] dt-bindings: spi: xilinx: Add clocks &
> > >>>> clock-names properties
> > >>>>
> > >>>> On Mon, Sep 23, 2024 at 06:02:42PM +0530, Amit Kumar Mahapatra wrote:
> > >>>>> Include the 'clocks' and 'clock-names' properties in the AXI
> > >>>>> Quad-SPI bindings. When the AXI4-Lite interface is enabled, the
> > >>>>> core operates in legacy mode, maintaining backward compatibility
> > >>>>> with version 1.00, and uses 's_axi_aclk' and 'ext_spi_clk'. For
> > >>>>> the AXI interface, it uses 's_axi4_aclk' and 'ext_spi_clk'.
>
> > >>>>> + properties:
> > >>>>> + clock-names:
> > >>>>> + items:
> > >>>>> + - const: s_axi_aclk
> > >>>>> + - const: ext_spi_clk
> > >>>>
> > >>>> These are all clocks, there should be no need to have "clk" in the names.
> > >>>
> > >>> These are the names exported by the IP and used by the DTG.
> > >>
> > >> So? This is a binding, not a verilog file.
> > >
> > > Axi Quad SPI is an FPGA-based IP, and the clock names are derived
> > > from the IP signal names as specified in the IP documentation [1].
> > > We chose these names to ensure alignment with the I/O signal names
> > > listed in Table 2-2 on page 19 of [1].
> > >
> > > [1]
> > > chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://www.amd.
> > > com/content/dam/xilinx/support/documents/ip_documentation/axi_quad_s
> > > pi/v3_2/pg153-axi-quad-spi.pdf
> >
> > So if hardware engineers call them "pink_pony_clk_aclk_really_clk" we
> > should follow...
> >
> > - bus or axi
> > - ext_spi or spi
> >
> > You have descriptions of each item to reference real signals. Conor's
> > comment is valid - do no make it verilog file.
> >
> > >
> > >>
> > >>>>> +
> > >>>>> + else:
> > >>>>> + properties:
> > >>>>> + clock-names:
> > >>>>> + items:
> > >>>>> + - const: s_axi4_aclk
> > >>>>> + - const: ext_spi_clk
> >
> > Nah, these are the same.
>
> They may be different, depending on whether or not the driver has to handle "axi4-
> lite" versus "axi" differently. That said, I find the commit message kinda odd in that it
> states that axi4-lite goes with the s_axi_aclk clock and axi goes with s_axi4_aclk.
Apologies for the typo. When the AXI4 interface is enabled, it uses s_axi4_aclk, and
when the AXI4-Lite interface is enabled, it uses s_axi_aclk.
In my next series I will update my commit message & change the clock-names
's_axi4_aclk', 's_axi_aclk' & 'ext_spi_clk' to 'axi4', 'axi' & 'ref' respectively
Regards,
Amit
> Seems backwards..
Powered by blists - more mailing lists