[PATCH] arm: boards: Add support for MYIR MYD-AM335X Development Board

Alexander Shiyan eagle.alexander923 at gmail.com
Thu Apr 21 01:58:26 PDT 2022


чт, 21 апр. 2022 г. в 11:30, Sascha Hauer <sha at pengutronix.de>:
> On Thu, Apr 21, 2022 at 10:57:33AM +0300, Alexander Shiyan wrote:
> > чт, 21 апр. 2022 г. в 10:35, Sascha Hauer <sha at pengutronix.de>:
> > >
> > > On Wed, Apr 20, 2022 at 12:53:22PM +0300, Alexander Shiyan wrote:
> > > > The MYD-AM335X Development Board designed by MYIR is a high-performance
> > > > ARM Evaluation Module (EVM) using the MYC-AM335X CPU module as the core
> > > > controller board. It is based on up to 1GHz Texas Instruments (TI)
> > > > Sitara AM335x family of ARM Cortex-A8 Microprocessors (MPUs) that deliver
> > > > high DMIPs at a low cost while also delivering optional 3D graphics
> > > > acceleration and key peripherals.
> > > >
> > > > Signed-off-by: Alexander Shiyan <eagle.alexander923 at gmail.com>
> > ...
> > > > +static struct omap_barebox_part myir_barebox_part = {
> > > > +     .nand_offset = SZ_128K,
> > > > +     .nand_size = SZ_512K,
> > > > +};
> > >
> > > Isn't this a bit too small? Putting a barebox into 128k seems very
> > > ambicous.
> >
> > The boot image (MLO) should not exceed 128 KB in size,
> > so I think this should be enough.
> > In any case this size is already specified for the boot partition
> > in kernel DTS for this board. :)
>
> Oh, I accidently misread nand_size as SZ_128K. Indeed it's SZ_512K which
> is small, but sufficient. Anyway, the BBU code assumes 4 eraseblocks for
> the four MLO images. That would be a nand_offset of SZ_512K also, which
> would then also match the upstream device tree for the board:
>
> &nand0 {
>         partition at 0 {
>                 label = "MLO";
>                 reg = <0x00000 0x20000>;
>         };
>
>         partition at 20000 {
>                 label = "boot";
>                 reg = <0x20000 0x80000>;
>         };
> };
>
> 512k is still very small nowadays. The images built with omap_defconfig
> are just a bit smaller than that.

Do you prefer that we fix this in the local overlay now,
or leave it as is until kernel DTS upstream changes?



More information about the barebox mailing list