[U-Boot] [PATCHv1] arm: stm32mp1: deploy spl in root folder
Patrick DELAUNAY
patrick.delaunay at st.com
Fri Jan 25 09:23:06 UTC 2019
Hi Sean,
> From: Sean Nyekjaer <sean at geanix.com>
> Sent: vendredi 25 janvier 2019 08:11
> To: Patrick DELAUNAY <patrick.delaunay at st.com>; ravi at prevas.dk
> Cc: Patrice CHOTARD <patrice.chotard at st.com>; u-boot at lists.denx.de
> Subject: Re: [U-Boot] [PATCHv1] arm: stm32mp1: deploy spl in root folder
> Importance: High
>
> Hi Patrick
>
> >
> > I checked with Patrice the other boards and most of the cases the
> > modified spl binay (u-boot-spl.bin/ u-boot-spl.ais / u-boot-spl.img /
> > u-boot-spl.pbl / u-boot-spl.gph) are generated in spl directory.
> >
> > So I done the same for u-boot-spl.stm32.
> All the NXP iMX6 variant boards deploy the SPL to the root folder :-)
Ok, I checked and the generated files are
1 - spl/u-boot-spl.bin
2 - SPL = spl/u-boot-spl.bin + $(PLUGIN).bin + imx header
3 - u-boot.bin
SPL is no only spl but include plugin,
so it is normal for for that it is not in spl sub-directory
> >
> > The only case where it is populated in u-boot directory it is when it
> > is concatenated with u-boot image (so the generated files depends on u-
> boot.bin) , but is not the case for us.
> >
> > Perhaps I miss so other cases ?
> >
> > Moreover I am working with YOCTO and using a sub-directory is not
> > issue with latest version and with patch in
> > poky/meta/recipes-bsp/u-boot/u-boot.inc
> > ------------------------------------------------------------------------------------------
> > commit c6ab82882cd49be5510d1f8c967d0dc2da2490c2
> > Author: Nathan Rossi <nathan at nathanrossi.com> Fri Mar 25 10:07:12
> 2016
> > Committer: Richard Purdie <richard.purdie at linuxfoundation.org> Mon Mar
> 28 16:55:48 2016
> >
> > u-boot.inc: Add sub-dir support for SPL_BINARY
> >
> > Add support for the SPL_BINARY variable to handle sub directories. In
> > some cases the SPL binary that needs to be deployed is only built to the
> > spl/ directory in U-Boot. So that a sub directory can be specified in
> > the SPL_BINARY variable, handle the case so that the deploy code uses
> > the basename of the path specified in SPL_BINARY.
> >
> > (From OE-Core rev: eb90d1c8fc7b82ca2593185930b3bf175f40ae13)
> >
> > Signed-off-by: Nathan Rossi <nathan at nathanrossi.com>
> > Signed-off-by: Richard Purdie <richard.purdie at linuxfoundation.org>
> > ----------------------------------------------------------------------
> > --------------------
> >
> > So, I don't understood the blocking point, and other platform do the same (the
> reason of the patch).
> >
> > Even I agree that copy the generated u-boot-spl.stm32 in u-boot build
> > directory could be easier (for BUILDROOT ?), I want keep the file
> > spl/u-boot-spl.stm32 to avoid to break the current upstream effort of YOCTO
> BSP support for STM32MP157 board (which expect spl file in spl directory).
>
> I got a new job (therefore new email address), so i don't have access to the
> Yocto source where I got into troubles with this.
> If i recall correctly we where using the SPI_BINARY variable when deploying the
> SPL and we should have been using the SPL_BINARYNAME.
>
> So no need for this patch.
Ok, I think the patch can be abandon as with latest YOCTO it is working with machine value:
SPL_BINARY = " spl/u-boot-spl.stm32"
Because path is supported in the next variables
SPL_BINARYNAME ?= "${@os.path.basename(d.getVar("SPL_BINARY"))}"
SPL_IMAGE ?= "${SPL_BINARYNAME}-${MACHINE}-${PV}-${PR}"
SPL_SYMLINK ?= "${SPL_BINARYNAME}-${MACHINE}"
PS: It can also work with previous version YOCTO with
SPL_BINARY = " spl/u-boot-spl.stm32"
SPL_BINARYNAME = "u-boot-spl.stm32"
But I will kept my proposal in mind if other requests/complaint come in the future.
Thanks for your time and
good luck for your new job.
>
> /Sean
Best regards
Patrick
More information about the U-Boot
mailing list