Am 15. September 2025 19:42:31 MESZ schrieb Tom Rini <tr...@konsulko.com>: >On Sat, Sep 13, 2025 at 06:49:46PM +0200, Heinrich Schuchardt wrote: >> Am 13. September 2025 17:18:07 MESZ schrieb Tom Rini <tr...@konsulko.com>: >> >On Sat, Sep 13, 2025 at 02:32:26AM +0200, Heinrich Schuchardt wrote: >> >> On 9/13/25 01:02, Tom Rini wrote: >> >> > This converts the existing README.commands.spl file to rST as >> >> > doc/usage/cmd/spl.rst and adds it to the command index. Minimal >> >> > rewording is done to the previous contents, to make it fit better with >> >> > the overall style of documentation now. >> >> > >> >> > Signed-off-by: Tom Rini <tr...@konsulko.com> >> >> > --- >> >> > Cc: Heinrich Schuchardt <xypron.g...@gmx.de> >> >> > --- >> >> > doc/README.commands.spl | 31 -------------------------- >> >> > doc/usage/cmd/spl.rst | 48 +++++++++++++++++++++++++++++++++++++++++ >> >> > doc/usage/index.rst | 1 + >> >> > 3 files changed, 49 insertions(+), 31 deletions(-) >> >> > delete mode 100644 doc/README.commands.spl >> >> > create mode 100644 doc/usage/cmd/spl.rst >> >> > >> >> > diff --git a/doc/README.commands.spl b/doc/README.commands.spl >> >> > deleted file mode 100644 >> >> > index ecfd3ca9ee58..000000000000 >> >> > --- a/doc/README.commands.spl >> >> > +++ /dev/null >> >> > @@ -1,31 +0,0 @@ >> >> > -The spl command is used to export a boot parameter image to RAM. Later >> >> > -it may implement more functions connected to the SPL. >> >> > - >> >> > -SUBCOMMAND EXPORT >> >> > -To execute the command everything has to be in place as if bootm >> >> > should be >> >> > -used. (kernel image, initrd-image, fdt-image etc.) >> >> > - >> >> > -export has two subcommands: >> >> > - atags: exports the ATAGS >> >> > - fdt: exports the FDT >> >> > - >> >> > -Call is: >> >> > -spl export <fdt|atags> [kernel_addr] [initrd_addr] [fdt_addr if fdt] >> >> > - >> >> > - >> >> > -TYPICAL CALL >> >> > - >> >> > -on OMAP3: >> >> > -nandecc hw >> >> > -nand read 0x82000000 0x280000 0x400000 /* Read kernel image from NAND*/ >> >> > -spl export atags /* export ATAGS */ >> >> > -nand erase 0x680000 0x20000 /* erase - one page */ >> >> > -nand write 0x80000100 0x680000 0x20000 /* write the image - one page */ >> >> > - >> >> > -call with FDT: >> >> > -nandecc hw >> >> > -nand read 0x82000000 0x280000 0x400000 /* Read kernel image from NAND*/ >> >> > -tftpboot 0x80000100 devkit8000.dtb /* Read fdt */ >> >> > -spl export fdt 0x82000000 - 0x80000100 /* export FDT */ >> >> > -nand erase 0x680000 0x20000 /* erase - one page */ >> >> > -nand write <adress shown by spl export> 0x680000 0x20000 >> >> > diff --git a/doc/usage/cmd/spl.rst b/doc/usage/cmd/spl.rst >> >> > new file mode 100644 >> >> > index 000000000000..59471c65a623 >> >> > --- /dev/null >> >> > +++ b/doc/usage/cmd/spl.rst >> >> > @@ -0,0 +1,48 @@ >> >> > +.. SPDX-License-Identifier: GPL-2.0+: >> >> >> >> %s/GPL-2.0+/GPL-2.0-OR-LATER/ >> >> >> >> Cf. https://spdx.org/licenses/GPL-2.0-or-later.html >> > >> >Yes, and it's not a hard requirement to use the new tag, checkpatch.pl >> >doesn't even complain. I got this from the existing command I copied, >> >FWIW. >> > >> >> >> >> > + >> >> > +.. index:: >> >> > + single: spl (command) >> >> > + >> >> > +spl command >> >> > +=========== >> >> > + >> >> > +Synopsis >> >> > +-------- >> >> > + >> >> > +:: >> >> > + >> >> > + spl export <fdt|atags> [kernel_addr] [initrd_addr] [fdt_addr if >> >> > fdt] >> >> >> >> Those brackets cannot be correct. I definitively cannot remove the >> >> kernel_addr and provide an initrd_addr instead. Do you mean: >> >> >> >> spl export <fdt|atags> [kernel_addr [initrd_addr [fdt_addr if fdt]]] >> > >> >It's what the command itself says currently, however. >> > >> >I'd be happy to do a follow-up patch to fix both. >> >> I am fine with the current patch if you supply the follow up. > >OK. But I also think that based on some of your questions, the first >step is to just rip out ATAGS from here. We have zero cases in tree of >ATAGS and SPL_OS_BOOT support, and this is an extremely legacy case. >Deleting that will make things a bit clearer overall. The next part is >that yes, two sentences doesn't explain what's going on here well if you >don't already know what's going on here. >
From what I read ATAGs are used to boot ARM systems that do not yet support device-trees. This seems to be incompatible with our driver model. Thus, eliminating all ATAG related code is a reasonable way forward. Best regards Heinrich
