On Mon, Oct 21, 2024 at 04:18:01PM +0200, Sascha Hauer wrote:
> On Mon, Oct 21, 2024 at 10:55:45AM +0200, Ahmad Fatoum wrote:
> > Hello Sascha,
> >
> > On 21.10.24 09:21, Sascha Hauer wrote:
> > > Both LS1046a boards supported by barebox have been switched to TF-A, so
> > > drop no longer used PPA support.
> > >
> > > Signed-off-by: Sascha Hauer <s.hauer@xxxxxxxxxxxxxx>
> > > ---
> > > arch/arm/mach-layerscape/Kconfig | 12 ---
> > > arch/arm/mach-layerscape/Makefile | 1 -
> > > arch/arm/mach-layerscape/ppa-entry.S | 32 --------
> > > arch/arm/mach-layerscape/ppa.c | 147 -----------------------------------
> > > firmware/Makefile | 1 -
> > > include/mach/layerscape/layerscape.h | 10 ---
> >
> > generate-dummy-fw.sh still has a firmware/ppa-ls1046a.bin
> > reference.
> >
> > > -static int of_psci_do_fixup(struct device_node *root, void *unused)
> > > -{
> > > - unsigned long psci_version;
> > > - struct device_node *np;
> > > - struct arm_smccc_res res = {};
> > > -
> > > - arm_smccc_smc(ARM_PSCI_0_2_FN_PSCI_VERSION, 0, 0, 0, 0, 0, 0, 0, &res);
> > > - psci_version = res.a0;
> > > -
> > > - for_each_compatible_node_from(np, root, NULL, "fsl,sec-v4.0-job-ring") {
> >
> > The TF-A doesn't expect one of the job rings to be for its own use?
>
> Good question. It looks like U-Boot still has this fixup code eventhough
> they have converted to TF-A.
>
> I wonder what the effects are when we don't apply this quirk. At least
> starting a kernel with PPA and without this quirk doesn't cause any
> problems.
Looking at the TF-A code the CAAM JR3 indeed seems to be used, at least
under certain circumstances, so I'll put the code to disable JR3 in the
kernel back in.
Sascha
--
Pengutronix e.K. | |
Steuerwalder Str. 21 | http://www.pengutronix.de/ |
31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
