Commit fd1fbc15d5b9 ("ARM: i.MX: move handoff-data out of OCRAM") added
the support to store early handoff-data within DRAM. As the commit says
this says, this was to be prepared for early-board handoff data passing.
The commit added the support only for i.MX8MM and i.MX93. All other
i.MX8MQ/P/N platforms are left out. This isn't very intuitiv given that
the commit prepared the common code base for future usage.
Furthermore putting the handoff data infront of the bootloader caused a
regression on a custom i.MX8MM board which makes use of the
__imx8mm_load_and_start_image_via_tfa() because barebox is directly
loaded behind secure marked memory.
This partly reverts fd1fbc15d5b9 ("ARM: i.MX: move handoff-data out of
OCRAM") because the early board handoff-data is still unused after a
year and it aligns the common bootflow for i.MX8MQ/P/N/M again.
Signed-off-by: Marco Felsch <m.fel...@pengutronix.de>
---
Hi,
I wasn't shure how to handle this regression, but in the end I went this
way. Of course I could adapt the lowlevel board code as well to honor
the handoff_data_size(). But I went this way, since this was implemented
for i.MX8MM SoCs only and there is actual no user of it.
That beeing said, IMHO the early board lowlevel handoff-data handling
should be done by the board lowlevel code because the location is very
board specific.
Regards,
Marco
arch/arm/mach-imx/atf.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/arch/arm/mach-imx/atf.c b/arch/arm/mach-imx/atf.c
index e6c21107a54a..11fe0334059d 100644
--- a/arch/arm/mach-imx/atf.c
+++ b/arch/arm/mach-imx/atf.c
@@ -112,8 +112,6 @@ void imx8mm_load_bl33(void *bl33)
hang();
}
- handoff_data_move(bl33 - ALIGN(handoff_data_size(), 0x1000));
-
/*
* On completion the TF-A will jump to MX8M_ATF_BL33_BASE_ADDR
* in EL2. Copy the image there, but replace the PBL part of
--
2.39.5
