Hi Michael,
Am 02.05.2022 um 16:00 schrieb Michael Tretter:
On Sun, 01 May 2022 20:26:07 +0200, Matthias Fend wrote:
Since pmu-fw release 2018.3, the ZYNQMP_FPGA_BIT_ONLY_BIN flag is no
longer used. This wasn't a problem for a while, but in newer versions a
validation sequence will fail if this flag is set. This means that the PL
can no longer be loaded.
Do not set the ZYNQMP_FPGA_BIT_ONLY_BIN flag unless absolutely necessary
to avoid this problem.
Thanks for the patch.
Signed-off-by: Matthias Fend <matthias.fend@xxxxxxxxx>
---
drivers/firmware/zynqmp-fpga.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/firmware/zynqmp-fpga.c b/drivers/firmware/zynqmp-fpga.c
index 63d7398fd..db34ac2be 100644
--- a/drivers/firmware/zynqmp-fpga.c
+++ b/drivers/firmware/zynqmp-fpga.c
@@ -261,9 +261,10 @@ static int fpgamgr_program_finish(struct firmware_handler *fh)
goto err_free_dma;
}
- if (mgr->features & ZYNQMP_PM_FEATURE_SIZE_NOT_NEEDED)
+ if (mgr->features & ZYNQMP_PM_FEATURE_SIZE_NOT_NEEDED) {
+ flags &= ~ZYNQMP_FPGA_BIT_ONLY_BIN;
I would prefer if the flag is initialized as unset and only set, if the
bitstream does not have headers. This would be a lot clearer, than resetting
the flag based on a version dependent feature flag with a different name.
If I understand correctly, the newer versions of the PMUFW don't support
loading a bitstream without headers at all. Maybe we should have a feature
flag for bitstreams without headers, but maybe it is enough to just let the
PMUFW reject the bitstream in these cases.
I don't know if the PMU firmware now somehow support header-less
bitstreams, but I can tell that it will fail in any case if this bit is set.
So, with recent versions, this bit simply never should be set.
Here is the check that will fail if this bit is set:
https://github.com/Xilinx/embeddedsw/blob/master/lib/sw_services/xilfpga/src/xilfpga.c#L643
Since I can't test the different combinations for old (pre-2018.3) PMU
firmware, I would suggest leaving the bit set to avoid breaking something.
~Matthias
Michael
buf_size = body_length;
- else
+ } else
buf_size = addr + body_length;
status = mgr->eemi_ops->fpga_load((u64)addr, buf_size, flags);
--
2.25.1
_______________________________________________
barebox mailing list
barebox@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/barebox
_______________________________________________
barebox mailing list
barebox@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/barebox
