Hi Heinrich,
On Mon, 4 Nov 2024 at 13:37, Heinrich Schuchardt xypron.glpk@gmx.de wrote:
Am 4. November 2024 14:39:45 MEZ schrieb Simon Glass sjg@chromium.org:
The cache-flush function is incorrect which causes a crash in the remoteproc tests with arm64.
Fix both problems by using map_sysmem() to convert an address to a pointer and map_to_sysmem() to convert a pointer to an address.
Also update the image-loader's cache-flushing logic.
As discussed some time ago, it would be good to update the way EFI_LOADER uses addresses and pointers, particularly around memory allocation. Ideally we would use addresses internally, with pointers only exposed via the external API, even where pointers are in fact u64 values.
No, we should get rid of sandbox virtual addresses where possible, e.g. in LMB.
They are are only relevant for the sandbox's command line interface and device-tree.
This will minimize the number of conversions, which only satisfy the needs of the sandbox and have no relevance on real systems including QEMU.
Do not wag the dog with the tail.
For now I'll just remove that comment from this commit. We use ulong in U-Boot for addresses. Everything works fine and we only get in trouble when we use a ulong as a pointer. Let's discuss this in more detail at some point.
Best regards
Heinrich
Signed-off-by: Simon Glass sjg@chromium.org Fixes: 3286d223fd7 ("sandbox: implement invalidate_icache_all()")
arch/sandbox/cpu/cache.c | 8 +++++++- drivers/remoteproc/rproc-elf-loader.c | 18 +++++++++++------- lib/efi_loader/efi_image_loader.c | 3 ++- 3 files changed, 20 insertions(+), 9 deletions(-)
Regards, Simon