On Wed, Apr 10, 2019 at 11:02:57AM +0200, Patrick Delaunay wrote:
Avoid ram_end = 0 on 32bit targets with ram_end at 4G.
Signed-off-by: Patrick Delaunay patrick.delaunay@st.com
example of issue in stm32mp1 board ev1: ram_start = c0000000 size = 40000000 ram_end = 100000000 ram_end &= ~EFI_PAGE_MASK => result is 0
lib/efi_loader/efi_memory.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/lib/efi_loader/efi_memory.c b/lib/efi_loader/efi_memory.c index 55622d2..81dc5fc 100644 --- a/lib/efi_loader/efi_memory.c +++ b/lib/efi_loader/efi_memory.c @@ -574,6 +574,10 @@ __weak void efi_add_known_memory(void)
/* Remove partial pages */ ram_end &= ~EFI_PAGE_MASK;
/* Fix for 32bit targets with ram_top at 4G */if (!ram_end)ram_end = 0x100000000ULL;ram_start = (ram_start + EFI_PAGE_MASK) & ~EFI_PAGE_MASK;
if (ram_end <= ram_start) {
-- 2.7.4
U-Boot mailing list U-Boot@lists.denx.de https://lists.denx.de/listinfo/u-boot
Hi,
I had the exact same issue yesterday and posted a diff which I think is a better approach:
https://patchwork.ozlabs.org/patch/1082739/
Best regards, another Patrick