[PATCH] efi_loader: fix AllocatePages overlap status

[Harsimran Singh Tungal]

On 2026-04-27 19:04 +0200, Heinrich Schuchardt wrote:
> Am 27. April 2026 17:48:21 MESZ schrieb Ilias Apalodimas <ilias.apalodimas at linaro.org>:
> >Hi Harsiman,
> >
> >On Mon, 27 Apr 2026 at 18:05, Harsimran Singh Tungal
> ><harsimransingh.tungal at arm.com> wrote:
> >>
> >> Return EFI_NOT_FOUND for EFI_ALLOCATE_ADDRESS overlap
> >>
> >> When efi_allocate_pages() is called with EFI_ALLOCATE_ADDRESS, UEFI
> >> expects EFI_NOT_FOUND if the requested address range is already
> >> allocated or unavailable.
> >>  U-Boot currently returns
> >> EFI_OUT_OF_RESOURCES when efi_update_memory_map() detects an overlap
> >> after a successful lmb_alloc_mem(), which does not match
> >> EFI_ALLOCATE_ADDRESS semantics.
> >>
> >> Return EFI_NOT_FOUND for EFI_ALLOCATE_ADDRESS requests that fail due
> >> to an overlapping EFI memory descriptor, while keeping
> >> EFI_OUT_OF_RESOURCES for other allocation types.
> >>
> >> The UEFI specification [1] specifies that
> >> EFI_BOOT_SERVICES.AllocatePages must return EFI_NOT_FOUND when the
> >> requested address range is unavailable or already allocated;
> >> EFI_OUT_OF_RESOURCES applies to non‑address‑specific allocation
> >> failures.
> >
> >Where does the EFI spec say that? EDKII indeed returns EFI_NOT_FOUND
> >for AllocateAddress but looking at the function description, I can't
> >find anything specific.
> >
> >Thanks
> >/Ilias
> 
> Thank you Harsimran for investigating this difference to EDJ II.
> 
> The spec has:
> 
> EFI_NOT_FOUND
> The requested pages could not be found.
> 
> and
> 
> EFI_OUT_OF_RESOURCEST
> The pages could not be allocated.
> 
> Looking at the text one could assume that EFI_NOT_FOUND should be returned if the pages requested with AllocateAddress don't exist in the memory map.
> 
> If the pages exist and are already allocated EFI_NOT_FOUND does not look like the right return code according to the spec.
> 
> It is unclear if for AllocateMaxAddress a comparison to the memory bottom is required.
> 
> The specification should be clarified in the UEFI forum.
> 
> Best regards
> 
> Heinrich 
> 

Hi Ilias, Hi Heinrich,

Thanks for the clarification.

You are right.
The AllocatePages() text only lists the generic status codes, and it
does not explicitly say that an AllocateAddress request on an already
allocated range must return EFI_NOT_FOUND.

My motivation for this change was tlso he UEFI SCT expectation rather just a 
direct statement in the UEFI base spec. The SCT AllocatePages tests expect
EFI_NOT_FOUND for AllocateAddress both with non-existent memory and with
allocated memory, and that is the behavior I was trying to align U-Boot
with, which matches with EDK II as well.

Is this solution acceptable for now?

Thanks,
Harsimran Singh Tungal 
> 
> 
> >
> >>
> >> [1] https://uefi.org/specs/UEFI/2.10_A/07_Services_Boot_Services.html
> >>
> >> Signed-off-by: Harsimran Singh Tungal <harsimransingh.tungal at arm.com>
> >> ---
> >>  lib/efi_loader/efi_memory.c | 4 +++-
> >>  1 file changed, 3 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/lib/efi_loader/efi_memory.c b/lib/efi_loader/efi_memory.c
> >> index b77c2f980cc..63cf1a7277f 100644
> >> --- a/lib/efi_loader/efi_memory.c
> >> +++ b/lib/efi_loader/efi_memory.c
> >> @@ -510,7 +510,9 @@ efi_status_t efi_allocate_pages(enum efi_allocate_type type,
> >>                 /* Map would overlap, bail out */
> >>                 lmb_free(addr, (u64)pages << EFI_PAGE_SHIFT, flags);
> >>                 unmap_sysmem((void *)(uintptr_t)efi_addr);
> >> -               return  EFI_OUT_OF_RESOURCES;
> >> +               if (type == EFI_ALLOCATE_ADDRESS)
> >> +                       return EFI_NOT_FOUND;
> >> +               return EFI_OUT_OF_RESOURCES;
> >>         }
> >>
> >>         *memory = efi_addr;
> >> --
> >> 2.34.1
> >>
> 
>