| Message ID | 20170726183405.27841-1-robdclark@gmail.com |
|---|---|
| State | Accepted |
| Commit | a1b24823b6c22ede11d311b05ea862965ea86b0b |
| Delegated to: | Alexander Graf |
| Headers | show |
On 26.07.17 20:34, Rob Clark wrote: > When booting shim -> fallback -> shim -> grub -> linux the memory map is > a bit larger than the size linux passes in on the first call. But in > the EFI_BUFFER_TOO_SMALL case we were not passing back the updated size > to linux so it would loop forever. > > Signed-off-by: Rob Clark <robdclark@gmail.com> The spec is actually very explicit about this case. It says in the EFI_BUFFER_TOO_SMALL case, we *have* to return the map size. Alex
On Wed, Jul 26, 2017 at 4:10 PM, Alexander Graf <agraf@suse.de> wrote: > > > On 26.07.17 20:34, Rob Clark wrote: >> >> When booting shim -> fallback -> shim -> grub -> linux the memory map is >> a bit larger than the size linux passes in on the first call. But in >> the EFI_BUFFER_TOO_SMALL case we were not passing back the updated size >> to linux so it would loop forever. >> >> Signed-off-by: Rob Clark <robdclark@gmail.com> > > > The spec is actually very explicit about this case. It says in the > EFI_BUFFER_TOO_SMALL case, we *have* to return the map size. > yes, that is what I fixed. We *weren't* returning the required buffer size before :-) BR, -R
On Mittwoch, 26. Juli 2017 22:25:29 CEST Rob Clark wrote: > On Wed, Jul 26, 2017 at 4:10 PM, Alexander Graf <agraf@suse.de> wrote: > > On 26.07.17 20:34, Rob Clark wrote: > >> When booting shim -> fallback -> shim -> grub -> linux the memory map is > >> a bit larger than the size linux passes in on the first call. But in > >> the EFI_BUFFER_TOO_SMALL case we were not passing back the updated size > >> to linux so it would loop forever. > >> > >> Signed-off-by: Rob Clark <robdclark@gmail.com> > > > > The spec is actually very explicit about this case. It says in the > > EFI_BUFFER_TOO_SMALL case, we *have* to return the map size. > > yes, that is what I fixed. We *weren't* returning the required buffer > size before :-) Sigh, yes, this was correct in the first 3 versions of the patch series, but unfortunately broken in v4 which was actually committed ... See: https://lists.denx.de/pipermail/u-boot/2016-October/268766.html Actually, the map_size variable is no longer needed, if you assign to *memory_map_size directly. Anyway, this patch is: Reviewed-by: Stefan Brüns <stefan.bruens@rwth-aachen.de> Kind regards, Stefan
> When booting shim -> fallback -> shim -> grub -> linux the memory map is > a bit larger than the size linux passes in on the first call. But in > the EFI_BUFFER_TOO_SMALL case we were not passing back the updated size > to linux so it would loop forever. > > Signed-off-by: Rob Clark <robdclark@gmail.com> Thanks, applied to efi-next Alex
diff --git a/lib/efi_loader/efi_memory.c b/lib/efi_loader/efi_memory.c index f59e3ea327..9e079f1fa3 100644 --- a/lib/efi_loader/efi_memory.c +++ b/lib/efi_loader/efi_memory.c @@ -407,11 +407,11 @@ efi_status_t efi_get_memory_map(unsigned long *memory_map_size, map_size = map_entries * sizeof(struct efi_mem_desc); + *memory_map_size = map_size; + if (provided_map_size < map_size) return EFI_BUFFER_TOO_SMALL; - *memory_map_size = map_size; - if (descriptor_size) *descriptor_size = sizeof(struct efi_mem_desc);
When booting shim -> fallback -> shim -> grub -> linux the memory map is a bit larger than the size linux passes in on the first call. But in the EFI_BUFFER_TOO_SMALL case we were not passing back the updated size to linux so it would loop forever. Signed-off-by: Rob Clark <robdclark@gmail.com> --- lib/efi_loader/efi_memory.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)