On 07/28/15 18:42, Ard Biesheuvel wrote:
> Now that GenFw correctly propagates the minimum alignment of the ELF
> input sections to the PE/COFF binary, we can simply select 'auto'
> alignment in the FDF Rule section instead of tweaking it by hand.
>
> Also add the FIXED FFS attribute to the module types that may execute
> in place. This enables a newly added optimization in GenFfs that strips
> redundant padding, preventing excessive waste of FV space if the section
> alignment is considerable (i.e., 2 KB or 4 KB)
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Ard Biesheuvel <ard.biesheu...@linaro.org>
> ---
> ArmVirtPkg/ArmVirtQemu.fdf | 12 ++++++------
> ArmVirtPkg/ArmVirtXen.fdf | 12 ++++++------
> 2 files changed, 12 insertions(+), 12 deletions(-)
>
> diff --git a/ArmVirtPkg/ArmVirtQemu.fdf b/ArmVirtPkg/ArmVirtQemu.fdf
> index 4ef0f8bb6aa4..3c0487cd95b6 100644
> --- a/ArmVirtPkg/ArmVirtQemu.fdf
> +++ b/ArmVirtPkg/ArmVirtQemu.fdf
> @@ -309,20 +309,20 @@ [FV.FVMAIN_COMPACT]
> ############################################################################
>
> [Rule.Common.SEC]
> - FILE SEC = $(NAMED_GUID) RELOCS_STRIPPED {
> - TE TE Align = 128 $(INF_OUTPUT)/$(MODULE_NAME).efi
> + FILE SEC = $(NAMED_GUID) RELOCS_STRIPPED FIXED {
> + TE TE Align = Auto $(INF_OUTPUT)/$(MODULE_NAME).efi
> }
>
> [Rule.Common.PEI_CORE]
> - FILE PEI_CORE = $(NAMED_GUID) {
> - TE TE Align = 8 $(INF_OUTPUT)/$(MODULE_NAME).efi
> + FILE PEI_CORE = $(NAMED_GUID) FIXED {
> + TE TE Align = Auto $(INF_OUTPUT)/$(MODULE_NAME).efi
> UI STRING ="$(MODULE_NAME)" Optional
> }
>
> [Rule.Common.PEIM]
> - FILE PEIM = $(NAMED_GUID) {
> + FILE PEIM = $(NAMED_GUID) FIXED {
> PEI_DEPEX PEI_DEPEX Optional $(INF_OUTPUT)/$(MODULE_NAME).depex
> - TE TE Align = 8 $(INF_OUTPUT)/$(MODULE_NAME).efi
> + TE TE Align = Auto $(INF_OUTPUT)/$(MODULE_NAME).efi
> UI STRING="$(MODULE_NAME)" Optional
> }
>
> diff --git a/ArmVirtPkg/ArmVirtXen.fdf b/ArmVirtPkg/ArmVirtXen.fdf
> index f98772b7191d..97cab4b058f2 100644
> --- a/ArmVirtPkg/ArmVirtXen.fdf
> +++ b/ArmVirtPkg/ArmVirtXen.fdf
> @@ -220,20 +220,20 @@ [FV.FVMAIN_COMPACT]
> ############################################################################
>
> [Rule.Common.SEC]
> - FILE SEC = $(NAMED_GUID) RELOCS_STRIPPED {
> - TE TE Align = 4K $(INF_OUTPUT)/$(MODULE_NAME).efi
> + FILE SEC = $(NAMED_GUID) RELOCS_STRIPPED FIXED {
> + TE TE Align = Auto $(INF_OUTPUT)/$(MODULE_NAME).efi
> }
>
> [Rule.Common.PEI_CORE]
> - FILE PEI_CORE = $(NAMED_GUID) {
> - TE TE Align = 8 $(INF_OUTPUT)/$(MODULE_NAME).efi
> + FILE PEI_CORE = $(NAMED_GUID) FIXED {
> + TE TE Align = Auto $(INF_OUTPUT)/$(MODULE_NAME).efi
> UI STRING ="$(MODULE_NAME)" Optional
> }
>
> [Rule.Common.PEIM]
> - FILE PEIM = $(NAMED_GUID) {
> + FILE PEIM = $(NAMED_GUID) FIXED {
> PEI_DEPEX PEI_DEPEX Optional $(INF_OUTPUT)/$(MODULE_NAME).depex
> - TE TE Align = 8 $(INF_OUTPUT)/$(MODULE_NAME).efi
> + TE TE Align = Auto $(INF_OUTPUT)/$(MODULE_NAME).efi
> UI STRING="$(MODULE_NAME)" Optional
> }
>
>
So, if I understand correctly, FIXED has no drawback even when building
with gcc, and when used in a clang build, it enables an optimization
that is *then* significant. And, because there is no penalty when using
gcc, it's simpler to add FIXED in a fixed way (pun intended) than to
create clang-specific rules (eg. by duplicating the current rules and
restricting them with *_XCODE5_* and *_GCC4x_* respectively).
... I hope the above mishmash can be called "review". :)
Reviewed-by: Laszlo Ersek <ler...@redhat.com>
_______________________________________________
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
