Re: [Buildroot] [PATCH 3/3] boot/edk2: bump to version edk2-stable202511

[Julien Olivain via buildroot <buildroot@buildroot.org>]

Hi Romain, All,

On 02/02/2026 10:51, Romain Naour via buildroot wrote:
> Hello Julien, All,
> 
> Le 01/02/2026 à 22:53, Julien Olivain a écrit :
>> Hi Romain,
>> 
>> On 01/02/2026 22:21, Romain Naour via buildroot wrote:
>>> Hello Julien,
>>> 
>>> Le 24/01/2026 à 14:13, Julien Olivain a écrit :
>>>> For release notes since edk2-stable202508, see:
>>>> https://github.com/tianocore/edk2/releases/tag/edk2-stable202511
>>>> 
>>>> This commit also updates the edk2-platforms packages with the last
>>>> commit merged at the edk2 release date (2025-11-19), which 
>>>> corresponds
>>>> to commit [1].
>>>> The edk2-non-osi package is also updated the same way, which
>>>> corresponds to commit [2].
>>>> 
>>>> This commit also adds an edk2-platforms package patch to fix a 
>>>> Marvell
>>>> build failure seen with the SolidRun MacchiatoBin platform.
>>>> 
>>>> This commit has been runtime tested with tests using EDK2 package,
>>>> with commands:
>>>> 
>>>>     support/testing/run-tests \
>>>>         -d dl -o output_folder \
>>>>             tests.boot.test_edk2 \
>>>>             tests.boot.test_grub.TestGrubAArch64EFI \
>>>>             tests.boot.test_grub.TestGrubRiscV64EFI \
>>>>             tests.boot.test_grub.TestGrubX8664EFI \
>>>>             tests.package.test_fwts
>>>> 
>>>> It has also been runtime tested (by booting in qemu) with defconfigs
>>>> using EDK2 package:
>>>> qemu_aarch64_sbsa_defconfig
>>>> qemu_loongarch64_virt_efi_defconfig
>>>> qemu_riscv64_virt_efi_defconfig
>>>> qemu_x86_64_efi_defconfig
>>>> 
>>>> [1] https://github.com/tianocore/edk2-platforms/
>>>> commit/1e64c1109ae2ac0185351e8d0c19732970d4fecf
>>>> [2] https://github.com/tianocore/edk2-non-osi/
>>>> commit/94d048981116e2e3eda52dad1a89958ee404098d
>>>> 
>>>> Cc: Dick Olsson <hi@senzilla.io>
>>>> Cc: Romain Naour <romain.naour@smile.fr>
>>>> Cc: Vincent Stehlé <vincent.stehle@arm.com>
>>>> Signed-off-by: Julien Olivain <ju.o@free.fr>
>>>> ---
>>>> The edk2-platforms patch contains CR-LF line endings, which might be 
>>>> changed
>>>> by the mailing list to LF. This can make the patch not apply 
>>>> cleanly. For
>>>> convenience, I also published those patches at:
>>>> https://gitlab.com/jolivain/buildroot/-/commits/bump_edk2-stable202511
>>>> ---
>>>>  Config.in.legacy                              |  7 ++
>>>>  boot/edk2/Config.in                           | 12 +--
>>>>  boot/edk2/edk2.hash                           |  2 +-
>>>>  boot/edk2/edk2.mk                             | 15 +---
>>>>  package/edk2-non-osi/edk2-non-osi.hash        |  2 +-
>>>>  package/edk2-non-osi/edk2-non-osi.mk          |  2 +-
>>>>  ...Marvell-fix-SMBIOS-type-7-structures.patch | 80 
>>>> +++++++++++++++++++
>>>>  package/edk2-platforms/edk2-platforms.hash    |  2 +-
>>>>  package/edk2-platforms/edk2-platforms.mk      |  2 +-
>>>>  9 files changed, 97 insertions(+), 27 deletions(-)
>>>>  create mode 100644 
>>>> package/edk2-platforms/0001-Silicon-Marvell-fix-SMBIOS-
>>>> type-7-structures.patch
>>>> 
>>> 
>>> ...
>>> 
>>>> diff --git a/boot/edk2/edk2.mk b/boot/edk2/edk2.mk
>>>> index 2c089aefca..dba4aed513 100644
>>>> --- a/boot/edk2/edk2.mk
>>>> +++ b/boot/edk2/edk2.mk
>>>> @@ -4,7 +4,7 @@
>>>>  #
>>>>  
>>>> ################################################################################
>>>> 
>>>> -EDK2_VERSION = edk2-stable202508
>>>> +EDK2_VERSION = edk2-stable202511
>>>>  EDK2_SITE = https://github.com/tianocore/edk2
>>>>  EDK2_SITE_METHOD = git
>>>>  EDK2_LICENSE = BSD-2-Clause-Patent
>>>> @@ -56,14 +56,7 @@ EDK2_GIT_SUBMODULES = YES
>>>>  EDK2_BUILD_PACKAGES = $(@D)/Build/Buildroot
>>>>  EDK2_PACKAGES_PATHS = $(@D) $(EDK2_BUILD_PACKAGES) 
>>>> $(STAGING_DIR)/usr/share/
>>>> edk2-platforms
>>>> 
>>>> -ifeq ($(BR2_TARGET_EDK2_PLATFORM_OVMF_I386),y)
>>>> -EDK2_ARCH = IA32
>>>> -EDK2_DEPENDENCIES += host-nasm
>>>> -EDK2_PACKAGE_NAME = OvmfPkg
>>>> -EDK2_PLATFORM_NAME = OvmfPkgIa32
>>>> -EDK2_BUILD_DIR = OvmfIa32
>>>> -
>>>> -else ifeq ($(BR2_TARGET_EDK2_PLATFORM_OVMF_X64),y)
>>>> +ifeq ($(BR2_TARGET_EDK2_PLATFORM_OVMF_X64),y)
>>>>  EDK2_ARCH = X64
>>>>  EDK2_DEPENDENCIES += host-nasm
>>>>  EDK2_PACKAGE_NAME = OvmfPkg
>>>> @@ -74,13 +67,13 @@ else ifeq 
>>>> ($(BR2_TARGET_EDK2_PLATFORM_ARM_VIRT_QEMU),y)
>>>>  EDK2_ARCH = AARCH64
>>>>  EDK2_PACKAGE_NAME = ArmVirtPkg
>>>>  EDK2_PLATFORM_NAME = ArmVirtQemu
>>>> -EDK2_BUILD_DIR = $(EDK2_PLATFORM_NAME)-$(EDK2_ARCH)
>>>> +EDK2_BUILD_DIR = $(EDK2_PLATFORM_NAME)-AArch64
>>>> 
>>>>  else ifeq ($(BR2_TARGET_EDK2_PLATFORM_ARM_VIRT_QEMU_KERNEL),y)
>>>>  EDK2_ARCH = AARCH64
>>>>  EDK2_PACKAGE_NAME = ArmVirtPkg
>>>>  EDK2_PLATFORM_NAME = ArmVirtQemuKernel
>>>> -EDK2_BUILD_DIR = $(EDK2_PLATFORM_NAME)-$(EDK2_ARCH)
>>>> +EDK2_BUILD_DIR = $(EDK2_PLATFORM_NAME)-AArch64
>>> 
>>> It seems the commit log doesn't explain this change.
>> 
>> You are correct. I forgot to include the explanation.
>> 
>> The following text should be added in the commit log:
>> 
>> """
>> Upstream commit [3] removed the 32bit ARM Virtual Platforms. The
>> specific change [4] changed a "$(ARCH)" to "AArch64". Possible
>> values for "$(ARCH)" were "AARCH64" or "ARM". Since Linux has
>> a case sensitive filesystem, the EDK2_BUILD_DIR values needs to be
>> updated accordingly.
>> 
>> [3] https://github.com/tianocore/edk2/commit/
>> b471ed29694c375ea1b6b484eb71af73ed61fece
>> [4] https://github.com/tianocore/edk2/commit/
>> b471ed29694c375ea1b6b484eb71af73ed61fece#diff-55200ac934bbe6e81bb4158bfff4797992b13fb1afb95fe982853024c6018430L20-R21
>> """
>> 
>> For convenience, I've also published this updated commit log at:
>> https://gitlab.com/jolivain/buildroot/-/commits/bump_edk2-stable202511_v2
>> 
>> Tell me if you want me to send a v2 on the list.
> 
> Thanks for the update.
> 
> For the upcoming 2026.02 release, I'm not sure if we should merge this 
> series
> right now. Indeed upstream EDK2 removed the IA32 support, so we have to 
> follow
> upstream change for the long-term but we suddenly remove it from 
> Buildroot while
> upstream still provide a specific branch for "any long term consumers".
> 
> In an other hand, the 2026.02 release will not be an LTS release (see
> https://buildroot.org/lts.html).
> 
> Thoughts?

Since 2026.02 will not be a Buildroot LTS version, it does not really 
matter
if this patch is applied now of just after the upcoming 2026.02 tag. It 
will
just extend by this IA32 and ARM32 support by 3 months (which is not 
much).

I am generally inclined to follow the upstream decisions and to keep
packages up to date. So I believe we will remove this support sooner of
later. I have no opinion if we should do that before or after the 
2026.02.
In any cases, the previous EDK2 version that still supports those
architectures are in the Buildroot LTS. So we could eventually pick
those EDK2 maintenance versions for Buildroot LTS branches.

I let others comment on that topic, if needed.

> Best regards,
> Romain
> 
> 
>> 
>>> Best regards,
>>> Romain
>> 
>> Best regards,
>> 
>> Julien.

Best regards,

Julien.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot