[PATCH] ARM: multi_v7_defconfig: enable CONFIG_EFI

[PATCH] ARM: multi_v7_defconfig: enable CONFIG_EFI

[Ard Biesheuvel]

This enables CONFIG_EFI for multi_v7_defconfig, which adds support for
booting via EFI, and for the EFI framebuffer as builtin options. It
also enables the EFI rtc, the EFI variable pseudo-filesystem and the
EFI capsule loader as modules.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---

We have been happily booting multi_v7_defconfig+CONFIG_EFI=y kernels on
kernelci for months now, so please consider enabling this by default.
The increase in compressed kernel footprint is ~30 KB, for the uncompressed
kernel it's ~10 KB, some of which is .init code.

 arch/arm/configs/multi_v7_defconfig | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/arch/arm/configs/multi_v7_defconfig b/arch/arm/configs/multi_v7_defconfig
index 2c8665cd9dc5..f30e8b44ea0d 100644
--- a/arch/arm/configs/multi_v7_defconfig
+++ b/arch/arm/configs/multi_v7_defconfig
@@ -129,6 +129,7 @@ CONFIG_CMA=y
 CONFIG_ARM_APPENDED_DTB=y
 CONFIG_ARM_ATAG_DTB_COMPAT=y
 CONFIG_KEXEC=y
+CONFIG_EFI=y
 CONFIG_CPU_FREQ=y
 CONFIG_CPU_FREQ_STAT_DETAILS=y
 CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND=y
@@ -593,6 +594,7 @@ CONFIG_DRM_PANEL_SIMPLE=y
 CONFIG_DRM_STI=m
 CONFIG_DRM_VC4=y
 CONFIG_FB_ARMCLCD=y
+CONFIG_FB_EFI=y
 CONFIG_FB_WM8505=y
 CONFIG_FB_SH_MOBILE_LCDC=y
 CONFIG_FB_SIMPLE=y
@@ -750,6 +752,7 @@ CONFIG_RTC_DRV_S35390A=m
 CONFIG_RTC_DRV_RX8581=m
 CONFIG_RTC_DRV_EM3027=y
 CONFIG_RTC_DRV_DA9063=m
+CONFIG_RTC_DRV_EFI=m
 CONFIG_RTC_DRV_DIGICOLOR=m
 CONFIG_RTC_DRV_S5M=m
 CONFIG_RTC_DRV_S3C=m
@@ -867,6 +870,8 @@ CONFIG_NVMEM=y
 CONFIG_NVMEM_SUNXI_SID=y
 CONFIG_BCM2835_MBOX=y
 CONFIG_RASPBERRYPI_FIRMWARE=y
+CONFIG_EFI_VARS=m
+CONFIG_EFI_CAPSULE_LOADER=m
 CONFIG_EXT4_FS=y
 CONFIG_AUTOFS4_FS=y
 CONFIG_MSDOS_FS=y
-- 
2.7.4

[PATCH] ARM: multi_v7_defconfig: enable CONFIG_EFI

[Arnd Bergmann]

On Thursday, September 15, 2016 3:28:20 PM CEST Ard Biesheuvel wrote:
> This enables CONFIG_EFI for multi_v7_defconfig, which adds support for
> booting via EFI, and for the EFI framebuffer as builtin options. It
> also enables the EFI rtc, the EFI variable pseudo-filesystem and the
> EFI capsule loader as modules.
> 
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> ---
> 
> We have been happily booting multi_v7_defconfig+CONFIG_EFI=y kernels on
> kernelci for months now, so please consider enabling this by default.
> The increase in compressed kernel footprint is ~30 KB, for the uncompressed
> kernel it's ~10 KB, some of which is .init code.
> 

Applied to next/defconfig, thanks!

	Arnd

[PATCH] ARM: multi_v7_defconfig: enable CONFIG_EFI

[Jon Hunter]

Hi Ard,

On 15/09/16 15:28, Ard Biesheuvel wrote:
> This enables CONFIG_EFI for multi_v7_defconfig, which adds support for
> booting via EFI, and for the EFI framebuffer as builtin options. It
> also enables the EFI rtc, the EFI variable pseudo-filesystem and the
> EFI capsule loader as modules.
> 
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> ---
> 
> We have been happily booting multi_v7_defconfig+CONFIG_EFI=y kernels on
> kernelci for months now, so please consider enabling this by default.
> The increase in compressed kernel footprint is ~30 KB, for the uncompressed
> kernel it's ~10 KB, some of which is .init code.

After this patch was applied the Tegra automated building for
multi_v7_defconfig has been failing on -next [0]. The error I am seeing
is as follows ...

  STUBCPY drivers/firmware/efi/libstub/lib-sort.stub.o
00000000 R_ARM_ABS32       sort
00000004 R_ARM_ABS32       __ksymtab_strings
drivers/firmware/efi/libstub/lib-sort.stub.o: absolute symbol references
not allowed in the EFI stub
make[4]: *** [drivers/firmware/efi/libstub/lib-sort.stub.o] Error 1
make[3]: *** [drivers/firmware/efi/libstub] Error 2
make[2]: *** [drivers/firmware/efi] Error 2
make[1]: *** [drivers/firmware] Error 2
make: *** [drivers] Error 2

The toolchain this builder uses (which is quite old admittedly) is ...

arm-linux-gnueabi-gcc (Ubuntu/Linaro 4.7.3-12ubuntu1) 4.7.3

I am not seeing this error on other machines but for the life of me, I
cannot figure out why this one is failing. Any thoughts?

Cheers
Jon

[0] http://nvtb.github.io/linux-next/		

-- 
nvpublic