clk: clk-fixed-factor: compile clk of provider only if it is enabled
Subsystems (4)
PATCH 1/1 clk: clk-fixed-factor: compile clk of provider only if it is enabled
From: Chali Anis <[email protected]>
This permits to not compile when it's not used, fix undefined reference error.
--- drivers/clk/clk-fixed-factor.c | 3 +++ 1 file changed, 3 insertions(+)
drivers/clk/clk-fixed-factor.c +3 -0
@@ -97,6 +97,8 @@ struct clk_hw *clk_hw_register_fixed_factor(struct device *dev, flags, mult, div)); } +#ifdef CONFIG_COMMON_CLK_OF_PROVIDER +/** * of_fixed_factor_clk_setup() - Setup function for simple fixed factor clock */
@@ -130,3 +132,4 @@ static int of_fixed_factor_clk_setup(struct device_node *node) }CLK_OF_DECLARE (fixed_factor_clk,"fixed-factor-clock" , of_fixed_factor_clk_setup); +#endif
Signature · 1 line
2.34.1
PATCH 1/1 arm: efi: add a dummy efi arch on ARM64
From: Chali Anis <[email protected]>
this arch will forces a generic arm 64 cpu, which will produce an efi payload on ARM 64, this works with a virtual machine using qemu-system-aarch64, need further work to support real hardware.
Submit the patch in case someone is interrested in working to have a full support for efi payload on real hardware.
--- Documentation/boards/efi.rst | 11 +++- arch/arm/Kconfig | 1 + arch/arm/configs/efi_defconfig | 101 +++++++++++++++++++++++++++++++++ arch/arm/mach-efi/Kconfig | 10 ++++ arch/arm/mach-efi/Makefile | 1 + 5 files changed, 123 insertions(+), 1 deletion(-) create mode 100644 arch/arm/configs/efi_defconfig create mode 100644 arch/arm/mach-efi/Kconfig create mode 100644 arch/arm/mach-efi/Makefile
Documentation/boards/efi.rst +10 -1
@@ -23,7 +23,7 @@ x86_64 binary currently is tested. Building barebox for EFI ------------------------ -Use the following to build barebox for EFI: +Use the following to build barebox for EFI for x86: .. code-block:: sh
@@ -31,6 +31,15 @@ Use the following to build barebox for EFI: make efi_defconfig make +for EFI on arm 64 use: + +.. code-block:: sh + + export ARCH=arm + export CROSS_COMPILE=aarch64-linux-gnu- + make efi_defconfig + make + The resulting EFI image is ``barebox.efi`` (or the barebox-flash-image link). Running barebox on EFI systems
arch/arm/Kconfig +1 -0
@@ -276,6 +276,7 @@ source "arch/arm/mach-vexpress/Kconfig" source "arch/arm/mach-tegra/Kconfig" source "arch/arm/mach-zynq/Kconfig" source "arch/arm/mach-zynqmp/Kconfig" +source "arch/arm/mach-efi/Kconfig" config BOARD_ARM_VIRT bool
arch/arm/configs/efi_defconfig new file +101 -0
101 lines not on this page · raw message
arch/arm/mach-efi/Kconfig new file +10 -0
@@ -0,0 +1,10 @@ +config EFI_ARM64 + bool "EFI on ARM64" + select CPU_V8 + select CPU_SUPPORTS_64BIT_KERNEL + select HAVE_EFI_PAYLOAD + select EFI_PAYLOAD + select CLOCKSOURCE_EFI + depends on 64BIT + default y +
arch/arm/mach-efi/Makefile new file +1 -0
@@ -0,0 +1 @@ +obj- := __dummy__.o
Signature · 1 line
2.34.1
Re: [PATCH] arm: efi: add a dummy efi arch on ARM64
Hi,
Quoted 8 lines · On 30.07.25 05:50, [email protected] wrote:
From: Chali Anis <[email protected]>
this arch will forces a generic arm 64 cpu, which will produce an efi payload on ARM 64, this works with a virtual machine using qemu-system-aarch64, need further work to support real hardware.
Submit the patch in case someone is interrested in working to have a full support for efi payload on real hardware.
Thanks for your patch.
For the Virt platforms we decided to use the multi defconfigs to cut down on the number of configs. I think efi_defconfig is useful to have, because it shares little with other existing configs, but it might confuse users to provide it before it is a viable target, i.e. can actually boot a kernel via EFI...
I just submitted an alternative patch for multi_v8_efi_defconfig, which could be used as stepping stone for further work.
What do you think?
Quoted 36 lines
Documentation/boards/efi.rst +10
-1-Use the following to build barebox for EFI: +Use the following to build barebox for EFI for x86: .. code-block:: sh@@ -31,6 +31,15 @@ Use the following to build barebox for EFI: make efi_defconfig make +for EFI on arm 64 use: + +.. code-block:: sh + + export ARCH=arm + export CROSS_COMPILE=aarch64-linux-gnu- + make efi_defconfig + make + The resulting EFI image is ``barebox.efi`` (or the barebox-flash-image link). Running barebox on EFI systemsarch/arm/Kconfig +1
-0@@ -276,6 +276,7 @@ source "arch/arm/mach-vexpress/Kconfig" source "arch/arm/mach-tegra/Kconfig" source "arch/arm/mach-zynq/Kconfig" source "arch/arm/mach-zynqmp/Kconfig" +source "arch/arm/mach-efi/Kconfig" config BOARD_ARM_VIRT boolarch/arm/configs/efi_defconfig new file +0
-0
arch/arm/configs/efi_defconfig line 84
Quoted 4 lines
arch/arm/configs/efi_defconfig new file +3
-0+CONFIG_VIDEO=y +CONFIG_FRAMEBUFFER_CONSOLE=y +CONFIG_DRIVER_VIDEO_EFI_GOP=y
This doesn't work correctly on ARM yet
arch/arm/configs/efi_defconfig line 86
Quoted 2 lines
arch/arm/configs/efi_defconfig new file +2
-0+CONFIG_FINTEK_SUPERIO=y +CONFIG_SMSC_SUPERIO=y
These are likely not used with ARM systems
arch/arm/configs/efi_defconfig line 91
Quoted 5 lines
arch/arm/configs/efi_defconfig new file +5
-0+CONFIG_STATE_DRV=y +CONFIG_WATCHDOG=y +CONFIG_WATCHDOG_EFI=y +CONFIG_F71808E_WDT=y +CONFIG_ITCO_WDT=y
Not relevant to ARM systems.
arch/arm/configs/efi_defconfig line 6
Quoted 8 lines
+++ b/arch/arm/mach-efi/Kconfig
arch/arm/configs/efi_defconfig new file +6
-0@@ -0,0 +1,10 @@ +config EFI_ARM64 + bool "EFI on ARM64" + select CPU_V8 + select CPU_SUPPORTS_64BIT_KERNEL + select HAVE_EFI_PAYLOAD + select EFI_PAYLOAD
EFI_PAYLOAD is still gated behind COMPILE_TEST, so I think this should result in a warning.
arch/arm/configs/efi_defconfig line 7
Quoted 1 line
arch/arm/configs/efi_defconfig new file +1
-0+ select CLOCKSOURCE_EFI
Wrong indentation.
Cheers,
Ahmad
Signature · 4 lines
Pengutronix e.K. | | Steuerwalder Str. 21 | http://www.pengutronix.de/ | 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
Hi Anis,
Quoted 4 lines · On 30.07.25 05:50, [email protected] wrote:
From: Chali Anis <[email protected]>
This permits to not compile when it's not used, fix undefined reference error.
Normally undefined references are garbage collected by the linker if the code that references them is going to be discarded as well.
We have configurations with and without OF support that make use of clocks in CI I think, so I am curious to know what configuration you used to provoke the error.
Cheers,
Ahma
Quoted 24 lines
Signed-off-by Chali Anis <[email protected]>--- drivers/clk/clk-fixed-factor.c | 3 +++ 1 file changed, 3 insertions(+)drivers/clk/clk-fixed-factor.c +3
-0@@ -97,6 +97,8 @@ struct clk_hw *clk_hw_register_fixed_factor(struct device *dev, flags, mult, div)); } +#ifdef CONFIG_COMMON_CLK_OF_PROVIDER +/** * of_fixed_factor_clk_setup() - Setup function for simple fixed factor clock */ @@ -130,3 +132,4 @@ static int of_fixed_factor_clk_setup(struct device_node *node) }CLK_OF_DECLARE (fixed_factor_clk,"fixed-factor-clock" , of_fixed_factor_clk_setup); +#endif
Signature · 4 lines
Pengutronix e.K. | | Steuerwalder Str. 21 | http://www.pengutronix.de/ | 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
A message that is not in the archive (sent off-list or not yet received): CAL+1fyCC0V3MO4tmgBJumbdPn7_BhQc-60s3QhGjye0rQQVpPw@mail.gmail.com
Hi,
[adding back barebox mailing list]
Quoted 7 lines · anis chali
Hi Ahmad,
I will check back tomorrow, but I faced the issue when I compiled efi on arm64 using the clocksource efi will compile the stuff inside the fixed clock and there is a code using of variant that raised an undefined reference but maybe it's all good since arm64 support device trees, so adding the config to enable of tree support will eliminate the issue.
If there is a build error, it should be fixed. I was just wondering how you ran into it.
Cheers,
Ahmad
Quoted 69 lines
Cheers,
AnisLe mer. 30 juil. 2025, 02:23, Ahmad Fatoum <[email protected] <mailto:[email protected]>> a écrit :
Hi Anis, On 30.07.25 05:50, [email protected] <mailto:[email protected]> wrote: > From: Chali Anis <[email protected] <mailto:[email protected]>> > > This permits to not compile when it's not used, fix undefined > reference error. Normally undefined references are garbage collected by the linker if the code that references them is going to be discarded as well. We have configurations with and without OF support that make use of clocks in CI I think, so I am curious to know what configuration you used to provoke the error. Cheers, Ahma > > Signed-off-by: Chali Anis <[email protected] <mailto:[email protected]>> > --- > drivers/clk/clk-fixed-factor.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/drivers/clk/clk-fixed-factor.c b/drivers/clk/clk- fixed-factor.c > index d2c808d40c63..ea04a4bf7f0f 100644 > --- a/drivers/clk/clk-fixed-factor.c > +++ b/drivers/clk/clk-fixed-factor.c > @@ -97,6 +97,8 @@ struct clk_hw *clk_hw_register_fixed_factor(struct device *dev, > flags, mult, div)); > } > > +#ifdef CONFIG_COMMON_CLK_OF_PROVIDER > + > /** > * of_fixed_factor_clk_setup() - Setup function for simple fixed factor clock > */ > @@ -130,3 +132,4 @@ static int of_fixed_factor_clk_setup(struct device_node *node) > } > CLK_OF_DECLARE(fixed_factor_clk, "fixed-factor-clock", > of_fixed_factor_clk_setup); > +#endif -- Pengutronix e.K. | | Steuerwalder Str. 21 | http:// www.pengutronix.de/ <http://www.pengutronix.de/> | 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
Signature · 4 lines
Pengutronix e.K. | | Steuerwalder Str. 21 | http://www.pengutronix.de/ | 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |