RISC-V: kconfig.socs cleanup, part 1
Subsystems (6)
Conor Dooley
maintainercover RISC-V: kconfig.socs cleanup, part 1
From: Conor Dooley <[email protected]>
Hey all,
Following on from my RFC series a little bit ago, I'm back again with an attempted conversion of our SOC_FOO symbols to a more uniform ARCH_FOO. Last time around, I had about 30 patches [0], but here I have a more modest offering.
Patch 1 adds stubs, that ideally would be applied for v6.2-rc1 (but after week 1, see the note in that patch for more), which would allow for converting drivers etc over a release cycle (or more) while allowing everything else to keep working in the meantime.
Christoph Hellwig & Damien voiced some objections in the original thread, hopefully the gradual conversion assuages Damien's concerns about breaking defconfigs - but I am not sure what to do about Christoph's dislike of the ARCH_FOO stuff. A change the world approach is unlikely to be well received & introducing "duplicate" symbols for cross-arch vendors would surely be similarly well received? I don't mind doing it, but my instinct is that that is a futile use of time.
The rest of this series contains v6.3 material, converting SOC_FOO usages in arch/riscv to ARCH_FOO.
Thanks,
Conor.
[0] - https://lore.kernel.org/linux-riscv/[email protected]/
Conor Dooley (7):
RISC-V: introduce ARCH_FOO kconfig aliases for SOC_FOO symbols
RISC-V: kconfig.socs: convert usage of SOC_CANAAN to ARCH_CANAAN
RISC-V: kbuild: convert all use of SOC_FOO to ARCH_FOO
RISC-V: stop selecting SIFIVE_PLIC at the SoC level
RISC-V: stop selecting the PolarFire SoC clock driver
RISC-V: stop selecting SiFive clock and serial drivers directly
RISC-V: stop directly selecting drivers for SOC_CANAAN
arch/riscv/Kconfig | 1 + arch/riscv/Kconfig.socs | 44 +++++++++++++++----------- arch/riscv/Makefile | 2 +- arch/riscv/boot/dts/Makefile | 2 +- arch/riscv/boot/dts/canaan/Makefile | 14 ++++---- arch/riscv/boot/dts/microchip/Makefile | 8 ++--- arch/riscv/boot/dts/sifive/Makefile | 4 +-- arch/riscv/boot/dts/starfive/Makefile | 2 +- 8 files changed, 43 insertions(+), 34 deletions(-)
Signature · 7 lines
2.37.2 _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Conor Dooley
maintainerPATCH 2/7 RISC-V: kconfig.socs: convert usage of SOC_CANAAN to ARCH_CANAAN
From: Conor Dooley <[email protected]>
While we cannot yet drop the SOC_ prefixed symbols, we can convert uses of these symbols within Kconfig.socs to the ARCH_ variants.
--- arch/riscv/Kconfig.socs | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
arch/riscv/Kconfig.socs +5 -5
@@ -75,14 +75,14 @@ config SOC_CANAAN help This enables support for Canaan Kendryte K210 SoC platform hardware. -if SOC_CANAAN +if ARCH_CANAAN config ARCH_CANAAN_K210_DTB_BUILTIN def_bool SOC_CANAAN_K210_DTB_BUILTIN config SOC_CANAAN_K210_DTB_BUILTIN bool "Builtin device tree for the Canaan Kendryte K210" - depends on SOC_CANAAN + depends on ARCH_CANAAN default y select OF select BUILTIN_DTB
@@ -96,14 +96,14 @@ config ARCH_CANAAN_K210_DTB_SOURCE config SOC_CANAAN_K210_DTB_SOURCE string "Source file for the Canaan Kendryte K210 builtin DTB" - depends on SOC_CANAAN - depends on SOC_CANAAN_K210_DTB_BUILTIN + depends on ARCH_CANAAN + depends on ARCH_CANAAN_K210_DTB_BUILTIN default "k210_generic" help Base name (without suffix, relative to arch/riscv/boot/dts/canaan) for the DTS file that will be used to produce the DTB linked into the kernel. -endif # SOC_CANAAN +endif # ARCH_CANAAN endmenu # "SoC selection"
Signature · 7 lines
2.37.2 _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Conor Dooley
maintainerPATCH 3/7 RISC-V: kbuild: convert all use of SOC_FOO to ARCH_FOO
In reply to Conor DooleyFrom: Conor Dooley <[email protected]>
Convert all non user visible use of SOC_FOO symbols to their ARCH_FOO variants. The canaan DTs are an outlier in that they're gated at the directory and the file level. Drop the directory level gating while we are swapping the symbol names over.
--- arch/riscv/Makefile | 2 +- arch/riscv/boot/dts/Makefile | 2 +- arch/riscv/boot/dts/canaan/Makefile | 14 +++++++------- arch/riscv/boot/dts/microchip/Makefile | 8 ++++---- arch/riscv/boot/dts/sifive/Makefile | 4 ++-- arch/riscv/boot/dts/starfive/Makefile | 2 +- 6 files changed, 16 insertions(+), 16 deletions(-)
arch/riscv/Makefile +1 -1
@@ -131,7 +131,7 @@ endif endif ifneq ($(CONFIG_XIP_KERNEL),y) -ifeq ($(CONFIG_RISCV_M_MODE)$(CONFIG_SOC_CANAAN),yy) +ifeq ($(CONFIG_RISCV_M_MODE)$(CONFIG_ARCH_CANAAN),yy) KBUILD_IMAGE := $(boot)/loader.bin else ifeq ($(CONFIG_EFI_ZBOOT),)
arch/riscv/boot/dts/Makefile +1 -1
@@ -1,7 +1,7 @@ # SPDX-License-Identifier: GPL-2.0 subdir-y += sifive subdir-y += starfive -subdir-$(CONFIG_SOC_CANAAN_K210_DTB_BUILTIN) += canaan +subdir-y += canaan subdir-y += microchip subdir-y += renesas
arch/riscv/boot/dts/canaan/Makefile +7 -7
@@ -1,9 +1,9 @@ # SPDX-License-Identifier: GPL-2.0 -dtb-$(CONFIG_SOC_CANAAN) += canaan_kd233.dtb -dtb-$(CONFIG_SOC_CANAAN) += k210_generic.dtb -dtb-$(CONFIG_SOC_CANAAN) += sipeed_maix_bit.dtb -dtb-$(CONFIG_SOC_CANAAN) += sipeed_maix_dock.dtb -dtb-$(CONFIG_SOC_CANAAN) += sipeed_maix_go.dtb -dtb-$(CONFIG_SOC_CANAAN) += sipeed_maixduino.dtb +dtb-$(CONFIG_ARCH_CANAAN) += canaan_kd233.dtb +dtb-$(CONFIG_ARCH_CANAAN) += k210_generic.dtb +dtb-$(CONFIG_ARCH_CANAAN) += sipeed_maix_bit.dtb +dtb-$(CONFIG_ARCH_CANAAN) += sipeed_maix_dock.dtb +dtb-$(CONFIG_ARCH_CANAAN) += sipeed_maix_go.dtb +dtb-$(CONFIG_ARCH_CANAAN) += sipeed_maixduino.dtb -obj-$(CONFIG_SOC_CANAAN_K210_DTB_BUILTIN) += $(addsuffix .dtb.o, $(CONFIG_SOC_CANAAN_K210_DTB_SOURCE)) +obj-$(CONFIG_ARCH_CANAAN_K210_DTB_BUILTIN) += $(addsuffix .dtb.o, $(CONFIG_ARCH_CANAAN_K210_DTB_SOURCE))
arch/riscv/boot/dts/microchip/Makefile +4 -4
@@ -1,6 +1,6 @@ # SPDX-License-Identifier: GPL-2.0 -dtb-$(CONFIG_SOC_MICROCHIP_POLARFIRE) += mpfs-icicle-kit.dtb -dtb-$(CONFIG_SOC_MICROCHIP_POLARFIRE) += mpfs-m100pfsevp.dtb -dtb-$(CONFIG_SOC_MICROCHIP_POLARFIRE) += mpfs-polarberry.dtb -dtb-$(CONFIG_SOC_MICROCHIP_POLARFIRE) += mpfs-sev-kit.dtb +dtb-$(CONFIG_ARCH_MICROCHIP_POLARFIRE) += mpfs-icicle-kit.dtb +dtb-$(CONFIG_ARCH_MICROCHIP_POLARFIRE) += mpfs-m100pfsevp.dtb +dtb-$(CONFIG_ARCH_MICROCHIP_POLARFIRE) += mpfs-polarberry.dtb +dtb-$(CONFIG_ARCH_MICROCHIP_POLARFIRE) += mpfs-sev-kit.dtb obj-$(CONFIG_BUILTIN_DTB) += $(addsuffix .o, $(dtb-y))
arch/riscv/boot/dts/sifive/Makefile +2 -2
@@ -1,4 +1,4 @@ # SPDX-License-Identifier: GPL-2.0 -dtb-$(CONFIG_SOC_SIFIVE) += hifive-unleashed-a00.dtb \ - hifive-unmatched-a00.dtb +dtb-$(CONFIG_ARCH_SIFIVE) += hifive-unleashed-a00.dtb \ + hifive-unmatched-a00.dtb obj-$(CONFIG_BUILTIN_DTB) += $(addsuffix .o, $(dtb-y))
arch/riscv/boot/dts/starfive/Makefile +1 -1
@@ -1,2 +1,2 @@ # SPDX-License-Identifier: GPL-2.0 -dtb-$(CONFIG_SOC_STARFIVE) += jh7100-beaglev-starlight.dtb jh7100-starfive-visionfive-v1.dtb +dtb-$(CONFIG_ARCH_STARFIVE) += jh7100-beaglev-starlight.dtb jh7100-starfive-visionfive-v1.dtb
Signature · 7 lines
2.37.2 _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Conor Dooley
maintainerPATCH 4/7 RISC-V: stop selecting SIFIVE_PLIC at the SoC level
In reply to Conor DooleyFrom: Conor Dooley <[email protected]>
The SIFIVE_PLIC driver is used by all current RISC-V SoCs & will be, where possible, used for future implementations. Rather than having each SoC select the option in Kconfig.socs, do so at the arch level.
--- arch/riscv/Kconfig | 1 + arch/riscv/Kconfig.socs | 5 ----- 2 files changed, 1 insertion(+), 5 deletions(-)
arch/riscv/Kconfig +1 -0
@@ -125,6 +125,7 @@ config RISCV select PCI_MSI if PCI select RISCV_INTC select RISCV_TIMER if RISCV_SBI + select SIFIVE_PLIC select SPARSE_IRQ select SYSCTL_EXCEPTION_TRACE select THREAD_INFO_IN_TASK
arch/riscv/Kconfig.socs +0 -5
@@ -6,7 +6,6 @@ config ARCH_MICROCHIP_POLARFIRE config SOC_MICROCHIP_POLARFIRE bool "Microchip PolarFire SoCs" select MCHP_CLK_MPFS - select SIFIVE_PLIC help This enables support for Microchip PolarFire SoC platforms.
@@ -24,7 +23,6 @@ config SOC_SIFIVE select SERIAL_SIFIVE_CONSOLE if TTY select CLK_SIFIVE select CLK_SIFIVE_PRCI - select SIFIVE_PLIC select ERRATA_SIFIVE if !XIP_KERNEL help This enables support for SiFive SoC platform hardware.
@@ -36,7 +34,6 @@ config SOC_STARFIVE bool "StarFive SoCs" select PINCTRL select RESET_CONTROLLER - select SIFIVE_PLIC help This enables support for StarFive SoC platform hardware.
@@ -51,7 +48,6 @@ config SOC_VIRT select POWER_RESET_SYSCON_POWEROFF select GOLDFISH select RTC_DRV_GOLDFISH if RTC_CLASS - select SIFIVE_PLIC select PM_GENERIC_DOMAINS if PM select PM_GENERIC_DOMAINS_OF if PM && OF select RISCV_SBI_CPUIDLE if CPU_IDLE && RISCV_SBI
@@ -67,7 +63,6 @@ config SOC_CANAAN select CLINT_TIMER if RISCV_M_MODE select SERIAL_SIFIVE if TTY select SERIAL_SIFIVE_CONSOLE if TTY - select SIFIVE_PLIC select ARCH_HAS_RESET_CONTROLLER select PINCTRL select COMMON_CLK
Signature · 7 lines
2.37.2 _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Conor Dooley
maintainerPATCH 5/7 RISC-V: stop selecting the PolarFire SoC clock driver
In reply to Conor DooleyFrom: Conor Dooley <[email protected]>
The driver is now enabled by default if SOC_MICROCHIP_POLARFIRE so there is no longer a need to select it in Kconfig.socs
--- arch/riscv/Kconfig.socs | 1 - 1 file changed, 1 deletion(-)
arch/riscv/Kconfig.socs +0 -1
@@ -5,7 +5,6 @@ config ARCH_MICROCHIP_POLARFIRE config SOC_MICROCHIP_POLARFIRE bool "Microchip PolarFire SoCs" - select MCHP_CLK_MPFS help This enables support for Microchip PolarFire SoC platforms.
Signature · 7 lines
2.37.2 _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Conor Dooley
maintainerPATCH 6/7 RISC-V: stop selecting SiFive clock and serial drivers directly
In reply to Conor DooleyFrom: Conor Dooley <[email protected]>
The SiFive clock and serial drivers will now default to the value of SOC_SIFIVE so there is no need to directly select their symbols anymore.
--- arch/riscv/Kconfig.socs | 4 ---- 1 file changed, 4 deletions(-)
arch/riscv/Kconfig.socs +0 -4
@@ -18,10 +18,6 @@ config ARCH_SIFIVE config SOC_SIFIVE bool "SiFive SoCs" - select SERIAL_SIFIVE if TTY - select SERIAL_SIFIVE_CONSOLE if TTY - select CLK_SIFIVE - select CLK_SIFIVE_PRCI select ERRATA_SIFIVE if !XIP_KERNEL help This enables support for SiFive SoC platform hardware.
Signature · 7 lines
2.37.2 _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Conor Dooley
maintainerPATCH 1/7 RISC-V: introduce ARCH_FOO kconfig aliases for SOC_FOO symbols
In reply to Conor DooleyFrom: Conor Dooley <[email protected]>
To facilitate a transfer from SOC_FOO to ARCH_FOO, over a release cycle, introduce some aliases so that drivers etc that use the SOC_FOO symbols can be converted.
---
To me, the most straight-forward conversion looks like so:
- this patch is applied in week 2 of the merge window, to avoid any conflicts with the Renesas tree
- all users of the SOC_ variants can be converted over a release cycle (or more) & no trees need to merge an immutable branch.
- we convert defconfig etc over after all users are converted
- doing it over at least one release cycle means that make oldconfig will keep people's configs working as they upgrade
- any new SoC families added uses ARCH_FOO
--- arch/riscv/Kconfig.socs | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+)
arch/riscv/Kconfig.socs +21 -0
@@ -1,5 +1,8 @@ menu "SoC selection" +config ARCH_MICROCHIP_POLARFIRE + def_bool SOC_MICROCHIP_POLARFIRE + config SOC_MICROCHIP_POLARFIRE bool "Microchip PolarFire SoCs" select MCHP_CLK_MPFS
@@ -12,6 +15,9 @@ config ARCH_RENESAS help This enables support for the RISC-V based Renesas SoCs. +config ARCH_SIFIVE + def_bool SOC_SIFIVE + config SOC_SIFIVE bool "SiFive SoCs" select SERIAL_SIFIVE if TTY
@@ -23,6 +29,9 @@ config SOC_SIFIVE help This enables support for SiFive SoC platform hardware. +config ARCH_STARFIVE + def_bool SOC_STARFIVE + config SOC_STARFIVE bool "StarFive SoCs" select PINCTRL
@@ -31,6 +40,9 @@ config SOC_STARFIVE help This enables support for StarFive SoC platform hardware. +config ARCH_VIRT + def_bool SOC_VIRT + config SOC_VIRT bool "QEMU Virt Machine" select CLINT_TIMER if RISCV_M_MODE
@@ -46,6 +58,9 @@ config SOC_VIRT help This enables support for QEMU Virt Machine. +config ARCH_CANAAN + def_bool SOC_CANAAN + config SOC_CANAAN bool "Canaan Kendryte K210 SoC" depends on !MMU
@@ -62,6 +77,9 @@ config SOC_CANAAN if SOC_CANAAN +config ARCH_CANAAN_K210_DTB_BUILTIN + def_bool SOC_CANAAN_K210_DTB_BUILTIN + config SOC_CANAAN_K210_DTB_BUILTIN bool "Builtin device tree for the Canaan Kendryte K210" depends on SOC_CANAAN
@@ -73,6 +91,9 @@ config SOC_CANAAN_K210_DTB_BUILTIN This option should be selected if no bootloader is being used. If unsure, say Y. +config ARCH_CANAAN_K210_DTB_SOURCE + def_bool SOC_CANAAN_K210_DTB_SOURCE + config SOC_CANAAN_K210_DTB_SOURCE string "Source file for the Canaan Kendryte K210 builtin DTB" depends on SOC_CANAAN
Signature · 7 lines
2.37.2 _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Conor Dooley
maintainerPATCH 7/7 RISC-V: stop directly selecting drivers for SOC_CANAAN
In reply to Conor DooleyFrom: Conor Dooley <[email protected]>
The serial and clock drivers will be enabled by default if the symbol itself is enabled, so stop directly selecting the drivers in Kconfigs.socs.
--- arch/riscv/Kconfig.socs | 3 --- 1 file changed, 3 deletions(-)
arch/riscv/Kconfig.socs +0 -3
@@ -56,12 +56,9 @@ config SOC_CANAAN bool "Canaan Kendryte K210 SoC" depends on !MMU select CLINT_TIMER if RISCV_M_MODE - select SERIAL_SIFIVE if TTY - select SERIAL_SIFIVE_CONSOLE if TTY select ARCH_HAS_RESET_CONTROLLER select PINCTRL select COMMON_CLK - select COMMON_CLK_K210 help This enables support for Canaan Kendryte K210 SoC platform hardware.
Signature · 7 lines
2.37.2 _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Re: [PATCH v1 3/7] RISC-V: kbuild: convert all use of SOC_FOO to ARCH_FOO
In reply to Conor DooleyQuoted 6 lines · Conor Dooley
From: Conor Dooley <[email protected]>
Convert all non user visible use of
SOC_FOOsymbols to theirARCH_FOOvariants. The canaan DTs are an outlier in that they're gated at the directory and the file level. Drop the directory level gating while we are swapping the symbol names over.
I still very much disagree with this. ARCH_FOO for something that is not a Linux architecture is a really really bad idea. We need to fix arm and not convert riscv, which got this right.
_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv
Re: [PATCH v1 5/7] RISC-V: stop selecting the PolarFire SoC clock driver
In reply to Conor DooleyHi Conor,
Thanks for your patch!
Quoted 4 lines · Conor Dooley
From: Conor Dooley <[email protected]>
The driver is now enabled by default if
SOC_MICROCHIP_POLARFIREso there is no longer a need to select it inKconfig.socs
Where is it enabled by default?
I cannot seem to find that location...
arch/riscv/Kconfig.socs config ARCH_MICROCHIP_POLARFIRE · line 9
Quoted 12 lines
Signed-off-by Conor Dooley <[email protected]>arch/riscv/Kconfig.socs +0
-1@@ -5,7 +5,6 @@ config ARCH_MICROCHIP_POLARFIRE config SOC_MICROCHIP_POLARFIRE bool "Microchip PolarFire SoCs" - select MCHP_CLK_MPFS help This enables support for Microchip PolarFire SoC platforms.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]
In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv
[email protected]
maintainerRe: [PATCH v1 5/7] RISC-V: stop selecting the PolarFire SoC clock driver
Quoted 8 lines · Geert Uytterhoeven
Quoted 4 lines · Conor Dooley
From: Conor Dooley <[email protected]>
The driver is now enabled by default if
SOC_MICROCHIP_POLARFIREso there is no longer a need to select it inKconfig.socsWhere is it enabled by default?
I cannot seem to find that location...
Huh, you're right.. I could have sworn I had submitted a patch for that but it turns out I only submitted a patch to make the SiFive clk stuff default on. I suppose my defconfig has an explicit set in it and I never noticed.
Thanks Geert!
_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv
Palmer Dabbelt
maintainerRe: [PATCH v1 3/7] RISC-V: kbuild: convert all use of SOC_FOO to ARCH_FOO
In reply to Christoph HellwigQuoted 9 lines · Christoph Hellwig
Quoted 6 lines · Conor Dooley
From: Conor Dooley <[email protected]>
Convert all non user visible use of
SOC_FOOsymbols to theirARCH_FOOvariants. The canaan DTs are an outlier in that they're gated at the directory and the file level. Drop the directory level gating while we are swapping the symbol names over.
Quoted 3 lines
I still very much disagree with this.
ARCH_FOOfor something that is not a Linux architecture is a really really bad idea. We need to fix arm and not convert riscv, which got this right.
Makes sense, though I guess I'd argue that SOC is also a wrong name (but at least it doesn't conflict with the ARCH_* symbols that already exist). That'll touch a bunch of trees, adding the Arm folks to see if they're on board.
I don't care a ton either way, my only opinion here is that RISC-V and Arm should use the same scheme. If the Arm folks want to move over to SOC_* that works for me, and has the advantage of being no work on my end ;)
Re: [PATCH v1 1/7] RISC-V: introduce ARCH_FOO kconfig aliases for SOC_FOO symbols
In reply to Conor DooleyHi Conor,
Quoted 7 lines · Conor Dooley
From: Conor Dooley <[email protected]>
To facilitate a transfer from
SOC_FOOtoARCH_FOO, over a release cycle, introduce some aliases so that drivers etc that use theSOC_FOOsymbols can be converted.Signed-off-by Conor Dooley <[email protected]>
Thanks for your patch, which is now commit fc43211939bb6874 ("RISC-V: kconfig.socs: convert usage of SOC_CANAAN to ARCH_CANAAN") in riscv/for-next
arch/riscv/Kconfig.socs config SOC_CANAAN_K210_DTB_BUILTIN · line 95
Quoted 9 lines
arch/riscv/Kconfig.socs +2
-0@@ -73,6 +91,9 @@ config SOC_CANAAN_K210_DTB_BUILTIN This option should be selected if no bootloader is being used. If unsure, say Y. +config ARCH_CANAAN_K210_DTB_SOURCE + def_bool SOC_CANAAN_K210_DTB_SOURCE
This is not correct, as SOC_CANAAN_K210_DTB_SOURCE below is not a bool, but a string.
arch/riscv/Kconfig.socs config SOC_CANAAN_K210_DTB_BUILTIN · line 99
Quoted 4 lines
arch/riscv/Kconfig.socs +1
-0+ config SOC_CANAAN_K210_DTB_SOURCE string "Source file for the Canaan Kendryte K210 builtin DTB" depends on SOC_CANAAN
Hence
obj-$(CONFIG_ARCH_CANAAN_K210_DTB_BUILTIN) += $(addsuffix .dtb.o,
$(CONFIG_ARCH_CANAAN_K210_DTB_SOURCE))
will do the wrong thing later, and I get a non-bootable system (no output) on my MAiX-BiT.
Unfortunately there is no def_string, so I don't think we can fix this in a backwards-compatible way, and have to replace all SOC_CANAAN_K210_DTB_SOURCE by ARCH_CANAAN_K210_DTB_SOURCE, and urging users to update their .config manually.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]
In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv
Conor Dooley
maintainerRe: [PATCH v1 1/7] RISC-V: introduce ARCH_FOO kconfig aliases for SOC_FOO symbols
Hi Geert...
Quoted 7 lines · Geert Uytterhoeven
Quoted 6 lines · Conor Dooley
From: Conor Dooley <[email protected]>
To facilitate a transfer from
SOC_FOOtoARCH_FOO, over a release cycle, introduce some aliases so that drivers etc that use theSOC_FOOsymbols can be converted.
Quoted 5 lines
Quoted 1 line
Signed-off-by Conor Dooley <[email protected]>Thanks for your patch, which is now commit
fc43211939bb6874("RISC-V: kconfig.socs: convert usage ofSOC_CANAANtoARCH_CANAAN") in riscv/for-next
I see this and I immediately know it is going to be bad news!
Quoted 23 lines
arch/riscv/Kconfig.socs config SOC_CANAAN_K210_DTB_BUILTIN · line 95
Quoted 9 lines
arch/riscv/Kconfig.socs +2
-0@@ -73,6 +91,9 @@ config SOC_CANAAN_K210_DTB_BUILTIN This option should be selected if no bootloader is being used. If unsure, say Y. +config ARCH_CANAAN_K210_DTB_SOURCE + def_bool SOC_CANAAN_K210_DTB_SOURCEThis is not correct, as
SOC_CANAAN_K210_DTB_SOURCEbelow is not a bool, but a string.arch/riscv/Kconfig.socs config SOC_CANAAN_K210_DTB_BUILTIN · line 99
Quoted 4 lines
arch/riscv/Kconfig.socs +1
-0+ config SOC_CANAAN_K210_DTB_SOURCE string "Source file for the Canaan Kendryte K210 builtin DTB" depends on SOC_CANAANHence
obj-$(CONFIG_ARCH_CANAAN_K210_DTB_BUILTIN) += $(addsuffix .dtb.o,$(CONFIG_ARCH_CANAAN_K210_DTB_SOURCE))
Quoted 7 lines
will do the wrong thing later, and I get a non-bootable system (no output) on my MAiX-BiT.
Unfortunately there is no
def_string, so I don't think we can fix this in a backwards-compatible way, and have to replace allSOC_CANAAN_K210_DTB_SOURCEbyARCH_CANAAN_K210_DTB_SOURCE, and urging users to update their .config manually.
That sucks. I'm not sure how I missed this - I had tested originally on my k210, but evidently there was something wrong with how I had done it. I must have not tested a subsequent revision on the k210. Mea cupla.
As it wasn't my intention to inflict this change without time for the symbols to appear in configs, my immediate feeling is that this part of the change should be reverted.
Of course, we could so something like:
arch/riscv/Kconfig.socs +2 -1
@@ -79,7 +79,8 @@ config SOC_CANAAN_K210_DTB_BUILTIN If unsure, say Y. config ARCH_CANAAN_K210_DTB_SOURCE - def_bool SOC_CANAAN_K210_DTB_SOURCE + string + default SOC_CANAAN_K210_DTB_SOURCE config SOC_CANAAN_K210_DTB_SOURCE string "Source file for the Canaan Kendryte K210 builtin DTB"
But I am not sure of how that interacts with the various methods of updating ones config.
From, admittedly limited, testing it does get updated if SOC_CANAAN_K210_DTB_SOURCE is changed using menuconfig. Similarly, if one alters that symbol and does olddefconfig it also gets updated.
I'm sure I am overlooking something here though, what is it?
Thanks & apologies,
Conor.
_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv
Re: [PATCH v1 1/7] RISC-V: introduce ARCH_FOO kconfig aliases for SOC_FOO symbols
In reply to Geert UytterhoevenQuoted 38 lines · Geert Uytterhoeven
Hi Conor,
Quoted 7 lines · Conor Dooley
From: Conor Dooley <[email protected]>
To facilitate a transfer from
SOC_FOOtoARCH_FOO, over a release cycle, introduce some aliases so that drivers etc that use theSOC_FOOsymbols can be converted.Signed-off-by Conor Dooley <[email protected]>Thanks for your patch, which is now commit
fc43211939bb6874("RISC-V: kconfig.socs: convert usage ofSOC_CANAANtoARCH_CANAAN") in riscv/for-nextarch/riscv/Kconfig.socs config SOC_CANAAN_K210_DTB_BUILTIN · line 95
Quoted 9 lines
arch/riscv/Kconfig.socs +2
-0@@ -73,6 +91,9 @@ config SOC_CANAAN_K210_DTB_BUILTIN This option should be selected if no bootloader is being used. If unsure, say Y. +config ARCH_CANAAN_K210_DTB_SOURCE + def_bool SOC_CANAAN_K210_DTB_SOURCEThis is not correct, as
SOC_CANAAN_K210_DTB_SOURCEbelow is not a bool, but a string.arch/riscv/Kconfig.socs config SOC_CANAAN_K210_DTB_BUILTIN · line 99
Quoted 4 lines
arch/riscv/Kconfig.socs +1
-0+ config SOC_CANAAN_K210_DTB_SOURCE string "Source file for the Canaan Kendryte K210 builtin DTB" depends on SOC_CANAANHence
obj-$(CONFIG_ARCH_CANAAN_K210_DTB_BUILTIN) += $(addsuffix .dtb.o,$(CONFIG_ARCH_CANAAN_K210_DTB_SOURCE))
Quoted 7 lines
will do the wrong thing later, and I get a non-bootable system (no output) on my MAiX-BiT.
Unfortunately there is no
def_string, so I don't think we can fix this in a backwards-compatible way, and have to replace allSOC_CANAAN_K210_DTB_SOURCEbyARCH_CANAAN_K210_DTB_SOURCE, and urging users to update their .config manually.
Yes, the default built-in dtb is set to k210_generic.dts since we have a only generic nommu_k210[_sdcard] defconfig. We could add defconfigs per board type, but that is not the common approach.
By the way, I recall some pushback with the renaming of SOC_XXX to ARCH_XXX. I personally do not like it and Christoph suggested the revers approach of renaming arm ARCH_XXX to something else instead. I did not follow further on that discussion, but it looks like the riscv SOC_XXX renaming went through ?
Quoted 11 lines
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
Signature · 8 lines
Damien Le Moal Western Digital Research _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Conor Dooley
maintainerRe: [PATCH v1 1/7] RISC-V: introduce ARCH_FOO kconfig aliases for SOC_FOO symbols
Quoted 5 lines · Damien Le Moal
By the way, I recall some pushback with the renaming of
SOC_XXXtoARCH_XXX. I personally do not like it and Christoph suggested the revers approach of renaming armARCH_XXXto something else instead. I did not follow further on that discussion, but it looks like the riscvSOC_XXXrenaming went through ?
We asked the ARM guys about it, I think on IRC as they'd not replied to the thread, and cold water was poured on the idea.
It's Murphy's Law that the thing with the problem would be the k210 too isn't it.
I could've sworn I booted one of each vendor so kinda kicking myself right now.
_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv
Re: [PATCH v1 1/7] RISC-V: introduce ARCH_FOO kconfig aliases for SOC_FOO symbols
Quoted 9 lines · Conor Dooley
Quoted 5 lines · Damien Le Moal
By the way, I recall some pushback with the renaming of
SOC_XXXtoARCH_XXX. I personally do not like it and Christoph suggested the revers approach of renaming armARCH_XXXto something else instead. I did not follow further on that discussion, but it looks like the riscvSOC_XXXrenaming went through ?We asked the ARM guys about it, I think on IRC as they'd not replied to the thread, and cold water was poured on the idea.
That is really too bad. Naming something ARCH_XXX that is not a CPU architecture is simply very confusing I think.
Quoted 4 lines
It's Murphy's Law that the thing with the problem would be the k210 too isn't it.
I could've sworn I booted one of each vendor so kinda kicking myself right now.
Signature · 8 lines
Damien Le Moal Western Digital Research _______________________________________________ linux-riscv mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-riscv
Re: [PATCH v1 1/7] RISC-V: introduce ARCH_FOO kconfig aliases for SOC_FOO symbols
In reply to Conor DooleyHi Conor,
Quoted 61 lines · Conor Dooley
Quoted 7 lines · Geert Uytterhoeven
Quoted 6 lines · Conor Dooley
From: Conor Dooley <[email protected]>
To facilitate a transfer from
SOC_FOOtoARCH_FOO, over a release cycle, introduce some aliases so that drivers etc that use theSOC_FOOsymbols can be converted.Quoted 5 lines
Quoted 1 line
Signed-off-by Conor Dooley <[email protected]>Thanks for your patch, which is now commit
fc43211939bb6874("RISC-V: kconfig.socs: convert usage ofSOC_CANAANtoARCH_CANAAN") in riscv/for-nextI see this and I immediately know it is going to be bad news!
Quoted 23 lines
arch/riscv/Kconfig.socs config SOC_CANAAN_K210_DTB_BUILTIN · line 95
Quoted 9 lines
arch/riscv/Kconfig.socs +2
-0@@ -73,6 +91,9 @@ config SOC_CANAAN_K210_DTB_BUILTIN This option should be selected if no bootloader is being used. If unsure, say Y. +config ARCH_CANAAN_K210_DTB_SOURCE + def_bool SOC_CANAAN_K210_DTB_SOURCEThis is not correct, as
SOC_CANAAN_K210_DTB_SOURCEbelow is not a bool, but a string.arch/riscv/Kconfig.socs config SOC_CANAAN_K210_DTB_BUILTIN
Quoted 4 lines
arch/riscv/Kconfig.socs +1
-0+ config SOC_CANAAN_K210_DTB_SOURCE string "Source file for the Canaan Kendryte K210 builtin DTB" depends on SOC_CANAANHence
obj-$(CONFIG_ARCH_CANAAN_K210_DTB_BUILTIN) += $(addsuffix .dtb.o,$(CONFIG_ARCH_CANAAN_K210_DTB_SOURCE))
Quoted 7 lines
will do the wrong thing later, and I get a non-bootable system (no output) on my MAiX-BiT.
Unfortunately there is no
def_string, so I don't think we can fix this in a backwards-compatible way, and have to replace allSOC_CANAAN_K210_DTB_SOURCEbyARCH_CANAAN_K210_DTB_SOURCE, and urging users to update their .config manually.That sucks. I'm not sure how I missed this - I had tested originally on my k210, but evidently there was something wrong with how I had done it. I must have not tested a subsequent revision on the k210. Mea cupla.
As it wasn't my intention to inflict this change without time for the symbols to appear in configs, my immediate feeling is that this part of the change should be reverted.
Of course, we could so something like:
arch/riscv/Kconfig.socs +0
-0
arch/riscv/Kconfig.socs config SOC_CANAAN_K210_DTB_BUILTIN · line 83
Quoted 6 lines
arch/riscv/Kconfig.socs +2
-1@@ -79,7 +79,8 @@ config SOC_CANAAN_K210_DTB_BUILTIN If unsure, say Y. config ARCH_CANAAN_K210_DTB_SOURCE - def_bool SOC_CANAAN_K210_DTB_SOURCE + string + default SOC_CANAAN_K210_DTB_SOURCE
Thanks, that works!
I guess it was too late in the evening for me, to realize that the lack of "def_string" can be worked around using "string" and "default".
Quoted 8 lines
arch/riscv/Kconfig.socs +0
-0config SOC_CANAAN_K210_DTB_SOURCE string "Source file for the Canaan Kendryte K210 builtin DTB"But I am not sure of how that interacts with the various methods of updating ones config.
From, admittedly limited, testing it does get updated ifSOC_CANAAN_K210_DTB_SOURCEis changed using menuconfig. Similarly, if one alters that symbol and does olddefconfig it also gets updated.
With the above, the value of SOC_CANAAN_K210_DTB_SOURCE in existing config files will be copied to ARCH_CANAAN_K210_DTB_SOURCE when running "make oldconfig".
Hence in v6.4 you can drop the old SOC_CANAAN_K210_DTB_SOURCE. As that symbol is not present in defconfigs, there is nothing to update there,
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]
In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv
Re: [PATCH v1 1/7] RISC-V: introduce ARCH_FOO kconfig aliases for SOC_FOO symbols
In reply to Damien Le MoalHi Damien,
Quoted 8 lines · Damien Le Moal
Quoted 7 lines · Conor Dooley
Quoted 5 lines · Damien Le Moal
By the way, I recall some pushback with the renaming of
SOC_XXXtoARCH_XXX. I personally do not like it and Christoph suggested the revers approach of renaming armARCH_XXXto something else instead. I did not follow further on that discussion, but it looks like the riscvSOC_XXXrenaming went through ?
Quoted 5 lines
Quoted 2 lines
We asked the ARM guys about it, I think on IRC as they'd not replied to the thread, and cold water was poured on the idea.
That is really too bad. Naming something
ARCH_XXXthat is not a CPU architecture is simply very confusing I think.
I agree it is.
However, fixing that means adding backwards-compatiblity symbols, updating the few thousand existing users, and dropping the old symbols. And handling all the fall-out of people who didn't run "make oldconfig" in between, now wondering why their kernel no longer boots.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]
In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv