[U-Boot] [PATCH v5 0/5] Add support for Thumb-1 builds

This series implements essential changes for thumb-1 support and activates thumb-1 build for openrd and tricorder as a proof of concept as well as a fix to bring the image sizes of openrd and tricorder targets back under an acceptable limit.
For other targets, some additional files might need to be forced to ARM building, notably architecture-specific instances of lowlevel_init.S or of cpu.c. This should be handled on a case-by-case basis.
IMPORTANT NOTE:
The whole set has been build-tested across all ARM targets and no build regression was detected.
Patch 1 has been confirmed to be binary-invariant.
Patch 2 has been build- and run-tested on ED Mini V2 (switched temporarily to Thumb build) and confirmed to be binary-invariant for non-thumb target Wireless Space.
Patches 3, 4 and 5 have been build-tested BUT HAVE NOT BEEN RUN-TESTED.
********************************************************* Open-RD and Tricorder owners, PLEASE run-test this series and report on the mailing list whether the target works. *********************************************************
Changes in v5: - switched tricorder[_flash] to Thumb-1 build
Changes in v4: - Made stm32f429-discovery explicitly state that it builds for Thumb
Changes in v3: - added arch/arm/lib/mem{set,cpy}.S to the list of modules which should always be built in ARM state. - Selected HAS_THUMB2 for CPU_V7M. - Fixed invalidate_l2_cache() when building for Thumb-1.
Changes in v2: - fixed a typo in the commit message - added file arch/arm/thumb1/include/asm/proc-armv/system.h, which overrides arch/arm/include/asm/proc-armv/system.h when building for Thumb-1 and provides non-functional but Thumb-compilable IRQ and FIQ related macros and functions.
Albert ARIBAUD (5): stm32f429-discovery: add CONFIG_SYS_THUMB_BUILD arm: support Thumb-1 with CONFIG_SYS_THUMB_BUILD kirkwood: support CONFIG_SYS_THUMB_BUILD openrd: switch to CONFIG_SYS_THUMB_BUILD tricorder: switch to CONFIG_SYS_THUMB_BUILD
Makefile | 2 + arch/arm/Kconfig | 5 ++ arch/arm/cpu/arm926ejs/Makefile | 11 ++++ arch/arm/cpu/arm926ejs/cache.c | 5 ++ arch/arm/include/asm/cache.h | 4 ++ arch/arm/lib/Makefile | 24 +++++++++ arch/arm/lib/cache.c | 11 ++++ arch/arm/lib/memcpy.S | 4 +- arch/arm/lib/memset.S | 2 +- arch/arm/mach-kirkwood/Makefile | 6 +++ arch/arm/mach-orion5x/Makefile | 10 ++++ arch/arm/thumb1/include/asm/proc-armv/system.h | 69 ++++++++++++++++++++++++++ examples/standalone/Makefile | 10 ++++ include/configs/openrd.h | 1 + include/configs/stm32f429-discovery.h | 1 + include/configs/tricorder.h | 1 + scripts/Makefile.lib | 3 +- 17 files changed, 165 insertions(+), 4 deletions(-) create mode 100644 arch/arm/thumb1/include/asm/proc-armv/system.h

This target is ARMv7-M therefore can only build for Thumb, but it did not #define CONFIG_SYS_THUMB_BUILD, so the U-Boot code did not know it had to build for Thumb(2), not ARM.
This patch is binary-invariant: builds of stm32f429-discovery with and without this patch were compared and found to differ only by their U-Boot version strings.
Signed-off-by: Albert ARIBAUD albert.u.boot@aribaud.net ---
Changes in v5: None Changes in v4: - Made stm32f429-discovery explicitly state that it builds for Thumb
Changes in v3: None Changes in v2: None
include/configs/stm32f429-discovery.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/include/configs/stm32f429-discovery.h b/include/configs/stm32f429-discovery.h index 1b4fd21..f481767 100644 --- a/include/configs/stm32f429-discovery.h +++ b/include/configs/stm32f429-discovery.h @@ -9,6 +9,7 @@ #define __CONFIG_H
#define CONFIG_STM32F4 +#define CONFIG_SYS_THUMB_BUILD #define CONFIG_STM32F4DISCOVERY #define CONFIG_SYS_GENERIC_BOARD

When building a Thumb-1-only target with CONFIG_SYS_THUMB_BUILD, some files fail to build, most of the time because they include mcr instructions, which only exist for Thumb-2.
This patch introduces a Kconfig option CONFIG_THUMB2 and uses it to select between Thumb-2 and ARM mode for the aforementioned files.
Signed-off-by: Albert ARIBAUD albert.u.boot@aribaud.net --- This patch has been build-tested and run-tested on ED Mini V2, above the "edmini: switch to SPL" patch, and found to reduce U-Boot size by 25% and SPL size by 14%... and to run fine. :)
This patch has also been tested against side effects on the non-Thumb wireless_space target. The binaries produced with and without this patch were found to differ only by their version string.
Changes in v5: None Changes in v4: None Changes in v3: - added arch/arm/lib/mem{set,cpy}.S to the list of modules which should always be built in ARM state. - Selected HAS_THUMB2 for CPU_V7M. - Fixed invalidate_l2_cache() when building for Thumb-1.
Changes in v2: - fixed a typo in the commit message - added file arch/arm/thumb1/include/asm/proc-armv/system.h, which overrides arch/arm/include/asm/proc-armv/system.h when building for Thumb-1 and provides non-functional but Thumb-compilable IRQ and FIQ related macros and functions.
Makefile | 2 + arch/arm/Kconfig | 5 ++ arch/arm/cpu/arm926ejs/Makefile | 11 ++++ arch/arm/cpu/arm926ejs/cache.c | 5 ++ arch/arm/include/asm/cache.h | 4 ++ arch/arm/lib/Makefile | 24 +++++++++ arch/arm/lib/cache.c | 11 ++++ arch/arm/lib/memcpy.S | 4 +- arch/arm/lib/memset.S | 2 +- arch/arm/mach-orion5x/Makefile | 10 ++++ arch/arm/thumb1/include/asm/proc-armv/system.h | 69 ++++++++++++++++++++++++++ examples/standalone/Makefile | 10 ++++ scripts/Makefile.lib | 3 +- 13 files changed, 156 insertions(+), 4 deletions(-) create mode 100644 arch/arm/thumb1/include/asm/proc-armv/system.h
diff --git a/Makefile b/Makefile index b6f83a5..739940c 100644 --- a/Makefile +++ b/Makefile @@ -607,6 +607,8 @@ KBUILD_CFLAGS += $(KCFLAGS) UBOOTINCLUDE := \ -Iinclude \ $(if $(KBUILD_SRC), -I$(srctree)/include) \ + $(if $(CONFIG_SYS_THUMB_BUILD), $(if $(CONFIG_HAS_THUMB2),, \ + -I$(srctree)/arch/$(ARCH)/thumb1/include),) \ -I$(srctree)/arch/$(ARCH)/include \ -include $(srctree)/include/linux/kconfig.h
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 506463c..b29923c 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -10,6 +10,9 @@ config ARM64 config HAS_VBAR bool
+config HAS_THUMB2 + bool + config CPU_ARM720T bool
@@ -32,9 +35,11 @@ config CPU_ARM1176 config CPU_V7 bool select HAS_VBAR + select HAS_THUMB2
config CPU_V7M bool + select HAS_THUMB2
config CPU_PXA bool diff --git a/arch/arm/cpu/arm926ejs/Makefile b/arch/arm/cpu/arm926ejs/Makefile index 63fa159..fe78922 100644 --- a/arch/arm/cpu/arm926ejs/Makefile +++ b/arch/arm/cpu/arm926ejs/Makefile @@ -20,3 +20,14 @@ obj-$(CONFIG_MX25) += mx25/ obj-$(CONFIG_MX27) += mx27/ obj-$(if $(filter mxs,$(SOC)),y) += mxs/ obj-$(if $(filter spear,$(SOC)),y) += spear/ + +# some files can only build in ARM or THUMB2, not THUMB1 + +ifdef CONFIG_SYS_THUMB_BUILD +ifndef CONFIG_HAS_THUMB2 + +CFLAGS_cpu.o := -marm +CFLAGS_cache.o := -marm + +endif +endif diff --git a/arch/arm/cpu/arm926ejs/cache.c b/arch/arm/cpu/arm926ejs/cache.c index 8d7873c..f12e365 100644 --- a/arch/arm/cpu/arm926ejs/cache.c +++ b/arch/arm/cpu/arm926ejs/cache.c @@ -99,4 +99,9 @@ void flush_cache(unsigned long start, unsigned long size) /* * Stub implementations for l2 cache operations */ + __weak void l2_cache_disable(void) {} + +#if defined CONFIG_SYS_THUMB_BUILD +__weak void invalidate_l2_cache(void) {} +#endif diff --git a/arch/arm/include/asm/cache.h b/arch/arm/include/asm/cache.h index a836e9f..1f63127 100644 --- a/arch/arm/include/asm/cache.h +++ b/arch/arm/include/asm/cache.h @@ -16,6 +16,9 @@ /* * Invalidate L2 Cache using co-proc instruction */ +#ifdef CONFIG_SYS_THUMB_BUILD +void invalidate_l2_cache(void); +#else static inline void invalidate_l2_cache(void) { unsigned int val=0; @@ -24,6 +27,7 @@ static inline void invalidate_l2_cache(void) : : "r" (val) : "cc"); isb(); } +#endif
void l2_cache_enable(void); void l2_cache_disable(void); diff --git a/arch/arm/lib/Makefile b/arch/arm/lib/Makefile index 31a5c8d..4c7e3f9 100644 --- a/arch/arm/lib/Makefile +++ b/arch/arm/lib/Makefile @@ -60,3 +60,27 @@ obj-$(CONFIG_DEBUG_LL) += debug.o ifneq (,$(findstring -mabi=aapcs-linux,$(PLATFORM_CPPFLAGS))) extra-y += eabi_compat.o endif + +# some files can only build in ARM or THUMB2, not THUMB1 + +ifdef CONFIG_SYS_THUMB_BUILD +ifndef CONFIG_HAS_THUMB2 + +# for C files, just apend -marm, which will override previous -mthumb* + +CFLAGS_cache.o := -marm +CFLAGS_cache-cp15.o := -marm + +# For .S, drop -mthumb* and other thumb-related options. +# CFLAGS_REMOVE_* would not have an effet, so AFLAGS_REMOVE_* +# was implemented and is used here. +# Also, define ${target}_NO_THUMB_BUILD for these two targets +# so that the code knows it should not use Thumb. + +AFLAGS_REMOVE_memset.o := -mthumb -mthumb-interwork +AFLAGS_REMOVE_memcpy.o := -mthumb -mthumb-interwork +AFLAGS_memset.o := -DMEMSET_NO_THUMB_BUILD +AFLAGS_memcpy.o := -DMEMCPY_NO_THUMB_BUILD + +endif +endif diff --git a/arch/arm/lib/cache.c b/arch/arm/lib/cache.c index 74cfde6..7c8878d 100644 --- a/arch/arm/lib/cache.c +++ b/arch/arm/lib/cache.c @@ -95,3 +95,14 @@ phys_addr_t noncached_alloc(size_t size, size_t align) return next; } #endif /* CONFIG_SYS_NONCACHED_MEMORY */ + +#if defined(CONFIG_SYS_THUMB_BUILD) +void invalidate_l2_cache(void) +{ + unsigned int val = 0; + + asm volatile("mcr p15, 1, %0, c15, c11, 0 @ invl l2 cache" + : : "r" (val) : "cc"); + isb(); +} +#endif diff --git a/arch/arm/lib/memcpy.S b/arch/arm/lib/memcpy.S index eeaf003..7d9fc0f 100644 --- a/arch/arm/lib/memcpy.S +++ b/arch/arm/lib/memcpy.S @@ -13,7 +13,7 @@ #include <linux/linkage.h> #include <asm/assembler.h>
-#ifdef CONFIG_SYS_THUMB_BUILD +#if defined(CONFIG_SYS_THUMB_BUILD) && !defined(MEMCPY_NO_THUMB_BUILD) #define W(instr) instr.w #else #define W(instr) instr @@ -62,7 +62,7 @@
/* Prototype: void *memcpy(void *dest, const void *src, size_t n); */ .syntax unified -#ifdef CONFIG_SYS_THUMB_BUILD +#if defined(CONFIG_SYS_THUMB_BUILD) && !defined(MEMCPY_NO_THUMB_BUILD) .thumb .thumb_func #endif diff --git a/arch/arm/lib/memset.S b/arch/arm/lib/memset.S index 7208f20..df053a3 100644 --- a/arch/arm/lib/memset.S +++ b/arch/arm/lib/memset.S @@ -16,7 +16,7 @@ .align 5
.syntax unified -#ifdef CONFIG_SYS_THUMB_BUILD +#if defined(CONFIG_SYS_THUMB_BUILD) && !defined(MEMSET_NO_THUMB_BUILD) .thumb .thumb_func #endif diff --git a/arch/arm/mach-orion5x/Makefile b/arch/arm/mach-orion5x/Makefile index 546ebcb..33dcad4 100644 --- a/arch/arm/mach-orion5x/Makefile +++ b/arch/arm/mach-orion5x/Makefile @@ -16,3 +16,13 @@ obj-y += timer.o ifndef CONFIG_SKIP_LOWLEVEL_INIT obj-y += lowlevel_init.o endif + +# some files can only build in ARM or THUMB2, not THUMB1 + +ifdef CONFIG_SYS_THUMB_BUILD +ifndef CONFIG_HAS_THUMB2 + +CFLAGS_cpu.o := -marm + +endif +endif diff --git a/arch/arm/thumb1/include/asm/proc-armv/system.h b/arch/arm/thumb1/include/asm/proc-armv/system.h new file mode 100644 index 0000000..7dfbf3d --- /dev/null +++ b/arch/arm/thumb1/include/asm/proc-armv/system.h @@ -0,0 +1,69 @@ +/* + * Thumb-1 drop-in for the linux/include/asm-arm/proc-armv/system.h + * + * (C) Copyright 2015 + * Albert ARIBAUD albert.u.boot@aribaud.net + * + * The original file does not build in Thumb mode. However, in U-Boot + * we don't use interrupt context, so we can redefine these as empty + * memory barriers, which makes Thumb-1 compiler happy. + * + * SPDX-License-Identifier: GPL-2.0+ + */ + +/* + * Use the same macro name as linux/include/asm-arm/proc-armv/system.h + * here, so that if the original ever gets included after us, it won't + * try to re-redefine anything. + */ + +#ifndef __ASM_PROC_SYSTEM_H +#define __ASM_PROC_SYSTEM_H + +/* + * Redefine all original macros with static inline functions containing + * a simple memory barrier, so that they produce the same instruction + * ordering constraints as their original counterparts. + * We use static inline functions rather than macros so that we can tell + * the compiler to not complain about unused arguments. + */ + +static inline void local_irq_save( + unsigned long flags __attribute__((unused))) +{ + __asm__ __volatile__ ("" : : : "memory"); +} + +static inline void local_irq_enable(void) +{ + __asm__ __volatile__ ("" : : : "memory"); +} + +static inline void local_irq_disable(void) +{ + __asm__ __volatile__ ("" : : : "memory"); +} + +static inline void __stf(void) +{ + __asm__ __volatile__ ("" : : : "memory"); +} + +static inline void __clf(void) +{ + __asm__ __volatile__ ("" : : : "memory"); +} + +static inline void local_save_flags( + unsigned long flags __attribute__((unused))) +{ + __asm__ __volatile__ ("" : : : "memory"); +} + +static inline void local_irq_restore( + unsigned long flags __attribute__((unused))) +{ + __asm__ __volatile__ ("" : : : "memory"); +} + +#endif /* __ASM_PROC_SYSTEM_H */ diff --git a/examples/standalone/Makefile b/examples/standalone/Makefile index 0863a8c..5a6ae00 100644 --- a/examples/standalone/Makefile +++ b/examples/standalone/Makefile @@ -73,3 +73,13 @@ $(obj)/%.srec: $(obj)/% FORCE $(obj)/%.bin: OBJCOPYFLAGS := -O binary $(obj)/%.bin: $(obj)/% FORCE $(call if_changed,objcopy) + +# some files can only build in ARM or THUMB2, not THUMB1 + +ifdef CONFIG_SYS_THUMB_BUILD +ifndef CONFIG_HAS_THUMB2 + +CFLAGS_stubs.o := -marm + +endif +endif diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib index 8731fc6..3d5c872 100644 --- a/scripts/Makefile.lib +++ b/scripts/Makefile.lib @@ -104,8 +104,9 @@ modname_flags = $(if $(filter 1,$(words $(modname))),\ orig_c_flags = $(KBUILD_CPPFLAGS) $(KBUILD_CFLAGS) $(KBUILD_SUBDIR_CCFLAGS) \ $(ccflags-y) $(CFLAGS_$(basetarget).o) _c_flags = $(filter-out $(CFLAGS_REMOVE_$(basetarget).o), $(orig_c_flags)) -_a_flags = $(KBUILD_CPPFLAGS) $(KBUILD_AFLAGS) $(KBUILD_SUBDIR_ASFLAGS) \ +orig_a_flags = $(KBUILD_CPPFLAGS) $(KBUILD_AFLAGS) $(KBUILD_SUBDIR_ASFLAGS) \ $(asflags-y) $(AFLAGS_$(basetarget).o) +_a_flags = $(filter-out $(AFLAGS_REMOVE_$(basetarget).o), $(orig_a_flags)) _cpp_flags = $(KBUILD_CPPFLAGS) $(cppflags-y) $(CPPFLAGS_$(@F))
#

Hi Albert,
On 17.07.2015 22:50, Albert ARIBAUD wrote:
When building a Thumb-1-only target with CONFIG_SYS_THUMB_BUILD, some files fail to build, most of the time because they include mcr instructions, which only exist for Thumb-2.
This patch introduces a Kconfig option CONFIG_THUMB2 and uses it to select between Thumb-2 and ARM mode for the aforementioned files.
Signed-off-by: Albert ARIBAUD albert.u.boot@aribaud.net
This patch has been build-tested and run-tested on ED Mini V2, above the "edmini: switch to SPL" patch, and found to reduce U-Boot size by 25% and SPL size by 14%... and to run fine. :)
This patch has also been tested against side effects on the non-Thumb wireless_space target. The binaries produced with and without this patch were found to differ only by their version string.
Changes in v5: None Changes in v4: None Changes in v3:
- added arch/arm/lib/mem{set,cpy}.S to the list of modules which should always be built in ARM state.
- Selected HAS_THUMB2 for CPU_V7M.
- Fixed invalidate_l2_cache() when building for Thumb-1.
Changes in v2:
- fixed a typo in the commit message
- added file arch/arm/thumb1/include/asm/proc-armv/system.h, which overrides arch/arm/include/asm/proc-armv/system.h when building for Thumb-1 and provides non-functional but Thumb-compilable IRQ and FIQ related macros and functions.
Whats the current status of this patch(set)? Do you plan to pull it in soon (when)? Or is something missing?
I'm asking since I have a ST SPEAr600 related patchset here which requires this THUMB1 support. And would like to commit this series soon as well.
Thanks, Stefan

Hi Albert,
On 14.08.2015 09:23, Stefan Roese wrote:
On 17.07.2015 22:50, Albert ARIBAUD wrote:
When building a Thumb-1-only target with CONFIG_SYS_THUMB_BUILD, some files fail to build, most of the time because they include mcr instructions, which only exist for Thumb-2.
This patch introduces a Kconfig option CONFIG_THUMB2 and uses it to select between Thumb-2 and ARM mode for the aforementioned files.
Signed-off-by: Albert ARIBAUD albert.u.boot@aribaud.net
This patch has been build-tested and run-tested on ED Mini V2, above the "edmini: switch to SPL" patch, and found to reduce U-Boot size by 25% and SPL size by 14%... and to run fine. :)
This patch has also been tested against side effects on the non-Thumb wireless_space target. The binaries produced with and without this patch were found to differ only by their version string.
Changes in v5: None Changes in v4: None Changes in v3:
- added arch/arm/lib/mem{set,cpy}.S to the list of modules which should always be built in ARM state.
- Selected HAS_THUMB2 for CPU_V7M.
- Fixed invalidate_l2_cache() when building for Thumb-1.
Changes in v2:
- fixed a typo in the commit message
- added file arch/arm/thumb1/include/asm/proc-armv/system.h, which overrides arch/arm/include/asm/proc-armv/system.h when building for Thumb-1 and provides non-functional but Thumb-compilable IRQ and FIQ related macros and functions.
Whats the current status of this patch(set)? Do you plan to pull it in soon (when)? Or is something missing?
I'm asking since I have a ST SPEAr600 related patchset here which requires this THUMB1 support. And would like to commit this series soon as well.
Sorry for bugging you again on this. Any updates? What are the plans?
Thanks, Stefan

Kirkwood files cpu.c and cache.c cannot build in Thumb state; force them in ARM state even under CONFIG_SYS_THUMB_BUILD.
Signed-off-by: Albert ARIBAUD albert.u.boot@aribaud.net ---
Changes in v5: None Changes in v4: None Changes in v3: None Changes in v2: None
arch/arm/mach-kirkwood/Makefile | 6 ++++++ 1 file changed, 6 insertions(+)
diff --git a/arch/arm/mach-kirkwood/Makefile b/arch/arm/mach-kirkwood/Makefile index df4756e..5abcf70 100644 --- a/arch/arm/mach-kirkwood/Makefile +++ b/arch/arm/mach-kirkwood/Makefile @@ -9,3 +9,9 @@ obj-y = cpu.o obj-y += cache.o obj-y += mpp.o + +# cpu.o and cache.o contain CP15 instructions which cannot be run in +# Thumb state, so build them for ARM state even with CONFIG_SYS_THUMB_BUILD + +CFLAGS_cpu.o := -marm +CFLAGS_cache.o := -marm

Now that kirkwood supports CONFIG_SYS_THUMB_BUILD, switch to it for openrd. This brings U-Boot back to an acceptable size.
Signed-off-by: Albert ARIBAUD albert.u.boot@aribaud.net ---
Changes in v5: None Changes in v4: None Changes in v3: None Changes in v2: None
include/configs/openrd.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/include/configs/openrd.h b/include/configs/openrd.h index 7211314..62f15b7 100644 --- a/include/configs/openrd.h +++ b/include/configs/openrd.h @@ -38,6 +38,7 @@ #define CONFIG_KW88F6281 1 /* SOC Name */ #define CONFIG_MACH_OPENRD_BASE /* Machine type */ #define CONFIG_SKIP_LOWLEVEL_INIT /* disable board lowlevel_init */ +#define CONFIG_SYS_THUMB_BUILD
/* * Commands configuration

The tricorder and tricorder_flash boards have grown too big. Reduce their size by building them with CONFIG_SYS_THUMB_BUILD.
Signed-off-by: Albert ARIBAUD albert.u.boot@aribaud.net ---
Changes in v5: - switched tricorder[_flash] to Thumb-1 build
Changes in v4: None Changes in v3: None Changes in v2: None
include/configs/tricorder.h | 1 + 1 file changed, 1 insertion(+)
diff --git a/include/configs/tricorder.h b/include/configs/tricorder.h index 23bf599..7a61fb6 100644 --- a/include/configs/tricorder.h +++ b/include/configs/tricorder.h @@ -17,6 +17,7 @@ #define __CONFIG_H
/* High Level Configuration Options */ +#define CONFIG_SYS_THUMB_BUILD #define CONFIG_OMAP /* in a TI OMAP core */ #define CONFIG_OMAP_COMMON /* Common ARM Erratas */
participants (2)
-
Albert ARIBAUD
-
Stefan Roese