arch: Kconfig: Let all little endian architectures define CPU_LITTLE_ENDIAN explicitly

PATCH Superseded Chen Gang 5 participants 9 messages Aug 12, 2014 – Aug 23, 2014

Subsystems (5)

Arm (32-bit) x86 Kbuild arm64 +1 more subsystems

lore

Share & tools

lore

Chen Gang

author

2014-08-12 22:48 UTC #0 lore raw

PATCH v2 1/1 arch: Kconfig: Let all little endian architectures define CPU_LITTLE_ENDIAN explicitly

x86 and ia64 are always little endian. And another architectures may be little endian: mips, sh, powerpc, and m32r (may mark CPU_LITTLE_ENDIAN explicitly); also arm(64) and c6x (which may be !CPU_BIG_ENDIAN).

Some drivers (e.g. some of "drivers/isdn/hisax") may only support little endian (CPU_LITTLE_ENDIAN), and some drivers may only support big endian (!CPU_LITTLE_ENDIAN).

So export all little endian architectures within kernel wide, so can let Kconfig easier for the modules which only support little endian or only for big endian (assume !CPU_LITTLE_ENDIAN is same as CPU_BIG_ENDIAN).

Signed-off-by Chen Gang <[email protected]>

---
 arch/arm/Kconfig   | 5 +++++
 arch/arm64/Kconfig | 4 ++++
 arch/c6x/Kconfig   | 4 ++++
 arch/ia64/Kconfig  | 3 +++
 arch/x86/Kconfig   | 3 +++
 5 files changed, 19 insertions(+)

arch/arm/Kconfig +5 -0

index c49a775..0510a5d 100644
@@ -199,6 +199,11 @@ config NEED_DMA_MAP_STATE
 config ARCH_SUPPORTS_UPROBES
 	def_bool y
 
+config CPU_LITTLE_ENDIAN
+	depends on !CPU_BIG_ENDIAN
+	def_bool y
+
+
 config ARCH_HAS_DMA_SET_COHERENT_MASK
 	bool

arch/arm64/Kconfig +4 -0

index 62b4ae1..290a253 100644
@@ -129,6 +129,10 @@ config KERNEL_MODE_NEON
 config FIX_EARLYCON_MEM
 	def_bool y
 
+config CPU_LITTLE_ENDIAN
+	depends on !CPU_BIG_ENDIAN
+	def_bool y
+
 source "init/Kconfig"
 
 source "kernel/Kconfig.freezer"

arch/c6x/Kconfig +4 -0

index 77ea09b..56a8f53 100644
@@ -48,6 +48,10 @@ config C6X_BIG_KERNEL
 
 	  If unsure, say N.
 
+config CPU_LITTLE_ENDIAN
+	depends on !CPU_BIG_ENDIAN
+	def_bool y
+
 source "init/Kconfig"
 
 # Use the generic interrupt handling code in kernel/irq/

arch/ia64/Kconfig +3 -0

index 64aefb7..fa12d34 100644
@@ -133,6 +133,9 @@ config AUDIT_ARCH
 	bool
 	default y
 
+config CPU_LITTLE_ENDIAN
+	def_bool y
+
 menuconfig PARAVIRT_GUEST
 	bool "Paravirtualized guest support"
 	depends on BROKEN

arch/x86/Kconfig +3 -0

index 541e7cc..a1b3568 100644
@@ -270,6 +270,9 @@ config ARCH_SUPPORTS_UPROBES
 config FIX_EARLYCON_MEM
 	def_bool y
 
+config CPU_LITTLE_ENDIAN
+	def_bool y
+
 source "init/Kconfig"
 source "kernel/Kconfig.freezer"

Signature · 1 line

1.7.11.7

Sam Ravnborg

2014-08-13 05:08 UTC #1 lore raw

Quoted 11 lines · Chen Gang

x86 and ia64 are always little endian. And another architectures may be little endian: mips, sh, powerpc, and m32r (may mark CPU_LITTLE_ENDIAN explicitly); also arm(64) and c6x (which may be !CPU_BIG_ENDIAN).
Some drivers (e.g. some of "drivers/isdn/hisax") may only support little endian (CPU_LITTLE_ENDIAN), and some drivers may only support big endian (!CPU_LITTLE_ENDIAN).
So export all little endian architectures within kernel wide, so can let Kconfig easier for the modules which only support little endian or only for big endian (assume !CPU_LITTLE_ENDIAN is same as CPU_BIG_ENDIAN).

We need to cover three cases here:
- An arch supports only little endian
- An arch supports only big endian
- An arch may be both little and big endian

The solution you suggest assumes that an arch is either little or big endian. But we cannot ignore the hybriads that can do both.

Sam

Chen Gang

author

2014-08-13 05:27 UTC #2 lore raw

Quoted 9 lines · Sam Ravnborg

Quoted 8 lines · Chen Gang
x86 and ia64 are always little endian. And another architectures may be little endian: mips, sh, powerpc, and m32r (may mark CPU_LITTLE_ENDIAN explicitly); also arm(64) and c6x (which may be !CPU_BIG_ENDIAN).
Some drivers (e.g. some of "drivers/isdn/hisax") may only support little endian (CPU_LITTLE_ENDIAN), and some drivers may only support big endian (!CPU_LITTLE_ENDIAN).

Quoted 8 lines

Quoted 3 lines
So export all little endian architectures within kernel wide, so can let Kconfig easier for the modules which only support little endian or only for big endian (assume !CPU_LITTLE_ENDIAN is same as CPU_BIG_ENDIAN).
We need to cover three cases here:
- An arch supports only little endian
- An arch supports only big endian
- An arch may be both little and big endian

Yeah, we need.

Quoted 3 lines

The solution you suggest assumes that an arch is either little or big endian. But we cannot ignore the hybriads that can do both.

For the architectures which can do both, for me, they are: sh, powerpc, m32r, and mips (may mark CPU_LITTLE_ENDIAN), also are: arm/arm64, and c6x (may mark !CPU_BIG_ENDIAN).

For the architectures which only support little endian: x86, and ia64.

The left, I assumes they are big endian (no any 'ENDIAN' can be found in their Kconfig files). In my memory, except related with Intel, all are (or support) big endian.

Thanks.

Signature · 3 lines

Chen Gang

Open, share, and attitude like air, water, and life which God blessed

H. Peter Anvin

reviewer

2014-08-13 06:51 UTC #3 lore raw

Quoted 11 lines · Chen Gang

Yeah, we need.
Quoted 3 lines
The solution you suggest assumes that an arch is either little or big endian. But we cannot ignore the hybriads that can do both.
For the architectures which can do both, for me, they are: sh, powerpc, m32r, and mips (may mark CPU_LITTLE_ENDIAN), also are: arm/arm64, and c6x (may mark !CPU_BIG_ENDIAN).

Quoted 6 lines

For the architectures which only support little endian: x86, and ia64.
The left, I assumes they are big endian (no any 'ENDIAN' can be found in their Kconfig files). In my memory, except related with Intel, all are (or support) big endian.

I'm quite certain that is wrong. From memory, for example, I believe CRIS is littleendian, and sure enough:

: tazenda 103 ; less arch/cris/include/uapi/asm/byteorder.h #ifndef _CRIS_BYTEORDER_H
#define _CRIS_BYTEORDER_H

#include <linux/byteorder/little_endian.h>

#endif

As far as I know, endianism is always a compile-time option on Linux, so we should be able to relatively easily define the architecture so that we either have hardwired littleendian, bigendian, or prompt.

-hpa

Chen Gang

author

2014-08-13 07:39 UTC #4 lore raw

Quoted 22 lines · H. Peter Anvin

Quoted 11 lines · Chen Gang
Yeah, we need.
Quoted 3 lines
The solution you suggest assumes that an arch is either little or big endian. But we cannot ignore the hybriads that can do both.
For the architectures which can do both, for me, they are: sh, powerpc, m32r, and mips (may mark CPU_LITTLE_ENDIAN), also are: arm/arm64, and c6x (may mark !CPU_BIG_ENDIAN).
Quoted 6 lines
For the architectures which only support little endian: x86, and ia64.
The left, I assumes they are big endian (no any 'ENDIAN' can be found in their Kconfig files). In my memory, except related with Intel, all are (or support) big endian.
I'm quite certain that is wrong. From memory, for example, I believe CRIS is littleendian, and sure enough:

Quoted 8 lines

: tazenda 103 ; less arch/cris/include/uapi/asm/byteorder.h #ifndef _CRIS_BYTEORDER_H
#define _CRIS_BYTEORDER_H
#include <linux/byteorder/little_endian.h>
#endif

OK, what you said sounds reasonable to me. After search "endian" in "./arch", xtensa also can be support both big endian or little endian.

Quoted 4 lines

As far as I know, endianism is always a compile-time option on Linux, so we should be able to relatively easily define the architecture so that we either have hardwired littleendian, bigendian, or prompt.

Yeah, but that may break building and re-config manually. So it is still necessary to check little endian or big endian during configuration time ("make *config").

Thanks.

Signature · 3 lines

Chen Gang

Open, share, and attitude like air, water, and life which God blessed

Geert Uytterhoeven

2014-08-13 08:03 UTC #5 lore raw

In reply to Chen Gang

CC kbuild

Quoted 5 lines · Chen Gang

arch/arm/Kconfig +0 -0
index c49a775..0510a5d 100644

arch/arm/Kconfig config NEED_DMA_MAP_STATE · line 208

Quoted 10 lines

arch/arm/Kconfig +5 -0

@@ -199,6 +199,11 @@ config NEED_DMA_MAP_STATE
 config ARCH_SUPPORTS_UPROBES
        def_bool y
 
+config CPU_LITTLE_ENDIAN
+       depends on !CPU_BIG_ENDIAN
+       def_bool y
+
+
 config ARCH_HAS_DMA_SET_COHERENT_MASK
        bool

As this is a common symbol, and replicated for all affected architectures, I'm wondering if we should have the "config CPU_LITTLE_ENDIAN" in common Kconfig code instead, and make the individual architectures do a "select CPU_LITTLE_ENDIAN"?

Also we could have "config CPU_BIG_ENDIAN", too, and error out if none or both are selected (can Kconfig error out?).

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

Chen Gang

author

2014-08-13 08:11 UTC #6 lore raw

Quoted 11 lines · Geert Uytterhoeven

CC kbuild

Quoted 8 lines · Chen Gang

arch/arm/Kconfig +0 -0

index c49a775..0510a5d 100644
@@ -199,6 +199,11 @@ config NEED_DMA_MAP_STATE
 config ARCH_SUPPORTS_UPROBES
        def_bool y

Quoted 13 lines

arch/arm/Kconfig config NEED_DMA_MAP_STATE · line 208
Quoted 7 lines
arch/arm/Kconfig +5 -0
+config CPU_LITTLE_ENDIAN
+       depends on !CPU_BIG_ENDIAN
+       def_bool y
+
+
 config ARCH_HAS_DMA_SET_COHERENT_MASK
        bool
As this is a common symbol, and replicated for all affected architectures, I'm wondering if we should have the "config CPU_LITTLE_ENDIAN" in common Kconfig code instead, and make the individual architectures do a "select CPU_LITTLE_ENDIAN"?

That sounds a good idea to me.

Quoted 3 lines

Also we could have "config CPU_BIG_ENDIAN", too, and error out if none or both are selected (can Kconfig error out?).

It is a good idea to me, welcome any members ideas, especially Kbuild members' ideas, suggestions or completions.

Thanks.

Signature · 3 lines

Chen Gang

Open, share, and attitude like air, water, and life which God blessed

Michal Marek

2014-08-22 11:09 UTC #7 lore raw

In reply to Geert Uytterhoeven

Dne 13.8.2014 10:03, Geert Uytterhoeven napsal(a):

Quoted 11 lines

CC kbuild

Quoted 8 lines · Chen Gang

arch/arm/Kconfig +0 -0

index c49a775..0510a5d 100644
@@ -199,6 +199,11 @@ config NEED_DMA_MAP_STATE
 config ARCH_SUPPORTS_UPROBES
        def_bool y

Quoted 12 lines

arch/arm/Kconfig config NEED_DMA_MAP_STATE · line 208
Quoted 7 lines
arch/arm/Kconfig +5 -0
+config CPU_LITTLE_ENDIAN
+       depends on !CPU_BIG_ENDIAN
+       def_bool y
+
+
 config ARCH_HAS_DMA_SET_COHERENT_MASK
        bool
As this is a common symbol, and replicated for all affected architectures, I'm wondering if we should have the "config CPU_LITTLE_ENDIAN" in common Kconfig code instead, and make the individual architectures do a "select CPU_LITTLE_ENDIAN"?

Yes!

Quoted 2 lines

Also we could have "config CPU_BIG_ENDIAN", too, and error out if none or both are selected (can Kconfig error out?).

We can error out in the Makefile, if there is consensus that we should be doing so.

Michal

Chen Gang

author

2014-08-23 02:26 UTC #8 lore raw

Quoted 25 lines · Michal Marek

Dne 13.8.2014 10:03, Geert Uytterhoeven napsal(a):
Quoted 11 lines
CC kbuild
Quoted 8 lines · Chen Gang
arch/arm/Kconfig +0 -0
index c49a775..0510a5d 100644
@@ -199,6 +199,11 @@ config NEED_DMA_MAP_STATE
 config ARCH_SUPPORTS_UPROBES
        def_bool y
 
Quoted 12 lines
arch/arm/Kconfig config NEED_DMA_MAP_STATE · line 208
Quoted 7 lines
arch/arm/Kconfig +5 -0
+config CPU_LITTLE_ENDIAN
+       depends on !CPU_BIG_ENDIAN
+       def_bool y
+
+
 config ARCH_HAS_DMA_SET_COHERENT_MASK
        bool
As this is a common symbol, and replicated for all affected architectures, I'm wondering if we should have the "config CPU_LITTLE_ENDIAN" in common Kconfig code instead, and make the individual architectures do a "select CPU_LITTLE_ENDIAN"?

Quoted 9 lines

Yes!
Quoted 2 lines
Also we could have "config CPU_BIG_ENDIAN", too, and error out if none or both are selected (can Kconfig error out?).
We can error out in the Makefile, if there is consensus that we should be doing so.

OK, thanks. I have sent patch v3 for it, and did not check related error in Kconfig files.

Please check the related patches (excuse me, I send the patch v3 only according to "scripts/maintainers.pl", maybe not cc to you explicitly).

Thanks.

Signature · 3 lines

Chen Gang

Open, share, and attitude like air, water, and life which God blessed