linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 1/2] x86: check CONFIG options instead of __arch_um__
@ 2020-04-21  7:39 Masahiro Yamada
  2020-04-21  7:39 ` [PATCH 2/2] raid6/algos.c: refactor raid6_algos array for x86 Masahiro Yamada
  0 siblings, 1 reply; 2+ messages in thread
From: Masahiro Yamada @ 2020-04-21  7:39 UTC (permalink / raw)
  To: x86, Ingo Molnar, Thomas Gleixner, Borislav Petkov, H . Peter Anvin
  Cc: Jason A . Donenfeld, linux-crypto, Masahiro Yamada, Anton Ivanov,
	Jeff Dike, Richard Weinberger, linux-kernel, linux-um

If the intention is to check i386/x86_64 excluding UML, we can use
CONFIG options instead.

There are only some users of __arch_um__. This commit replaces them,
then removes the __arch_um__ definition.

The original reason for checking __i386__ / __x86_64__ was perhaps
because lib/raid6/algos.c is built not only for the kernel but also
for the user-space test program.

However, lib/raid6/test/Makefile passes -DCONFIG_X86, -DCONFIG_X86_32,
and -DCONFIG_X86_64 for this case. So, I do not see a good reason to
not use CONFIG options here.

What is confusing is, CONFIG_X86_{32,64} is defined not only in
arch/x86/Kconfig, but also in arch/x86/um/Kconfig. To exlucde UML,
we need to check CONFIG_X86 too.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Acked-by: H. Peter Anvin (Intel) <hpa@zytor.com>
---

 arch/um/Makefile  | 2 +-
 kernel/signal.c   | 2 +-
 lib/raid6/algos.c | 6 ++++--
 lib/raid6/x86.h   | 2 +-
 4 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/arch/um/Makefile b/arch/um/Makefile
index d2daa206872d..064fbed7a4e9 100644
--- a/arch/um/Makefile
+++ b/arch/um/Makefile
@@ -62,7 +62,7 @@ KBUILD_CPPFLAGS += -I$(srctree)/$(HOST_DIR)/um
 #
 # These apply to USER_CFLAGS to.
 
-KBUILD_CFLAGS += $(CFLAGS) $(CFLAGS-y) -D__arch_um__ \
+KBUILD_CFLAGS += $(CFLAGS) $(CFLAGS-y) \
 	$(ARCH_INCLUDE) $(MODE_INCLUDE) -Dvmap=kernel_vmap	\
 	-Dlongjmp=kernel_longjmp -Dsetjmp=kernel_setjmp \
 	-Din6addr_loopback=kernel_in6addr_loopback \
diff --git a/kernel/signal.c b/kernel/signal.c
index 713104884414..1af3ad707b02 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -1246,7 +1246,7 @@ static void print_fatal_signal(int signr)
 	struct pt_regs *regs = signal_pt_regs();
 	pr_info("potentially unexpected fatal signal %d.\n", signr);
 
-#if defined(__i386__) && !defined(__arch_um__)
+#if defined(CONFIG_X86) && defined(CONFIG_X86_32)
 	pr_info("code at %08lx: ", regs->ip);
 	{
 		int i;
diff --git a/lib/raid6/algos.c b/lib/raid6/algos.c
index 6d5e5000fdd7..978fa19efbbf 100644
--- a/lib/raid6/algos.c
+++ b/lib/raid6/algos.c
@@ -29,7 +29,8 @@ struct raid6_calls raid6_call;
 EXPORT_SYMBOL_GPL(raid6_call);
 
 const struct raid6_calls * const raid6_algos[] = {
-#if defined(__i386__) && !defined(__arch_um__)
+#ifdef CONFIG_X86
+#if defined(CONFIG_X86_32)
 #ifdef CONFIG_AS_AVX512
 	&raid6_avx512x2,
 	&raid6_avx512x1,
@@ -43,7 +44,7 @@ const struct raid6_calls * const raid6_algos[] = {
 	&raid6_mmxx2,
 	&raid6_mmxx1,
 #endif
-#if defined(__x86_64__) && !defined(__arch_um__)
+#if defined(CONFIG_X86_64)
 #ifdef CONFIG_AS_AVX512
 	&raid6_avx512x4,
 	&raid6_avx512x2,
@@ -56,6 +57,7 @@ const struct raid6_calls * const raid6_algos[] = {
 	&raid6_sse2x2,
 	&raid6_sse2x1,
 #endif
+#endif /* CONFIG_X86 */
 #ifdef CONFIG_ALTIVEC
 	&raid6_vpermxor8,
 	&raid6_vpermxor4,
diff --git a/lib/raid6/x86.h b/lib/raid6/x86.h
index 9a6ff37115e7..0436b32f7709 100644
--- a/lib/raid6/x86.h
+++ b/lib/raid6/x86.h
@@ -14,7 +14,7 @@
 #ifndef LINUX_RAID_RAID6X86_H
 #define LINUX_RAID_RAID6X86_H
 
-#if (defined(__i386__) || defined(__x86_64__)) && !defined(__arch_um__)
+#ifdef CONFIG_X86
 
 #ifdef __KERNEL__ /* Real code */
 
-- 
2.25.1


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* [PATCH 2/2] raid6/algos.c: refactor raid6_algos array for x86
  2020-04-21  7:39 [PATCH 1/2] x86: check CONFIG options instead of __arch_um__ Masahiro Yamada
@ 2020-04-21  7:39 ` Masahiro Yamada
  0 siblings, 0 replies; 2+ messages in thread
From: Masahiro Yamada @ 2020-04-21  7:39 UTC (permalink / raw)
  To: x86, Ingo Molnar, Thomas Gleixner, Borislav Petkov, H . Peter Anvin
  Cc: Jason A . Donenfeld, linux-crypto, Masahiro Yamada, linux-kernel

Factor out the common pointer references.

Common for i386 and x86_64:
  raid6_avx512x2
  raid6_avx512x1
  raid6_avx2x2
  raid6_avx2x1
  raid6_sse2x2
  raid6_sse2x1
  raid6_sse1x2

Only for i386:
  raid6_sse1x1
  raid6_mmxx2
  raid6_mmxx1

Only for x86_64:
  raid6_avx512x4
  raid6_avx2x4
  raid6_sse2x4

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
---

 lib/raid6/algos.c | 24 ++++++++++--------------
 1 file changed, 10 insertions(+), 14 deletions(-)

diff --git a/lib/raid6/algos.c b/lib/raid6/algos.c
index 978fa19efbbf..c0f2fd6e0621 100644
--- a/lib/raid6/algos.c
+++ b/lib/raid6/algos.c
@@ -30,32 +30,28 @@ EXPORT_SYMBOL_GPL(raid6_call);
 
 const struct raid6_calls * const raid6_algos[] = {
 #ifdef CONFIG_X86
-#if defined(CONFIG_X86_32)
 #ifdef CONFIG_AS_AVX512
-	&raid6_avx512x2,
-	&raid6_avx512x1,
-#endif
-	&raid6_avx2x2,
-	&raid6_avx2x1,
-	&raid6_sse2x2,
-	&raid6_sse2x1,
-	&raid6_sse1x2,
-	&raid6_sse1x1,
-	&raid6_mmxx2,
-	&raid6_mmxx1,
-#endif
 #if defined(CONFIG_X86_64)
-#ifdef CONFIG_AS_AVX512
 	&raid6_avx512x4,
+#endif
 	&raid6_avx512x2,
 	&raid6_avx512x1,
 #endif
+#ifdef CONFIG_X86_64
 	&raid6_avx2x4,
+#endif
 	&raid6_avx2x2,
 	&raid6_avx2x1,
+#ifdef CONFIG_X86_64
 	&raid6_sse2x4,
+#endif
 	&raid6_sse2x2,
 	&raid6_sse2x1,
+	&raid6_sse1x2,
+#ifdef CONFIG_X86_32
+	&raid6_sse1x1,
+	&raid6_mmxx2,
+	&raid6_mmxx1,
 #endif
 #endif /* CONFIG_X86 */
 #ifdef CONFIG_ALTIVEC
-- 
2.25.1


^ permalink raw reply related	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2020-04-21  7:42 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-04-21  7:39 [PATCH 1/2] x86: check CONFIG options instead of __arch_um__ Masahiro Yamada
2020-04-21  7:39 ` [PATCH 2/2] raid6/algos.c: refactor raid6_algos array for x86 Masahiro Yamada

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).