Skip to contentSkip to conversation

lib/raid6: use vdupq_n_u8 to avoid endianness warnings

PATCH Applied [email protected] 4 participants 7 messages Feb 26, 2019 – Feb 28, 2019

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 lib/raid6: use vdupq_n_u8 to avoid endianness warnings

Clang warns: vector initializers are not compatible with NEON intrinsics in big endian mode [-Wnonportable-vector-initialization]

While this is usually the case, it's not an issue for this case since we're initializing the uint8x16_t (16x uint8_t's) with the same value.

Instead, use vdupq_n_u8 which both compilers lower into a single movi
instruction: https://godbolt.org/z/vBrgzt

This avoids the static storage for a constant value.

---
 lib/raid6/neon.uc            | 5 ++---
 lib/raid6/recov_neon_inner.c | 7 ++-----
 2 files changed, 4 insertions(+), 8 deletions(-)
lib/raid6/neon.uc +2 -3
index d5242f544551..b7c68030da4f 100644
@@ -28,7 +28,6 @@
 
 typedef uint8x16_t unative_t;
 
-#define NBYTES(x) ((unative_t){x,x,x,x, x,x,x,x, x,x,x,x, x,x,x,x})
 #define NSIZE	sizeof(unative_t)
 
 /*
@@ -61,7 +60,7 @@ void raid6_neon$#_gen_syndrome_real(int disks, unsigned long bytes, void **ptrs)
 	int d, z, z0;
 
 	register unative_t wd$$, wq$$, wp$$, w1$$, w2$$;
-	const unative_t x1d = NBYTES(0x1d);
+	const unative_t x1d = vdupq_n_u8(0x1d);
 
 	z0 = disks - 3;		/* Highest data disk */
 	p = dptr[z0+1];		/* XOR parity */
@@ -92,7 +91,7 @@ void raid6_neon$#_xor_syndrome_real(int disks, int start, int stop,
 	int d, z, z0;
 
 	register unative_t wd$$, wq$$, wp$$, w1$$, w2$$;
-	const unative_t x1d = NBYTES(0x1d);
+	const unative_t x1d = vdupq_n_u8(0x1d);
 
 	z0 = stop;		/* P/Q right side optimization */
 	p = dptr[disks-2];	/* XOR parity */
lib/raid6/recov_neon_inner.c +2 -5
index 8cd20c9f834a..7d00c31a6547 100644
@@ -10,11 +10,6 @@
 
 #include <arm_neon.h>
 
-static const uint8x16_t x0f = {
-	0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f,
-	0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f,
-};
-
 #ifdef CONFIG_ARM
 /*
  * AArch32 does not provide this intrinsic natively because it does not
@@ -41,6 +36,7 @@ void __raid6_2data_recov_neon(int bytes, uint8_t *p, uint8_t *q, uint8_t *dp,
 	uint8x16_t pm1 = vld1q_u8(pbmul + 16);
 	uint8x16_t qm0 = vld1q_u8(qmul);
 	uint8x16_t qm1 = vld1q_u8(qmul + 16);
+	uint8x16_t x0f = vdupq_n_u8(0x0f);
 
 	/*
 	 * while ( bytes-- ) {
@@ -87,6 +83,7 @@ void __raid6_datap_recov_neon(int bytes, uint8_t *p, uint8_t *q, uint8_t *dq,
 {
 	uint8x16_t qm0 = vld1q_u8(qmul);
 	uint8x16_t qm1 = vld1q_u8(qmul + 16);
+	uint8x16_t x0f = vdupq_n_u8(0x0f);
 
 	/*
 	 * while (bytes--) {
Signature · 1 line
2.21.0.rc2.261.ga7da99ff1b-goog
Reviewed-by on v1 1/1
#1 lore raw
Quoted 7 lines · On Tue, 26 Feb 2019 at 05:03, <[email protected]> wrote:

Clang warns: vector initializers are not compatible with NEON intrinsics in big endian mode [-Wnonportable-vector-initialization]

While this is usually the case, it's not an issue for this case since we're initializing the uint8x16_t (16x uint8_t's) with the same value.

Quoted 8 lines

Instead, use vdupq_n_u8 which both compilers lower into a single movi
instruction: https://godbolt.org/z/vBrgzt

This avoids the static storage for a constant value.

Much better, thanks,

Did you double check that the intrinsic exists on 32-bit ARM as well? I assume it does, but please make sure if you haven't yet.

If so,

Quoted 70 lines
---
 lib/raid6/neon.uc            | 5 ++---
 lib/raid6/recov_neon_inner.c | 7 ++-----
 2 files changed, 4 insertions(+), 8 deletions(-)
lib/raid6/neon.uc +2 -3
index d5242f544551..b7c68030da4f 100644
@@ -28,7 +28,6 @@
 
 typedef uint8x16_t unative_t;
 
-#define NBYTES(x) ((unative_t){x,x,x,x, x,x,x,x, x,x,x,x, x,x,x,x})
 #define NSIZE  sizeof(unative_t)
 
 /*
@@ -61,7 +60,7 @@ void raid6_neon$#_gen_syndrome_real(int disks, unsigned long bytes, void **ptrs)
        int d, z, z0;
 
        register unative_t wd$$, wq$$, wp$$, w1$$, w2$$;
-       const unative_t x1d = NBYTES(0x1d);
+       const unative_t x1d = vdupq_n_u8(0x1d);
 
        z0 = disks - 3;         /* Highest data disk */
        p = dptr[z0+1];         /* XOR parity */
@@ -92,7 +91,7 @@ void raid6_neon$#_xor_syndrome_real(int disks, int start, int stop,
        int d, z, z0;
 
        register unative_t wd$$, wq$$, wp$$, w1$$, w2$$;
-       const unative_t x1d = NBYTES(0x1d);
+       const unative_t x1d = vdupq_n_u8(0x1d);
 
        z0 = stop;              /* P/Q right side optimization */
        p = dptr[disks-2];      /* XOR parity */
lib/raid6/recov_neon_inner.c +2 -6
index 8cd20c9f834a..7d00c31a6547 100644
@@ -10,11 +10,6 @@
 
 #include <arm_neon.h>
 
-static const uint8x16_t x0f = {
-       0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f,
-       0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f,
-};
-
 #ifdef CONFIG_ARM
 /*
  * AArch32 does not provide this intrinsic natively because it does not
@@ -41,6 +36,7 @@ void __raid6_2data_recov_neon(int bytes, uint8_t *p, uint8_t *q, uint8_t *dp,
        uint8x16_t pm1 = vld1q_u8(pbmul + 16);
        uint8x16_t qm0 = vld1q_u8(qmul);
        uint8x16_t qm1 = vld1q_u8(qmul + 16);
+       uint8x16_t x0f = vdupq_n_u8(0x0f);
 
        /*
         * while ( bytes-- ) {
@@ -87,6 +83,7 @@ void __raid6_datap_recov_neon(int bytes, uint8_t *p, uint8_t *q, uint8_t *dq,
 {
        uint8x16_t qm0 = vld1q_u8(qmul);
        uint8x16_t qm1 = vld1q_u8(qmul + 16);
+       uint8x16_t x0f = vdupq_n_u8(0x0f);
 
        /*
         * while (bytes--) {
--

2.21.0.rc2.261.ga7da99ff1b-goog

#2 lore raw
Quoted 14 lines · Ard Biesheuvel
Quoted 12 lines · On Tue, 26 Feb 2019 at 05:03, <[email protected]> wrote:

Clang warns: vector initializers are not compatible with NEON intrinsics in big endian mode [-Wnonportable-vector-initialization]

While this is usually the case, it's not an issue for this case since we're initializing the uint8x16_t (16x uint8_t's) with the same value.

Instead, use vdupq_n_u8 which both compilers lower into a single movi
instruction: https://godbolt.org/z/vBrgzt

This avoids the static storage for a constant value.

Quoted 8 lines
Quoted 3 lines

Much better, thanks,

Did you double check that the intrinsic exists on 32-bit ARM as well? I assume it does, but please make sure if you haven't yet.

Thanks for the review!
Looking through Clang's generated arm_neon.h, vdupq_n_u8 seems to have 2 definitions predicated on __LITTLE_ENDIAN__ (not __arch64__ or __ARM_ARCH >= 8 like some of the other types and functions).

So NEON got some additions in v8? Is there a doc that lists them?
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0491g/BABDBBJB.html
is where I found vdupq_n_u8, but it doesn't seem to mention compatibility (so I assume it's been around since the introduction of NEON?).

Signature · 2 lines
Thanks,
~Nick Desaulniers
#3 lore raw
Quoted 21 lines · Nick Desaulniers
Quoted 18 lines · Ard Biesheuvel
Quoted 15 lines · On Tue, 26 Feb 2019 at 05:03, <[email protected]> wrote:

Clang warns: vector initializers are not compatible with NEON intrinsics in big endian mode [-Wnonportable-vector-initialization]

While this is usually the case, it's not an issue for this case since we're initializing the uint8x16_t (16x uint8_t's) with the same value.

Instead, use vdupq_n_u8 which both compilers lower into a single movi
instruction: https://godbolt.org/z/vBrgzt

This avoids the static storage for a constant value.

Quoted 6 lines
Quoted 4 lines

Much better, thanks,

Did you double check that the intrinsic exists on 32-bit ARM as well? I assume it does, but please make sure if you haven't yet.

Thanks for the review!

My pleasure.

Quoted 5 lines

Looking through Clang's generated arm_neon.h, vdupq_n_u8 seems to have 2 definitions predicated on __LITTLE_ENDIAN__ (not __arch64__ or __ARM_ARCH >= 8 like some of the other types and functions).

So NEON got some additions in v8?

Basically, yes. One example is right there in the NEON recovery code, guarded by #ifdef CONFIG_ARM.

New intrinsics were also introduced for the crypto instructions, although I think those were also added to the 32-bit version of arm_neon.h

Quoted 1 line

Is there a doc that lists them?

Not that I know of.

Quoted 4 lines

http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0491g/BABDBBJB.html
is where I found vdupq_n_u8, but it doesn't seem to mention compatibility (so I assume it's been around since the introduction of NEON?).

Yes, it appears that the 32-bit arm_neon.h header on my box defines this intrinsic, so this should be fine.

#4 lore raw
In reply to Nick Desaulniers
Quoted 30 lines · Nick Desaulniers
Quoted 14 lines · Ard Biesheuvel
Quoted 12 lines · On Tue, 26 Feb 2019 at 05:03, <[email protected]> wrote:

Clang warns: vector initializers are not compatible with NEON intrinsics in big endian mode [-Wnonportable-vector-initialization]

While this is usually the case, it's not an issue for this case since we're initializing the uint8x16_t (16x uint8_t's) with the same value.

Instead, use vdupq_n_u8 which both compilers lower into a single movi
instruction: https://godbolt.org/z/vBrgzt

This avoids the static storage for a constant value.

Quoted 8 lines
Quoted 3 lines

Much better, thanks,

Did you double check that the intrinsic exists on 32-bit ARM as well? I assume it does, but please make sure if you haven't yet.

Thanks for the review!
Looking through Clang's generated arm_neon.h, vdupq_n_u8 seems to have 2 definitions predicated on __LITTLE_ENDIAN__ (not __arch64__ or __ARM_ARCH >= 8 like some of the other types and functions).

Quoted 5 lines

So NEON got some additions in v8? Is there a doc that lists them?
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0491g/BABDBBJB.html
is where I found vdupq_n_u8, but it doesn't seem to mention compatibility (so I assume it's been around since the introduction of NEON?).

FWIW the most recent 'proper' spec document I know of is this one:

http://infocenter.arm.com/help/topic/com.arm.doc.ihi0073b/index.html

Apparently we have a more interactive playground on the new site, too:

https://developer.arm.com/technologies/neon/intrinsics

Robin.

#5 lore raw
Quoted 38 lines · Robin Murphy
Quoted 30 lines · Nick Desaulniers
Quoted 14 lines · Ard Biesheuvel
Quoted 12 lines · On Tue, 26 Feb 2019 at 05:03, <[email protected]> wrote:

Clang warns: vector initializers are not compatible with NEON intrinsics in big endian mode [-Wnonportable-vector-initialization]

While this is usually the case, it's not an issue for this case since we're initializing the uint8x16_t (16x uint8_t's) with the same value.

Instead, use vdupq_n_u8 which both compilers lower into a single movi
instruction: https://godbolt.org/z/vBrgzt

This avoids the static storage for a constant value.

Quoted 8 lines
Quoted 3 lines

Much better, thanks,

Did you double check that the intrinsic exists on 32-bit ARM as well? I assume it does, but please make sure if you haven't yet.

Thanks for the review!
Looking through Clang's generated arm_neon.h, vdupq_n_u8 seems to have 2 definitions predicated on __LITTLE_ENDIAN__ (not __arch64__ or __ARM_ARCH >= 8 like some of the other types and functions).

Quoted 5 lines

So NEON got some additions in v8? Is there a doc that lists them?
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0491g/BABDBBJB.html
is where I found vdupq_n_u8, but it doesn't seem to mention compatibility (so I assume it's been around since the introduction of NEON?).

Quoted 3 lines

FWIW the most recent 'proper' spec document I know of is this one:

http://infocenter.arm.com/help/topic/com.arm.doc.ihi0073b/index.html

Bookmarked, thanks!
Ard, page 171 mentions armv7, armv8 for supported architectures for vdupq_n_u8.

Quoted 5 lines

Apparently we have a more interactive playground on the new site, too:

https://developer.arm.com/technologies/neon/intrinsics

Also bookmarked! I'm also super happy to see this; I'm familiar with Intel's equivalent:
https://software.intel.com/sites/landingpage/IntrinsicsGuide/

Interactive sites like these are quite useful. Reading a post
recently: https://www.sigarch.org/simd-instructions-considered-harmful/

"The IA-32 instruction set has grown from 80 to around 1400 instructions since 1978, largely fueled by SIMD."

reminded me how useful and almost necessary the interactive sites are for navigating the large swathes of SIMD extensions. (no comment on the title of that article)

Signature · 2 lines
Thanks,
~Nick Desaulniers
#6 lore raw
In reply to [email protected]
Quoted 11 lines · On Mon, Feb 25, 2019 at 08:03:42PM -0800, [email protected] wrote:

Clang warns: vector initializers are not compatible with NEON intrinsics in big endian mode [-Wnonportable-vector-initialization]

While this is usually the case, it's not an issue for this case since we're initializing the uint8x16_t (16x uint8_t's) with the same value.

Instead, use vdupq_n_u8 which both compilers lower into a single movi
instruction: https://godbolt.org/z/vBrgzt

This avoids the static storage for a constant value.

Quoted 7 lines
---
 lib/raid6/neon.uc            | 5 ++---
 lib/raid6/recov_neon_inner.c | 7 ++-----
 2 files changed, 4 insertions(+), 8 deletions(-)

Queued for 5.1. Thanks.

Signature · 1 line
Catalin