This is an automated email from the ASF dual-hosted git repository.
xiaoxiang781216 pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/nuttx.git
The following commit(s) were added to refs/heads/master by this push:
new 52e6c0abcab libc/machine/arm64: Use .d rather than .2d for element mov
instructions
52e6c0abcab is described below
commit 52e6c0abcab576c93f4f3526c852e3ddee57be35
Author: wangmingrong1 <[email protected]>
AuthorDate: Mon Nov 2 18:59:49 2026 +0800
libc/machine/arm64: Use .d rather than .2d for element mov instructions
Use .d rather than .2d for element mov instructions in string routines
so the assembly compiles with clang too.The same changes are also found in
github arm-software:
https://github.com/ARM-software/optimized-routines/commit/4f4e530b2fda475a37e583bc113ad2fd3e83dc7f
Signed-off-by: wangmingrong1 <[email protected]>
---
libs/libc/machine/arm64/arch_memchr.S | 6 +++---
libs/libc/machine/arm64/arch_strchr.S | 6 +++---
libs/libc/machine/arm64/arch_strchrnul.S | 6 +++---
libs/libc/machine/arm64/arch_strrchr.S | 10 +++++-----
4 files changed, 14 insertions(+), 14 deletions(-)
diff --git a/libs/libc/machine/arm64/arch_memchr.S
b/libs/libc/machine/arm64/arch_memchr.S
index 576a31fc824..68081a1c4e4 100644
--- a/libs/libc/machine/arm64/arch_memchr.S
+++ b/libs/libc/machine/arm64/arch_memchr.S
@@ -114,7 +114,7 @@ def_fn ARCH_LIBCFUN(memchr)
and vhas_chr2.16b, vhas_chr2.16b, vrepmask.16b
addp vend.16b, vhas_chr1.16b, vhas_chr2.16b /* 256->128 */
addp vend.16b, vend.16b, vend.16b /* 128->64 */
- mov synd, vend.2d[0]
+ mov synd, vend.d[0]
/* Clear the soff*2 lower bits */
lsl tmp, soff, #1
lsr synd, synd, tmp
@@ -134,7 +134,7 @@ def_fn ARCH_LIBCFUN(memchr)
/* Use a fast check for the termination condition */
orr vend.16b, vhas_chr1.16b, vhas_chr2.16b
addp vend.2d, vend.2d, vend.2d
- mov synd, vend.2d[0]
+ mov synd, vend.d[0]
/* We're not out of data, loop if we haven't found the character */
cbz synd, .Lloop
@@ -144,7 +144,7 @@ def_fn ARCH_LIBCFUN(memchr)
and vhas_chr2.16b, vhas_chr2.16b, vrepmask.16b
addp vend.16b, vhas_chr1.16b, vhas_chr2.16b /* 256->128 */
addp vend.16b, vend.16b, vend.16b /* 128->64 */
- mov synd, vend.2d[0]
+ mov synd, vend.d[0]
/* Only do the clear for the last possible block */
b.hi .Ltail
diff --git a/libs/libc/machine/arm64/arch_strchr.S
b/libs/libc/machine/arm64/arch_strchr.S
index 0f62d19344e..573d5937cbc 100644
--- a/libs/libc/machine/arm64/arch_strchr.S
+++ b/libs/libc/machine/arm64/arch_strchr.S
@@ -121,7 +121,7 @@ def_fn ARCH_LIBCFUN(strchr)
addp vend1.16b, vend1.16b, vend2.16b // 128->64
lsr tmp1, tmp3, tmp1
- mov tmp3, vend1.2d[0]
+ mov tmp3, vend1.d[0]
bic tmp1, tmp3, tmp1 // Mask padding bits.
cbnz tmp1, .Ltail
@@ -136,7 +136,7 @@ def_fn ARCH_LIBCFUN(strchr)
orr vend2.16b, vhas_nul2.16b, vhas_chr2.16b
orr vend1.16b, vend1.16b, vend2.16b
addp vend1.2d, vend1.2d, vend1.2d
- mov tmp1, vend1.2d[0]
+ mov tmp1, vend1.d[0]
cbz tmp1, .Lloop
/* Termination condition found. Now need to establish exactly why
@@ -150,7 +150,7 @@ def_fn ARCH_LIBCFUN(strchr)
addp vend1.16b, vend1.16b, vend2.16b // 256->128
addp vend1.16b, vend1.16b, vend2.16b // 128->64
- mov tmp1, vend1.2d[0]
+ mov tmp1, vend1.d[0]
.Ltail:
/* Count the trailing zeros, by bit reversing... */
rbit tmp1, tmp1
diff --git a/libs/libc/machine/arm64/arch_strchrnul.S
b/libs/libc/machine/arm64/arch_strchrnul.S
index 3c3ea8c818a..a041ef41578 100644
--- a/libs/libc/machine/arm64/arch_strchrnul.S
+++ b/libs/libc/machine/arm64/arch_strchrnul.S
@@ -113,7 +113,7 @@ def_fn ARCH_LIBCFUN(strchrnul)
addp vend1.16b, vend1.16b, vend1.16b // 128->64
lsr tmp1, tmp3, tmp1
- mov tmp3, vend1.2d[0]
+ mov tmp3, vend1.d[0]
bic tmp1, tmp3, tmp1 // Mask padding bits.
cbnz tmp1, .Ltail
@@ -128,7 +128,7 @@ def_fn ARCH_LIBCFUN(strchrnul)
orr vhas_chr2.16b, vhas_nul2.16b, vhas_chr2.16b
orr vend1.16b, vhas_chr1.16b, vhas_chr2.16b
addp vend1.2d, vend1.2d, vend1.2d
- mov tmp1, vend1.2d[0]
+ mov tmp1, vend1.d[0]
cbz tmp1, .Lloop
/* Termination condition found. Now need to establish exactly why
@@ -138,7 +138,7 @@ def_fn ARCH_LIBCFUN(strchrnul)
addp vend1.16b, vhas_chr1.16b, vhas_chr2.16b // 256->128
addp vend1.16b, vend1.16b, vend1.16b // 128->64
- mov tmp1, vend1.2d[0]
+ mov tmp1, vend1.d[0]
.Ltail:
/* Count the trailing zeros, by bit reversing... */
rbit tmp1, tmp1
diff --git a/libs/libc/machine/arm64/arch_strrchr.S
b/libs/libc/machine/arm64/arch_strrchr.S
index 5e4ae80f18d..513946f1ee9 100644
--- a/libs/libc/machine/arm64/arch_strrchr.S
+++ b/libs/libc/machine/arm64/arch_strrchr.S
@@ -124,10 +124,10 @@ def_fn ARCH_LIBCFUN(strrchr)
addp vhas_chr1.16b, vhas_chr1.16b, vhas_chr2.16b // 256->128
addp vhas_nul1.16b, vhas_nul1.16b, vhas_nul1.16b // 128->64
addp vhas_chr1.16b, vhas_chr1.16b, vhas_chr1.16b // 128->64
- mov nul_match, vhas_nul1.2d[0]
+ mov nul_match, vhas_nul1.d[0]
lsl tmp1, tmp1, #1
mov const_m1, #~0
- mov chr_match, vhas_chr1.2d[0]
+ mov chr_match, vhas_chr1.d[0]
lsr tmp3, const_m1, tmp1
bic nul_match, nul_match, tmp3 // Mask padding bits.
@@ -150,15 +150,15 @@ def_fn ARCH_LIBCFUN(strrchr)
addp vhas_chr1.16b, vhas_chr1.16b, vhas_chr2.16b // 256->128
addp vend1.16b, vend1.16b, vend1.16b // 128->64
addp vhas_chr1.16b, vhas_chr1.16b, vhas_chr1.16b // 128->64
- mov nul_match, vend1.2d[0]
- mov chr_match, vhas_chr1.2d[0]
+ mov nul_match, vend1.d[0]
+ mov chr_match, vhas_chr1.d[0]
cbz nul_match, .Lloop
and vhas_nul1.16b, vhas_nul1.16b, vrepmask_0.16b
and vhas_nul2.16b, vhas_nul2.16b, vrepmask_0.16b
addp vhas_nul1.16b, vhas_nul1.16b, vhas_nul2.16b
addp vhas_nul1.16b, vhas_nul1.16b, vhas_nul1.16b
- mov nul_match, vhas_nul1.2d[0]
+ mov nul_match, vhas_nul1.d[0]
.Ltail:
/* Work out exactly where the string ends. */
