* [RESEND PATCH] ARM: vfp: Improve data types in vfp_estimate_div128to64()
@ 2025-02-09 8:21 Thorsten Blum
0 siblings, 0 replies; 3+ messages in thread
From: Thorsten Blum @ 2025-02-09 8:21 UTC (permalink / raw)
To: Russell King; +Cc: Thorsten Blum, linux-arm-kernel, linux-kernel
The divisors mh and ml can both be u32 instead of u64.
Since do_div() implicitly casts the divisors from u64 to u32, changing
their data types to u32 also removes the following Coccinelle warnings
reported by do_div.cocci:
arch/arm/vfp/vfp.h:121:2-8: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead.
arch/arm/vfp/vfp.h:135:2-8: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead.
Use upper_32_bits() to compare 32-bit numbers instead of 64-bit numbers
and to prevent the warning:
left shift count >= width of type [-Wshift-count-overflow]
Compile-tested only.
Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
---
arch/arm/vfp/vfp.h | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/arch/arm/vfp/vfp.h b/arch/arm/vfp/vfp.h
index e43a630f8a16..29073d89ef00 100644
--- a/arch/arm/vfp/vfp.h
+++ b/arch/arm/vfp/vfp.h
@@ -109,12 +109,13 @@ static inline u64 vfp_hi64multiply64(u64 n, u64 m)
static inline u64 vfp_estimate_div128to64(u64 nh, u64 nl, u64 m)
{
- u64 mh, ml, remh, reml, termh, terml, z;
+ u64 remh, reml, termh, terml, z;
+ u32 mh, ml;
if (nh >= m)
return ~0ULL;
mh = m >> 32;
- if (mh << 32 <= nh) {
+ if (mh <= upper_32_bits(nh)) {
z = 0xffffffff00000000ULL;
} else {
z = nh;
@@ -129,7 +130,7 @@ static inline u64 vfp_estimate_div128to64(u64 nh, u64 nl, u64 m)
add128(&remh, &reml, remh, reml, mh, ml);
}
remh = (remh << 32) | (reml >> 32);
- if (mh << 32 <= remh) {
+ if (mh <= upper_32_bits(remh)) {
z |= 0xffffffff;
} else {
do_div(remh, mh);
--
2.48.1
^ permalink raw reply related [flat|nested] 3+ messages in thread* [RESEND PATCH] ARM: vfp: Improve data types in vfp_estimate_div128to64()
@ 2025-01-13 9:08 Thorsten Blum
0 siblings, 0 replies; 3+ messages in thread
From: Thorsten Blum @ 2025-01-13 9:08 UTC (permalink / raw)
To: Russell King; +Cc: Thorsten Blum, linux-arm-kernel, linux-kernel
The divisors mh and ml can both be u32 instead of u64.
Since do_div() implicitly casts the divisors from u64 to u32, changing
their data types to u32 also removes the following Coccinelle warnings
reported by do_div.cocci:
arch/arm/vfp/vfp.h:121:2-8: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead.
arch/arm/vfp/vfp.h:135:2-8: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead.
Use upper_32_bits() to compare 32-bit numbers instead of 64-bit numbers
and to prevent the warning:
left shift count >= width of type [-Wshift-count-overflow]
Compile-tested only.
Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
---
arch/arm/vfp/vfp.h | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/arch/arm/vfp/vfp.h b/arch/arm/vfp/vfp.h
index e43a630f8a16..29073d89ef00 100644
--- a/arch/arm/vfp/vfp.h
+++ b/arch/arm/vfp/vfp.h
@@ -109,12 +109,13 @@ static inline u64 vfp_hi64multiply64(u64 n, u64 m)
static inline u64 vfp_estimate_div128to64(u64 nh, u64 nl, u64 m)
{
- u64 mh, ml, remh, reml, termh, terml, z;
+ u64 remh, reml, termh, terml, z;
+ u32 mh, ml;
if (nh >= m)
return ~0ULL;
mh = m >> 32;
- if (mh << 32 <= nh) {
+ if (mh <= upper_32_bits(nh)) {
z = 0xffffffff00000000ULL;
} else {
z = nh;
@@ -129,7 +130,7 @@ static inline u64 vfp_estimate_div128to64(u64 nh, u64 nl, u64 m)
add128(&remh, &reml, remh, reml, mh, ml);
}
remh = (remh << 32) | (reml >> 32);
- if (mh << 32 <= remh) {
+ if (mh <= upper_32_bits(remh)) {
z |= 0xffffffff;
} else {
do_div(remh, mh);
--
2.47.1
^ permalink raw reply related [flat|nested] 3+ messages in thread* [RESEND PATCH] ARM: vfp: Improve data types in vfp_estimate_div128to64()
@ 2024-07-10 0:34 Thorsten Blum
0 siblings, 0 replies; 3+ messages in thread
From: Thorsten Blum @ 2024-07-10 0:34 UTC (permalink / raw)
To: linux; +Cc: linux-arm-kernel, linux-kernel, Thorsten Blum
The divisors mh and ml can both be u32 instead of u64.
Since do_div() implicitly casts the divisors from u64 to u32, changing
their data types to u32 also removes the following Coccinelle warnings
reported by do_div.cocci:
arch/arm/vfp/vfp.h:121:2-8: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead.
arch/arm/vfp/vfp.h:135:2-8: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead.
Use upper_32_bits() to compare 32-bit numbers instead of 64-bit numbers
and to prevent the warning:
left shift count >= width of type [-Wshift-count-overflow]
Compile-tested only.
Signed-off-by: Thorsten Blum <thorsten.blum@toblux.com>
---
arch/arm/vfp/vfp.h | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/arch/arm/vfp/vfp.h b/arch/arm/vfp/vfp.h
index e43a630f8a16..29073d89ef00 100644
--- a/arch/arm/vfp/vfp.h
+++ b/arch/arm/vfp/vfp.h
@@ -109,12 +109,13 @@ static inline u64 vfp_hi64multiply64(u64 n, u64 m)
static inline u64 vfp_estimate_div128to64(u64 nh, u64 nl, u64 m)
{
- u64 mh, ml, remh, reml, termh, terml, z;
+ u64 remh, reml, termh, terml, z;
+ u32 mh, ml;
if (nh >= m)
return ~0ULL;
mh = m >> 32;
- if (mh << 32 <= nh) {
+ if (mh <= upper_32_bits(nh)) {
z = 0xffffffff00000000ULL;
} else {
z = nh;
@@ -129,7 +130,7 @@ static inline u64 vfp_estimate_div128to64(u64 nh, u64 nl, u64 m)
add128(&remh, &reml, remh, reml, mh, ml);
}
remh = (remh << 32) | (reml >> 32);
- if (mh << 32 <= remh) {
+ if (mh <= upper_32_bits(remh)) {
z |= 0xffffffff;
} else {
do_div(remh, mh);
--
2.45.2
^ permalink raw reply related [flat|nested] 3+ messages in thread
end of thread, other threads:[~2025-02-09 8:23 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-02-09 8:21 [RESEND PATCH] ARM: vfp: Improve data types in vfp_estimate_div128to64() Thorsten Blum
-- strict thread matches above, loose matches on Subject: below --
2025-01-13 9:08 Thorsten Blum
2024-07-10 0:34 Thorsten Blum
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).