On Wed, Aug 28, 2019 at 06:11:23PM +0200, Peter Zijlstra wrote:
> On Wed, Aug 28, 2019 at 05:19:21PM +0200, Peter Zijlstra wrote:
> > On Mon, Aug 26, 2019 at 07:47:35AM -0700, kan.li...@linux.intel.com wrote:
>
> > > + return mul_u64_u32_div(slots, val, 0xff);
> >
> > But also; x86_64 seems to lack a sane implementation of that function,
> > and it currently compiles into utter crap (it can be 2 instructions).
This one actually builds defconfig :-)
---
Subject: x86/math64: Provide a sane mul_u64_u32_div() implementation for x86_64
From: Peter Zijlstra <pet...@infradead.org>
Date: Wed Aug 28 17:39:46 CEST 2019
On x86_64 we can do a u64 * u64 -> u128 widening multiply followed by
a u128 / u64 -> u64 division to implement a sane version of
mul_u64_u32_div().
Signed-off-by: Peter Zijlstra (Intel) <pet...@infradead.org>
---
arch/x86/include/asm/div64.h | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/arch/x86/include/asm/div64.h b/arch/x86/include/asm/div64.h
index 20a46150e0a8..9b8cb50768c2 100644
--- a/arch/x86/include/asm/div64.h
+++ b/arch/x86/include/asm/div64.h
@@ -73,6 +73,19 @@ static inline u64 mul_u32_u32(u32 a, u32 b)
#else
# include <asm-generic/div64.h>
+
+static inline u64 mul_u64_u32_div(u64 a, u32 mul, u32 div)
+{
+ u64 q;
+
+ asm ("mulq %2; divq %3" : "=a" (q)
+ : "a" (a), "rm" ((u64)mul), "rm" ((u64)div)
+ : "rdx");
+
+ return q;
+}
+#define mul_u64_u32_div mul_u64_u32_div
+
#endif /* CONFIG_X86_32 */
#endif /* _ASM_X86_DIV64_H */
