[PATCH v2 0/3] Improve bitmap_empty and bitmap_full

[PATCH v2 0/3] Improve bitmap_empty and bitmap_full

[Jia He]

find_fisrt_{zero_}bit are too heavy for bitmap_{full,empty}. We don't 
need to calculate and compare the position of bitmap. This set of patch
instroduces lightweight api and replaces the heavy one.

v2: Move the declarations to linux/bitops.h for compilation

Jia He (3):
  Move 2 mask macro from bitmap.h to bitops.h
  Introduce 2 bit ops api: all_is_bit_{one,zero}
  Replace find_fisrt_{zero_}bit with the new lightweight api

 include/linux/bitmap.h |  7 ++-----
 include/linux/bitops.h |  7 +++++++
 lib/find_bit.c         | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 59 insertions(+), 5 deletions(-)

-- 
2.5.0

[PATCH v2 1/3] linux/bitmap: Move 2 mask macro from bitmap.h to bitops.h

[Jia He]

This patch moves the mask macro to bitops.h and then the new introduced
api in find_bit.c can use it.

Signed-off-by: Jia He <hejianet@gmail.com>
---
 include/linux/bitmap.h | 3 ---
 include/linux/bitops.h | 4 ++++
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/include/linux/bitmap.h b/include/linux/bitmap.h
index 9653fdb..15524f6 100644
--- a/include/linux/bitmap.h
+++ b/include/linux/bitmap.h
@@ -172,9 +172,6 @@ extern unsigned int bitmap_ord_to_pos(const unsigned long *bitmap, unsigned int
 extern int bitmap_print_to_pagebuf(bool list, char *buf,
 				   const unsigned long *maskp, int nmaskbits);
 
-#define BITMAP_FIRST_WORD_MASK(start) (~0UL << ((start) & (BITS_PER_LONG - 1)))
-#define BITMAP_LAST_WORD_MASK(nbits) (~0UL >> (-(nbits) & (BITS_PER_LONG - 1)))
-
 #define small_const_nbits(nbits) \
 	(__builtin_constant_p(nbits) && (nbits) <= BITS_PER_LONG)
 
diff --git a/include/linux/bitops.h b/include/linux/bitops.h
index 2b8ed12..b881028 100644
--- a/include/linux/bitops.h
+++ b/include/linux/bitops.h
@@ -24,6 +24,10 @@
 #define GENMASK_ULL(h, l) \
 	(((~0ULL) << (l)) & (~0ULL >> (BITS_PER_LONG_LONG - 1 - (h))))
 
+/* For bitmap ops*/
+#define BITMAP_FIRST_WORD_MASK(start) (~0UL << ((start) & (BITS_PER_LONG - 1)))
+#define BITMAP_LAST_WORD_MASK(nbits) (~0UL >> (-(nbits) & (BITS_PER_LONG - 1)))
+
 extern unsigned int __sw_hweight8(unsigned int w);
 extern unsigned int __sw_hweight16(unsigned int w);
 extern unsigned int __sw_hweight32(unsigned int w);
-- 
2.5.0

[PATCH v2 2/3] lib: Introduce 2 bit ops api: all_is_bit_{one,zero}

[Jia He]

This patch introduces 2 lightweight bit api.
all_bit_is_zero return 1 if the bit string is all zero.
The addr is the start address, the size is the bit size of the bit string.
all_bit_is_one is the opposite.

Signed-off-by: Jia He <hejianet@gmail.com>
---
 lib/find_bit.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)

diff --git a/lib/find_bit.c b/lib/find_bit.c
index 18072ea..1d56d8d 100644
--- a/lib/find_bit.c
+++ b/lib/find_bit.c
@@ -131,6 +131,56 @@ unsigned long find_last_bit(const unsigned long *addr, unsigned long size)
 EXPORT_SYMBOL(find_last_bit);
 #endif
 
+#ifndef all_bit_is_zero
+/*
+ * return val: 1 means all bit is zero
+ */
+unsigned int all_bit_is_zero(const unsigned long *addr, unsigned size)
+{
+	unsigned long idx;
+	unsigned long mask = size;
+
+	if (unlikely(size == 0))
+		return 1;
+
+	if (size > BITS_PER_LONG) {
+		for (idx = 0; idx * BITS_PER_LONG < size; idx++)
+			if (addr[idx])
+				return 0;
+
+		mask = size - (idx - 1) * BITS_PER_LONG;
+	}
+
+	return !(*addr & BITMAP_LAST_WORD_MASK(mask));
+}
+EXPORT_SYMBOL(all_bit_is_zero);
+#endif
+
+#ifndef all_bit_is_one
+/*
+ * return val: 1 means all bit is one
+ */
+unsigned int all_bit_is_one(const unsigned long *addr, unsigned size)
+{
+	unsigned long idx;
+	unsigned long mask = size;
+
+	if (unlikely(size == 0))
+		return 1;
+
+	if (size > BITS_PER_LONG) {
+		for (idx = 0; idx * BITS_PER_LONG < size; idx++)
+			if (~addr[idx])
+				return 0;
+
+		mask = size - (idx - 1) * BITS_PER_LONG;
+	}
+
+	return !(~(*addr) & BITMAP_LAST_WORD_MASK(mask));
+}
+EXPORT_SYMBOL(all_bit_is_one);
+#endif
+
 #ifdef __BIG_ENDIAN
 
 /* include/linux/byteorder does not support "unsigned long" type */
-- 
2.5.0

[PATCH v3 3/3] linux/bitmap: Replace find_fisrt_{zero_}bit with the new lightweight api

[Jia He]

This is to replace find_fisrt_{zero_}bit with the new lightweight api 
all_bit_is_{one,zero} in bitmap_{full,empty}

Signed-off-by: Jia He <hejianet@gmail.com>
---
 include/linux/bitmap.h | 4 ++--
 include/linux/bitops.h | 3 +++
 2 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/include/linux/bitmap.h b/include/linux/bitmap.h
index 15524f6..8877a68 100644
--- a/include/linux/bitmap.h
+++ b/include/linux/bitmap.h
@@ -281,7 +281,7 @@ static inline int bitmap_empty(const unsigned long *src, unsigned nbits)
 	if (small_const_nbits(nbits))
 		return ! (*src & BITMAP_LAST_WORD_MASK(nbits));
 
-	return find_first_bit(src, nbits) == nbits;
+	return all_bit_is_zero(src, nbits);
 }
 
 static inline int bitmap_full(const unsigned long *src, unsigned int nbits)
@@ -289,7 +289,7 @@ static inline int bitmap_full(const unsigned long *src, unsigned int nbits)
 	if (small_const_nbits(nbits))
 		return ! (~(*src) & BITMAP_LAST_WORD_MASK(nbits));
 
-	return find_first_zero_bit(src, nbits) == nbits;
+	return all_bit_is_one(src, nbits);
 }
 
 static __always_inline int bitmap_weight(const unsigned long *src, unsigned int nbits)
diff --git a/include/linux/bitops.h b/include/linux/bitops.h
index b881028..e34f2ff 100644
--- a/include/linux/bitops.h
+++ b/include/linux/bitops.h
@@ -214,6 +214,9 @@ static inline unsigned long __ffs64(u64 word)
 	return __ffs((unsigned long)word);
 }
 
+extern unsigned int all_bit_is_one(const unsigned long *addr, unsigned size);
+extern unsigned int all_bit_is_zero(const unsigned long *addr, unsigned size);
+
 #ifdef __KERNEL__
 
 #ifndef set_mask_bits
-- 
2.5.0

Re: [PATCH v2 2/3] lib: Introduce 2 bit ops api: all_is_bit_{one,zero}

[xinhui]

hi, jia
	Nice patch. But I have one minor question. see inline comments.

On 2015/11/19 14:48, Jia He wrote:
> This patch introduces 2 lightweight bit api.
> all_bit_is_zero return 1 if the bit string is all zero.
> The addr is the start address, the size is the bit size of the bit string.
> all_bit_is_one is the opposite.
>
> Signed-off-by: Jia He <hejianet@gmail.com>
> ---
>   lib/find_bit.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++
>   1 file changed, 50 insertions(+)
>
> diff --git a/lib/find_bit.c b/lib/find_bit.c
> index 18072ea..1d56d8d 100644
> --- a/lib/find_bit.c
> +++ b/lib/find_bit.c
> @@ -131,6 +131,56 @@ unsigned long find_last_bit(const unsigned long *addr, unsigned long size)
>   EXPORT_SYMBOL(find_last_bit);
>   #endif
>
> +#ifndef all_bit_is_zero
> +/*
> + * return val: 1 means all bit is zero
> + */
> +unsigned int all_bit_is_zero(const unsigned long *addr, unsigned size)
> +{
  Seems better that size should be type of "unsigned long". Otherwise I'm afraid when we compare idx * BITS_PER_LONG with size, there might be overflow issue.

> +	unsigned long idx;
> +	unsigned long mask = size;
> +
> +	if (unlikely(size == 0))
> +		return 1;
> +
> +	if (size > BITS_PER_LONG) {
> +		for (idx = 0; idx * BITS_PER_LONG < size; idx++)
> +			if (addr[idx])
> +				return 0;
> +
> +		mask = size - (idx - 1) * BITS_PER_LONG;
> +	}
> +
> +	return !(*addr & BITMAP_LAST_WORD_MASK(mask));
> +}
> +EXPORT_SYMBOL(all_bit_is_zero);
> +#endif
> +
> +#ifndef all_bit_is_one
> +/*
> + * return val: 1 means all bit is one
> + */
> +unsigned int all_bit_is_one(const unsigned long *addr, unsigned size)
> +{
this argc of size should be type of "unsigned long", too.

thanks
xinhui

> +	unsigned long idx;
> +	unsigned long mask = size;
> +
> +	if (unlikely(size == 0))
> +		return 1;
> +
> +	if (size > BITS_PER_LONG) {
> +		for (idx = 0; idx * BITS_PER_LONG < size; idx++)
> +			if (~addr[idx])
> +				return 0;
> +
> +		mask = size - (idx - 1) * BITS_PER_LONG;
> +	}
> +
> +	return !(~(*addr) & BITMAP_LAST_WORD_MASK(mask));
> +}
> +EXPORT_SYMBOL(all_bit_is_one);
> +#endif
> +
>   #ifdef __BIG_ENDIAN
>
>   /* include/linux/byteorder does not support "unsigned long" type */
>

Re: [PATCH v2 2/3] lib: Introduce 2 bit ops api: all_is_bit_{one,zero}

[yalin wang]

> On Nov 19, 2015, at 14:48, Jia He <hejianet@gmail.com> wrote:
> 
> 
why not use memcmp() to compare with  0x0000000 or 0xffffffff  ?
memcmp() have better performance on some platforms .

Thanks

Re: [PATCH v2 2/3] lib: Introduce 2 bit ops api: all_is_bit_{one,zero}

[hejianet]

Thanks， I will add it in next verison
B.R.
Justin

在 11/19/15 4:40 PM, xinhui 写道:
> hi, jia
>     Nice patch. But I have one minor question. see inline comments.
>
> On 2015/11/19 14:48, Jia He wrote:
>> This patch introduces 2 lightweight bit api.
>> all_bit_is_zero return 1 if the bit string is all zero.
>> The addr is the start address, the size is the bit size of the bit 
>> string.
>> all_bit_is_one is the opposite.
>>
>> Signed-off-by: Jia He <hejianet@gmail.com>
>> ---
>>   lib/find_bit.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++
>>   1 file changed, 50 insertions(+)
>>
>> diff --git a/lib/find_bit.c b/lib/find_bit.c
>> index 18072ea..1d56d8d 100644
>> --- a/lib/find_bit.c
>> +++ b/lib/find_bit.c
>> @@ -131,6 +131,56 @@ unsigned long find_last_bit(const unsigned long 
>> *addr, unsigned long size)
>>   EXPORT_SYMBOL(find_last_bit);
>>   #endif
>>
>> +#ifndef all_bit_is_zero
>> +/*
>> + * return val: 1 means all bit is zero
>> + */
>> +unsigned int all_bit_is_zero(const unsigned long *addr, unsigned size)
>> +{
>  Seems better that size should be type of "unsigned long". Otherwise 
> I'm afraid when we compare idx * BITS_PER_LONG with size, there might 
> be overflow issue.
>
>> +    unsigned long idx;
>> +    unsigned long mask = size;
>> +
>> +    if (unlikely(size == 0))
>> +        return 1;
>> +
>> +    if (size > BITS_PER_LONG) {
>> +        for (idx = 0; idx * BITS_PER_LONG < size; idx++)
>> +            if (addr[idx])
>> +                return 0;
>> +
>> +        mask = size - (idx - 1) * BITS_PER_LONG;
>> +    }
>> +
>> +    return !(*addr & BITMAP_LAST_WORD_MASK(mask));
>> +}
>> +EXPORT_SYMBOL(all_bit_is_zero);
>> +#endif
>> +
>> +#ifndef all_bit_is_one
>> +/*
>> + * return val: 1 means all bit is one
>> + */
>> +unsigned int all_bit_is_one(const unsigned long *addr, unsigned size)
>> +{
> this argc of size should be type of "unsigned long", too.
>
> thanks
> xinhui
>
>> +    unsigned long idx;
>> +    unsigned long mask = size;
>> +
>> +    if (unlikely(size == 0))
>> +        return 1;
>> +
>> +    if (size > BITS_PER_LONG) {
>> +        for (idx = 0; idx * BITS_PER_LONG < size; idx++)
>> +            if (~addr[idx])
>> +                return 0;
>> +
>> +        mask = size - (idx - 1) * BITS_PER_LONG;
>> +    }
>> +
>> +    return !(~(*addr) & BITMAP_LAST_WORD_MASK(mask));
>> +}
>> +EXPORT_SYMBOL(all_bit_is_one);
>> +#endif
>> +
>>   #ifdef __BIG_ENDIAN
>>
>>   /* include/linux/byteorder does not support "unsigned long" type */
>>
>
> -- 
> To unsubscribe from this list: send the line "unsubscribe 
> linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: [PATCH v2 0/3] Improve bitmap_empty and bitmap_full

[Rasmus Villemoes]

On Thu, Nov 19 2015, Jia He <hejianet@gmail.com> wrote:

> find_fisrt_{zero_}bit are too heavy for bitmap_{full,empty}. We don't 
> need to calculate and compare the position of bitmap. This set of patch
> instroduces lightweight api and replaces the heavy one.
>

Please check the history of the code you're modifying. git blame
include/linux/bitmap.h would immediately point you to 2afe27c718b
"lib/bitmap.c: bitmap_[empty,full]: remove code duplication". While it's
obviously true that find_first_bit does slightly more work than strictly
necessary to establish whether the bitmap is empty, it does the same
number of memory accesses, so I wouldn't consider it particularly
heavy-weight. Getting rid of .text as 2afe27c718b did is a good thing,
so you'd have to explain why we should reintroduce specialized functions
for this.

Your code is also buggy :-(

Rasmus