mm/zsmalloc: remove unnecessary check

RFC PATCH In review Mahendran Ganesh 2 participants 7 messages Nov 20, 2014 – Nov 24, 2014

Subsystems (2)

Memory management Zsmalloc compressed slab memory allocator

lore

Share & tools

lore

Mahendran Ganesh

author

2014-11-20 13:22 UTC #0 lore raw

RFC PATCH 1/1 mm/zsmalloc: remove unnecessary check

ZS_SIZE_CLASSES is calc by:
((ZS_MAX_ALLOC_SIZE - ZS_MIN_ALLOC_SIZE) / ZS_SIZE_CLASS_DELTA + 1)

So when i is in [0, ZS_SIZE_CLASSES - 1), the size:
size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA
will not be greater than ZS_MAX_ALLOC_SIZE

This patch removes the unnecessary check.

Signed-off-by Mahendran Ganesh <[email protected]>

---
 mm/zsmalloc.c |    2 --
 1 file changed, 2 deletions(-)

mm/zsmalloc.c +0 -2

index b3b57ef..f2279e2 100644
@@ -973,8 +973,6 @@ struct zs_pool *zs_create_pool(gfp_t flags)
 		struct size_class *prev_class;
 
 		size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA;
-		if (size > ZS_MAX_ALLOC_SIZE)
-			size = ZS_MAX_ALLOC_SIZE;
 		pages_per_zspage = get_pages_per_zspage(size);
 
 		/*

Signature · 7 lines

1.7.9.5

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Minchan Kim

maintainer

2014-11-21 03:54 UTC #1 lore raw

Quoted 8 lines · Mahendran Ganesh

ZS_SIZE_CLASSES is calc by:
((ZS_MAX_ALLOC_SIZE - ZS_MIN_ALLOC_SIZE) / ZS_SIZE_CLASS_DELTA + 1)
So when i is in [0, ZS_SIZE_CLASSES - 1), the size:
size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA
will not be greater than ZS_MAX_ALLOC_SIZE
This patch removes the unnecessary check.

It depends on ZS_MIN_ALLOC_SIZE.
For example, we would change min to 8 but MAX is still 4096. ZS_SIZE_CLASSES is (4096 - 8) / 16 + 1 = 256 so 8 + 255 * 16 = 4088, which exceeds the max.

Quoted 27 lines

Signed-off-by Mahendran Ganesh <[email protected]>

---
 mm/zsmalloc.c |    2 --
 1 file changed, 2 deletions(-)

mm/zsmalloc.c +0 -2

index b3b57ef..f2279e2 100644
@@ -973,8 +973,6 @@ struct zs_pool *zs_create_pool(gfp_t flags)
 		struct size_class *prev_class;
 
 		size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA;
-		if (size > ZS_MAX_ALLOC_SIZE)
-			size = ZS_MAX_ALLOC_SIZE;
 		pages_per_zspage = get_pages_per_zspage(size);
 
 		/*

Signature · 7 lines

1.7.9.5

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Signature · 2 lines

Kind regards,
Minchan Kim

Ganesh Mahendran

author

2014-11-21 05:33 UTC #2 lore raw

Hello

2014-11-21 11:54 GMT+08:00 Minchan Kim <[email protected]>:

Quoted 8 lines · Mahendran Ganesh

Quoted 7 lines
ZS_SIZE_CLASSES is calc by:
((ZS_MAX_ALLOC_SIZE - ZS_MIN_ALLOC_SIZE) / ZS_SIZE_CLASS_DELTA + 1)
So when i is in [0, ZS_SIZE_CLASSES - 1), the size:
size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA
will not be greater than ZS_MAX_ALLOC_SIZE

Quoted 6 lines

Quoted 1 line
This patch removes the unnecessary check.
It depends on ZS_MIN_ALLOC_SIZE.
For example, we would change min to 8 but MAX is still 4096. ZS_SIZE_CLASSES is (4096 - 8) / 16 + 1 = 256 so 8 + 255 * 16 = 4088, which exceeds the max.

Here, 4088 is less than MAX(4096).

ZS_SIZE_CLASSES = (MAX - MIN) / Delta + 1
So, I think the value of
MIN + (ZS_SIZE_CLASSES - 1) * Delta =
MIN + ((MAX - MIN) / Delta) * Delta =
MAX
will not exceed the MAX

Thanks.

Quoted 23 lines

Signed-off-by Mahendran Ganesh <[email protected]>

---
 mm/zsmalloc.c |    2 --
 1 file changed, 2 deletions(-)

mm/zsmalloc.c +0 -3

index b3b57ef..f2279e2 100644
@@ -973,8 +973,6 @@ struct zs_pool *zs_create_pool(gfp_t flags)
              struct size_class *prev_class;
 
              size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA;
-             if (size > ZS_MAX_ALLOC_SIZE)
-                     size = ZS_MAX_ALLOC_SIZE;
              pages_per_zspage = get_pages_per_zspage(size);
 
              /*
--

1.7.9.5

Quoted 9 lines

Quoted 5 lines
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
--
Kind regards,
Minchan Kim

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Minchan Kim

maintainer Acked-by

2014-11-21 06:48 UTC #3 lore raw

Quoted 17 lines · Ganesh Mahendran

Hello
2014-11-21 11:54 GMT+08:00 Minchan Kim <[email protected]>:
Quoted 8 lines · Mahendran Ganesh
Quoted 7 lines
ZS_SIZE_CLASSES is calc by:
((ZS_MAX_ALLOC_SIZE - ZS_MIN_ALLOC_SIZE) / ZS_SIZE_CLASS_DELTA + 1)
So when i is in [0, ZS_SIZE_CLASSES - 1), the size:
size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA
will not be greater than ZS_MAX_ALLOC_SIZE
Quoted 6 lines
Quoted 1 line
This patch removes the unnecessary check.
It depends on ZS_MIN_ALLOC_SIZE.
For example, we would change min to 8 but MAX is still 4096. ZS_SIZE_CLASSES is (4096 - 8) / 16 + 1 = 256 so 8 + 255 * 16 = 4088, which exceeds the max.

Quoted 8 lines

Here, 4088 is less than MAX(4096).
ZS_SIZE_CLASSES = (MAX - MIN) / Delta + 1
So, I think the value of
MIN + (ZS_SIZE_CLASSES - 1) * Delta =
MIN + ((MAX - MIN) / Delta) * Delta =
MAX
will not exceed the MAX

You're right. It was complext math for me.
I should go back to elementary school.

Thanks!

Acked-by Minchan Kim <[email protected]>

Minchan Kim

maintainer

2014-11-21 10:32 UTC #4 lore raw

Quoted 27 lines · Minchan Kim

Quoted 17 lines · Ganesh Mahendran
Hello
2014-11-21 11:54 GMT+08:00 Minchan Kim <[email protected]>:
Quoted 8 lines · Mahendran Ganesh
Quoted 7 lines
ZS_SIZE_CLASSES is calc by:
((ZS_MAX_ALLOC_SIZE - ZS_MIN_ALLOC_SIZE) / ZS_SIZE_CLASS_DELTA + 1)
So when i is in [0, ZS_SIZE_CLASSES - 1), the size:
size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA
will not be greater than ZS_MAX_ALLOC_SIZE
Quoted 6 lines
Quoted 1 line
This patch removes the unnecessary check.
It depends on ZS_MIN_ALLOC_SIZE.
For example, we would change min to 8 but MAX is still 4096. ZS_SIZE_CLASSES is (4096 - 8) / 16 + 1 = 256 so 8 + 255 * 16 = 4088, which exceeds the max.
Quoted 8 lines
Here, 4088 is less than MAX(4096).
ZS_SIZE_CLASSES = (MAX - MIN) / Delta + 1
So, I think the value of
MIN + (ZS_SIZE_CLASSES - 1) * Delta =
MIN + ((MAX - MIN) / Delta) * Delta =
MAX
will not exceed the MAX

Quoted 6 lines

You're right. It was complext math for me.
I should go back to elementary school.
Thanks!
Acked-by Minchan Kim <[email protected]>

I catch a nasty cold but above my poor math makes me think more. ZS_SIZE_CLASSES is broken. In above my example, current code cannot allocate 4096 size class so we should correct ZS_SIZE_CLASSES at first.

zs_size_classes = zs_max - zs_min / delta + 1;
if ((zs_max - zs_min) % delta)

	zs_size_classes += 1;

Then, we need to code piece you removed.
As well, we need to fix below.

- area->vm_buf = (char *)__get_free_page(GFP_KERNEL);
+ area->vm_buf = kmalloc(ZS_MAX_ALLOC_SIZE);

Hope I am sane in this time :(

Ganesh Mahendran

author

2014-11-21 14:56 UTC #5 lore raw

Hello Minchan

2014-11-21 18:32 GMT+08:00 Minchan Kim <[email protected]>:

Quoted 35 lines · Minchan Kim

Quoted 27 lines · Ganesh Mahendran
Quoted 17 lines
Hello
2014-11-21 11:54 GMT+08:00 Minchan Kim <[email protected]>:
Quoted 8 lines · Mahendran Ganesh
Quoted 7 lines
ZS_SIZE_CLASSES is calc by:
((ZS_MAX_ALLOC_SIZE - ZS_MIN_ALLOC_SIZE) / ZS_SIZE_CLASS_DELTA + 1)
So when i is in [0, ZS_SIZE_CLASSES - 1), the size:
size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA
will not be greater than ZS_MAX_ALLOC_SIZE
Quoted 6 lines
Quoted 1 line
This patch removes the unnecessary check.
It depends on ZS_MIN_ALLOC_SIZE.
For example, we would change min to 8 but MAX is still 4096. ZS_SIZE_CLASSES is (4096 - 8) / 16 + 1 = 256 so 8 + 255 * 16 = 4088, which exceeds the max.
Quoted 8 lines
Here, 4088 is less than MAX(4096).
ZS_SIZE_CLASSES = (MAX - MIN) / Delta + 1
So, I think the value of
MIN + (ZS_SIZE_CLASSES - 1) * Delta =
MIN + ((MAX - MIN) / Delta) * Delta =
MAX
will not exceed the MAX
Quoted 6 lines
You're right. It was complext math for me.
I should go back to elementary school.
Thanks!
Acked-by Minchan Kim <[email protected]>

Quoted 8 lines

I catch a nasty cold but above my poor math makes me think more. ZS_SIZE_CLASSES is broken. In above my example, current code cannot allocate 4096 size class so we should correct ZS_SIZE_CLASSES at first.
zs_size_classes = zs_max - zs_min / delta + 1;
if ((zs_max - zs_min) % delta)
        zs_size_classes += 1;

Yes, you are right.
When the zs_min is less than delta, we can not allocate PAGE_SIZE size class.

Quoted 6 lines

Then, we need to code piece you removed.
As well, we need to fix below.
- area->vm_buf = (char *)__get_free_page(GFP_KERNEL);
+ area->vm_buf = kmalloc(ZS_MAX_ALLOC_SIZE);

If our purpose is to allocate the max obj size as len of PAGE_SIZE, we do not need to
change this line. Since the ZS_MAX_ALLOC_SIZE will always be PAGE_SIZE

Thanks.

Quoted 2 lines

Hope I am sane in this time :(

Minchan Kim

maintainer

2014-11-24 07:55 UTC #6 lore raw

Hello Ganesh,

Quoted 50 lines · Ganesh Mahendran

Hello Minchan
2014-11-21 18:32 GMT+08:00 Minchan Kim <[email protected]>:
Quoted 35 lines · Minchan Kim
Quoted 27 lines · Ganesh Mahendran
Quoted 17 lines
Hello
2014-11-21 11:54 GMT+08:00 Minchan Kim <[email protected]>:
Quoted 8 lines · Mahendran Ganesh
Quoted 7 lines
ZS_SIZE_CLASSES is calc by:
((ZS_MAX_ALLOC_SIZE - ZS_MIN_ALLOC_SIZE) / ZS_SIZE_CLASS_DELTA + 1)
So when i is in [0, ZS_SIZE_CLASSES - 1), the size:
size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA
will not be greater than ZS_MAX_ALLOC_SIZE
Quoted 6 lines
Quoted 1 line
This patch removes the unnecessary check.
It depends on ZS_MIN_ALLOC_SIZE.
For example, we would change min to 8 but MAX is still 4096. ZS_SIZE_CLASSES is (4096 - 8) / 16 + 1 = 256 so 8 + 255 * 16 = 4088, which exceeds the max.
Quoted 8 lines
Here, 4088 is less than MAX(4096).
ZS_SIZE_CLASSES = (MAX - MIN) / Delta + 1
So, I think the value of
MIN + (ZS_SIZE_CLASSES - 1) * Delta =
MIN + ((MAX - MIN) / Delta) * Delta =
MAX
will not exceed the MAX
Quoted 6 lines
You're right. It was complext math for me.
I should go back to elementary school.
Thanks!
Acked-by Minchan Kim <[email protected]>
Quoted 8 lines
I catch a nasty cold but above my poor math makes me think more. ZS_SIZE_CLASSES is broken. In above my example, current code cannot allocate 4096 size class so we should correct ZS_SIZE_CLASSES at first.
zs_size_classes = zs_max - zs_min / delta + 1;
if ((zs_max - zs_min) % delta)
        zs_size_classes += 1;
Yes, you are right.
When the zs_min is less than delta, we can not allocate PAGE_SIZE size class.
Quoted 1 line

Quoted 8 lines

Quoted 5 lines
Then, we need to code piece you removed.
As well, we need to fix below.
- area->vm_buf = (char *)__get_free_page(GFP_KERNEL);
+ area->vm_buf = kmalloc(ZS_MAX_ALLOC_SIZE);
If our purpose is to allocate the max obj size as len of PAGE_SIZE, we do not need to
change this line. Since the ZS_MAX_ALLOC_SIZE will always be PAGE_SIZE

No, please don't assume ZS_MAX_ALLOC_SIZE is PAGE_SIZE forever. Some reason could make buffer larger than PAGE_SIZE.
For example, we might put allocator's metadata into each object's head/tail.

Signature · 8 lines

Kind regards,
Minchan Kim

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>