[PATCH] xvmalloc: total_pages can be ulong

[PATCH] xvmalloc: total_pages can be ulong

[Ajeet Yadav]

>From 4763266cb0077d1082d4a1ef941448e9b4eb55ca Mon Sep 17 00:00:00 2001
From: Ajeet Yadav <ajeet.yadav.77@gmail.com>
Date: Wed, 28 Dec 2011 11:08:34 +0530
Subject: [PATCH] xvmalloc: total_pages can be ulong

total_pages is u64, its too large value considering
that its in page unit on both 32-bit and 64-bit system
, therefore make it ulong.

Having done that we can use normal ++ and -- operators
instead of stat_inc() and stat_dec(), therefore
remove these functions as well.

Signed-off-by: Ajeet Yadav <ajeet.yadav.77@gmail.com>
---
 drivers/staging/zram/xvmalloc.c     |   16 +++-------------
 drivers/staging/zram/xvmalloc_int.h |    2 +-
 2 files changed, 4 insertions(+), 14 deletions(-)

diff --git a/drivers/staging/zram/xvmalloc.c b/drivers/staging/zram/xvmalloc.c
index 1f9c508..2f989cf 100644
--- a/drivers/staging/zram/xvmalloc.c
+++ b/drivers/staging/zram/xvmalloc.c
@@ -26,16 +26,6 @@
 #include "xvmalloc.h"
 #include "xvmalloc_int.h"

-static void stat_inc(u64 *value)
-{
-	*value = *value + 1;
-}
-
-static void stat_dec(u64 *value)
-{
-	*value = *value - 1;
-}
-
 static int test_flag(struct block_header *block, enum blockflags flag)
 {
 	return block->prev & BIT(flag);
@@ -281,7 +271,7 @@ static int grow_pool(struct xv_pool *pool, gfp_t flags)
 	if (unlikely(!page))
 		return -ENOMEM;

-	stat_inc(&pool->total_pages);
+	pool->total_pages++;

 	spin_lock(&pool->lock);
 	block = get_ptr_atomic(page, 0, KM_USER0);
@@ -472,7 +462,7 @@ void xv_free(struct xv_pool *pool, struct page
*page, u32 offset)
 		spin_unlock(&pool->lock);

 		__free_page(page);
-		stat_dec(&pool->total_pages);
+		pool->total_pages--;
 		return;
 	}

@@ -505,6 +495,6 @@ EXPORT_SYMBOL_GPL(xv_get_object_size);
  */
 u64 xv_get_total_size_bytes(struct xv_pool *pool)
 {
-	return pool->total_pages << PAGE_SHIFT;
+	return (u64)pool->total_pages << PAGE_SHIFT;
 }
 EXPORT_SYMBOL_GPL(xv_get_total_size_bytes);
diff --git a/drivers/staging/zram/xvmalloc_int.h
b/drivers/staging/zram/xvmalloc_int.h
index b5f1f7f..fcb0868 100644
--- a/drivers/staging/zram/xvmalloc_int.h
+++ b/drivers/staging/zram/xvmalloc_int.h
@@ -87,7 +87,7 @@ struct block_header {
 struct xv_pool {
 	ulong flbitmap;
 	ulong slbitmap[MAX_FLI];
-	u64 total_pages;	/* stats */
+	ulong total_pages;	/* stats */
 	struct freelist_entry freelist[NUM_FREE_LISTS];
 	spinlock_t lock;
 };
-- 
1.6.0.3

Re: [PATCH] xvmalloc: total_pages can be ulong

[Nitin Gupta]

Hi Ajeet,

On 12/28/2011 12:50 AM, Ajeet Yadav wrote:

> From 4763266cb0077d1082d4a1ef941448e9b4eb55ca Mon Sep 17 00:00:00 2001
> From: Ajeet Yadav <ajeet.yadav.77@gmail.com>
> Date: Wed, 28 Dec 2011 11:08:34 +0530
> Subject: [PATCH] xvmalloc: total_pages can be ulong
> 
> total_pages is u64, its too large value considering
> that its in page unit on both 32-bit and 64-bit system
> , therefore make it ulong.
> 
> Having done that we can use normal ++ and -- operators
> instead of stat_inc() and stat_dec(), therefore
> remove these functions as well.
> 


I'm soon going to send patches which replace xvmalloc with a new allocator.

Thanks,
Nitin

Re: [PATCH] xvmalloc: total_pages can be ulong

[Ajeet Yadav]

Hi Nitin,

I wish if I could somehow know your plans for next release, so that I
can schedule my work accordingly, else my work on zram will be total
waste of my time making patches and your time replying and reviewing
it. Finally I have decided to delete my zram git.

Thanks and wishes for new year in advance.

On Thu, Dec 29, 2011 at 7:28 PM, Nitin Gupta <ngupta@vflare.org> wrote:
> Hi Ajeet,
>
> On 12/28/2011 12:50 AM, Ajeet Yadav wrote:
>
>> From 4763266cb0077d1082d4a1ef941448e9b4eb55ca Mon Sep 17 00:00:00 2001
>> From: Ajeet Yadav <ajeet.yadav.77@gmail.com>
>> Date: Wed, 28 Dec 2011 11:08:34 +0530
>> Subject: [PATCH] xvmalloc: total_pages can be ulong
>>
>> total_pages is u64, its too large value considering
>> that its in page unit on both 32-bit and 64-bit system
>> , therefore make it ulong.
>>
>> Having done that we can use normal ++ and -- operators
>> instead of stat_inc() and stat_dec(), therefore
>> remove these functions as well.
>>
>
>
> I'm soon going to send patches which replace xvmalloc with a new allocator.
>
> Thanks,
> Nitin

Re: [PATCH] xvmalloc: total_pages can be ulong

[Greg KH]

On Wed, Dec 28, 2011 at 11:20:51AM +0530, Ajeet Yadav wrote:
> >From 4763266cb0077d1082d4a1ef941448e9b4eb55ca Mon Sep 17 00:00:00 2001
> From: Ajeet Yadav <ajeet.yadav.77@gmail.com>
> Date: Wed, 28 Dec 2011 11:08:34 +0530
> Subject: [PATCH] xvmalloc: total_pages can be ulong
> 
> total_pages is u64, its too large value considering
> that its in page unit on both 32-bit and 64-bit system
> , therefore make it ulong.

Your patch is line-wrapped and can not be applied.

Please use the git send-email command to send the patch, otherwise bad
things like this can happen :(

Care to redo it?

thanks,

greg k-h

Re: [PATCH] xvmalloc: total_pages can be ulong

[Ajeet Yadav]

sure, thanks for feedback

On Thu, Feb 9, 2012 at 6:41 AM, Greg KH <gregkh@linuxfoundation.org> wrote:
> On Wed, Dec 28, 2011 at 11:20:51AM +0530, Ajeet Yadav wrote:
>> >From 4763266cb0077d1082d4a1ef941448e9b4eb55ca Mon Sep 17 00:00:00 2001
>> From: Ajeet Yadav <ajeet.yadav.77@gmail.com>
>> Date: Wed, 28 Dec 2011 11:08:34 +0530
>> Subject: [PATCH] xvmalloc: total_pages can be ulong
>>
>> total_pages is u64, its too large value considering
>> that its in page unit on both 32-bit and 64-bit system
>> , therefore make it ulong.
>
> Your patch is line-wrapped and can not be applied.
>
> Please use the git send-email command to send the patch, otherwise bad
> things like this can happen :(
>
> Care to redo it?
>
> thanks,
>
> greg k-h

Re: [PATCH] xvmalloc: total_pages can be ulong

[Nitin Gupta]

On 2/8/12 8:11 PM, Greg KH wrote:
> On Wed, Dec 28, 2011 at 11:20:51AM +0530, Ajeet Yadav wrote:
>> > From 4763266cb0077d1082d4a1ef941448e9b4eb55ca Mon Sep 17 00:00:00 2001
>> From: Ajeet Yadav<ajeet.yadav.77@gmail.com>
>> Date: Wed, 28 Dec 2011 11:08:34 +0530
>> Subject: [PATCH] xvmalloc: total_pages can be ulong
>>
>> total_pages is u64, its too large value considering
>> that its in page unit on both 32-bit and 64-bit system
>> , therefore make it ulong.
>
> Your patch is line-wrapped and can not be applied.
>
> Please use the git send-email command to send the patch, otherwise bad
> things like this can happen :(
>
> Care to redo it?
>

Since xvmalloc has now been replaced with zsmalloc, this patch is no 
longer required.

Thanks,
Nitin

Re: [PATCH] xvmalloc: total_pages can be ulong

[Ajeet Yadav]

thanks

On Fri, Feb 10, 2012 at 7:28 PM, Nitin Gupta <ngupta@vflare.org> wrote:
> On 2/8/12 8:11 PM, Greg KH wrote:
>>
>> On Wed, Dec 28, 2011 at 11:20:51AM +0530, Ajeet Yadav wrote:
>>>
>>> > From 4763266cb0077d1082d4a1ef941448e9b4eb55ca Mon Sep 17 00:00:00 2001
>>> From: Ajeet Yadav<ajeet.yadav.77@gmail.com>
>>> Date: Wed, 28 Dec 2011 11:08:34 +0530
>>> Subject: [PATCH] xvmalloc: total_pages can be ulong
>>>
>>> total_pages is u64, its too large value considering
>>> that its in page unit on both 32-bit and 64-bit system
>>> , therefore make it ulong.
>>
>>
>> Your patch is line-wrapped and can not be applied.
>>
>> Please use the git send-email command to send the patch, otherwise bad
>> things like this can happen :(
>>
>> Care to redo it?
>>
>
> Since xvmalloc has now been replaced with zsmalloc, this patch is no longer
> required.
>
> Thanks,
> Nitin
>