[PATCH] slob:Use _safe funtion to iterate partially free list.

[PATCH] slob:Use _safe funtion to iterate partially free list.

[Bob Liu]

Since a list entry may be removed, so use list_for_each_entry_safe
instead of list_for_each_entry.

Signed-off-by: Bob Liu <lliubbo@gmail.com>
---
 mm/slob.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/mm/slob.c b/mm/slob.c
index 3f19a34..e2af18b 100644
--- a/mm/slob.c
+++ b/mm/slob.c
@@ -320,7 +320,7 @@ static void *slob_page_alloc(struct slob_page *sp, size_t size, int align)
  */
 static void *slob_alloc(size_t size, gfp_t gfp, int align, int node)
 {
-	struct slob_page *sp;
+	struct slob_page *sp, *tmp;
 	struct list_head *prev;
 	struct list_head *slob_list;
 	slob_t *b = NULL;
@@ -335,7 +335,7 @@ static void *slob_alloc(size_t size, gfp_t gfp, int align, int node)
 
 	spin_lock_irqsave(&slob_lock, flags);
 	/* Iterate through each partially free page, try to find room */
-	list_for_each_entry(sp, slob_list, list) {
+	list_for_each_entry_safe(sp, tmp, slob_list, list) {    
 #ifdef CONFIG_NUMA
 		/*
 		 * If there's a node specification, search for a partial
-- 
1.5.6.3

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

Re: [PATCH] slob:Use _safe funtion to iterate partially free list.

[Johannes Weiner]

On Sun, Jul 04, 2010 at 05:22:33PM +0800, Bob Liu wrote:
> Since a list entry may be removed, so use list_for_each_entry_safe
> instead of list_for_each_entry.
> 
> Signed-off-by: Bob Liu <lliubbo@gmail.com>
> ---
>  mm/slob.c |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/slob.c b/mm/slob.c
> index 3f19a34..e2af18b 100644
> --- a/mm/slob.c
> +++ b/mm/slob.c
> @@ -320,7 +320,7 @@ static void *slob_page_alloc(struct slob_page *sp, size_t size, int align)
>   */
>  static void *slob_alloc(size_t size, gfp_t gfp, int align, int node)
>  {
> -	struct slob_page *sp;
> +	struct slob_page *sp, *tmp;
>  	struct list_head *prev;
>  	struct list_head *slob_list;
>  	slob_t *b = NULL;
> @@ -335,7 +335,7 @@ static void *slob_alloc(size_t size, gfp_t gfp, int align, int node)
>  
>  	spin_lock_irqsave(&slob_lock, flags);
>  	/* Iterate through each partially free page, try to find room */
> -	list_for_each_entry(sp, slob_list, list) {
> +	list_for_each_entry_safe(sp, tmp, slob_list, list) {    
>  #ifdef CONFIG_NUMA

sp's list head is only modified if an allocation was successful, but
then the iteration stops as well.  So I see no reason for your patch.
Did I overlook something?

	Hannes

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

Re: [PATCH] slob:Use _safe funtion to iterate partially free list.

[Bob Liu]

On Sun, Jul 4, 2010 at 6:16 PM, Johannes Weiner <hannes@cmpxchg.org> wrote:
> On Sun, Jul 04, 2010 at 05:22:33PM +0800, Bob Liu wrote:
>> Since a list entry may be removed, so use list_for_each_entry_safe
>> instead of list_for_each_entry.
>>
>> Signed-off-by: Bob Liu <lliubbo@gmail.com>
>> ---
>>  mm/slob.c |    4 ++--
>>  1 files changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/mm/slob.c b/mm/slob.c
>> index 3f19a34..e2af18b 100644
>> --- a/mm/slob.c
>> +++ b/mm/slob.c
>> @@ -320,7 +320,7 @@ static void *slob_page_alloc(struct slob_page *sp, size_t size, int align)
>>   */
>>  static void *slob_alloc(size_t size, gfp_t gfp, int align, int node)
>>  {
>> -     struct slob_page *sp;
>> +     struct slob_page *sp, *tmp;
>>       struct list_head *prev;
>>       struct list_head *slob_list;
>>       slob_t *b = NULL;
>> @@ -335,7 +335,7 @@ static void *slob_alloc(size_t size, gfp_t gfp, int align, int node)
>>
>>       spin_lock_irqsave(&slob_lock, flags);
>>       /* Iterate through each partially free page, try to find room */
>> -     list_for_each_entry(sp, slob_list, list) {
>> +     list_for_each_entry_safe(sp, tmp, slob_list, list) {
>>  #ifdef CONFIG_NUMA
>
> sp's list head is only modified if an allocation was successful, but
> then the iteration stops as well.  So I see no reason for your patch.
> Did I overlook something?
>

Sorry, I am wrong. Please ignore this patch.
Thanks for your comment.

But It seems that the slob_list maybe have some member's next pointer NULL.
Because I triggered a NULL pointer access error.

And after I changed spin_lock_irqsave(&slob_lock, flags) before set slob_list.
This bug seems disappeared.

I will resend a patch, Please review.
Thanks!

-- 
Regards,
--Bob

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