[PATCH 03/31] mm: expose gfp_to_alloc_flags()

[PATCH 03/31] mm: expose gfp_to_alloc_flags()

[Suresh Jayaraman]

From: Peter Zijlstra <a.p.zijlstra@chello.nl> 

Expose the gfp to alloc_flags mapping, so we can use it in other parts
of the vm.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
---
 mm/internal.h   |   15 +++++++++++++++
 mm/page_alloc.c |   16 +---------------
 2 files changed, 16 insertions(+), 15 deletions(-)

Index: mmotm/mm/internal.h
===================================================================
--- mmotm.orig/mm/internal.h
+++ mmotm/mm/internal.h
@@ -194,6 +194,21 @@ static inline struct page *mem_map_next(
 #define __paginginit __init
 #endif
 
+/* The ALLOC_WMARK bits are used as an index to zone->watermark */
+#define ALLOC_WMARK_MIN		WMARK_MIN
+#define ALLOC_WMARK_LOW		WMARK_LOW
+#define ALLOC_WMARK_HIGH	WMARK_HIGH
+#define ALLOC_NO_WATERMARKS	0x04 /* don't check watermarks at all */
+
+/* Mask to get the watermark bits */
+#define ALLOC_WMARK_MASK	(ALLOC_NO_WATERMARKS-1)
+
+#define ALLOC_HARDER		0x10 /* try to alloc harder */
+#define ALLOC_HIGH		0x20 /* __GFP_HIGH set */
+#define ALLOC_CPUSET		0x40 /* check for correct cpuset */
+
+int gfp_to_alloc_flags(gfp_t gfp_mask);
+
 /* Memory initialisation debug and verification */
 enum mminit_level {
 	MMINIT_WARNING,
Index: mmotm/mm/page_alloc.c
===================================================================
--- mmotm.orig/mm/page_alloc.c
+++ mmotm/mm/page_alloc.c
@@ -1190,19 +1190,6 @@ failed:
 	return NULL;
 }
 
-/* The ALLOC_WMARK bits are used as an index to zone->watermark */
-#define ALLOC_WMARK_MIN		WMARK_MIN
-#define ALLOC_WMARK_LOW		WMARK_LOW
-#define ALLOC_WMARK_HIGH	WMARK_HIGH
-#define ALLOC_NO_WATERMARKS	0x04 /* don't check watermarks at all */
-
-/* Mask to get the watermark bits */
-#define ALLOC_WMARK_MASK	(ALLOC_NO_WATERMARKS-1)
-
-#define ALLOC_HARDER		0x10 /* try to alloc harder */
-#define ALLOC_HIGH		0x20 /* __GFP_HIGH set */
-#define ALLOC_CPUSET		0x40 /* check for correct cpuset */
-
 #ifdef CONFIG_FAIL_PAGE_ALLOC
 
 static struct fail_page_alloc_attr {
@@ -1691,8 +1678,7 @@ void wake_all_kswapd(unsigned int order,
 		wakeup_kswapd(zone, order);
 }
 
-static inline int
-gfp_to_alloc_flags(gfp_t gfp_mask)
+int gfp_to_alloc_flags(gfp_t gfp_mask)
 {
 	struct task_struct *p = current;
 	int alloc_flags = ALLOC_WMARK_MIN | ALLOC_CPUSET;

--
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 03/31] mm: expose gfp_to_alloc_flags()

[David Rientjes]

On Thu, 1 Oct 2009, Suresh Jayaraman wrote:

> From: Peter Zijlstra <a.p.zijlstra@chello.nl> 
> 
> Expose the gfp to alloc_flags mapping, so we can use it in other parts
> of the vm.
> 
> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>

Nack, these flags are internal to the page allocator and exporting them to 
generic VM code is unnecessary.

The only bit you actually use in your patchset is ALLOC_NO_WATERMARKS to 
determine whether a particular allocation can use memory reserves.  I'd 
suggest adding a bool function that returns whether the current context is 
given access to reserves including your new __GFP_MEMALLOC flag and 
exporting that instead.

--
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 03/31] mm: expose gfp_to_alloc_flags()

[Neil Brown]

On Thursday October 1, rientjes@google.com wrote:
> On Thu, 1 Oct 2009, Suresh Jayaraman wrote:
> 
> > From: Peter Zijlstra <a.p.zijlstra@chello.nl> 
> > 
> > Expose the gfp to alloc_flags mapping, so we can use it in other parts
> > of the vm.
> > 
> > Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
> > Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
> 
> Nack, these flags are internal to the page allocator and exporting them to 
> generic VM code is unnecessary.
> 
> The only bit you actually use in your patchset is ALLOC_NO_WATERMARKS to 
> determine whether a particular allocation can use memory reserves.  I'd 
> suggest adding a bool function that returns whether the current context is 
> given access to reserves including your new __GFP_MEMALLOC flag and 
> exporting that instead.

That sounds like a very appropriate suggestion, thanks.

So something like this?
Then change every occurrence of
+		if (!(gfp_to_alloc_flags(gfpflags) & ALLOC_NO_WATERMARKS))
to
+		if (!(gfp_has_no_watermarks(gfpflags)))

??

Thanks,
NeilBrown



diff --git a/mm/internal.h b/mm/internal.h
index 22ec8d2..7ff78d6 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -195,6 +195,8 @@ static inline struct page *mem_map_next(struct page *iter,
 #define __paginginit __init
 #endif
 
+int gfp_has_no_watermarks(gfp_t gfp_mask);
+
 /* Memory initialisation debug and verification */
 enum mminit_level {
 	MMINIT_WARNING,
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index bf72055..4b4292a 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1782,6 +1782,11 @@ gfp_to_alloc_flags(gfp_t gfp_mask)
 	return alloc_flags;
 }
 
+int gfp_has_no_watermarks(gfp_t gfp_mask)
+{
+	return (gfp_to_alloc_flags(gfp_mask) & ALLOC_NO_WATERMARKS);
+}
+
 static inline struct page *
 __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
 	struct zonelist *zonelist, enum zone_type high_zoneidx,

--
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 03/31] mm: expose gfp_to_alloc_flags()

[Suresh Jayaraman]

David Rientjes wrote:
> On Thu, 1 Oct 2009, Suresh Jayaraman wrote:
> 
>> From: Peter Zijlstra <a.p.zijlstra@chello.nl> 
>>
>> Expose the gfp to alloc_flags mapping, so we can use it in other parts
>> of the vm.
>>
>> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
>> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
> 
> Nack, these flags are internal to the page allocator and exporting them to 
> generic VM code is unnecessary.

Yes, you're right.

> The only bit you actually use in your patchset is ALLOC_NO_WATERMARKS to 
> determine whether a particular allocation can use memory reserves.  I'd 
> suggest adding a bool function that returns whether the current context is 
> given access to reserves including your new __GFP_MEMALLOC flag and 
> exporting that instead.

Makes sense and Neil already posted a patch citing the suggested
changes, will incorporate the change.

Thanks,

-- 
Suresh Jayaraman

--
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 03/31] mm: expose gfp_to_alloc_flags()

[David Rientjes]

On Fri, 2 Oct 2009, Neil Brown wrote:

> So something like this?
> Then change every occurrence of
> +		if (!(gfp_to_alloc_flags(gfpflags) & ALLOC_NO_WATERMARKS))
> to
> +		if (!(gfp_has_no_watermarks(gfpflags)))
> 
> ??
> 

No, it's not even necessary to call gfp_to_alloc_flags() at all, just 
create a globally exported function such as can_alloc_use_reserves() and 
use it in gfp_to_alloc_flags().

 [ Using 'p' in gfp_to_alloc_flags() is actually wrong since
   test_thread_flag() only works on current anyway, so it would be
   inconsistent if p were set to anything other than current; we can
   get rid of that auto variable. ]

Something like the following, which you can fold into this patch proposal 
and modify later for GFP_MEMALLOC.

Signed-off-by: David Rientjes <rientjes@google.com>
---
diff --git a/include/linux/gfp.h b/include/linux/gfp.h
index 557bdad..7dd62a0 100644
--- a/include/linux/gfp.h
+++ b/include/linux/gfp.h
@@ -265,6 +265,8 @@ static inline void arch_free_page(struct page *page, int order) { }
 static inline void arch_alloc_page(struct page *page, int order) { }
 #endif
 
+int can_alloc_use_reserves(void);
+
 struct page *
 __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order,
 		       struct zonelist *zonelist, nodemask_t *nodemask);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index bf72055..cf1d765 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1744,10 +1744,19 @@ void wake_all_kswapd(unsigned int order, struct zonelist *zonelist,
 		wakeup_kswapd(zone, order);
 }
 
+/*
+ * Does the current context allow the allocation to utilize memory reserves
+ * by ignoring watermarks for all zones?
+ */
+int can_alloc_use_reserves(void)
+{
+	return !in_interrupt() && ((current->flags & PF_MEMALLOC) ||
+				   unlikely(test_thread_flag(TIF_MEMDIE)));
+}
+
 static inline int
 gfp_to_alloc_flags(gfp_t gfp_mask)
 {
-	struct task_struct *p = current;
 	int alloc_flags = ALLOC_WMARK_MIN | ALLOC_CPUSET;
 	const gfp_t wait = gfp_mask & __GFP_WAIT;
 
@@ -1769,15 +1778,12 @@ gfp_to_alloc_flags(gfp_t gfp_mask)
 		 * See also cpuset_zone_allowed() comment in kernel/cpuset.c.
 		 */
 		alloc_flags &= ~ALLOC_CPUSET;
-	} else if (unlikely(rt_task(p)))
+	} else if (unlikely(rt_task(current)))
 		alloc_flags |= ALLOC_HARDER;
 
-	if (likely(!(gfp_mask & __GFP_NOMEMALLOC))) {
-		if (!in_interrupt() &&
-		    ((p->flags & PF_MEMALLOC) ||
-		     unlikely(test_thread_flag(TIF_MEMDIE))))
+	if (likely(!(gfp_mask & __GFP_NOMEMALLOC)))
+		if (can_alloc_use_reserves())
 			alloc_flags |= ALLOC_NO_WATERMARKS;
-	}
 
 	return alloc_flags;
 }

--
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>