[PATCH] hugetlb: restore interleaving of bootmem huge pages

[PATCH] hugetlb: restore interleaving of bootmem huge pages

[Lee Schermerhorn]

PATCH restore interleaving of bootmem huge pages

Against: 2.6.31-rc1-mmotm-090625-1549
atop the "hugetlb-balance-freeing-of-huge-pages-across-node" series

I noticed that alloc_bootmem_huge_page() will only advance to the
next node on failure to allocate a huge page, potentially filling 
nodes with huge-pages.  I asked about this on linux-mm and linux-numa,
cc'ing the usual huge page suspects.

Mel Gorman responded:

	I strongly suspect that the same node being used until allocation
	failure instead of round-robin is an oversight and not deliberate
	at all. It appears to be a side-effect of a fix made way back in
	commit 63b4613c3f0d4b724ba259dc6c201bb68b884e1a ["hugetlb: fix
	hugepage allocation with memoryless nodes"]. Prior to that patch
	it looked like allocations would always round-robin even when
	allocation was successful.

This patch--factored out of my "hugetlb mempolicy" series--moves the
advance of the hstate next node from which to allocate up before the
test for success of the attempted allocation.

Note that alloc_bootmem_huge_page() is only used for order > MAX_ORDER
huge pages.

I'll post a separate patch for mainline/stable, as the above mentioned
"balance freeing" series renamed the next node to alloc function.

Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>

 mm/hugetlb.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c
===================================================================
--- linux-2.6.31-rc1-mmotm-090625-1549.orig/mm/hugetlb.c	2009-07-13 09:05:22.000000000 -0400
+++ linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c	2009-07-13 09:06:22.000000000 -0400
@@ -1030,6 +1030,7 @@ int __weak alloc_bootmem_huge_page(struc
 				NODE_DATA(h->next_nid_to_alloc),
 				huge_page_size(h), huge_page_size(h), 0);
 
+		hstate_next_node_to_alloc(h);
 		if (addr) {
 			/*
 			 * Use the beginning of the huge page to store the
@@ -1039,7 +1040,6 @@ int __weak alloc_bootmem_huge_page(struc
 			m = addr;
 			goto found;
 		}
-		hstate_next_node_to_alloc(h);
 		nr_nodes--;
 	}
 	return 0;

[PATCH] hugetlb: restore interleaving of bootmem huge pages

[Lee Schermerhorn]

PATCH restore interleaving of bootmem huge pages

Against: 2.6.31-rc1-mmotm-090625-1549
atop the "hugetlb-balance-freeing-of-huge-pages-across-node" series

I noticed that alloc_bootmem_huge_page() will only advance to the
next node on failure to allocate a huge page, potentially filling 
nodes with huge-pages.  I asked about this on linux-mm and linux-numa,
cc'ing the usual huge page suspects.

Mel Gorman responded:

	I strongly suspect that the same node being used until allocation
	failure instead of round-robin is an oversight and not deliberate
	at all. It appears to be a side-effect of a fix made way back in
	commit 63b4613c3f0d4b724ba259dc6c201bb68b884e1a ["hugetlb: fix
	hugepage allocation with memoryless nodes"]. Prior to that patch
	it looked like allocations would always round-robin even when
	allocation was successful.

This patch--factored out of my "hugetlb mempolicy" series--moves the
advance of the hstate next node from which to allocate up before the
test for success of the attempted allocation.

Note that alloc_bootmem_huge_page() is only used for order > MAX_ORDER
huge pages.

I'll post a separate patch for mainline/stable, as the above mentioned
"balance freeing" series renamed the next node to alloc function.

Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>

 mm/hugetlb.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c
===================================================================
--- linux-2.6.31-rc1-mmotm-090625-1549.orig/mm/hugetlb.c	2009-07-13 09:05:22.000000000 -0400
+++ linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c	2009-07-13 09:06:22.000000000 -0400
@@ -1030,6 +1030,7 @@ int __weak alloc_bootmem_huge_page(struc
 				NODE_DATA(h->next_nid_to_alloc),
 				huge_page_size(h), huge_page_size(h), 0);
 
+		hstate_next_node_to_alloc(h);
 		if (addr) {
 			/*
 			 * Use the beginning of the huge page to store the
@@ -1039,7 +1040,6 @@ int __weak alloc_bootmem_huge_page(struc
 			m = addr;
 			goto found;
 		}
-		hstate_next_node_to_alloc(h);
 		nr_nodes--;
 	}
 	return 0;


--
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] hugetlb: restore interleaving of bootmem huge pages

[Mel Gorman]

On Thu, Jul 16, 2009 at 10:31:02AM -0400, Lee Schermerhorn wrote:
> PATCH restore interleaving of bootmem huge pages
> 
> Against: 2.6.31-rc1-mmotm-090625-1549
> atop the "hugetlb-balance-freeing-of-huge-pages-across-node" series
> 
> I noticed that alloc_bootmem_huge_page() will only advance to the
> next node on failure to allocate a huge page, potentially filling 
> nodes with huge-pages.  I asked about this on linux-mm and linux-numa,
> cc'ing the usual huge page suspects.
> 
> Mel Gorman responded:
> 
> 	I strongly suspect that the same node being used until allocation
> 	failure instead of round-robin is an oversight and not deliberate
> 	at all. It appears to be a side-effect of a fix made way back in
> 	commit 63b4613c3f0d4b724ba259dc6c201bb68b884e1a ["hugetlb: fix
> 	hugepage allocation with memoryless nodes"]. Prior to that patch
> 	it looked like allocations would always round-robin even when
> 	allocation was successful.
> 
> This patch--factored out of my "hugetlb mempolicy" series--moves the
> advance of the hstate next node from which to allocate up before the
> test for success of the attempted allocation.
> 
> Note that alloc_bootmem_huge_page() is only used for order > MAX_ORDER
> huge pages.
> 
> I'll post a separate patch for mainline/stable, as the above mentioned
> "balance freeing" series renamed the next node to alloc function.
> 
> Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>

Reviewed-by: Mel Gorman <mel@csn.ul.ie>

> 
>  mm/hugetlb.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> Index: linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c
> ===================================================================
> --- linux-2.6.31-rc1-mmotm-090625-1549.orig/mm/hugetlb.c	2009-07-13 09:05:22.000000000 -0400
> +++ linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c	2009-07-13 09:06:22.000000000 -0400
> @@ -1030,6 +1030,7 @@ int __weak alloc_bootmem_huge_page(struc
>  				NODE_DATA(h->next_nid_to_alloc),
>  				huge_page_size(h), huge_page_size(h), 0);
>  
> +		hstate_next_node_to_alloc(h);
>  		if (addr) {
>  			/*
>  			 * Use the beginning of the huge page to store the
> @@ -1039,7 +1040,6 @@ int __weak alloc_bootmem_huge_page(struc
>  			m = addr;
>  			goto found;
>  		}
> -		hstate_next_node_to_alloc(h);
>  		nr_nodes--;
>  	}
>  	return 0;
> 
> 

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab

Re: [PATCH] hugetlb: restore interleaving of bootmem huge pages

[Mel Gorman]

On Thu, Jul 16, 2009 at 10:31:02AM -0400, Lee Schermerhorn wrote:
> PATCH restore interleaving of bootmem huge pages
> 
> Against: 2.6.31-rc1-mmotm-090625-1549
> atop the "hugetlb-balance-freeing-of-huge-pages-across-node" series
> 
> I noticed that alloc_bootmem_huge_page() will only advance to the
> next node on failure to allocate a huge page, potentially filling 
> nodes with huge-pages.  I asked about this on linux-mm and linux-numa,
> cc'ing the usual huge page suspects.
> 
> Mel Gorman responded:
> 
> 	I strongly suspect that the same node being used until allocation
> 	failure instead of round-robin is an oversight and not deliberate
> 	at all. It appears to be a side-effect of a fix made way back in
> 	commit 63b4613c3f0d4b724ba259dc6c201bb68b884e1a ["hugetlb: fix
> 	hugepage allocation with memoryless nodes"]. Prior to that patch
> 	it looked like allocations would always round-robin even when
> 	allocation was successful.
> 
> This patch--factored out of my "hugetlb mempolicy" series--moves the
> advance of the hstate next node from which to allocate up before the
> test for success of the attempted allocation.
> 
> Note that alloc_bootmem_huge_page() is only used for order > MAX_ORDER
> huge pages.
> 
> I'll post a separate patch for mainline/stable, as the above mentioned
> "balance freeing" series renamed the next node to alloc function.
> 
> Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>

Reviewed-by: Mel Gorman <mel@csn.ul.ie>

> 
>  mm/hugetlb.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> Index: linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c
> ===================================================================
> --- linux-2.6.31-rc1-mmotm-090625-1549.orig/mm/hugetlb.c	2009-07-13 09:05:22.000000000 -0400
> +++ linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c	2009-07-13 09:06:22.000000000 -0400
> @@ -1030,6 +1030,7 @@ int __weak alloc_bootmem_huge_page(struc
>  				NODE_DATA(h->next_nid_to_alloc),
>  				huge_page_size(h), huge_page_size(h), 0);
>  
> +		hstate_next_node_to_alloc(h);
>  		if (addr) {
>  			/*
>  			 * Use the beginning of the huge page to store the
> @@ -1039,7 +1040,6 @@ int __weak alloc_bootmem_huge_page(struc
>  			m = addr;
>  			goto found;
>  		}
> -		hstate_next_node_to_alloc(h);
>  		nr_nodes--;
>  	}
>  	return 0;
> 
> 

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab

--
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] hugetlb: restore interleaving of bootmem huge pages

[Andy Whitcroft]

On Thu, Jul 16, 2009 at 10:31:02AM -0400, Lee Schermerhorn wrote:
> PATCH restore interleaving of bootmem huge pages
> 
> Against: 2.6.31-rc1-mmotm-090625-1549
> atop the "hugetlb-balance-freeing-of-huge-pages-across-node" series
> 
> I noticed that alloc_bootmem_huge_page() will only advance to the
> next node on failure to allocate a huge page, potentially filling 
> nodes with huge-pages.  I asked about this on linux-mm and linux-numa,
> cc'ing the usual huge page suspects.
> 
> Mel Gorman responded:
> 
> 	I strongly suspect that the same node being used until allocation
> 	failure instead of round-robin is an oversight and not deliberate
> 	at all. It appears to be a side-effect of a fix made way back in
> 	commit 63b4613c3f0d4b724ba259dc6c201bb68b884e1a ["hugetlb: fix
> 	hugepage allocation with memoryless nodes"]. Prior to that patch
> 	it looked like allocations would always round-robin even when
> 	allocation was successful.
> 
> This patch--factored out of my "hugetlb mempolicy" series--moves the
> advance of the hstate next node from which to allocate up before the
> test for success of the attempted allocation.
> 
> Note that alloc_bootmem_huge_page() is only used for order > MAX_ORDER
> huge pages.
> 
> I'll post a separate patch for mainline/stable, as the above mentioned
> "balance freeing" series renamed the next node to alloc function.
> 
> Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>

It looks like this behaviour was in the original implementation to my eye.
It does indeed seem to prefer taking all it can from one node before moving
on to the next.  Your change seems reasonable to my eye though it may be
worth asking Andi if it was intended.  The intent of this change seems
to bring the behaviour into line with that of alloc_fresh_huge_page()
used for orders less than MAX_ORDER.

Reviewed-by: Andy Whitcroft <apw@canonical.com>

-apw

> 
>  mm/hugetlb.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> Index: linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c
> ===================================================================
> --- linux-2.6.31-rc1-mmotm-090625-1549.orig/mm/hugetlb.c	2009-07-13 09:05:22.000000000 -0400
> +++ linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c	2009-07-13 09:06:22.000000000 -0400
> @@ -1030,6 +1030,7 @@ int __weak alloc_bootmem_huge_page(struc
>  				NODE_DATA(h->next_nid_to_alloc),
>  				huge_page_size(h), huge_page_size(h), 0);
>  
> +		hstate_next_node_to_alloc(h);
>  		if (addr) {
>  			/*
>  			 * Use the beginning of the huge page to store the
> @@ -1039,7 +1040,6 @@ int __weak alloc_bootmem_huge_page(struc
>  			m = addr;
>  			goto found;
>  		}
> -		hstate_next_node_to_alloc(h);
>  		nr_nodes--;
>  	}
>  	return 0;
> 
> 

Re: [PATCH] hugetlb: restore interleaving of bootmem huge pages

[Andy Whitcroft]

On Thu, Jul 16, 2009 at 10:31:02AM -0400, Lee Schermerhorn wrote:
> PATCH restore interleaving of bootmem huge pages
> 
> Against: 2.6.31-rc1-mmotm-090625-1549
> atop the "hugetlb-balance-freeing-of-huge-pages-across-node" series
> 
> I noticed that alloc_bootmem_huge_page() will only advance to the
> next node on failure to allocate a huge page, potentially filling 
> nodes with huge-pages.  I asked about this on linux-mm and linux-numa,
> cc'ing the usual huge page suspects.
> 
> Mel Gorman responded:
> 
> 	I strongly suspect that the same node being used until allocation
> 	failure instead of round-robin is an oversight and not deliberate
> 	at all. It appears to be a side-effect of a fix made way back in
> 	commit 63b4613c3f0d4b724ba259dc6c201bb68b884e1a ["hugetlb: fix
> 	hugepage allocation with memoryless nodes"]. Prior to that patch
> 	it looked like allocations would always round-robin even when
> 	allocation was successful.
> 
> This patch--factored out of my "hugetlb mempolicy" series--moves the
> advance of the hstate next node from which to allocate up before the
> test for success of the attempted allocation.
> 
> Note that alloc_bootmem_huge_page() is only used for order > MAX_ORDER
> huge pages.
> 
> I'll post a separate patch for mainline/stable, as the above mentioned
> "balance freeing" series renamed the next node to alloc function.
> 
> Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>

It looks like this behaviour was in the original implementation to my eye.
It does indeed seem to prefer taking all it can from one node before moving
on to the next.  Your change seems reasonable to my eye though it may be
worth asking Andi if it was intended.  The intent of this change seems
to bring the behaviour into line with that of alloc_fresh_huge_page()
used for orders less than MAX_ORDER.

Reviewed-by: Andy Whitcroft <apw@canonical.com>

-apw

> 
>  mm/hugetlb.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> Index: linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c
> ===================================================================
> --- linux-2.6.31-rc1-mmotm-090625-1549.orig/mm/hugetlb.c	2009-07-13 09:05:22.000000000 -0400
> +++ linux-2.6.31-rc1-mmotm-090625-1549/mm/hugetlb.c	2009-07-13 09:06:22.000000000 -0400
> @@ -1030,6 +1030,7 @@ int __weak alloc_bootmem_huge_page(struc
>  				NODE_DATA(h->next_nid_to_alloc),
>  				huge_page_size(h), huge_page_size(h), 0);
>  
> +		hstate_next_node_to_alloc(h);
>  		if (addr) {
>  			/*
>  			 * Use the beginning of the huge page to store the
> @@ -1039,7 +1040,6 @@ int __weak alloc_bootmem_huge_page(struc
>  			m = addr;
>  			goto found;
>  		}
> -		hstate_next_node_to_alloc(h);
>  		nr_nodes--;
>  	}
>  	return 0;
> 
> 

--
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] hugetlb: restore interleaving of bootmem huge pages

[Andi Kleen]

> It looks like this behaviour was in the original implementation to my eye.
> It does indeed seem to prefer taking all it can from one node before moving
> on to the next.  Your change seems reasonable to my eye though it may be
> worth asking Andi if it was intended.  The intent of this change seems
> to bring the behaviour into line with that of alloc_fresh_huge_page()
> used for orders less than MAX_ORDER.

I don't remember intending it this way. The intention was always
standard round robin one by one. If it didn't do that it wasn't extended.

-Andi

-- 
ak@linux.intel.com -- Speaking for myself only.

--
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>

[PATCH] hugetlb: restore interleaving of bootmem huge pages

[Lee Schermerhorn]

PATCH fix distribution of bootmem huge pages

Against: 2.6.31-rc3

For stable, applicable back through 2.6.27

[A version of this patch has been added to the -mm tree.  That patch
won't apply cleanly to mainline nor stable trees because of function
renaming.  This one should.]

I noticed that alloc_bootmem_huge_page() will only advance to the
next node on failure to allocate a huge page.  I asked about this
on linux-mm and linux-numa, cc'ing the usual huge page suspects.
Mel Gorman responded:

	I strongly suspect that the same node being used until allocation
	failure instead of round-robin is an oversight and not deliberate
	at all. It appears to be a side-effect of a fix made way back in
	commit 63b4613c3f0d4b724ba259dc6c201bb68b884e1a ["hugetlb: fix
	hugepage allocation with memoryless nodes"]. Prior to that patch
	it looked like allocations would always round-robin even when
	allocation was successful.

Andy Whitcroft countered that the existing behavior looked like Andi
Kleen's original implementation and suggested that we ask him.  We
did and Andy replied that his intention was to interleave the
allocations.  So, ...

This patch moves the advance of the hstate next node from which to
allocate up before the test for success of the attempted allocation.
This will unconditionally advance the next node from which to alloc,
interleaving successful allocations over the nodes with sufficient
contiguous memory, and skipping over nodes that fail the huge page
allocation attempt.

Note that alloc_bootmem_huge_page() will only be called for huge pages
of order > MAX_ORDER.

Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>

 mm/hugetlb.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: linux-2.6.31-rc3-mmotm-090715-2057/mm/hugetlb.c
===================================================================
--- linux-2.6.31-rc3-mmotm-090715-2057.orig/mm/hugetlb.c	2009-07-17 11:30:20.000000000 -0400
+++ linux-2.6.31-rc3-mmotm-090715-2057/mm/hugetlb.c	2009-07-17 11:37:30.000000000 -0400
@@ -1010,6 +1010,7 @@ int __weak alloc_bootmem_huge_page(struc
 				NODE_DATA(h->hugetlb_next_nid),
 				huge_page_size(h), huge_page_size(h), 0);
 
+		hstate_next_node(h);
 		if (addr) {
 			/*
 			 * Use the beginning of the huge page to store the
@@ -1019,7 +1020,6 @@ int __weak alloc_bootmem_huge_page(struc
 			m = addr;
 			goto found;
 		}
-		hstate_next_node(h);
 		nr_nodes--;
 	}
 	return 0;

Re: [PATCH] hugetlb: restore interleaving of bootmem huge pages

[Andi Kleen]

> Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>

Reviewed-by: Andi Kleen <ak@linux.intel.com>

-Andi

-- 
ak@linux.intel.com -- Speaking for myself only.