[Patch] New zone ZONE_EASY_RECLAIM take 3. (change build_zonelists)[3/5]

[Patch] New zone ZONE_EASY_RECLAIM take 3. (change build_zonelists)[3/5]

[Yasunori Goto]

This is changing build_zonelists for new zone.

__GFP_xxxs are flag for requires of page allocation which zone
is prefered. But, it is used as an index number for zonelists[] too.
But after my patch, __GFP_xxx might be set at same time. So,
last set bit number of __GFP is recognized for zonelists' index
by this patch.

Note:
 This patch is modified take 3 to avoid panic on i386.
 __GFP_DMA32 is 0 for i386. So, ZONE_DMA32 is selected 
 if zone_bits is 0 which means Zone_normal. 
 Zone_DMA32 is not allocated on i386, so kernel paniced 
 by no normal memory.
 In this patch, even if zone_bits is 0 adn __GFP_DMA32 is 0,
 Zone_Normal is selected.


Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>

Index: zone_reclaim/mm/page_alloc.c
===================================================================
--- zone_reclaim.orig/mm/page_alloc.c	2005-12-06 14:11:20.000000000 +0900
+++ zone_reclaim/mm/page_alloc.c	2005-12-06 15:41:50.000000000 +0900
@@ -1574,6 +1574,10 @@ static int __init build_zonelists_node(p
 		struct zone *zone;
 	default:
 		BUG();
+	case ZONE_EASY_RECLAIM:
+		zone = pgdat->node_zones + ZONE_EASY_RECLAIM;
+		if (zone->present_pages)
+			zonelist->zones[j++] = zone;
 	case ZONE_HIGHMEM:
 		zone = pgdat->node_zones + ZONE_HIGHMEM;
 		if (populated_zone(zone)) {
@@ -1602,12 +1606,16 @@ static int __init build_zonelists_node(p
 static inline int highest_zone(int zone_bits)
 {
 	int res = ZONE_NORMAL;
-	if (zone_bits & (__force int)__GFP_HIGHMEM)
-		res = ZONE_HIGHMEM;
-	if (zone_bits & (__force int)__GFP_DMA32)
-		res = ZONE_DMA32;
+
 	if (zone_bits & (__force int)__GFP_DMA)
 		res = ZONE_DMA;
+	if (zone_bits & (__force int)__GFP_DMA32)
+		res = ZONE_DMA32;
+	if (zone_bits & (__force int)__GFP_HIGHMEM)
+		res = ZONE_HIGHMEM;
+	if (zone_bits & (__force int)__GFP_EASY_RECLAIM)
+		res = ZONE_EASY_RECLAIM;
+
 	return res;
 }
 
Index: zone_reclaim/include/linux/gfp.h
===================================================================
--- zone_reclaim.orig/include/linux/gfp.h	2005-12-06 14:12:43.000000000 +0900
+++ zone_reclaim/include/linux/gfp.h	2005-12-06 14:12:44.000000000 +0900
@@ -80,7 +80,7 @@ struct vm_area_struct;
 
 static inline int gfp_zone(gfp_t gfp)
 {
-	int zone = GFP_ZONEMASK & (__force int) gfp;
+	int zone = fls(GFP_ZONEMASK & (__force int) gfp);
 	BUG_ON(zone >= GFP_ZONETYPES);
 	return zone;
 }

-- 
Yasunori Goto 


--
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] New zone ZONE_EASY_RECLAIM take 3. (change build_zonelists)[3/5]

[Joel Schopp]

> @@ -1602,12 +1606,16 @@ static int __init build_zonelists_node(p
>  static inline int highest_zone(int zone_bits)
>  {
>  	int res = ZONE_NORMAL;
> -	if (zone_bits & (__force int)__GFP_HIGHMEM)
> -		res = ZONE_HIGHMEM;
> -	if (zone_bits & (__force int)__GFP_DMA32)
> -		res = ZONE_DMA32;
> +
>  	if (zone_bits & (__force int)__GFP_DMA)
>  		res = ZONE_DMA;
> +	if (zone_bits & (__force int)__GFP_DMA32)
> +		res = ZONE_DMA32;
> +	if (zone_bits & (__force int)__GFP_HIGHMEM)
> +		res = ZONE_HIGHMEM;
> +	if (zone_bits & (__force int)__GFP_EASY_RECLAIM)
> +		res = ZONE_EASY_RECLAIM;
> +
>  	return res;
>  }
>  

These look to come in the wrong order here.  You want ZONE_EASY_RECLAIM to be 
the highest zone, but this puts HIGHMEM, DMA32, and DMA ahead of it.  It in fact 
seems to get the order exactly backward.

> Index: zone_reclaim/include/linux/gfp.h
> ===================================================================
> --- zone_reclaim.orig/include/linux/gfp.h	2005-12-06 14:12:43.000000000 +0900
> +++ zone_reclaim/include/linux/gfp.h	2005-12-06 14:12:44.000000000 +0900
> @@ -80,7 +80,7 @@ struct vm_area_struct;
>  
>  static inline int gfp_zone(gfp_t gfp)
>  {
> -	int zone = GFP_ZONEMASK & (__force int) gfp;
> +	int zone = fls(GFP_ZONEMASK & (__force int) gfp);
>  	BUG_ON(zone >= GFP_ZONETYPES);
>  	return zone;
>  }
> 

Does this have endian issues?  I'm not too familiar with it.

--
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] New zone ZONE_EASY_RECLAIM take 3. (change build_zonelists)[3/5]

[Yasunori Goto]

> > @@ -1602,12 +1606,16 @@ static int __init build_zonelists_node(p
> >  static inline int highest_zone(int zone_bits)
> >  {
> >  	int res = ZONE_NORMAL;
> > -	if (zone_bits & (__force int)__GFP_HIGHMEM)
> > -		res = ZONE_HIGHMEM;
> > -	if (zone_bits & (__force int)__GFP_DMA32)
> > -		res = ZONE_DMA32;
> > +
> >  	if (zone_bits & (__force int)__GFP_DMA)
> >  		res = ZONE_DMA;
> > +	if (zone_bits & (__force int)__GFP_DMA32)
> > +		res = ZONE_DMA32;
> > +	if (zone_bits & (__force int)__GFP_HIGHMEM)
> > +		res = ZONE_HIGHMEM;
> > +	if (zone_bits & (__force int)__GFP_EASY_RECLAIM)
> > +		res = ZONE_EASY_RECLAIM;
> > +
> >  	return res;
> >  }
> >  
> 
> These look to come in the wrong order here.  You want ZONE_EASY_RECLAIM to be 
> the highest zone, but this puts HIGHMEM, DMA32, and DMA ahead of it.  It in fact 
> seems to get the order exactly backward.

I think this is correct about it.

If these are used "else if", you are correct.
But, if zone_bits is __GFP_EASY_RECLAIM, then it fall through to
last line.

> 
> > Index: zone_reclaim/include/linux/gfp.h
> > ===================================================================
> > --- zone_reclaim.orig/include/linux/gfp.h	2005-12-06 14:12:43.000000000 +0900
> > +++ zone_reclaim/include/linux/gfp.h	2005-12-06 14:12:44.000000000 +0900
> > @@ -80,7 +80,7 @@ struct vm_area_struct;
> >  
> >  static inline int gfp_zone(gfp_t gfp)
> >  {
> > -	int zone = GFP_ZONEMASK & (__force int) gfp;
> > +	int zone = fls(GFP_ZONEMASK & (__force int) gfp);
> >  	BUG_ON(zone >= GFP_ZONETYPES);
> >  	return zone;
> >  }
> > 
> 
> Does this have endian issues?  I'm not too familiar with it.

I'm also not too familiar with it. But, if there is a difference.
I suppose generic_ffs() and generic_fls() is not generic. :-P

However, Kame-san tells me this patch is still wrong.
__GFP_EASY_RECLAIM is 0x04 on i386, so fls(__GFP_EASY_RECLAIM)
is 3. zone 3 is ZONE_HIGHMEM, not ZONE_EASY_RECLAIM.
This patch should be more considered. Ah.... 

Bye.

-- 
Yasunori Goto 


--
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>