[RFC] reduce usage of mem_map

[RFC] reduce usage of mem_map

[Martin J. Bligh]

I've tried to cut down the usage of mem_map somewhat.
There's already macros to do the conversion between
pfns to pages, and it doesn't work the way they've
embedded it for discontigmem systems. Comments?
Please don't apply - not tested yet ;-)

M.

diff -urN virgin-2.5.30/arch/i386/mm/init.c linux-2.5.30-pfn_to_page/arch/i386/mm/init.c
--- virgin-2.5.30/arch/i386/mm/init.c	Thu Aug  1 14:16:13 2002
+++ linux-2.5.30-pfn_to_page/arch/i386/mm/init.c	Thu Aug  1 22:39:17 2002
@@ -217,7 +217,7 @@
 {
 	int pfn;
 	for (pfn = highstart_pfn; pfn < highend_pfn; pfn++) {
-		struct page *page = mem_map + pfn;
+		struct page *page = pfn_to_page(pfn);
 
 		if (!page_is_ram(pfn)) {
 			SetPageReserved(page);
@@ -419,7 +419,7 @@
 	bad_ppro = ppro_with_ram_bug();
 
 #ifdef CONFIG_HIGHMEM
-	highmem_start_page = mem_map + highstart_pfn;
+	highmem_start_page = pfn_to_page(highstart_pfn);
 	max_mapnr = num_physpages = highend_pfn;
 #else
 	max_mapnr = num_physpages = max_low_pfn;
@@ -437,7 +437,7 @@
 		/*
 		 * Only count reserved RAM pages
 		 */
-		if (page_is_ram(tmp) && PageReserved(mem_map+tmp))
+		if (page_is_ram(tmp) && PageReserved(pfn_to_page(tmp)))
 			reservedpages++;
 
 	set_highmem_pages_init(bad_ppro);
diff -urN virgin-2.5.30/arch/i386/mm/pgtable.c linux-2.5.30-pfn_to_page/arch/i386/mm/pgtable.c
--- virgin-2.5.30/arch/i386/mm/pgtable.c	Thu Aug  1 14:17:27 2002
+++ linux-2.5.30-pfn_to_page/arch/i386/mm/pgtable.c	Thu Aug  1 17:55:07 2002
@@ -22,24 +22,26 @@
 
 void show_mem(void)
 {
-	int i, total = 0, reserved = 0;
+	int pfn, total = 0, reserved = 0;
 	int shared = 0, cached = 0;
 	int highmem = 0;
+	struct page *page;
 
 	printk("Mem-info:\n");
 	show_free_areas();
 	printk("Free swap:       %6dkB\n",nr_swap_pages<<(PAGE_SHIFT-10));
-	i = max_mapnr;
-	while (i-- > 0) {
+	pfn = max_mapnr;
+	while (pfn-- > 0) {
+		page = pfn_to_page(pfn);
 		total++;
-		if (PageHighMem(mem_map+i))
+		if (PageHighMem(page)
 			highmem++;
-		if (PageReserved(mem_map+i))
+		if (PageReserved(page))
 			reserved++;
-		else if (PageSwapCache(mem_map+i))
+		else if (PageSwapCache(page))
 			cached++;
-		else if (page_count(mem_map+i))
-			shared += page_count(mem_map+i) - 1;
+		else if (page_count(page))
+			shared += page_count(page) - 1;
 	}
 	printk("%d pages of RAM\n", total);
 	printk("%d pages of HIGHMEM\n",highmem);
diff -urN virgin-2.5.30/drivers/net/ns83820.c linux-2.5.30-pfn_to_page/drivers/net/ns83820.c
--- virgin-2.5.30/drivers/net/ns83820.c	Thu Aug  1 14:16:44 2002
+++ linux-2.5.30-pfn_to_page/drivers/net/ns83820.c	Thu Aug  1 22:43:41 2002
@@ -1081,7 +1081,7 @@
 				   frag->page_offset,
 				   frag->size, PCI_DMA_TODEVICE);
 		dprintk("frag: buf=%08Lx  page=%08lx offset=%08lx\n",
-			(long long)buf, (long)(frag->page - mem_map),
+			(long long)buf, (long) page_to_pfn(frag->page),
 			frag->page_offset);
 		len = frag->size;
 		frag++;
diff -urN virgin-2.5.30/include/asm-i386/pci.h linux-2.5.30-pfn_to_page/include/asm-i386/pci.h
--- virgin-2.5.30/include/asm-i386/pci.h	Thu Aug  1 14:16:22 2002
+++ linux-2.5.30-pfn_to_page/include/asm-i386/pci.h	Thu Aug  1 23:00:43 2002
@@ -109,7 +109,7 @@
 	if (direction == PCI_DMA_NONE)
 		BUG();
 
-	return (dma_addr_t)(page - mem_map) * PAGE_SIZE + offset;
+	return (dma_addr_t)(page_to_pfn(page)) * PAGE_SIZE + offset;
 }
 
 static inline void pci_unmap_page(struct pci_dev *hwdev, dma_addr_t dma_address,
@@ -240,7 +240,7 @@
 {
 	unsigned long poff = (dma_addr >> PAGE_SHIFT);
 
-	return mem_map + poff;
+	return pfn_to_page(poff);
 }
 
 static __inline__ unsigned long
diff -urN virgin-2.5.30/include/asm-i386/pgtable.h linux-2.5.30-pfn_to_page/include/asm-i386/pgtable.h
--- virgin-2.5.30/include/asm-i386/pgtable.h	Thu Aug  1 14:16:32 2002
+++ linux-2.5.30-pfn_to_page/include/asm-i386/pgtable.h	Thu Aug  1 23:11:30 2002
@@ -235,7 +235,7 @@
 ((unsigned long) __va(pmd_val(pmd) & PAGE_MASK))
 
 #define pmd_page(pmd) \
-	(mem_map + (pmd_val(pmd) >> PAGE_SHIFT))
+	(pfn_to_page(pmd_val(pmd) >> PAGE_SHIFT))
 
 #define pmd_large(pmd) \
 	((pmd_val(pmd) & (_PAGE_PSE|_PAGE_PRESENT)) == (_PAGE_PSE|_PAGE_PRESENT))
diff -urN virgin-2.5.30/kernel/suspend.c linux-2.5.30-pfn_to_page/kernel/suspend.c
--- virgin-2.5.30/kernel/suspend.c	Thu Aug  1 14:16:15 2002
+++ linux-2.5.30-pfn_to_page/kernel/suspend.c	Thu Aug  1 23:17:51 2002
@@ -468,31 +468,33 @@
 {
 	int chunk_size;
 	int nr_copy_pages = 0;
-	int loop;
+	int pfn;
+	struct page *page;
 	
 	if (max_mapnr != num_physpages)
 		panic("mapnr is not expected");
-	for (loop = 0; loop < max_mapnr; loop++) {
-		if (PageHighMem(mem_map+loop))
+	for (pfn = 0; pfn < max_mapnr; pfn++) {
+		page = pfn_to_page(pfn);
+		if (PageHighMem(page))
 			panic("Swsusp not supported on highmem boxes. Send 1GB of RAM to <pavel@ucw.cz> and try again ;-).");
-		if (!PageReserved(mem_map+loop)) {
-			if (PageNosave(mem_map+loop))
+		if (!PageReserved(page)) {
+			if (PageNosave(page))
 				continue;
 
-			if ((chunk_size=is_head_of_free_region(mem_map+loop))!=0) {
-				loop += chunk_size - 1;
+			if ((chunk_size=is_head_of_free_region(page))!=0) {
+				pfn += chunk_size - 1;
 				continue;
 			}
-		} else if (PageReserved(mem_map+loop)) {
-			BUG_ON (PageNosave(mem_map+loop));
+		} else if (PageReserved(page)) {
+			BUG_ON (PageNosave(page));
 
 			/*
 			 * Just copy whole code segment. Hopefully it is not that big.
 			 */
-			if (ADDRESS(loop) >= (unsigned long)
-				&__nosave_begin && ADDRESS(loop) < 
+			if (ADDRESS(pfn) >= (unsigned long)
+				&__nosave_begin && ADDRESS(pfn) < 
 				(unsigned long)&__nosave_end) {
-				PRINTK("[nosave %x]", ADDRESS(loop));
+				PRINTK("[nosave %x]", ADDRESS(pfn));
 				continue;
 			}
 			/* Hmm, perhaps copying all reserved pages is not too healthy as they may contain 
@@ -501,7 +503,7 @@
 
 		nr_copy_pages++;
 		if (pagedir_p) {
-			pagedir_p->orig_address = ADDRESS(loop);
+			pagedir_p->orig_address = ADDRESS(pfn);
 			copy_page(pagedir_p->address, pagedir_p->orig_address);
 			pagedir_p++;
 		}
diff -urN virgin-2.5.30/mm/page_alloc.c linux-2.5.30-pfn_to_page/mm/page_alloc.c
--- virgin-2.5.30/mm/page_alloc.c	Thu Aug  1 14:16:06 2002
+++ linux-2.5.30-pfn_to_page/mm/page_alloc.c	Thu Aug  1 23:25:13 2002
@@ -47,9 +47,9 @@
  */
 static inline int bad_range(zone_t *zone, struct page *page)
 {
-	if (page - mem_map >= zone->zone_start_mapnr + zone->size)
+	if (page_to_pfn(page) >= zone->zone_start_mapnr + zone->size)
 		return 1;
-	if (page - mem_map < zone->zone_start_mapnr)
+	if (page_to_pfn(page) < zone->zone_start_mapnr)
 		return 1;
 	if (zone != page_zone(page))
 		return 1;

--
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/

Re: [RFC] reduce usage of mem_map

[William Lee Irwin III]

jn Thu, Aug 01, 2002 at 11:38:07PM -0700, Martin J. Bligh wrote:
> I've tried to cut down the usage of mem_map somewhat.
> There's already macros to do the conversion between
> pfns to pages, and it doesn't work the way they've
> embedded it for discontigmem systems. Comments?
> Please don't apply - not tested yet ;-)
> M.

mem_map should be eliminated anyway. The concept is a legacy interface
for indexing into the older contiguous core map. The only pseudo-useful
thing it does is creating some kind of base address for address
calculation in page_to_pfn() etc., not that they're any kind of
performance-critical bit of the kernel. Those kinds of things can
circumvent the core VM without any significant ugliness impact.
Worse comes to worse, just do Roman Zippel's thing and shove the
page_address() bits into arch code where they'd belong if arches weren't
forcing us to keep ->virtual all the time (necessitating core control).

In the meantime, the necessity of laying out the pageframe maps in such
a manner that indexing from mem_map is sort of valid is an irritating
constraint not satisfiable without virtual remapping tricks and/or
MAP_NR_DENSE() on many platforms. Kill it. Kill it dead.


Cheers,
Bill
--
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/