[PATCH 1/3] x86, mm, 64bit: Put early page table high.

[PATCH 1/3] x86, mm, 64bit: Put early page table high.

[Yinghai Lu]

While dubug kdump, found current kernel will have problem with crashkernel=512M.

It turns out that initial mapping is to 512M, and later initial mapping to 4G
(acutally is 2040M in my platform), will put page table near 512M. 
then initial mapping to 128g will be near 2g.

before this patch:
[    0.000000] initial memory mapped : 0 - 20000000
[    0.000000] init_memory_mapping: [0x00000000000000-0x0000007f74ffff]
[    0.000000]  0000000000 - 007f600000 page 2M
[    0.000000]  007f600000 - 007f750000 page 4k
[    0.000000] kernel direct mapping tables up to 7f750000 @ [0x1fffc000-0x1fffffff]
[    0.000000]     memblock_x86_reserve_range: [0x1fffc000-0x1fffdfff]          PGTABLE
[    0.000000] init_memory_mapping: [0x00000100000000-0x0000207fffffff]
[    0.000000]  0100000000 - 2080000000 page 2M
[    0.000000] kernel direct mapping tables up to 2080000000 @ [0x7bc01000-0x7bc83fff]
[    0.000000]     memblock_x86_reserve_range: [0x7bc01000-0x7bc7efff]          PGTABLE
[    0.000000] RAMDISK: 7bc84000 - 7f745000
[    0.000000] crashkernel reservation failed - No suitable area found.

after patch:
[    0.000000] initial memory mapped : 0 - 20000000
[    0.000000] init_memory_mapping: [0x00000000000000-0x0000007f74ffff]
[    0.000000]  0000000000 - 007f600000 page 2M
[    0.000000]  007f600000 - 007f750000 page 4k
[    0.000000] kernel direct mapping tables up to 7f750000 @ [0x7f74c000-0x7f74ffff]
[    0.000000]     memblock_x86_reserve_range: [0x7f74c000-0x7f74dfff]          PGTABLE
[    0.000000] init_memory_mapping: [0x00000100000000-0x0000207fffffff]
[    0.000000]  0100000000 - 2080000000 page 2M
[    0.000000] kernel direct mapping tables up to 2080000000 @ [0x207ff7d000-0x207fffffff]
[    0.000000]     memblock_x86_reserve_range: [0x207ff7d000-0x207fffafff]          PGTABLE
[    0.000000] RAMDISK: 7bc84000 - 7f745000
[    0.000000]     memblock_x86_reserve_range: [0x17000000-0x36ffffff]     CRASH KERNEL
[    0.000000] Reserving 512MB of memory at 368MB for crashkernel (System RAM: 133120MB)

It means with the patch, page table for [0, 2g) will need 2g, instead of under 512M,
page table for [4g, 128g) will be near 128g, instead of under 2g.

That would good, if we have lots of memory above 4g, like 1024g, or 2048g or 16T, will not put
related page table under 2g. that would be have chance to fill the under 2g if 1G or 2M page is
not used.

the code change will use add map_low_page() and update unmap_low_page() for 64bit, and use them
to get access the corresponding high memory for page table setting.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/mm/init.c    |    9 ++-----
 arch/x86/mm/init_64.c |   63 +++++++++++++++++++++-----------------------------
 2 files changed, 30 insertions(+), 42 deletions(-)

Index: linux-2.6/arch/x86/mm/init.c
===================================================================
--- linux-2.6.orig/arch/x86/mm/init.c
+++ linux-2.6/arch/x86/mm/init.c
@@ -33,7 +33,7 @@ int direct_gbpages
 static void __init find_early_table_space(unsigned long end, int use_pse,
 					  int use_gbpages)
 {
-	unsigned long puds, pmds, ptes, tables, start;
+	unsigned long puds, pmds, ptes, tables, start = 0, good_end = end;
 	phys_addr_t base;
 
 	puds = (end + PUD_SIZE - 1) >> PUD_SHIFT;
@@ -73,12 +73,9 @@ static void __init find_early_table_spac
 	 * need roughly 0.5KB per GB.
 	 */
 #ifdef CONFIG_X86_32
-	start = 0x7000;
-#else
-	start = 0x8000;
+	good_end = max_pfn_mapped << PAGE_SHIFT;
 #endif
-	base = memblock_find_in_range(start, max_pfn_mapped<<PAGE_SHIFT,
-					tables, PAGE_SIZE);
+	base = memblock_find_in_range(start, good_end, tables, PAGE_SIZE);
 	if (base == MEMBLOCK_ERROR)
 		panic("Cannot find space for the kernel page tables");
 
Index: linux-2.6/arch/x86/mm/init_64.c
===================================================================
--- linux-2.6.orig/arch/x86/mm/init_64.c
+++ linux-2.6/arch/x86/mm/init_64.c
@@ -333,12 +333,28 @@ static __ref void *alloc_low_page(unsign
 	return adr;
 }
 
+static __ref void *map_low_page(void *virt)
+{
+	void *adr;
+	unsigned long phys, left;
+
+	if (after_bootmem)
+		return virt;
+
+	phys = __pa(virt);
+	left = phys & (PAGE_SIZE - 1);
+	adr = early_memremap(phys & PAGE_MASK, PAGE_SIZE);
+	adr = (void *)(((unsigned long)adr) | left);
+
+	return adr;
+}
+
 static __ref void unmap_low_page(void *adr)
 {
 	if (after_bootmem)
 		return;
 
-	early_iounmap(adr, PAGE_SIZE);
+	early_iounmap((void *)((unsigned long)adr & PAGE_MASK), PAGE_SIZE);
 }
 
 static unsigned long __meminit
@@ -386,15 +402,6 @@ phys_pte_init(pte_t *pte_page, unsigned
 }
 
 static unsigned long __meminit
-phys_pte_update(pmd_t *pmd, unsigned long address, unsigned long end,
-		pgprot_t prot)
-{
-	pte_t *pte = (pte_t *)pmd_page_vaddr(*pmd);
-
-	return phys_pte_init(pte, address, end, prot);
-}
-
-static unsigned long __meminit
 phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
 	      unsigned long page_size_mask, pgprot_t prot)
 {
@@ -420,8 +427,10 @@ phys_pmd_init(pmd_t *pmd_page, unsigned
 		if (pmd_val(*pmd)) {
 			if (!pmd_large(*pmd)) {
 				spin_lock(&init_mm.page_table_lock);
-				last_map_addr = phys_pte_update(pmd, address,
+				pte = map_low_page((pte_t *)pmd_page_vaddr(*pmd));
+				last_map_addr = phys_pte_init(pte, address,
 								end, prot);
+				unmap_low_page(pte);
 				spin_unlock(&init_mm.page_table_lock);
 				continue;
 			}
@@ -468,18 +477,6 @@ phys_pmd_init(pmd_t *pmd_page, unsigned
 }
 
 static unsigned long __meminit
-phys_pmd_update(pud_t *pud, unsigned long address, unsigned long end,
-		unsigned long page_size_mask, pgprot_t prot)
-{
-	pmd_t *pmd = pmd_offset(pud, 0);
-	unsigned long last_map_addr;
-
-	last_map_addr = phys_pmd_init(pmd, address, end, page_size_mask, prot);
-	__flush_tlb_all();
-	return last_map_addr;
-}
-
-static unsigned long __meminit
 phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
 			 unsigned long page_size_mask)
 {
@@ -504,8 +501,11 @@ phys_pud_init(pud_t *pud_page, unsigned
 
 		if (pud_val(*pud)) {
 			if (!pud_large(*pud)) {
-				last_map_addr = phys_pmd_update(pud, addr, end,
+				pmd = map_low_page(pmd_offset(pud, 0));
+				last_map_addr = phys_pmd_init(pmd, addr, end,
 							 page_size_mask, prot);
+				unmap_low_page(pmd);
+				__flush_tlb_all();
 				continue;
 			}
 			/*
@@ -553,17 +553,6 @@ phys_pud_init(pud_t *pud_page, unsigned
 	return last_map_addr;
 }
 
-static unsigned long __meminit
-phys_pud_update(pgd_t *pgd, unsigned long addr, unsigned long end,
-		 unsigned long page_size_mask)
-{
-	pud_t *pud;
-
-	pud = (pud_t *)pgd_page_vaddr(*pgd);
-
-	return phys_pud_init(pud, addr, end, page_size_mask);
-}
-
 unsigned long __meminit
 kernel_physical_mapping_init(unsigned long start,
 			     unsigned long end,
@@ -587,8 +576,10 @@ kernel_physical_mapping_init(unsigned lo
 			next = end;
 
 		if (pgd_val(*pgd)) {
-			last_map_addr = phys_pud_update(pgd, __pa(start),
+			pud = map_low_page((pud_t *)pgd_page_vaddr(*pgd));
+			last_map_addr = phys_pud_init(pud, __pa(start),
 						 __pa(end), page_size_mask);
+			unmap_low_page(pud);
 			continue;
 		}
 

[PATCH 2/3] x86, 64bit, gart: Fix allocation with memblock

[Yinghai Lu]

When trying to change alloc_bootmem with memblock to go with real top-down
Found one old system:
[    0.000000] Node 0: aperture @ ac000000 size 64 MB
[    0.000000] Aperture pointing to e820 RAM. Ignoring.
[    0.000000] Your BIOS doesn't leave a aperture memory hole
[    0.000000] Please enable the IOMMU option in the BIOS setup
[    0.000000] This costs you 64 MB of RAM
[    0.000000]     memblock_x86_reserve_range: [0x2020000000-0x2023ffffff]       aperture64
[    0.000000] Cannot allocate aperture memory hole (ffff882020000000,65536K)
[    0.000000]        memblock_x86_free_range: [0x2020000000-0x2023ffffff]
[    0.000000] Kernel panic - not syncing: Not enough memory for aperture
[    0.000000] Pid: 0, comm: swapper Not tainted 2.6.37-rc5-tip-yh-06229-gb792dc2-dirty #331
[    0.000000] Call Trace:
[    0.000000]  [<ffffffff81cf50fe>] ? panic+0x91/0x1a3
[    0.000000]  [<ffffffff827c66b2>] ? gart_iommu_hole_init+0x3d7/0x4a3
[    0.000000]  [<ffffffff81d026a9>] ? _etext+0x0/0x3
[    0.000000]  [<ffffffff827ba940>] ? pci_iommu_alloc+0x47/0x71
[    0.000000]  [<ffffffff827c820b>] ? mem_init+0x19/0xec
[    0.000000]  [<ffffffff827b3c40>] ? start_kernel+0x20a/0x3e8
[    0.000000]  [<ffffffff827b32cc>] ? x86_64_start_reservations+0x9c/0xa0
[    0.000000]  [<ffffffff827b33e4>] ? x86_64_start_kernel+0x114/0x11b

it means __alloc_bootmem_nopanic() get too high for that aperture.

Use memblock_find_in_range() with limit directly.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 arch/x86/kernel/aperture_64.c |   33 ++++++++++++++++-----------------
 1 file changed, 16 insertions(+), 17 deletions(-)

Index: linux-2.6/arch/x86/kernel/aperture_64.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/aperture_64.c
+++ linux-2.6/arch/x86/kernel/aperture_64.c
@@ -13,7 +13,7 @@
 #include <linux/kernel.h>
 #include <linux/types.h>
 #include <linux/init.h>
-#include <linux/bootmem.h>
+#include <linux/memblock.h>
 #include <linux/mmzone.h>
 #include <linux/pci_ids.h>
 #include <linux/pci.h>
@@ -69,7 +69,7 @@ static void __init insert_aperture_resou
 static u32 __init allocate_aperture(void)
 {
 	u32 aper_size;
-	void *p;
+	unsigned long addr;
 
 	/* aper_size should <= 1G */
 	if (fallback_aper_order > 5)
@@ -95,27 +95,26 @@ static u32 __init allocate_aperture(void
 	 * so don't use 512M below as gart iommu, leave the space for kernel
 	 * code for safe
 	 */
-	p = __alloc_bootmem_nopanic(aper_size, aper_size, 512ULL<<20);
+	addr = memblock_find_in_range(0, 1ULL<<32, aper_size, 512ULL<<20);
+	if (addr == MEMBLOCK_ERROR || addr + aper_size > 0xffffffff) {
+		printk(KERN_ERR
+			"Cannot allocate aperture memory hole (%lx,%uK)\n",
+				addr, aper_size>>10);
+		return 0;
+	}
+	memblock_x86_reserve_range(addr, addr + aper_size, "aperture64");
 	/*
 	 * Kmemleak should not scan this block as it may not be mapped via the
 	 * kernel direct mapping.
 	 */
-	kmemleak_ignore(p);
-	if (!p || __pa(p)+aper_size > 0xffffffff) {
-		printk(KERN_ERR
-			"Cannot allocate aperture memory hole (%p,%uK)\n",
-				p, aper_size>>10);
-		if (p)
-			free_bootmem(__pa(p), aper_size);
-		return 0;
-	}
+	kmemleak_ignore(phys_to_virt(addr));
 	printk(KERN_INFO "Mapping aperture over %d KB of RAM @ %lx\n",
-			aper_size >> 10, __pa(p));
-	insert_aperture_resource((u32)__pa(p), aper_size);
-	register_nosave_region((u32)__pa(p) >> PAGE_SHIFT,
-				(u32)__pa(p+aper_size) >> PAGE_SHIFT);
+			aper_size >> 10, addr);
+	insert_aperture_resource((u32)addr, aper_size);
+	register_nosave_region(addr >> PAGE_SHIFT,
+			       (addr+aper_size) >> PAGE_SHIFT);
 
-	return (u32)__pa(p);
+	return (u32)addr;
 }
 
 

[PATCH 3/3] memblock: Make find_memory_core_early() find from top-down

[Yinghai Lu]

That is used for find ram in node or bootmem type.

We should make it top-down so it will be consistent to memblock_find.

Suggested-by: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 mm/page_alloc.c |   34 +++++++++++++++++++++++++++++++++-
 1 file changed, 33 insertions(+), 1 deletion(-)

Index: linux-2.6/mm/page_alloc.c
===================================================================
--- linux-2.6.orig/mm/page_alloc.c
+++ linux-2.6/mm/page_alloc.c
@@ -3556,6 +3556,34 @@ static int __meminit next_active_region_
 	return -1;
 }
 
+/*
+ * Basic iterator support. Return the last range of PFNs for a node
+ * Note: nid == MAX_NUMNODES returns last region regardless of node
+ */
+static int __meminit last_active_region_index_in_nid(int nid)
+{
+	int i;
+
+	for (i = nr_nodemap_entries - 1; i >= 0; i--)
+		if (nid == MAX_NUMNODES || early_node_map[i].nid == nid)
+			return i;
+
+	return -1;
+}
+
+/*
+ * Basic iterator support. Return the previous active range of PFNs for a node
+ * Note: nid == MAX_NUMNODES returns next region regardless of node
+ */
+static int __meminit previous_active_region_index_in_nid(int index, int nid)
+{
+	for (index = index - 1; index >= 0; index--)
+		if (nid == MAX_NUMNODES || early_node_map[index].nid == nid)
+			return index;
+
+	return -1;
+}
+
 #ifndef CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID
 /*
  * Required by SPARSEMEM. Given a PFN, return what node the PFN is on.
@@ -3607,6 +3635,10 @@ bool __meminit early_pfn_in_nid(unsigned
 	for (i = first_active_region_index_in_nid(nid); i != -1; \
 				i = next_active_region_index_in_nid(i, nid))
 
+#define for_each_active_range_index_in_nid_reverse(i, nid) \
+	for (i = last_active_region_index_in_nid(nid); i != -1; \
+				i = previous_active_region_index_in_nid(i, nid))
+
 /**
  * free_bootmem_with_active_regions - Call free_bootmem_node for each active range
  * @nid: The node to free memory on. If MAX_NUMNODES, all nodes are freed.
@@ -3645,7 +3677,7 @@ u64 __init find_memory_core_early(int ni
 	int i;
 
 	/* Need to go over early_node_map to find out good range for node */
-	for_each_active_range_index_in_nid(i, nid) {
+	for_each_active_range_index_in_nid_reverse(i, nid) {
 		u64 addr;
 		u64 ei_start, ei_last;
 		u64 final_start, final_end;

[PATCH 3/3] memblock: Make find_memory_core_early() find from top-down

[Yinghai Lu]

That is used for find ram in node or bootmem type.

We should make it top-down so it will be consistent to memblock_find.

Suggested-by: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>

---
 mm/page_alloc.c |   34 +++++++++++++++++++++++++++++++++-
 1 file changed, 33 insertions(+), 1 deletion(-)

Index: linux-2.6/mm/page_alloc.c
===================================================================
--- linux-2.6.orig/mm/page_alloc.c
+++ linux-2.6/mm/page_alloc.c
@@ -3556,6 +3556,34 @@ static int __meminit next_active_region_
 	return -1;
 }
 
+/*
+ * Basic iterator support. Return the last range of PFNs for a node
+ * Note: nid == MAX_NUMNODES returns last region regardless of node
+ */
+static int __meminit last_active_region_index_in_nid(int nid)
+{
+	int i;
+
+	for (i = nr_nodemap_entries - 1; i >= 0; i--)
+		if (nid == MAX_NUMNODES || early_node_map[i].nid == nid)
+			return i;
+
+	return -1;
+}
+
+/*
+ * Basic iterator support. Return the previous active range of PFNs for a node
+ * Note: nid == MAX_NUMNODES returns next region regardless of node
+ */
+static int __meminit previous_active_region_index_in_nid(int index, int nid)
+{
+	for (index = index - 1; index >= 0; index--)
+		if (nid == MAX_NUMNODES || early_node_map[index].nid == nid)
+			return index;
+
+	return -1;
+}
+
 #ifndef CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID
 /*
  * Required by SPARSEMEM. Given a PFN, return what node the PFN is on.
@@ -3607,6 +3635,10 @@ bool __meminit early_pfn_in_nid(unsigned
 	for (i = first_active_region_index_in_nid(nid); i != -1; \
 				i = next_active_region_index_in_nid(i, nid))
 
+#define for_each_active_range_index_in_nid_reverse(i, nid) \
+	for (i = last_active_region_index_in_nid(nid); i != -1; \
+				i = previous_active_region_index_in_nid(i, nid))
+
 /**
  * free_bootmem_with_active_regions - Call free_bootmem_node for each active range
  * @nid: The node to free memory on. If MAX_NUMNODES, all nodes are freed.
@@ -3645,7 +3677,7 @@ u64 __init find_memory_core_early(int ni
 	int i;
 
 	/* Need to go over early_node_map to find out good range for node */
-	for_each_active_range_index_in_nid(i, nid) {
+	for_each_active_range_index_in_nid_reverse(i, nid) {
 		u64 addr;
 		u64 ei_start, ei_last;
 		u64 final_start, final_end;

--
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/ .
Fight unfair telecom policy in Canada: sign http://dissolvethecrtc.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[tip:x86/mm] x86-64, mm: Put early page table high

[tip-bot for Yinghai Lu]

Commit-ID:  4b239f458c229de044d6905c2b0f9fe16ed9e01e
Gitweb:     http://git.kernel.org/tip/4b239f458c229de044d6905c2b0f9fe16ed9e01e
Author:     Yinghai Lu <yinghai@kernel.org>
AuthorDate: Fri, 17 Dec 2010 16:58:28 -0800
Committer:  H. Peter Anvin <hpa@linux.intel.com>
CommitDate: Wed, 29 Dec 2010 14:46:54 -0800

x86-64, mm: Put early page table high

While dubug kdump, found current kernel will have problem with crashkernel=512M.

It turns out that initial mapping is to 512M, and later initial mapping to 4G
(acutally is 2040M in my platform), will put page table near 512M.
then initial mapping to 128g will be near 2g.

before this patch:
[    0.000000] initial memory mapped : 0 - 20000000
[    0.000000] init_memory_mapping: [0x00000000000000-0x0000007f74ffff]
[    0.000000]  0000000000 - 007f600000 page 2M
[    0.000000]  007f600000 - 007f750000 page 4k
[    0.000000] kernel direct mapping tables up to 7f750000 @ [0x1fffc000-0x1fffffff]
[    0.000000]     memblock_x86_reserve_range: [0x1fffc000-0x1fffdfff]          PGTABLE
[    0.000000] init_memory_mapping: [0x00000100000000-0x0000207fffffff]
[    0.000000]  0100000000 - 2080000000 page 2M
[    0.000000] kernel direct mapping tables up to 2080000000 @ [0x7bc01000-0x7bc83fff]
[    0.000000]     memblock_x86_reserve_range: [0x7bc01000-0x7bc7efff]          PGTABLE
[    0.000000] RAMDISK: 7bc84000 - 7f745000
[    0.000000] crashkernel reservation failed - No suitable area found.

after patch:
[    0.000000] initial memory mapped : 0 - 20000000
[    0.000000] init_memory_mapping: [0x00000000000000-0x0000007f74ffff]
[    0.000000]  0000000000 - 007f600000 page 2M
[    0.000000]  007f600000 - 007f750000 page 4k
[    0.000000] kernel direct mapping tables up to 7f750000 @ [0x7f74c000-0x7f74ffff]
[    0.000000]     memblock_x86_reserve_range: [0x7f74c000-0x7f74dfff]          PGTABLE
[    0.000000] init_memory_mapping: [0x00000100000000-0x0000207fffffff]
[    0.000000]  0100000000 - 2080000000 page 2M
[    0.000000] kernel direct mapping tables up to 2080000000 @ [0x207ff7d000-0x207fffffff]
[    0.000000]     memblock_x86_reserve_range: [0x207ff7d000-0x207fffafff]          PGTABLE
[    0.000000] RAMDISK: 7bc84000 - 7f745000
[    0.000000]     memblock_x86_reserve_range: [0x17000000-0x36ffffff]     CRASH KERNEL
[    0.000000] Reserving 512MB of memory at 368MB for crashkernel (System RAM: 133120MB)

It means with the patch, page table for [0, 2g) will need 2g, instead of under 512M,
page table for [4g, 128g) will be near 128g, instead of under 2g.

That would good, if we have lots of memory above 4g, like 1024g, or 2048g or 16T, will not put
related page table under 2g. that would be have chance to fill the under 2g if 1G or 2M page is
not used.

the code change will use add map_low_page() and update unmap_low_page() for 64bit, and use them
to get access the corresponding high memory for page table setting.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <4D0C0734.7060900@kernel.org>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
---
 arch/x86/mm/init.c    |    9 ++----
 arch/x86/mm/init_64.c |   63 +++++++++++++++++++++----------------------------
 2 files changed, 30 insertions(+), 42 deletions(-)

diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c
index c0e28a1..5863950 100644
--- a/arch/x86/mm/init.c
+++ b/arch/x86/mm/init.c
@@ -33,7 +33,7 @@ int direct_gbpages
 static void __init find_early_table_space(unsigned long end, int use_pse,
 					  int use_gbpages)
 {
-	unsigned long puds, pmds, ptes, tables, start;
+	unsigned long puds, pmds, ptes, tables, start = 0, good_end = end;
 	phys_addr_t base;
 
 	puds = (end + PUD_SIZE - 1) >> PUD_SHIFT;
@@ -73,12 +73,9 @@ static void __init find_early_table_space(unsigned long end, int use_pse,
 	 * need roughly 0.5KB per GB.
 	 */
 #ifdef CONFIG_X86_32
-	start = 0x7000;
-#else
-	start = 0x8000;
+	good_end = max_pfn_mapped << PAGE_SHIFT;
 #endif
-	base = memblock_find_in_range(start, max_pfn_mapped<<PAGE_SHIFT,
-					tables, PAGE_SIZE);
+	base = memblock_find_in_range(start, good_end, tables, PAGE_SIZE);
 	if (base == MEMBLOCK_ERROR)
 		panic("Cannot find space for the kernel page tables");
 
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index 71a5929..024847d 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -333,12 +333,28 @@ static __ref void *alloc_low_page(unsigned long *phys)
 	return adr;
 }
 
+static __ref void *map_low_page(void *virt)
+{
+	void *adr;
+	unsigned long phys, left;
+
+	if (after_bootmem)
+		return virt;
+
+	phys = __pa(virt);
+	left = phys & (PAGE_SIZE - 1);
+	adr = early_memremap(phys & PAGE_MASK, PAGE_SIZE);
+	adr = (void *)(((unsigned long)adr) | left);
+
+	return adr;
+}
+
 static __ref void unmap_low_page(void *adr)
 {
 	if (after_bootmem)
 		return;
 
-	early_iounmap(adr, PAGE_SIZE);
+	early_iounmap((void *)((unsigned long)adr & PAGE_MASK), PAGE_SIZE);
 }
 
 static unsigned long __meminit
@@ -386,15 +402,6 @@ phys_pte_init(pte_t *pte_page, unsigned long addr, unsigned long end,
 }
 
 static unsigned long __meminit
-phys_pte_update(pmd_t *pmd, unsigned long address, unsigned long end,
-		pgprot_t prot)
-{
-	pte_t *pte = (pte_t *)pmd_page_vaddr(*pmd);
-
-	return phys_pte_init(pte, address, end, prot);
-}
-
-static unsigned long __meminit
 phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
 	      unsigned long page_size_mask, pgprot_t prot)
 {
@@ -420,8 +427,10 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
 		if (pmd_val(*pmd)) {
 			if (!pmd_large(*pmd)) {
 				spin_lock(&init_mm.page_table_lock);
-				last_map_addr = phys_pte_update(pmd, address,
+				pte = map_low_page((pte_t *)pmd_page_vaddr(*pmd));
+				last_map_addr = phys_pte_init(pte, address,
 								end, prot);
+				unmap_low_page(pte);
 				spin_unlock(&init_mm.page_table_lock);
 				continue;
 			}
@@ -468,18 +477,6 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
 }
 
 static unsigned long __meminit
-phys_pmd_update(pud_t *pud, unsigned long address, unsigned long end,
-		unsigned long page_size_mask, pgprot_t prot)
-{
-	pmd_t *pmd = pmd_offset(pud, 0);
-	unsigned long last_map_addr;
-
-	last_map_addr = phys_pmd_init(pmd, address, end, page_size_mask, prot);
-	__flush_tlb_all();
-	return last_map_addr;
-}
-
-static unsigned long __meminit
 phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
 			 unsigned long page_size_mask)
 {
@@ -504,8 +501,11 @@ phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
 
 		if (pud_val(*pud)) {
 			if (!pud_large(*pud)) {
-				last_map_addr = phys_pmd_update(pud, addr, end,
+				pmd = map_low_page(pmd_offset(pud, 0));
+				last_map_addr = phys_pmd_init(pmd, addr, end,
 							 page_size_mask, prot);
+				unmap_low_page(pmd);
+				__flush_tlb_all();
 				continue;
 			}
 			/*
@@ -553,17 +553,6 @@ phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
 	return last_map_addr;
 }
 
-static unsigned long __meminit
-phys_pud_update(pgd_t *pgd, unsigned long addr, unsigned long end,
-		 unsigned long page_size_mask)
-{
-	pud_t *pud;
-
-	pud = (pud_t *)pgd_page_vaddr(*pgd);
-
-	return phys_pud_init(pud, addr, end, page_size_mask);
-}
-
 unsigned long __meminit
 kernel_physical_mapping_init(unsigned long start,
 			     unsigned long end,
@@ -587,8 +576,10 @@ kernel_physical_mapping_init(unsigned long start,
 			next = end;
 
 		if (pgd_val(*pgd)) {
-			last_map_addr = phys_pud_update(pgd, __pa(start),
+			pud = map_low_page((pud_t *)pgd_page_vaddr(*pgd));
+			last_map_addr = phys_pud_init(pud, __pa(start),
 						 __pa(end), page_size_mask);
+			unmap_low_page(pud);
 			continue;
 		}
 

[tip:x86/mm] x86-64, gart: Fix allocation with memblock

[tip-bot for Yinghai Lu]

Commit-ID:  32e3f2b00c529477d26895c5428ed95bba537443
Gitweb:     http://git.kernel.org/tip/32e3f2b00c529477d26895c5428ed95bba537443
Author:     Yinghai Lu <yinghai@kernel.org>
AuthorDate: Fri, 17 Dec 2010 16:58:40 -0800
Committer:  H. Peter Anvin <hpa@linux.intel.com>
CommitDate: Wed, 29 Dec 2010 14:46:54 -0800

x86-64, gart: Fix allocation with memblock

When trying to change alloc_bootmem with memblock to go with real top-down
Found one old system:
[    0.000000] Node 0: aperture @ ac000000 size 64 MB
[    0.000000] Aperture pointing to e820 RAM. Ignoring.
[    0.000000] Your BIOS doesn't leave a aperture memory hole
[    0.000000] Please enable the IOMMU option in the BIOS setup
[    0.000000] This costs you 64 MB of RAM
[    0.000000]     memblock_x86_reserve_range: [0x2020000000-0x2023ffffff]       aperture64
[    0.000000] Cannot allocate aperture memory hole (ffff882020000000,65536K)
[    0.000000]        memblock_x86_free_range: [0x2020000000-0x2023ffffff]
[    0.000000] Kernel panic - not syncing: Not enough memory for aperture
[    0.000000] Pid: 0, comm: swapper Not tainted 2.6.37-rc5-tip-yh-06229-gb792dc2-dirty #331
[    0.000000] Call Trace:
[    0.000000]  [<ffffffff81cf50fe>] ? panic+0x91/0x1a3
[    0.000000]  [<ffffffff827c66b2>] ? gart_iommu_hole_init+0x3d7/0x4a3
[    0.000000]  [<ffffffff81d026a9>] ? _etext+0x0/0x3
[    0.000000]  [<ffffffff827ba940>] ? pci_iommu_alloc+0x47/0x71
[    0.000000]  [<ffffffff827c820b>] ? mem_init+0x19/0xec
[    0.000000]  [<ffffffff827b3c40>] ? start_kernel+0x20a/0x3e8
[    0.000000]  [<ffffffff827b32cc>] ? x86_64_start_reservations+0x9c/0xa0
[    0.000000]  [<ffffffff827b33e4>] ? x86_64_start_kernel+0x114/0x11b

it means __alloc_bootmem_nopanic() get too high for that aperture.

Use memblock_find_in_range() with limit directly.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <4D0C0740.90104@kernel.org>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
---
 arch/x86/kernel/aperture_64.c |   33 ++++++++++++++++-----------------
 1 files changed, 16 insertions(+), 17 deletions(-)

diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c
index dcd7c83..85f66b4 100644
--- a/arch/x86/kernel/aperture_64.c
+++ b/arch/x86/kernel/aperture_64.c
@@ -13,7 +13,7 @@
 #include <linux/kernel.h>
 #include <linux/types.h>
 #include <linux/init.h>
-#include <linux/bootmem.h>
+#include <linux/memblock.h>
 #include <linux/mmzone.h>
 #include <linux/pci_ids.h>
 #include <linux/pci.h>
@@ -69,7 +69,7 @@ static void __init insert_aperture_resource(u32 aper_base, u32 aper_size)
 static u32 __init allocate_aperture(void)
 {
 	u32 aper_size;
-	void *p;
+	unsigned long addr;
 
 	/* aper_size should <= 1G */
 	if (fallback_aper_order > 5)
@@ -95,27 +95,26 @@ static u32 __init allocate_aperture(void)
 	 * so don't use 512M below as gart iommu, leave the space for kernel
 	 * code for safe
 	 */
-	p = __alloc_bootmem_nopanic(aper_size, aper_size, 512ULL<<20);
+	addr = memblock_find_in_range(0, 1ULL<<32, aper_size, 512ULL<<20);
+	if (addr == MEMBLOCK_ERROR || addr + aper_size > 0xffffffff) {
+		printk(KERN_ERR
+			"Cannot allocate aperture memory hole (%lx,%uK)\n",
+				addr, aper_size>>10);
+		return 0;
+	}
+	memblock_x86_reserve_range(addr, addr + aper_size, "aperture64");
 	/*
 	 * Kmemleak should not scan this block as it may not be mapped via the
 	 * kernel direct mapping.
 	 */
-	kmemleak_ignore(p);
-	if (!p || __pa(p)+aper_size > 0xffffffff) {
-		printk(KERN_ERR
-			"Cannot allocate aperture memory hole (%p,%uK)\n",
-				p, aper_size>>10);
-		if (p)
-			free_bootmem(__pa(p), aper_size);
-		return 0;
-	}
+	kmemleak_ignore(phys_to_virt(addr));
 	printk(KERN_INFO "Mapping aperture over %d KB of RAM @ %lx\n",
-			aper_size >> 10, __pa(p));
-	insert_aperture_resource((u32)__pa(p), aper_size);
-	register_nosave_region((u32)__pa(p) >> PAGE_SHIFT,
-				(u32)__pa(p+aper_size) >> PAGE_SHIFT);
+			aper_size >> 10, addr);
+	insert_aperture_resource((u32)addr, aper_size);
+	register_nosave_region(addr >> PAGE_SHIFT,
+			       (addr+aper_size) >> PAGE_SHIFT);
 
-	return (u32)__pa(p);
+	return (u32)addr;
 }
 
 

[tip:x86/mm] memblock: Make find_memory_core_early() find from top-down

[tip-bot for Yinghai Lu]

Commit-ID:  1a4a678b12c84db9ae5dce424e0e97f0559bb57c
Gitweb:     http://git.kernel.org/tip/1a4a678b12c84db9ae5dce424e0e97f0559bb57c
Author:     Yinghai Lu <yinghai@kernel.org>
AuthorDate: Fri, 17 Dec 2010 16:59:07 -0800
Committer:  H. Peter Anvin <hpa@linux.intel.com>
CommitDate: Wed, 29 Dec 2010 14:46:55 -0800

memblock: Make find_memory_core_early() find from top-down

That is used for find ram in node or bootmem type.

We should make it top-down so it will be consistent to memblock_find,
and to avoid allocating potentially valuable low memory before we
actually need it.

Suggested-by: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <4D0C075B.3040501@kernel.org>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
---
 mm/page_alloc.c |   34 +++++++++++++++++++++++++++++++++-
 1 files changed, 33 insertions(+), 1 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 07a6544..19413bf 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3555,6 +3555,34 @@ static int __meminit next_active_region_index_in_nid(int index, int nid)
 	return -1;
 }
 
+/*
+ * Basic iterator support. Return the last range of PFNs for a node
+ * Note: nid == MAX_NUMNODES returns last region regardless of node
+ */
+static int __meminit last_active_region_index_in_nid(int nid)
+{
+	int i;
+
+	for (i = nr_nodemap_entries - 1; i >= 0; i--)
+		if (nid == MAX_NUMNODES || early_node_map[i].nid == nid)
+			return i;
+
+	return -1;
+}
+
+/*
+ * Basic iterator support. Return the previous active range of PFNs for a node
+ * Note: nid == MAX_NUMNODES returns next region regardless of node
+ */
+static int __meminit previous_active_region_index_in_nid(int index, int nid)
+{
+	for (index = index - 1; index >= 0; index--)
+		if (nid == MAX_NUMNODES || early_node_map[index].nid == nid)
+			return index;
+
+	return -1;
+}
+
 #ifndef CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID
 /*
  * Required by SPARSEMEM. Given a PFN, return what node the PFN is on.
@@ -3606,6 +3634,10 @@ bool __meminit early_pfn_in_nid(unsigned long pfn, int node)
 	for (i = first_active_region_index_in_nid(nid); i != -1; \
 				i = next_active_region_index_in_nid(i, nid))
 
+#define for_each_active_range_index_in_nid_reverse(i, nid) \
+	for (i = last_active_region_index_in_nid(nid); i != -1; \
+				i = previous_active_region_index_in_nid(i, nid))
+
 /**
  * free_bootmem_with_active_regions - Call free_bootmem_node for each active range
  * @nid: The node to free memory on. If MAX_NUMNODES, all nodes are freed.
@@ -3644,7 +3676,7 @@ u64 __init find_memory_core_early(int nid, u64 size, u64 align,
 	int i;
 
 	/* Need to go over early_node_map to find out good range for node */
-	for_each_active_range_index_in_nid(i, nid) {
+	for_each_active_range_index_in_nid_reverse(i, nid) {
 		u64 addr;
 		u64 ei_start, ei_last;
 		u64 final_start, final_end;