Skip to contentSkip to conversation

sparsemem: more separation between NUMA and DISCONTIG

RFC PATCH In review Dave Hansen 2 participants 3 messages Mar 17, 2005 – Mar 18, 2005

Subsystems (5)
Memory management MM core Generic include/asm header files Page allocator +1 more subsystems
lore
Share & tools
lore

Dave Hansen

author
#0 lore raw

RFC PATCH 5/6 sparsemem: more separation between NUMA and DISCONTIG

There is some confusion with the SPARSEMEM patch between what is needed for DISCONTIG vs. NUMA. For instance, the NODE_DATA() macro needs to be switched on NUMA, but not on FLATMEM.

This patch is required if the previous patch is applied.

Signed-off-by Dave Hansen <[email protected]>

---

 memhotplug-dave/arch/i386/mm/init.c       |   12 ++++++------
 memhotplug-dave/include/asm-i386/mmzone.h |    4 ++--
 memhotplug-dave/include/linux/mmzone.h    |    8 ++++++--
 memhotplug-dave/mm/page_alloc.c           |    2 +-
 mm/memory.c                               |    0 
 5 files changed, 15 insertions(+), 11 deletions(-)

diff -puN arch/i386/Kconfig~B-sparse-161-i386-separate-flatmem-and-numa arch/i386/Kconfig

include/asm-i386/mmzone.h renamed from include/asm-i386/mmzone.h~B-sparse-161-i386-separate-flatmem-and-numa +2 -2
diff -puN include/asm-i386/mmzone.h~B-sparse-161-i386-separate-flatmem-and-numa include/asm-i386/mmzone.h
@@ -8,7 +8,7 @@
 
 #include <asm/smp.h>
 
-#if defined(CONFIG_DISCONTIGMEM) || defined(CONFIG_SPARSEMEM)
+#if CONFIG_NUMA
 extern struct pglist_data *node_data[];
 #define NODE_DATA(nid)	(node_data[nid])
 

@@ -42,7 +42,7 @@ static inline void get_memcfg_numa(void)
 	get_memcfg_numa_flat();
 }
 
-#endif /* !CONFIG_DISCONTIGMEM || !CONFIG_SPARSEMEM */
+#endif /* CONFIG_NUMA */
 
 #ifdef CONFIG_DISCONTIGMEM

diff -puN include/linux/bootmem.h~B-sparse-161-i386-separate-flatmem-and-numa include/linux/bootmem.h

include/linux/mmzone.h renamed from include/linux/mmzone.h~B-sparse-161-i386-separate-flatmem-and-numa +6 -2
diff -puN include/linux/mmzone.h~B-sparse-161-i386-separate-flatmem-and-numa include/linux/mmzone.h
@@ -383,10 +383,13 @@ int lowmem_reserve_ratio_sysctl_handler(
 /* Returns the number of the current Node. */
 #define numa_node_id()		(cpu_to_node(_smp_processor_id()))
 
+#ifndef CONFIG_NUMA
+#define NODE_DATA(nid)		(&contig_page_data)
+extern struct pglist_data contig_page_data;
+#endif
+
 #ifdef CONFIG_FLATMEM
 
-extern struct pglist_data contig_page_data;
-#define NODE_DATA(nid)		(&contig_page_data)
 #define NODE_MEM_MAP(nid)	mem_map
 #define MAX_NODES_SHIFT		1
 #define pfn_to_nid(pfn)		(0)

@@ -521,6 +524,7 @@ static inline int pfn_valid(unsigned lon
 #define pfn_to_nid		early_pfn_to_nid
 #else
 #define pfn_to_nid(pfn) 0
+#define early_pfn_to_nid(pfn)	0
 #endif
 
 #define pfn_to_pgdat(pfn)						\
mm/page_alloc.c renamed from mm/page_alloc.c~B-sparse-161-i386-separate-flatmem-and-numa +1 -1
diff -puN mm/page_alloc.c~B-sparse-161-i386-separate-flatmem-and-numa mm/page_alloc.c
@@ -1784,7 +1784,7 @@ void __init free_area_init_node(int nid,
 	free_area_init_core(pgdat, zones_size, zholes_size);
 }
 
-#ifdef CONFIG_FLATMEM
+#ifndef CONFIG_DISCONTIGMEM
 static bootmem_data_t contig_bootmem_data;
 struct pglist_data contig_page_data = { .bdata = &contig_bootmem_data };

diff -puN arch/i386/kernel/setup.c~B-sparse-161-i386-separate-flatmem-and-numa arch/i386/kernel/setup.c

arch/i386/mm/init.c renamed from arch/i386/mm/init.c~B-sparse-161-i386-separate-flatmem-and-numa +6 -6
diff -puN arch/i386/mm/init.c~B-sparse-161-i386-separate-flatmem-and-numa arch/i386/mm/init.c
@@ -277,7 +277,9 @@ void __init one_highpage_init(struct pag
 		SetPageReserved(page);
 }
 
-#ifdef CONFIG_FLATMEM
+#ifdef CONFIG_NUMA
+extern void set_highmem_pages_init(int);
+#else
 static void __init set_highmem_pages_init(int bad_ppro)
 {
 	int pfn;

@@ -285,8 +287,6 @@ static void __init set_highmem_pages_ini
 		one_highpage_init(pfn_to_page(pfn), pfn, bad_ppro);
 	totalram_pages += totalhigh_pages;
 }
-#else
-extern void set_highmem_pages_init(int);
 #endif /* CONFIG_FLATMEM */
 
 #else

@@ -298,10 +298,10 @@ extern void set_highmem_pages_init(int);
 unsigned long long __PAGE_KERNEL = _PAGE_KERNEL;
 unsigned long long __PAGE_KERNEL_EXEC = _PAGE_KERNEL_EXEC;
 
-#ifdef CONFIG_FLATMEM
-#define remap_numa_kva() do {} while (0)
-#else
+#ifdef CONFIG_NUMA
 extern void __init remap_numa_kva(void);
+#else
+#define remap_numa_kva() do {} while (0)
 #endif
 
 static void __init pagetable_init (void)

diff -puN mm/memory.c~B-sparse-161-i386-separate-flatmem-and-numa mm/memory.c _

#1 lore raw
Quoted 6 lines · Dave Hansen

There is some confusion with the SPARSEMEM patch between what is needed for DISCONTIG vs. NUMA. For instance, the NODE_DATA() macro needs to be switched on NUMA, but not on FLATMEM.

This patch is required if the previous patch is applied.

This patch breaks !CONFIG_NUMA ppc64:

include/linux/mmzone.h:387:1: warning: "NODE_DATA" redefined include/asm/mmzone.h:55:1: warning: this is the location of the previous definition

I'll hack around it for now.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

author
#2 lore raw
Quoted 6 lines · Andrew Morton
Quoted 5 lines · Dave Hansen

There is some confusion with the SPARSEMEM patch between what is needed for DISCONTIG vs. NUMA. For instance, the NODE_DATA() macro needs to be switched on NUMA, but not on FLATMEM.

Quoted 8 lines
Quoted 1 line

This patch is required if the previous patch is applied.

This patch breaks !CONFIG_NUMA ppc64:

include/linux/mmzone.h:387:1: warning: "NODE_DATA" redefined include/asm/mmzone.h:55:1: warning: this is the location of the previous definition

I'll hack around it for now.

I'll make sure to have it fixed properly in my copy.

Could I have a copy of your .config? I'm keeping a growing collection.

-- Dave

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>