The patch titled
memory hotplug: register section/node id to free
has been added to the -mm tree. Its filename is
memory-hotplug-register-section-node-id-to-free.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this
The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/
------------------------------------------------------
Subject: memory hotplug: register section/node id to free
From: Yasunori Goto <y-goto@xxxxxxxxxxxxxx>
This patchset is to free pages which are allocated by bootmem for
memory-hotremove. Some structures of memory management are allocated from
bootmem, such as the memmap.
To remove memory physically, some of them must be freed according to
circumstance. This patch set makes basis to free those pages, and free
memmaps.
Basically my idea is using some remaining members of struct page to remember
information about the users of bootmem (section number or node id).
When the section is removed, the kernel can use this information to solve some
issues:
1) When the memmap of the to-be-removed section is allocated on another
section by bootmem, it should/can be free.
2) When the memmap of the to-be-removed section is allocated on the same
section, it shouldn't be freed. Because the section has to be offlined
already and all pages must be isolated against the page allocater.
3) When the to-be-removed section has another section's memmap, the kernel
will be able to show easily which section should be removed before it.
(Not implemented yet)
4) In case 2), the page migrator will be able to check and skip the memmap
during page isolation when page offline. Current page migration fails in
this case because this page is just a reserved page and it can't determine
whether this page can be removed or not. But, it will be able to do so
with the infrastructure which this patch adds. (Not implemented yet.)
5) The node information such as the pgdat has similar issues. And this
will also be able to be solved with this infrastructure. (Not implemented
yet, but, remembering node id in the pages.)
Fortunately, the current bootmem allocator just keeps PageReserved flags, and
doesn't use any other members of page struct. The users of bootmem doesn't
use them either.
This patch:
This is to register information which is a node or section's id. The kernel
can distinguish which node/section uses the pages allocated by bootmem. This
is the basis for hot-remove sections or nodes.
Signed-off-by: Yasunori Goto <y-goto@xxxxxxxxxxxxxx>
Cc: Yinghai Lu <yhlu.kernel@xxxxxxxxx>
Cc: Badari Pulavarty <pbadari@xxxxxxxxxx>
Cc: Christoph Lameter <clameter@xxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/linux/memory_hotplug.h | 18 +++++
include/linux/mmzone.h | 1
mm/bootmem.c | 1
mm/memory_hotplug.c | 97 ++++++++++++++++++++++++++++++-
mm/sparse.c | 3
5 files changed, 117 insertions(+), 3 deletions(-)
diff -puN include/linux/memory_hotplug.h~memory-hotplug-register-section-node-id-to-free include/linux/memory_hotplug.h
--- a/include/linux/memory_hotplug.h~memory-hotplug-register-section-node-id-to-free
+++ a/include/linux/memory_hotplug.h
@@ -11,6 +11,15 @@ struct pglist_data;
struct mem_section;
#ifdef CONFIG_MEMORY_HOTPLUG
+
+/*
+ * Magic number for free bootmem.
+ * The normal smallest mapcount is -1. Here is smaller value than it.
+ */
+#define SECTION_INFO 0xfffffffe
+#define MIX_INFO 0xfffffffd
+#define NODE_INFO 0xfffffffc
+
/*
* pgdat resizing functions
*/
@@ -145,6 +154,9 @@ static inline void arch_refresh_nodedata
#endif /* CONFIG_NUMA */
#endif /* CONFIG_HAVE_ARCH_NODEDATA_EXTENSION */
+extern void register_page_bootmem_info_node(struct pglist_data *pgdat);
+extern void put_page_bootmem(struct page *page);
+
#else /* ! CONFIG_MEMORY_HOTPLUG */
/*
* Stub functions for when hotplug is off
@@ -172,6 +184,10 @@ static inline int mhp_notimplemented(con
return -ENOSYS;
}
+static inline void register_page_bootmem_info_node(struct pglist_data *pgdat)
+{
+}
+
#endif /* ! CONFIG_MEMORY_HOTPLUG */
extern int add_memory(int nid, u64 start, u64 size);
@@ -180,5 +196,7 @@ extern int remove_memory(u64 start, u64
extern int sparse_add_one_section(struct zone *zone, unsigned long start_pfn,
int nr_pages);
extern void sparse_remove_one_section(struct zone *zone, struct mem_section *ms);
+extern struct page *sparse_decode_mem_map(unsigned long coded_mem_map,
+ unsigned long pnum);
#endif /* __LINUX_MEMORY_HOTPLUG_H */
diff -puN include/linux/mmzone.h~memory-hotplug-register-section-node-id-to-free include/linux/mmzone.h
--- a/include/linux/mmzone.h~memory-hotplug-register-section-node-id-to-free
+++ a/include/linux/mmzone.h
@@ -879,6 +879,7 @@ static inline struct mem_section *__nr_t
return &mem_section[SECTION_NR_TO_ROOT(nr)][nr & SECTION_ROOT_MASK];
}
extern int __section_nr(struct mem_section* ms);
+extern unsigned long usemap_size(void);
/*
* We use the lower bits of the mem_map pointer to store
diff -puN mm/bootmem.c~memory-hotplug-register-section-node-id-to-free mm/bootmem.c
--- a/mm/bootmem.c~memory-hotplug-register-section-node-id-to-free
+++ a/mm/bootmem.c
@@ -458,6 +458,7 @@ void __init free_bootmem_node(pg_data_t
unsigned long __init free_all_bootmem_node(pg_data_t *pgdat)
{
+ register_page_bootmem_info_node(pgdat);
return free_all_bootmem_core(pgdat);
}
diff -puN mm/memory_hotplug.c~memory-hotplug-register-section-node-id-to-free mm/memory_hotplug.c
--- a/mm/memory_hotplug.c~memory-hotplug-register-section-node-id-to-free
+++ a/mm/memory_hotplug.c
@@ -58,8 +58,103 @@ static void release_memory_resource(stru
return;
}
-
#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
+static void get_page_bootmem(unsigned long info, struct page *page, int magic)
+{
+ atomic_set(&page->_mapcount, magic);
+ SetPagePrivate(page);
+ set_page_private(page, info);
+ atomic_inc(&page->_count);
+}
+
+void put_page_bootmem(struct page *page)
+{
+ int magic;
+
+ magic = atomic_read(&page->_mapcount);
+ BUG_ON(magic >= -1);
+
+ if (atomic_dec_return(&page->_count) == 1) {
+ ClearPagePrivate(page);
+ set_page_private(page, 0);
+ reset_page_mapcount(page);
+ __free_pages_bootmem(page, 0);
+ }
+
+}
+
+void register_page_bootmem_info_section(unsigned long start_pfn)
+{
+ unsigned long *usemap, mapsize, section_nr, i;
+ struct mem_section *ms;
+ struct page *page, *memmap;
+
+ if (!pfn_valid(start_pfn))
+ return;
+
+ section_nr = pfn_to_section_nr(start_pfn);
+ ms = __nr_to_section(section_nr);
+
+ /* Get section's memmap address */
+ memmap = sparse_decode_mem_map(ms->section_mem_map, section_nr);
+
+ /*
+ * Get page for the memmap's phys address
+ * XXX: need more consideration for sparse_vmemmap...
+ */
+ page = virt_to_page(memmap);
+ mapsize = sizeof(struct page) * PAGES_PER_SECTION;
+ mapsize = PAGE_ALIGN(mapsize) >> PAGE_SHIFT;
+
+ /* remember memmap's page */
+ for (i = 0; i < mapsize; i++, page++)
+ get_page_bootmem(section_nr, page, SECTION_INFO);
+
+ usemap = __nr_to_section(section_nr)->pageblock_flags;
+ page = virt_to_page(usemap);
+
+ mapsize = PAGE_ALIGN(usemap_size()) >> PAGE_SHIFT;
+
+ for (i = 0; i < mapsize; i++, page++)
+ get_page_bootmem(section_nr, page, MIX_INFO);
+
+}
+
+void register_page_bootmem_info_node(struct pglist_data *pgdat)
+{
+ unsigned long i, pfn, end_pfn, nr_pages;
+ int node = pgdat->node_id;
+ struct page *page;
+ struct zone *zone;
+
+ nr_pages = PAGE_ALIGN(sizeof(struct pglist_data)) >> PAGE_SHIFT;
+ page = virt_to_page(pgdat);
+
+ for (i = 0; i < nr_pages; i++, page++)
+ get_page_bootmem(node, page, NODE_INFO);
+
+ zone = &pgdat->node_zones[0];
+ for (; zone < pgdat->node_zones + MAX_NR_ZONES - 1; zone++) {
+ if (zone->wait_table) {
+ nr_pages = zone->wait_table_hash_nr_entries
+ * sizeof(wait_queue_head_t);
+ nr_pages = PAGE_ALIGN(nr_pages) >> PAGE_SHIFT;
+ page = virt_to_page(zone->wait_table);
+
+ for (i = 0; i < nr_pages; i++, page++)
+ get_page_bootmem(node, page, NODE_INFO);
+ }
+ }
+
+ pfn = pgdat->node_start_pfn;
+ end_pfn = pfn + pgdat->node_spanned_pages;
+
+ /* register_section info */
+ for (; pfn < end_pfn; pfn += PAGES_PER_SECTION)
+ register_page_bootmem_info_section(pfn);
+
+}
+
static int __add_zone(struct zone *zone, unsigned long phys_start_pfn)
{
struct pglist_data *pgdat = zone->zone_pgdat;
diff -puN mm/sparse.c~memory-hotplug-register-section-node-id-to-free mm/sparse.c
--- a/mm/sparse.c~memory-hotplug-register-section-node-id-to-free
+++ a/mm/sparse.c
@@ -200,7 +200,6 @@ static unsigned long sparse_encode_mem_m
/*
* Decode mem_map from the coded memmap
*/
-static
struct page *sparse_decode_mem_map(unsigned long coded_mem_map, unsigned long pnum)
{
/* mask off the extra low bits of information */
@@ -223,7 +222,7 @@ static int __meminit sparse_init_one_sec
return 1;
}
-static unsigned long usemap_size(void)
+unsigned long usemap_size(void)
{
unsigned long size_bytes;
size_bytes = roundup(SECTION_BLOCKFLAGS_BITS, 8) / 8;
_
Patches currently in -mm which might be from y-goto@xxxxxxxxxxxxxx are
hotplug-memory-remove-generic-__remove_pages-support.patch
powerpc-hotplug-memory-notifications-for-ppc.patch
powerpc-update-lmb-for-hotplug-memory-add-remove.patch
powerpc-provide-walk_memory_resource-for-ppc.patch
block-fix-memory-hotplug-and-bouncing-in-block-layer.patch
mm-make-mem_map-allocation-continuous-v2.patch
mm-fix-alloc_bootmem_core-to-use-fast-searching-for-all-nodes.patch
mm-offset-align-in-alloc_bootmem.patch
mm-make-reserve_bootmem-can-crossed-the-nodes.patch
memory-hotplug-register-section-node-id-to-free.patch
memory-hotplug-align-memmap-to-page-size.patch
memory-hotplug-create-alloc_bootmem_section.patch
memory-hotplug-allocate-usemap-on-the-section-with-pgdat.patch
memory-hotplug-free-memmaps-allocated-by-bootmem.patch
ipc-scale-msgmni-to-the-amount-of-lowmem.patch
ipc-scale-msgmni-to-the-number-of-ipc-namespaces.patch
ipc-define-the-slab_memory_callback-priority-as-a-constant.patch
ipc-recompute-msgmni-on-memory-add--remove.patch
ipc-invoke-the-ipcns-notifier-chain-as-a-work-item.patch
ipc-recompute-msgmni-on-ipc-namespace-creation-removal.patch
ipc-do-not-recompute-msgmni-anymore-if-explicitly-set-by-user.patch
ipc-re-enable-msgmni-automatic-recomputing-msgmni-if-set-to-negative.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
