memory hotplug: allow setting of phys_device

/sys/devices/system/memory/memoryX/phys_device is supposed to contain the
number of the physical device that the corresponding piece of memory
belongs to.

In case a physical device should be replaced or taken offline for whatever
reason it is necessary to set all corresponding memory pieces offline.
The current implementation always sets phys_device to '0' and there is no
way or hook to change that.  Seems like there was a plan to implement that
but it wasn't finished for whatever reason.

So add a weak function which architectures can override to actually set
the phys_device from within add_memory_block().

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/drivers/base/memory.c b/drivers/base/memory.c
index 2f86915..db0848e 100644 (file)
--- a/drivers/base/memory.c
+++ b/drivers/base/memory.c
@@ -429,12 +429,16 @@ static inline int memory_fail_init(void)
  * differentiation between which *physical* devices each
  * section belongs to...
  */
+int __weak arch_get_memory_phys_device(unsigned long start_pfn)
+{
+       return 0;
+}
 
 static int add_memory_block(int nid, struct mem_section *section,
-                       unsigned long state, int phys_device,
-                       enum mem_add_context context)
+                       unsigned long state, enum mem_add_context context)
 {
        struct memory_block *mem = kzalloc(sizeof(*mem), GFP_KERNEL);
+       unsigned long start_pfn;
        int ret = 0;
 
        if (!mem)
@@ -443,7 +447,8 @@ static int add_memory_block(int nid, struct mem_section *section,
        mem->phys_index = __section_nr(section);
        mem->state = state;
        mutex_init(&mem->state_mutex);
-       mem->phys_device = phys_device;
+       start_pfn = section_nr_to_pfn(mem->phys_index);
+       mem->phys_device = arch_get_memory_phys_device(start_pfn);
 
        ret = register_memory(mem, section);
        if (!ret)
@@ -515,7 +520,7 @@ int remove_memory_block(unsigned long node_id, struct mem_section *section,
  */
 int register_new_memory(int nid, struct mem_section *section)
 {
-       return add_memory_block(nid, section, MEM_OFFLINE, 0, HOTPLUG);
+       return add_memory_block(nid, section, MEM_OFFLINE, HOTPLUG);
 }
 
 int unregister_memory_section(struct mem_section *section)
@@ -548,7 +553,7 @@ int __init memory_dev_init(void)
                if (!present_section_nr(i))
                        continue;
                err = add_memory_block(0, __nr_to_section(i), MEM_ONLINE,
-                                       0, BOOT);
+                                      BOOT);
                if (!ret)
                        ret = err;
        }

diff --git a/include/linux/memory.h b/include/linux/memory.h
index 1adfe77..85582e1 100644 (file)
--- a/include/linux/memory.h
+++ b/include/linux/memory.h
@@ -36,6 +36,8 @@ struct memory_block {
        struct sys_device sysdev;
 };
 
+int arch_get_memory_phys_device(unsigned long start_pfn);
+
 /* These states are exposed to userspace as text strings in sysfs */
 #define        MEM_ONLINE              (1<<0) /* exposed to userspace */
 #define        MEM_GOING_OFFLINE       (1<<1) /* exposed to userspace */