[RFC] mm/hotplug: Make get_nid_for_pfn() work with HAVE_ARCH_PFN_VALID

From: Anshuman Khandual <anshuman.khandual@arm.com>
To: linux-mm@kvack.org, linux-kernel@vger.kernel.org
Cc: logang@deltatee.com, osalvador@suse.de, hannes@cmpxchg.org,
	mhocko@suse.com, akpm@linux-foundation.org,
	richard.weiyang@gmail.com, rientjes@google.com,
	zi.yan@cs.rutgers.edu
Subject: [RFC] mm/hotplug: Make get_nid_for_pfn() work with HAVE_ARCH_PFN_VALID
Date: Thu, 21 Mar 2019 13:38:20 +0530	[thread overview]
Message-ID: <1553155700-3414-1-git-send-email-anshuman.khandual@arm.com> (raw)

Memory hot remove uses get_nid_for_pfn() while tearing down linked sysfs
entries between memory block and node. It first checks pfn validity with
pfn_valid_within() before fetching nid. With CONFIG_HOLES_IN_ZONE config
(arm64 has this enabled) pfn_valid_within() calls pfn_valid().

pfn_valid() is an arch implementation on arm64 (CONFIG_HAVE_ARCH_PFN_VALID)
which scans all mapped memblock regions with memblock_is_map_memory(). This
creates a problem in memory hot remove path which has already removed given
memory range from memory block with memblock_[remove|free] before arriving
at unregister_mem_sect_under_nodes().

During runtime memory hot remove get_nid_for_pfn() needs to validate that
given pfn has a struct page mapping so that it can fetch required nid. This
can be achieved just by looking into it's section mapping information. This
adds a new helper pfn_section_valid() for this purpose. Its same as generic
pfn_valid().

This maintains existing behaviour for deferred struct page init case.

Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
---
This is a preparatory patch for memory hot-remove enablement on arm64. I
will appreciate some early feedback on this approach.

 drivers/base/node.c    | 15 ++++++++++++---
 include/linux/mmzone.h |  9 +++++++--
 2 files changed, 19 insertions(+), 5 deletions(-)

diff --git a/drivers/base/node.c b/drivers/base/node.c
index 86d6cd92ce3d..9e944b71e352 100644
--- a/drivers/base/node.c
+++ b/drivers/base/node.c
@@ -394,11 +394,20 @@ int unregister_cpu_under_node(unsigned int cpu, unsigned int nid)
 #ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
 static int __ref get_nid_for_pfn(unsigned long pfn)
 {
-	if (!pfn_valid_within(pfn))
-		return -1;
 #ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
-	if (system_state < SYSTEM_RUNNING)
+	if (system_state < SYSTEM_RUNNING) {
+		if (!pfn_valid_within(pfn))
+			return -1;
 		return early_pfn_to_nid(pfn);
+	}
+#endif
+
+#if defined(CONFIG_HAVE_ARCH_PFN_VALID) && defined(CONFIG_HOLES_IN_ZONE)
+	if (!pfn_section_valid(pfn))
+		return -1;
+#else
+	if (!pfn_valid_within(pfn))
+		return -1;
 #endif
 	return pfn_to_nid(pfn);
 }
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 842f9189537b..9cf4c1111b95 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -1242,13 +1242,18 @@ static inline struct mem_section *__pfn_to_section(unsigned long pfn)
 
 extern int __highest_present_section_nr;
 
-#ifndef CONFIG_HAVE_ARCH_PFN_VALID
-static inline int pfn_valid(unsigned long pfn)
+static inline int pfn_section_valid(unsigned long pfn)
 {
 	if (pfn_to_section_nr(pfn) >= NR_MEM_SECTIONS)
 		return 0;
 	return valid_section(__nr_to_section(pfn_to_section_nr(pfn)));
 }
+
+#ifndef CONFIG_HAVE_ARCH_PFN_VALID
+static inline int pfn_valid(unsigned long pfn)
+{
+	return pfn_section_valid(pfn);
+}
 #endif
 
 static inline int pfn_present(unsigned long pfn)
-- 
2.20.1

