Linux-mm Archive on lore.kernel.org
 help / color / Atom feed
From: Oscar Salvador <osalvador@suse.de>
To: akpm@linux-foundation.org
Cc: dan.j.williams@intel.com, david@redhat.com,
	pasha.tatashin@soleen.com, mhocko@suse.com,
	aneesh.kumar@linux.ibm.com, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org, Oscar Salvador <osalvador@suse.de>
Subject: [PATCH 2/2] mm,memory_hotplug: Fix shrink_{zone,node}_span
Date: Mon, 15 Jul 2019 10:15:49 +0200
Message-ID: <20190715081549.32577-3-osalvador@suse.de> (raw)
In-Reply-To: <20190715081549.32577-1-osalvador@suse.de>

Since [1], shrink_{zone,node}_span work on PAGES_PER_SUBSECTION granularity.
The problem is that deactivation of the section occurs later on in
sparse_remove_section, so pfn_valid()->pfn_section_valid() will always return
true before we deactivate the {sub}section.

I spotted this during hotplug hotremove tests, there I always saw that
spanned_pages was, at least, left with PAGES_PER_SECTION, even if we
removed all memory linked to that zone.

Fix this by decoupling section_deactivate from sparse_remove_section, and
re-order the function calls.

Now, __remove_section will:

1) deactivate section
2) shrink {zone,node}'s pages
3) remove section

[1] https://patchwork.kernel.org/patch/11003467/

Fixes: mmotm ("mm/hotplug: prepare shrink_{zone, pgdat}_span for sub-section removal")
Signed-off-by: Oscar Salvador <osalvador@suse.de>
---
 include/linux/memory_hotplug.h |  7 ++--
 mm/memory_hotplug.c            |  6 +++-
 mm/sparse.c                    | 77 +++++++++++++++++++++++++++++-------------
 3 files changed, 62 insertions(+), 28 deletions(-)

diff --git a/include/linux/memory_hotplug.h b/include/linux/memory_hotplug.h
index f46ea71b4ffd..d2eb917aad5f 100644
--- a/include/linux/memory_hotplug.h
+++ b/include/linux/memory_hotplug.h
@@ -348,9 +348,10 @@ extern void move_pfn_range_to_zone(struct zone *zone, unsigned long start_pfn,
 extern bool is_memblock_offlined(struct memory_block *mem);
 extern int sparse_add_section(int nid, unsigned long pfn,
 		unsigned long nr_pages, struct vmem_altmap *altmap);
-extern void sparse_remove_section(struct mem_section *ms,
-		unsigned long pfn, unsigned long nr_pages,
-		unsigned long map_offset, struct vmem_altmap *altmap);
+int sparse_deactivate_section(unsigned long pfn, unsigned long nr_pages);
+void sparse_remove_section(unsigned long pfn, unsigned long nr_pages,
+                           unsigned long map_offset, struct vmem_altmap *altmap,
+                           int section_empty);
 extern struct page *sparse_decode_mem_map(unsigned long coded_mem_map,
 					  unsigned long pnum);
 extern bool allow_online_pfn_range(int nid, unsigned long pfn, unsigned long nr_pages,
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index b9ba5b85f9f7..03d535eee60d 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -517,12 +517,16 @@ static void __remove_section(struct zone *zone, unsigned long pfn,
 		struct vmem_altmap *altmap)
 {
 	struct mem_section *ms = __nr_to_section(pfn_to_section_nr(pfn));
+	int ret;
 
 	if (WARN_ON_ONCE(!valid_section(ms)))
 		return;
 
+	ret = sparse_deactivate_section(pfn, nr_pages);
 	__remove_zone(zone, pfn, nr_pages);
-	sparse_remove_section(ms, pfn, nr_pages, map_offset, altmap);
+	if (ret >= 0)
+		sparse_remove_section(pfn, nr_pages, map_offset, altmap,
+				      ret);
 }
 
 /**
diff --git a/mm/sparse.c b/mm/sparse.c
index 1e224149aab6..d4953ee1d087 100644
--- a/mm/sparse.c
+++ b/mm/sparse.c
@@ -732,16 +732,47 @@ static void free_map_bootmem(struct page *memmap)
 }
 #endif /* CONFIG_SPARSEMEM_VMEMMAP */
 
-static void section_deactivate(unsigned long pfn, unsigned long nr_pages,
-		struct vmem_altmap *altmap)
+static void section_remove(unsigned long pfn, unsigned long nr_pages,
+			   struct vmem_altmap *altmap, int section_empty)
+{
+	struct mem_section *ms = __pfn_to_section(pfn);
+	bool section_early = early_section(ms);
+	struct page *memmap = NULL;
+
+	if (section_empty) {
+		unsigned long section_nr = pfn_to_section_nr(pfn);
+
+		if (!section_early) {
+			kfree(ms->usage);
+			ms->usage = NULL;
+		}
+		memmap = sparse_decode_mem_map(ms->section_mem_map, section_nr);
+		ms->section_mem_map = sparse_encode_mem_map(NULL, section_nr);
+	}
+
+        if (section_early && memmap)
+		free_map_bootmem(memmap);
+        else
+		depopulate_section_memmap(pfn, nr_pages, altmap);
+}
+
+/**
+ * section_deactivate: Deactivate a {sub}section.
+ *
+ * Return:
+ * * -1         - {sub}section has already been deactivated.
+ * * 0          - Section is not empty
+ * * 1          - Section is empty
+ */
+
+static int section_deactivate(unsigned long pfn, unsigned long nr_pages)
 {
 	DECLARE_BITMAP(map, SUBSECTIONS_PER_SECTION) = { 0 };
 	DECLARE_BITMAP(tmp, SUBSECTIONS_PER_SECTION) = { 0 };
 	struct mem_section *ms = __pfn_to_section(pfn);
-	bool section_is_early = early_section(ms);
-	struct page *memmap = NULL;
 	unsigned long *subsection_map = ms->usage
 		? &ms->usage->subsection_map[0] : NULL;
+	int section_empty = 0;
 
 	subsection_mask_set(map, pfn, nr_pages);
 	if (subsection_map)
@@ -750,7 +781,7 @@ static void section_deactivate(unsigned long pfn, unsigned long nr_pages,
 	if (WARN(!subsection_map || !bitmap_equal(tmp, map, SUBSECTIONS_PER_SECTION),
 				"section already deactivated (%#lx + %ld)\n",
 				pfn, nr_pages))
-		return;
+		return -1;
 
 	/*
 	 * There are 3 cases to handle across two configurations
@@ -770,21 +801,10 @@ static void section_deactivate(unsigned long pfn, unsigned long nr_pages,
 	 * For 2/ and 3/ the SPARSEMEM_VMEMMAP={y,n} cases are unified
 	 */
 	bitmap_xor(subsection_map, map, subsection_map, SUBSECTIONS_PER_SECTION);
-	if (bitmap_empty(subsection_map, SUBSECTIONS_PER_SECTION)) {
-		unsigned long section_nr = pfn_to_section_nr(pfn);
-
-		if (!section_is_early) {
-			kfree(ms->usage);
-			ms->usage = NULL;
-		}
-		memmap = sparse_decode_mem_map(ms->section_mem_map, section_nr);
-		ms->section_mem_map = sparse_encode_mem_map(NULL, section_nr);
-	}
+	if (bitmap_empty(subsection_map, SUBSECTIONS_PER_SECTION))
+		section_empty = 1;
 
-	if (section_is_early && memmap)
-		free_map_bootmem(memmap);
-	else
-		depopulate_section_memmap(pfn, nr_pages, altmap);
+	return section_empty;
 }
 
 static struct page * __meminit section_activate(int nid, unsigned long pfn,
@@ -834,7 +854,11 @@ static struct page * __meminit section_activate(int nid, unsigned long pfn,
 
 	memmap = populate_section_memmap(pfn, nr_pages, nid, altmap);
 	if (!memmap) {
-		section_deactivate(pfn, nr_pages, altmap);
+		int ret;
+
+		ret = section_deactivate(pfn, nr_pages);
+		if (ret >= 0)
+			section_remove(pfn, nr_pages, altmap, ret);
 		return ERR_PTR(-ENOMEM);
 	}
 
@@ -919,12 +943,17 @@ static inline void clear_hwpoisoned_pages(struct page *memmap, int nr_pages)
 }
 #endif
 
-void sparse_remove_section(struct mem_section *ms, unsigned long pfn,
-		unsigned long nr_pages, unsigned long map_offset,
-		struct vmem_altmap *altmap)
+int sparse_deactivate_section(unsigned long pfn, unsigned long nr_pages)
+{
+	return section_deactivate(pfn, nr_pages);
+}
+
+void sparse_remove_section(unsigned long pfn, unsigned long nr_pages,
+			   unsigned long map_offset, struct vmem_altmap *altmap,
+			   int section_empty)
 {
 	clear_hwpoisoned_pages(pfn_to_page(pfn) + map_offset,
 			nr_pages - map_offset);
-	section_deactivate(pfn, nr_pages, altmap);
+	section_remove(pfn, nr_pages, altmap, section_empty);
 }
 #endif /* CONFIG_MEMORY_HOTPLUG */
-- 
2.12.3


  parent reply index

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-15  8:15 [PATCH 0/2] Fixes for sub-section hotplug Oscar Salvador
2019-07-15  8:15 ` [PATCH 1/2] mm,sparse: Fix deactivate_section for early sections Oscar Salvador
2019-07-15 16:02   ` Aneesh Kumar K.V
2019-07-16  4:33   ` Dan Williams
2019-07-18 12:07   ` David Hildenbrand
2019-07-15  8:15 ` Oscar Salvador [this message]
2019-07-15 16:11   ` [PATCH 2/2] mm,memory_hotplug: Fix shrink_{zone,node}_span Aneesh Kumar K.V
2019-07-15 21:24     ` Oscar Salvador
2019-07-17  2:28       ` Dan Williams
2019-07-17  7:38         ` Oscar Salvador
2019-07-17  8:01           ` David Hildenbrand
2019-07-17  8:08             ` Oscar Salvador
2019-07-17  5:38       ` Aneesh Kumar K.V
2019-07-17  7:42         ` Oscar Salvador
2019-07-18 12:05   ` Oscar Salvador

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190715081549.32577-3-osalvador@suse.de \
    --to=osalvador@suse.de \
    --cc=akpm@linux-foundation.org \
    --cc=aneesh.kumar@linux.ibm.com \
    --cc=dan.j.williams@intel.com \
    --cc=david@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@suse.com \
    --cc=pasha.tatashin@soleen.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-mm Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-mm/0 linux-mm/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-mm linux-mm/ https://lore.kernel.org/linux-mm \
		linux-mm@kvack.org
	public-inbox-index linux-mm

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kvack.linux-mm


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git