Linux-mm Archive on lore.kernel.org
 help / color / Atom feed
From: Matthew Wilcox <willy@infradead.org>
To: linux-mm@kvack.org
Cc: Andrew Morton <akpm@linux-foundation.org>,
	Hugh Dickins <hughd@google.com>,
	William Kucharski <william.kucharski@oracle.com>,
	Jani Nikula <jani.nikula@linux.intel.com>,
	Alexey Dobriyan <adobriyan@gmail.com>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Chris Wilson <chris@chris-wilson.co.uk>,
	Matthew Auld <matthew.auld@intel.com>,
	Huang Ying <ying.huang@intel.com>,
	intel-gfx@lists.freedesktop.org, cgroups@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 7/8] mm/shmem: Return head page from find_lock_entry
Date: Sat, 12 Sep 2020 04:20:42 +0100
Message-ID: <20200912032042.GA6583@casper.infradead.org> (raw)
In-Reply-To: <20200910183318.20139-8-willy@infradead.org>

On Thu, Sep 10, 2020 at 07:33:17PM +0100, Matthew Wilcox (Oracle) wrote:
> Convert shmem_getpage_gfp() (the only remaining caller of
> find_lock_entry()) to cope with a head page being returned instead of
> the subpage for the index.

This version was buggy.  Apparently I was too focused on running the test suite against XFS and neglected to run it against tmpfs, which crashed instantly.

Here's the patch I should have sent.

commit 7bfa655881da76f3386e6d4c07e38a165b4a6ca8
Author: Matthew Wilcox (Oracle) <willy@infradead.org>
Date:   Sun Aug 2 07:22:34 2020 -0400

    mm/shmem: Return head page from find_lock_entry
    
    Convert shmem_getpage_gfp() (the only remaining caller of
    find_lock_entry()) to cope with a head page being returned instead of
    the subpage for the index.
    
    Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>

diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index 905a64030647..f374618b2c93 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -371,6 +371,15 @@ static inline struct page *grab_cache_page_nowait(struct address_space *mapping,
 			mapping_gfp_mask(mapping));
 }
 
+/* Does this page contain this index? */
+static inline bool thp_contains(struct page *head, pgoff_t index)
+{
+	/* HugeTLBfs indexes the page cache in units of hpage_size */
+	if (PageHuge(head))
+		return head->index == index;
+	return page_index(head) == (index & ~(thp_nr_pages(head) - 1UL));
+}
+
 /*
  * Given the page we found in the page cache, return the page corresponding
  * to this index in the file
diff --git a/mm/filemap.c b/mm/filemap.c
index 2f134383b0ae..453535170b8d 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1614,37 +1614,34 @@ struct page *find_get_entry(struct address_space *mapping, pgoff_t index)
 }
 
 /**
- * find_lock_entry - locate, pin and lock a page cache entry
- * @mapping: the address_space to search
- * @offset: the page cache index
+ * find_lock_entry - Locate and lock a page cache entry.
+ * @mapping: The address_space to search.
+ * @index: The page cache index.
  *
- * Looks up the page cache slot at @mapping & @offset.  If there is a
- * page cache page, it is returned locked and with an increased
- * refcount.
+ * Looks up the page at @mapping & @index.  If there is a page in the
+ * cache, the head page is returned locked and with an increased refcount.
  *
  * If the slot holds a shadow entry of a previously evicted page, or a
  * swap entry from shmem/tmpfs, it is returned.
  *
- * find_lock_entry() may sleep.
- *
- * Return: the found page or shadow entry, %NULL if nothing is found.
+ * Context: May sleep.
+ * Return: The head page or shadow entry, %NULL if nothing is found.
  */
-struct page *find_lock_entry(struct address_space *mapping, pgoff_t offset)
+struct page *find_lock_entry(struct address_space *mapping, pgoff_t index)
 {
 	struct page *page;
 
 repeat:
-	page = find_get_entry(mapping, offset);
+	page = find_get_entry(mapping, index);
 	if (page && !xa_is_value(page)) {
 		lock_page(page);
 		/* Has the page been truncated? */
-		if (unlikely(page_mapping(page) != mapping)) {
+		if (unlikely(page->mapping != mapping)) {
 			unlock_page(page);
 			put_page(page);
 			goto repeat;
 		}
-		page = find_subpage(page, offset);
-		VM_BUG_ON_PAGE(page_to_pgoff(page) != offset, page);
+		VM_BUG_ON_PAGE(!thp_contains(page, index), page);
 	}
 	return page;
 }
diff --git a/mm/shmem.c b/mm/shmem.c
index 271548ca20f3..58bc9e326d0d 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1822,6 +1822,8 @@ static int shmem_getpage_gfp(struct inode *inode, pgoff_t index,
 		return error;
 	}
 
+	if (page)
+		hindex = page->index;
 	if (page && sgp == SGP_WRITE)
 		mark_page_accessed(page);
 
@@ -1832,11 +1834,10 @@ static int shmem_getpage_gfp(struct inode *inode, pgoff_t index,
 		unlock_page(page);
 		put_page(page);
 		page = NULL;
+		hindex = index;
 	}
-	if (page || sgp == SGP_READ) {
-		*pagep = page;
-		return 0;
-	}
+	if (page || sgp == SGP_READ)
+		goto out;
 
 	/*
 	 * Fast cache lookup did not find it:
@@ -1961,14 +1962,13 @@ static int shmem_getpage_gfp(struct inode *inode, pgoff_t index,
 	 * it now, lest undo on failure cancel our earlier guarantee.
 	 */
 	if (sgp != SGP_WRITE && !PageUptodate(page)) {
-		struct page *head = compound_head(page);
 		int i;
 
-		for (i = 0; i < compound_nr(head); i++) {
-			clear_highpage(head + i);
-			flush_dcache_page(head + i);
+		for (i = 0; i < compound_nr(page); i++) {
+			clear_highpage(page + i);
+			flush_dcache_page(page + i);
 		}
-		SetPageUptodate(head);
+		SetPageUptodate(page);
 	}
 
 	/* Perhaps the file has been truncated since we checked */
@@ -1984,6 +1984,7 @@ static int shmem_getpage_gfp(struct inode *inode, pgoff_t index,
 		error = -EINVAL;
 		goto unlock;
 	}
+out:
 	*pagep = page + index - hindex;
 	return 0;
 


  reply index

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-10 18:33 [PATCH v2 0/8] Return head pages from find_*_entry Matthew Wilcox (Oracle)
2020-09-10 18:33 ` [PATCH v2 1/8] mm: Factor find_get_incore_page out of mincore_page Matthew Wilcox (Oracle)
2020-09-10 18:33 ` [PATCH v2 2/8] mm: Use find_get_incore_page in memcontrol Matthew Wilcox (Oracle)
2020-09-10 18:33 ` [PATCH v2 3/8] mm: Optimise madvise WILLNEED Matthew Wilcox (Oracle)
2020-09-14 16:17   ` Qian Cai
2020-09-14 16:47     ` Qian Cai
2020-09-14 16:50     ` Matthew Wilcox
2020-09-14 19:44       ` Qian Cai
2020-09-10 18:33 ` [PATCH v2 4/8] proc: Optimise smaps for shmem entries Matthew Wilcox (Oracle)
2020-09-10 18:33 ` [PATCH v2 5/8] i915: Use find_lock_page instead of find_lock_entry Matthew Wilcox (Oracle)
2020-09-10 18:33 ` [PATCH v2 6/8] mm: Convert find_get_entry to return the head page Matthew Wilcox (Oracle)
2020-09-14  8:55   ` [mm] 2037ab69a5: BUG:KASAN:null-ptr-deref_in_t kernel test robot
2020-09-14 11:27     ` Matthew Wilcox
2020-09-10 18:33 ` [PATCH v2 7/8] mm/shmem: Return head page from find_lock_entry Matthew Wilcox (Oracle)
2020-09-12  3:20   ` Matthew Wilcox [this message]
2020-09-10 18:33 ` [PATCH v2 8/8] mm: Add find_lock_head Matthew Wilcox (Oracle)
2020-09-15  8:10 ` [PATCH v2 0/8] Return head pages from find_*_entry Hugh Dickins
2020-09-15 12:53   ` Naresh Kamboju
2020-09-15 13:08     ` Matthew Wilcox

Reply instructions:

You may reply publicly 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=20200912032042.GA6583@casper.infradead.org \
    --to=willy@infradead.org \
    --cc=adobriyan@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=cgroups@vger.kernel.org \
    --cc=chris@chris-wilson.co.uk \
    --cc=hannes@cmpxchg.org \
    --cc=hughd@google.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=jani.nikula@linux.intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=matthew.auld@intel.com \
    --cc=william.kucharski@oracle.com \
    --cc=ying.huang@intel.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