[PATCH] BLOCK: Dissociate generic_writepages() from mpage stuff [try #6]

Dissociate the generic_writepages() function from the mpage stuff, moving its declaration to linux/mm.h and actually emitting a full implementation into mm/page-writeback.c. The implementation is a partial duplicate of mpage_writepages() with all BIO references removed. It is used by NFS to do writeback. Signed-Off-By: David Howells <dhowells@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>

[PATCH] BLOCK: Dissociate generic_writepages() from mpage stuff [try #6]
Dissociate the generic_writepages() function from the mpage stuff, moving its declaration to linux/mm.h and actually emitting a full implementation into mm/page-writeback.c. The implementation is a partial duplicate of mpage_writepages() with all BIO references removed. It is used by NFS to do writeback. Signed-Off-By: David Howells <dhowells@redhat.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
David Howells · Jens Axboe
1 parent 7b0de42d7c
Showing 5 changed files with 139 additions and 6 deletions Side-by-side Diff
fs/block_dev.c
fs/mpage.c
include/linux/mpage.h
include/linux/writeback.h
mm/page-writeback.c
@@ -17,6 +17,7 @@
 #include <linux/module.h>
 #include <linux/blkpg.h>
 #include <linux/buffer_head.h>
+#include <linux/writeback.h>
 #include <linux/mpage.h>
 #include <linux/mount.h>
 #include <linux/uio.h>
@@ -693,6 +693,8 @@
  * the call was made get new I/O started against them.  If wbc->sync_mode is
  * WB_SYNC_ALL then we were called for data integrity and we must wait for
  * existing IO to complete.
+ *
+ * If you fix this you should check generic_writepages() also!
  */
 int
 mpage_writepages(struct address_space *mapping,
@@ -20,10 +20,4 @@
 		struct writeback_control *wbc, get_block_t get_block);
 int mpage_writepage(struct page *page, get_block_t *get_block,
 		struct writeback_control *wbc);
-
-static inline int
-generic_writepages(struct address_space *mapping, struct writeback_control *wbc)
-{
-	return mpage_writepages(mapping, wbc, NULL);
-}
@@ -111,6 +111,8 @@
 }
  
 int pdflush_operation(void (*fn)(unsigned long), unsigned long arg0);
+extern int generic_writepages(struct address_space *mapping,
+			      struct writeback_control *wbc);
 int do_writepages(struct address_space *mapping, struct writeback_control *wbc);
 int sync_page_range(struct inode *inode, struct address_space *mapping,
 			loff_t pos, loff_t count);
@@ -31,6 +31,7 @@
 #include <linux/cpu.h>
 #include <linux/syscalls.h>
 #include <linux/buffer_head.h>
+#include <linux/pagevec.h>
  
 /*
  * The maximum number of pages to writeout in a single bdflush/kupdate
@@ -550,6 +551,139 @@
 	writeback_set_ratelimit();
 	register_cpu_notifier(&ratelimit_nb);
 }
+
+/**
+ * generic_writepages - walk the list of dirty pages of the given
+ *                      address space and writepage() all of them.
+ *
+ * @mapping: address space structure to write
+ * @wbc: subtract the number of written pages from *@wbc->nr_to_write
+ *
+ * This is a library function, which implements the writepages()
+ * address_space_operation.
+ *
+ * If a page is already under I/O, generic_writepages() skips it, even
+ * if it's dirty.  This is desirable behaviour for memory-cleaning writeback,
+ * but it is INCORRECT for data-integrity system calls such as fsync().  fsync()
+ * and msync() need to guarantee that all the data which was dirty at the time
+ * the call was made get new I/O started against them.  If wbc->sync_mode is
+ * WB_SYNC_ALL then we were called for data integrity and we must wait for
+ * existing IO to complete.
+ *
+ * Derived from mpage_writepages() - if you fix this you should check that
+ * also!
+ */
+int generic_writepages(struct address_space *mapping,
+		       struct writeback_control *wbc)
+{
+	struct backing_dev_info *bdi = mapping->backing_dev_info;
+	int ret = 0;
+	int done = 0;
+	int (*writepage)(struct page *page, struct writeback_control *wbc);
+	struct pagevec pvec;
+	int nr_pages;
+	pgoff_t index;
+	pgoff_t end;		/* Inclusive */
+	int scanned = 0;
+	int range_whole = 0;
+
+	if (wbc->nonblocking && bdi_write_congested(bdi)) {
+		wbc->encountered_congestion = 1;
+		return 0;
+	}
+
+	writepage = mapping->a_ops->writepage;
+
+	/* deal with chardevs and other special file */
+	if (!writepage)
+		return 0;
+
+	pagevec_init(&pvec, 0);
+	if (wbc->range_cyclic) {
+		index = mapping->writeback_index; /* Start from prev offset */
+		end = -1;
+	} else {
+		index = wbc->range_start >> PAGE_CACHE_SHIFT;
+		end = wbc->range_end >> PAGE_CACHE_SHIFT;
+		if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
+			range_whole = 1;
+		scanned = 1;
+	}
+retry:
+	while (!done && (index <= end) &&
+	       (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
+					      PAGECACHE_TAG_DIRTY,
+					      min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
+		unsigned i;
+
+		scanned = 1;
+		for (i = 0; i < nr_pages; i++) {
+			struct page *page = pvec.pages[i];
+
+			/*
+			 * At this point we hold neither mapping->tree_lock nor
+			 * lock on the page itself: the page may be truncated or
+			 * invalidated (changing page->mapping to NULL), or even
+			 * swizzled back from swapper_space to tmpfs file
+			 * mapping
+			 */
+			lock_page(page);
+
+			if (unlikely(page->mapping != mapping)) {
+				unlock_page(page);
+				continue;
+			}
+
+			if (!wbc->range_cyclic && page->index > end) {
+				done = 1;
+				unlock_page(page);
+				continue;
+			}
+
+			if (wbc->sync_mode != WB_SYNC_NONE)
+				wait_on_page_writeback(page);
+
+			if (PageWriteback(page) ||
+			    !clear_page_dirty_for_io(page)) {
+				unlock_page(page);
+				continue;
+			}
+
+			ret = (*writepage)(page, wbc);
+			if (ret) {
+				if (ret == -ENOSPC)
+					set_bit(AS_ENOSPC, &mapping->flags);
+				else
+					set_bit(AS_EIO, &mapping->flags);
+			}
+
+			if (unlikely(ret == AOP_WRITEPAGE_ACTIVATE))
+				unlock_page(page);
+			if (ret || (--(wbc->nr_to_write) <= 0))
+				done = 1;
+			if (wbc->nonblocking && bdi_write_congested(bdi)) {
+				wbc->encountered_congestion = 1;
+				done = 1;
+			}
+		}
+		pagevec_release(&pvec);
+		cond_resched();
+	}
+	if (!scanned && !done) {
+		/*
+		 * We hit the last page and there is more work to be done: wrap
+		 * back to the start of the file
+		 */
+		scanned = 1;
+		index = 0;
+		goto retry;
+	}
+	if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
+		mapping->writeback_index = index;
+	return ret;
+}
+
+EXPORT_SYMBOL(generic_writepages);
  
 int do_writepages(struct address_space *mapping, struct writeback_control *wbc)
 {
...	...	@@ -17,6 +17,7 @@
17	17	#include <linux/module.h>
18	18	#include <linux/blkpg.h>
19	19	#include <linux/buffer_head.h>
	20	+#include <linux/writeback.h>
20	21	#include <linux/mpage.h>
21	22	#include <linux/mount.h>
22	23	#include <linux/uio.h>
...	...	@@ -693,6 +693,8 @@
693	693	* the call was made get new I/O started against them. If wbc->sync_mode is
694	694	* WB_SYNC_ALL then we were called for data integrity and we must wait for
695	695	* existing IO to complete.
	696	+ *
	697	+ * If you fix this you should check generic_writepages() also!
696	698	*/
697	699	int
698	700	mpage_writepages(struct address_space *mapping,
...	...	@@ -20,10 +20,4 @@
20	20	struct writeback_control *wbc, get_block_t get_block);
21	21	int mpage_writepage(struct page page, get_block_t get_block,
22	22	struct writeback_control *wbc);
23		-
24		-static inline int
25		-generic_writepages(struct address_space mapping, struct writeback_control wbc)
26		-{
27		- return mpage_writepages(mapping, wbc, NULL);
28		-}
...	...	@@ -111,6 +111,8 @@
111	111	}
112	112
113	113	int pdflush_operation(void (*fn)(unsigned long), unsigned long arg0);
	114	+extern int generic_writepages(struct address_space *mapping,
	115	+ struct writeback_control *wbc);
114	116	int do_writepages(struct address_space mapping, struct writeback_control wbc);
115	117	int sync_page_range(struct inode inode, struct address_space mapping,
116	118	loff_t pos, loff_t count);
...	...	@@ -31,6 +31,7 @@
31	31	#include <linux/cpu.h>
32	32	#include <linux/syscalls.h>
33	33	#include <linux/buffer_head.h>
	34	+#include <linux/pagevec.h>
34	35
35	36	/*
36	37	* The maximum number of pages to writeout in a single bdflush/kupdate
...	...	@@ -550,6 +551,139 @@
550	551	writeback_set_ratelimit();
551	552	register_cpu_notifier(&ratelimit_nb);
552	553	}
	554	+
	555	+/**
	556	+ * generic_writepages - walk the list of dirty pages of the given
	557	+ * address space and writepage() all of them.
	558	+ *
	559	+ * @mapping: address space structure to write
	560	+ * @wbc: subtract the number of written pages from *@wbc->nr_to_write
	561	+ *
	562	+ * This is a library function, which implements the writepages()
	563	+ * address_space_operation.
	564	+ *
	565	+ * If a page is already under I/O, generic_writepages() skips it, even
	566	+ * if it's dirty. This is desirable behaviour for memory-cleaning writeback,
	567	+ * but it is INCORRECT for data-integrity system calls such as fsync(). fsync()
	568	+ * and msync() need to guarantee that all the data which was dirty at the time
	569	+ * the call was made get new I/O started against them. If wbc->sync_mode is
	570	+ * WB_SYNC_ALL then we were called for data integrity and we must wait for
	571	+ * existing IO to complete.
	572	+ *
	573	+ * Derived from mpage_writepages() - if you fix this you should check that
	574	+ * also!
	575	+ */
	576	+int generic_writepages(struct address_space *mapping,
	577	+ struct writeback_control *wbc)
	578	+{
	579	+ struct backing_dev_info *bdi = mapping->backing_dev_info;
	580	+ int ret = 0;
	581	+ int done = 0;
	582	+ int (writepage)(struct page page, struct writeback_control *wbc);
	583	+ struct pagevec pvec;
	584	+ int nr_pages;
	585	+ pgoff_t index;
	586	+ pgoff_t end; /* Inclusive */
	587	+ int scanned = 0;
	588	+ int range_whole = 0;
	589	+
	590	+ if (wbc->nonblocking && bdi_write_congested(bdi)) {
	591	+ wbc->encountered_congestion = 1;
	592	+ return 0;
	593	+ }
	594	+
	595	+ writepage = mapping->a_ops->writepage;
	596	+
	597	+ /* deal with chardevs and other special file */
	598	+ if (!writepage)
	599	+ return 0;
	600	+
	601	+ pagevec_init(&pvec, 0);
	602	+ if (wbc->range_cyclic) {
	603	+ index = mapping->writeback_index; /* Start from prev offset */
	604	+ end = -1;
	605	+ } else {
	606	+ index = wbc->range_start >> PAGE_CACHE_SHIFT;
	607	+ end = wbc->range_end >> PAGE_CACHE_SHIFT;
	608	+ if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
	609	+ range_whole = 1;
	610	+ scanned = 1;
	611	+ }
	612	+retry:
	613	+ while (!done && (index <= end) &&
	614	+ (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
	615	+ PAGECACHE_TAG_DIRTY,
	616	+ min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
	617	+ unsigned i;
	618	+
	619	+ scanned = 1;
	620	+ for (i = 0; i < nr_pages; i++) {
	621	+ struct page *page = pvec.pages[i];
	622	+
	623	+ /*
	624	+ * At this point we hold neither mapping->tree_lock nor
	625	+ * lock on the page itself: the page may be truncated or
	626	+ * invalidated (changing page->mapping to NULL), or even
	627	+ * swizzled back from swapper_space to tmpfs file
	628	+ * mapping
	629	+ */
	630	+ lock_page(page);
	631	+
	632	+ if (unlikely(page->mapping != mapping)) {
	633	+ unlock_page(page);
	634	+ continue;
	635	+ }
	636	+
	637	+ if (!wbc->range_cyclic && page->index > end) {
	638	+ done = 1;
	639	+ unlock_page(page);
	640	+ continue;
	641	+ }
	642	+
	643	+ if (wbc->sync_mode != WB_SYNC_NONE)
	644	+ wait_on_page_writeback(page);
	645	+
	646	+ if (PageWriteback(page) \|\|
	647	+ !clear_page_dirty_for_io(page)) {
	648	+ unlock_page(page);
	649	+ continue;
	650	+ }
	651	+
	652	+ ret = (*writepage)(page, wbc);
	653	+ if (ret) {
	654	+ if (ret == -ENOSPC)
	655	+ set_bit(AS_ENOSPC, &mapping->flags);
	656	+ else
	657	+ set_bit(AS_EIO, &mapping->flags);
	658	+ }
	659	+
	660	+ if (unlikely(ret == AOP_WRITEPAGE_ACTIVATE))
	661	+ unlock_page(page);
	662	+ if (ret \|\| (--(wbc->nr_to_write) <= 0))
	663	+ done = 1;
	664	+ if (wbc->nonblocking && bdi_write_congested(bdi)) {
	665	+ wbc->encountered_congestion = 1;
	666	+ done = 1;
	667	+ }
	668	+ }
	669	+ pagevec_release(&pvec);
	670	+ cond_resched();
	671	+ }
	672	+ if (!scanned && !done) {
	673	+ /*
	674	+ * We hit the last page and there is more work to be done: wrap
	675	+ * back to the start of the file
	676	+ */
	677	+ scanned = 1;
	678	+ index = 0;
	679	+ goto retry;
	680	+ }
	681	+ if (wbc->range_cyclic \|\| (range_whole && wbc->nr_to_write > 0))
	682	+ mapping->writeback_index = index;
	683	+ return ret;
	684	+}
	685	+
	686	+EXPORT_SYMBOL(generic_writepages);
553	687
554	688	int do_writepages(struct address_space mapping, struct writeback_control wbc)
555	689	{