diff mbox series

[RFC,V4] iomap: add support to track dirty state of sub pages

Message ID 20200821123306.1658495-1-yukuai3@huawei.com (mailing list archive)
State New
Headers show
Series [RFC,V4] iomap: add support to track dirty state of sub pages | expand

Commit Message

Yu Kuai Aug. 21, 2020, 12:33 p.m. UTC
changes from v3:
 - add IOMAP_STATE_ARRAY_SIZE
 - replace set_bit / clear_bit with bitmap_set / bitmap_clear
 - move iomap_set_page_dirty() out of 'iop->state_lock'
 - merge iomap_set/clear_range_dirty() and iomap_iop_clear/clear_range_dirty()

changes from v2:
 as suggested by Mathew:
 - move iomap_set_page_dirty() into iomap_set_range_dirty()
 - add DIRTY_BITS()
 - move ioamp_set_rannge_dirty() from iomap_page_mkwrite_actor() to
   iomap_page_mkwrite()
 - clear the dirty bits of entire page in iomap_writepage_map

changes from v1:
 - separate set dirty and clear dirty functions
 - don't test uptodate bit in iomap_writepage_map()
 - use one bitmap array for uptodate and dirty.

commit 9dc55f1389f9 ("iomap: add support for sub-pagesize buffered I/O
without buffer heads") replace the per-block structure buffer_head with
the per-page structure iomap_page. However, iomap_page can't track the
dirty state of sub pages, which will cause performance issue since sub
pages will be writeback even if they are not dirty.

For example, if block size is 4k and page size is 64k:

dd if=/dev/zero of=testfile bs=4k count=16 oflag=sync

With buffer_head implementation, the above dd cmd will writeback 4k in
each round. However, with iomap_page implementation, the range of
writeback in each round is from the start of the page to the end offset
we just wrote.

Thus add support to track dirty state in iomap_page.

I tested this path with:
test environment:
	platform:	arm64
	kernel:		v5.8
	pagesize:	64k
	blocksize:	4k

test case:
	dd if=/dev/zero of=/mnt/testfile bs=1M count=128
	fio --ioengine=sync --rw=randwrite --iodepth=64 --name=test --filename=/mnt/testfile --bs=4k --fsync=1

The test result is:
a. with patch

```
Jobs: 1 (f=1): [w(1)][100.0%][r=0KiB/s,w=4576KiB/s][r=0,w=1144 IOPS][eta 00m:00s]
test: (groupid=0, jobs=1): err= 0: pid=233608: Fri Aug 21 08:09:21 2020
  write: IOPS=1152, BW=4609KiB/s (4720kB/s)(128MiB/28437msec)
    clat (nsec): min=3000, max=23580, avg=4944.15, stdev=1589.23
     lat (nsec): min=3160, max=24500, avg=5109.51, stdev=1592.11
    clat percentiles (nsec):
     |  1.00th=[ 3376],  5.00th=[ 3600], 10.00th=[ 3824], 20.00th=[ 4016],
     | 30.00th=[ 4128], 40.00th=[ 4192], 50.00th=[ 4256], 60.00th=[ 4320],
     | 70.00th=[ 4448], 80.00th=[ 7200], 90.00th=[ 7968], 95.00th=[ 8160],
     | 99.00th=[ 8512], 99.50th=[ 8768], 99.90th=[10688], 99.95th=[12224],
     | 99.99th=[20352]
   bw (  KiB/s): min= 1920, max= 4800, per=99.97%, avg=4607.39, stdev=395.03, samples=56
   iops        : min=  480, max= 1200, avg=1151.84, stdev=98.76, samples=56
  lat (usec)   : 4=18.41%, 10=81.45%, 20=0.13%, 50=0.01%
  fsync/fdatasync/sync_file_range:
    sync (usec): min=651, max=22374, avg=852.24, stdev=430.25
    sync percentiles (usec):
     |  1.00th=[  660],  5.00th=[  660], 10.00th=[  668], 20.00th=[  668],
     | 30.00th=[  676], 40.00th=[  676], 50.00th=[  685], 60.00th=[  685],
     | 70.00th=[  693], 80.00th=[ 1401], 90.00th=[ 1418], 95.00th=[ 1434],
     | 99.00th=[ 1467], 99.50th=[ 1680], 99.90th=[ 7767], 99.95th=[ 7767],
     | 99.99th=[ 7767]
  cpu          : usr=0.41%, sys=2.18%, ctx=98400, majf=0, minf=4
  IO depths    : 1=200.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=0,32768,0,32767 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=64

Run status group 0 (all jobs):
  WRITE: bw=4609KiB/s (4720kB/s), 4609KiB/s-4609KiB/s (4720kB/s-4720kB/s), io=128MiB (134MB), run=28437-28437msec

Disk stats (read/write):
  sda: ios=4/65441, merge=0/5, ticks=3/28859, in_queue=54849, util=99.70%
```

b. without patch

```
Jobs: 1 (f=1): [w(1)][100.0%][r=0KiB/s,w=3003KiB/s][r=0,w=750 IOPS][eta 00m:00s]
test: (groupid=0, jobs=1): err= 0: pid=9174: Tue Aug 18 04:17:16 2020
  write: IOPS=678, BW=2714KiB/s (2780kB/s)(128MiB/48286msec)
    clat (nsec): min=3420, max=26240, avg=5898.60, stdev=1824.49
     lat (nsec): min=3600, max=26860, avg=6065.21, stdev=1826.90
    clat percentiles (nsec):
     |  1.00th=[ 3792],  5.00th=[ 4128], 10.00th=[ 4320], 20.00th=[ 4512],
     | 30.00th=[ 4576], 40.00th=[ 4704], 50.00th=[ 4832], 60.00th=[ 4960],
     | 70.00th=[ 7968], 80.00th=[ 8256], 90.00th=[ 8512], 95.00th=[ 8768],
     | 99.00th=[ 9152], 99.50th=[ 9408], 99.90th=[11840], 99.95th=[13376],
     | 99.99th=[18560]
   bw (  KiB/s): min= 1016, max= 3128, per=99.92%, avg=2711.92, stdev=357.89, samples=96
   iops        : min=  254, max=  782, avg=677.98, stdev=89.47, samples=96
  lat (usec)   : 4=3.14%, 10=96.66%, 20=0.20%, 50=0.01%
  fsync/fdatasync/sync_file_range:
    sync (usec): min=814, max=24221, avg=1456.82, stdev=543.48
    sync percentiles (usec):
     |  1.00th=[  988],  5.00th=[  996], 10.00th=[  996], 20.00th=[ 1012],
     | 30.00th=[ 1029], 40.00th=[ 1221], 50.00th=[ 1270], 60.00th=[ 1287],
     | 70.00th=[ 1795], 80.00th=[ 1844], 90.00th=[ 2245], 95.00th=[ 2278],
     | 99.00th=[ 2442], 99.50th=[ 2737], 99.90th=[ 5407], 99.95th=[ 5538],
     | 99.99th=[ 5735]
  cpu          : usr=0.19%, sys=1.54%, ctx=98412, majf=0, minf=4
  IO depths    : 1=200.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=0,32768,0,32767 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=64

Run status group 0 (all jobs):
  WRITE: bw=2714KiB/s (2780kB/s), 2714KiB/s-2714KiB/s (2780kB/s-2780kB/s), io=128MiB (134MB), run=48286-48286msec

Disk stats (read/write):
  sda: ios=4/65344, merge=0/5, ticks=2/48198, in_queue=88938, util=99.83%
```

c. ext4

```
Jobs: 1 (f=1): [w(1)][100.0%][r=0KiB/s,w=3919KiB/s][r=0,w=979 IOPS][eta 00m:00s]
test: (groupid=0, jobs=1): err= 0: pid=8682: Tue Aug 18 04:15:43 2020
  write: IOPS=960, BW=3840KiB/s (3932kB/s)(128MiB/34133msec)
    clat (usec): min=4, max=349, avg= 8.92, stdev= 2.94
     lat (usec): min=4, max=349, avg= 9.06, stdev= 2.94
    clat percentiles (nsec):
     |  1.00th=[ 6112],  5.00th=[ 6624], 10.00th=[ 6880], 20.00th=[ 7200],
     | 30.00th=[ 7456], 40.00th=[ 7712], 50.00th=[ 8032], 60.00th=[ 8384],
     | 70.00th=[ 9024], 80.00th=[11712], 90.00th=[12608], 95.00th=[13120],
     | 99.00th=[14272], 99.50th=[14656], 99.90th=[17536], 99.95th=[20352],
     | 99.99th=[33536]
   bw (  KiB/s): min= 1344, max= 3992, per=100.00%, avg=3839.88, stdev=314.69, samples=68
   iops        : min=  336, max=  998, avg=959.97, stdev=78.67, samples=68
  lat (usec)   : 10=74.64%, 20=25.31%, 50=0.05%, 100=0.01%, 500=0.01%
  fsync/fdatasync/sync_file_range:
    sync (usec): min=666, max=25174, avg=1021.69, stdev=871.62
    sync percentiles (usec):
     |  1.00th=[  685],  5.00th=[  693], 10.00th=[  701], 20.00th=[  701],
     | 30.00th=[  709], 40.00th=[  717], 50.00th=[  717], 60.00th=[  725],
     | 70.00th=[  734], 80.00th=[ 1500], 90.00th=[ 1516], 95.00th=[ 1532],
     | 99.00th=[ 6128], 99.50th=[ 6128], 99.90th=[ 7832], 99.95th=[ 8225],
     | 99.99th=[ 9634]
  cpu          : usr=0.32%, sys=2.87%, ctx=90254, majf=0, minf=4
  IO depths    : 1=200.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=0,32768,0,32767 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=64

Run status group 0 (all jobs):
  WRITE: bw=3840KiB/s (3932kB/s), 3840KiB/s-3840KiB/s (3932kB/s-3932kB/s), io=128MiB (134MB), run=34133-34133msec

Disk stats (read/write):
  sda: ios=0/75055, merge=0/8822, ticks=0/40565, in_queue=68469, util=99.80%
```

Signed-off-by: Yu Kuai <yukuai3@huawei.com>
---
 fs/iomap/buffered-io.c | 88 ++++++++++++++++++++++++++++++++++--------
 1 file changed, 72 insertions(+), 16 deletions(-)

Comments

Matthew Wilcox Aug. 21, 2020, 12:44 p.m. UTC | #1
On Fri, Aug 21, 2020 at 08:33:06PM +0800, Yu Kuai wrote:
> changes from v3:
>  - add IOMAP_STATE_ARRAY_SIZE
>  - replace set_bit / clear_bit with bitmap_set / bitmap_clear
>  - move iomap_set_page_dirty() out of 'iop->state_lock'
>  - merge iomap_set/clear_range_dirty() and iomap_iop_clear/clear_range_dirty()

I'm still working on the iomap parts of the THP series (fixing up
invalidatepage right now), but here are some of the relevant bits (patch
series to follow)
Yu Kuai Sept. 11, 2020, 8:27 a.m. UTC | #2
On 2020/08/21 20:44, Matthew Wilcox wrote:
> On Fri, Aug 21, 2020 at 08:33:06PM +0800, Yu Kuai wrote:
>> changes from v3: - add IOMAP_STATE_ARRAY_SIZE - replace set_bit / 
>> clear_bit with bitmap_set / bitmap_clear - move 
>> iomap_set_page_dirty() out of 'iop->state_lock' - merge 
>> iomap_set/clear_range_dirty() and 
>> iomap_iop_clear/clear_range_dirty()
> 
> I'm still working on the iomap parts of the THP series (fixing up 
> invalidatepage right now), but here are some of the relevant bits 
> (patch series to follow)
> 

Hi, Matthew

Since your THP iomap patches were reviewed, I made some modifications
based on these patches.

Best regards,
Yu Kuai

---
  fs/iomap/buffered-io.c | 92 +++++++++++++++++++++++++++++++++---------
  1 file changed, 74 insertions(+), 18 deletions(-)

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index edf5eea56cf5..bc7f57748be8 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -23,13 +23,17 @@

  /*
   * Structure allocated for each page or THP when block size < page size
- * to track sub-page uptodate status and I/O completions.
+ * to track sub-page status and I/O completions.
   */
  struct iomap_page {
  	atomic_t		read_bytes_pending;
  	atomic_t		write_bytes_pending;
-	spinlock_t		uptodate_lock;
-	unsigned long		uptodate[];
+	spinlock_t		state_lock;
+	/*
+	 * The first half bits are used to track sub-page uptodate status,
+	 * the second half bits are for dirty status.
+	 */
+	unsigned long		state[];
  };

  static inline struct iomap_page *to_iomap_page(struct page *page)
@@ -57,9 +61,9 @@ iomap_page_create(struct inode *inode, struct page *page)
  	if (iop || nr_blocks <= 1)
  		return iop;

-	iop = kzalloc(struct_size(iop, uptodate, BITS_TO_LONGS(nr_blocks)),
+	iop = kzalloc(struct_size(iop, state, BITS_TO_LONGS(2 * nr_blocks)),
  			GFP_NOFS | __GFP_NOFAIL);
-	spin_lock_init(&iop->uptodate_lock);
+	spin_lock_init(&iop->state_lock);
  	attach_page_private(page, iop);
  	return iop;
  }
@@ -74,7 +78,7 @@ iomap_page_release(struct page *page)
  		return;
  	WARN_ON_ONCE(atomic_read(&iop->read_bytes_pending));
  	WARN_ON_ONCE(atomic_read(&iop->write_bytes_pending));
-	WARN_ON_ONCE(bitmap_full(iop->uptodate, nr_blocks) !=
+	WARN_ON_ONCE(bitmap_full(iop->state, nr_blocks) !=
  			PageUptodate(page));
  	kfree(iop);
  }
@@ -105,7 +109,7 @@ iomap_adjust_read_range(struct inode *inode, struct 
iomap_page *iop,

  		/* move forward for each leading block marked uptodate */
  		for (i = first; i <= last; i++) {
-			if (!test_bit(i, iop->uptodate))
+			if (!test_bit(i, iop->state))
  				break;
  			*pos += block_size;
  			poff += block_size;
@@ -115,7 +119,7 @@ iomap_adjust_read_range(struct inode *inode, struct 
iomap_page *iop,

  		/* truncate len if we find any trailing uptodate block(s) */
  		for ( ; i <= last; i++) {
-			if (test_bit(i, iop->uptodate)) {
+			if (test_bit(i, iop->state)) {
  				plen -= (last - i + 1) * block_size;
  				last = i - 1;
  				break;
@@ -139,6 +143,55 @@ iomap_adjust_read_range(struct inode *inode, struct 
iomap_page *iop,
  	*lenp = plen;
  }

+static void
+iomap_set_range_dirty(struct page *page, unsigned int off,
+		unsigned int len)
+{
+	struct inode *inode = page->mapping->host;
+	unsigned int blocks_per_page = i_blocks_per_page(inode, page);
+	unsigned int first = (off >> inode->i_blkbits) + blocks_per_page;
+	unsigned int last = ((off + len - 1) >> inode->i_blkbits) + 
blocks_per_page;
+	unsigned long flags;
+	struct iomap_page *iop;
+
+	if (PageError(page))
+		return;
+
+	if (len)
+		iomap_set_page_dirty(page);
+
+	if (!page_has_private(page))
+		return;
+
+	iop = to_iomap_page(page);
+	spin_lock_irqsave(&iop->state_lock, flags);
+	bitmap_set(iop->state, first, last - first + 1);
+	spin_unlock_irqrestore(&iop->state_lock, flags);
+}
+
+static void
+iomap_clear_range_dirty(struct page *page, unsigned int off,
+		unsigned int len)
+{
+	struct inode *inode = page->mapping->host;
+	unsigned int blocks_per_page = i_blocks_per_page(inode, page);
+	unsigned int first = (off >> inode->i_blkbits) + blocks_per_page;
+	unsigned int last = ((off + len - 1) >> inode->i_blkbits) + 
blocks_per_page;
+	unsigned long flags;
+	struct iomap_page *iop;
+
+	if (PageError(page))
+		return;
+
+	if (!page_has_private(page))
+		return;
+
+	iop = to_iomap_page(page);
+	spin_lock_irqsave(&iop->state_lock, flags);
+	bitmap_clear(iop->state, first, last - first + 1);
+	spin_unlock_irqrestore(&iop->state_lock, flags);
+}
+
  static void
  iomap_iop_set_range_uptodate(struct page *page, unsigned off, unsigned 
len)
  {
@@ -148,11 +201,11 @@ iomap_iop_set_range_uptodate(struct page *page, 
unsigned off, unsigned len)
  	unsigned last = (off + len - 1) >> inode->i_blkbits;
  	unsigned long flags;

-	spin_lock_irqsave(&iop->uptodate_lock, flags);
-	bitmap_set(iop->uptodate, first, last - first + 1);
-	if (bitmap_full(iop->uptodate, i_blocks_per_page(inode, page)))
+	spin_lock_irqsave(&iop->state_lock, flags);
+	bitmap_set(iop->state, first, last - first + 1);
+	if (bitmap_full(iop->state, i_blocks_per_page(inode, page)))
  		SetPageUptodate(page);
-	spin_unlock_irqrestore(&iop->uptodate_lock, flags);
+	spin_unlock_irqrestore(&iop->state_lock, flags);
  }

  static void
@@ -445,7 +498,7 @@ iomap_is_partially_uptodate(struct page *page, 
unsigned long from,

  	if (iop) {
  		for (i = first; i <= last; i++)
-			if (!test_bit(i, iop->uptodate))
+			if (!test_bit(i, iop->state))
  				return 0;
  		return 1;
  	}
@@ -683,7 +736,7 @@ static size_t __iomap_write_end(struct inode *inode, 
loff_t pos, size_t len,
  	if (unlikely(copied < len && !PageUptodate(page)))
  		return 0;
  	iomap_set_range_uptodate(page, offset_in_page(pos), len);
-	iomap_set_page_dirty(page);
+	iomap_set_range_dirty(page, offset_in_page(pos), len);
  	return copied;
  }

@@ -997,7 +1050,6 @@ iomap_page_mkwrite_actor(struct inode *inode, 
loff_t pos, loff_t length,
  	} else {
  		WARN_ON_ONCE(!PageUptodate(page));
  		iomap_page_create(inode, page);
-		set_page_dirty(page);
  	}

  	return length;
@@ -1007,7 +1059,7 @@ vm_fault_t iomap_page_mkwrite(struct vm_fault 
*vmf, const struct iomap_ops *ops)
  {
  	struct page *page = vmf->page;
  	struct inode *inode = file_inode(vmf->vma->vm_file);
-	unsigned long length;
+	unsigned int length, dirty_bits;
  	loff_t offset;
  	ssize_t ret;

@@ -1016,6 +1068,7 @@ vm_fault_t iomap_page_mkwrite(struct vm_fault 
*vmf, const struct iomap_ops *ops)
  	if (ret < 0)
  		goto out_unlock;
  	length = ret;
+	dirty_bits = ret;

  	offset = page_offset(page);
  	while (length > 0) {
@@ -1028,6 +1081,7 @@ vm_fault_t iomap_page_mkwrite(struct vm_fault 
*vmf, const struct iomap_ops *ops)
  		length -= ret;
  	}

+	iomap_set_range_dirty(page, 0, dirty_bits);
  	wait_for_stable_page(page);
  	return VM_FAULT_LOCKED;
  out_unlock:
@@ -1340,11 +1394,12 @@ iomap_writepage_map(struct iomap_writepage_ctx *wpc,
  	struct iomap_page *iop = to_iomap_page(page);
  	struct iomap_ioend *ioend, *next;
  	unsigned len = i_blocksize(inode);
+	unsigned int blocks_per_page = i_blocks_per_page(inode, page);
  	u64 file_offset; /* file offset of page */
  	int error = 0, count = 0, i;
  	LIST_HEAD(submit_list);

-	WARN_ON_ONCE(i_blocks_per_page(inode, page) > 1 && !iop);
+	WARN_ON_ONCE(blocks_per_page > 1 && !iop);
  	WARN_ON_ONCE(iop && atomic_read(&iop->write_bytes_pending) != 0);

  	/*
@@ -1355,7 +1410,7 @@ iomap_writepage_map(struct iomap_writepage_ctx *wpc,
  	for (i = 0, file_offset = page_offset(page);
  	     i < (PAGE_SIZE >> inode->i_blkbits) && file_offset < end_offset;
  	     i++, file_offset += len) {
-		if (iop && !test_bit(i, iop->uptodate))
+		if (iop && !test_bit(i, iop->state + blocks_per_page))
  			continue;

  		error = wpc->ops->map_blocks(wpc, inode, file_offset);
@@ -1404,6 +1459,7 @@ iomap_writepage_map(struct iomap_writepage_ctx *wpc,
  		 */
  		set_page_writeback_keepwrite(page);
  	} else {
+		iomap_clear_range_dirty(page, 0, PAGE_SIZE);
  		clear_page_dirty_for_io(page);
  		set_page_writeback(page);
  	}
Yu Kuai Sept. 11, 2020, 8:35 a.m. UTC | #3
Hi,

Sorry that after copy and paste, the content of the patch somehow 
changed and looks strange.

Best regards,
Yu Kuai

On 2020/09/11 16:27, yukuai (C) wrote:
> On 2020/08/21 20:44, Matthew Wilcox wrote:
>> On Fri, Aug 21, 2020 at 08:33:06PM +0800, Yu Kuai wrote:
>>> changes from v3: - add IOMAP_STATE_ARRAY_SIZE - replace set_bit / 
>>> clear_bit with bitmap_set / bitmap_clear - move 
>>> iomap_set_page_dirty() out of 'iop->state_lock' - merge 
>>> iomap_set/clear_range_dirty() and iomap_iop_clear/clear_range_dirty()
>>
>> I'm still working on the iomap parts of the THP series (fixing up 
>> invalidatepage right now), but here are some of the relevant bits 
>> (patch series to follow)
>>
> 
> Hi, Matthew
> 
> Since your THP iomap patches were reviewed, I made some modifications
> based on these patches.
> 
> Best regards,
> Yu Kuai
> 
> ---
>   fs/iomap/buffered-io.c | 92 +++++++++++++++++++++++++++++++++---------
>   1 file changed, 74 insertions(+), 18 deletions(-)
> 
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index edf5eea56cf5..bc7f57748be8 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -23,13 +23,17 @@
> 
>   /*
>    * Structure allocated for each page or THP when block size < page size
> - * to track sub-page uptodate status and I/O completions.
> + * to track sub-page status and I/O completions.
>    */
>   struct iomap_page {
>       atomic_t        read_bytes_pending;
>       atomic_t        write_bytes_pending;
> -    spinlock_t        uptodate_lock;
> -    unsigned long        uptodate[];
> +    spinlock_t        state_lock;
> +    /*
> +     * The first half bits are used to track sub-page uptodate status,
> +     * the second half bits are for dirty status.
> +     */
> +    unsigned long        state[];
>   };
> 
>   static inline struct iomap_page *to_iomap_page(struct page *page)
> @@ -57,9 +61,9 @@ iomap_page_create(struct inode *inode, struct page *page)
>       if (iop || nr_blocks <= 1)
>           return iop;
> 
> -    iop = kzalloc(struct_size(iop, uptodate, BITS_TO_LONGS(nr_blocks)),
> +    iop = kzalloc(struct_size(iop, state, BITS_TO_LONGS(2 * nr_blocks)),
>               GFP_NOFS | __GFP_NOFAIL);
> -    spin_lock_init(&iop->uptodate_lock);
> +    spin_lock_init(&iop->state_lock);
>       attach_page_private(page, iop);
>       return iop;
>   }
> @@ -74,7 +78,7 @@ iomap_page_release(struct page *page)
>           return;
>       WARN_ON_ONCE(atomic_read(&iop->read_bytes_pending));
>       WARN_ON_ONCE(atomic_read(&iop->write_bytes_pending));
> -    WARN_ON_ONCE(bitmap_full(iop->uptodate, nr_blocks) !=
> +    WARN_ON_ONCE(bitmap_full(iop->state, nr_blocks) !=
>               PageUptodate(page));
>       kfree(iop);
>   }
> @@ -105,7 +109,7 @@ iomap_adjust_read_range(struct inode *inode, struct 
> iomap_page *iop,
> 
>           /* move forward for each leading block marked uptodate */
>           for (i = first; i <= last; i++) {
> -            if (!test_bit(i, iop->uptodate))
> +            if (!test_bit(i, iop->state))
>                   break;
>               *pos += block_size;
>               poff += block_size;
> @@ -115,7 +119,7 @@ iomap_adjust_read_range(struct inode *inode, struct 
> iomap_page *iop,
> 
>           /* truncate len if we find any trailing uptodate block(s) */
>           for ( ; i <= last; i++) {
> -            if (test_bit(i, iop->uptodate)) {
> +            if (test_bit(i, iop->state)) {
>                   plen -= (last - i + 1) * block_size;
>                   last = i - 1;
>                   break;
> @@ -139,6 +143,55 @@ iomap_adjust_read_range(struct inode *inode, struct 
> iomap_page *iop,
>       *lenp = plen;
>   }
> 
> +static void
> +iomap_set_range_dirty(struct page *page, unsigned int off,
> +        unsigned int len)
> +{
> +    struct inode *inode = page->mapping->host;
> +    unsigned int blocks_per_page = i_blocks_per_page(inode, page);
> +    unsigned int first = (off >> inode->i_blkbits) + blocks_per_page;
> +    unsigned int last = ((off + len - 1) >> inode->i_blkbits) + 
> blocks_per_page;
> +    unsigned long flags;
> +    struct iomap_page *iop;
> +
> +    if (PageError(page))
> +        return;
> +
> +    if (len)
> +        iomap_set_page_dirty(page);
> +
> +    if (!page_has_private(page))
> +        return;
> +
> +    iop = to_iomap_page(page);
> +    spin_lock_irqsave(&iop->state_lock, flags);
> +    bitmap_set(iop->state, first, last - first + 1);
> +    spin_unlock_irqrestore(&iop->state_lock, flags);
> +}
> +
> +static void
> +iomap_clear_range_dirty(struct page *page, unsigned int off,
> +        unsigned int len)
> +{
> +    struct inode *inode = page->mapping->host;
> +    unsigned int blocks_per_page = i_blocks_per_page(inode, page);
> +    unsigned int first = (off >> inode->i_blkbits) + blocks_per_page;
> +    unsigned int last = ((off + len - 1) >> inode->i_blkbits) + 
> blocks_per_page;
> +    unsigned long flags;
> +    struct iomap_page *iop;
> +
> +    if (PageError(page))
> +        return;
> +
> +    if (!page_has_private(page))
> +        return;
> +
> +    iop = to_iomap_page(page);
> +    spin_lock_irqsave(&iop->state_lock, flags);
> +    bitmap_clear(iop->state, first, last - first + 1);
> +    spin_unlock_irqrestore(&iop->state_lock, flags);
> +}
> +
>   static void
>   iomap_iop_set_range_uptodate(struct page *page, unsigned off, unsigned 
> len)
>   {
> @@ -148,11 +201,11 @@ iomap_iop_set_range_uptodate(struct page *page, 
> unsigned off, unsigned len)
>       unsigned last = (off + len - 1) >> inode->i_blkbits;
>       unsigned long flags;
> 
> -    spin_lock_irqsave(&iop->uptodate_lock, flags);
> -    bitmap_set(iop->uptodate, first, last - first + 1);
> -    if (bitmap_full(iop->uptodate, i_blocks_per_page(inode, page)))
> +    spin_lock_irqsave(&iop->state_lock, flags);
> +    bitmap_set(iop->state, first, last - first + 1);
> +    if (bitmap_full(iop->state, i_blocks_per_page(inode, page)))
>           SetPageUptodate(page);
> -    spin_unlock_irqrestore(&iop->uptodate_lock, flags);
> +    spin_unlock_irqrestore(&iop->state_lock, flags);
>   }
> 
>   static void
> @@ -445,7 +498,7 @@ iomap_is_partially_uptodate(struct page *page, 
> unsigned long from,
> 
>       if (iop) {
>           for (i = first; i <= last; i++)
> -            if (!test_bit(i, iop->uptodate))
> +            if (!test_bit(i, iop->state))
>                   return 0;
>           return 1;
>       }
> @@ -683,7 +736,7 @@ static size_t __iomap_write_end(struct inode *inode, 
> loff_t pos, size_t len,
>       if (unlikely(copied < len && !PageUptodate(page)))
>           return 0;
>       iomap_set_range_uptodate(page, offset_in_page(pos), len);
> -    iomap_set_page_dirty(page);
> +    iomap_set_range_dirty(page, offset_in_page(pos), len);
>       return copied;
>   }
> 
> @@ -997,7 +1050,6 @@ iomap_page_mkwrite_actor(struct inode *inode, 
> loff_t pos, loff_t length,
>       } else {
>           WARN_ON_ONCE(!PageUptodate(page));
>           iomap_page_create(inode, page);
> -        set_page_dirty(page);
>       }
> 
>       return length;
> @@ -1007,7 +1059,7 @@ vm_fault_t iomap_page_mkwrite(struct vm_fault 
> *vmf, const struct iomap_ops *ops)
>   {
>       struct page *page = vmf->page;
>       struct inode *inode = file_inode(vmf->vma->vm_file);
> -    unsigned long length;
> +    unsigned int length, dirty_bits;
>       loff_t offset;
>       ssize_t ret;
> 
> @@ -1016,6 +1068,7 @@ vm_fault_t iomap_page_mkwrite(struct vm_fault 
> *vmf, const struct iomap_ops *ops)
>       if (ret < 0)
>           goto out_unlock;
>       length = ret;
> +    dirty_bits = ret;
> 
>       offset = page_offset(page);
>       while (length > 0) {
> @@ -1028,6 +1081,7 @@ vm_fault_t iomap_page_mkwrite(struct vm_fault 
> *vmf, const struct iomap_ops *ops)
>           length -= ret;
>       }
> 
> +    iomap_set_range_dirty(page, 0, dirty_bits);
>       wait_for_stable_page(page);
>       return VM_FAULT_LOCKED;
>   out_unlock:
> @@ -1340,11 +1394,12 @@ iomap_writepage_map(struct iomap_writepage_ctx 
> *wpc,
>       struct iomap_page *iop = to_iomap_page(page);
>       struct iomap_ioend *ioend, *next;
>       unsigned len = i_blocksize(inode);
> +    unsigned int blocks_per_page = i_blocks_per_page(inode, page);
>       u64 file_offset; /* file offset of page */
>       int error = 0, count = 0, i;
>       LIST_HEAD(submit_list);
> 
> -    WARN_ON_ONCE(i_blocks_per_page(inode, page) > 1 && !iop);
> +    WARN_ON_ONCE(blocks_per_page > 1 && !iop);
>       WARN_ON_ONCE(iop && atomic_read(&iop->write_bytes_pending) != 0);
> 
>       /*
> @@ -1355,7 +1410,7 @@ iomap_writepage_map(struct iomap_writepage_ctx *wpc,
>       for (i = 0, file_offset = page_offset(page);
>            i < (PAGE_SIZE >> inode->i_blkbits) && file_offset < end_offset;
>            i++, file_offset += len) {
> -        if (iop && !test_bit(i, iop->uptodate))
> +        if (iop && !test_bit(i, iop->state + blocks_per_page))
>               continue;
> 
>           error = wpc->ops->map_blocks(wpc, inode, file_offset);
> @@ -1404,6 +1459,7 @@ iomap_writepage_map(struct iomap_writepage_ctx *wpc,
>            */
>           set_page_writeback_keepwrite(page);
>       } else {
> +        iomap_clear_range_dirty(page, 0, PAGE_SIZE);
>           clear_page_dirty_for_io(page);
>           set_page_writeback(page);
>       }
Matthew Wilcox Sept. 11, 2020, 12:05 p.m. UTC | #4
On Fri, Sep 11, 2020 at 04:27:08PM +0800, yukuai (C) wrote:
> Since your THP iomap patches were reviewed, I made some modifications
> based on these patches.

Thanks!  This looks good to me!

> +static void
> +iomap_set_range_dirty(struct page *page, unsigned int off,
> +		unsigned int len)
> +{
> +	struct inode *inode = page->mapping->host;
> +	unsigned int blocks_per_page = i_blocks_per_page(inode, page);

I might call this nr_blocks.

> +	unsigned int first = (off >> inode->i_blkbits) + blocks_per_page;
> +	unsigned int last = ((off + len - 1) >> inode->i_blkbits) +
> blocks_per_page;
> +	unsigned long flags;
> +	struct iomap_page *iop;
> +
> +	if (PageError(page))
> +		return;

I think this is wrong.  If PageError is set then we've seen an I/O error
on this page (today either a read or a write, although I intend to make
that for writes only soon).  But I don't see a reason for declining to
make a page dirty if we've seen an eror -- indeed, that seems likely to
lead to further data loss as we then won't even try to write parts of
the page back to storage.

> +	if (len)
> +		iomap_set_page_dirty(page);
> +
> +	if (!page_has_private(page))
> +		return;
> +
> +	iop = to_iomap_page(page);

We usually do the last three lines as:

	iop = to_iomap_page(page);
	if (!iop)
		return;

(actually, we usually call to_iomap_page() at the start of the function
and then check it here)

> +static void
> +iomap_clear_range_dirty(struct page *page, unsigned int off,
> +		unsigned int len)
> +{
> +	struct inode *inode = page->mapping->host;
> +	unsigned int blocks_per_page = i_blocks_per_page(inode, page);
> +	unsigned int first = (off >> inode->i_blkbits) + blocks_per_page;
> +	unsigned int last = ((off + len - 1) >> inode->i_blkbits) +
> blocks_per_page;
> +	unsigned long flags;
> +	struct iomap_page *iop;
> +
> +	if (PageError(page))
> +		return;

It does make sense to avoid clearing the dirty state of the page here;
if the write failed, then the page is still dirty, and it'd be nice to
retry writes until they succeed.  So if you take out the PageError in
set_range_dirty, don't take it out here!

> +	if (!page_has_private(page))
> +		return;
> +
> +	iop = to_iomap_page(page);
> +	spin_lock_irqsave(&iop->state_lock, flags);
> +	bitmap_clear(iop->state, first, last - first + 1);
> +	spin_unlock_irqrestore(&iop->state_lock, flags);
> +}
> +
>  static void
>  iomap_iop_set_range_uptodate(struct page *page, unsigned off, unsigned len)
>  {
> @@ -148,11 +201,11 @@ iomap_iop_set_range_uptodate(struct page *page,
> unsigned off, unsigned len)
>  	unsigned last = (off + len - 1) >> inode->i_blkbits;
>  	unsigned long flags;
> 
> -	spin_lock_irqsave(&iop->uptodate_lock, flags);
> -	bitmap_set(iop->uptodate, first, last - first + 1);
> -	if (bitmap_full(iop->uptodate, i_blocks_per_page(inode, page)))
> +	spin_lock_irqsave(&iop->state_lock, flags);
> +	bitmap_set(iop->state, first, last - first + 1);
> +	if (bitmap_full(iop->state, i_blocks_per_page(inode, page)))
>  		SetPageUptodate(page);
> -	spin_unlock_irqrestore(&iop->uptodate_lock, flags);
> +	spin_unlock_irqrestore(&iop->state_lock, flags);
>  }
> 
>  static void
> @@ -445,7 +498,7 @@ iomap_is_partially_uptodate(struct page *page, unsigned
> long from,
> 
>  	if (iop) {
>  		for (i = first; i <= last; i++)
> -			if (!test_bit(i, iop->uptodate))
> +			if (!test_bit(i, iop->state))
>  				return 0;
>  		return 1;
>  	}
> @@ -683,7 +736,7 @@ static size_t __iomap_write_end(struct inode *inode,
> loff_t pos, size_t len,
>  	if (unlikely(copied < len && !PageUptodate(page)))
>  		return 0;
>  	iomap_set_range_uptodate(page, offset_in_page(pos), len);
> -	iomap_set_page_dirty(page);
> +	iomap_set_range_dirty(page, offset_in_page(pos), len);

I _think_ the call to set_range_uptodate here is now unnecessary.  The
blocks should already have been set uptodate in write_begin.  But please
check!

> @@ -1007,7 +1059,7 @@ vm_fault_t iomap_page_mkwrite(struct vm_fault *vmf,
> const struct iomap_ops *ops)
>  {
>  	struct page *page = vmf->page;
>  	struct inode *inode = file_inode(vmf->vma->vm_file);
> -	unsigned long length;
> +	unsigned int length, dirty_bits;

This is dirty_bytes, surely?
Yu Kuai Sept. 25, 2020, 2:30 a.m. UTC | #5
On 2020/09/11 20:05, Matthew Wilcox wrote:
>> @@ -683,7 +736,7 @@ static size_t __iomap_write_end(struct inode *inode,
>> loff_t pos, size_t len,
>>   	if (unlikely(copied < len && !PageUptodate(page)))
>>   		return 0;
>>   	iomap_set_range_uptodate(page, offset_in_page(pos), len);
>> -	iomap_set_page_dirty(page);
>> +	iomap_set_range_dirty(page, offset_in_page(pos), len);
> I_think_  the call to set_range_uptodate here is now unnecessary.  The
> blocks should already have been set uptodate in write_begin.  But please
> check!

Hi, Matthew

Sorry it took me so long to get back to this.

I found that set_range_uptodate() might be skipped in write_begin()
in this case:

                  if (!(flags & IOMAP_WRITE_F_UNSHARE) &&
                   ┊   (from <= poff || from >= poff + plen) &&
                   ┊   (to <= poff || to >= poff + plen))
                           continue;

And iomap_adjust_read_range() can set 'poff' greater than 'from'
and 'poff + len' less than 'to'.

Thanks
Yu Kuai
diff mbox series

Patch

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index bcfc288dba3f..4539f98f3c12 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -21,15 +21,22 @@ 
 
 #include "../internal.h"
 
+#define DIRTY_BITS(x)	((x) + PAGE_SIZE / SECTOR_SIZE)
+#define IOMAP_STATE_ARRAY_SIZE	(PAGE_SIZE * 2 / SECTOR_SIZE)
+
 /*
  * Structure allocated for each page when block size < PAGE_SIZE to track
- * sub-page uptodate status and I/O completions.
+ * sub-page status and I/O completions.
  */
 struct iomap_page {
 	atomic_t		read_count;
 	atomic_t		write_count;
-	spinlock_t		uptodate_lock;
-	DECLARE_BITMAP(uptodate, PAGE_SIZE / 512);
+	spinlock_t		state_lock;
+	/*
+	 * The first half bits are used to track sub-page uptodate status,
+	 * the second half bits are for dirty status.
+	 */
+	DECLARE_BITMAP(state, IOMAP_STATE_ARRAY_SIZE);
 };
 
 static inline struct iomap_page *to_iomap_page(struct page *page)
@@ -52,8 +59,8 @@  iomap_page_create(struct inode *inode, struct page *page)
 	iop = kmalloc(sizeof(*iop), GFP_NOFS | __GFP_NOFAIL);
 	atomic_set(&iop->read_count, 0);
 	atomic_set(&iop->write_count, 0);
-	spin_lock_init(&iop->uptodate_lock);
-	bitmap_zero(iop->uptodate, PAGE_SIZE / SECTOR_SIZE);
+	spin_lock_init(&iop->state_lock);
+	bitmap_zero(iop->state, IOMAP_STATE_ARRAY_SIZE);
 
 	/*
 	 * migrate_page_move_mapping() assumes that pages with private data have
@@ -101,7 +108,7 @@  iomap_adjust_read_range(struct inode *inode, struct iomap_page *iop,
 
 		/* move forward for each leading block marked uptodate */
 		for (i = first; i <= last; i++) {
-			if (!test_bit(i, iop->uptodate))
+			if (!test_bit(i, iop->state))
 				break;
 			*pos += block_size;
 			poff += block_size;
@@ -111,7 +118,7 @@  iomap_adjust_read_range(struct inode *inode, struct iomap_page *iop,
 
 		/* truncate len if we find any trailing uptodate block(s) */
 		for ( ; i <= last; i++) {
-			if (test_bit(i, iop->uptodate)) {
+			if (test_bit(i, iop->state)) {
 				plen -= (last - i + 1) * block_size;
 				last = i - 1;
 				break;
@@ -135,6 +142,53 @@  iomap_adjust_read_range(struct inode *inode, struct iomap_page *iop,
 	*lenp = plen;
 }
 
+static void
+iomap_set_range_dirty(struct page *page, unsigned int off,
+		unsigned int len)
+{
+	struct inode *inode = page->mapping->host;
+	unsigned int first = DIRTY_BITS(off >> inode->i_blkbits);
+	unsigned int last = DIRTY_BITS((off + len - 1) >> inode->i_blkbits);
+	unsigned long flags;
+	struct iomap_page *iop;
+
+	if (PageError(page))
+		return;
+
+	if (len)
+		iomap_set_page_dirty(page);
+
+	if (!page_has_private(page))
+		return;
+
+	iop = to_iomap_page(page);
+	spin_lock_irqsave(&iop->state_lock, flags);
+	bitmap_set(iop->state, first, last - first + 1);
+	spin_unlock_irqrestore(&iop->state_lock, flags);
+}
+
+static void
+iomap_clear_range_dirty(struct page *page, unsigned int off,
+		unsigned int len)
+{
+	struct inode *inode = page->mapping->host;
+	unsigned int first = DIRTY_BITS(off >> inode->i_blkbits);
+	unsigned int last = DIRTY_BITS((off + len - 1) >> inode->i_blkbits);
+	unsigned long flags;
+	struct iomap_page *iop;
+
+	if (PageError(page))
+		return;
+
+	if (!page_has_private(page))
+		return;
+
+	iop = to_iomap_page(page);
+	spin_lock_irqsave(&iop->state_lock, flags);
+	bitmap_clear(iop->state, first, last - first + 1);
+	spin_unlock_irqrestore(&iop->state_lock, flags);
+}
+
 static void
 iomap_iop_set_range_uptodate(struct page *page, unsigned off, unsigned len)
 {
@@ -146,17 +200,17 @@  iomap_iop_set_range_uptodate(struct page *page, unsigned off, unsigned len)
 	unsigned long flags;
 	unsigned int i;
 
-	spin_lock_irqsave(&iop->uptodate_lock, flags);
+	spin_lock_irqsave(&iop->state_lock, flags);
 	for (i = 0; i < PAGE_SIZE / i_blocksize(inode); i++) {
 		if (i >= first && i <= last)
-			set_bit(i, iop->uptodate);
-		else if (!test_bit(i, iop->uptodate))
+			set_bit(i, iop->state);
+		else if (!test_bit(i, iop->state))
 			uptodate = false;
 	}
 
 	if (uptodate)
 		SetPageUptodate(page);
-	spin_unlock_irqrestore(&iop->uptodate_lock, flags);
+	spin_unlock_irqrestore(&iop->state_lock, flags);
 }
 
 static void
@@ -466,7 +520,7 @@  iomap_is_partially_uptodate(struct page *page, unsigned long from,
 
 	if (iop) {
 		for (i = first; i <= last; i++)
-			if (!test_bit(i, iop->uptodate))
+			if (!test_bit(i, iop->state))
 				return 0;
 		return 1;
 	}
@@ -705,7 +759,7 @@  __iomap_write_end(struct inode *inode, loff_t pos, unsigned len,
 	if (unlikely(copied < len && !PageUptodate(page)))
 		return 0;
 	iomap_set_range_uptodate(page, offset_in_page(pos), len);
-	iomap_set_page_dirty(page);
+	iomap_set_range_dirty(page, offset_in_page(pos), len);
 	return copied;
 }
 
@@ -1029,7 +1083,6 @@  iomap_page_mkwrite_actor(struct inode *inode, loff_t pos, loff_t length,
 	} else {
 		WARN_ON_ONCE(!PageUptodate(page));
 		iomap_page_create(inode, page);
-		set_page_dirty(page);
 	}
 
 	return length;
@@ -1039,7 +1092,7 @@  vm_fault_t iomap_page_mkwrite(struct vm_fault *vmf, const struct iomap_ops *ops)
 {
 	struct page *page = vmf->page;
 	struct inode *inode = file_inode(vmf->vma->vm_file);
-	unsigned long length;
+	unsigned int length, bytes_in_page;
 	loff_t offset;
 	ssize_t ret;
 
@@ -1048,6 +1101,7 @@  vm_fault_t iomap_page_mkwrite(struct vm_fault *vmf, const struct iomap_ops *ops)
 	if (ret < 0)
 		goto out_unlock;
 	length = ret;
+	bytes_in_page = ret;
 
 	offset = page_offset(page);
 	while (length > 0) {
@@ -1060,6 +1114,7 @@  vm_fault_t iomap_page_mkwrite(struct vm_fault *vmf, const struct iomap_ops *ops)
 		length -= ret;
 	}
 
+	iomap_set_range_dirty(page, 0, bytes_in_page);
 	wait_for_stable_page(page);
 	return VM_FAULT_LOCKED;
 out_unlock:
@@ -1386,7 +1441,7 @@  iomap_writepage_map(struct iomap_writepage_ctx *wpc,
 	for (i = 0, file_offset = page_offset(page);
 	     i < (PAGE_SIZE >> inode->i_blkbits) && file_offset < end_offset;
 	     i++, file_offset += len) {
-		if (iop && !test_bit(i, iop->uptodate))
+		if (iop && !test_bit(DIRTY_BITS(i), iop->state))
 			continue;
 
 		error = wpc->ops->map_blocks(wpc, inode, file_offset);
@@ -1435,6 +1490,7 @@  iomap_writepage_map(struct iomap_writepage_ctx *wpc,
 		 */
 		set_page_writeback_keepwrite(page);
 	} else {
+		iomap_clear_range_dirty(page, 0, PAGE_SIZE);
 		clear_page_dirty_for_io(page);
 		set_page_writeback(page);
 	}