Skip to contentSkip to conversation

iomap: add dirty page control to iomap_zero_iter

PATCH Superseded Chi Zhiling 2 participants 3 messages May 10 – May 11

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 iomap: add dirty page control to iomap_zero_iter

From: Chi Zhiling <[email protected]>

This patch prepares the iomap framework for exFAT's upcoming migration to iomap. During testing of the exFAT iomap branch with xfstests generic/299 on a VM with 8GB RAM and a 40GB disk, system unresponsiveness was observed.

iomap_zero_iter() lacked dirty page throttling, which could cause memory pressure when exFAT's valid_size mechanism triggers large-scale zeroing operations during writes beyond valid_size.

Align iomap_zero_iter() with iomap_write_iter() by adding balance_dirty_pages_ratelimited() to throttle dirty page generation during large zeroing operations. Also add cond_resched() to provide voluntary rescheduling points during long-running loops.

---

exFAT iomap migration:
https://lore.kernel.org/linux-fsdevel/[email protected]/

 fs/iomap/buffered-io.c | 4 ++++
 1 file changed, 4 insertions(+)
fs/iomap/buffered-io.c +4 -0
index d7b648421a70..f6955786c8ad 100644
@@ -1543,6 +1543,8 @@ static int iomap_zero_iter(struct iomap_iter *iter, bool *did_zero,
 		size_t offset;
 		bool ret;
 
+		balance_dirty_pages_ratelimited(iter->inode->i_mapping);
+
 		bytes = min_t(u64, SIZE_MAX, bytes);
 		status = iomap_write_begin(iter, write_ops, &folio, &offset,
 				&bytes);
@@ -1571,6 +1573,8 @@ static int iomap_zero_iter(struct iomap_iter *iter, bool *did_zero,
 		if (WARN_ON_ONCE(!ret))
 			return -EIO;
 
+		cond_resched();
+
 		status = iomap_iter_advance(iter, bytes);
 		if (status)
 			break;
Signature · 1 line
2.43.0
#1 lore raw
Quoted 6 lines · Chi Zhiling

From: Chi Zhiling <[email protected]>

This patch prepares the iomap framework for exFAT's upcoming migration to iomap. During testing of the exFAT iomap branch with xfstests generic/299 on a VM with 8GB RAM and a 40GB disk, system unresponsiveness was observed.

Quoted 8 lines

iomap_zero_iter() lacked dirty page throttling, which could cause memory pressure when exFAT's valid_size mechanism triggers large-scale zeroing operations during writes beyond valid_size.

Align iomap_zero_iter() with iomap_write_iter() by adding balance_dirty_pages_ratelimited() to throttle dirty page generation during large zeroing operations. Also add cond_resched() to provide voluntary rescheduling points during long-running loops.

The balance_dirty_pages_ratelimited looks good. Now with lazy preempt we really should not need more cond_resched calls, though.

Quoted 35 lines

---

exFAT iomap migration:
https://lore.kernel.org/linux-fsdevel/[email protected]/

 fs/iomap/buffered-io.c | 4 ++++
 1 file changed, 4 insertions(+)
fs/iomap/buffered-io.c +4 -0
index d7b648421a70..f6955786c8ad 100644
@@ -1543,6 +1543,8 @@ static int iomap_zero_iter(struct iomap_iter *iter, bool *did_zero,
 		size_t offset;
 		bool ret;
 
+		balance_dirty_pages_ratelimited(iter->inode->i_mapping);
+
 		bytes = min_t(u64, SIZE_MAX, bytes);
 		status = iomap_write_begin(iter, write_ops, &folio, &offset,
 				&bytes);
@@ -1571,6 +1573,8 @@ static int iomap_zero_iter(struct iomap_iter *iter, bool *did_zero,
 		if (WARN_ON_ONCE(!ret))
 			return -EIO;
 
+		cond_resched();
+
 		status = iomap_iter_advance(iter, bytes);
 		if (status)
 			break;
Signature · 1 line
2.43.0

---end quoted text---

#2 lore raw
Quoted 11 lines · Christoph Hellwig
Quoted 10 lines · Chi Zhiling

From: Chi Zhiling <[email protected]>

This patch prepares the iomap framework for exFAT's upcoming migration to iomap. During testing of the exFAT iomap branch with xfstests generic/299 on a VM with 8GB RAM and a 40GB disk, system unresponsiveness was observed.

iomap_zero_iter() lacked dirty page throttling, which could cause memory pressure when exFAT's valid_size mechanism triggers large-scale zeroing operations during writes beyond valid_size.

Quoted 7 lines
Quoted 4 lines

Align iomap_zero_iter() with iomap_write_iter() by adding balance_dirty_pages_ratelimited() to throttle dirty page generation during large zeroing operations. Also add cond_resched() to provide voluntary rescheduling points during long-running loops.

The balance_dirty_pages_ratelimited looks good. Now with lazy preempt we really should not need more cond_resched calls, though.

Okay, will drop it in next version.

Thanks,

Quoted 37 lines
Quoted 36 lines

---

exFAT iomap migration:
https://lore.kernel.org/linux-fsdevel/[email protected]/

  fs/iomap/buffered-io.c | 4 ++++
  1 file changed, 4 insertions(+)
fs/iomap/buffered-io.c +4 -0
index d7b648421a70..f6955786c8ad 100644
@@ -1543,6 +1543,8 @@ static int iomap_zero_iter(struct iomap_iter *iter, bool *did_zero,
  		size_t offset;
  		bool ret;
 
+		balance_dirty_pages_ratelimited(iter->inode->i_mapping);
+
  		bytes = min_t(u64, SIZE_MAX, bytes);
  		status = iomap_write_begin(iter, write_ops, &folio, &offset,
  				&bytes);
@@ -1571,6 +1573,8 @@ static int iomap_zero_iter(struct iomap_iter *iter, bool *did_zero,
  		if (WARN_ON_ONCE(!ret))
  			return -EIO;
 
+		cond_resched();
+
  		status = iomap_iter_advance(iter, bytes);
  		if (status)
  			break;
Signature · 1 line
2.43.0

---end quoted text---