Uncached buffered IO

Hi,

5 years ago I posted patches adding support for RWF_UNCACHED, as a way to do buffered IO that isn't page cache persistent. The approach back then was to have private pages for IO, and then get rid of them once IO was done. But that then runs into all the issues that O_DIRECT has, in terms of synchronizing with the page cache.

So here's a new approach to the same concent, but using the page cache as synchronization. Due to excessive bike shedding on the naming, this is now named RWF_DONTCACHE, and is less special in that it's just page cache IO, except it prunes the ranges once IO is completed.

Why do this, you may ask? The tldr is that device speeds are only getting faster, while reclaim is not. Doing normal buffered IO can be very unpredictable, and suck up a lot of resources on the reclaim side. This leads people to use O_DIRECT as a work-around, which has its own set of restrictions in terms of size, offset, and length of IO. It's also inherently synchronous, and now you need async IO as well. While the latter isn't necessarily a big problem as we have good options available there, it also should not be a requirement when all you want to do is read or write some data without caching.

Even on desktop type systems, a normal NVMe device can fill the entire page cache in seconds. On the big system I used for testing, there's a lot more RAM, but also a lot more devices. As can be seen in some of the results in the following patches, you can still fill RAM in seconds even when there's 1TB of it. Hence this problem isn't solely a "big hyperscaler system" issue, it's common across the board.

Common for both reads and writes with RWF_DONTCACHE is that they use the page cache for IO. Reads work just like a normal buffered read would, with the only exception being that the touched ranges will get pruned after data has been copied. For writes, the ranges will get writeback kicked off before the syscall returns, and then writeback completion will prune the range. Hence writes aren't synchronous, and it's easy to pipeline writes using RWF_DONTCACHE. Folios that aren't instantiated by RWF_DONTCACHE IO are left untouched. This means you that uncached IO will take advantage of the page cache for uptodate data, but not leave anything it instantiated/created in cache.

File systems need to support this. This patchset adds support for the generic read path, which covers file systems like ext4. Patches exist to add support for iomap/XFS and btrfs as well, which sit on top of this series. If RWF_DONTCACHE IO is attempted on a file system that doesn't support it, -EOPNOTSUPP is returned. Hence the user can rely on it either working as designed, or flagging and error if that's not the case. The intent here is to give the application a sensible fallback path - eg, it may fall back to O_DIRECT if appropriate, or just live with the fact that uncached IO isn't available and do normal buffered IO.

Adding "support" to other file systems should be trivial, most of the time just a one-liner adding FOP_DONTCACHE to the fop_flags in the file_operations struct, if the file system is using either iomap or the generic filemap helpers for reading and writing.

Performance results are in patch 8 for reads, and you can find the write side results in the XFS patch adding support for DONTCACHE writes for XFS:

https://git.kernel.dk/cgit/linux/commit/?h=buffered-uncached-fs.10&id=257e92de795fdff7d7e256501e024fac6da6a7f4

with the tldr being that I see about a 65% improvement in performance for both, with fully predictable IO times. CPU reduction is substantial as well, with no kswapd activity at all for reclaim when using uncached IO.

Using it from applications is trivial - just set RWF_DONTCACHE for the read or write, using pwritev2(2) or preadv2(2). For io_uring, same thing, just set RWF_DONTCACHE in sqe->rw_flags for a buffered read/write operation. And that's it.

Patches 1..7 are just prep patches, and should have no functional changes at all. Patch 8 adds support for the filemap path for RWF_DONTCACHE reads, and patches 9..12 are just prep patches for supporting the write side of uncached writes. In the below mentioned branch, there are then patches to adopt uncached reads and writes for xfs, btrfs, and ext4. The latter currently relies on bit of a hack for passing whether this is an uncached write or not through ->write_begin(), which can hopefully go away once ext4 adopts iomap for buffered writes. I say this is a hack as it's not the prettiest way to do it, however it is fully solid and will work just fine.

Passes full xfstests and fsx overnight runs, no issues observed. That includes the vm running the testing also using RWF_DONTCACHE on the host. I'll post fsstress and fsx patches for RWF_DONTCACHE separately. As far as I'm concerned, no further work needs doing here.

And git tree for the patches is here:

https://git.kernel.dk/cgit/linux/log/?h=buffered-uncached.10

with the file system patches on top adding support for xfs/btrfs/ext4 here:

https://git.kernel.dk/cgit/linux/log/?h=buffered-uncached-fs.10

 include/linux/fs.h             |  21 ++++++-
 include/linux/page-flags.h     |   5 ++
 include/linux/pagemap.h        |   3 +
 include/trace/events/mmflags.h |   3 +-
 include/uapi/linux/fs.h        |   6 +-
 mm/filemap.c                   | 102 ++++++++++++++++++++++++++++-----
 mm/internal.h                  |   2 +
 mm/readahead.c                 |  22 +++++--
 mm/swap.c                      |   2 +
 mm/truncate.c                  |  53 +++++++++--------
 10 files changed, 173 insertions(+), 46 deletions(-)

Since v7
- Rename filemap_uncached_read() to filemap_end_dropbehind_read()
- Rename folio_end_dropbehind() to folio_end_dropbehind_write()
- Make the "mm: add FGP_DONTCACHE folio creation flag" patch part of the base patches series, to avoid dependencies with btrfs/xfs/iomap
- Remove now dead IOMAP_F_DONTCACHE define and setting on xfs/iomap
- Re-instate mistakenly deleted VM_BUG_ON_FOLIO() in invalidate_inode_pages2_range()
- Add reviewed-by's
- Add separate fs patch branch that sits on top of the core branch
- Rebase on current -git master

Jens Axboe

Rather than pass in both the file and position directly from the kiocb, just take a struct kiocb instead. With the kiocb being passed in, skip passing in the address_space separately as well. While doing so, move the ki_flags checking into filemap_create_folio() as well. In preparation for actually needing the kiocb in the function.

No functional changes in this patch.

---
 mm/filemap.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)

index f61cf51c2238..8b29323b15d7 100644
@@ -2459,15 +2459,17 @@ static int filemap_update_page(struct kiocb *iocb,
 	return error;
 }
 
-static int filemap_create_folio(struct file *file,
-		struct address_space *mapping, loff_t pos,
-		struct folio_batch *fbatch)
+static int filemap_create_folio(struct kiocb *iocb, struct folio_batch *fbatch)
 {
+	struct address_space *mapping = iocb->ki_filp->f_mapping;
 	struct folio *folio;
 	int error;
 	unsigned int min_order = mapping_min_folio_order(mapping);
 	pgoff_t index;
 
+	if (iocb->ki_flags & (IOCB_NOWAIT | IOCB_WAITQ))
+		return -EAGAIN;
+
 	folio = filemap_alloc_folio(mapping_gfp_mask(mapping), min_order);
 	if (!folio)
 		return -ENOMEM;

@@ -2486,7 +2488,7 @@ static int filemap_create_folio(struct file *file,
 	 * well to keep locking rules simple.
 	 */
 	filemap_invalidate_lock_shared(mapping);
-	index = (pos >> (PAGE_SHIFT + min_order)) << min_order;
+	index = (iocb->ki_pos >> (PAGE_SHIFT + min_order)) << min_order;
 	error = filemap_add_folio(mapping, folio, index,
 			mapping_gfp_constraint(mapping, GFP_KERNEL));
 	if (error == -EEXIST)

@@ -2494,7 +2496,8 @@ static int filemap_create_folio(struct file *file,
 	if (error)
 		goto error;
 
-	error = filemap_read_folio(file, mapping->a_ops->read_folio, folio);
+	error = filemap_read_folio(iocb->ki_filp, mapping->a_ops->read_folio,
+					folio);
 	if (error)
 		goto error;
 

@@ -2550,9 +2553,7 @@ static int filemap_get_pages(struct kiocb *iocb, size_t count,
 		filemap_get_read_batch(mapping, index, last_index - 1, fbatch);
 	}
 	if (!folio_batch_count(fbatch)) {
-		if (iocb->ki_flags & (IOCB_NOWAIT | IOCB_WAITQ))
-			return -EAGAIN;
-		err = filemap_create_folio(filp, mapping, iocb->ki_pos, fbatch);
+		err = filemap_create_folio(iocb, fbatch);
 		if (err == AOP_TRUNCATED_PAGE)
 			goto retry;
 		return err;

2.45.2

Rather than use the page_cache_sync_readahead() helper, define our own ractl and use page_cache_sync_ra() directly. In preparation for needing to modify ractl inside filemap_get_pages().

No functional changes in this patch.

---
 mm/filemap.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

index 8b29323b15d7..220dc7c6e12f 100644
@@ -2527,7 +2527,6 @@ static int filemap_get_pages(struct kiocb *iocb, size_t count,
 {
 	struct file *filp = iocb->ki_filp;
 	struct address_space *mapping = filp->f_mapping;
-	struct file_ra_state *ra = &filp->f_ra;
 	pgoff_t index = iocb->ki_pos >> PAGE_SHIFT;
 	pgoff_t last_index;
 	struct folio *folio;

@@ -2542,12 +2541,13 @@ static int filemap_get_pages(struct kiocb *iocb, size_t count,
 
 	filemap_get_read_batch(mapping, index, last_index - 1, fbatch);
 	if (!folio_batch_count(fbatch)) {
+		DEFINE_READAHEAD(ractl, filp, &filp->f_ra, mapping, index);
+
 		if (iocb->ki_flags & IOCB_NOIO)
 			return -EAGAIN;
 		if (iocb->ki_flags & IOCB_NOWAIT)
 			flags = memalloc_noio_save();
-		page_cache_sync_readahead(mapping, ra, filp, index,
-				last_index - index);
+		page_cache_sync_ra(&ractl, last_index - index);
 		if (iocb->ki_flags & IOCB_NOWAIT)
 			memalloc_noio_restore(flags);
 		filemap_get_read_batch(mapping, index, last_index - 1, fbatch);

2.45.2

Just a wrapper around filemap_alloc_folio() for now, but add it in preparation for modifying the folio based on the 'ractl' being passed in.

No functional changes in this patch.

---
 mm/readahead.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

index ea650b8b02fb..8a62ad4106ff 100644
@@ -188,6 +188,12 @@ static void read_pages(struct readahead_control *rac)
 	BUG_ON(readahead_count(rac));
 }
 
+static struct folio *ractl_alloc_folio(struct readahead_control *ractl,
+				       gfp_t gfp_mask, unsigned int order)
+{
+	return filemap_alloc_folio(gfp_mask, order);
+}
+
 /**
  * page_cache_ra_unbounded - Start unchecked readahead.
  * @ractl: Readahead control.

@@ -265,8 +271,8 @@ void page_cache_ra_unbounded(struct readahead_control *ractl,
 			continue;
 		}
 
-		folio = filemap_alloc_folio(gfp_mask,
-					    mapping_min_folio_order(mapping));
+		folio = ractl_alloc_folio(ractl, gfp_mask,
+					mapping_min_folio_order(mapping));
 		if (!folio)
 			break;
 

@@ -436,7 +442,7 @@ static inline int ra_alloc_folio(struct readahead_control *ractl, pgoff_t index,
 		pgoff_t mark, unsigned int order, gfp_t gfp)
 {
 	int err;
-	struct folio *folio = filemap_alloc_folio(gfp, order);
+	struct folio *folio = ractl_alloc_folio(ractl, gfp, order);
 
 	if (!folio)
 		return -ENOMEM;

@@ -750,7 +756,7 @@ void readahead_expand(struct readahead_control *ractl,
 		if (folio && !xa_is_value(folio))
 			return; /* Folio apparently present */
 
-		folio = filemap_alloc_folio(gfp_mask, min_order);
+		folio = ractl_alloc_folio(ractl, gfp_mask, min_order);
 		if (!folio)
 			return;
 

@@ -779,7 +785,7 @@ void readahead_expand(struct readahead_control *ractl,
 		if (folio && !xa_is_value(folio))
 			return; /* Folio apparently present */
 
-		folio = filemap_alloc_folio(gfp_mask, min_order);
+		folio = ractl_alloc_folio(ractl, gfp_mask, min_order);
 		if (!folio)
 			return;
 

2.45.2

Add a folio flag that file IO can use to indicate that the cached IO being done should be dropped from the page cache upon completion.

---
 include/linux/page-flags.h     | 5 +++++
 include/trace/events/mmflags.h | 3 ++-
 2 files changed, 7 insertions(+), 1 deletion(-)

index cf46ac720802..16607f02abd0 100644
@@ -110,6 +110,7 @@ enum pageflags {
 	PG_reclaim,		/* To be reclaimed asap */
 	PG_swapbacked,		/* Page is backed by RAM/swap */
 	PG_unevictable,		/* Page is "unevictable"  */
+	PG_dropbehind,		/* drop pages on IO completion */
 #ifdef CONFIG_MMU
 	PG_mlocked,		/* Page is vma mlocked */
 #endif

@@ -562,6 +563,10 @@ PAGEFLAG(Reclaim, reclaim, PF_NO_TAIL)
 FOLIO_FLAG(readahead, FOLIO_HEAD_PAGE)
 	FOLIO_TEST_CLEAR_FLAG(readahead, FOLIO_HEAD_PAGE)
 
+FOLIO_FLAG(dropbehind, FOLIO_HEAD_PAGE)
+	FOLIO_TEST_CLEAR_FLAG(dropbehind, FOLIO_HEAD_PAGE)
+	__FOLIO_SET_FLAG(dropbehind, FOLIO_HEAD_PAGE)
+
 #ifdef CONFIG_HIGHMEM
 /*
  * Must use a macro here due to header dependency issues. page_zone() is not

index bb8a59c6caa2..3bc8656c8359 100644
@@ -116,7 +116,8 @@
 	DEF_PAGEFLAG_NAME(head),					\
 	DEF_PAGEFLAG_NAME(reclaim),					\
 	DEF_PAGEFLAG_NAME(swapbacked),					\
-	DEF_PAGEFLAG_NAME(unevictable)					\
+	DEF_PAGEFLAG_NAME(unevictable),					\
+	DEF_PAGEFLAG_NAME(dropbehind)					\
 IF_HAVE_PG_MLOCK(mlocked)						\
 IF_HAVE_PG_HWPOISON(hwpoison)						\
 IF_HAVE_PG_IDLE(idle)							\

2.45.2

If ractl->dropbehind is set to true, then folios created are marked as dropbehind as well.

---
 include/linux/pagemap.h | 1 +
 mm/readahead.c          | 8 +++++++-
 2 files changed, 8 insertions(+), 1 deletion(-)

index bcf0865a38ae..5da4b6d42fae 100644
@@ -1353,6 +1353,7 @@ struct readahead_control {
 	pgoff_t _index;
 	unsigned int _nr_pages;
 	unsigned int _batch_count;
+	bool dropbehind;
 	bool _workingset;
 	unsigned long _pflags;
 };

index 8a62ad4106ff..c0a6dc5d5686 100644
@@ -191,7 +191,13 @@ static void read_pages(struct readahead_control *rac)
 static struct folio *ractl_alloc_folio(struct readahead_control *ractl,
 				       gfp_t gfp_mask, unsigned int order)
 {
-	return filemap_alloc_folio(gfp_mask, order);
+	struct folio *folio;
+
+	folio = filemap_alloc_folio(gfp_mask, order);
+	if (folio && ractl->dropbehind)
+		__folio_set_dropbehind(folio);
+
+	return folio;
 }
 
 /**

2.45.2

Add a folio_unmap_invalidate() helper, which unmaps and invalidates a given folio. The caller must already have locked the folio. Embed the old invalidate_complete_folio2() helper in there as well, as nobody else calls it.

Use this new helper in invalidate_inode_pages2_range(), rather than duplicate the code there.

In preparation for using this elsewhere as well, have it take a gfp_t mask rather than assume GFP_KERNEL is the right choice. This bubbles back to invalidate_complete_folio2() as well.

---
 mm/internal.h |  2 ++
 mm/truncate.c | 53 +++++++++++++++++++++++++++------------------------
 2 files changed, 30 insertions(+), 25 deletions(-)

index cb8d8e8e3ffa..ed3c3690eb03 100644
@@ -392,6 +392,8 @@ void unmap_page_range(struct mmu_gather *tlb,
 			     struct vm_area_struct *vma,
 			     unsigned long addr, unsigned long end,
 			     struct zap_details *details);
+int folio_unmap_invalidate(struct address_space *mapping, struct folio *folio,
+			   gfp_t gfp);
 
 void page_cache_ra_order(struct readahead_control *, struct file_ra_state *,
 		unsigned int order);

2.45.2

It's simply doing what the code did before. I do agree the mapping check is a bit odd at that point, but that's how invalidate_inode_pages2_range() and folio_launder() was setup. We can certainly clean that up after the merge of these helpers, but I didn't want to introduce any potential changes with this merge.

-EBUSY was the return from a 0 return from those two helpers before.

Jens Axboe

Any further concerns with this? Trying to nudge this patchset forward... It's not like there's a lot of time left for 6.14.

Jens Axboe

If a file system supports uncached buffered IO, it may set FOP_DONTCACHE and enable support for RWF_DONTCACHE. If RWF_DONTCACHE is attempted without the file system supporting it, it'll get errored with -EOPNOTSUPP.

---
 include/linux/fs.h      | 14 +++++++++++++-
 include/uapi/linux/fs.h |  6 +++++-
 2 files changed, 18 insertions(+), 2 deletions(-)

index 7e29433c5ecc..6a838b5479a6 100644
@@ -322,6 +322,7 @@ struct readahead_control;
 #define IOCB_NOWAIT		(__force int) RWF_NOWAIT
 #define IOCB_APPEND		(__force int) RWF_APPEND
 #define IOCB_ATOMIC		(__force int) RWF_ATOMIC
+#define IOCB_DONTCACHE		(__force int) RWF_DONTCACHE
 
 /* non-RWF related bits - start at 16 */
 #define IOCB_EVENTFD		(1 << 16)

@@ -356,7 +357,8 @@ struct readahead_control;
 	{ IOCB_SYNC,		"SYNC" }, \
 	{ IOCB_NOWAIT,		"NOWAIT" }, \
 	{ IOCB_APPEND,		"APPEND" }, \
-	{ IOCB_ATOMIC,		"ATOMIC"}, \
+	{ IOCB_ATOMIC,		"ATOMIC" }, \
+	{ IOCB_DONTCACHE,	"DONTCACHE" }, \
 	{ IOCB_EVENTFD,		"EVENTFD"}, \
 	{ IOCB_DIRECT,		"DIRECT" }, \
 	{ IOCB_WRITE,		"WRITE" }, \

@@ -2127,6 +2129,8 @@ struct file_operations {
 #define FOP_UNSIGNED_OFFSET	((__force fop_flags_t)(1 << 5))
 /* Supports asynchronous lock callbacks */
 #define FOP_ASYNC_LOCK		((__force fop_flags_t)(1 << 6))
+/* File system supports uncached read/write buffered IO */
+#define FOP_DONTCACHE		((__force fop_flags_t)(1 << 7))
 
 /* Wrap a directory iterator that needs exclusive inode access */
 int wrap_directory_iterator(struct file *, struct dir_context *,

@@ -3614,6 +3618,14 @@ static inline int kiocb_set_rw_flags(struct kiocb *ki, rwf_t flags,
 		if (!(ki->ki_filp->f_mode & FMODE_CAN_ATOMIC_WRITE))
 			return -EOPNOTSUPP;
 	}
+	if (flags & RWF_DONTCACHE) {
+		/* file system must support it */
+		if (!(ki->ki_filp->f_op->fop_flags & FOP_DONTCACHE))
+			return -EOPNOTSUPP;
+		/* DAX mappings not supported */
+		if (IS_DAX(ki->ki_filp->f_mapping->host))
+			return -EOPNOTSUPP;
+	}
 	kiocb_flags |= (__force int) (flags & RWF_SUPPORTED);
 	if (flags & RWF_SYNC)
 		kiocb_flags |= IOCB_DSYNC;

index 753971770733..56a4f93a08f4 100644
@@ -332,9 +332,13 @@ typedef int __bitwise __kernel_rwf_t;
 /* Atomic Write */
 #define RWF_ATOMIC	((__force __kernel_rwf_t)0x00000040)
 
+/* buffered IO that drops the cache after reading or writing data */
+#define RWF_DONTCACHE	((__force __kernel_rwf_t)0x00000080)
+
 /* mask of flags supported by the kernel */
 #define RWF_SUPPORTED	(RWF_HIPRI | RWF_DSYNC | RWF_SYNC | RWF_NOWAIT |\
-			 RWF_APPEND | RWF_NOAPPEND | RWF_ATOMIC)
+			 RWF_APPEND | RWF_NOAPPEND | RWF_ATOMIC |\
+			 RWF_DONTCACHE)
 
 #define PROCFS_IOCTL_MAGIC 'f'
 

2.45.2

Jens, you can use this as a stable branch for the VFS changes.

---

Applied to the vfs-6.14.uncached_buffered_io branch of the vfs/vfs.git tree. Patches in the vfs-6.14.uncached_buffered_io branch should appear in linux-next soon.

Please report any outstanding bugs that were missed during review in a new review to the original patch series allowing us to drop it.

It's encouraged to provide Acked-bys and Reviewed-bys even though the patch has now been applied. If possible patch trailers will be updated.

Note that commit hashes shown below are subject to change due to rebase, trailer updates or similar. If in doubt, please check the listed branch.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs-6.14.uncached_buffered_io

[07/12] fs: add RWF_DONTCACHE iocb and FOP_DONTCACHE file_operations flag

        https://git.kernel.org/vfs/vfs/c/af6505e5745b

Thanks, but not sure it's super useful for a patch in the middle. It just adds a dependency, and I'd need to rebase. Providing an ack or reviewed by would probably be more useful.

But given how slow the mm side is going, honestly not sure what the target or plan is for inclusion at this point in time.

Jens Axboe

Add RWF_DONTCACHE as a read operation flag, which means that any data read wil be removed from the page cache upon completion. Uses the page cache to synchronize, and simply prunes folios that were instantiated when the operation completes. While it would be possible to use private pages for this, using the page cache as synchronization is handy for a variety of reasons:

1) No special truncate magic is needed
2) Async buffered reads need some place to serialize, using the page cache is a lot easier than writing extra code for this
3) The pruning cost is pretty reasonable

and the code to support this is much simpler as a result.

You can think of uncached buffered IO as being the much more attractive cousin of O_DIRECT - it has none of the restrictions of O_DIRECT. Yes, it will copy the data, but unlike regular buffered IO, it doesn't run into the unpredictability of the page cache in terms of reclaim. As an example, on a test box with 32 drives, reading them with buffered IO looks as follows:

Reading bs 65536, uncached 0
1s: 145945MB/sec
2s: 158067MB/sec
3s: 157007MB/sec
4s: 148622MB/sec
5s: 118824MB/sec
6s: 70494MB/sec
7s: 41754MB/sec
8s: 90811MB/sec
9s: 92204MB/sec
10s: 95178MB/sec
11s: 95488MB/sec
12s: 95552MB/sec
13s: 96275MB/sec

where it's quite easy to see where the page cache filled up, and performance went from good to erratic, and finally settles at a much lower rate. Looking at top while this is ongoing, we see:

 PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND
7535 root      20   0  267004      0      0 S  3199   0.0   8:40.65 uncached
3326 root      20   0       0      0      0 R 100.0   0.0   0:16.40 kswapd4
3327 root      20   0       0      0      0 R 100.0   0.0   0:17.22 kswapd5
3328 root      20   0       0      0      0 R 100.0   0.0   0:13.29 kswapd6
3332 root      20   0       0      0      0 R 100.0   0.0   0:11.11 kswapd10
3339 root      20   0       0      0      0 R 100.0   0.0   0:16.25 kswapd17
3348 root      20   0       0      0      0 R 100.0   0.0   0:16.40 kswapd26
3343 root      20   0       0      0      0 R 100.0   0.0   0:16.30 kswapd21
3344 root      20   0       0      0      0 R 100.0   0.0   0:11.92 kswapd22
3349 root      20   0       0      0      0 R 100.0   0.0   0:16.28 kswapd27
3352 root      20   0       0      0      0 R  99.7   0.0   0:11.89 kswapd30
3353 root      20   0       0      0      0 R  96.7   0.0   0:16.04 kswapd31
3329 root      20   0       0      0      0 R  96.4   0.0   0:11.41 kswapd7
3345 root      20   0       0      0      0 R  96.4   0.0   0:13.40 kswapd23
3330 root      20   0       0      0      0 S  91.1   0.0   0:08.28 kswapd8
3350 root      20   0       0      0      0 S  86.8   0.0   0:11.13 kswapd28
3325 root      20   0       0      0      0 S  76.3   0.0   0:07.43 kswapd3
3341 root      20   0       0      0      0 S  74.7   0.0   0:08.85 kswapd19
3334 root      20   0       0      0      0 S  71.7   0.0   0:10.04 kswapd12
3351 root      20   0       0      0      0 R  60.5   0.0   0:09.59 kswapd29
3323 root      20   0       0      0      0 R  57.6   0.0   0:11.50 kswapd1

[...]

which is just showing a partial list of the 32 kswapd threads that are running mostly full tilt, burning ~28 full CPU cores.

If the same test case is run with RWF_DONTCACHE set for the buffered read, the output looks as follows:

Reading bs 65536, uncached 0
1s: 153144MB/sec
2s: 156760MB/sec
3s: 158110MB/sec
4s: 158009MB/sec
5s: 158043MB/sec
6s: 157638MB/sec
7s: 157999MB/sec
8s: 158024MB/sec
9s: 157764MB/sec
10s: 157477MB/sec
11s: 157417MB/sec
12s: 157455MB/sec
13s: 157233MB/sec
14s: 156692MB/sec

which is just chugging along at ~155GB/sec of read performance. Looking at top, we see:

 PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND
7961 root      20   0  267004      0      0 S  3180   0.0   5:37.95 uncached
8024 axboe     20   0   14292   4096      0 R   1.0   0.0   0:00.13 top

where just the test app is using CPU, no reclaim is taking place outside of the main thread. Not only is performance 65% better, it's also using half the CPU to do it.

---
 mm/filemap.c | 28 ++++++++++++++++++++++++++--
 mm/swap.c    |  2 ++
 2 files changed, 28 insertions(+), 2 deletions(-)

index 220dc7c6e12f..dd563208d09d 100644
@@ -2473,6 +2473,8 @@ static int filemap_create_folio(struct kiocb *iocb, struct folio_batch *fbatch)
 	folio = filemap_alloc_folio(mapping_gfp_mask(mapping), min_order);
 	if (!folio)
 		return -ENOMEM;
+	if (iocb->ki_flags & IOCB_DONTCACHE)
+		__folio_set_dropbehind(folio);
 
 	/*
 	 * Protect against truncate / hole punch. Grabbing invalidate_lock

@@ -2518,6 +2520,8 @@ static int filemap_readahead(struct kiocb *iocb, struct file *file,
 
 	if (iocb->ki_flags & IOCB_NOIO)
 		return -EAGAIN;
+	if (iocb->ki_flags & IOCB_DONTCACHE)
+		ractl.dropbehind = 1;
 	page_cache_async_ra(&ractl, folio, last_index - folio->index);
 	return 0;
 }

@@ -2547,6 +2551,8 @@ static int filemap_get_pages(struct kiocb *iocb, size_t count,
 			return -EAGAIN;
 		if (iocb->ki_flags & IOCB_NOWAIT)
 			flags = memalloc_noio_save();
+		if (iocb->ki_flags & IOCB_DONTCACHE)
+			ractl.dropbehind = 1;
 		page_cache_sync_ra(&ractl, last_index - index);
 		if (iocb->ki_flags & IOCB_NOWAIT)
 			memalloc_noio_restore(flags);

@@ -2594,6 +2600,20 @@ static inline bool pos_same_folio(loff_t pos1, loff_t pos2, struct folio *folio)
 	return (pos1 >> shift == pos2 >> shift);
 }
 
+static void filemap_end_dropbehind_read(struct address_space *mapping,
+					struct folio *folio)
+{
+	if (!folio_test_dropbehind(folio))
+		return;
+	if (folio_test_writeback(folio) || folio_test_dirty(folio))
+		return;
+	if (folio_trylock(folio)) {
+		if (folio_test_clear_dropbehind(folio))
+			folio_unmap_invalidate(mapping, folio, 0);
+		folio_unlock(folio);
+	}
+}
+
 /**
  * filemap_read - Read data from the page cache.
  * @iocb: The iocb to read.

@@ -2707,8 +2727,12 @@ ssize_t filemap_read(struct kiocb *iocb, struct iov_iter *iter,
 			}
 		}
 put_folios:
-		for (i = 0; i < folio_batch_count(&fbatch); i++)
-			folio_put(fbatch.folios[i]);
+		for (i = 0; i < folio_batch_count(&fbatch); i++) {
+			struct folio *folio = fbatch.folios[i];
+
+			filemap_end_dropbehind_read(mapping, folio);
+			folio_put(folio);
+		}
 		folio_batch_init(&fbatch);
 	} while (iov_iter_count(iter) && iocb->ki_pos < isize && !error);
 

index 10decd9dffa1..ba02bd5ba145 100644
@@ -427,6 +427,8 @@ static void folio_inc_refs(struct folio *folio)
  */
 void folio_mark_accessed(struct folio *folio)
 {
+	if (folio_test_dropbehind(folio))
+		return;
 	if (lru_gen_enabled()) {
 		folio_inc_refs(folio);
 		return;

2.45.2

If the folio is marked as streaming, drop pages when writeback completes. Intended to be used with RWF_DONTCACHE, to avoid needing sync writes for uncached IO.

---
 mm/filemap.c | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

index dd563208d09d..aa0b3af6533d 100644
@@ -1599,6 +1599,27 @@ int folio_wait_private_2_killable(struct folio *folio)
 }
 EXPORT_SYMBOL(folio_wait_private_2_killable);
 
+/*
+ * If folio was marked as dropbehind, then pages should be dropped when writeback
+ * completes. Do that now. If we fail, it's likely because of a big folio -
+ * just reset dropbehind for that case and latter completions should invalidate.
+ */
+static void folio_end_dropbehind_write(struct folio *folio)
+{
+	/*
+	 * Hitting !in_task() should not happen off RWF_DONTCACHE writeback,
+	 * but can happen if normal writeback just happens to find dirty folios
+	 * that were created as part of uncached writeback, and that writeback
+	 * would otherwise not need non-IRQ handling. Just skip the
+	 * invalidation in that case.
+	 */
+	if (in_task() && folio_trylock(folio)) {
+		if (folio->mapping)
+			folio_unmap_invalidate(folio->mapping, folio, 0);
+		folio_unlock(folio);
+	}
+}
+
 /**
  * folio_end_writeback - End writeback against a folio.
  * @folio: The folio.

@@ -1609,6 +1630,8 @@ EXPORT_SYMBOL(folio_wait_private_2_killable);
  */
 void folio_end_writeback(struct folio *folio)
 {
+	bool folio_dropbehind = false;
+
 	VM_BUG_ON_FOLIO(!folio_test_writeback(folio), folio);
 
 	/*

@@ -1630,9 +1653,14 @@ void folio_end_writeback(struct folio *folio)
 	 * reused before the folio_wake_bit().
 	 */
 	folio_get(folio);
+	if (!folio_test_dirty(folio))
+		folio_dropbehind = folio_test_clear_dropbehind(folio);
 	if (__folio_end_writeback(folio))
 		folio_wake_bit(folio, PG_writeback);
 	acct_reclaim_writeback(folio);
+
+	if (folio_dropbehind)
+		folio_end_dropbehind_write(folio);
 	folio_put(folio);
 }
 EXPORT_SYMBOL(folio_end_writeback);

2.45.2

Works like filemap_fdatawrite_range(), except it's a non-integrity data writeback and hence only starts writeback on the specified range. Will help facilitate generically starting uncached writeback from generic_write_sync(), as header dependencies preclude doing this inline from fs.h.

---
 include/linux/fs.h |  2 ++
 mm/filemap.c       | 18 ++++++++++++++++++
 2 files changed, 20 insertions(+)

index 6a838b5479a6..653b5efa3d3f 100644
@@ -2878,6 +2878,8 @@ extern int __must_check file_fdatawait_range(struct file *file, loff_t lstart,
 extern int __must_check file_check_and_advance_wb_err(struct file *file);
 extern int __must_check file_write_and_wait_range(struct file *file,
 						loff_t start, loff_t end);
+int filemap_fdatawrite_range_kick(struct address_space *mapping, loff_t start,
+		loff_t end);
 
 static inline int file_write_and_wait(struct file *file)
 {

index aa0b3af6533d..9842258ba343 100644
@@ -449,6 +449,24 @@ int filemap_fdatawrite_range(struct address_space *mapping, loff_t start,
 }
 EXPORT_SYMBOL(filemap_fdatawrite_range);
 
+/**
+ * filemap_fdatawrite_range_kick - start writeback on a range
+ * @mapping:	target address_space
+ * @start:	index to start writeback on
+ * @end:	last (non-inclusive) index for writeback
+ *
+ * This is a non-integrity writeback helper, to start writing back folios
+ * for the indicated range.
+ *
+ * Return: %0 on success, negative error code otherwise.
+ */
+int filemap_fdatawrite_range_kick(struct address_space *mapping, loff_t start,
+				  loff_t end)
+{
+	return __filemap_fdatawrite_range(mapping, start, end, WB_SYNC_NONE);
+}
+EXPORT_SYMBOL_GPL(filemap_fdatawrite_range_kick);
+
 /**
  * filemap_flush - mostly a non-blocking flush
  * @mapping:	target address_space

2.45.2

When a buffered write submitted with IOCB_DONTCACHE has been successfully submitted, call filemap_fdatawrite_range_kick() to kick off the IO. File systems call generic_write_sync() for any successful buffered write submission, hence add the logic here rather than needing to modify the file system.

---
 include/linux/fs.h | 5 +++++
 1 file changed, 5 insertions(+)

index 653b5efa3d3f..58a618853574 100644
@@ -2912,6 +2912,11 @@ static inline ssize_t generic_write_sync(struct kiocb *iocb, ssize_t count)
 				(iocb->ki_flags & IOCB_SYNC) ? 0 : 1);
 		if (ret)
 			return ret;
+	} else if (iocb->ki_flags & IOCB_DONTCACHE) {
+		struct address_space *mapping = iocb->ki_filp->f_mapping;
+
+		filemap_fdatawrite_range_kick(mapping, iocb->ki_pos,
+					      iocb->ki_pos + count);
 	}
 
 	return count;

2.45.2

Callers can pass this in for uncached folio creation, in which case if a folio is newly created it gets marked as uncached. If a folio exists for this index and lookup succeeds, then it will not get marked as uncached. If an !uncached lookup finds a cached folio, clear the flag. For that case, there are competeting uncached and cached users of the folio, and it should not get pruned.

---
 include/linux/pagemap.h | 2 ++
 mm/filemap.c            | 5 +++++
 2 files changed, 7 insertions(+)

index 5da4b6d42fae..64c6dada837e 100644
@@ -710,6 +710,7 @@ pgoff_t page_cache_prev_miss(struct address_space *mapping,
  * * %FGP_NOFS - __GFP_FS will get cleared in gfp.
  * * %FGP_NOWAIT - Don't block on the folio lock.
  * * %FGP_STABLE - Wait for the folio to be stable (finished writeback)
+ * * %FGP_DONTCACHE - Uncached buffered IO
  * * %FGP_WRITEBEGIN - The flags to use in a filesystem write_begin()
  *   implementation.
  */

@@ -723,6 +724,7 @@ typedef unsigned int __bitwise fgf_t;
 #define FGP_NOWAIT		((__force fgf_t)0x00000020)
 #define FGP_FOR_MMAP		((__force fgf_t)0x00000040)
 #define FGP_STABLE		((__force fgf_t)0x00000080)
+#define FGP_DONTCACHE		((__force fgf_t)0x00000100)
 #define FGF_GET_ORDER(fgf)	(((__force unsigned)fgf) >> 26)	/* top 6 bits */
 
 #define FGP_WRITEBEGIN		(FGP_LOCK | FGP_WRITE | FGP_CREAT | FGP_STABLE)

index 9842258ba343..68bdfff4117e 100644
@@ -2001,6 +2001,8 @@ struct folio *__filemap_get_folio(struct address_space *mapping, pgoff_t index,
 			/* Init accessed so avoid atomic mark_page_accessed later */
 			if (fgp_flags & FGP_ACCESSED)
 				__folio_set_referenced(folio);
+			if (fgp_flags & FGP_DONTCACHE)
+				__folio_set_dropbehind(folio);
 
 			err = filemap_add_folio(mapping, folio, index, gfp);
 			if (!err)

@@ -2023,6 +2025,9 @@ struct folio *__filemap_get_folio(struct address_space *mapping, pgoff_t index,
 
 	if (!folio)
 		return ERR_PTR(-ENOENT);
+	/* not an uncached lookup, clear uncached if set */
+	if (folio_test_dropbehind(folio) && !(fgp_flags & FGP_DONTCACHE))
+		folio_clear_dropbehind(folio);
 	return folio;
 }
 EXPORT_SYMBOL(__filemap_get_folio);

2.45.2

Of course, we're doing something here which userspace could itself do:
drop the pagecache after reading it (with appropriate chunk sizing) and for writes, sync the written area then invalidate it. Possible added benefits from using separate threads for this.

I suggest that diligence requires that we at least justify an in-kernel approach at this time, please.

And there's a possible middle-ground implementation where the kernel itself kicks off threads to do the drop-behind just before the read or write syscall returns, which will probably be simpler. Can we please describe why this also isn't acceptable?

Also, it seems wrong for a read(RWF_DONTCACHE) to drop cache if it was already present. Because it was presumably present for a reason. Does this implementation already take care of this? To make an application which does read(/etc/passwd, RWF_DONTCACHE) less annoying?

Also, consuming a new page flag isn't a minor thing. It would be nice to see some justification around this, and some decription of how many we have left.

(sorry missed this reply!)

Conceptually yes. But you'd end up doing extra work to do it. Some of that not so expensive, like system calls, and others more so, like LRU manipulation. Outside of that, I do think it makes sense to expose as a generic thing, rather than require applications needing to kick writeback manually, reclaim manually, etc.

That's more of an implementation detail. I didn't test anything like that, though we surely could. If it's better, there's no reason why it can't just be changed to do that. My gut tells me you want the task/CPU that just did the page cache additions to do the pruning to, that should be more efficient than having a kworker or similar do it.

The implementation doesn't drop pages that were already present, only pages that got created/added to the page cache for the operation. So that part should already work as you expect.

For sure, though various discussions on this already occurred and Kirill posted patches for unifying some of this already. It's not something I wanted to tackle, as I think that should be left to people more familiar with the page/folio flags and they (sometimes odd) interactions.

Jens Axboe

Well, gut might be wrong ;)

There may be benefit in using different CPUs to perform the dropbehind, rather than making the read() caller do this synchronously.

If I understand correctly, the write() dropbehind is performed at interrupt (write completion) time so that's already async.

Matthew & Kirill: are you OK with merging this as-is and then revisiting the page-flag consumption at a later time?

A gut this big is rarely wrong ;-)

It does, but we could actually get rid of that, at least when called via io_uring. From the testing I've done, doing it inline it definitely superior. Though it will depend on if you care about overall efficiency or just sheer speed/overhead of the read/write itself.

Jens Axboe

I have tried to find a way to avoid adding a new flag bit, but I have not found one. I am okay with merging it as it is.

  Kiryl Shutsemau / Kirill A. Shutemov