Currently mapping_try_invalidate() and invalidate_inode_pages2_range()
traverses the xarray in batches and then for each batch, maintains and
set the flag named xa_has_values if the batch has a shadow entry to
clear the entries at the end of the iteration. However they forgot to
reset the flag at the end of the iteration which cause them to always
try to clear the shadow entries in the subsequent iterations where
there might not be any shadow entries. Fixing it.

Fixes: 61c663e020d2 ("mm/truncate: batch-clear shadow entries")
Signed-off-by: Shakeel Butt <shakeel.butt@linux.dev>
---
 mm/truncate.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/mm/truncate.c b/mm/truncate.c
index 520c8cf8f58f..e5151703ba04 100644
--- a/mm/truncate.c
+++ b/mm/truncate.c
@@ -463,10 +463,10 @@ unsigned long mapping_try_invalidate(struct address_space *mapping,
 	unsigned long ret;
 	unsigned long count = 0;
 	int i;
-	bool xa_has_values = false;
 
 	folio_batch_init(&fbatch);
 	while (find_lock_entries(mapping, &index, end, &fbatch, indices)) {
+		bool xa_has_values = false;
 		int nr = folio_batch_count(&fbatch);
 
 		for (i = 0; i < nr; i++) {
@@ -592,7 +592,6 @@ int invalidate_inode_pages2_range(struct address_space *mapping,
 	int ret = 0;
 	int ret2 = 0;
 	int did_range_unmap = 0;
-	bool xa_has_values = false;
 
 	if (mapping_empty(mapping))
 		return 0;
@@ -600,6 +599,7 @@ int invalidate_inode_pages2_range(struct address_space *mapping,
 	folio_batch_init(&fbatch);
 	index = start;
 	while (find_get_entries(mapping, &index, end, &fbatch, indices)) {
+		bool xa_has_values = false;
 		int nr = folio_batch_count(&fbatch);
 
 		for (i = 0; i < nr; i++) {
-- 
2.43.5

On Wed, Oct 2, 2024 at 4:52 PM Shakeel Butt <shakeel.butt@linux.dev> wrote:
>
> Currently mapping_try_invalidate() and invalidate_inode_pages2_range()
> traverses the xarray in batches and then for each batch, maintains and
> set the flag named xa_has_values if the batch has a shadow entry to
> clear the entries at the end of the iteration. However they forgot to
> reset the flag at the end of the iteration which cause them to always
> try to clear the shadow entries in the subsequent iterations where
> there might not be any shadow entries. Fixing it.
>
> Fixes: 61c663e020d2 ("mm/truncate: batch-clear shadow entries")
> Signed-off-by: Shakeel Butt <shakeel.butt@linux.dev>

Acked-by: Yu Zhao <yuzhao@google.com>

On Wed,  2 Oct 2024 15:51:50 -0700 Shakeel Butt <shakeel.butt@linux.dev> wrote:

> Currently mapping_try_invalidate() and invalidate_inode_pages2_range()
> traverses the xarray in batches and then for each batch, maintains and
> set the flag named xa_has_values if the batch has a shadow entry to
> clear the entries at the end of the iteration. However they forgot to
> reset the flag at the end of the iteration which cause them to always
> try to clear the shadow entries in the subsequent iterations where
> there might not be any shadow entries. Fixing it.
> 

So this is an efficiency thing, no other effects expected?

> --- a/mm/truncate.c
> +++ b/mm/truncate.c
> @@ -463,10 +463,10 @@ unsigned long mapping_try_invalidate(struct address_space *mapping,
>  	unsigned long ret;
>  	unsigned long count = 0;
>  	int i;
> -	bool xa_has_values = false;
>  
>  	folio_batch_init(&fbatch);
>  	while (find_lock_entries(mapping, &index, end, &fbatch, indices)) {
> +		bool xa_has_values = false;
>  		int nr = folio_batch_count(&fbatch);
>  
>  		for (i = 0; i < nr; i++) {
> @@ -592,7 +592,6 @@ int invalidate_inode_pages2_range(struct address_space *mapping,
>  	int ret = 0;
>  	int ret2 = 0;
>  	int did_range_unmap = 0;
> -	bool xa_has_values = false;
>  
>  	if (mapping_empty(mapping))
>  		return 0;
> @@ -600,6 +599,7 @@ int invalidate_inode_pages2_range(struct address_space *mapping,
>  	folio_batch_init(&fbatch);
>  	index = start;
>  	while (find_get_entries(mapping, &index, end, &fbatch, indices)) {
> +		bool xa_has_values = false;
>  		int nr = folio_batch_count(&fbatch);
>  
>  		for (i = 0; i < nr; i++) {
> -- 
> 2.43.5

On Wed, Oct 02, 2024 at 03:55:55PM GMT, Andrew Morton wrote:
> On Wed,  2 Oct 2024 15:51:50 -0700 Shakeel Butt <shakeel.butt@linux.dev> wrote:
> 
> > Currently mapping_try_invalidate() and invalidate_inode_pages2_range()
> > traverses the xarray in batches and then for each batch, maintains and
> > set the flag named xa_has_values if the batch has a shadow entry to
> > clear the entries at the end of the iteration. However they forgot to
> > reset the flag at the end of the iteration which cause them to always
> > try to clear the shadow entries in the subsequent iterations where
> > there might not be any shadow entries. Fixing it.
> > 
> 
> So this is an efficiency thing, no other effects expected?
> 

Correct, just an efficiency thing.

On Wed, 2 Oct 2024 16:09:11 -0700 Shakeel Butt <shakeel.butt@linux.dev> wrote:

> On Wed, Oct 02, 2024 at 03:55:55PM GMT, Andrew Morton wrote:
> > On Wed,  2 Oct 2024 15:51:50 -0700 Shakeel Butt <shakeel.butt@linux.dev> wrote:
> > 
> > > Currently mapping_try_invalidate() and invalidate_inode_pages2_range()
> > > traverses the xarray in batches and then for each batch, maintains and
> > > set the flag named xa_has_values if the batch has a shadow entry to
> > > clear the entries at the end of the iteration. However they forgot to
> > > reset the flag at the end of the iteration which cause them to always
> > > try to clear the shadow entries in the subsequent iterations where
> > > there might not be any shadow entries. Fixing it.
> > > 
> > 
> > So this is an efficiency thing, no other effects expected?
> > 
> 
> Correct, just an efficiency thing.

Thanks.  I'm assuming the benfits are sufficiently small that a
backport is inappropriate.

On Thu, Oct 03, 2024 at 01:01:33PM GMT, Andrew Morton wrote:
> On Wed, 2 Oct 2024 16:09:11 -0700 Shakeel Butt <shakeel.butt@linux.dev> wrote:
> 
> > On Wed, Oct 02, 2024 at 03:55:55PM GMT, Andrew Morton wrote:
> > > On Wed,  2 Oct 2024 15:51:50 -0700 Shakeel Butt <shakeel.butt@linux.dev> wrote:
> > > 
> > > > Currently mapping_try_invalidate() and invalidate_inode_pages2_range()
> > > > traverses the xarray in batches and then for each batch, maintains and
> > > > set the flag named xa_has_values if the batch has a shadow entry to
> > > > clear the entries at the end of the iteration. However they forgot to
> > > > reset the flag at the end of the iteration which cause them to always
> > > > try to clear the shadow entries in the subsequent iterations where
> > > > there might not be any shadow entries. Fixing it.
> > > > 
> > > 
> > > So this is an efficiency thing, no other effects expected?
> > > 
> > 
> > Correct, just an efficiency thing.
> 
> Thanks.  I'm assuming the benfits are sufficiently small that a
> backport is inappropriate.

I agree.