diff mbox series

[v3,5/6] gfs2: Replace kmap() by kmap_local_page() in gfs2_read_super

Message ID 7ad72f9d1a97d673483c206d8ac9a88f2e32d3b9.1688073459.git.drv@mailo.com
State New
Headers
Series gfs2: kmap{_atomic} conversion to kmap_local_{page/folio} |

Commit Message

Deepak R Varma June 29, 2023, 9:51 p.m. UTC 
  The use of kmap() is being deprecated in favor of kmap_local_page().

There are two main problems with kmap(): (1) It comes with an overhead as
the mapping space is restricted and protected by a global lock for
synchronization and (2) it also requires global TLB invalidation when the
kmap’s pool wraps and it might block when the mapping space is fully
utilized until a slot becomes available.

With kmap_local_page() the mappings are per thread, CPU local, can take
page faults, and can be called from any context (including interrupts).
It is faster than kmap() in kernels with HIGHMEM enabled. Furthermore,
the tasks can be preempted and, when they are scheduled to run again, the
kernel virtual addresses are restored and still valid.

Therefore, replace kmap() with kmap_local_page() in gfs2_read_super().

Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Deepak R Varma <drv@mailo.com>
---
Changes in v3:
   - Patch included in patch set

Changes in v2:
   - None

 fs/gfs2/ops_fstype.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Comments

Fabio M. De Francesco July 1, 2023, 1:45 p.m. UTC | #1 
On giovedì 29 giugno 2023 23:51:53 CEST Deepak R Varma wrote:
> The use of kmap() is being deprecated in favor of kmap_local_page().
> 
> There are two main problems with kmap(): (1) It comes with an overhead as
> the mapping space is restricted and protected by a global lock for
> synchronization and (2) it also requires global TLB invalidation when the
> kmap’s pool wraps and it might block when the mapping space is fully
> utilized until a slot becomes available.
> 
> With kmap_local_page() the mappings are per thread, CPU local, can take
> page faults, and can be called from any context (including interrupts).
> It is faster than kmap() in kernels with HIGHMEM enabled. Furthermore,
> the tasks can be preempted and, when they are scheduled to run again, the
> kernel virtual addresses are restored and still valid.
> 
> Therefore, replace kmap() with kmap_local_page() in gfs2_read_super().
> 
> Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> Signed-off-by: Deepak R Varma <drv@mailo.com>
> ---
> Changes in v3:
>    - Patch included in patch set
> 
> Changes in v2:
>    - None
> 
>  fs/gfs2/ops_fstype.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
> index 8a27957dbfee..80fe61662412 100644
> --- a/fs/gfs2/ops_fstype.c
> +++ b/fs/gfs2/ops_fstype.c
> @@ -264,9 +264,9 @@ static int gfs2_read_super(struct gfs2_sbd *sdp, 
sector_t
> sector, int silent) __free_page(page);
>  		return -EIO;
>  	}
> -	p = kmap(page);
> +	p = kmap_local_page(page);
>
Deepak,

Could this page ever come from ZONE_HIGHMEM? Can you please check whether we 
really need to kmap*() it or not? 

Fabio
>
>  	gfs2_sb_in(sdp, p);
> -	kunmap(page);
> +	kunmap_local(p);
>  	__free_page(page);
>  	return gfs2_check_sb(sdp, silent);
>  }
> --
> 2.34.1
diff mbox series

Patch

diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index 8a27957dbfee..80fe61662412 100644
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -264,9 +264,9 @@  static int gfs2_read_super(struct gfs2_sbd *sdp, sector_t sector, int silent)
 		__free_page(page);
 		return -EIO;
 	}
-	p = kmap(page);
+	p = kmap_local_page(page);
 	gfs2_sb_in(sdp, p);
-	kunmap(page);
+	kunmap_local(p);
 	__free_page(page);
 	return gfs2_check_sb(sdp, silent);
 }