[PATCH 1/2] loop: remove a pointless blk_queue_max_hw_sectors call

[PATCH 1/2] loop: remove a pointless blk_queue_max_hw_sectors call

[Christoph Hellwig]

BLK_DEF_MAX_SECTORS is (as the name implies) already the default.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/loop.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 5bc2b4fcfa772d..371a318e691d02 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -2038,8 +2038,6 @@ static int loop_add(int i)
 	}
 	lo->lo_queue = lo->lo_disk->queue;
 
-	blk_queue_max_hw_sectors(lo->lo_queue, BLK_DEF_MAX_SECTORS);
-
 	/*
 	 * By default, we do buffer IO, so it doesn't make sense to enable
 	 * merge because the I/O submitted to backing file is handled page by
-- 
2.39.2

[PATCH 2/2] loop: don't update discard limits from loop_set_status

[Christoph Hellwig]

loop_set_status doesn't change anything relevant to the discard and
write_zeroes setting, so don't bother calling loop_config_discard.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/loop.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 371a318e691d02..794bc92da165c3 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -1303,8 +1303,6 @@ loop_set_status(struct loop_device *lo, const struct loop_info64 *info)
 		loop_set_size(lo, new_size);
 	}
 
-	loop_config_discard(lo);
-
 	/* update dio if lo_offset or transfer is changed */
 	__loop_update_dio(lo, lo->use_dio);
 
-- 
2.39.2

Re: [PATCH 1/2] loop: remove a pointless blk_queue_max_hw_sectors call

[Yu Kuai]

Hi,

在 2023/12/26 17:14, Christoph Hellwig 写道:
> BLK_DEF_MAX_SECTORS is (as the name implies) already the default.

Looks like this not true. From blk_set_default_limits():

lim->max_sectors = lim->max_hw_sectors = BLK_SAFE_MAX_SECTORS;

So, the default is not BLK_DEF_MAX_SECTORS. Or am I missing something?

Thanks,
Kuai
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>   drivers/block/loop.c | 2 --
>   1 file changed, 2 deletions(-)
> 
> diff --git a/drivers/block/loop.c b/drivers/block/loop.c
> index 5bc2b4fcfa772d..371a318e691d02 100644
> --- a/drivers/block/loop.c
> +++ b/drivers/block/loop.c
> @@ -2038,8 +2038,6 @@ static int loop_add(int i)
>   	}
>   	lo->lo_queue = lo->lo_disk->queue;
>   
> -	blk_queue_max_hw_sectors(lo->lo_queue, BLK_DEF_MAX_SECTORS);
> -
>   	/*
>   	 * By default, we do buffer IO, so it doesn't make sense to enable
>   	 * merge because the I/O submitted to backing file is handled page by
> 

Re: [PATCH 1/2] loop: remove a pointless blk_queue_max_hw_sectors call

[Christoph Hellwig]

On Tue, Dec 26, 2023 at 05:38:31PM +0800, Yu Kuai wrote:
> Hi,
>
> 在 2023/12/26 17:14, Christoph Hellwig 写道:
>> BLK_DEF_MAX_SECTORS is (as the name implies) already the default.
>
> Looks like this not true. From blk_set_default_limits():
>
> lim->max_sectors = lim->max_hw_sectors = BLK_SAFE_MAX_SECTORS;
>
> So, the default is not BLK_DEF_MAX_SECTORS. Or am I missing something?

No, you're not.  The naming and our code is just a desaster..