[PATCH] block: rename min_segment_size

[PATCH] block: rename min_segment_size

[Keith Busch]

From: Keith Busch <kbusch@kernel.org>

Despite its name, the block layer is fine with segments smaller that the
"min_segment_size" limit. The value is an optimization limit indicating
the largest aligned segment that can be used without considering segment
boundary limits, so give it a name that reflects that.

Signed-off-by: Keith Busch <kbusch@kernel.org>
---
 block/blk-merge.c      | 2 +-
 block/blk-settings.c   | 4 ++--
 block/blk.h            | 2 +-
 include/linux/blkdev.h | 2 +-
 4 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/block/blk-merge.c b/block/blk-merge.c
index 70d704615be52..863a4a6d3da62 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -307,7 +307,7 @@ int bio_split_rw_at(struct bio *bio, const struct queue_limits *lim,
 
 		if (nsegs < lim->max_segments &&
 		    bytes + bv.bv_len <= max_bytes &&
-		    bv.bv_offset + bv.bv_len <= lim->min_segment_size) {
+		    bv.bv_offset + bv.bv_len <= lim->max_aligned_segment) {
 			nsegs++;
 			bytes += bv.bv_len;
 		} else {
diff --git a/block/blk-settings.c b/block/blk-settings.c
index d6438e6c276dc..a61d676ff4c49 100644
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -442,12 +442,12 @@ int blk_validate_limits(struct queue_limits *lim)
 			return -EINVAL;
 	}
 
-	/* setup min segment size for building new segment in fast path */
+	/* setup max aligned segment size for building new segment in fast path */
 	if (lim->seg_boundary_mask > lim->max_segment_size - 1)
 		seg_size = lim->max_segment_size;
 	else
 		seg_size = lim->seg_boundary_mask + 1;
-	lim->min_segment_size = min_t(unsigned int, seg_size, PAGE_SIZE);
+	lim->max_aligned_segment = min_t(unsigned int, seg_size, PAGE_SIZE);
 
 	/*
 	 * We require drivers to at least do logical block aligned I/O, but
diff --git a/block/blk.h b/block/blk.h
index 46f566f9b1266..ed60d3dcdf93c 100644
--- a/block/blk.h
+++ b/block/blk.h
@@ -376,7 +376,7 @@ static inline bool bio_may_need_split(struct bio *bio,
 	if (bio->bi_vcnt != 1)
 		return true;
 	return bio->bi_io_vec->bv_len + bio->bi_io_vec->bv_offset >
-		lim->min_segment_size;
+		lim->max_aligned_segment;
 }
 
 /**
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index fe1797bbec420..b8d4333716e59 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -378,7 +378,7 @@ struct queue_limits {
 	unsigned int		max_sectors;
 	unsigned int		max_user_sectors;
 	unsigned int		max_segment_size;
-	unsigned int		min_segment_size;
+	unsigned int		max_aligned_segment;
 	unsigned int		physical_block_size;
 	unsigned int		logical_block_size;
 	unsigned int		alignment_offset;
-- 
2.47.3

Re: [PATCH] block: rename min_segment_size

[Ming Lei]

On Fri, Aug 22, 2025 at 10:10:38AM -0700, Keith Busch wrote:
> From: Keith Busch <kbusch@kernel.org>
> 
> Despite its name, the block layer is fine with segments smaller that the
> "min_segment_size" limit. The value is an optimization limit indicating
> the largest aligned segment that can be used without considering segment
> boundary limits, so give it a name that reflects that.

Yeah, it is right.

Reviewed-by: Ming Lei <ming.lei@redhat.com>


Thanks,
Ming

Re: [PATCH] block: rename min_segment_size

[Christoph Hellwig]

On Fri, Aug 22, 2025 at 10:10:38AM -0700, Keith Busch wrote:
> From: Keith Busch <kbusch@kernel.org>
> 
> Despite its name, the block layer is fine with segments smaller that the
> "min_segment_size" limit. The value is an optimization limit indicating
> the largest aligned segment that can be used without considering segment
> boundary limits, so give it a name that reflects that.

But max_aligned_segment also feels wrong for that.  It's not really
the maximum alignmnet, it is the fast path alignment.  Maybe something
like fast_segment_granularity or nosplit_segment_granularity?

Re: [PATCH] block: rename min_segment_size

[Martin K. Petersen]

Christoph,

> But max_aligned_segment also feels wrong for that. It's not really the
> maximum alignmnet, it is the fast path alignment. Maybe something like
> fast_segment_granularity or nosplit_segment_granularity?

Maybe just segment_granularity to match the other granularities we have?

-- 
Martin K. Petersen

Re: [PATCH] block: rename min_segment_size

[Keith Busch]

On Mon, Aug 25, 2025 at 03:18:43PM -0400, Martin K. Petersen wrote:
> 
> Christoph,
> 
> > But max_aligned_segment also feels wrong for that. It's not really the
> > maximum alignmnet, it is the fast path alignment. Maybe something like
> > fast_segment_granularity or nosplit_segment_granularity?
> 
> Maybe just segment_granularity to match the other granularities we have?

I'm not sure I like granularity for this limit. That sounds like it
defines segments to be sized to some multiple of that value, but it's
perfectly fine to use smaller segments.

The limit defines the largest aligned segment the block layer can know
for certain won't need to be split. This allows a short cut while
counting segments without checking against any other limits.

  max_aligned_unsplittable_segment_size

That's pretty verbose, but this limit is kind of unique in its purpose.

Re: [PATCH] block: rename min_segment_size

[Christoph Hellwig]

On Mon, Aug 25, 2025 at 01:50:09PM -0600, Keith Busch wrote:
> On Mon, Aug 25, 2025 at 03:18:43PM -0400, Martin K. Petersen wrote:
> > 
> > Christoph,
> > 
> > > But max_aligned_segment also feels wrong for that. It's not really the
> > > maximum alignmnet, it is the fast path alignment. Maybe something like
> > > fast_segment_granularity or nosplit_segment_granularity?
> > 
> > Maybe just segment_granularity to match the other granularities we have?
> 
> I'm not sure I like granularity for this limit. That sounds like it
> defines segments to be sized to some multiple of that value, but it's
> perfectly fine to use smaller segments.

I guess I stuck too much to nvme terminology where those boundaries
are called granularity.  We also do that for the discard granularity.
Maye it should be boundary?