[PATCH v4 05/11] block: Add core atomic write support

[Dave Chinner]

On Mon, Feb 19, 2024 at 01:01:03PM +0000, John Garry wrote:
> Add atomic write support as follows:
> diff --git a/block/blk-merge.c b/block/blk-merge.c
> index 74e9e775f13d..12a75a252ca2 100644
> --- a/block/blk-merge.c
> +++ b/block/blk-merge.c
> @@ -18,6 +18,42 @@
>  #include "blk-rq-qos.h"
>  #include "blk-throttle.h"
>  
> +static bool rq_straddles_atomic_write_boundary(struct request *rq,
> +					unsigned int front,
> +					unsigned int back)
> +{
> +	unsigned int boundary = queue_atomic_write_boundary_bytes(rq->q);
> +	unsigned int mask, imask;
> +	loff_t start, end;
> +
> +	if (!boundary)
> +		return false;
> +
> +	start = rq->__sector << SECTOR_SHIFT;
> +	end = start + rq->__data_len;
> +
> +	start -= front;
> +	end += back;
> +
> +	/* We're longer than the boundary, so must be crossing it */
> +	if (end - start > boundary)
> +		return true;
> +
> +	mask = boundary - 1;
> +
> +	/* start/end are boundary-aligned, so cannot be crossing */
> +	if (!(start & mask) || !(end & mask))
> +		return false;
> +
> +	imask = ~mask;
> +
> +	/* Top bits are different, so crossed a boundary */
> +	if ((start & imask) != (end & imask))
> +		return true;
> +
> +	return false;
> +}

I have no way of verifying this function is doing what it is
supposed to because it's function is undocumented. I have no idea
what the front/back variables are supposed to represent, and so no
clue if they are being applied properly.

That said, it's also applying unsigned 32 bit mask variables to
signed 64 bit quantities and trying to do things like "high bit changed"
checks on the 64 bit variable. This just smells like a future
source of "large offsets don't work like we expected!" bugs.

> diff --git a/block/blk-settings.c b/block/blk-settings.c
> index 06ea91e51b8b..176f26374abc 100644
> --- a/block/blk-settings.c
> +++ b/block/blk-settings.c
> @@ -59,6 +59,13 @@ void blk_set_default_limits(struct queue_limits *lim)
>  	lim->zoned = false;
>  	lim->zone_write_granularity = 0;
>  	lim->dma_alignment = 511;
> +	lim->atomic_write_hw_max_sectors = 0;
> +	lim->atomic_write_max_sectors = 0;
> +	lim->atomic_write_hw_boundary_sectors = 0;
> +	lim->atomic_write_hw_unit_min_sectors = 0;
> +	lim->atomic_write_unit_min_sectors = 0;
> +	lim->atomic_write_hw_unit_max_sectors = 0;
> +	lim->atomic_write_unit_max_sectors = 0;
>  }

Seems to me this function would do better to just

	memset(lim, 0, sizeof(*lim));

and then set all the non-zero fields.

>  
>  /**
> @@ -101,6 +108,44 @@ void blk_queue_bounce_limit(struct request_queue *q, enum blk_bounce bounce)
>  }
>  EXPORT_SYMBOL(blk_queue_bounce_limit);
>  
> +
> +/*
> + * Returns max guaranteed sectors which we can fit in a bio. For convenience of
> + * users, rounddown_pow_of_two() the return value.
> + *
> + * We always assume that we can fit in at least PAGE_SIZE in a segment, apart
> + * from first and last segments.
> + */
> +static unsigned int blk_queue_max_guaranteed_bio_sectors(
> +					struct queue_limits *limits,
> +					struct request_queue *q)
> +{
> +	unsigned int max_segments = min(BIO_MAX_VECS, limits->max_segments);
> +	unsigned int length;
> +
> +	length = min(max_segments, 2) * queue_logical_block_size(q);
> +	if (max_segments > 2)
> +		length += (max_segments - 2) * PAGE_SIZE;
> +
> +	return rounddown_pow_of_two(length >> SECTOR_SHIFT);
> +}
> +
> +static void blk_atomic_writes_update_limits(struct request_queue *q)
> +{
> +	struct queue_limits *limits = &q->limits;
> +	unsigned int max_hw_sectors =
> +		rounddown_pow_of_two(limits->max_hw_sectors);
> +	unsigned int unit_limit = min(max_hw_sectors,
> +		blk_queue_max_guaranteed_bio_sectors(limits, q));
> +
> +	limits->atomic_write_max_sectors =
> +		min(limits->atomic_write_hw_max_sectors, max_hw_sectors);
> +	limits->atomic_write_unit_min_sectors =
> +		min(limits->atomic_write_hw_unit_min_sectors, unit_limit);
> +	limits->atomic_write_unit_max_sectors =
> +		min(limits->atomic_write_hw_unit_max_sectors, unit_limit);
> +}
> +
>  /**
>   * blk_queue_max_hw_sectors - set max sectors for a request for this queue
>   * @q:  the request queue for the device
> @@ -145,6 +190,8 @@ void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_secto
>  				 limits->logical_block_size >> SECTOR_SHIFT);
>  	limits->max_sectors = max_sectors;
>  
> +	blk_atomic_writes_update_limits(q);
> +
>  	if (!q->disk)
>  		return;
>  	q->disk->bdi->io_pages = max_sectors >> (PAGE_SHIFT - 9);
> @@ -182,6 +229,62 @@ void blk_queue_max_discard_sectors(struct request_queue *q,
>  }
>  EXPORT_SYMBOL(blk_queue_max_discard_sectors);
>  
> +/**
> + * blk_queue_atomic_write_max_bytes - set max bytes supported by
> + * the device for atomic write operations.
> + * @q:  the request queue for the device
> + * @bytes: maximum bytes supported
> + */
> +void blk_queue_atomic_write_max_bytes(struct request_queue *q,
> +				      unsigned int bytes)
> +{
> +	q->limits.atomic_write_hw_max_sectors = bytes >> SECTOR_SHIFT;
> +	blk_atomic_writes_update_limits(q);
> +}
> +EXPORT_SYMBOL(blk_queue_atomic_write_max_bytes);

Ok, so this can silently set a limit that is different to what the
caller asked to have set?

How is the caller supposed to find this out if the smaller limit
that was set is not compatible with their configuration?

i.e. shouldn't this return an error if the requested size cannot
be set exactly as specified?

Same concern about the other limits that can be trimmed to be
smaller than requested.

-Dave.
-- 
Dave Chinner
david at fromorbit.com