[PATCH v2 2/6] drivers: block: Introduce blk_flush()/blk_dflush()

[Simon Glass]

uHi Denis,

On 2026-05-29T03:44:35, None <dmukhin at ford.com> wrote:
> drivers: block: Introduce blk_flush()/blk_dflush()
>
> Add generic flush operations for committing dirty data to the storage
> device.
>
> This provides a common block-layer interface for flushing pending
> writes, allowing callers to ensure that data buffered by the block
> device or its backing implementation is written out to the underlying
> storage.
>
> Signed-off-by: Denis Mukhin <dmukhin at ford.com>
>
> disk/disk-uclass.c         |  6 ++++++
>  drivers/block/blk-uclass.c | 18 ++++++++++++++++++
>  include/blk.h              | 28 ++++++++++++++++++++++++++++
>  include/part.h             |  8 ++++++++
>  4 files changed, 60 insertions(+)

> diff --git a/drivers/block/blk-uclass.c b/drivers/block/blk-uclass.c
> @@ -514,6 +514,19 @@ long blk_erase(struct udevice *dev, lbaint_t start, lbaint_t blkcnt)
> +long blk_flush(struct udevice *dev)
> +{
> +     struct blk_desc *desc = dev_get_uclass_plat(dev);
> +     const struct blk_ops *ops = blk_get_ops(dev);
> +
> +     if (!ops->flush)
> +             return -ENOSYS;
> +
> +     blkcache_invalidate(desc->uclass_id, desc->devnum);
> +
> +     return ops->flush(dev);
> +}

Why invalidate the read cache here? Flush pushes dirty data down to
the media; it doesn't change what a subsequent read returns, so the
cached contents are still valid. Erase/write mutate device state, but
flush does not. Please drop the blkcache_invalidate() call (and the
matching one in the non-DM inline below).

> diff --git a/include/blk.h b/include/blk.h
> @@ -99,6 +99,7 @@ struct blk_desc {
>       unsigned long   (*block_erase)(struct blk_desc *block_dev,
>                                      lbaint_t start,
>                                      lbaint_t blkcnt);
> +     unsigned long   (*block_flush)(struct blk_desc *block_dev);

This field lives in the !CONFIG_BLK (legacy) branch, but no legacy
driver in the series wires it up. The only consumer (NVMe) is DM-only.
Are you going to need the legacy plumbing?

> diff --git a/include/blk.h b/include/blk.h
> @@ -275,6 +276,14 @@ struct blk_ops {
> +     /**
> +      * flush() -  commit all dirty data to storage
> +      *
> +      * @dev:        Device to flush
> +      * @return 0 if OK, -ve on error
> +      */
> +     unsigned long (*flush)(struct udevice *dev);
>  };

Two small things: stray double space after the dash (flush() -
commit), and please place this next to .erase so the
read/write/erase/flush group is together.

> diff --git a/include/part.h b/include/part.h
> @@ -454,6 +454,14 @@ ulong disk_blk_write(struct udevice *dev, lbaint_t start, lbaint_t blkcnt,
> +/**
> + * disk_blk_flush() - commit data to a disk
> + *
> + * @dev:     Device to flush
> + * Return:   0 success, or -ve error number (see the IS_ERR_VALUE() macro
> + */
> +ulong disk_blk_flush(struct udevice *dev);

Missing closing parenthesis after IS_ERR_VALUE() macro.

> diff --git a/disk/disk-uclass.c b/disk/disk-uclass.c
> @@ -122,6 +122,11 @@ unsigned long disk_blk_erase(struct udevice *dev, lbaint_t start,
> +unsigned long disk_blk_flush(struct udevice *dev)
> +{
> +     return blk_flush(dev_get_parent(dev));
> +}

Please note in the commit message that flushing a partition flushes
the entire underlying device (the API has no start/blkcnt), so callers
don't expect partition-scoped semantics.

Regards,
Simon