Re: [PATCH v3 04/21] md: add helper to split bios at reshape offset

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Lockless read of the 64-bit `mddev->reshape_position` and missing memory barriers after `MD_RECOVERY_RESHAPE` cause torn reads and inconsistent state observation.
--

--- Patch [4]: [PATCH v3 04/21] md: add helper to split bios at reshape offset ---
commit df6c72a7c4a6271ac6b7e8dc1420ce49f5ef19d3
Author: Yu Kuai <[email protected]>

md: add helper to split bios at reshape offset

This commit adds mddev_bio_split_at_reshape_offset() so personalities can
share reshape-offset bio splitting instead of open-coding the same helper
in multiple places.

> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index 0b59c676f7c03..c18a7f3c27e65 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -9361,6 +9361,45 @@ void md_submit_discard_bio(struct mddev *mddev, struct md_rdev *rdev,
>  }
>  EXPORT_SYMBOL_GPL(md_submit_discard_bio);
>  
> +struct bio *mddev_bio_split_at_reshape_offset(struct mddev *mddev,
> +					      struct bio *bio,
> +					      unsigned int *max_sectors,
> +					      struct bio_set *bs)
> +{
> +	sector_t boundary;
> +	sector_t start;
> +	sector_t end;
> +	unsigned int split_sectors;
> +
> +	split_sectors = bio_sectors(bio);
> +	if (max_sectors && *max_sectors && *max_sectors < split_sectors)
> +		split_sectors = *max_sectors;
> +
> +	if (!test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery))
> +		goto split;
> +
> +	boundary = mddev->reshape_position;

[Severity: High]
Can this lockless read of mddev->reshape_position lead to torn reads?

Since reshape_position is a 64-bit value, reading it concurrently with
background reshape updates from sync_thread might result in a torn read on
32-bit architectures. A torn read could produce an arbitrary split boundary,
potentially causing misaligned bio splits and IO errors.

Additionally, are memory barriers needed after testing MD_RECOVERY_RESHAPE
to ensure a consistent value is observed on weakly ordered architectures?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=4