Re: [PATCH V4 5/9] dax/fsdev: use __va(phys) for kaddr in direct_access

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] `__fsdev_dax_direct_access()` returns `-EFAULT` instead of truncating the returned size when a request crosses a multi-range physical boundary.
--

commit 25e7e6c7d5f413efd9eb57096b8df2c8803d235a
Author: John Groves <John@Groves.net>

dax/fsdev: use __va(phys) for kaddr in direct_access

This commit updates __fsdev_dax_direct_access() to use __va(phys) for the
kaddr return, fixing an issue where multi-range devices with physical gaps
resulted in an incorrect kernel virtual address. It also removes the now
unused virt_addr field from struct dev_dax.

> diff --git a/drivers/dax/fsdev.c b/drivers/dax/fsdev.c
> index 0fd5e1293d725..af9ef80c05c6d 100644
> --- a/drivers/dax/fsdev.c
> +++ b/drivers/dax/fsdev.c
> @@ -51,9 +51,7 @@ static long __fsdev_dax_direct_access(struct dax_device *dax_dev, pgoff_t pgoff,
>  	struct dev_dax *dev_dax = dax_get_private(dax_dev);
>  	size_t size = nr_pages << PAGE_SHIFT;
>  	size_t offset = pgoff << PAGE_SHIFT;
> -	void *virt_addr = dev_dax->virt_addr + offset;
>  	phys_addr_t phys;
> -	unsigned long local_pfn;
>  
>  	phys = dax_pgoff_to_phys(dev_dax, pgoff, size);
>  	if (phys == -1) {

[Severity: High]
This isn't a bug introduced by this patch, but it appears to be a pre-existing
issue. If a direct_access request spans across a physical gap into the next
DAX range, dax_pgoff_to_phys() returns -1 because it checks if the entire
requested size fits within a single contiguous physical range.

This causes __fsdev_dax_direct_access() to fail the entire operation with
-EFAULT. Should it instead truncate the returned size to the boundary of the
available contiguous length, allowing the caller to safely issue subsequent
requests for the remainder of the pages? Failing completely with -EFAULT
might spuriously break valid I/O accesses on multi-range DAX devices.

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/0100019ea3929225-a0f8e6f7-30ae-4f8e-ae6f-19129666c4c3-000000@email.amazonses.com?part=5