Re: [PATCH v5 21/27] scsi-multipath: add delayed disk removal support

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] The delayed_work struct for removing the multipath head is never initialized for SCSI.
--

--- Patch [21]: [PATCH v5 21/27] scsi-multipath: add delayed disk removal support ---
commit 75d6c7bf920e841ac7f62e67cc466430605858b8
Author: John Garry <[email protected]>

scsi-multipath: add delayed disk removal support

Add support in core code for delayed disk removal support. In this, the
callback calls into the scsi_driver to do the necessary removal work.

> diff --git a/drivers/scsi/scsi_multipath.c b/drivers/scsi/scsi_multipath.c
> index eae5864ce5826..ed9c83189942a 100644
> --- a/drivers/scsi/scsi_multipath.c
> +++ b/drivers/scsi/scsi_multipath.c
[ ... ]
> @@ -360,7 +360,34 @@ static bool scsi_mpath_available_path(struct mpath_device *mpath_device)
[ ... ]
>  static struct mpath_head_template smpdt = {
> +	.remove_head = scsi_mpath_remove_head_work,

[Severity: High]
Does enabling this feature lead to an uninitialized work struct regression?

When delayed removal is triggered, it calls mod_delayed_work() on
mpath_head->remove_work. However, the initialization path for SCSI multipath
heads appears to omit INIT_DELAYED_WORK() for this work struct.

In drivers/scsi/scsi_multipath.c:scsi_mpath_alloc_head():

	if (mpath_head_init(&scsi_mpath_head->mpath_head))
		goto out_free;

mpath_head_init() zeroes the structure but does not initialize remove_work.

Could this lead to a panic in the workqueue core when operating on an
uninitialized work struct?

>  	.is_disabled = scsi_mpath_is_disabled,
>  	.is_optimized = scsi_mpath_is_optimized,
>  	.available_path = scsi_mpath_available_path,

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=21