Re: [PATCH v2 9/9] ata: Annotate the code that uses the host lock

[Niklas Cassel <cassel@kernel.org>]

On Thu, May 21, 2026 at 10:33:37AM -0700, Bart Van Assche wrote:
> Annotate all functions and also their direct and indirect callers with
> __must_hold() that have a lockdep_assert_held() statement in their
> function body for the host lock. As one can see in the comments added by
> this patch, locking is missing from the following two functions:
>  * Some of the ata_port_freeze() callers.
>  * nv_do_interrupt().
> 
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
> ---

(snip)

> diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c
> index 8b8081f4dc63..52bcd8c78774 100644
> --- a/drivers/ata/libata-core.c
> +++ b/drivers/ata/libata-core.c
> @@ -5027,6 +5027,9 @@ void ata_qc_complete(struct ata_port *ap, struct ata_queued_cmd *qc)
>  	struct ata_device *dev = qc->dev;
>  	struct ata_eh_info *ehi = &dev->link->eh_info;
>  
> +	/* Tell the compiler that qc->dev->link->ap == ap. */
> +	__assume_ctx_lock(qc->dev->link->ap->lock);

Is there any way we could avoid this __assume_ctx_lock() ?

Which is the function we call that needs this?

Could we avoid it if we add a ap / qc / link argument to that function?



> +
>  	/* Trigger the LED (if available) */
>  	ledtrig_disk_activity(!!(qc->tf.flags & ATA_TFLAG_WRITE));
>  
> @@ -5162,6 +5165,10 @@ void ata_qc_issue(struct ata_port *ap, struct ata_queued_cmd *qc)
>  	struct ata_link *link = qc->dev->link;
>  	u8 prot = qc->tf.protocol;
>  
> +	/* Tell the compiler that qc->ap == qc->dev->link->ap->lock == ap. */
> +	__assume_ctx_lock(qc->ap->lock);
> +	__assume_ctx_lock(qc->dev->link->ap->lock);

Is there any way we could avoid these two __assume_ctx_lock() ?

Which is the function we call that needs this?

Could we avoid it if we add a ap / qc / link argument to that function?



Kind regards,
Niklas