Re: [PATCH v4] io_uring: annotate remote tasks for kcoverage

[Jens Axboe <axboe@kernel.dk>]

On 6/24/26 3:01 AM, Robert Femmer wrote:
> diff --git a/io_uring/io-wq.c b/io_uring/io-wq.c
> index 8cc7b47d3089..173299dfc9c2 100644
> --- a/io_uring/io-wq.c
> +++ b/io_uring/io-wq.c
> @@ -19,6 +19,7 @@
>  #include <linux/mmu_context.h>
>  #include <linux/sched/sysctl.h>
>  #include <uapi/linux/io_uring.h>
> +#include <linux/kcov.h>
>  
>  #include "io-wq.h"
>  #include "slist.h"
> @@ -639,6 +640,7 @@ static void io_worker_handle_work(struct io_wq_acct *acct,
>  		/* handle a whole dependent link */
>  		do {
>  			struct io_wq_work *next_hashed, *linked;
> +			struct io_kiocb *req;
>  			unsigned int work_flags = atomic_read(&work->flags);
>  			unsigned int hash = __io_wq_is_hashed(work_flags)
>  				? __io_get_work_hash(work_flags)

We try to, roughly, keep this in reverse xmas tree order. I've fixed it
up.

> diff --git a/io_uring/sqpoll.c b/io_uring/sqpoll.c
> index 46c12afec73e..aafb640d3b2f 100644
> --- a/io_uring/sqpoll.c
> +++ b/io_uring/sqpoll.c
> @@ -13,6 +13,7 @@
>  #include <linux/cpuset.h>
>  #include <linux/sched/cputime.h>
>  #include <linux/io_uring.h>
> +#include <linux/kcov.h>
>  
>  #include <uapi/linux/io_uring.h>
>  
> @@ -342,10 +343,12 @@ static int io_sq_thread(void *data)
>  
>  		cap_entries = !list_is_singular(&sqd->ctx_list);
>  		list_for_each_entry(ctx, &sqd->ctx_list, sqd_list) {
> +			kcov_remote_start_common(ctx->kcov_handle);
>  			int ret = __io_sq_thread(ctx, sqd, cap_entries, &ist);

Never code before variable declarations in the scope. I've also fixed
that one up.

Also not sure what branch this was against, because it doesn't apply to
any current one. Guessing something old? I had applied 2 hunks, fwiw.
Please check the result!

-- 
Jens Axboe