netdev - Re: [PATCH 2/3] ptp: support multiple timestamp event readers

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Wed, 6 Sep 2023 20:13:49 +0200
From: Simon Horman <horms@...nel.org>
To: Xabier Marquiegui <reibax@...il.com>
Cc: richardcochran@...il.com, chrony-dev@...ony.tuxfamily.org,
	mlichvar@...hat.com, netdev@...r.kernel.org,
	ntp-lists@...tcorallo.com
Subject: Re: [PATCH 2/3] ptp: support multiple timestamp event readers

On Wed, Sep 06, 2023 at 12:47:53PM +0200, Xabier Marquiegui wrote:
> Use linked lists to create one event queue per open file. This enables
> simultaneous readers for timestamp event queues.
> 
> Signed-off-by: Xabier Marquiegui <reibax@...il.com>

Hi Xabier,

some minor feedback from my side

...

> diff --git a/drivers/ptp/ptp_chardev.c b/drivers/ptp/ptp_chardev.c
> index 1ea11f864abb..c65dc6fefaa6 100644
> --- a/drivers/ptp/ptp_chardev.c
> +++ b/drivers/ptp/ptp_chardev.c
> @@ -103,9 +103,39 @@ int ptp_set_pinfunc(struct ptp_clock *ptp, unsigned int pin,
>  
>  int ptp_open(struct posix_clock *pc, fmode_t fmode)
>  {
> +	struct ptp_clock *ptp = container_of(pc, struct ptp_clock, clock);
> +	struct timestamp_event_queue *queue;
> +
> +	queue = kzalloc(sizeof(struct timestamp_event_queue), GFP_KERNEL);
> +	if (queue == NULL)

nit: this could be: if (!queue)

> +		return -EINVAL;
> +	queue->reader_pid = task_pid_nr(current);
> +	list_add_tail(&queue->qlist, &ptp->tsevqs);
> +
>  	return 0;
>  }

...

> @@ -436,14 +466,25 @@ __poll_t ptp_poll(struct posix_clock *pc, struct file *fp, poll_table *wait)
>  {
>  	struct ptp_clock *ptp = container_of(pc, struct ptp_clock, clock);
>  	struct timestamp_event_queue *queue;
> +	struct list_head *pos, *n;
> +	bool found = false;
> +	pid_t reader_pid = task_pid_nr(current);
>  
>  	poll_wait(fp, &ptp->tsev_wq, wait);
>  
>  	/*
> -	 * Extract only the first element in the queue list
> -	 * TODO: Identify the relevant queue
> +	 * Extract only the desired element in the queue list
>  	 */
> -	queue = list_entry(&ptp->tsevqs, struct timestamp_event_queue, qlist);
> +	list_for_each_safe(pos, n, &ptp->tsevqs) {
> +		queue = list_entry(pos, struct timestamp_event_queue, qlist);
> +		if (queue->reader_pid == reader_pid) {
> +			found = true;
> +			break;
> +		}
> +	}
> +
> +	if (!found)
> +		return -EINVAL;

Sparse complains that the return type for this function is __poll_t,
but here an int is returned. Perhaps returning EPOLLERR is appropriate here?

>  
>  	return queue_cnt(queue) ? EPOLLIN : 0;
>  }

...