All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: linux-mtd@lists.infradead.org,
	David Woodhouse <dwmw2@infradead.org>,
	Brian Norris <computersforpeace@gmail.com>,
	Artem Bityutskiy <dedekind1@gmail.com>,
	Richard Weinberger <richard@nod.at>,
	tglx@linutronix.de
Subject: Re: [PATCH] mtd: nand: do FIFO processing in nand_get_device()
Date: Mon, 30 Nov 2015 17:15:49 +0100	[thread overview]
Message-ID: <20151130161549.GL17308@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <20151125173543.GA17151@linutronix.de>

On Wed, Nov 25, 2015 at 06:35:43PM +0100, Sebastian Andrzej Siewior wrote:
> +++ b/drivers/mtd/nand/nand_base.c
> @@ -137,8 +137,25 @@ static void nand_release_device(struct mtd_info *mtd)
>  	/* Release the controller and the chip */
>  	spin_lock(&chip->controller->lock);
>  	chip->controller->active = NULL;
> -	chip->state = FL_READY;
> -	wake_up(&chip->controller->wq);
> +
> +	if (waitqueue_active(&chip->controller->wq)) {
> +		wait_queue_head_t *q;
> +		wait_queue_t *waiter;
> +		unsigned long flags;
> +
> +		q = &chip->controller->wq;
> +		chip->state = FL_HANDOVER;
> +
> +		spin_lock_irqsave(&q->lock, flags);

This lock is actually not required, as your add/remove_wait_queue calls
are also under chip->controller->lock.

> +		waiter = list_first_entry(&q->task_list, wait_queue_t,
> +					  task_list);
> +		spin_unlock_irqrestore(&q->lock, flags);

And the one read instruction under a spinlock is a tad pointless anyway.

> +
> +		chip->controller->handover_waiter = waiter;

You could consider using ->active for this; as it stands you never use
both at the same time. Its a tad icky, but it avoids adding that new
field.

> +		wake_up(q);
> +	} else {
> +		chip->state = FL_READY;
> +	}
>  	spin_unlock(&chip->controller->lock);
>  }
>  
> @@ -843,10 +860,18 @@ nand_get_device(struct mtd_info *mtd, int new_state)
>  	if (!chip->controller->active)
>  		chip->controller->active = chip;
>  
> -	if (chip->controller->active == chip && chip->state == FL_READY) {
> -		chip->state = new_state;
> -		spin_unlock(lock);
> -		return 0;
> +	if (chip->controller->active == chip) {
> +		if (chip->state == FL_READY) {
> +			chip->state = new_state;
> +			spin_unlock(lock);
> +			return 0;
> +		}
> +		if (chip->state == FL_HANDOVER &&
> +		    chip->controller->handover_waiter == &wait) {
> +			chip->state = new_state;
> +			spin_unlock(lock);
> +			return 0;
> +		}
>  	}

That means this would become:

	if (chip->state == FL_HANDOVER && chip->controller->active == &wait) {
		chip->controller->active = chip;
		chip->state = new_state;
	}

	... existing code ...

  reply	other threads:[~2015-11-30 16:15 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-11-23 18:09 [RFC] avoid a live lock in wear_leveling_worker() Sebastian Andrzej Siewior
2015-11-23 18:09 ` [RFC PATCH 1/2] mtd: nand: schedule() after releasing the device Sebastian Andrzej Siewior
2015-11-23 18:18   ` Peter Zijlstra
2015-11-25 17:35     ` [PATCH] mtd: nand: do FIFO processing in nand_get_device() Sebastian Andrzej Siewior
2015-11-30 16:15       ` Peter Zijlstra [this message]
2015-12-06 14:17         ` Sebastian Andrzej Siewior
2015-12-06 14:23           ` [PATCH v2] " Sebastian Andrzej Siewior
2015-12-02 18:52       ` [PATCH] " Brian Norris
2015-12-02 20:41         ` Sebastian Andrzej Siewior
2015-11-23 18:09 ` [RFC PATCH 2/2] mtd: ubi: wl: avoid erasing a PEB which is empty Sebastian Andrzej Siewior
2015-11-23 21:30   ` Richard Weinberger
2015-11-23 21:50     ` Richard Weinberger
2015-11-24  8:26     ` Sebastian Andrzej Siewior
2015-11-24  8:39       ` Richard Weinberger
2015-11-24  8:42         ` Sebastian Andrzej Siewior
2015-11-24  9:02           ` Richard Weinberger
2015-11-24  9:07             ` Sebastian Andrzej Siewior
2015-11-24  9:16               ` Richard Weinberger
2015-11-24 12:58   ` Artem Bityutskiy
2015-11-24 13:33     ` Sebastian Andrzej Siewior
2015-11-24 13:40       ` Artem Bityutskiy
2015-11-24 13:57       ` Artem Bityutskiy
2015-11-26 14:56     ` Sebastian Andrzej Siewior

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20151130161549.GL17308@twins.programming.kicks-ass.net \
    --to=peterz@infradead.org \
    --cc=bigeasy@linutronix.de \
    --cc=computersforpeace@gmail.com \
    --cc=dedekind1@gmail.com \
    --cc=dwmw2@infradead.org \
    --cc=linux-mtd@lists.infradead.org \
    --cc=richard@nod.at \
    --cc=tglx@linutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.