devicetree.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Frank Rowand <frowand.list@gmail.com>
To: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: devicetree@vger.kernel.org, Rob Herring <robh+dt@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: Re: [PATCH v2] of: allocate / free phandle cache outside of the devtree_lock
Date: Fri, 29 Nov 2019 20:46:31 -0600	[thread overview]
Message-ID: <14b65198-fea8-377c-2147-86cb2680bee2@gmail.com> (raw)
In-Reply-To: <20191129140446.dqb5y4uzv57cislp@linutronix.de>

On 11/29/19 8:04 AM, Sebastian Andrzej Siewior wrote:
> The phandle cache code allocates memory while holding devtree_lock which
> is a raw_spinlock_t. Memory allocation (and free()) is not possible on
> RT while a raw_spinlock_t is held.
> Invoke the kfree() and kcalloc() while the lock is dropped.
> 
> Cc: Rob Herring <robh+dt@kernel.org>
> Cc: Frank Rowand <frowand.list@gmail.com>
> Cc: devicetree@vger.kernel.org
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
> v1…v2:
>   - Add comments regarding memory allocation under the raw_spinlock_t
>   - use a temporary variable for the memory allocation.
> 
>  drivers/of/base.c |   32 +++++++++++++++++++++++---------
>  1 file changed, 23 insertions(+), 9 deletions(-)
> 
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c
> @@ -136,33 +136,37 @@ static struct device_node **phandle_cach
>  static u32 phandle_cache_mask;
>  
>  /*
> - * Caller must hold devtree_lock.
> + * Caller must hold devtree_lock under which kfree() can't be used so caller
> + * needs to kfree() the returned pointer.
>   */
> -static void __of_free_phandle_cache(void)
> +static struct device_node **__of_free_phandle_cache(void)
>  {
>  	u32 cache_entries = phandle_cache_mask + 1;
>  	u32 k;
> +	struct device_node **shadow;
>  
>  	if (!phandle_cache)
> -		return;
> +		return NULL;
>  
>  	for (k = 0; k < cache_entries; k++)
>  		of_node_put(phandle_cache[k]);
>  
> -	kfree(phandle_cache);
> +	shadow = phandle_cache;
>  	phandle_cache = NULL;
> +	return shadow;
>  }
>  
>  int of_free_phandle_cache(void)
>  {
>  	unsigned long flags;
> +	struct device_node **shadow;
>  
>  	raw_spin_lock_irqsave(&devtree_lock, flags);
>  
> -	__of_free_phandle_cache();
> +	shadow = __of_free_phandle_cache();
>  
>  	raw_spin_unlock_irqrestore(&devtree_lock, flags);
> -
> +	kfree(shadow);
>  	return 0;
>  }
>  #if !defined(CONFIG_MODULES)
> @@ -197,10 +201,12 @@ void of_populate_phandle_cache(void)
>  	u32 cache_entries;
>  	struct device_node *np;
>  	u32 phandles = 0;
> +	struct device_node **shadow;
> +	struct device_node **tmp_phandle_cache;
>  
>  	raw_spin_lock_irqsave(&devtree_lock, flags);
>  
> -	__of_free_phandle_cache();
> +	shadow = __of_free_phandle_cache();
>  
>  	for_each_of_allnodes(np)
>  		if (np->phandle && np->phandle != OF_PHANDLE_ILLEGAL)
> @@ -208,12 +214,19 @@ void of_populate_phandle_cache(void)
>  
>  	if (!phandles)
>  		goto out;
> +	/*
> +	 * kcalloc() can't be used while a raw_spinlock_t held so it is dropped
> +	 * for the allocation.
> +	 */
> +	raw_spin_unlock_irqrestore(&devtree_lock, flags);
>  
>  	cache_entries = roundup_pow_of_two(phandles);
>  	phandle_cache_mask = cache_entries - 1;
>  
> -	phandle_cache = kcalloc(cache_entries, sizeof(*phandle_cache),
> -				GFP_ATOMIC);
> +	tmp_phandle_cache = kcalloc(cache_entries, sizeof(*phandle_cache),
> +				    GFP_ATOMIC);
> +	raw_spin_lock_irqsave(&devtree_lock, flags);
> +	phandle_cache = tmp_phandle_cache;
>  	if (!phandle_cache)
>  		goto out;
>  
> @@ -225,6 +238,7 @@ void of_populate_phandle_cache(void)
>  
>  out:
>  	raw_spin_unlock_irqrestore(&devtree_lock, flags);
> +	kfree(shadow);
>  }
>  
>  void __init of_core_init(void)
> 

Reviewed-by: Frank Rowand <frank.rowand@sony.com>

Compiled and booted on 5.4.  Did not do any more detailed testing.

-Frank

      reply	other threads:[~2019-11-30  2:46 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-11 17:21 [PATCH] of: allocate / free phandle cache outside of the devtree_lock Sebastian Andrzej Siewior
2019-11-12  3:35 ` Rob Herring
2019-11-12  9:10   ` Sebastian Andrzej Siewior
2019-11-12 15:55     ` Rob Herring
2019-11-12 23:46       ` Frank Rowand
2019-11-13  0:48         ` Rob Herring
2019-11-13 16:52           ` Frank Rowand
2019-11-12 22:48 ` Frank Rowand
2019-11-29 13:57   ` Sebastian Andrzej Siewior
2019-11-30  2:48     ` Frank Rowand
2019-11-29 14:04   ` [PATCH v2] " Sebastian Andrzej Siewior
2019-11-30  2:46     ` Frank Rowand [this message]

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=14b65198-fea8-377c-2147-86cb2680bee2@gmail.com \
    --to=frowand.list@gmail.com \
    --cc=bigeasy@linutronix.de \
    --cc=devicetree@vger.kernel.org \
    --cc=robh+dt@kernel.org \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).