All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jerome Brunet <jbrunet@baylibre.com>
To: Michael Turquette <mturquette@baylibre.com>,
	Stephen Boyd <sboyd@codeaurora.org>,
	Kevin Hilman <khilman@baylibre.com>
Cc: linux-clk@vger.kernel.org, linux-amlogic@lists.infradead.org,
	Linus Walleij <linus.walleij@linaro.org>,
	Boris Brezillon <boris.brezillon@free-electrons.com>
Subject: Re: [PATCH v2 11/11] clk: move CLK_SET_RATE_GATE protection from prepare to enable
Date: Mon, 29 May 2017 11:17:32 +0200	[thread overview]
Message-ID: <1496049452.7514.5.camel@baylibre.com> (raw)
In-Reply-To: <20170521215958.19743-12-jbrunet@baylibre.com>

On Sun, 2017-05-21 at 23:59 +0200, Jerome Brunet wrote:
> CLK_SET_RATE_GATE name suggest that the rate can be set when the provider
> is gated (disabled). With the current implementation, the rate can only be
> set when the provider is unprepared, while it should be allowed to set a
> prepared and disable provider.
> Fix this by moving the rate protection mechanism in the enable/disable
> core functions
> 

I'll drop this patch in the v3.
The protection count should only be touch while holding the prepare_lock.
This was disaster waiting to happen ... :(

> Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
> ---
>  drivers/clk/clk.c | 23 ++++++++++++-----------
>  1 file changed, 12 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
> index 6ee5fc59cf1f..e6e5048ce186 100644
> --- a/drivers/clk/clk.c
> +++ b/drivers/clk/clk.c
> @@ -491,9 +491,6 @@ static void clk_core_unprepare(struct clk_core *core)
>  	if (WARN_ON(core->prepare_count == 1 && core->flags &
> CLK_IS_CRITICAL))
>  		return;
>  
> -	if (core->flags & CLK_SET_RATE_GATE)
> -		clk_core_rate_unprotect(core);
> -
>  	if (--core->prepare_count > 0)
>  		return;
>  
> @@ -564,14 +561,6 @@ static int clk_core_prepare(struct clk_core *core)
>  
>  	core->prepare_count++;
>  
> -	/*
> -	 * CLK_SET_RATE_GATE is a special case of clock protection
> -	 * Instead of a consumer protection, the provider is protecting
> -	 * itself when prepared
> -	 */
> -	if (core->flags & CLK_SET_RATE_GATE)
> -		clk_core_rate_protect(core);
> -
>  	return 0;
>  }
>  
> @@ -716,6 +705,9 @@ static void clk_core_disable(struct clk_core *core)
>  	if (WARN_ON(core->enable_count == 1 && core->flags &
> CLK_IS_CRITICAL))
>  		return;
>  
> +	if (core->flags & CLK_SET_RATE_GATE)
> +		clk_core_rate_unprotect(core);
> +
>  	if (--core->enable_count > 0)
>  		return;
>  
> @@ -791,6 +783,15 @@ static int clk_core_enable(struct clk_core *core)
>  	}
>  
>  	core->enable_count++;
> +
> +	/*
> +	 * CLK_SET_RATE_GATE is a special case of clock protection
> +	 * Instead of a consumer protection, the provider is protecting
> +	 * itself when enabled
> +	 */
> +	if (core->flags & CLK_SET_RATE_GATE)
> +		clk_core_rate_protect(core);
> +
>  	return 0;
>  }
>  

WARNING: multiple messages have this Message-ID (diff)
From: jbrunet@baylibre.com (Jerome Brunet)
To: linus-amlogic@lists.infradead.org
Subject: [PATCH v2 11/11] clk: move CLK_SET_RATE_GATE protection from prepare to enable
Date: Mon, 29 May 2017 11:17:32 +0200	[thread overview]
Message-ID: <1496049452.7514.5.camel@baylibre.com> (raw)
In-Reply-To: <20170521215958.19743-12-jbrunet@baylibre.com>

On Sun, 2017-05-21 at 23:59 +0200, Jerome Brunet wrote:
> CLK_SET_RATE_GATE name suggest that the rate can be set when the provider
> is gated (disabled). With the current implementation, the rate can only be
> set when the provider is unprepared, while it should be allowed to set a
> prepared and disable provider.
> Fix this by moving the rate protection mechanism in the enable/disable
> core functions
> 

I'll drop this patch in the v3.
The protection count should only be touch while holding the prepare_lock.
This was disaster waiting to happen ... :(

> Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
> ---
> ?drivers/clk/clk.c | 23 ++++++++++++-----------
> ?1 file changed, 12 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
> index 6ee5fc59cf1f..e6e5048ce186 100644
> --- a/drivers/clk/clk.c
> +++ b/drivers/clk/clk.c
> @@ -491,9 +491,6 @@ static void clk_core_unprepare(struct clk_core *core)
> ?	if (WARN_ON(core->prepare_count == 1 && core->flags &
> CLK_IS_CRITICAL))
> ?		return;
> ?
> -	if (core->flags & CLK_SET_RATE_GATE)
> -		clk_core_rate_unprotect(core);
> -
> ?	if (--core->prepare_count > 0)
> ?		return;
> ?
> @@ -564,14 +561,6 @@ static int clk_core_prepare(struct clk_core *core)
> ?
> ?	core->prepare_count++;
> ?
> -	/*
> -	?* CLK_SET_RATE_GATE is a special case of clock protection
> -	?* Instead of a consumer protection, the provider is protecting
> -	?* itself when prepared
> -	?*/
> -	if (core->flags & CLK_SET_RATE_GATE)
> -		clk_core_rate_protect(core);
> -
> ?	return 0;
> ?}
> ?
> @@ -716,6 +705,9 @@ static void clk_core_disable(struct clk_core *core)
> ?	if (WARN_ON(core->enable_count == 1 && core->flags &
> CLK_IS_CRITICAL))
> ?		return;
> ?
> +	if (core->flags & CLK_SET_RATE_GATE)
> +		clk_core_rate_unprotect(core);
> +
> ?	if (--core->enable_count > 0)
> ?		return;
> ?
> @@ -791,6 +783,15 @@ static int clk_core_enable(struct clk_core *core)
> ?	}
> ?
> ?	core->enable_count++;
> +
> +	/*
> +	?* CLK_SET_RATE_GATE is a special case of clock protection
> +	?* Instead of a consumer protection, the provider is protecting
> +	?* itself when enabled
> +	?*/
> +	if (core->flags & CLK_SET_RATE_GATE)
> +		clk_core_rate_protect(core);
> +
> ?	return 0;
> ?}
> ?

  reply	other threads:[~2017-05-29  9:17 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-05-21 21:59 [PATCH v2 00/11] clk: implement clock rate protection mechanism Jerome Brunet
2017-05-21 21:59 ` Jerome Brunet
2017-05-21 21:59 ` [PATCH v2 01/11] clk: take the prepare lock out of clk_core_set_parent Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-25 18:54   ` Michael Turquette
2017-05-25 18:54     ` Michael Turquette
2017-05-29  9:35     ` Jerome Brunet
2017-05-29  9:35       ` Jerome Brunet
2017-05-21 21:59 ` [PATCH v2 02/11] clk: add clk_core_set_phase_nolock function Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-23  9:35   ` Adriana Reus
2017-05-23  9:35     ` Adriana Reus
2017-05-23  9:48     ` Jerome Brunet
2017-05-23  9:48       ` Jerome Brunet
2017-05-25 18:58       ` Michael Turquette
2017-05-25 18:58         ` Michael Turquette
2017-05-21 21:59 ` [PATCH v2 03/11] clk: rework calls to round and determine rate callbacks Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-25 20:13   ` Michael Turquette
2017-05-25 20:13     ` Michael Turquette
2017-05-21 21:59 ` [PATCH v2 04/11] clk: use round rate to bail out early in set_rate Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-25 20:20   ` Michael Turquette
2017-05-25 20:20     ` Michael Turquette
2017-05-29  9:12     ` Jerome Brunet
2017-05-29  9:12       ` Jerome Brunet
2017-05-21 21:59 ` [PATCH v2 05/11] clk: add support for clock protection Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-25 20:58   ` Michael Turquette
2017-05-25 20:58     ` Michael Turquette
2017-05-29  9:15     ` Jerome Brunet
2017-05-29  9:15       ` Jerome Brunet
2017-05-21 21:59 ` [PATCH v2 06/11] clk: add clk_set_rate_protect Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-21 21:59 ` [PATCH v2 07/11] clk: rollback set_rate_range changes on failure Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-21 21:59 ` [PATCH v2 08/11] clk: cosmetic changes to clk_summary debugfs entry Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-21 21:59 ` [PATCH v2 09/11] clk: fix incorrect usage of ENOSYS Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-21 21:59 ` [PATCH v2 10/11] clk: fix CLK_SET_RATE_GATE with clock rate protection Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-23 13:42   ` Adriana Reus
2017-05-23 13:42     ` Adriana Reus
2017-05-23 15:09     ` Jerome Brunet
2017-05-23 15:09       ` Jerome Brunet
2017-05-21 21:59 ` [PATCH v2 11/11] clk: move CLK_SET_RATE_GATE protection from prepare to enable Jerome Brunet
2017-05-21 21:59   ` Jerome Brunet
2017-05-29  9:17   ` Jerome Brunet [this message]
2017-05-29  9:17     ` Jerome Brunet

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=1496049452.7514.5.camel@baylibre.com \
    --to=jbrunet@baylibre.com \
    --cc=boris.brezillon@free-electrons.com \
    --cc=khilman@baylibre.com \
    --cc=linus.walleij@linaro.org \
    --cc=linux-amlogic@lists.infradead.org \
    --cc=linux-clk@vger.kernel.org \
    --cc=mturquette@baylibre.com \
    --cc=sboyd@codeaurora.org \
    /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.