public inbox for linux-arm-kernel@lists.infradead.org
 help / color / mirror / Atom feed
From: maxime.ripard@bootlin.com (Maxime Ripard)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH 2/2] clk: core: link consumer with clock driver
Date: Tue, 27 Nov 2018 13:38:58 +0100	[thread overview]
Message-ID: <20181127123858.3v344nonsuyqskrt@flea> (raw)
In-Reply-To: <20181122212212.16039-3-miquel.raynal@bootlin.com>

Hi Miquel,

On Thu, Nov 22, 2018 at 10:22:12PM +0100, Miquel Raynal wrote:
> One major concern when, for instance, suspending/resuming a platform
> is to never access registers before the underlying clock has been
> resumed, otherwise most of the time the kernel will just crash. One
> solution is to use syscore operations when registering clock drivers
> suspend/resume callbacks. One problem of using syscore_ops is that the
> suspend/resume scheduling will depend on the order of the
> registrations, which brings (unacceptable) randomness in the process.
> 
> A feature called device links has been introduced to handle such
> situation. It creates dependencies between consumers and providers,
> enforcing e.g. the suspend/resume order when needed. Such feature is
> already in use for regulators.
> 
> Add device links support in the clock subsystem by creating/deleting
> the links at get/put time.
> 
> Example of a boot (ESPRESSObin, A3700 SoC) with devices linked to clocks:
> ahci-mvebu d00e0000.sata: Linked as a consumer to d0013000.nb-periph-clk
> mvneta d0030000.ethernet: Linked as a consumer to d0018000.sb-periph-clk
> xhci-hcd d0058000.usb: Linked as a consumer to d0018000.sb-periph-clk
> xenon-sdhci d00d0000.sdhci: Linked as a consumer to d0013000.nb-periph-clk
> xenon-sdhci d00d0000.sdhci: Dropping the link to d0013000.nb-periph-clk
> advk-pcie d0070000.pcie: Linked as a consumer to d0018000.sb-periph-clk
> xenon-sdhci d00d0000.sdhci: Linked as a consumer to d0013000.nb-periph-clk
> xenon-sdhci d00d0000.sdhci: Linked as a consumer to regulator.1
> cpu cpu0: Linked as a consumer to d0013000.nb-periph-clk
> cpu cpu0: Dropping the link to d0013000.nb-periph-clk
> cpu cpu0: Linked as a consumer to d0013000.nb-periph-clk
> 
> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
> ---
>  drivers/clk/clk.c            | 20 ++++++++++++++++++++
>  drivers/clk/clkdev.c         | 13 ++++++++++---
>  include/linux/clk-provider.h |  2 ++
>  3 files changed, 32 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
> index b799347c5fd6..33a0f2b0533a 100644
> --- a/drivers/clk/clk.c
> +++ b/drivers/clk/clk.c
> @@ -90,6 +90,7 @@ struct clk {
>  	unsigned long max_rate;
>  	unsigned int exclusive_count;
>  	struct hlist_node clks_node;
> +	struct device_link *link;
>  };
>  
>  /***           runtime pm          ***/
> @@ -262,6 +263,25 @@ struct clk_hw *__clk_get_hw(struct clk *clk)
>  }
>  EXPORT_SYMBOL_GPL(__clk_get_hw);
>  
> +void __clk_device_link(struct device *consumer, struct clk *clk)
> +{
> +	if (!consumer || !clk || !clk->core)
> +		return;
> +
> +	clk->link = device_link_add(consumer, clk->core->dev,
> +				    DL_FLAG_STATELESS);
> +}
> +EXPORT_SYMBOL_GPL(__clk_device_link);
> +
> +void __clk_device_unlink(struct clk *clk)
> +{
> +	if (!clk || !clk->link)
> +		return;
> +
> +	device_link_del(clk->link);
> +}
> +EXPORT_SYMBOL_GPL(__clk_device_unlink);
> +
>  unsigned int clk_hw_get_num_parents(const struct clk_hw *hw)
>  {
>  	return hw->core->num_parents;
> diff --git a/drivers/clk/clkdev.c b/drivers/clk/clkdev.c
> index 9ab3db8b3988..fccfd4c01457 100644
> --- a/drivers/clk/clkdev.c
> +++ b/drivers/clk/clkdev.c
> @@ -194,20 +194,27 @@ EXPORT_SYMBOL(clk_get_sys);
>  struct clk *clk_get(struct device *dev, const char *con_id)
>  {
>  	const char *dev_id = dev ? dev_name(dev) : NULL;
> -	struct clk *clk;
> +	struct clk *clk = NULL;
>  
>  	if (dev && dev->of_node) {
>  		clk = __of_clk_get_by_name(dev->of_node, dev_id, con_id);
> -		if (!IS_ERR(clk) || PTR_ERR(clk) == -EPROBE_DEFER)
> +		if (PTR_ERR(clk) == -EPROBE_DEFER)
>  			return clk;
>  	}
>  
> -	return clk_get_sys(dev_id, con_id);
> +	if (IS_ERR_OR_NULL(clk))
> +		clk = clk_get_sys(dev_id, con_id);
> +
> +	if (!IS_ERR(clk))
> +		__clk_device_link(dev, clk);
> +
> +	return clk;

I think this doesn't address all the cases. In your case, where you
have one consumer that is not a clock, and one provider that is a
clock, it works just fine.

However, if you have clocks providers chained, for example with one
oscillator, a clock controller, and a device, the link will be created
between the device and the controller, but there will be no link
between the controller and the oscillator.

Adding a link in __clk_init_parent looks like it would address that
case.

Maxime

-- 
Maxime Ripard, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20181127/fe347eb3/attachment.sig>

  parent reply	other threads:[~2018-11-27 12:38 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-11-22 21:22 [PATCH 0/2] Link consumer with clock driver Miquel Raynal
2018-11-22 21:22 ` [PATCH 1/2] clk: core: clarify the check for runtime PM Miquel Raynal
2018-11-22 21:22 ` [PATCH 2/2] clk: core: link consumer with clock driver Miquel Raynal
2018-11-23  8:30   ` kbuild test robot
2018-11-23  9:11     ` Miquel Raynal
2018-11-30  9:26       ` Stephen Boyd
2018-11-30 10:20         ` Miquel Raynal
2018-12-03 19:20           ` Stephen Boyd
2018-12-03 22:16             ` Miquel Raynal
2018-12-03 22:28               ` Stephen Boyd
2018-11-27 12:38   ` Maxime Ripard [this message]
2018-11-29 16:03     ` Miquel Raynal
2018-11-30  9:24 ` [PATCH 0/2] Link " Stephen Boyd
2018-11-30 12:00   ` Miquel Raynal

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=20181127123858.3v344nonsuyqskrt@flea \
    --to=maxime.ripard@bootlin.com \
    --cc=linux-arm-kernel@lists.infradead.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox