ASoC: rsnd: SSI needs to prepare clocks before enabling
diff mbox

Message ID 1396009626-24699-1-git-send-email-ben.dooks@codethink.co.uk
State New, archived
Headers show

Commit Message

Ben Dooks March 28, 2014, 12:27 p.m. UTC
From: William Towle <william.towle@codethink.co.uk>

The rsnd_ssi_hw_{stop|start} functions need to call clk_prepare_enable()
and clk_disable_unprepare(), otherwise the following warning occurs on
attempt to play sound:

WARNING: CPU: 2 PID: 1176 at drivers/clk/clk.c:926 __clk_enable+0x30/0x9c()

due to drivers/clk/clk.c:

    926         if (WARN_ON(clk->prepare_count == 0))
    927                 return -ESHUTDOWN;

Signed-off-by: William Towle <william.towle@codethink.co.uk>
[ben.dooks@codethink.co.uk: add more bug info to report]
Tested-by: Ben Dooks <ben.dooks@codethink.co.uk>
---
 sound/soc/sh/rcar/ssi.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Comments

Mark Brown March 29, 2014, 10:09 a.m. UTC | #1
On Fri, Mar 28, 2014 at 12:27:06PM +0000, Ben Dooks wrote:
> From: William Towle <william.towle@codethink.co.uk>
> 
> The rsnd_ssi_hw_{stop|start} functions need to call clk_prepare_enable()
> and clk_disable_unprepare(), otherwise the following warning occurs on
> attempt to play sound:

This doesn't apply cleanly against current code, can you please check and
resend - I can't apply this anyway since...

> Signed-off-by: William Towle <william.towle@codethink.co.uk>
> [ben.dooks@codethink.co.uk: add more bug info to report]
> Tested-by: Ben Dooks <ben.dooks@codethink.co.uk>
> ---

...you forgot to sign off.  I'm also not seeing the debug info you
mention in the changelog, indeed there's no error checking.
Ben Dooks April 4, 2014, 11:42 a.m. UTC | #2
On Sat, Mar 29, 2014 at 10:09:06AM +0000, Mark Brown wrote:
> On Fri, Mar 28, 2014 at 12:27:06PM +0000, Ben Dooks wrote:
> > From: William Towle <william.towle@codethink.co.uk>
> > 
> > The rsnd_ssi_hw_{stop|start} functions need to call clk_prepare_enable()
> > and clk_disable_unprepare(), otherwise the following warning occurs on
> > attempt to play sound:
> 
> This doesn't apply cleanly against current code, can you please check and
> resend - I can't apply this anyway since...
> 
> > Signed-off-by: William Towle <william.towle@codethink.co.uk>
> > [ben.dooks@codethink.co.uk: add more bug info to report]
> > Tested-by: Ben Dooks <ben.dooks@codethink.co.uk>
> > ---
> 
> ...you forgot to sign off.  I'm also not seeing the debug info you
> mention in the changelog, indeed there's no error checking.

Ooops. 

I need to re-do this anyway as we need to prepare the clock separately
from the enable, as the enable code is occasionally being called from
an interrupt context.

I anm not sure what you mean by debug info... there is the log from
the WARN_ON() which is trigger from the clock driver.

Patch
diff mbox

diff --git a/sound/soc/sh/rcar/ssi.c b/sound/soc/sh/rcar/ssi.c
index 4b8cf7c..1c07ef7 100644
--- a/sound/soc/sh/rcar/ssi.c
+++ b/sound/soc/sh/rcar/ssi.c
@@ -250,7 +250,7 @@  static void rsnd_ssi_hw_start(struct rsnd_ssi *ssi,
 	u32 cr;
 
 	if (0 == ssi->usrcnt) {
-		clk_enable(ssi->clk);
+		clk_prepare_enable(ssi->clk);
 
 		if (rsnd_rdai_is_clk_master(rdai)) {
 			if (rsnd_ssi_clk_from_parent(ssi))
@@ -309,7 +309,7 @@  static void rsnd_ssi_hw_stop(struct rsnd_ssi *ssi,
 				rsnd_ssi_master_clk_stop(ssi);
 		}
 
-		clk_disable(ssi->clk);
+		clk_disable_unprepare(ssi->clk);
 	}
 
 	dev_dbg(dev, "ssi%d hw stopped\n", rsnd_mod_id(&ssi->mod));