ASoC: kirkwood: fix loss of external clock at probe time
PATCH v3 2/2 ASoC: kirkwood: fix loss of external clock at probe time
At probe time, when the clock driver is not yet initialized, the external clock of the kirkwood sound device will not be usable.
This patch fixes this problem defering the device probe.
It also removes the test about same internal and external clocks which can never occur.
---
v3: remove the test of same internal and external clocks
The associated patch:
[PATCH v3 1/2] ASoC: kirkwood: clk: probe defer when clock not yet ready is unchanged and not resent
--- sound/soc/kirkwood/kirkwood-i2s.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-)
sound/soc/kirkwood/kirkwood-i2s.c +7 -9
@@ -496,15 +496,13 @@ static int kirkwood_i2s_dev_probe(struct platform_device *pdev)return err; priv->extclk =devm_clk_get (&pdev->dev,"extclk" ); -if (!IS_ERR (priv->extclk)) { -if (priv->extclk == priv->clk) { -devm_clk_put (&pdev->dev, priv->extclk); - priv->extclk =ERR_PTR (-EINVAL ); - }else { -dev_info (&pdev->dev,"found external clock\n" ); -clk_prepare_enable (priv->extclk); - soc_dai = &kirkwood_i2s_dai_extclk; - } +if (IS_ERR (priv->extclk)) { +if (PTR_ERR (priv->extclk) == -EPROBE_DEFER ) +return -EPROBE_DEFER ; + }else { +dev_info (&pdev->dev,"found external clock\n" ); +clk_prepare_enable (priv->extclk); + soc_dai = kirkwood_i2s_dai_extclk; }/* Some sensible defaults - this reflects the powerup values */
Signature · 6 lines
Ken ar c'hentañ | ** Breizh ha Linux atav! ** Jef | http://moinejf.free.fr/ _______________________________________________ Alsa-devel mailing list [email protected] http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Hello,
Quoted 10 lines · Jean-Francois Moine
sound/soc/kirkwood/kirkwood-i2s.c +0
-2@@ -496,15 +496,13 @@ static int kirkwood_i2s_dev_probe(struct platform_device *pdev)return err; priv->extclk =devm_clk_get (&pdev->dev,"extclk" ); -if (!IS_ERR (priv->extclk)) { -if (priv->extclk == priv->clk) {
sound/soc/kirkwood/kirkwood-i2s.c static int kirkwood_i2s_dev_probe(struct platform_device *pdev) · line 501
Quoted 10 lines
sound/soc/kirkwood/kirkwood-i2s.c +3
-7-devm_clk_put (&pdev->dev, priv->extclk); - priv->extclk =ERR_PTR (-EINVAL ); - }else { -dev_info (&pdev->dev,"found external clock\n" ); -clk_prepare_enable (priv->extclk); - soc_dai = &kirkwood_i2s_dai_extclk; - } +if (IS_ERR (priv->extclk)) { +if (PTR_ERR (priv->extclk) == -EPROBE_DEFER ) +return -EPROBE_DEFER ;
Maybe the better logic here is:
if (!PTR_ERR(priv->extclk) == -ENOENT) return PTR_ERR(priv->extclk);
?
sound/soc/kirkwood/kirkwood-i2s.c static int kirkwood_i2s_dev_probe(struct platform_device *pdev) · line 505
Quoted 4 lines
sound/soc/kirkwood/kirkwood-i2s.c +4
-0+ }else { +dev_info (&pdev->dev,"found external clock\n" ); +clk_prepare_enable (priv->extclk); + soc_dai = kirkwood_i2s_dai_extclk;
Another todo (that is already in the patched code) is that you should check the return value of clk_prepare_enable.
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-K?nig | Industrial Linux Solutions | http://www.pengutronix.de/ |
Quoted 8 lines · On Fri, 18 Oct 2013 21:12:59 +0200 Uwe Kleine-König…
sound/soc/kirkwood/kirkwood-i2s.c static int kirkwood_i2s_dev_probe(struct platform_device *pdev) · line 501
Quoted 3 lines
sound/soc/kirkwood/kirkwood-i2s.c +3
-0+if (IS_ERR (priv->extclk)) { +if (PTR_ERR (priv->extclk) == -EPROBE_DEFER ) +return -EPROBE_DEFER ;Maybe the better logic here is:
if (!PTR_ERR(priv->extclk) == -ENOENT) return PTR_ERR(priv->extclk);?
No. This patch is associated with an other one which returns -EPROBE_DEFER when the external clock is declared in the DT and when the clock driver is not yet initialized. Then, the kirkwood modules must be probed later.
For any other error, the external clock is not used, and it is not easy to know if the error is due to the absence of external clock in the DT or to a DT parsing error or to anything else.
Quoted 6 lines
sound/soc/kirkwood/kirkwood-i2s.c static int kirkwood_i2s_dev_probe(struct platform_device *pdev) · line 505
Quoted 4 lines
sound/soc/kirkwood/kirkwood-i2s.c +4
-0+ }else { +dev_info (&pdev->dev,"found external clock\n" ); +clk_prepare_enable (priv->extclk); + soc_dai = kirkwood_i2s_dai_extclk;Another todo (that is already in the patched code) is that you should check the return value of
clk_prepare_enable.
You are right, but an error at this level should not often occur...
Signature · 6 lines
Ken ar c'hentañ | ** Breizh ha Linux atav! ** Jef | http://moinejf.free.fr/ _______________________________________________ Alsa-devel mailing list [email protected] http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
Quoted 10 lines · Jean-Francois Moine
Quoted 7 lines · On Fri, 18 Oct 2013 21:12:59 +0200 Uwe Kleine-K?nig…
sound/soc/kirkwood/kirkwood-i2s.c static int kirkwood_i2s_dev_probe(struct platform_device *pdev) · line 501
Quoted 3 lines
sound/soc/kirkwood/kirkwood-i2s.c +3
-0+if (IS_ERR (priv->extclk)) { +if (PTR_ERR (priv->extclk) == -EPROBE_DEFER ) +return -EPROBE_DEFER ;Maybe the better logic here is:
if (!PTR_ERR(priv->extclk) == -ENOENT) return PTR_ERR(priv->extclk);
Quoted 6 lines
Quoted 1 line
?
No. This patch is associated with an other one which returns -EPROBE_DEFER when the external clock is declared in the DT and when the clock driver is not yet initialized. Then, the kirkwood modules must be probed later.
Yes, that's understood. My suggestion behaves as your's for the return values -EPROBE_DEFER and -ENOENT, so the deferred probe should work, too. The question is only what you want to do for other errors (don't know if they can happen at this stage). I'd say, on a dt parsing error bail out, too.
Best regards
Uwe
Signature · 2 lines
Pengutronix e.K. | Uwe Kleine-K?nig | Industrial Linux Solutions | http://www.pengutronix.de/ |
Mark Brown
maintainerQuoted 5 lines · Uwe Kleine-König
Yes, that's understood. My suggestion behaves as your's for the return values -EPROBE_DEFER and
-ENOENT, so the deferred probe should work, too. The question is only what you want to do for other errors (don't know if they can happen at this stage). I'd say, on a dt parsing error bail out, too.
It seems like the best thing to do is defer if the error might resolve itself but otherwise carry on with the reduced functionality from the internal clocks. I'd expect most errors to be permanently fatal though, certainly a failure to parse the DT is unlikely to get fixed at runtime.
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Mark Brown
maintainerQuoted 2 lines · Jean-Francois Moine
At probe time, when the clock driver is not yet initialized, the external clock of the kirkwood sound device will not be usable.
This doesn't apply against -next, could you please check and resend?
git://git.kernel.org/pub/scm/linux/kernel/git/broone/sound.git topic/kirkwood
or the for-next branch.
Quoted 2 lines
It also removes the test about same internal and external clocks which can never occur.
It seems like reasonable defensiveness to have the test there in case
someone has a broken DT, though it'd be better to complain about the DT.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131020/3521b125/attachment.sig>
Quoted 9 lines · On Sun, 20 Oct 2013 17:38:19 +0100 Mark Brown <[email protected]> wrote:
Quoted 2 lines · Jean-Francois Moine
At probe time, when the clock driver is not yet initialized, the external clock of the kirkwood sound device will not be usable.
This doesn't apply against -next, could you please check and resend?
git://git.kernel.org/pub/scm/linux/kernel/git/broone/sound.gittopic/kirkwoodor the for-next branch.
Not easy to find :)
Quoted 5 lines
Quoted 2 lines
It also removes the test about same internal and external clocks which can never occur.
It seems like reasonable defensiveness to have the test there in case someone has a broken DT, though it'd be better to complain about the DT.
The removed test run into:
clocks = <&si5351 2>, <&si5351 2>; clock-names = "internal", "extclk";
where it seems that the user really wanted to make a mistake. Then, the actual code just wants the si5351 2 to work as the internal clock, and that cannot be.
Also, about errors, you may see that a normal error as:
clocks = <&si5351 2>, <&gate_clk 13>; clock-names = "internal", "extclk";
i.e. inversion of the internal and external clock phandle's, cannot be detected...
So, it is better to remove the useless test.
Signature · 6 lines
Ken ar c'hentañ | ** Breizh ha Linux atav! ** Jef | http://moinejf.free.fr/ _______________________________________________ Alsa-devel mailing list [email protected] http://mailman.alsa-project.org/mailman/listinfo/alsa-devel