| Message ID | 1402664146-14314-1-git-send-email-denis@eukrea.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
+ Alexander On Fri, Jun 13, 2014 at 02:55:45PM +0200, Denis Carikli wrote: > Without that fix, the board freeze during boot. > > This appeared after the following commit: > 496f065 ARM: i.MX: Use of_clk_get_by_name() for timer clocks for DT case. > > Signed-off-by: Denis Carikli <denis@eukrea.com> > --- > arch/arm/boot/dts/imx25.dtsi | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/arm/boot/dts/imx25.dtsi b/arch/arm/boot/dts/imx25.dtsi > index bb74d95..98783f5 100644 > --- a/arch/arm/boot/dts/imx25.dtsi > +++ b/arch/arm/boot/dts/imx25.dtsi > @@ -312,7 +312,7 @@ > gpt4: timer@53f84000 { > compatible = "fsl,imx25-gpt", "fsl,imx31-gpt"; > reg = <0x53f84000 0x4000>; > - clocks = <&clks 9>, <&clks 45>; > + clocks = <&clks 9>, <&clks 47>; So this changes the clock from esdhc1_ipg_per to gpt_ipg_per. It looks to me this is a bug fix on its own. I need some help to understand why the bug is only seen after Alexander's of_clk_get_by_name() patch. Shawn > clock-names = "ipg", "per"; > interrupts = <1>; > }; > -- > 1.7.9.5 >
Sat, 21 Jun 2014 14:36:34 +0800 ?? Shawn Guo <shawn.guo@linaro.org>: > + Alexander > > On Fri, Jun 13, 2014 at 02:55:45PM +0200, Denis Carikli wrote: > > Without that fix, the board freeze during boot. > > > > This appeared after the following commit: > > 496f065 ARM: i.MX: Use of_clk_get_by_name() for timer clocks for DT case. > > > > Signed-off-by: Denis Carikli <denis@eukrea.com> > > --- > > arch/arm/boot/dts/imx25.dtsi | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/arch/arm/boot/dts/imx25.dtsi b/arch/arm/boot/dts/imx25.dtsi > > index bb74d95..98783f5 100644 > > --- a/arch/arm/boot/dts/imx25.dtsi > > +++ b/arch/arm/boot/dts/imx25.dtsi > > @@ -312,7 +312,7 @@ > > gpt4: timer@53f84000 { > > compatible = "fsl,imx25-gpt", "fsl,imx31-gpt"; > > reg = <0x53f84000 0x4000>; > > - clocks = <&clks 9>, <&clks 45>; > > + clocks = <&clks 9>, <&clks 47>; > > So this changes the clock from esdhc1_ipg_per to gpt_ipg_per. It looks > to me this is a bug fix on its own. I need some help to understand why > the bug is only seen after Alexander's of_clk_get_by_name() patch. Clk45 points to esdhc1_ipg_per. Both (DT and non-DT cases) define clocks for GPT as: clk_register_clkdev(clk[ipg], "ipg", "imx-gpt.0"); clk_register_clkdev(clk[gpt_ipg_per], "per", "imx-gpt.0"); ie to valid clocks. Nevertheless, the definition is still not correct. I think that should be so: clocks = <&clks 92>, <&clks 47>; and clk_register_clkdev(clk[gpt1_ipg], "ipg", "imx-gpt.0"); clk_register_clkdev(clk[gpt_ipg_per], "per", "imx-gpt.0"); ---
On Sat, Jun 21, 2014 at 10:53:28AM +0400, Alexander Shiyan wrote: > Sat, 21 Jun 2014 14:36:34 +0800 ?? Shawn Guo <shawn.guo@linaro.org>: > > + Alexander > > > > On Fri, Jun 13, 2014 at 02:55:45PM +0200, Denis Carikli wrote: > > > Without that fix, the board freeze during boot. > > > > > > This appeared after the following commit: > > > 496f065 ARM: i.MX: Use of_clk_get_by_name() for timer clocks for DT case. > > > > > > Signed-off-by: Denis Carikli <denis@eukrea.com> > > > --- > > > arch/arm/boot/dts/imx25.dtsi | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/arch/arm/boot/dts/imx25.dtsi b/arch/arm/boot/dts/imx25.dtsi > > > index bb74d95..98783f5 100644 > > > --- a/arch/arm/boot/dts/imx25.dtsi > > > +++ b/arch/arm/boot/dts/imx25.dtsi > > > @@ -312,7 +312,7 @@ > > > gpt4: timer@53f84000 { > > > compatible = "fsl,imx25-gpt", "fsl,imx31-gpt"; > > > reg = <0x53f84000 0x4000>; > > > - clocks = <&clks 9>, <&clks 45>; > > > + clocks = <&clks 9>, <&clks 47>; > > > > So this changes the clock from esdhc1_ipg_per to gpt_ipg_per. It looks > > to me this is a bug fix on its own. I need some help to understand why > > the bug is only seen after Alexander's of_clk_get_by_name() patch. > > Clk45 points to esdhc1_ipg_per. > Both (DT and non-DT cases) define clocks for GPT as: > clk_register_clkdev(clk[ipg], "ipg", "imx-gpt.0"); > clk_register_clkdev(clk[gpt_ipg_per], "per", "imx-gpt.0"); > ie to valid clocks. Ah, yes. > > Nevertheless, the definition is still not correct. I think that should be so: > clocks = <&clks 92>, <&clks 47>; Right. Seems all gpt nodes need the fix. Shawn > > and > clk_register_clkdev(clk[gpt1_ipg], "ipg", "imx-gpt.0"); > clk_register_clkdev(clk[gpt_ipg_per], "per", "imx-gpt.0"); > > --- >
diff --git a/arch/arm/boot/dts/imx25.dtsi b/arch/arm/boot/dts/imx25.dtsi index bb74d95..98783f5 100644 --- a/arch/arm/boot/dts/imx25.dtsi +++ b/arch/arm/boot/dts/imx25.dtsi @@ -312,7 +312,7 @@ gpt4: timer@53f84000 { compatible = "fsl,imx25-gpt", "fsl,imx31-gpt"; reg = <0x53f84000 0x4000>; - clocks = <&clks 9>, <&clks 45>; + clocks = <&clks 9>, <&clks 47>; clock-names = "ipg", "per"; interrupts = <1>; };
Without that fix, the board freeze during boot. This appeared after the following commit: 496f065 ARM: i.MX: Use of_clk_get_by_name() for timer clocks for DT case. Signed-off-by: Denis Carikli <denis@eukrea.com> --- arch/arm/boot/dts/imx25.dtsi | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)