diff mbox

[v5] ARM: davinci: da8xx: Fix sleeping function called from invalid context

Message ID 1480947224-26222-1-git-send-email-abailon@baylibre.com (mailing list archive)
State New, archived
Headers show

Commit Message

Alexandre Bailon Dec. 5, 2016, 2:13 p.m. UTC
Everytime the usb20 phy is enabled, there is a
"sleeping function called from invalid context" BUG.

clk_enable() from arch/arm/mach-davinci/clock.c uses spin_lock_irqsave()
before to invoke the callback usb20_phy_clk_enable().
usb20_phy_clk_enable() uses clk_get() and clk_enable_prepapre()
which may sleep.
Move clk_get() to da8xx_register_usb20_phy_clk() and
replace clk_prepare_enable() by clk_enable().

Signed-off-by: Alexandre Bailon <abailon@baylibre.com>
---
 arch/arm/mach-davinci/usb-da8xx.c | 29 +++++++++++++++--------------
 1 file changed, 15 insertions(+), 14 deletions(-)

Comments

Sekhar Nori Dec. 6, 2016, 9:33 a.m. UTC | #1
On Monday 05 December 2016 07:43 PM, Alexandre Bailon wrote:
> Everytime the usb20 phy is enabled, there is a
> "sleeping function called from invalid context" BUG.
> 
> clk_enable() from arch/arm/mach-davinci/clock.c uses spin_lock_irqsave()
> before to invoke the callback usb20_phy_clk_enable().
> usb20_phy_clk_enable() uses clk_get() and clk_enable_prepapre()
> which may sleep.
> Move clk_get() to da8xx_register_usb20_phy_clk() and
> replace clk_prepare_enable() by clk_enable().
> 
> Signed-off-by: Alexandre Bailon <abailon@baylibre.com>

This will still cause the recursive locking problem reported by David.
Not sure what the point of sending this version was.

Thanks,
Sekhar
Alexandre Bailon Dec. 6, 2016, 9:51 a.m. UTC | #2
On 12/06/2016 10:33 AM, Sekhar Nori wrote:
> On Monday 05 December 2016 07:43 PM, Alexandre Bailon wrote:
>> Everytime the usb20 phy is enabled, there is a
>> "sleeping function called from invalid context" BUG.
>>
>> clk_enable() from arch/arm/mach-davinci/clock.c uses spin_lock_irqsave()
>> before to invoke the callback usb20_phy_clk_enable().
>> usb20_phy_clk_enable() uses clk_get() and clk_enable_prepapre()
>> which may sleep.
>> Move clk_get() to da8xx_register_usb20_phy_clk() and
>> replace clk_prepare_enable() by clk_enable().
>>
>> Signed-off-by: Alexandre Bailon <abailon@baylibre.com>
> 
> This will still cause the recursive locking problem reported by David.
> Not sure what the point of sending this version was.
> 
> Thanks,
> Sekhar
> 
What am I supposed to do ?

Thanks,
Alexandre
Sekhar Nori Dec. 6, 2016, 9:57 a.m. UTC | #3
On Tuesday 06 December 2016 03:21 PM, Alexandre Bailon wrote:
> On 12/06/2016 10:33 AM, Sekhar Nori wrote:
>> On Monday 05 December 2016 07:43 PM, Alexandre Bailon wrote:
>>> Everytime the usb20 phy is enabled, there is a
>>> "sleeping function called from invalid context" BUG.
>>>
>>> clk_enable() from arch/arm/mach-davinci/clock.c uses spin_lock_irqsave()
>>> before to invoke the callback usb20_phy_clk_enable().
>>> usb20_phy_clk_enable() uses clk_get() and clk_enable_prepapre()
>>> which may sleep.
>>> Move clk_get() to da8xx_register_usb20_phy_clk() and
>>> replace clk_prepare_enable() by clk_enable().
>>>
>>> Signed-off-by: Alexandre Bailon <abailon@baylibre.com>
>>
>> This will still cause the recursive locking problem reported by David.
>> Not sure what the point of sending this version was.
>>
>> Thanks,
>> Sekhar
>>

> What am I supposed to do ?

That needs to be resolved between you and David. Perhaps convert the fix
sent by David into a proper patch and base this patch on that. Or wait
for David to send it himself and let him also make the modifications
needed in this patch.

David ?

Thanks,
Sekhar
David Lechner Dec. 6, 2016, 6:56 p.m. UTC | #4
On 12/06/2016 03:57 AM, Sekhar Nori wrote:
> On Tuesday 06 December 2016 03:21 PM, Alexandre Bailon wrote:
>> On 12/06/2016 10:33 AM, Sekhar Nori wrote:
>>> On Monday 05 December 2016 07:43 PM, Alexandre Bailon wrote:
>>>> Everytime the usb20 phy is enabled, there is a
>>>> "sleeping function called from invalid context" BUG.
>>>>
>>>> clk_enable() from arch/arm/mach-davinci/clock.c uses spin_lock_irqsave()
>>>> before to invoke the callback usb20_phy_clk_enable().
>>>> usb20_phy_clk_enable() uses clk_get() and clk_enable_prepapre()
>>>> which may sleep.
>>>> Move clk_get() to da8xx_register_usb20_phy_clk() and
>>>> replace clk_prepare_enable() by clk_enable().
>>>>
>>>> Signed-off-by: Alexandre Bailon <abailon@baylibre.com>
>>>
>>> This will still cause the recursive locking problem reported by David.
>>> Not sure what the point of sending this version was.
>>>
>>> Thanks,
>>> Sekhar
>>>
>
>> What am I supposed to do ?
>
> That needs to be resolved between you and David. Perhaps convert the fix
> sent by David into a proper patch and base this patch on that. Or wait
> for David to send it himself and let him also make the modifications
> needed in this patch.
>
> David ?
>
> Thanks,
> Sekhar
>

Alexandre, I was hoping that you would just squash my patch with your 
patch and take Sekhar's suggestion about a separate patch to make the 
private __clk_enable() public as davinci_clk_enable() when you re-submit.

You can add "Suggested-By: David Lechner <david@lechnology.com>" to the 
commit message if you would like to give me some credit for my ideas.
Alexandre Bailon Dec. 7, 2016, 9:07 a.m. UTC | #5
On 12/06/2016 07:56 PM, David Lechner wrote:
> On 12/06/2016 03:57 AM, Sekhar Nori wrote:
>> On Tuesday 06 December 2016 03:21 PM, Alexandre Bailon wrote:
>>> On 12/06/2016 10:33 AM, Sekhar Nori wrote:
>>>> On Monday 05 December 2016 07:43 PM, Alexandre Bailon wrote:
>>>>> Everytime the usb20 phy is enabled, there is a
>>>>> "sleeping function called from invalid context" BUG.
>>>>>
>>>>> clk_enable() from arch/arm/mach-davinci/clock.c uses
>>>>> spin_lock_irqsave()
>>>>> before to invoke the callback usb20_phy_clk_enable().
>>>>> usb20_phy_clk_enable() uses clk_get() and clk_enable_prepapre()
>>>>> which may sleep.
>>>>> Move clk_get() to da8xx_register_usb20_phy_clk() and
>>>>> replace clk_prepare_enable() by clk_enable().
>>>>>
>>>>> Signed-off-by: Alexandre Bailon <abailon@baylibre.com>
>>>>
>>>> This will still cause the recursive locking problem reported by David.
>>>> Not sure what the point of sending this version was.
>>>>
>>>> Thanks,
>>>> Sekhar
>>>>
>>
>>> What am I supposed to do ?
>>
>> That needs to be resolved between you and David. Perhaps convert the fix
>> sent by David into a proper patch and base this patch on that. Or wait
>> for David to send it himself and let him also make the modifications
>> needed in this patch.
>>
>> David ?
>>
>> Thanks,
>> Sekhar
>>
> 
> Alexandre, I was hoping that you would just squash my patch with your
> patch and take Sekhar's suggestion about a separate patch to make the
> private __clk_enable() public as davinci_clk_enable() when you re-submit.
OK.
> 
> You can add "Suggested-By: David Lechner <david@lechnology.com>" to the
> commit message if you would like to give me some credit for my ideas.
That was my concern. I will do that.

Thanks,
Alexandre
diff mbox

Patch

diff --git a/arch/arm/mach-davinci/usb-da8xx.c b/arch/arm/mach-davinci/usb-da8xx.c
index b010e5f..f62f9c2 100644
--- a/arch/arm/mach-davinci/usb-da8xx.c
+++ b/arch/arm/mach-davinci/usb-da8xx.c
@@ -22,6 +22,8 @@ 
 #define DA8XX_USB0_BASE		0x01e00000
 #define DA8XX_USB1_BASE		0x01e25000
 
+static struct clk *usb20_clk;
+
 static struct platform_device da8xx_usb_phy = {
 	.name		= "da8xx-usb-phy",
 	.id		= -1,
@@ -158,24 +160,16 @@  int __init da8xx_register_usb_refclkin(int rate)
 
 static void usb20_phy_clk_enable(struct clk *clk)
 {
-	struct clk *usb20_clk;
 	int err;
 	u32 val;
 	u32 timeout = 500000; /* 500 msec */
 
 	val = readl(DA8XX_SYSCFG0_VIRT(DA8XX_CFGCHIP2_REG));
 
-	usb20_clk = clk_get(&da8xx_usb20_dev.dev, "usb20");
-	if (IS_ERR(usb20_clk)) {
-		pr_err("could not get usb20 clk: %ld\n", PTR_ERR(usb20_clk));
-		return;
-	}
-
 	/* The USB 2.O PLL requires that the USB 2.O PSC is enabled as well. */
-	err = clk_prepare_enable(usb20_clk);
+	err = clk_enable(usb20_clk);
 	if (err) {
 		pr_err("failed to enable usb20 clk: %d\n", err);
-		clk_put(usb20_clk);
 		return;
 	}
 
@@ -197,8 +191,7 @@  static void usb20_phy_clk_enable(struct clk *clk)
 
 	pr_err("Timeout waiting for USB 2.0 PHY clock good\n");
 done:
-	clk_disable_unprepare(usb20_clk);
-	clk_put(usb20_clk);
+	clk_disable(usb20_clk);
 }
 
 static void usb20_phy_clk_disable(struct clk *clk)
@@ -285,11 +278,19 @@  static struct clk_lookup usb20_phy_clk_lookup =
 int __init da8xx_register_usb20_phy_clk(bool use_usb_refclkin)
 {
 	struct clk *parent;
-	int ret = 0;
+	int ret;
+
+	usb20_clk = clk_get(&da8xx_usb20_dev.dev, "usb20");
+	ret = PTR_ERR_OR_ZERO(usb20_clk);
+	if (ret)
+		return ret;
 
 	parent = clk_get(NULL, use_usb_refclkin ? "usb_refclkin" : "pll0_aux");
-	if (IS_ERR(parent))
-		return PTR_ERR(parent);
+	ret = PTR_ERR_OR_ZERO(parent);
+	if (ret) {
+		clk_put(usb20_clk);
+		return ret;
+	}
 
 	usb20_phy_clk.parent = parent;
 	ret = clk_register(&usb20_phy_clk);