Revert "ARM: bcm283x: Switch V3D over to using the PM driver instead of firmware."

Submitted by Stefan Wahren on Sept. 8, 2019, 3:44 p.m.

Details

Message ID 1567957493-4567-1-git-send-email-wahrenst@gmx.net
State New
Headers show
Series "Revert "ARM: bcm283x: Switch V3D over to using the PM driver instead of firmware."" ( rev: 1 ) in DRI devel

Not browsing as part of any series.

Commit Message

Stefan Wahren Sept. 8, 2019, 3:44 p.m.
Since release of the new BCM2835 PM driver there has been several reports
of V3D probing issues. This is caused by timeouts during powering-up the
GRAFX PM domain:

  bcm2835-power: Timeout waiting for grafx power OK

I was able to reproduce this reliable on my Raspberry Pi 3B+ after setting
force_turbo=1 in the firmware configuration. Since there are no issues
using the firmware PM driver with the same setup, there must be an issue
in the BCM2835 PM driver.

Unfortunately there hasn't been much progress in identifying the root cause
since June (mostly in the lack of documentation), so i decided to switch
back until the issue in the BCM2835 PM driver is fixed.

Link: https://github.com/raspberrypi/linux/issues/3046
Fixes: e1dc2b2e1bef (" ARM: bcm283x: Switch V3D over to using the PM driver instead of firmware.")
Cc: stable@vger.kernel.org
Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
---
 arch/arm/boot/dts/bcm2835-rpi.dtsi | 4 ++++
 arch/arm/boot/dts/bcm283x.dtsi     | 4 +---
 2 files changed, 5 insertions(+), 3 deletions(-)

--
2.7.4

Patch hide | download patch | download mbox

diff --git a/arch/arm/boot/dts/bcm2835-rpi.dtsi b/arch/arm/boot/dts/bcm2835-rpi.dtsi
index 6c6a7f6..b909e3b 100644
--- a/arch/arm/boot/dts/bcm2835-rpi.dtsi
+++ b/arch/arm/boot/dts/bcm2835-rpi.dtsi
@@ -67,6 +67,10 @@ 
 	power-domains = <&power RPI_POWER_DOMAIN_USB>;
 };

+&v3d {
+	power-domains = <&power RPI_POWER_DOMAIN_V3D>;
+};
+
 &vec {
 	power-domains = <&power RPI_POWER_DOMAIN_VEC>;
 	status = "okay";
diff --git a/arch/arm/boot/dts/bcm283x.dtsi b/arch/arm/boot/dts/bcm283x.dtsi
index 2d191fc..b238567 100644
--- a/arch/arm/boot/dts/bcm283x.dtsi
+++ b/arch/arm/boot/dts/bcm283x.dtsi
@@ -3,7 +3,6 @@ 
 #include <dt-bindings/clock/bcm2835-aux.h>
 #include <dt-bindings/gpio/gpio.h>
 #include <dt-bindings/interrupt-controller/irq.h>
-#include <dt-bindings/soc/bcm2835-pm.h>

 /* firmware-provided startup stubs live here, where the secondary CPUs are
  * spinning.
@@ -121,7 +120,7 @@ 
 			#interrupt-cells = <2>;
 		};

-		pm: watchdog@7e100000 {
+		watchdog@7e100000 {
 			compatible = "brcm,bcm2835-pm", "brcm,bcm2835-pm-wdt";
 			#power-domain-cells = <1>;
 			#reset-cells = <1>;
@@ -641,7 +640,6 @@ 
 			compatible = "brcm,bcm2835-v3d";
 			reg = <0x7ec00000 0x1000>;
 			interrupts = <1 10>;
-			power-domains = <&pm BCM2835_POWER_DOMAIN_GRAFX_V3D>;
 		};

 		vc4: gpu {

Comments


On 9/8/19 8:44 AM, Stefan Wahren wrote:
> Since release of the new BCM2835 PM driver there has been several reports
> of V3D probing issues. This is caused by timeouts during powering-up the
> GRAFX PM domain:
> 
>   bcm2835-power: Timeout waiting for grafx power OK
> 
> I was able to reproduce this reliable on my Raspberry Pi 3B+ after setting
> force_turbo=1 in the firmware configuration. Since there are no issues
> using the firmware PM driver with the same setup, there must be an issue
> in the BCM2835 PM driver.
> 
> Unfortunately there hasn't been much progress in identifying the root cause
> since June (mostly in the lack of documentation), so i decided to switch
> back until the issue in the BCM2835 PM driver is fixed.
> 
> Link: https://github.com/raspberrypi/linux/issues/3046
> Fixes: e1dc2b2e1bef (" ARM: bcm283x: Switch V3D over to using the PM driver instead of firmware.")
> Cc: stable@vger.kernel.org
> Signed-off-by: Stefan Wahren <wahrenst@gmx.net>

Do you want me to pick up this change directly, or would you want to
issue a pull request for 5.4-rcX with other fixes?
Hi Florian,

Am 20.09.19 um 19:52 schrieb Florian Fainelli:
> On 9/8/19 8:44 AM, Stefan Wahren wrote:
>> Since release of the new BCM2835 PM driver there has been several reports
>> of V3D probing issues. This is caused by timeouts during powering-up the
>> GRAFX PM domain:
>>
>>   bcm2835-power: Timeout waiting for grafx power OK
>>
>> I was able to reproduce this reliable on my Raspberry Pi 3B+ after setting
>> force_turbo=1 in the firmware configuration. Since there are no issues
>> using the firmware PM driver with the same setup, there must be an issue
>> in the BCM2835 PM driver.
>>
>> Unfortunately there hasn't been much progress in identifying the root cause
>> since June (mostly in the lack of documentation), so i decided to switch
>> back until the issue in the BCM2835 PM driver is fixed.
>>
>> Link: https://github.com/raspberrypi/linux/issues/3046
>> Fixes: e1dc2b2e1bef (" ARM: bcm283x: Switch V3D over to using the PM driver instead of firmware.")
>> Cc: stable@vger.kernel.org
>> Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
> Do you want me to pick up this change directly, or would you want to
> issue a pull request for 5.4-rcX with other fixes?

there aren't any other fixes, please pick up this directly.

Thanks

Stefan
On 9/8/2019 8:44 AM, Stefan Wahren wrote:
> Since release of the new BCM2835 PM driver there has been several reports
> of V3D probing issues. This is caused by timeouts during powering-up the
> GRAFX PM domain:
> 
>   bcm2835-power: Timeout waiting for grafx power OK
> 
> I was able to reproduce this reliable on my Raspberry Pi 3B+ after setting
> force_turbo=1 in the firmware configuration. Since there are no issues
> using the firmware PM driver with the same setup, there must be an issue
> in the BCM2835 PM driver.
> 
> Unfortunately there hasn't been much progress in identifying the root cause
> since June (mostly in the lack of documentation), so i decided to switch
> back until the issue in the BCM2835 PM driver is fixed.
> 
> Link: https://github.com/raspberrypi/linux/issues/3046
> Fixes: e1dc2b2e1bef (" ARM: bcm283x: Switch V3D over to using the PM driver instead of firmware.")
> Cc: stable@vger.kernel.org
> Signed-off-by: Stefan Wahren <wahrenst@gmx.net>

Applied to devicetree/fixes, thanks!
Hi Florian,

Am 23.09.19 um 20:56 schrieb Florian Fainelli:
>
> On 9/8/2019 8:44 AM, Stefan Wahren wrote:
>> Since release of the new BCM2835 PM driver there has been several reports
>> of V3D probing issues. This is caused by timeouts during powering-up the
>> GRAFX PM domain:
>>
>>   bcm2835-power: Timeout waiting for grafx power OK
>>
>> I was able to reproduce this reliable on my Raspberry Pi 3B+ after setting
>> force_turbo=1 in the firmware configuration. Since there are no issues
>> using the firmware PM driver with the same setup, there must be an issue
>> in the BCM2835 PM driver.
>>
>> Unfortunately there hasn't been much progress in identifying the root cause
>> since June (mostly in the lack of documentation), so i decided to switch
>> back until the issue in the BCM2835 PM driver is fixed.
>>
>> Link: https://github.com/raspberrypi/linux/issues/3046
>> Fixes: e1dc2b2e1bef (" ARM: bcm283x: Switch V3D over to using the PM driver instead of firmware.")
>> Cc: stable@vger.kernel.org
>> Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
> Applied to devicetree/fixes, thanks!

i noticed that X hangs with recent Raspbian and Linux 5.4-rc2 after this
revert has been applied.

Is there a chance to drop the revert?

Stefan
On 10/9/19 2:33 PM, Stefan Wahren wrote:
> Hi Florian,
> 
> Am 23.09.19 um 20:56 schrieb Florian Fainelli:
>>
>> On 9/8/2019 8:44 AM, Stefan Wahren wrote:
>>> Since release of the new BCM2835 PM driver there has been several reports
>>> of V3D probing issues. This is caused by timeouts during powering-up the
>>> GRAFX PM domain:
>>>
>>>   bcm2835-power: Timeout waiting for grafx power OK
>>>
>>> I was able to reproduce this reliable on my Raspberry Pi 3B+ after setting
>>> force_turbo=1 in the firmware configuration. Since there are no issues
>>> using the firmware PM driver with the same setup, there must be an issue
>>> in the BCM2835 PM driver.
>>>
>>> Unfortunately there hasn't been much progress in identifying the root cause
>>> since June (mostly in the lack of documentation), so i decided to switch
>>> back until the issue in the BCM2835 PM driver is fixed.
>>>
>>> Link: https://github.com/raspberrypi/linux/issues/3046
>>> Fixes: e1dc2b2e1bef (" ARM: bcm283x: Switch V3D over to using the PM driver instead of firmware.")
>>> Cc: stable@vger.kernel.org
>>> Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
>> Applied to devicetree/fixes, thanks!
> 
> i noticed that X hangs with recent Raspbian and Linux 5.4-rc2 after this
> revert has been applied.
> 
> Is there a chance to drop the revert?

I have not sent anything yet as I wanted to get some testing done on
other platforms, so yes, I can definitively drop that particular changes
from devicetree/fixes and not send it for a 5.4-rc3 pull request. Does
that work for you?
On 09.10.19 23:42, Florian Fainelli wrote:
> On 10/9/19 2:33 PM, Stefan Wahren wrote:
>> Hi Florian,
>>
>> Am 23.09.19 um 20:56 schrieb Florian Fainelli:
>>> On 9/8/2019 8:44 AM, Stefan Wahren wrote:
>>>> Since release of the new BCM2835 PM driver there has been several reports
>>>> of V3D probing issues. This is caused by timeouts during powering-up the
>>>> GRAFX PM domain:
>>>>
>>>>   bcm2835-power: Timeout waiting for grafx power OK
>>>>
>>>> I was able to reproduce this reliable on my Raspberry Pi 3B+ after setting
>>>> force_turbo=1 in the firmware configuration. Since there are no issues
>>>> using the firmware PM driver with the same setup, there must be an issue
>>>> in the BCM2835 PM driver.
>>>>
>>>> Unfortunately there hasn't been much progress in identifying the root cause
>>>> since June (mostly in the lack of documentation), so i decided to switch
>>>> back until the issue in the BCM2835 PM driver is fixed.
>>>>
>>>> Link: https://github.com/raspberrypi/linux/issues/3046
>>>> Fixes: e1dc2b2e1bef (" ARM: bcm283x: Switch V3D over to using the PM driver instead of firmware.")
>>>> Cc: stable@vger.kernel.org
>>>> Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
>>> Applied to devicetree/fixes, thanks!
>> i noticed that X hangs with recent Raspbian and Linux 5.4-rc2 after this
>> revert has been applied.
>>
>> Is there a chance to drop the revert?
> I have not sent anything yet as I wanted to get some testing done on
> other platforms, so yes, I can definitively drop that particular changes
> from devicetree/fixes and not send it for a 5.4-rc3 pull request. Does
> that work for you?

Yes, definitely

Thanks Stefan