n900 in 4.2-rc0: repeating oopses
diff mbox

Message ID 20150630092404.GA19285@amd
State New
Headers show

Commit Message

Pavel Machek June 30, 2015, 9:24 a.m. UTC
Hi!

> Just tried booting 4.2-rc0 on n900 (commit
> 4a10a91756ef381bced7b88cfb9232f660b92d93) and it is broken. Previous
> -rc0 version worked. This time, there's some output on console, but
> too fast for me to read.
> 
> It seems oopses happen before mounting root. If you have serial
> console, they should be easy to see.

I tried again according to pali's instructions, and it is still
broken the same way. Any other ideas? Does it work for you?


									Pavel

Comments

Pali Rohár June 30, 2015, 9:53 a.m. UTC | #1
On Tuesday 30 June 2015 11:24:04 Pavel Machek wrote:
> Hi!
> 
> > Just tried booting 4.2-rc0 on n900 (commit
> > 4a10a91756ef381bced7b88cfb9232f660b92d93) and it is broken. Previous
> > -rc0 version worked. This time, there's some output on console, but
> > too fast for me to read.
> > 
> > It seems oopses happen before mounting root. If you have serial
> > console, they should be easy to see.
> 
> I tried again according to pali's instructions, and it is still
> broken the same way. Any other ideas? Does it work for you?
> 
> diff --git a/arch/arm/boot/dts/omap3.dtsi b/arch/arm/boot/dts/omap3.dtsi
> index 69a40cf..ff6f2bf 100644
> --- a/arch/arm/boot/dts/omap3.dtsi
> +++ b/arch/arm/boot/dts/omap3.dtsi
> @@ -570,6 +570,7 @@
>  			interrupts = <49>;
>  			dmas = <&sdma 69>;
>  			dma-names = "rx";
> +			status = "disabled";
>  		};
>  
>  		smartreflex_core: smartreflex@480cb000 {
> 
> 									Pavel
> 

Ok, then it is not problem with omap crypto and HS device...

I will try 4.2 at the end of week.
Tony Lindgren July 1, 2015, 6:59 a.m. UTC | #2
* Pali Rohár <pali.rohar@gmail.com> [150630 02:55]:
> 
> I will try 4.2 at the end of week.

At least today's 4.1.0-11549-g05a8256 boots just fine on my n900.

Regards,

Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Pali Rohár July 1, 2015, 7:22 a.m. UTC | #3
On Tuesday 30 June 2015 23:59:33 Tony Lindgren wrote:
> * Pali Rohár <pali.rohar@gmail.com> [150630 02:55]:
> > 
> > I will try 4.2 at the end of week.
> 
> At least today's 4.1.0-11549-g05a8256 boots just fine on my n900.
> 
> Regards,
> 
> Tony

So, Pavel can you re-test? Maybe there can be problem with some driver
which Tony did not compiled into zImage? Just speculation...
Pavel Machek July 1, 2015, 9:59 a.m. UTC | #4
On Wed 2015-07-01 09:22:55, Pali Rohár wrote:
> On Tuesday 30 June 2015 23:59:33 Tony Lindgren wrote:
> > * Pali Rohár <pali.rohar@gmail.com> [150630 02:55]:
> > > 
> > > I will try 4.2 at the end of week.
> > 
> > At least today's 4.1.0-11549-g05a8256 boots just fine on my n900.
> > 
> > Regards,
> > 
> > Tony
> 
> So, Pavel can you re-test? Maybe there can be problem with some driver
> which Tony did not compiled into zImage? Just speculation...

I re-tested with today's git, and it seems to boot. Thanks for help...

Now. "echo mem > /sys/power/state" is broken, as in "returns
immediately in about 50% cases". The messages are

Powerdomain (per_pwrdm) didn't enter target state 1
Powerdomain (core_pwerdm) didn't enter target state 1

Any ideas? Thanks,

									Pavel
Tony Lindgren July 1, 2015, 10:34 a.m. UTC | #5
* Pavel Machek <pavel@ucw.cz> [150701 03:02]:
> On Wed 2015-07-01 09:22:55, Pali Rohár wrote:
> > On Tuesday 30 June 2015 23:59:33 Tony Lindgren wrote:
> > > * Pali Rohár <pali.rohar@gmail.com> [150630 02:55]:
> > > > 
> > > > I will try 4.2 at the end of week.
> > > 
> > > At least today's 4.1.0-11549-g05a8256 boots just fine on my n900.
> > > 
> > > Regards,
> > > 
> > > Tony
> > 
> > So, Pavel can you re-test? Maybe there can be problem with some driver
> > which Tony did not compiled into zImage? Just speculation...
> 
> I re-tested with today's git, and it seems to boot. Thanks for help...

OK good to hear.
 
> Now. "echo mem > /sys/power/state" is broken, as in "returns
> immediately in about 50% cases". The messages are
> 
> Powerdomain (per_pwrdm) didn't enter target state 1
> Powerdomain (core_pwerdm) didn't enter target state 1
> 
> Any ideas? Thanks,

Works for me after enabling the idle timeouts with the following
script and blanking the screen and disconnecting USB:

Also both keyboard LEDs should start blinking after the idle
timeout with screen blanked and USB disconnected. If not, you
have some module loaded that blocks the deeper idle states.

Regrads,

Tony

8< ----
#!/bin/bash

modprobe leds-gpio
modprobe ledtrig-default-on

uarts=$(find /sys/class/tty/ttyO*/device/power/ -type d)
for uart in $uarts; do
        echo 3000 > $uart/autosuspend_delay_ms
        #echo -1 > $uart/autosuspend_delay_ms
done

uarts=$(find /sys/class/tty/ttyO*/power/ -type d)
for uart in $uarts; do
        echo enabled > $uart/wakeup
        echo auto > $uart/control
done

echo 255 > /sys/class/backlight/acx565akm/brightness
echo 1 > /sys/devices/platform/omapfb/graphics/fb0/blank

echo 1 > /sys/kernel/debug/pm_debug/enable_off_mode
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Pavel Machek July 1, 2015, 1:09 p.m. UTC | #6
On Wed 2015-07-01 03:34:22, Tony Lindgren wrote:
> * Pavel Machek <pavel@ucw.cz> [150701 03:02]:
> > On Wed 2015-07-01 09:22:55, Pali Rohár wrote:
> > > On Tuesday 30 June 2015 23:59:33 Tony Lindgren wrote:
> > > > * Pali Rohár <pali.rohar@gmail.com> [150630 02:55]:
> > > > > 
> > > > > I will try 4.2 at the end of week.
> > > > 
> > > > At least today's 4.1.0-11549-g05a8256 boots just fine on my n900.
> > > > 
> > > > Regards,
> > > > 
> > > > Tony
> > > 
> > > So, Pavel can you re-test? Maybe there can be problem with some driver
> > > which Tony did not compiled into zImage? Just speculation...
> > 
> > I re-tested with today's git, and it seems to boot. Thanks for help...
> 
> OK good to hear.
>  
> > Now. "echo mem > /sys/power/state" is broken, as in "returns
> > immediately in about 50% cases". The messages are
> > 
> > Powerdomain (per_pwrdm) didn't enter target state 1
> > Powerdomain (core_pwerdm) didn't enter target state 1
> > 
> > Any ideas? Thanks,
> 
> Works for me after enabling the idle timeouts with the following
> script and blanking the screen and disconnecting USB:

Um. I'm forcing suspend with "echo mem > /sys/power/state" . (It
worked in 4.1). That should just make it sleep, no autosuspend-related
trickery... (But yes, I guess I should set up the leds and try
autosuspend, too.)

Regards,
									Pavel
Tony Lindgren July 2, 2015, 6:16 a.m. UTC | #7
* Pavel Machek <pavel@ucw.cz> [150701 06:11]:
> On Wed 2015-07-01 03:34:22, Tony Lindgren wrote:
> > 
> > Works for me after enabling the idle timeouts with the following
> > script and blanking the screen and disconnecting USB:
> 
> Um. I'm forcing suspend with "echo mem > /sys/power/state" . (It
> worked in 4.1). That should just make it sleep, no autosuspend-related
> trickery... (But yes, I guess I should set up the leds and try
> autosuspend, too.)

That too works just fine for me with omap2plus_defconfig after
echo enabled > /sys/class/tty/ttyO2/power/wakeup as I have n900
in my test rack.

You probably have some additional drivers enabled?

Regards,

Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Pavel Machek July 2, 2015, 8:25 p.m. UTC | #8
On Wed 2015-07-01 03:34:22, Tony Lindgren wrote:
> * Pavel Machek <pavel@ucw.cz> [150701 03:02]:
> > On Wed 2015-07-01 09:22:55, Pali Rohár wrote:
> > > On Tuesday 30 June 2015 23:59:33 Tony Lindgren wrote:
> > > > * Pali Rohár <pali.rohar@gmail.com> [150630 02:55]:
> > > > > 
> > > > > I will try 4.2 at the end of week.
> > > > 
> > > > At least today's 4.1.0-11549-g05a8256 boots just fine on my n900.
> > > > 
> > > > Regards,
> > > > 
> > > > Tony
> > > 
> > > So, Pavel can you re-test? Maybe there can be problem with some driver
> > > which Tony did not compiled into zImage? Just speculation...
> > 
> > I re-tested with today's git, and it seems to boot. Thanks for help...
> 
> OK good to hear.

Hmm. Tried to toggle brightness on 4.1, and after few changes:

root@n900:~# echo 255 > /sys/class/backlight/acx565akm/brightness
root@n900:~# echo 120 > /sys/class/backlight/acx565akm/brightness
root@n900:~# echo 10 > /sys/class/backlight/acx565akm/brightness
root@n900:~# echo 1 > /sys/class/backlight/acx565akm/brightness



^C^C
^Z

seems like acx565akm hung :-(.
									Pavel
Pavel Machek July 26, 2015, 9:26 a.m. UTC | #9
Hi!

> > > So, Pavel can you re-test? Maybe there can be problem with some driver
> > > which Tony did not compiled into zImage? Just speculation...
> > 
> > I re-tested with today's git, and it seems to boot. Thanks for help...
> 
> OK good to hear.
>  
> > Now. "echo mem > /sys/power/state" is broken, as in "returns
> > immediately in about 50% cases". The messages are
> > 
> > Powerdomain (per_pwrdm) didn't enter target state 1
> > Powerdomain (core_pwerdm) didn't enter target state 1
> > 
> > Any ideas? Thanks,
> 
> Works for me after enabling the idle timeouts with the following
> script and blanking the screen and disconnecting USB:
> 
> Also both keyboard LEDs should start blinking after the idle
> timeout with screen blanked and USB disconnected. If not, you
> have some module loaded that blocks the deeper idle states.

Ok, tried that, I had to do:

 cd /sys/class/gpio
 echo 162 > export
 cd gpio162
 echo out > direction
 echo 1 > value

to get the debug lights to work. But I could not get those leds to
blink.

> modprobe leds-gpio
> modprobe ledtrig-default-on

Is this actually neccessary/relevant?

> echo 255 > /sys/class/backlight/acx565akm/brightness

And this?

Thanks,
									Pavel
Pavel Machek July 26, 2015, 10:31 p.m. UTC | #10
On Wed 2015-07-01 23:16:21, Tony Lindgren wrote:
> * Pavel Machek <pavel@ucw.cz> [150701 06:11]:
> > On Wed 2015-07-01 03:34:22, Tony Lindgren wrote:
> > > 
> > > Works for me after enabling the idle timeouts with the following
> > > script and blanking the screen and disconnecting USB:
> > 
> > Um. I'm forcing suspend with "echo mem > /sys/power/state" . (It
> > worked in 4.1). That should just make it sleep, no autosuspend-related
> > trickery... (But yes, I guess I should set up the leds and try
> > autosuspend, too.)
> 
> That too works just fine for me with omap2plus_defconfig after
> echo enabled > /sys/class/tty/ttyO2/power/wakeup as I have n900
> in my test rack.

Going through that. AFAICT, you are not using devicetree on n900?

CONFIG_MACH_NOKIA_RX51=y

Unfortunately, not having serials, it is tricky to get any output from
that cnofiguration, so I don't know what I got wrong...
									Pavel
Pali Rohár July 27, 2015, 7:12 a.m. UTC | #11
On Monday 27 July 2015 00:31:19 Pavel Machek wrote:
> On Wed 2015-07-01 23:16:21, Tony Lindgren wrote:
> > * Pavel Machek <pavel@ucw.cz> [150701 06:11]:
> > > On Wed 2015-07-01 03:34:22, Tony Lindgren wrote:
> > > > 
> > > > Works for me after enabling the idle timeouts with the following
> > > > script and blanking the screen and disconnecting USB:
> > > 
> > > Um. I'm forcing suspend with "echo mem > /sys/power/state" . (It
> > > worked in 4.1). That should just make it sleep, no autosuspend-related
> > > trickery... (But yes, I guess I should set up the leds and try
> > > autosuspend, too.)
> > 
> > That too works just fine for me with omap2plus_defconfig after
> > echo enabled > /sys/class/tty/ttyO2/power/wakeup as I have n900
> > in my test rack.
> 
> Going through that. AFAICT, you are not using devicetree on n900?
> 
> CONFIG_MACH_NOKIA_RX51=y
> 
> Unfortunately, not having serials, it is tricky to get any output from
> that cnofiguration, so I don't know what I got wrong...
> 									Pavel

You can try to test it in qemu. It has emulation also of serial.
Tony Lindgren July 28, 2015, 6:01 a.m. UTC | #12
* Pavel Machek <pavel@ucw.cz> [150726 02:29]:
> Hi!
> 
> > > > So, Pavel can you re-test? Maybe there can be problem with some driver
> > > > which Tony did not compiled into zImage? Just speculation...
> > > 
> > > I re-tested with today's git, and it seems to boot. Thanks for help...
> > 
> > OK good to hear.
> >  
> > > Now. "echo mem > /sys/power/state" is broken, as in "returns
> > > immediately in about 50% cases". The messages are
> > > 
> > > Powerdomain (per_pwrdm) didn't enter target state 1
> > > Powerdomain (core_pwerdm) didn't enter target state 1
> > > 
> > > Any ideas? Thanks,
> > 
> > Works for me after enabling the idle timeouts with the following
> > script and blanking the screen and disconnecting USB:
> > 
> > Also both keyboard LEDs should start blinking after the idle
> > timeout with screen blanked and USB disconnected. If not, you
> > have some module loaded that blocks the deeper idle states.
> 
> Ok, tried that, I had to do:
> 
>  cd /sys/class/gpio
>  echo 162 > export
>  cd gpio162
>  echo out > direction
>  echo 1 > value
> 
> to get the debug lights to work. But I could not get those leds to
> blink.

I think that's the GPIO override for them? And in that case the
PMIC sys_clkreq and sys_off_mode signals are not cutting off the
LEDs I believe.

Also not that any driver blocking the deeper idle states will
also cause the LEDs to stay on. And USB cable being plugged in
too at least for core off.. So you if you're using NFSroot over
USB Ethernet, you probably can only see the left LED ever go off.
 
> > modprobe leds-gpio
> > modprobe ledtrig-default-on
> 
> Is this actually neccessary/relevant?

Yes at least with omap2plus_defconfig.
 
> > echo 255 > /sys/class/backlight/acx565akm/brightness
> 
> And this?

It used to be the backlight did not come on initially as modules
or something.. May not be needed any longer.

Regards,

Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch
diff mbox

diff --git a/arch/arm/boot/dts/omap3.dtsi b/arch/arm/boot/dts/omap3.dtsi
index 69a40cf..ff6f2bf 100644
--- a/arch/arm/boot/dts/omap3.dtsi
+++ b/arch/arm/boot/dts/omap3.dtsi
@@ -570,6 +570,7 @@ 
 			interrupts = <49>;
 			dmas = <&sdma 69>;
 			dma-names = "rx";
+			status = "disabled";
 		};
 
 		smartreflex_core: smartreflex@480cb000 {