[U-Boot] [RFC] mx28: Fix reboot from the kernel

[U-Boot] [RFC] mx28: Fix reboot from the kernel

[Fabio Estevam]

When running Linux on a mx28evk and issuing a 'reboot'command the following behaviour is observed:

root at freescale /$ reboot                                                        
starting pid 329, tty '': '/etc/rc.d/rcS stop'                                  
root at freescale /$ Stopping inetd:                                               
Terminated                                                                      
Unmounting filesystems                                                          
umount: tmpfs busy - remounted read-only                                        
chown: /home/user/.rhosts: Read-only file system                                
chown: /home/user: Read-only file system                                        
chown: /home/user: Read-only file system                                        
cat: can't open '/proc/devices': No such file or directory                      
The system is going down NOW!                                                   
Sent SIGTERM to all processes                                                   
Sent SIGKILL to all processes                                                   
Requesting system reboot                                                        
[   24.240000] Restarting system.                                               
                                                                          
U-Boot 2011.12-06887-g586ab2f-dirty (Jan 24 2012 - 12:58:50)                    
                                                                                
Freescale i.MX28 family                                                         
DRAM:  0

, and then U-boot hangs.

Calling mx28_power_init() prior to mxs_iomux_setup_multiple_pads() fixes the problem and
make U-boot to re-start succesfully.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
---
This patch fixes the reboot issue, but I am not able to provide a good explanation for it, so
I am marking it as RFC.

 arch/arm/cpu/arm926ejs/mx28/spl_boot.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/arch/arm/cpu/arm926ejs/mx28/spl_boot.c b/arch/arm/cpu/arm926ejs/mx28/spl_boot.c
index dfb8309..0e57252 100644
--- a/arch/arm/cpu/arm926ejs/mx28/spl_boot.c
+++ b/arch/arm/cpu/arm926ejs/mx28/spl_boot.c
@@ -49,8 +49,8 @@ void early_delay(int delay)
 void mx28_common_spl_init(const iomux_cfg_t *iomux_setup,
 			const unsigned int iomux_size)
 {
-	mxs_iomux_setup_multiple_pads(iomux_setup, iomux_size);
 	mx28_power_init();
+	mxs_iomux_setup_multiple_pads(iomux_setup, iomux_size);
 	mx28_mem_init();
 	mx28_power_wait_pswitch();
 }
-- 
1.7.1

[U-Boot] [RFC] mx28: Fix reboot from the kernel

[Marek Vasut]

> When running Linux on a mx28evk and issuing a 'reboot'command the following
> behaviour is observed:
> 
> root at freescale /$ reboot
> starting pid 329, tty '': '/etc/rc.d/rcS stop'
> root at freescale /$ Stopping inetd:
> Terminated
> Unmounting filesystems
> umount: tmpfs busy - remounted read-only
> chown: /home/user/.rhosts: Read-only file system
> chown: /home/user: Read-only file system
> chown: /home/user: Read-only file system
> cat: can't open '/proc/devices': No such file or directory
> The system is going down NOW!
> Sent SIGTERM to all processes
> Sent SIGKILL to all processes
> Requesting system reboot
> [   24.240000] Restarting system.
> 
> U-Boot 2011.12-06887-g586ab2f-dirty (Jan 24 2012 - 12:58:50)
> 
> Freescale i.MX28 family
> DRAM:  0
> 
> , and then U-boot hangs.
> 
> Calling mx28_power_init() prior to mxs_iomux_setup_multiple_pads() fixes
> the problem and make U-boot to re-start succesfully.
> 
> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
> ---
> This patch fixes the reboot issue, but I am not able to provide a good
> explanation for it, so I am marking it as RFC.
> 
>  arch/arm/cpu/arm926ejs/mx28/spl_boot.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/arch/arm/cpu/arm926ejs/mx28/spl_boot.c
> b/arch/arm/cpu/arm926ejs/mx28/spl_boot.c index dfb8309..0e57252 100644
> --- a/arch/arm/cpu/arm926ejs/mx28/spl_boot.c
> +++ b/arch/arm/cpu/arm926ejs/mx28/spl_boot.c
> @@ -49,8 +49,8 @@ void early_delay(int delay)
>  void mx28_common_spl_init(const iomux_cfg_t *iomux_setup,
>  			const unsigned int iomux_size)
>  {
> -	mxs_iomux_setup_multiple_pads(iomux_setup, iomux_size);
>  	mx28_power_init();
> +	mxs_iomux_setup_multiple_pads(iomux_setup, iomux_size);
>  	mx28_mem_init();
>  	mx28_power_wait_pswitch();
>  }

Can you explain why this patch fixes anything ? It's hard to understand why 
doing this would fix it.

[U-Boot] [RFC] mx28: Fix reboot from the kernel

[Fabio Estevam]

On Tue, Jan 24, 2012 at 9:48 PM, Marek Vasut <marek.vasut@gmail.com> wrote:

> Can you explain why this patch fixes anything ? It's hard to understand why
> doing this would fix it.

Yes, I agree and that's why I marked it as RFC and please also see my
comments below the --- line.

I don't understand the reason for the fix as well. Matthias' patch for
adding the boot mode information in mx28 magically fixed the reboot
issue and playing with his patch I realized that any instruction prior
to mxs_iomux_setup_multiple_pads "fixes" the reboot problem.

Regards,

Fabio Estevam

[U-Boot] [RFC] mx28: Fix reboot from the kernel

[Marek Vasut]

> On Tue, Jan 24, 2012 at 9:48 PM, Marek Vasut <marek.vasut@gmail.com> wrote:
> > Can you explain why this patch fixes anything ? It's hard to understand
> > why doing this would fix it.
> 
> Yes, I agree and that's why I marked it as RFC and please also see my
> comments below the --- line.
> 
> I don't understand the reason for the fix as well. Matthias' patch for
> adding the boot mode information in mx28 magically fixed the reboot
> issue and playing with his patch I realized that any instruction prior
> to mxs_iomux_setup_multiple_pads "fixes" the reboot problem.

Now that IS CREEPY. You mean like ... just anything ? Or does the code executed 
before need to be exactly long or longer than something ?

M
> 
> Regards,
> 
> Fabio Estevam

[U-Boot] [RFC] mx28: Fix reboot from the kernel

[Fabio Estevam]

On Tue, Jan 24, 2012 at 11:25 PM, Marek Vasut <marek.vasut@gmail.com> wrote:

> Now that IS CREEPY. You mean like ... just anything ? Or does the code executed
> before need to be exactly long or longer than something ?

Yes, my tests show that anything prior to
mxs_iomux_setup_multiple_pads makes reboot to work.

I added more details here: http://marc.info/?l=u-boot&m=132741813307150&w=2

That?s a very strange behaviour.

Regards,

Fabio Estevam

[U-Boot] [RFC] mx28: Fix reboot from the kernel

[Marek Vasut]

> On Tue, Jan 24, 2012 at 11:25 PM, Marek Vasut <marek.vasut@gmail.com> wrote:
> > Now that IS CREEPY. You mean like ... just anything ? Or does the code
> > executed before need to be exactly long or longer than something ?
> 
> Yes, my tests show that anything prior to
> mxs_iomux_setup_multiple_pads makes reboot to work.

Your @G jabber is botched, can you try pushing through on the other side? :-(
> 
> I added more details here: http://marc.info/?l=u-boot&m=132741813307150&w=2
> 
> That?s a very strange behaviour.

Let's hope UPS works on time as the weather forecast in Back to the Future II, 
if that's the case, I'd be able to test myself on thursday :-)

> 
> Regards,
> 
> Fabio Estevam

[U-Boot] [RFC] mx28: Fix reboot from the kernel

[Matthias Fuchs]

On 24.01.2012 17:23, Fabio Estevam wrote:
> When running Linux on a mx28evk and issuing a 'reboot'command the following behaviour is observed:
> 
> root at freescale /$ reboot                                                        
> starting pid 329, tty '': '/etc/rc.d/rcS stop'                                  
> root at freescale /$ Stopping inetd:                                               
> Terminated                                                                      
> Unmounting filesystems                                                          
> umount: tmpfs busy - remounted read-only                                        
> chown: /home/user/.rhosts: Read-only file system                                
> chown: /home/user: Read-only file system                                        
> chown: /home/user: Read-only file system                                        
> cat: can't open '/proc/devices': No such file or directory                      
> The system is going down NOW!                                                   
> Sent SIGTERM to all processes                                                   
> Sent SIGKILL to all processes                                                   
> Requesting system reboot                                                        
> [   24.240000] Restarting system.                                               
>                                                                           
> U-Boot 2011.12-06887-g586ab2f-dirty (Jan 24 2012 - 12:58:50)                    
>                                                                                 
> Freescale i.MX28 family                                                         
> DRAM:  0
> 
> , and then U-boot hangs.
> 
> Calling mx28_power_init() prior to mxs_iomux_setup_multiple_pads() fixes the problem and
> make U-boot to re-start succesfully.

Hmm, I did not observe that issue on my board.
> 
> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
> ---
> This patch fixes the reboot issue, but I am not able to provide a good explanation for it, so
> I am marking it as RFC.
> 
>  arch/arm/cpu/arm926ejs/mx28/spl_boot.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/arch/arm/cpu/arm926ejs/mx28/spl_boot.c b/arch/arm/cpu/arm926ejs/mx28/spl_boot.c
> index dfb8309..0e57252 100644
> --- a/arch/arm/cpu/arm926ejs/mx28/spl_boot.c
> +++ b/arch/arm/cpu/arm926ejs/mx28/spl_boot.c
> @@ -49,8 +49,8 @@ void early_delay(int delay)
>  void mx28_common_spl_init(const iomux_cfg_t *iomux_setup,
>  			const unsigned int iomux_size)
>  {
> -	mxs_iomux_setup_multiple_pads(iomux_setup, iomux_size);
>  	mx28_power_init();
> +	mxs_iomux_setup_multiple_pads(iomux_setup, iomux_size);
>  	mx28_mem_init();
>  	mx28_power_wait_pswitch();
>  }
Without an explanation I woutld not accept this patch.

So when adding "some" instructions before mx28_mem_init() you could
check for the first steps in that function are and if really delaying
the first instructions or same later stuff helps. Perhaps this brings us
a clue.

I will see if your reordering has any impact on my "USB recovery boot
issue".

Matthias

[U-Boot] [RFC] mx28: Fix reboot from the kernel

[Stefano Babic]

On 24/01/2012 17:23, Fabio Estevam wrote:
> When running Linux on a mx28evk and issuing a 'reboot'command the following behaviour is observed:
> 

Hi Fabio,

> 
> Calling mx28_power_init() prior to mxs_iomux_setup_multiple_pads() fixes the problem and
> make U-boot to re-start succesfully.

This does not explain why it hangs, and why it does not hangs on
M28....apparently, checking both functions there is no evident reason
why they should be called in a certain order. I think we have to dig
more deeply inside this strange behavior.

Best regards,
SWtefano Babic

-- 
=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office at denx.de
=====================================================================