NOTE: hwmod changes merged to linux-omap master for testing

DISCUSSION Tony Lindgren 5 participants 19 messages Oct 2, 2010 – Oct 7, 2010

Share & tools

Tony Lindgren

author

2010-10-02 00:48 UTC #0 lore raw

Hi all,

So I've finally merged Kevin's hwmod branch into the linux-omap master so we can give it some testing to see if we can merge it to the mainline. Might take few minutes to hit the mirrors.

Please note that for omap2plus machines you need to change:

- u-boot bootargs for kernel cmdline needs to change with setenv
- if you use hardcoded CONFIG_CMDLINE, change it in your .config
- /etc/inittab in your omap root file system if using serial console

This is because otherwise the serial-omap would be conflicting with the 8250 driver. This should not affect omap1 machines.

Please give it a try and report any issues on this list! Now all the machines except ams_delta_defconfig should build. That one is still missing some camera driver changes to build.

Regards,

Tony

G, Manjunath Kondaiah

2010-10-04 07:49 UTC #1 lore raw

Hi Tony,
With latest merges on master branch with commit ID:
0882b1455797b0a104978000a622c3f2412e7cf5 omap: READ CAREFULLY: Fix build after merging in hwmod support for testing

Booting with zoom3 seems to be fine but observed issues with omap4 blaze board.

Looks like multiple warning dumps and finally it crashes after "Freeing init memory" with both NFS and ramdisk file systems.

The complete bootlogs are available at:
NFS bootlog:
http://pastebin.com/nrWQ5FGG

Ramdisk bootlog:
http://pastebin.com/1F75d2Ew

-Manjunath

Quoted 37 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Tony Lindgren
Sent: Saturday, October 02, 2010 6:18 AM
To: [email protected]
Subject: NOTE: hwmod changes merged to linux-omap master for testing
Hi all,
So I've finally merged Kevin's hwmod branch into the linux-omap master so we can give it some testing to see if we can merge it to the mainline. Might take few minutes to hit the mirrors.
Please note that for omap2plus machines you need to change:
- u-boot bootargs for kernel cmdline needs to change with setenv
- if you use hardcoded CONFIG_CMDLINE, change it in your .config
- /etc/inittab in your omap root file system if using serial console
This is because otherwise the serial-omap would be conflicting with the 8250 driver. This should not affect omap1 machines.
Please give it a try and report any issues on this list! Now all the machines except ams_delta_defconfig should build. That one is still missing some camera driver changes to build.
Regards,
Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html

Govindraj

2010-10-04 09:20 UTC #2 lore raw

Quoted 7 lines · Manjunath Kondaiah

Hi Tony,
With latest merges on master branch with commit ID:
0882b1455797b0a104978000a622c3f2412e7cf5 omap: READ CAREFULLY: Fix build after merging in hwmod support for testing
Booting with zoom3 seems to be fine but observed issues with omap4 blaze board.
Looks like multiple warning dumps and finally it crashes after "Freeing init memory" with both NFS and ramdisk file systems.

This issue though seems to exist much before hwmod merge, I tried below commit on master

commit 2365f1f7e5544b531ccd3e07fd06bb12bf7a62a7
Merge: acc012f 391bcdd
Author: Tony Lindgren <[email protected]>
Date: Wed Sep 29 11:19:10 2010 -0700

Issue still reproducible.
I used 4430SDP with es2.0

Not sure what is happening though.

http://pastebin.com/LZKaAztT

---
Regards,
Govindraj.R

Quoted 47 lines

The complete bootlogs are available at:
NFS bootlog:
http://pastebin.com/nrWQ5FGG
Ramdisk bootlog:
http://pastebin.com/1F75d2Ew
-Manjunath
Quoted 30 lines
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Tony Lindgren
Sent: Saturday, October 02, 2010 6:18 AM
To: [email protected]
Subject: NOTE: hwmod changes merged to linux-omap master for testing
Hi all,
So I've finally merged Kevin's hwmod branch into the linux-omap master so we can give it some testing to see if we can merge it to the mainline. Might take few minutes to hit the mirrors.
Please note that for omap2plus machines you need to change:
- u-boot bootargs for kernel cmdline needs to change with setenv
- if you use hardcoded CONFIG_CMDLINE, change it in your .config
- /etc/inittab in your omap root file system if using serial console
This is because otherwise the serial-omap would be conflicting with the 8250 driver. This should not affect omap1 machines.
Please give it a try and report any issues on this list! Now all the machines except ams_delta_defconfig should build. That one is still missing some camera driver changes to build.
Regards,
Quoted 7 lines
Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
--

Quoted 4 lines

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

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

G, Manjunath Kondaiah

2010-10-04 11:34 UTC #3 lore raw

Quoted 17 lines

-----Original Message-----
From: Govindraj [mailto:[email protected]]
Sent: Monday, October 04, 2010 2:50 PM
To: G, Manjunath Kondaiah
Cc: Tony Lindgren; [email protected]
Subject: Re: NOTE: hwmod changes merged to linux-omap master for testing
Quoted 5 lines · Manjunath Kondaiah
Hi Tony,
With latest merges on master branch with commit ID:
0882b1455797b0a104978000a622c3f2412e7cf5 omap: READ CAREFULLY: Fix build after merging in hwmod support for testing

Quoted 8 lines

Quoted 1 line
Booting with zoom3 seems to be fine but observed issues
with omap4 blaze board.
Quoted 2 lines
Looks like multiple warning dumps and finally it crashes
after "Freeing init memory" with both NFS and ramdisk file systems.
This issue though seems to exist much before hwmod merge, I tried below commit on master

Mainline line kernel(2.6.36-rc6) merged only with 'omap_for_2.6.37' of git://dev.omapzoom.org/pub/scm/santosh/kernel-omap4-base
will work fine on omap4.

-Manjunath

Ohad Ben-Cohen

2010-10-04 12:09 UTC #4 lore raw

In reply to G, Manjunath Kondaiah

Quoted 1 line · Manjunath Kondaiah

Booting with zoom3 seems to be fine but observed issues with omap4 blaze board.

FWIW, for me it worked just fine with blaze

G, Manjunath Kondaiah

2010-10-04 12:13 UTC #5 lore raw

Quoted 11 lines

-----Original Message-----
From: Ohad Ben-Cohen [mailto:[email protected]]
Sent: Monday, October 04, 2010 5:39 PM
To: G, Manjunath Kondaiah
Cc: Tony Lindgren; [email protected]
Subject: Re: NOTE: hwmod changes merged to linux-omap master for testing

Quoted 6 lines

Quoted 1 line · Manjunath Kondaiah
Booting with zoom3 seems to be fine but observed issues
with omap4 blaze board.
FWIW, for me it worked just fine with blaze

ES2.0 or ES1.0?

-Manjunath

Ohad Ben-Cohen

2010-10-04 12:14 UTC #6 lore raw

Quoted 1 line · Manjunath Kondaiah

ES2.0 or ES1.0?

2.0

Quoted 2 lines

-Manjunath

G, Manjunath Kondaiah

2010-10-04 12:20 UTC #7 lore raw

Quoted 10 lines

-----Original Message-----
From: Ohad Ben-Cohen [mailto:[email protected]]
Sent: Monday, October 04, 2010 5:44 PM
To: G, Manjunath Kondaiah
Cc: Tony Lindgren; [email protected]
Subject: Re: NOTE: hwmod changes merged to linux-omap master for testing

Quoted 5 lines

Quoted 1 line · Manjunath Kondaiah
ES2.0 or ES1.0?
2.0

Can you post complete boot log? so far, it failed on one blaze and one 3430sdp both of them are having es2.0

-Manjunath

Ohad Ben-Cohen

2010-10-04 12:33 UTC #8 lore raw

Quoted 2 lines · Manjunath Kondaiah

Can you post complete boot log? so far, it failed on one blaze and one 3430sdp both of them are having es2.0

Sure:

wizery.com/ohad/blaze-linux-omap.txt.gz

Summary of my setup:
- Blaze ES2.0
- linux-omap-2.6 master branch (commit 0882b1455797b0a104978000a622c3f2412e7cf5 "omap: READ CAREFULLY: Fix build after merging in hwmod support for testing")
- omap2plus_defconfig
- kernel boots (with 2-3 warnings)

G, Manjunath Kondaiah

2010-10-04 12:53 UTC #9 lore raw

Quoted 12 lines

-----Original Message-----
From: Ohad Ben-Cohen [mailto:[email protected]]
Sent: Monday, October 04, 2010 6:03 PM
To: G, Manjunath Kondaiah
Cc: Tony Lindgren; [email protected]
Subject: Re: NOTE: hwmod changes merged to linux-omap master for testing
On Mon, Oct 4, 2010 at 2:20 PM, G, Manjunath Kondaiah <[email protected]> wrote:

Quoted 7 lines

Quoted 1 line
Can you post complete boot log? so far, it failed on one
blaze and one
Quoted 1 line
3430sdp both of them are having es2.0
Sure:
wizery.com/ohad/blaze-linux-omap.txt.gz

Here is difference: you are booting with nosmp..

console=ttyO2,115200n8 root=/dev/nfs rw rootdelay=1 mem=460M@0x80000000 mem=512M@0xA0000000 noinitrd vram=12M omapfb.vram=0:5M,1:5M nosmp nfsroot=192.168.1.4:/home/ohad/embedded/myfs_blaze,nolock,devfs=mount ip=192.168.1.160

Remove nosmp from your bootargs and try.

-Manjunath

Ohad Ben-Cohen

2010-10-04 22:30 UTC #10 lore raw

Quoted 1 line · Manjunath Kondaiah

Remove nosmp from your bootargs and try.

Indeed, it doesn't boot.

wizery.com/ohad/blaze-linux-omap-smp.txt.gz

Quoted 5 lines

-Manjunath

Tony Lindgren

author

2010-10-07 00:04 UTC #11 lore raw

Quoted 6 lines · Ohad Ben-Cohen

Quoted 1 line · Manjunath Kondaiah
Remove nosmp from your bootargs and try.
Indeed, it doesn't boot.
wizery.com/ohad/blaze-linux-omap-smp.txt.gz

FYI, posted a fix for this:

http://patchwork.kernel.org/patch/237401/

Regards,

Tony

G, Manjunath Kondaiah

2010-10-07 08:17 UTC #12 lore raw

Quoted 15 lines

-----Original Message-----
From: Tony Lindgren [mailto:[email protected]]
Sent: Thursday, October 07, 2010 5:34 AM
To: Ohad Ben-Cohen
Cc: G, Manjunath Kondaiah; [email protected]
Subject: Re: NOTE: hwmod changes merged to linux-omap master for testing
Quoted 1 line · Ohad Ben-Cohen
On Mon, Oct 4, 2010 at 2:52 PM, G, Manjunath Kondaiah
Quoted 2 lines · <[email protected]> wrote:
Quoted 1 line
Remove nosmp from your bootargs and try.

Quoted 8 lines

Quoted 3 lines
Indeed, it doesn't boot.
wizery.com/ohad/blaze-linux-omap-smp.txt.gz
FYI, posted a fix for this:
http://patchwork.kernel.org/patch/237401/

Thanks! It boots up(both with smp and nosmp options in bootargs) on omap4 blaze. -Manjunath

Kevin Hilman

2010-10-04 15:06 UTC #13 lore raw

In reply to G, Manjunath Kondaiah

Quoted 11 lines · Manjunath Kondaiah

Hi Tony,
With latest merges on master branch with commit ID:
0882b1455797b0a104978000a622c3f2412e7cf5 omap: READ CAREFULLY: Fix build after merging in hwmod support for testing
Booting with zoom3 seems to be fine but observed issues with omap4 blaze board.
Looks like multiple warning dumps and finally it crashes after "Freeing init memory" with both NFS and ramdisk file systems.
The complete bootlogs are available at:
NFS bootlog:
http://pastebin.com/nrWQ5FGG

In the future, it helps greatly if you state what defconfig you're using, and specifically what you've changed from the default defconfig.

Looking at this log, you've enabled CPUidle in your config, which is not (yet) supported for OMAP4 in mainline.

Kevin

Quoted 45 lines

Ramdisk bootlog:
http://pastebin.com/1F75d2Ew
-Manjunath
Quoted 30 lines
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Tony Lindgren
Sent: Saturday, October 02, 2010 6:18 AM
To: [email protected]
Subject: NOTE: hwmod changes merged to linux-omap master for testing
Hi all,
So I've finally merged Kevin's hwmod branch into the linux-omap master so we can give it some testing to see if we can merge it to the mainline. Might take few minutes to hit the mirrors.
Please note that for omap2plus machines you need to change:
- u-boot bootargs for kernel cmdline needs to change with setenv
- if you use hardcoded CONFIG_CMDLINE, change it in your .config
- /etc/inittab in your omap root file system if using serial console
This is because otherwise the serial-omap would be conflicting with the 8250 driver. This should not affect omap1 machines.
Please give it a try and report any issues on this list! Now all the machines except ams_delta_defconfig should build. That one is still missing some camera driver changes to build.
Regards,
Quoted 7 lines
Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

G, Manjunath Kondaiah

2010-10-04 16:46 UTC #14 lore raw

Quoted 21 lines

-----Original Message-----
From: Kevin Hilman [mailto:[email protected]]
Sent: Monday, October 04, 2010 8:36 PM
To: G, Manjunath Kondaiah
Cc: Tony Lindgren; [email protected]
Subject: Re: NOTE: hwmod changes merged to linux-omap master for testing
Quoted 6 lines · Manjunath Kondaiah
Hi Tony,
With latest merges on master branch with commit ID:
0882b1455797b0a104978000a622c3f2412e7cf5 omap: READ CAREFULLY: Fix build after merging in hwmod support for testing
Booting with zoom3 seems to be fine but observed issues
with omap4 blaze board.
Quoted 2 lines
Looks like multiple warning dumps and finally it crashes

Quoted 8 lines

after "Freeing init memory" with both NFS and ramdisk file systems.
Quoted 4 lines
The complete bootlogs are available at:
NFS bootlog:
http://pastebin.com/nrWQ5FGG
In the future, it helps greatly if you state what defconfig you're using,

ok. I have used omap2plus_defconfig

Quoted 2 lines

and specifically what you've changed from the
default defconfig.

No changes to defconfig

Quoted 4 lines

Looking at this log, you've enabled CPUidle in your config, which is not
(yet) supported for OMAP4 in mainline.

As I mentioned no changes to defconfig and it looks like cpu idle is disabled and cpu idle wait is enabled in .config after running make omap2plus_defconfig

$ cat arch/arm/configs/omap2plus_defconfig | grep "CPU_IDLE"
$
$ make omap2plus_defconfig
#
# configuration written to .config
#
$ cat .config | grep "CPU_IDLE"
CONFIG_ARCH_HAS_CPU_IDLE_WAIT=y
# CONFIG_CPU_IDLE is not set
$

Even if I disable CPU_IDLE_WAIT manually, behaviour is same

-Manjunath

G, Manjunath Kondaiah

2010-10-04 18:26 UTC #15 lore raw

Quoted 38 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of G, Manjunath Kondaiah
Sent: Monday, October 04, 2010 10:16 PM
To: Kevin Hilman
Cc: Tony Lindgren; [email protected]
Subject: RE: NOTE: hwmod changes merged to linux-omap master for testing
Quoted 8 lines
-----Original Message-----
From: Kevin Hilman [mailto:[email protected]]
Sent: Monday, October 04, 2010 8:36 PM
To: G, Manjunath Kondaiah
Cc: Tony Lindgren; [email protected]
Subject: Re: NOTE: hwmod changes merged to linux-omap master for testing
Quoted 5 lines
Quoted 3 lines · Manjunath Kondaiah
Hi Tony,
With latest merges on master branch with commit ID:
0882b1455797b0a104978000a622c3f2412e7cf5 omap: READ
CAREFULLY: Fix
Quoted 8 lines
Quoted 3 lines
build after merging in hwmod support for testing
Booting with zoom3 seems to be fine but observed issues
with omap4 blaze board.
Quoted 2 lines
Looks like multiple warning dumps and finally it crashes
after "Freeing init memory" with both NFS and ramdisk file systems.
Quoted 1 line

Quoted 9 lines

Quoted 6 lines
Quoted 3 lines
The complete bootlogs are available at:
NFS bootlog:
http://pastebin.com/nrWQ5FGG
In the future, it helps greatly if you state what defconfig you're using,
ok. I have used omap2plus_defconfig

Since this is for linux-omap master, now we have only omap2plus_defconfig for all omap2plus hence the reason for not mentioning defconfig.

-Manjunath

Kevin Hilman

2010-10-04 19:34 UTC #16 lore raw

In reply to G, Manjunath Kondaiah

Quoted 27 lines · Manjunath Kondaiah

Quoted 11 lines
Quoted 6 lines · Manjunath Kondaiah
Hi Tony,
With latest merges on master branch with commit ID:
0882b1455797b0a104978000a622c3f2412e7cf5 omap: READ CAREFULLY: Fix build after merging in hwmod support for testing
Booting with zoom3 seems to be fine but observed issues
with omap4 blaze board.
Quoted 2 lines
Looks like multiple warning dumps and finally it crashes
Quoted 8 lines
after "Freeing init memory" with both NFS and ramdisk file systems.
Quoted 4 lines
The complete bootlogs are available at:
NFS bootlog:
http://pastebin.com/nrWQ5FGG
In the future, it helps greatly if you state what defconfig you're using,
ok. I have used omap2plus_defconfig
Quoted 2 lines
and specifically what you've changed from the
default defconfig.
No changes to defconfig
Quoted 1 line

Quoted 7 lines

Quoted 3 lines
Looking at this log, you've enabled CPUidle in your config, which is not
(yet) supported for OMAP4 in mainline.
As I mentioned no changes to defconfig and it looks like cpu idle is disabled and cpu idle wait is enabled in .config after running make omap2plus_defconfig

OK, my fault. I misread your bootlog, and now see that it's the default idle path that is triggering the fault, not CPUidle.

I tried on my es2.0 panda and see the same thing now, dump below. The misc. PM warnings on boot are expected as OMAP4 does not have
functioning PM support in mainline yet. The error that crashes is
somewhat different, but also likely related to PM as it appears to be triggered the upon the first attempt at idle.

I haven't debugged any further, but you might want to try again with enabling more verbose kernel debugging, specifically:

CONFIG_DEBUG_BUGVERBOSE=y
CONFIG_DEBUG_ERRORS=y

as you shoul hopefully get some more information as to what causes the initial panic.

Kevin

[    3.495330] Waiting for root device /dev/mmcblk0p2...                        
[    3.508880] mmc0: new high speed SDHC card at address d555                   
[    3.516265] mmcblk0: mmc0:d555 SU08G 7.60 GiB                                
[    3.523223]  mmcblk0: p1 p2                                                  
[    3.607177] EXT3-fs: barriers not enabled                                    
[    3.612609] kjournald starting.  Commit interval 5 seconds                   
[    3.624664] EXT3-fs (mmcblk0p2): using internal journal                      
[    3.630401] EXT3-fs (mmcblk0p2): mounted filesystem with ordered data mode   
[    3.637878] VFS: Mounted root (ext3 filesystem) on device 179:2.             
[    3.644287] Freeing init memory: 268K                                        
[    3.716003] Kernel panic - not syncing: Attempted to kill init!              
[    3.722381] CPU0: stopping                                                   
[    3.725250] [<c0052540>] (unwind_backtrace+0x0/0xec) from [<c004b2f8>] (do_I)
[    3.733917] [<c004b2f8>] (do_IPI+0xfc/0x184) from [<c03efdfc>] (__irq_svc+0x)
[    3.742095] Exception stack(0xc056df78 to 0xc056dfc0)                        
[    3.747375] df60:                                                       00000
[    3.755920] df80: 00000000 0000348f c056c000 c05e8aa8 c003503c c0580234 80032
[    3.764465] dfa0: 0000001f 00000000 00000000 c056dfc0 c00b7b4c c004cd44 2000f
[    3.773040] [<c03efdfc>] (__irq_svc+0x9c/0x160) from [<c004cd44>] (default_i)
[    3.781677] [<c004cd44>] (default_idle+0x30/0x38) from [<c004d2f0>] (cpu_idl)
[    3.790130] [<c004d2f0>] (cpu_idle+0x9c/0xf8) from [<c0008ba8>] (start_kerne)
[    3.798767] [<c0008ba8>] (start_kernel+0x2a4/0x300) from [<80008084>] (0x800)

G, Manjunath Kondaiah

2010-10-05 06:36 UTC #17 lore raw

Quoted 11 lines

-----Original Message-----
From: Kevin Hilman [mailto:[email protected]]
Sent: Tuesday, October 05, 2010 1:05 AM
To: G, Manjunath Kondaiah
Cc: Tony Lindgren; [email protected]
Subject: Re: NOTE: hwmod changes merged to linux-omap master for testing
"G, Manjunath Kondaiah" <[email protected]> writes:

...

Quoted 14 lines

Quoted 1 line
As I mentioned no changes to defconfig and it looks like
cpu idle is
Quoted 2 lines
disabled and cpu idle wait is enabled in .config after running make omap2plus_defconfig
OK, my fault. I misread your bootlog, and now see that it's the default idle path that is triggering the fault, not CPUidle.
I tried on my es2.0 panda and see the same thing now, dump below. The misc. PM warnings on boot are expected as OMAP4 does not have
functioning PM support in mainline yet. The error that crashes is
somewhat different, but also likely related to PM as it appears to be triggered the upon the first attempt at idle.

Quoted 8 lines

I haven't debugged any further, but you might want to try again with enabling more verbose kernel debugging, specifically:
CONFIG_DEBUG_BUGVERBOSE=y
CONFIG_DEBUG_ERRORS=y
as you shoul hopefully get some more information as to what causes the initial panic.

After enabling above debug options also, I don't see any change in the log.
http://pastebin.com/mf8HvQCX

-Manjunath

Kevin Hilman

2010-10-05 14:57 UTC #18 lore raw

Quoted 28 lines · Manjunath Kondaiah

Quoted 9 lines
-----Original Message-----
From: Kevin Hilman [mailto:[email protected]]
Sent: Tuesday, October 05, 2010 1:05 AM
To: G, Manjunath Kondaiah
Cc: Tony Lindgren; [email protected]
Subject: Re: NOTE: hwmod changes merged to linux-omap master for testing
"G, Manjunath Kondaiah" <[email protected]> writes:
...
Quoted 17 lines
Quoted 1 line
As I mentioned no changes to defconfig and it looks like
cpu idle is
Quoted 2 lines
disabled and cpu idle wait is enabled in .config after running make omap2plus_defconfig
OK, my fault. I misread your bootlog, and now see that it's the default idle path that is triggering the fault, not CPUidle.
I tried on my es2.0 panda and see the same thing now, dump below. The misc. PM warnings on boot are expected as OMAP4 does not have
functioning PM support in mainline yet. The error that crashes is
somewhat different, but also likely related to PM as it appears to be triggered the upon the first attempt at idle.
I haven't debugged any further, but you might want to try again with enabling more verbose kernel debugging, specifically:

Quoted 8 lines

Quoted 5 lines
CONFIG_DEBUG_BUGVERBOSE=y
CONFIG_DEBUG_ERRORS=y
as you shoul hopefully get some more information as to what causes the initial panic.
After enabling above debug options also, I don't see any change in the log.
http://pastebin.com/mf8HvQCX

Do you have 'debug' on your kernel command line?

Kevin