[HOWTO] Compiling new kernels on Harmattan device - Page 13 - maemo.org

Active Topics

The N900 is now 15 years old! (9)
to Nokia N900 by Arno_11 - 4 hrs, 52 mins ago
Sailfish OS for the Motorola Moto G7 Power (XT1955-5) - (ocean) (4)
to SailfishOS by Maemish - 5 days, 17 hrs ago
more...

Page 13 of 19

Thread Tools

juiceme	2013-09-10 , 14:36
Community Council \| Posts: 4,920 \| Thanked: 12,867 times \| Joined on May 2012 @ Southerrn Finland	#121

Originally Posted by juiceme

I am getting unresolved symbols....

Okay it was just as I thought, NFSD depends on lockd.ko, sunrpc.ko and exportfs.ko modules. I did not realise that exportfs.ko was missing...

Fortunately you do not need to download all the modules, it is enough to load the exportfs module: http://www.swagman.org/juice/exportfs.ko (md5sum=5f7718f81737d7e5c0acf80903172f28)

And remember to run /sbin/depmod

Quote & Reply |

	peterleinchen	2013-09-15 , 20:22
	Posts: 4,118 \| Thanked: 8,901 times \| Joined on Aug 2010 @ Ruhrgebiet, Germany	#122

Hej juiceme.

I have problems to run your nat enabled kernels with ubiboot.
I extracted your tar, edited conf and nothing. Black screen after selection.
Checked boot (kernel is there), checked modules folder (looks okay), reran depmod... But all to no avail.
Log file tells me kernel selection was okay and successfully kexec loaded. [dmesg looks okay too, and it only tells up to execution of second kernel, right?).
So what else could I check? Or do you have any other idea?

(tried with nat_extfat kernel and new modules, also failing)

__________________
SIM-Switcher, automated SIM switching with a Double (Dual) SIM adapter
--
Thank you all for voting me into the Community Council 2014-2016!

Please consider your membership / supporting Maemo e.V. and help to spread this by following/copying this link to your TMO signature:
[MC eV] Maemo Community eV membership application, http://talk.maemo.org/showthread.php?t=94257

editsignature, http://talk.maemo.org/profile.php?do=editsignature

Quote & Reply |

juiceme	2013-09-16 , 11:57
Community Council \| Posts: 4,920 \| Thanked: 12,867 times \| Joined on May 2012 @ Southerrn Finland	#123

Originally Posted by peterleinchen

Hej juiceme.

I have problems to run your nat enabled kernels with ubiboot.
I extracted your tar, edited conf and nothing. Black screen after selection.
Checked boot (kernel is there), checked modules folder (looks okay), reran depmod... But all to no avail.
Log file tells me kernel selection was okay and successfully kexec loaded. [dmesg looks okay too, and it only tells up to execution of second kernel, right?).
So what else could I check? Or do you have any other idea?

(tried with nat_extfat kernel and new modules, also failing)

Hmm well, this is strange indeed.

Just about only thing that can go wrong is the modules, and usually that is when you have the wrong name for the modules directory or have depmod mixed up somehow.
The system is pretty tolerant on trivial stuff like file permissions since the modules are accessed as root anyways...

But if you have installed the modules correctly in correct directory then I cannot easily say what's the matter...

This is what I have, and it worked the correctly out-of-the-box:

~ # uname -a
Linux RM696 2.6.32.59-plus_nat-20121301 #1 PREEMPT Tue Sep 3 20:09:42 EEST 2013 armv7l GNU/Linux
~ #
~ # ls -la /lib/modules/ | grep plus
drwxrwxr-x 2 root root 4096 Sep 10 17:26 2.6.32.59-plus_nat-20121301
drwxrwxr-x 2 root root 4096 Sep 10 17:01 2.6.32.59-plus_nat-20121301_njuu
~ #

How does your device behave, exactly? Does it reboot, does it shut down or does it just hang there with a black screen?

Your problem might also depend on something else as the kernel, do you have some unorthodox configuration that would depend on some kernel feature not implemented on the 2.6.32.59-plus_nat kernel? (I doubt that since I am fairly sure I left no option off the config...)

You might try to look into the log partition (with command "strings /dev/mtd3") as sometimes the device writes a boot log there on a failing boot. It's a ring buffer so you'd need to dig it a bit to find out which entry is the latest, if there at all. (and it does not have any dates in the logs, bummer...)

Other than that, well there's always the serial console, the last hope of the damned. I have not yet run on any problems that I could not debug with that medicine

Quote & Reply |

The Following User Says Thank You to juiceme For This Useful Post:
peterleinchen

	peterleinchen	2013-09-16 , 19:42
	Posts: 4,118 \| Thanked: 8,901 times \| Joined on Aug 2010 @ Ruhrgebiet, Germany	#124

juiceme, thanks for reminder of /dev/mtd3.
But I did not find any occurence of 2.6.32.59, so I doubt there is a log at all.
I select the kernel and then the device just stay with a black (slighly dimmed) screen, just as a first nemo boot. I need to push 8s power.
I do not have anything special (just this little aegis/sip sylimking, you know

, but this would appear later). Also syslog on harmattan does not have any entry. So it looks like the kernel does not start at all. When it would do and had problems with some modules, I would find some entry in syslog or mtde, right?
I am out of ideas...

--edit

Originally Posted by juiceme

~ # ls -la /lib/modules/ | grep plus
drwxrwxr-x 2 root root 4096 Sep 10 17:26 2.6.32.59-plus_nat-20121301
drwxrwxr-x 2 root root 4096 Sep 10 17:01 2.6.32.59-plus_nat-20121301_njuu
~ # [/FONT][/INDENT]

What is that _njuu folder? I do not have this one.

And about serial: yeah sure, but I do not have such stuff

Last edited by peterleinchen; 2013-09-16 at 19:59.

Quote & Reply |

juiceme	2013-09-16 , 20:24
Community Council \| Posts: 4,920 \| Thanked: 12,867 times \| Joined on May 2012 @ Southerrn Finland	#125

Originally Posted by peterleinchen

juiceme, thanks for reminder of /dev/mtd3.
But I did not find any occurence of 2.6.32.59, so I doubt there is a log at all.
I select the kernel and then the device just stay with a black (slighly dimmed) screen, just as a first nemo boot. I need to push 8s power.

OK, this means the kernel is running, since if it was powered off (as also could sometimes happen) just 1s push would be enouh to start it.
It's running, but hanging somewhere, propably there'd be some nice kernel message waiting for you on the serial console

This is propably the reason you don't get anything on MTD3, since it does not panic/restart it won't create the log entry.
(Jonni might confirm if my thinking is straight on this, but AFAIK the MTD3 log is written by the boot loader if the kernel panics, and only then.)

Originally Posted by peterleinchen

I do not have anything special (just this little aegis/sip sylimking, you know , but this would appear later). Also syslog on harmattan does not have any entry. So it looks like the kernel does not start at all. When it would do and had problems with some modules, I would find some entry in syslog or mtde, right?
I am out of ideas...

Well I belive it starts OK but does not get so far as it will write anything to syslog. As it goes, kernel log is only stored in memory for the first part of boot, and only when it mounts /root it will write the log out to syslog.
You can look at the logs, see when kernel starts the first part does not have any date tags, only microseconds since kernel starts. That is because before mounting /root, starting syslogd and initializing RTC the kernel knows nothing about the environment it starts in.

Originally Posted by peterleinchen

What is that _njuu folder? I do not have this one.

Well yeah, it's just another kernel build I am having there, some different parameters meant I had to recompile modules.
You know, have a new build every other day

Originally Posted by peterleinchen

And about serial: yeah sure, but I do not have such stuff

Well I know but you should, since it's the best thing since sliced bread really! And it's not that difficult, see the info I gave Jackburton earlier..

PS: One thing what comes to my mind... (but sure you hd this done OK...)
When you did run depmod, you were in the "/lib/modules/2.6.32.59-plus_nat-20121301" directory, and you spelled it correctly, "depmod -a /lib/modules/2.6.32.59-plus_nat-20121301"

If you misspelled that, it won't work correctly, and if I remember correctly depmod does not give any error messages...

Quote & Reply |

	peterleinchen	2013-09-16 , 20:50
	Posts: 4,118 \| Thanked: 8,901 times \| Joined on Aug 2010 @ Ruhrgebiet, Germany	#126

Originally Posted by juiceme

PS: One thing what comes to my mind... (but sure you hd this done OK...)
When you did run depmod, you were in the "/lib/modules/2.6.32.59-plus_nat-20121301" directory, and you spelled it correctly, "depmod -a /lib/modules/2.6.32.59-plus_nat-20121301"

If you misspelled that, it won't work correctly, and if I remember correctly depmod does not give any error messages...

I was in /root dir and I typed
/sbin/depmod 2.6.32.59-plus_nat-20121301 and all modules.* files were written with a new date stamp.
Hmm...

Quote & Reply |

juiceme	2013-09-16 , 20:57
Community Council \| Posts: 4,920 \| Thanked: 12,867 times \| Joined on May 2012 @ Southerrn Finland	#127

Originally Posted by peterleinchen

I was in /root dir and I typed
/sbin/depmod 2.6.32.59-plus_nat-20121301 and all modules.* files were written with a new date stamp.
Hmm...

I'm fairly sure you need the "-a" there...

Quote & Reply |

	peterleinchen	2013-09-16 , 21:05
	Posts: 4,118 \| Thanked: 8,901 times \| Joined on Aug 2010 @ Ruhrgebiet, Germany	#128

I typed -a, but just read man page and it is the default, so does not matter.

But even then, your tar had all updated.. And what goes wrong untaring a folder and a kernel???

Quote & Reply |

juiceme	2013-09-16 , 21:21
Community Council \| Posts: 4,920 \| Thanked: 12,867 times \| Joined on May 2012 @ Southerrn Finland	#129

Originally Posted by peterleinchen

I typed -a, but just read man page and it is the default, so does not matter.

But even then, your tar had all updated.. And what goes wrong untaring a folder and a kernel???

Well I'd say nothing can go wrong but usually just saying it is prone to trigger the impossible...
But really, nothing should go wrong.

And all other kernels boot correctly, right?

Are you sure you do not have something non-standard I might have overlooked, I do remember you had some extra partitioning there, maybe something else too?

Quote & Reply |

	peterleinchen	2013-09-16 , 21:33
	Posts: 4,118 \| Thanked: 8,901 times \| Joined on Aug 2010 @ Ruhrgebiet, Germany	#130

Yep, no other problems (except the one just posted in ubiboot thread

).

And yes I do have manually partitioned device with additional p4/5/6/7 in between p1 and p2 (just like nemo, but extended with 3 partitions), so that is similar. And root dev is p2 as normal. Nothing special else I can remember right now.