Re: JFFS2 directory copy problem

[Cmmn Ml <cm...@ya...>]

Hi Paul,

Thanks for the reply, as I'm still in the process of getting the problem fixed. Let me allow to brief you about the problem I'm having.

As you have mentioned about my configuration file, it is based on SE7705 defconfig file. The board I'm working on is virtually the same as the Solution Engine board. On this board 2.6.16.9 kernel (with sh--stable--2.6.16--base-0 patch) is running well without any problem other than a problem I was experiencing with JFFS2 file system (trying to copy a directory which is larger than around 2 MB  makes the system to crash). Because of that I'm trying to use 2.6.21-rc7 kernel to see whether the JFFS2 problem is fixed in it.

To build the 2.6.21-rc7 kernel for SH7705, I had to add a patch from 2.6-git added by Nobuhiro (http://git.kernel.org/?p=linux/kernel/git/lethal/sh-2.6.git;a=commit;h=5452d25e28f365938b36f3a972fe2a1bb69a4c3e). This patched 2.6.21-rc7 kernel boots up fine, but fails to start the /sbin/init and stops running. To check at least the timer interrupt is handled after this point, I added a printk into the timer interrupt handler. With this, when the kernel boots up,  after the timer interrupt is registered, I'm getting the printk messages for each timer interrupt as expected. As you suggested that makes me sure that the timer IRQ has no problem.

But at the end of the initial kernel bootup process, when it tries to start the system console by calling run_init_process("/sbin/init"), everything goes wrong. As I mentioned, after calling run_init_process("/sbin/init") function, the kernel hangs at that point and there after I'm not getting the printk message from the timer interrupt handler. That means after this point, the kernel is not in a running state. Based on all the above observations, I suspect that we may have a problem in the kernel where it tries to start the /sbin/init.

I'm wondering to know that can it be a possibility that there may be a problem in the kernel when it tries to start /sbin/init ? if so where should I look to fix it up, and is there any one who has successfully used the 2.6.21-rc7 kernel on SH7705 without this problem ?

If any one of you can give me some thoughts about this, that will be a great help for me to fix it up.

Thanks in advance.

Cmmn





Paul Mundt <le...@li...> wrote: On Wed, May 09, 2007 at 07:33:58AM -0700, Cmmn Ml wrote:
> As a summary what happens is the kernel boots up fine up to the point
> where run_init_process("/sbin/init") is called. After the
> run_init_process("/sbin/init") function is called, kernel hangs itself
> and even the timer interrupt is not handled. As I feel the kernel fails
> to start a user space process and puts it self to a not running state.
> 
If you're never getting the timer interrupt, then there's certainly
something terribly wrong with the IRQ handling. That has to be fixed
first.

These should already be handled via the IPR map though, are you dealing
with a custom board? Do you have any unusual TMU configuration? Also make
sure that the timer IRQ is set to the proper vector.. You need to be
getting interrupts before you start worrying about userspace problems.

Your config looks pretty close to the SE7705 defconfig, so the timer
should certainly not be causing you any difficulty.


       
---------------------------------
Be a better Heartthrob. Get better relationship answers from someone who knows.
Yahoo! Answers - Check it out.