[okl4-developer] regarding interrupts

[Nelson Tam]

Hi Gabi,

On 21/04/2008, at 3:06 PM, Gabi Voiculescu wrote:
> Hello again.
>
> Thanks for the advice on the entering kdb. Your advices helped me to  
> rule out a set of causes for the freeze and pointed me to debug the  
> kernel timer driver.
>
> After I resolved the timer problems I have been able to enter kdb  
> from the console.
>
> Test_libs=all runs with 99% success. The only miss is at
>
> libs/iguana/test/memsect_tests.c:235:F:Memory Sections:MEMS0600:  
> Unable to create and direct section.
>
> Max_memsections is 0.
>
> Is this an error? Any thoughts on what can cause this error? I  
> should say I have mapped 128MB of RAM.

Can you explain a bit more... do you mean you have already mapping  
128MB of RAM somewhere else, on top of what's running in the memsect  
test?  How did you map this?

Generally speaking, we don't make public releases unless they pass all  
our regression tests at 100% on our supported platforms.

> Beside this I have a couple more questions for you. I would  
> appreciate even partial answers to these.
>
> 1 - in platform/..../plat.cc in OKL4 1.5.2 I have defined a vector  
> called arm_bootdesc which I see defines non RAM areas for the  
> particular platform.What is it's purpose?
>     1a - If I use the reserved type can I impose restrictions on a  
> memory area - say flash?

Reserved means that it won't be touched by the kernel, and won't be  
allocated for kernel memory or user memory.

>     1b - Do I define RAM sections in there?

Yes.

>     1c - Is there a particular order in which the entries have to be  
> declared? For example in every example platform except for one the  
> areas are defined in incremental order.

No, but is there any reason why you have to define them in random order?

> 2 - Is there a way to get rid of the v1 driver information in  
> platform/..../tools/machines.py, and still have the build create a  
> good image?
> Currently, on okl4-1.5.2, if I comment out the lines
>
> timer_driver = "arm1176_timer"
> serial_driver = "drv_l4_kdb"
> drivers = [timer_driver, serial_driver]

Are your platform are still using the v1 serial driver for console  
output?  Once you define the v2_driver variable, the v1 drivers should  
not have been built anyway so removing that configuration should cause  
no harmful effects.

> I get two warning messages in the beginning of the build, and more  
> importantly, I can no longer enter in okl4 (program halts somewhere  
> in init_tracebuffer() ). My only concern would be if there is a  
> conflict between the v1 and v2 drivers (currently I specify  
> different drivers for the kdb serial) that the test_libs don't pick  
> up.

--
(nt)

Nelson Tam

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1560 bytes
Desc: not available
Url : http://lists.okl4.org/pipermail/developer/attachments/20080429/43f0e5f4/attachment.bin