Re: Debugging nk.bin..

Tech-Archive recommends: Repair Windows Errors & Optimize Windows Performance

Thanks for your reply...

When the kernel start function is called from my assembly i should get the
debug messages in my PB. Is it right?? Or when does my PB starts giving
debug messages...

Since my PB is not giving messages i thought before calling kernelstart
itself (i mean at assembly) a crash was there...

But i have my lcd backlight connected to gpio20, i think i may use it for
debugging... But can you give me the flow of the program from kernelstart.. i
mean what are the functions that are called one by one after kernel start...
since kernel start is not available as source code.. how do we predict the
flow after it??

--
Ramesh.c


"Bruce Eitman [eMVP]" wrote:

Before OEMInit is called OEMInitDebugSerial is called. What you should do
is read the source code in the Kernel to see what it actually does, search
for things like OEMInit.

The next thing I would do is carefully add some code in the serial init to
toggle an I/O pin that you can scope. By carefully, you should make sure
that the code works, otherwise it won't tell you anything. A JTAG debugger
would come in handy right now.

--
Bruce Eitman (eMVP)
Senior Engineer
beitman AT applieddata DOT net

Applied Data Systems
www.applieddata.net
An ISO 9001:2000 Registered Company
Microsoft WEP Gold-level Member

"Rameshchandra" <Rameshchandra@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:B4971EA9-F241-46D0-A1CE-FD80EC70F35C@xxxxxxxxxxxxxxxx
Hi, thanks for your attention...

There were no such serious warnings or errors while doing the makeimg or
while compiling my bsp, everything went smooth on my compilation side...

In my bsp i predicted the flow to be like this...

1) Assembly code which does basic low level initialization
2) Then the code checks if we were in Kernel side or eboot side (startup
func)... and if we were on the kernel side then call kernelstart
function...
3)Kernelstart not found in the wince sources.. may be in lib. provided by
microsoft..
4)Then OEMINIT and other high level functions..
5)I put a retail message on the OEMINIT but not printed on my
hyperterminal.
6)I suspect the problem may be a crash on the assembly code...
Waiting for an emulator to solve this issue...

But the same low level stuff worked in my 4.2 bsp, which makes me to
confuse..

Is my proceeding is rt? or something i am making mistake here??
--
Ramesh.c


"Bruce Eitman [eMVP]" wrote:

Did you check the makeimg output for errors or serious warnings, like
ramimage overlaps ram.

--
Bruce Eitman (eMVP)
Senior Engineer
beitman AT applieddata DOT net

Applied Data Systems
www.applieddata.net
An ISO 9001:2000 Registered Company
Microsoft WEP Gold-level Member

"Rameshchandra" <Rameshchandra@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in
message
news:07C80F80-4809-47B3-BF28-C5673AECD1B3@xxxxxxxxxxxxxxxx
Hi all,

I am actually in the process of migrating 4.2 bsp to 5.0. My bsp was
working very well on 4.2. Now, i have migrated the catalog items, done
bsp
migration( linked my bsp against the wince 5.0 libraries). Now, created
the
nk.bin in debug mode after compiling the entire bsp without any errors.
I
am using the old eboot which was burnt in my board.(which worked
perfectly
for 4.2).

Now, i donwloaded my new nk.bin and on the hypterminal i am getting the
message that jumping to image at physical addres (address of my RAM)
from
the boot loader side and
then no more messages are displayed...

No retail messages, My PB's kitl is not connecting to the board.. and
no
debug messages on the PB. I was almost left without any clue on
debugging
my
new nk.bin...

Problem with my kitl?? or bsp migration failed?? I followed the
microsoft
documentation on migrating the bsp.. Where i can suspect the problem to
be...

Please any one of you guide me is solving this issue...
--
Ramesh.c






.

Relevant Pages

  • Re: Special schedulers, one CPU only kernel, one only userland
    ... already have, do some small thing and then reacquire the lock itself, ... this seems similar to the handling of interrupts: ... >> it get serviced my the BSP? ... Where kernel threads explicitely pinned to ...
    (freebsd-arch)
  • Re: OEM Adaption Layer and OAL stubs
    ... function build into the kernel at image creation time). ... > Currently im using the latest VIA cepc twister bsp for my target. ... > this OEM specific methods actually are used to create a bsp so that I may ... > Can anyone help me clearify or send a link that can explain the structure ...
    (microsoft.public.windowsce.embedded)
  • Re: Debugging nk.bin..
    ... Before OEMInit is called OEMInitDebugSerial is called. ... In my bsp i predicted the flow to be like this... ... Then the code checks if we were in Kernel side or eboot side (startup ... I am actually in the process of migrating 4.2 bsp to 5.0. ...
    (microsoft.public.windowsce.platbuilder)
  • Re: OEM Adaption Layer and OAL stubs
    ... > function build into the kernel at image creation time). ... >> Currently im using the latest VIA cepc twister bsp for my target. ... >> Among other things I want to be able to call the realtime clock to get ... >> Can anyone help me clearify or send a link that can explain the structure ...
    (microsoft.public.windowsce.embedded)
  • Re: [patch 0/3] dynamic_printk: new feature
    ... How about not just debug messages but setting a loglevel for each module ... kernel messages emitted beneath the loglevel value ... Replace the unconditional printk() messages with printkmessages that are ...
    (Linux-Kernel)