<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Grant,<br>
<br>
I've pulled the latest git and built a mpc5200_simple system with a
minimal dts.<br>
There's not a single char put on the console ....<br>
<br>
<br>
<br>
Grant Likely schrieb:
<blockquote
cite="mid:fa686aa40803161301r1685f2cale75d67b8a203cca3@mail.gmail.com"
type="cite">
<pre wrap="">On Sun, Mar 16, 2008 at 1:15 PM, André Schwarz
<a class="moz-txt-link-rfc2396E" href="mailto:Andre.Schwarz@matrix-vision.de"><Andre.Schwarz@matrix-vision.de></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">All,
I'm quite stuck in getting our MPC5200B based systems work on 2.6.24+
... maybe someone could give me some hints.
Up to now the systems have been running on 2.6.19 without any problems.
This is what I've done so far :
- get a recent system with 2.6.19 running and keep the toolchain (gcc
4.0) including RFS uImage.
- latest bootloader (u-boot 1.3.2) is running fine with dtb support.
- Kernel and RFS images are available via bootp/tftp
- device tree blob seems ok, i.e. u-boot can display the blob and fill
in all clocks correctly.
- Kernel gets called with Kernel, RFS and dtb on suitable adresses
(0x100000, 0x600000, 0x5F0000).
- nothing more happens. It looks like the console isn't working properly
- maybe worse ...
</pre>
</blockquote>
<pre wrap=""><!---->
Can you either halt the processor or reset the board and access memory
after this point? You should see if you can access __log_buf (Look in
System.map for the address and subtract 0xc0000000 to get the physical
address). __log_buf should give you clues about how far booting
progresses.
</pre>
</blockquote>
(c033d4b4 b __log_buf) yields all zeroes - nothing logged at all.<br>
<br>
Examining the CPU state with the bdi shows :<br>
<br>
Target CPU : MPC8280/8220/5200 (Zeppo)<br>
Target state : debug mode<br>
Debug entry cause : COP halt<br>
Current PC : 0xc000552c<br>
Current CR : 0x48000042<br>
Current MSR : 0x00009032<br>
Current LR : 0xc00166c0<br>
<br>
The program counter shows the adress of "irq_set_virq_count" which has
a BUG() macro. Looks like this one evaluates to true.<br>
<br>
What do you think ?<br>
<br>
<blockquote
cite="mid:fa686aa40803161301r1685f2cale75d67b8a203cca3@mail.gmail.com"
type="cite">
<pre wrap=""></pre>
<blockquote type="cite">
<pre wrap=""> The kernel has been build using the "MPC5200 simple" setup. I've added
our board identifier "mv,mvbluecougar" to the list of supported boards
in the "mpc52xx_simple.c" file. This identifier is used in the dtb as
well of course. This should be enough along with a valid device tree
blob, isn't it ? Did I miss anything already ?
</pre>
</blockquote>
<pre wrap=""><!---->
Yes, this should be enough (on a side note; it should probably be
"matrix-vision,bluecougar" unless of''mv' is your company's stock
ticker symbol).
</pre>
<blockquote type="cite">
<pre wrap=""> As far as I did understand by reading the docs it should be possible
running a general MPC52xx kernel with a custom dtb ? Is this correct ?
</pre>
</blockquote>
<pre wrap=""><!---->
That is correct.
</pre>
<blockquote type="cite">
<pre wrap=""> The dts file has already been reduced to a minimum, i.e. only the cpu,
pic, xlb and a few nodes like psc1 and psc2 are present. No bestcomm,
pci, ata or "advanced" features like mtd partitions. All I want to get
working is the serial console on "/dev/ttyPSC0".
</pre>
</blockquote>
<pre wrap=""><!---->
Cheers,
g.
</pre>
</blockquote>
<br>
regards,<br>
Andre<br>
<BR>
MATRIX VISION GmbH, Talstraße 16, DE-71570 Oppenweiler - Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschäftsführer: Gerhard Thullner, Werner Armingeon, Uwe Furtner
<BR>
</body>
</html>