[Qemu-devel] Re: Accessing the serial port from proll

[Blue Swirl]

> sparc-elf-ld: section .rodata [00000000ffd08000 -> 00000000ffd0a129] 
> overlaps section .text [00000000ffd00000 -> 00000000ffd08077]

This means that the space reserved for code is full. You can adjust it like 
this:

diff -ru proll-patch-16/qemu/Makefile proll-patch-16b/qemu/Makefile
--- proll-patch-16/qemu/Makefile        2005-08-14 10:25:06.000000000 +0000
+++ proll-patch-16b/qemu/Makefile       2005-11-20 12:23:46.000000000 +0000
@@ -27,8 +27,8 @@
# want to shift it to form a PGD entry. A relocatable label will not work.
# Linux kernel expects us to be at LINUX_OPPROM_BEGVM 
<asm-sparc/openprom.h>.
PROLBASE =   0xffd00000
-PROLRODATA = 0xffd08000
-PROLDATA =   0xffd0b000
+PROLRODATA = 0xffd09000
+PROLDATA =   0xffd0c000
PROLSIZE = 240*1024

# Linux

> Also, some additional background info.  I am actually using the -nographic 
> option.  This project is for a grad class on computer architecture.  The 
> prof. says he has a sparc processor implemented on an fpga and wants to 
> load linux on it, but it doesn't have a hard drive or network connection, 
> so that is why I'm writing a serial loader.

Real serial port hardware needs a bit more initialising than Qemu, for 
example programming the baud rate. In the original JavaStation case where 
Proll was first used, the real OBP did this before Proll booted, so Proll 
didn't have to do everything.

There are probably other similar gaps elsewhere, like setting up DRAM 
controllers, clearing special boot modes, initialising bridge devices, 
loading firmware etc. If you have complete specs for the hardware, these 
should not be too difficult to add if even necessary.

_________________________________________________________________
Don't just search. Find. Check out the new MSN Search! 
http://search.msn.com/