Hi mario,
thanks again for your support.
So when I create the image using mkifs the IPL is created as well? Sorry, but I don’t see 100% clear on this yet.
So let me get this straight. All I have to do is to alter the bootscript in my buildfile in order to get the image to a certain address in RAM?
My current buildfile (bootstrap section) looks like this:
[+compress]
[ //MY ATTRIBUTES, SEE BELOW// ]
[virtual=x86,bios] .bootstrap = {
startup-bios
PATH=/proc/boot LD_LIBRARY_PATH=/proc/boot procnto -vv
}
So all I have to do is to add … what exactly?
Reading through the documentations of QNX 6.2.1 I’ve found that one can specify
[image=0xXXX]
[ram=0xXXX]
[default_image=0xXXX]
[default_ram=0xXXX]
Please, which one do I need? Because I’ve been experimenting with those a bit. When adding [ram=0xXXX], the system reboots immediately after booting. This can’t be right. The addresses I used are randomly chosen and it doesn’t seem to make any difference.
When adding [image=0xXXX], the system boot normally, but I wasn’t able to find any differences as to where the image went in RAM. Using dumpifs on the image, I got the address of startup_vaddr, that’s, if I’m not mistaken the address in RAM the IPL jumps to when handing over control to the image. Conducting a very simple memory dump at that address, i.e. mapping the memory via mmap_device_memory() and printing it on screen, only brought lot’s of zeroes. So something must be pretty wrong here. Either I’m looking at the wrong address, or mmap_device_memory() is initializing the memory with zeroes. But I don’t believe it’s the latter one.
The attributes [default_image=0xXXX] and [default_ram=0xXXX], as they were stated out in the documentation for the bootscript are rejected all along by mkifs.
Please, can you straighten this out for me a bit, because this is really driving me nuts. I just started out with QNX, everything worked fine so far, only this thing seems to resist doing what I expect it to do.
Much obliged 