Re: PA-RISC 2.0 support

I finally got around to getting the cvs source and poking around.  From the 
looks of it, the issues you mentioned are still open.  Most of my development 
experience is in userland, and my low level experience is all on embedded 
devices (just a tiny bit different from a PA-RISC ;) so getting my feet wet 
is really slow going, so far.

In the __sigreturn14 thing, with a bit of reading I could do the FPU context 
saving stuff (unless there's a big issue that I missed), but I don't quite 
get the problem with the flag checking thing, mostly due to lack of 
experience with this level of the system.  Are there any decent docs that I 
could check out to fix this?

I also took a look at what TLB/PTE stuff I could find, and at my experience 
level, adding 2.0 support is not something I'd do in a couple weeks, 
nevermind an afternoon.

Anyway, I'll try my hand at doing wscons and fixing __sigreturn14, and 
hopefully that'll be enough of a warm up for me to do something bigger.

On March 31, 2004 11:55 am, Jochen Kunz wrote:
> On Wed, 31 Mar 2004 09:52:13 -0700
>
> Reid Orsten <[email protected]> wrote:
> > Roughly what needs to be done for the 2.0's to work?
>
> Support for the PA 2.0 TLB / PTE format.
>
> I had a short look at this. It seems that PA 2.0 is a "linear" extension
> of PA 1.1. It is possible to run 32 bit binaries (including the kernel)
> on PA 2.0 machines when the PSW W bit is 0. But, obviously, the TLB and
> everything related is extended to support the 64 bit physical addres
> space. So this is somthing for you when you are interrested in some
> quite low level, machine dependent virtual memory management work. This
> is nothing you do in an afternoon.
>
> See chapter 2 and 3 of parisc2.0.pdf. (can be found on
> http://www.openpa.net/)
>
> > I've got a couple lying around (C180, C360, J282).
>
> C240 here.
>
> > I've also got a 712-60 and a 715-65 that are being completely ignored;
> > so I could help out with some general development as well.
>
> Well. There are some other things like wscons to do, I still didn't get
> the "energy" and time to dig into the longjmp / __sigreturn14 issue, gdb
> for hppa / hp700 is broken, some MD pthread bits are missing... and an
> other developer was looking at the ld.elf_so problem.
>
> BTW: Thanks to Jens Konrad. He gave me K100 so I am now able to work on
> PA7200 support.