[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Re: Atari machines running Debian?

> Michael Schmitz pí?e v Út 22. 08. 2006 v 17:59 +0200:
> > AFAIR 2.5.15 still did not have discontig memory support, so wouldn't
> > boot with TT-RAM. (I did send patches for 2.6.13 which went on top of
> > Roman's patches, but didn't find any of these made it into 2.6.15)
> Luckily ARAnyM offers up to 14 MB of ST-RAM (if you disable FastRAM by
> settings its size to zero). Better than most TT/Falcons with just 4 MB
> of ST-RAM.

I'll try that. My 2.6.13 (the patched tree that used to run on the TT)
dies in calibrate_delay() with scheduling from the idle thread. I receive
no debug output whatsoever after time_init(). Investigating.

CVS 2.6.17 seems borked right now (lotsa nonexisting files messing up
make clean etc.). I'll keep poking around in 2.6.13.

> > With discontig memory support applied, 2.6.13 still had issues with
> > interrupts (hardirq_count wasn't properly increased)
> that's not good.

I still don't understand the reason.

"atari_slow_irq_" #n "_handler:\t"                                         \
"       addl    %6,%5\n"        /* preempt_count() += HARDIRQ_OFFSET */    \
        SAVE_ALL_INT "\n"                                                  \
        GET_CURRENT(%%d0) "\n"                                             \

(%5: "m" (preempt_count()), %6: "di" (HARDIRQ_OFFSET))

is what fails to work,

"atari_slow_irq_" #n "_handler:\t"                                         \
        SAVE_ALL_INT "\n"                                                  \
        GET_CURRENT(%%d0) "\n"                                             \
"       addqb   #1,%%a2@(%c5)\n"        /* hardirq_count + 1 */         \

( %5: "di" (TASK_INFO+TINFO_PREEMPT+2) )

is what worked for me. I _think_ I even moved the addl    %6,%5 after
get_current, but I cannot recall the details.

> > and mouse/keyboard support was missing.
> this might not be hard to add.

It wasn't. I'll put up my old patches at our web server ...

> > I did receive some patch from Petr, but that would not run either...
> yes, that must have been the patch I was talking about - it was written
> by Milan Jurik but without any testing, I think.

With Geert's clarification I'll give it a fresh try. I'm still unsure how
to acquire a working screen buffer with ST-RAM swap disabled, but that'll
eventually work out.


Reply to: