Ok.
A) disable PCI retry.
B)
Bool
xf86DisableInterrupts()
{
if (!ExtendedEnabled)
#if !defined(__mc68000__) && !defined(__powerpc__) &&
!defined(__sparc__) && !defined(__mips__)
if (iopl(3) || ioperm(0, 1024, 1))
^^^^^^^
replace with
0
Recompile X.
Michael has tried with nvidia and no problem.
On Sat, Nov 25, 2000 at 01:35:59AM +0100, Erwin Rol wrote:
> OK here is what i have tried in the last few days.
> The following part where the same in all the tests.
> XFree 4.0.1
> Athlon TB 800, 128MB PC133, ASUS K7V MB,
>
> 1)
> with a s3virge based card
> RTAI 1.5 on a 2.2.16 kernel
>
> running the latency example which resulted in in peaks of 50msec latency
> when i for exmple
> pressed enter in a console window (causeing it to scroll one line or
> more) also heavy X by dragging
> windows as fast as posible caused also latency.
>
> No change in the pci setup like sugested helped, appart from not using
> acceleration.
>
> 2)
> with a Matrox G400MAX
> RTAI 1.5 on a 2.2.16 kernel
>
> same latency example wich was way beter , no more problem with dragging
> and enter presses. But when i do a ls -Rl
> in a console window latency jumps up to 100-500usec with peaks at
> 1.5msec. When i minimize that window the latency goes
> back to 3-15usec with peaks of 30usec.
>
> 3)
> with a Matrox G400MAX
> RTL 3.0 pre9 on a 2.2.17 kernel
>
> the ./examples/measurements test results in the same max latencies as
> with RTAI.
>
> all 3 tests i also ran in failsafe mode with only xterminals (no gnome,
> WM etc.)
>
> sO the big teast left is trying all that with FXRee 3.3.6 , which i will
> have to install again first :-/ but i don't see
> a othe choice to rule out the hardware. An other thing would be when
> someone else could reproduce it , the person
> on comp.realtime sounded like he had the same problem.
>
> If there are any test i should run , or something else there is that you
> want to know , just give a shout.
>
> - eRwin (and i blamed Paolo for calling me Edwin , sory Paolo :-)
>
>
> [EMAIL PROTECTED] wrote:
> >
> > On Fri, Nov 24, 2000 at 09:04:48PM +0100, Erwin Rol wrote:
> > > The thing that worries me a bit is the fact that now 3 XFree 4.0.1
> > > drivers cause problems
> > > for RT-Linux , the nVidia, the mga and the s3virge one.
> >
> > Edwin:
> > I had the impression that you were running RTAI. While this may be
> > a harware/low-levelX problem, I'd like to get a precise bug report
> > before we start speculating about X4 and the PCI bus.
> > Version of RTLinux
> > Version of Linux
> > Hardware
> > Test code.
> >
> >
> >
> >
> > >
> > > I would really like to see someone else try Xfree 4.0.1 with RT-linux (i
> > > know it is a pain when you have
> > > a distribution with XFree 3.3.x :-/ )
> > >
> > > OK will keep you posted on what else i find out,
> > >
> > > Erwin
> > >
> > >
> > > Michel D�nzer wrote:
> > > >
> > > > Erwin Rol wrote:
> > > >
> > > > > > - Try "pci_burst" and "pci_retry" in the Device section. (Both on.)
> > > > > done that.
> > > >
> > > > What if you disable PCI retry? I'd have imagined that the PCI bus might stall
> > > > when the CPU tries to access a register while the accelerator is doing a
> > > > time-consuming task.
> > > >
> > > > Please forgive me if this is completely off...
> > > >
> > > > Michel
> > > >
> > > > --
> > > > Earthling Michel D�nzer (MrCooper) \ CS student and free software enthusiast
> > > > Debian GNU/Linux (powerpc,i386) user \ member of XFree86 and The DRI Project
> > > -- [rtl] ---
> > > To unsubscribe:
> > > echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
> > > echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
> > > ---
> > > For more information on Real-Time Linux see:
> > > http://www.rtlinux.org/rtlinux/
> >
> > --
> > ---------------------------------------------------------
> > Victor Yodaiken
> > Finite State Machine Labs: The RTLinux Company.
> > www.fsmlabs.com www.rtlinux.com
> -- [rtl] ---
> To unsubscribe:
> echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
> echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
> ---
> For more information on Real-Time Linux see:
> http://www.rtlinux.org/rtlinux/
--
---------------------------------------------------------
Victor Yodaiken
Finite State Machine Labs: The RTLinux Company.
www.fsmlabs.com www.rtlinux.com
-- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
---
For more information on Real-Time Linux see:
http://www.rtlinux.org/rtlinux/
