[Sheflug] XFree 4.0 - the saga continues

["Chris J/#6" <nospam [at] nccnet.co.uk>]

Well, XFree86 4.0.3 is still playing silly buggers, interrupting, 
probably right down at the hardware level, with the smooth running of 
XMMS.

Stuff so far played with...

I nudged an optionin XMMS that said "use realtime rights" (or words to 
that effect), was informed that this required XMMS to run as root, made 
it SUID, and got a rap on the knuckles by the GTK+ libs, which point 
blank refuse to run on a SETUID program. One piece of C later, and GTK+ 
can go take a jump of a cliff. Unfortunatly this didn't fix it, but its 
one I may revisit.

I played around with various XFree settings, and found that out of all 
the options I could play with (okay, about 4 or 5) only one actually 
fixed the problem. Unfortuantly it had the side effect of making X slower 
than a dead slug ... yep, you've guessed it, the option that fixed it is 
the one that switches off any hardware acceleration...so after a few days 
of seeing if I could cope, I decided I couldn't and removed the line that 
read "Option "NoAccel"".

So I did a kernel build - and wondered why during the build, "make 
depend" was spewing out stuff with "smp_" symbols. One CTRL-C and make 
xconfig later, SMP support was disabled and I tried again. This also 
failed to fix the problem, so I was left with kernel 2.2.19 sitting there 
and an XMMS that went clickity click still during heavy xterm output.

The final thing so far tried, is changing the kernel ... faced with a 
choice ... downgrade to 2.2.14, which is tried and trusted and I've ran 
since it came out up until the weekend when I reinstalled, or take the 
plunge and upgrade to 2.4.3.

I decided to act like a fool and plunged in with 2.4.3...after fiddling 
around with the new kernel options, it came up and ... it hadn't fixed 
the problem either, but it seems slightly better and less suscetible to 
small amounts of output (ie, just pressing enter once or twice, or small 
scrolls in vi).

I might drop down to 2.2.14 and do a comparison ... a less heavy-weight 
kernel could affect things quite a bit. Also, I seem to recall that 
2.2.14 was the last kernel they had good performance out of 
(comparatively, in the 2.2 series), so it'll be an interesting benchmark.

No time to look into it at weekend, as I'm away, so instead it'll prolly 
be start of week when I'll invade comp.os.linux.misc, comp.windows.x and 
the XFree mailing lists for ideas, hints, tips, and solutions, or lack 
thereof of this problem.

Chris...

-- 
Chris Johnson            \  "If not for me then, do it for yourself. If 
not
sixie@nccnet.co.uk        \  for then do it for the world." -- Stevie 
Nicks
www.nccnet.co.uk/~sixie/   ~---------------------------------------+
Redclaw chat - http://redclaw.org.uk - telnet redclaw.org.uk 2000   
\______


---------------------------------------------------------------------
Sheffield Linux User's Group - http://www.sheflug.co.uk
To unsubscribe from this list send mail to
- <sheflug-request [at] vuw.ac.nz> - with the word 
 "unsubscribe" in the body of the message. 

  GNU the choice of a complete generation.