Re: vesafb & less on console Bug ? and another problem

Petr Vandrovec Ing. VTEI (VANDROVE@vc.cvut.cz)
Wed, 7 Apr 1999 18:12:27 MET-1


Dietmar Kling wrote:
>> Dietmar Kling wrote:
>> > Hello,
>> > When I display a Textfile with less (version 332) larger than
>> > screensize and hit the key 'b' in less (Hint: hit space to scroll
>> > down some pages) less tries to scroll back the text. This takes so
>> > much cpu-power that sound (mpg123) stops until the 50 lines are
>> > scrolled and it needs severals seconds to scroll the lines on
>> > my 350 AMD !!. Every task on the machine seems to stop working as long
>> > as the lines are scrolled back.
> First: Thanks to all which replied.
> My Card is a Matrox G100. The matroxbuffer didn't work for me last time i
> tried it (2.1.128 I believe). It comes sometimes
> up but my fbtv program shows the wrong colors and I am to lazy to evaluate
> what goes wrong.
Hi Dietmar,
G100 works with matroxfb since 2.2.0 (drivers after Jan 11, 1999), multihead
since Jan 21.
fbtv 2.32 works fine in cfb8, 16, 24 and 32 with matroxfb/G100 (&G200, on
Intel, I did not ever try plug Bt848 into Motorola PPC (PReP)).
> Guys, i will try again matroxfb but i insist that scrolling worked for me in
> vesafb (I did work under X for a long so i didn't notice the behaviour on the
> console).
> The only thing i changed beetween 2.1.124 and today was the processor (AMD 200
> -> AMD 350).
> (Ok today i upgraded to 128 MB RAM, kill me).
> Scrolling worked well under the old kernels. When I find time, i will
> recompile some old 2.1.124 and try to prove this. And yes I did _not_ change
> my boot configuration.
Could you try to boot with
video=scrollback:0
to check, whether 2.1.124 (faster) behavior will reappear on 2.2.3+? I think
it could.
> Second Problem ( Or how to kill the ALT-Console Switch)
> I also see a slowdon of fbtv (the console version of xawtv) compared to
> older Kernels.
> ( But this could be a problem of the version of fbtv (2.34), though)
> But I can easily kill the ALT-F1/F2/F3... Console switch.
> Recipe:
> Start fbtv on Console 1
> Switch to Console 2
> Switch back to console 1
> There is ca. 1 second before the tv-screen comes back on my console.
> In that time try to switch back to console 2
> Bingo!
> You killed the alt console Switch.
> fbtv-Screen comes up and you cannot switch to another console anymore.
I'm using fbtv on matroxfb G100 (832x640x32bpp) without problems with fbtv
somewhere around 2.32 (binary does not want to reveal version; 2.37 from
debian-potato does not work, probably because it says `channel XX
(PAL/ntsc-cable-jp)' instead of PAL/pal-europe... (using same kernel
driver 0.6.3)). Console switch time is almost zero.
But I think that lock of console switching is caused by bug in fbtv (there
is another bug - sometime when switching from fbtv screen to another console,
fbtv does not correctly stop Bt848, so that part of picture is written into
new console).
> P.S. For those who want to know this is a AMD-350 Computer with 128 MB RAM a
> Matrox G100 Card
> a Bt848 Card a Soundblaster an Advansys SCSI-Controller, a Sony DAT-Tape some
> SCSI-Disks and a Big Fireball IDE. The machine is running fine,
> Captain :) .
PII350/128MB SDRAM, Bt848-AVerMedia TVPhone, OPTi82C930 sound, AIC7895 SCSI,
8GB SCSI Fujitsu, CDROM. There were problems - machine died hard (no-numlock,
no messages) in few hours (2-20) when grabbing teletext data; after upgrade
to aic7xxx-5.1.13 it grabs data for some days (~50 hrs) without problems,
so I hope that it is fixed now... (problem was not caused by locked bus,
other transfers (playing music) still repeated last ~0.5sec (?64KB) of sound
when crash occured)). Kernel is 2.2.5 - vger990329.
Best regards,
Petr Vandrovec
vandrove@vc.cvut.cz

P.S.: Machine grabs teletext 24hrs/day (unless I'm watching TV), see
http://platan.vc.cvut.cz/cgi-bin/httpttx if you are interested (czech
teletexts only).

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/