Puppy 2.12: 3D-Control-Center V2.00

[withasong]

You asked whether I installed the Kernelmodule and I did install it first. Then I installed the 3rd item in the list, the OpenGL. Then I installed the 2nd item in the list, the Graficscard-drivers. I did it in the order you indicated in the comment area of 3DCC.

Yes, I do have the radeon.ko file in the folder structure you indicated, and here's what the console window shows:

sh-3.00# depmod
sh-3.00# modprobe radeon
WARNING: Error inserting drm (/lib/modules/2.6.18.1/kernel/drivers/char/drm/drm.ko): Unknown symbol in module, or unknown parameter (see dmesg)
FATAL: Error inserting radeon (/lib/modules/2.6.18.1/kernel/drivers/char/drm/radeon.ko): Unknown symbol in module, or unknown parameter (see dmesg)
sh-3.00#

[MU]

these errors should not appear, I'm puzzled
According to your log, you have the correct kernel.
Anyone has an idea?

Please run and post the results:
modinfo drm
modinfo radeon

Mark

[withasong]

Mark, here's the info you requested in the last post:

sh-3.00# modinfo drm
filename: /lib/modules/2.6.18.1/kernel/drivers/char/drm/drm.ko
author: Gareth Hughes, Leif Delgass, José Fonseca, Jon Smirl
description: DRM shared core routines
license: GPL and additional rights
vermagic: 2.6.18.1 mod_unload 486 gcc-3.4
depends: agpgart
parm: debug:Enable debug output (int)
parm: cards_limit:Maximum number of graphics cards (int)
sh-3.00# modinfo radeon
filename: /lib/modules/2.6.18.1/kernel/drivers/char/drm/radeon.ko
author: Gareth Hughes, Keith Whitwell, others.
description: ATI Radeon
license: GPL and additional rights
vermagic: 2.6.18.1 mod_unload 486 gcc-3.4
depends: drm
parm: no_wb:Disable AGP writeback for scratch registers
(int)
sh-3.00#

I tried an experiement with a new burn of a 2.12 ISO, on which I tried to install the Ati-fglrx-8.28.8 driver on my 9500-based computer. I followed the instructions in the comment area of 3DCC and installed #1 and #3, then went to #2 for the Ati-fglrx-8.28.8 install. At that point, I noticed when I clicked on Ati-fglrx-8.28.8 that I shouldn't have installed #3 (the OpenGL drivers), but I didn't know that until after I had already installed them and it was too late at that point anyway. So I went ahead and tried to install Ati-fglrx-8.28.8. When I was asked if I wanted to overwrite the already installed files, I said OK, but then the installation responded with "Canceling, OK". I tried it repeatedly a few times with it always canceling on the overwrites, then decided maybe the OK and Cancel buttons were crossed (which I knew was a VERY longshot), and when I clicked Cancel, it just gave me another "Canceling, OK". Did I get this strange response because I installed #3, the OpenGL's, or is it something else? I could try another ISO burn, skipping #3 in the 3DCC list (since I now know not to do that), but if that doesn't work, I wonder why I can't install the Ati-fglrx-8.28.8 driver with my 9500 (NOTE: I didn't try installing any other driver with 3DCC before I tried to install Ati-fglrx-8.28.8 on this fresh 2.12 ISO burn). Should I just try another ISO burn, skipping #3 before I try #2's Ati-fglrx-8.28.8 -- or is that a waste of time because the canceling is related to some other issue (unknown at the moment)? Again, this was just an experiement to see if the Ati-fglrx-8.28.8 might resolve some issues.

BTW, on one of my other computers with the 8500 ATI Radeon, I installed 3DCC, installed steps #1, #3, and then #2 with the r200 driver which is appropriate to the 8500 series, and when I rebooted I got a blank, black screen and nothing more (and I did try to reboot a few times by pressing the computer's reset button, but the result was always the same: I couldn't see any desktop after the bootup cycle was completed). When I rebooted and stepped back one session on my multisession disk, I got the desktop back. It seems at the moment like the r200 can't be used with the 8500 (another BTW, it's an 8500 AIW, All-In-Wonder, if that makes any difference). I did note when I was trying to install the Ati-fglrx-8.28.8 driver on my 9500-based computer that 8500-AIW's are listed as included in the Ati-fglrx-8.28.8 driver; so perhaps on another day I'll try the Ati-fglrx-8.28.8 with my 8500 -- but right now, it appears that the r200 is incompatible with my 8500.

A last BTW, I wonder if in some future upgrade of 3DCC, whether the order of installing #1, #3, and then #2 could be changed to a top-down sequence by switching the arrangement of OpenGL and Graficdrivers (so then it'd be #1, #2 and #3).

Mark, thanks for all the time and attention you're giving to all the strange things that are happening here at my end of the world. You make me feel not alone, and that's SO appreciated!

[MU]

ok, the modinfo results show, that you have the correct kernelmodule.
But you get errors trying to modprobe them, so my only idea at moment is, that is really incompatible with your 8500. For whatever reason, I have no idea why.

So ok, try the Ati-driver.
It does not matter, that you installed OpenGL before, as the Ati-dotpup overwrites those files.
At least it should, but you get this strange behaviour, that it refuses to overwrite existing files.
I tried it again now myself (in Muppy006), and I have no problems.
It overwrites my already installed files.

The only idea I have at moment:
there might be a problem with xmessage.
This can happen in theory, when you have a wrong font-setting, or strange locale-setting. Or a broken theme.
It should not happen, but is the only idea I have at moment.

Please type these commands and post the result, maybe they help to find out what is wrong:

ls /root/.gtkrc*
echo $LANG
xmessage -v
Xdialog -v

Mark
Ps: at what end of this world do you live? (sorry, I'm just curious )

[tempestuous]

Mark, does the radeon module depend on the drm module?
Run "lsmod" to see what modules are loaded. We should see "agpgart" and "intel-agp/via-agp/..." as well.

It's feasible that the manuafacturer of that graphics card "messed" with the pci-id of the ATi chip, so the radeon module refuses to load because it thinks the hardware is incompatible.

As a last resort, try to force the module to load, like this -

modprobe -f radeon

Check for error messages, then run lsmod to see if it loaded.

[MU]

Mark, does the radeon module depend on the drm module?

yes. The relevant lines from "lsmod":

Code: Select all

radeon                110048  1 
drm                    68324  2 radeon

radeonfb              107008  0 
i2c_algo_bit            9144  1 radeonfb
i2c_piix4               8380  0 
i2c_core               21408  3 radeonfb,i2c_algo_bit,i2c_piix4
intel_agp              22668  1 
agpgart                34608  2 drm,intel_agp

I don't know what radeonfb is good for.

I wonder if "modprobe agpgart" will work for withasong?

Mark

[withasong]

Sorry to have been away so long (though it's only been a day). My problem was that I had SO many CDs with 2.12 test runs on my video cards (there were more than a dozen), and then I added more with 2.13beta test runs, that I couldn't tell which ones were which any more and though I kept notes, I just couldn't go on with what I had been doing. So, I started with a whole new group of CDs with freshly burned ISO's so I could sort things out.

On my 9500-based computer, I again (and again) installed 3DCC, doing steps #1, then #3, then #2, and I used the r300 driver. I had to do it again and again since I would get nutty multisession total blackout's when I tried to reboot after installing the r300 driver (but even after that worked, I'd still get total blackout's just doing other things too -- which is SO frustrating since everything that was done has to be repeated repeatedly since nothing is saved when I click on reboot and everything just goes totally blank and black with nothing further happening). Anyway, frustration aside (yeah, sure!), I got similar results in both 2.12 and 2.13beta on some things when things did happen successfully (but I don't mean I got the results I wanted here). I did some 2.12 and 2.13beta tests on both my 9500 and 8500-based computers today. Right now I'm in 2.12 with the r300 driver and want to start responding to some of the questions that were posed.

Mark, you asked me to do the following:

ls /root/.gtkrc*
echo $LANG
xmessage -v
Xdialog -v

I got:

sh-3.00# ls /root/.gtkrc*
ls: /root/.gtkrc*: No such file or directory
sh-3.00#
sh-3.00# echo $LANG
en_US
sh-3.00# xmessage -v
sh-3.00# Xdialog -v
2.1.1
sh-3.00#

When I entered "xmessage -v" I got a box with a "-v" in it and an okay button. I confirmed the "No such file or directory" to the ".gtkrc" in the root folder with uxplor, and indeed, there's no such thing there. To confirm I typed it correctly, you asked me to type "LS" as the command but in lower case, correct?...and there's a dot in front of the "gtkrc*" correct?

Since I'm afraid of instability problems (I've had them all day in both 2.12 and 2.13beta AND on both computers, and it's evening here now), so I'm going to post this that I have written so far, and I'll come back to tempestuous' posting, and then back to Mark's posting which was done after that, and then finally with some other stuff (but I might poop-out energetically and need to finish some of this tomorrow morning, since I've been at this since this morning -- so stay tuned if I don't get through all these itemized responses until a bit later). I've been at this testing stuff for more than 8 hours already today, and, yikes, I'm getting tired (not lacking in enthusiasm, just lacking in energy)!

[withasong]

Tempestuous, here's responses to what you asked me to do:

sh-3.00# modprobe -f radeon
WARNING: Error inserting drm (/lib/modules/2.6.18.1/kernel/drivers/char/drm/drm. ko): Unknown symbol in module, or unknown parameter (see dmesg)
FATAL: Error inserting radeon (/lib/modules/2.6.18.1/kernel/drivers/char/drm/rad eon.ko): Unknown symbol in module, or unknown parameter (see dmesg)
sh-3.00# lsmod
Module Size Used by
usb_storage 83904 0
snd_mixer_oss 17328 0
snd_seq_midi 8608 0
snd_emu10k1_synth 6800 0
snd_emux_synth 35760 1 snd_emu10k1_synth
snd_seq_virmidi 6896 1 snd_emux_synth
snd_seq_midi_event 6960 2 snd_seq_midi,snd_seq_virmidi
snd_seq_midi_emul 7344 1 snd_emux_synth
snd_seq 48592 5 snd_seq_midi,snd_emux_synth,snd_seq_virmidi,snd_ seq_midi_event,snd_seq_midi_emul
lp 12520 0
parport_pc 32100 1
parport 35080 2 lp,parport_pc
usbhid 40736 0
radeonfb 107008 0
i2c_algo_bit 9144 1 radeonfb
emu10k1_gp 3696 0
gameport 14904 2 emu10k1_gp
snd_emu10k1 112864 1 snd_emu10k1_synth
snd_rawmidi 24064 3 snd_seq_midi,snd_seq_virmidi,snd_emu10k1
snd_ac97_codec 90448 1 snd_emu10k1
snd_ac97_bus 2128 1 snd_ac97_codec
snd_pcm 77592 2 snd_emu10k1,snd_ac97_codec
snd_seq_device 7996 6 snd_seq_midi,snd_emu10k1_synth,snd_emux_synth,sn d_seq,snd_emu10k1,snd_rawmidi
snd_timer 23540 3 snd_seq,snd_emu10k1,snd_pcm
snd_page_alloc 10008 2 snd_emu10k1,snd_pcm
snd_util_mem 4560 2 snd_emux_synth,snd_emu10k1
snd_hwdep 9076 2 snd_emux_synth,snd_emu10k1
snd 51396 11 snd_mixer_oss,snd_emux_synth,snd_seq_virmidi,sn d_seq,snd_emu10k1,snd_rawmidi,snd_ac97_codec,snd_pcm,snd_seq_device,snd_timer,sn d_hwdep
soundcore 9408 1 snd
sis900 22288 0
mii 5680 1 sis900
i2c_sis96x 5460 0
i2c_core 21408 3 radeonfb,i2c_algo_bit,i2c_sis96x
fuse 44388 2
unionfs 77696 1
nls_iso8859_1 3984 0
nls_cp437 5648 0
sr_mod 17412 0
ide_cd 40100 0
cdrom 39424 2 sr_mod,ide_cd
ehci_hcd 30200 0
ohci_hcd 20564 0
uhci_hcd 23356 0
usbcore 126116 6 usb_storage,usbhid,ehci_hcd,ohci_hcd,uhci_hcd
sh-3.00#

[withasong]

Mark, here's what you suggested I try after tempestuous' posting:

sh-3.00# modprobe agpgart
FATAL: Module agpgart not found.
sh-3.00#

Up until now, I've been using the r300 video driver and reporting results on my 9500-based computer.

I'll have to hold the results on the ATI-fglrx-8.28.8 driver on 2.12 and on 2.13beta until tomorrow morning because I'm just too pooped to pop, but I'll preview the results as saying that I got dis-similar outcomes (which surprised me) -- AND I'm talking about both versions of Puppy on both my 9500-based computer and ALSO my 8500-based computer (I've been busy)!

Goodnight for now (yawn!)...time for dinner and a movie, then zzzzzzz

[Gn2]

Unknown symbols usually relates to module version dependency mis-matches.
OR it can be bad re-compiles of modules only (vs full kernel re-compiling)

If still having problems - try downloading kernel source (match the Puppy version number)
Then try reinstalling ATI proprietary drivers as per their recommendations for individual chipsets.
Use ATI instructions for configuring all -

User supplied Puppy wizards cannot cover all hardware variations
Esp if the packager didn't have access to the hardware to test all before uploading.
When additional enhancements (3d accell etc) are not needed - "generic" Vesa modelines work great.

Puppy has frame buffer choices (kernel makefile support * then subsequent boot probes) issues still not fully resolved !

For your expressed wants - you may need to bite_the_ bullet & really get into the inners for your own systems.
Do not expect one Cfg will work for all hardware.
Nice if it does - seldom are we are so lucky ?

[tempestuous]

sh-3.00# modprobe agpgart
FATAL: Module agpgart not found.
sh-3.00#

There's your problem. The proprietary graphics driver won't work without the agpgart module. And the opensource driver might also improve when it's loaded, too.

If your system can't find the agpgart module, it means it can't access the zdrv file. You should be using the full puppy-2.12-seamonkey-zdrv.iso

[Gn2]

The AGP support should be Cfg'd into supplied kernel image -
Proprietary AGP settings/driver MAY be used - one or the other not both.
AGP support should NOT be modularized - as such any (kmod) insmod or probe commands will fail & a lsmod will not (cannot) find.

This has been noted & posted B4

PCI & newer PCI express cards do not use AGP specs.
When AGP card cannot use BIOS settings - it is supposed to default back to old PCI standards.
See the kernel makefile options & help menus for usage .
Google AGP

Code: Select all

CONFIG_AGP=y
# CONFIG_AGP_ALI is not set
# CONFIG_AGP_ATI is not set
# CONFIG_AGP_AMD is not set
# CONFIG_AGP_AMD64 is not set
# CONFIG_AGP_INTEL is not set

# CONFIG_AGP_NVIDIA is not set  <--NOTE - using Nvidia proprietary support 

# CONFIG_AGP_SIS is not set
# CONFIG_AGP_SWORKS is not set
CONFIG_AGP_VIA=y
# CONFIG_AGP_EFFICEON is not set
CONFIG_DRM=y

FRAME BUFFER OPTIONS:

Code: Select all

# Graphics support
#
CONFIG_FB=y
CONFIG_FB_CFB_FILLRECT=m
CONFIG_FB_CFB_COPYAREA=m
CONFIG_FB_CFB_IMAGEBLIT=m
# CONFIG_FB_MACMODES is not set
CONFIG_FB_MODE_HELPERS=y
# CONFIG_FB_TILEBLITTING is not set
# CONFIG_FB_CIRRUS is not set
# CONFIG_FB_PM2 is not set
# CONFIG_FB_CYBER2000 is not set
# CONFIG_FB_ARC is not set
# CONFIG_FB_ASILIANT is not set
# CONFIG_FB_IMSTT is not set
# CONFIG_FB_VGA16 is not set
CONFIG_FB_VESA=m
# CONFIG_FB_VESA_STD is not set
CONFIG_FB_VESA_TNG=y
CONFIG_FB_VESA_DEFAULT_MODE="640x480@60"
CONFIG_VIDEO_SELECT=y
# CONFIG_FB_HGA is not set
# CONFIG_FB_S1D13XXX is not set

CONFIG_FB_NVIDIA=m  <----NOTE : That was accidently missed -necessitating modules/dep file edit
 to comment out "nvfb" line for   Nvidia propreitary use.!
(faster than re-compiling modules just to correct oversight when newer proprietary kernel wrapper & GLX was installed) 

CONFIG_FB_NVIDIA_I2C=y
# CONFIG_FB_RIVA is not set
# CONFIG_FB_I810 is not set
# CONFIG_FB_INTEL is not set
# CONFIG_FB_MATROX is not set
# CONFIG_FB_RADEON_OLD is not set
# CONFIG_FB_RADEON is not set
# CONFIG_FB_ATY128 is not set
# CONFIG_FB_ATY is not set
# CONFIG_FB_SAVAGE is not set
# CONFIG_FB_SIS is not set
# CONFIG_FB_NEOMAGIC is not set
~ 
~  
# CONFIG_FB_VIRTUAL is not set

[tempestuous]

Sorry withasong,
I got my facts reversed: all of the opensource 3D Xorg drivers need the agpgart module.
But the proprietary 3D drivers sometimes don't need this module.

But still, the fact that you cannot access modules which come standard with Puppy (in the zdrv file) is a problem you should address.

[Gn2]

XOrg or XFree - do not have own drivers -Those are X-server data files for any GUI Apps/tools .
They use generic 2d Vesa standards & generic drivers unless proprietary drivers are installed.
3d & other capabilities are enabled only if enhanced driver support is added via proprietary drivers.
DRI - DRM - GLX are Req'd for additional benefits
Dri is not used for Nvidia proprietary kernel GLX wrappers
Mesa is a generic standard

The separate or included Puppy (iso versions) zdr drivers are extras
If AGP support is compiled into kernel (modprobe agp) can not work (unless modprobe is now a Puppy script instead of the command)
= AGP will not be a module.

Code: Select all

# modprobe agp
FATAL: Module agp not found.

= It was not modularized

Code: Select all

~ # dmesg |grep agp
Linux agpgart interface v0.101 (c) Dave Jones

GENERIC makefile AGP support should be compiled in
Specific card support should NOT - as the frame buffer options - some MAY (and already have - refer to other posts) cause hardware conflicts .
Makefile specific card support is OK - IF it does not conflict to later added support
That is up to user - often requires kernel sources for needed headers. (wrapper install run shell probes for them)

If bootup procedes - and text mode available if/when desktop Cfg fails > Post_install edits may be used.
Otherwise - the user is out of luck - there is then no way to add or
correct initrd support unless a re-master is done.
AFAIK (tried) It cannot be accomplished in (puppy pfix=rdsh mode)

HTH

[MU]

withasong, it seems you downloaded the small version of Puppy, that does not include zdrv_212.sfs.
This addon includes "agpgart", that is needed by the free drivers for agp-graficscards.
Please download and save it in /mnt/home:
http://puptrix.org/isos/Puppylinux-offi ... rv_212.sfs

Then reboot, and type:
modprobe agpgart

If you get no errors, the 3DCC might work now with the rxx drivers.
If you still get errors, please report them again.

###########

Concerning your problem with the Ati-dotpup not overwriting files, I have still no idea what could be the reason, but I will create a workaround (alien package) tomorrow.
At moment my jaw hurts too much.

wait... please try this:
http://dotpups.de/dotpups/System_Utilit ... g-gtk2.pup
This replaces xdialog with a different version, what might fix this issue.

Mark

[withasong]

Thanks Mark, tempestuous, and Gn2 for all your great help thus far. You've been SO patient and helpful!

Last night as I was about to fall asleep, I had a thought that almost work me up (but not quite, but I did remember when I got up): I wondered what my BIOS settings are on the computer with the 9500 Pro ATI Radeon video card.

This morning, first thing, I checked (and it was in an area of my BIOS on my 9500-based computer that I had just never needed to check before). I found that the "Primary Graphics Adapter" was set to PCI (tuh duh!, I thought, this is the problem why I keep seeing my AGP board sometimes listed as PCI); so I set it to AGP, saved the setting, and rebooted with a Puppy 2.12 that I had last used (BTW, every version of 2.12 and 2.13beta that I have been using for some time is the zdrv ISO's consistently).

I found no change on the multisession I had used with the previous BIOS setting (rats!), so I figured that I probably needed to reinstall everything with the new BIOS settings in order to get any difference in the installed results, so I burned two fresh ISO 2.12's (one for r300 and one for ATI-fglrx) and two 2.13beta's (one for r300 and one for ATI-fglrx also). Again, these ISO's are all zdrv versions of Puppy! It's taken me over 4 hours to try out all the permutations to just to get to the results below (and I still keep getting reboot blackout's which means nothing gets saved from my session just run, which didn't help me timewise at all!):

9500-based Computer:

Bad news related to DRI: On all versions (2.12 and 2.13beta with either r300 or ATI-fglrx), DRI on the 3D demo is still NOT active and I still get 14.29 fps uniformly.

Bad news with r300: MPlayer plays avi's, mpg's, etc., with NO video display but the sound is fine. gXine starts to play DVDs with sound with NO video display too. Again, r300 displays NO video in either player, but the sound is just great.

Bad news with ATI-fglrx and MPlayer: I get "Error opening/initializing the selected video-out (-vo) device" for all avi's, mpg's and DVDs in both 2.12 and 2.13beta uniformly with MPlayer in GUI mode. Again, uniform "Error..." with no play attempted in MPlayer no matter what video format I try with ATI-fglrx as the driver.

Good news with ATI-fglrx and gXine: On both 2.12 and 2.13beta, gXine now plays DVDs with video and sound! This is a positive change since I reset the BIOS from PCI to AGP for the Primary Graphics Adapter! This is the first time I have seen video with my 9500 in Puppy!

More about MPlayer with ATI-fglrx: I've been using the Unichrome version of MPlayer ever since Mark first told me about it, and on one of the last installs (gee, I've installed this 20 or more times now), I actually read the pop-up thing that I thought was licensing stuff; so I decided to try the non-GUI approach with the ATI-fglrx driver that is now playing DVDs in gXine but on MPlayer instead. The pop-up says to play DVDs in MPlayer with this command line:

mplayer dvd://1 -vo xvmc -vc ffmpeg12mc -fs

So, I gave it a try and got this:

sh-3.00# mplayer dvd://1 -vo xvmc -vc ffmpeg12mc -fs
MPlayer 1.0rc1-3.4.4 (C) 2000-2006 MPlayer Team
CPU: AMD Athlon(tm) XP 1800+ (Family: 6, Model: 6, Stepping: 2)
MMX2 supported but disabled
SSE supported but disabled
3DNow supported but disabled
3DNowExt supported but disabled
CPUflags: MMX: 1 MMX2: 0 3DNow: 0 3DNow2: 0 SSE: 0 SSE2: 0
Compiled for x86 CPU with extensions: MMX

Playing dvd://1.
There are 14 titles on this DVD.
There are 5 chapters in this DVD title.
There are 1 angles in this DVD title.
audio stream: 0 format: ac3 (stereo) language: en aid: 128.
number of audio channels on disk: 1.
number of subtitles on disk: 0
MPEG-PS file format detected.
VIDEO: MPEG2 720x480 (aspect 2) 29.970 fps 9800.0 kbps (1225.0 kbyte/s)
vo_xvmc: X-Video extension 2.2
vo_xvmc: No X-Video MotionCompensation Extension on :0.0
Error opening/initializing the selected video_out (-vo) device.

Exiting... (End of file)
sh-3.00#

Gee, the error line near the bottom is what I've got uniformly in the GUI too for some time now, and I get it in the command line version too. However, there's some kind of interesting stuff that seems disabled near the top of the listing, but I don't know if it applies to videos (that's why I'm showing it in this posting). I checked the config and gui.config files and see that both have "xv" like tempestuous suggested; so I'm not sure why avi's, mpg's, and DVDs don't play in MPlayer even though they do now play in gXine. Hmm...I just don't know on this, but thought I'd report it.

8500-Based Computer:

I know I said this before, but I want to say it here again: I tried the r200 driver on my 8500, since the r200 is appropriate for it, but all I got was no display at all in Puppy with the r200 (a totally blank desktop, even with repeated reboots), until I stepped back to a multisession before I installed the r200 (then the desktop reappeared). I wonder why the r200 won't work with my 8500 AIW?

So, on my 8500, I tried the ATI-fglrx (since it had positive results with gXine on my 9500) and found DRI was now ENABLED in the 3D demo on my 8500! I got 50 fps this time! That's another improvement !

BTW, on my 8500, MPlayer plays avi's, mpg's, and DVDs with the standard video setup that is configured when Puppy 2.12 and 2.13beta are both first installed (I don't know how this compares to r200 or r300 or ATI-fglrx). This is SUCH a contrast to my 9500 computer (I just haven't had the time to experiment with my 9600 computer, my 3rd computer, since I've been spending SO much time on my troubled 9500 computer, but as I remember, my 9600 has no trouble just like my 8500 computer with MPlayer playing everything too)!

Some Remaining Issues:

Why doesn't MPlayer play videos on my 9500-based computer when MPlayer plays videos in all formats on my 8500?...AND why doesn't MPlayer play DVDs when gXine plays them with the ATI-fglrx driver on my 9500?

Why is the video display slightly degraded with my 9500 in gXine with the ATI-fglrx driver over what I get in Windows? I switched to Windows and played the same DVDs I played with gXine since it started working, and the display in Windows is so wonderfully fine-pitched and movements have no horizontal-line after-images or jerkiness with my 9500. I know it's got to be driver related. It's certainly not a Linux problem, because on my older video card, my 8500 (which is SO much older than my 9500), Puppy Linux shows great videos with BOTH gXine AND MPlayer (nearly as good as Windows).

I've often wondered what is the default graphics standard in Puppy and how does it compare to the options in 3DCC if I've not installed 3DCC or any of the drivers in it? Is it like the r300 or ATI-fglrx, or what?

Why is DRI still NOT active no matter what BIOS settings or video driver I use on my 9500-based computer?

Now that I've got my AGP BIOS'd right, are there any tests I should re-run?

PS: Mark, the overwriting problem only occurred in 2.13beta with the ATI-fglrx driver when my BIOS was set to PCI instead of AGP. I repeatedly tested this yesterday with careful note-taking when my BIOS was mismatched to my AGP card (but I didn't know it was mismatched at that point) and in 2.12 the overwriting problem didn't happen, but in 2.13beta it did happen repeatedly. Now, today, ATI-fglrx overwrites files in 2.13beta with the right BIOS setting without self-canceling. Gee, did 24 hours go by; it's almost time for dinner and a movie again...

PPS: Mark, New Castle, California

[Gn2]

W.A.S.

Up to you - but it is suggested:
Work on ONE system at a time - get each configured correctly - do not confuse/mix issues to varied hardware.
As to this thread > It is now so mixed up WITH wrong uploads, failed compiles, missing dependencies .....no hardware to test
3d - but for for ATI only ?

Including & mixing up old generic 2d drivers, Mesa - to varied Radeon. Mach. Rage etal chipsets ....
versus 3d which it was supposed to enable - or why not just use generic drivers & VESA standards.

The sure way - which could have been done in less time -
Download kernel sources, install to either default /usr/src/linux <--symlink to linux_version_xx
OR create a working directory anywhere in path (run to check path)

Code: Select all

 env

Download recomended ATI driver for specific chipset.
Run ATI installer
DRI must be used for ATI
See your /etc/X11/xorg.conf (settings)

To get most out of O/System -Download - Re-compile Kernel

To get most out of CPU & any App - re-compile with proper Gcc Flag Optimizations Taken from HERE
AND > Gcc Manuals

To find CPU capability

Code: Select all

cat /proc/cpuinfo

Amongst other data will be optimization flags that may be set in Gcc upper-level makefile
Example

flags:: fpu vme de pse tsc msr pae mce cx8 sep mtrr pge
mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow ts

Not all Apps can use all- these are OPTIONAL dependency flags.

(The reason a sources-based O/System is more customizable/powerful vs any binary-based platform),
did you read the supplied URLs Esp ATI Release
For most ANY Linux Related Info

[tempestuous]

That's good news.
Now that your bios is in a better state, for both your 8500 and 9500 setups I would still like to know if the agpgart module has successfully loaded (and also the xxx-agp module). During the development of the first Puppy2 version, Mark and Barry worked hard to get these modules to load automatically at boot up.
Run "lsmod" to find out. If it's not listed, you need to fix this.

Sorry to have to ask this question, but during boot up, you chose Xorg, not Xvesa, right?
I wonder if your Xorg driver is falling back to the generic "vesa" (not Xvesa) particularly with your 9500. This would explain the modest performance, and output problems with MPlayer.
What is listed in Section "Device" > Driver in your /etc/X11/xorg.conf file?

The pop-up says to play DVDs in MPlayer with this command line:
mplayer dvd://1 -vo xvmc -vc ffmpeg12mc -fs

Well, I wrote that popup. If you read it more carefully, you will see that it only applies to graphics cards which support "xvmc" video output. ATi cards are not compatible with this output type.
There is a detailed thread on this subject http://www.murga-linux.com/puppy/viewtopic.php?t=13216

[withasong]

Thanks tempestuous for your response and questions.

I ran lsmod after I ran modprobe agpgart, and here's the output:

sh-3.00# modprobe agpgart
FATAL: Module agpgart not found.
sh-3.00# lsmod
Module Size Used by
usb_storage 83904 0
snd_mixer_oss 17328 0
snd_seq_midi 8608 0
snd_emu10k1_synth 6800 0
snd_emux_synth 35760 1 snd_emu10k1_synth
snd_seq_virmidi 6896 1 snd_emux_synth
snd_seq_midi_event 6960 2 snd_seq_midi,snd_seq_virmidi
snd_seq_midi_emul 7344 1 snd_emux_synth
snd_seq 48592 5 snd_seq_midi,snd_emux_synth,snd_seq_virmidi,snd_seq_midi_event,snd_seq_midi_emul
lp 12520 0
parport_pc 32100 1
parport 35080 2 lp,parport_pc
usbhid 40736 0
radeonfb 107008 0
i2c_algo_bit 9144 1 radeonfb
emu10k1_gp 3696 0
gameport 14904 2 emu10k1_gp
snd_emu10k1 112864 1 snd_emu10k1_synth
snd_rawmidi 24064 3 snd_seq_midi,snd_seq_virmidi,snd_emu10k1
snd_ac97_codec 90448 1 snd_emu10k1
snd_ac97_bus 2128 1 snd_ac97_codec
snd_pcm 77592 2 snd_emu10k1,snd_ac97_codec
snd_seq_device 7996 6 snd_seq_midi,snd_emu10k1_synth,snd_emux_synth,snd_seq,snd_emu10k1,snd_rawmidi
snd_timer 23540 3 snd_seq,snd_emu10k1,snd_pcm
snd_page_alloc 10008 2 snd_emu10k1,snd_pcm
snd_util_mem 4560 2 snd_emux_synth,snd_emu10k1
snd_hwdep 9076 2 snd_emux_synth,snd_emu10k1
snd 51396 11 snd_mixer_oss,snd_emux_synth,snd_seq_virmidi,snd_seq,snd_emu10k1,snd_rawmidi,snd_ac97_codec,snd_pcm,snd_seq_device,snd_timer,snd_hwdep
soundcore 9408 1 snd
sis900 22288 0
mii 5680 1 sis900
i2c_sis96x 5460 0
i2c_core 21408 3 radeonfb,i2c_algo_bit,i2c_sis96x
fuse 44388 2
unionfs 77696 1
nls_iso8859_1 3984 1
nls_cp437 5648 1
sr_mod 17412 0
ide_cd 40100 0
cdrom 39424 2 sr_mod,ide_cd
ehci_hcd 30200 0
ohci_hcd 20564 0
uhci_hcd 23356 0
usbcore 126116 6 usb_storage,usbhid,ehci_hcd,ohci_hcd,uhci_hcd
sh-3.00#

I didn't notice agpgart in the list from the lsmod, besides the modprobe says it's not found anyway (but I wanted to provide what I found). You told me I would need to fix this as though I know how to do such a thing, yet I don't think I know how to fix such a problem yet (I'd like to say "Yeah, I know what to do" -- but I don't). If it's easy to say how to fix it, I'd love to know; but if it's not easy, I'll understand if I get no feedback.

Yes, during the original install of Puppy, I certainly did choose the Xorg configuation, but now that I have installed the ATI-fglrx that is part of 3DCC, I see "vesa" listed as my driver in the Graficsdriver's panel of 3DCC (whereas I usually see "ati" before I install any drivers in 3DCC). On all 3 of my computers, I always see "vesa" as a result of the ATI-fglrx installation within the 3DCC driver section.

You ask "what is listed in Section "Device" > Driver in your /etc/X11/xorg.conf file", and I see two sections marked as Device, so here's what I get from near the end of the first Device through the following Device (they both contain some interesting looking stuff to me):

...
Driver "vesa" #card0driver
Identifier "Card0"
Driver "ati"
VendorName "ATI Technologies Inc"
BoardName "R300 AD [Radeon 9500 Pro]"
BusID "PCI:1:0:0"
EndSection

Section "Device"
Identifier "aticonfig-Device[0]"
Driver "fglrx"
EndSection

I note with puzzlement that the BusID still says PCI in the first Device section above, though I rechecked my BIOS setting before I booted up today and it is still set from yesterday as AGP, which is the state I used to make the current installation with the 3DCC device driver it says is installed in the second Device section above. Again, the current installation was done after I set my BIOS to AGP! Since BIOS is absolutely set to AGP now, I don't understand the persistence of PCI in these reports (but it's got to be part of what is a problem here in Linux -- but not Windows). I've done the BIOS, I've checked to be sure I'm in an AGP slot (and yup, the card can't possibly fit in a PCI slot because it's totally shaped differently), and I'm open to any suggestions here.

About the popup you wrote, please don't take offense if I don't understand the terms I read since I'm still new to Linux lingo. A careful reading, as you suggest, doesn't benefit me if I don't know what things mean; it just makes me feel more stupid and not informed. However, I did remember you saying to me that I must be sure to use "xv", and when I read " XvMC support for VIA Unichrome graphics" is what this version of the Unichome MPlayer was setup for according to the popup, I thought it might be my case (not knowing what the "MC" added to "Xv" was meaning -- cause I just don't know these things). Because I didn't understand, I tried both the GUI approach and the command line approach, just in case something might work or it might yield something. Now that I have the thread you provided, I do have more information to help me sort out what was said in the popup (but I didn't have the benefit of your kindly provided link previously -- and thanks for it). I trust that my trying both approaches hurt nothing and I will indeed try to be more careful in the future. Please be patient with my lack of understanding and experience; I'll improve in time.

On Dec. 29th at 12:11 p.m. (according to the info I see in the posting at my end today), Mark asked that I download zdrv_212.sfs and place it in /mnt/home (and since I found no "home" folder in "mnt", I had to create it in "mnt"). I have finished the download and I am willing to give it a try, though like I said, I've only used the puppy-2.12-seamonkey-zdrv.iso for all the tests I've run recently. I'm for anything to work and will try whatever I'm asked to do, especially since I'm doing everything multisession, so nothing of lasting consequence can go wrong (but I kind of wonder about the placement of this driver in a folder I had to create, though I confess I don't know about these things). I'll reboot in a few moments after I place this posting.

Armed with 3DCC, I ran 6 tests on my 9600-based computer (my 3rd computer) this morning and got some interesting results that I will post separately (the 3DCC demo and drivers provide a wealth of varied outcomes!). I really appreciate the help all 3 of you (tempestuous, Gn2, and Mark) are providing to me, but I fear Gn2's remark about this thread meandering away from 3DCC into my personal ATI problems may be right (though 3DCC and its drivers are center-stage in what I'm testing right now). I did start in another thread that I started, but the link provided by tempestuous to the 3DCC here in this thread really started giving me more help and the tools I needed to get rolling than I was getting previously; so I've just stayed here where the help and tools have been maximal. I don't know how to reweave these threads into something more appropriate. Until guided elsewhere, I'll continue to post my adventures with 3DCC and my video cards here (but I'm certainly open to suggestions!).

[Gn2]

Hi

If it's possible - perhaps a Mod could split your posts & replies to same - into another thread ?
Preferably where it was First Posted
Before "migrating" to here.
It would then keep the 3dcc specific topics separate & dealing only on that utility usage ?
This thread is in Forum Header ANNOUNCEMENTS - no problems or replies to same should be here.
If for no other reason - the obvious:
(A manual search for solutions would be in Beginners/Users/ or Bugs Headers)

In meantime, please try

Code: Select all

find / -name "agpgart" (*and) dmesg | grep agpgart 

You need either the ATI version or more generic AGP support as supplied by
(either initrd compiled in OR ) - As loaded by 3dcc add-ons.
Examples:

LSMOD > amd_k7_agp 8572 1
agpgart 34608 1 amd_k7_agp
---------------------------------------------------------
DMESG Grep >Linux agpgart interface v0.101 (c) Dave Jones
agpgart: Detected AMD 761 chipset
agpgart: AGP aperture is 64M @ 0xd8000000
----------------------------------------------------------------
MODPROBE AGPGART > NIL !

BTW - The only lines needed to paste in lsmod ouput were those relevant to vid modes - all else have no bearing.
The xorg.conf file needs to have ATI as the driver
NOT (vesa) - Yours shows TWO driver lines !
Once the first USEABLE is loaded, next is ignored.
(or desktop startup gets aborted, dumping you back to text mode with a screen error Msg )
which also gets logged in /var/log/Xorg_xx file > May also be in /tmp folder.

In (xorg.conf) "device section" (PCI Bus is OK) AGP is not a separate bus it is an extension to PCI.
It's allowable to have multiple "device" sections - they will be listed as > :0 , then 1,2, etc - which allows for variables in running any desired desktop environment.
It is just then needed to create own .xinitrc file to execute -same as using variables or alternate Desktop or alternate Window Mgrs.
As explained before a WINDOW Mgr is NOT the GUI desktop App E.G.> KDE - Flux - Xfce4- E17, etal.

Once AGP support is enabled in Puppy (BIOS is now OK) the wizard should work - if not, paste in your xorg.conf file
It was not necessary to re-burn ISOs - just start in

Code: Select all

puppy pfix=ram

for pristine state (saving old persistent data file by renaming it for later possible reference)