Page 8 of 19
savage video-card doesn't work
Posted: Wed 10 Apr 2013, 15:50
by rodin.s
Savage video-driver doesn't work for me anymore. It worked in latest Upup Precise. I have some distorted picture. Managed to switch to vesa by editing xorg.conf. Posting from it now. Maybe my card is too old for this kernel.
Posted: Wed 10 Apr 2013, 17:57
by pemasu
rodin.s. I would need /var/log/Xorg.0.log after failed xwin X start using savage_drv.so
dmesg output would also be useful.
I would like to see if there is something about savage.ko kernel module or savage_drv.so failure in those logs.
I have both of them included in the build...so they should be in use...when you boot without xorg.conf modification. Something clearly is not ok for you.
savage video-card doesn't work
Posted: Wed 10 Apr 2013, 21:13
by rodin.s
I have made a screenshot. But it's much worse than it is on screenshot and pdiag might have all needed information.
Posted: Wed 10 Apr 2013, 21:44
by pemasu
rodin.s
Pdiag does not collect Xorg.0.log. It possibly contains most useful information, maybe you could post it.
I noticed that you dont have drm or savage kernel modules loaded.
Do you have them loaded in Puppies where your savage card works ok.
Also....libgl1-mesa-dri does not provide savage_dri.so anymore so no 3D acceleration from that part also. Mesa-9.0.3 does not seem to include it anymore.
Tested
Posted: Thu 11 Apr 2013, 00:17
by darry1966
Machine AMD Sempron 3000+
Ram 1G
boots nicely and has correct screen setting.
Only thing wrong is with my p/s2 mouse which shows weird mouse pointer on screen. sound works.
savage card
Posted: Thu 11 Apr 2013, 07:53
by rodin.s
I have uploaded Xorg log to my previous post and I have savage and drm modules loaded for example in Precise-5.4.2
Code: Select all
# lsmod | grep savage
savage 20554 1
drm 121458 2 savage
Posted: Thu 11 Apr 2013, 13:40
by pemasu
rodin.s. What happens if you logout to the console state and load manually drm and savage.
modprobe drm
modprobe savage
and then start X again.
savage card
Posted: Thu 11 Apr 2013, 14:24
by rodin.s
Almost same thing.
Raring 3.8.4.2
Posted: Thu 11 Apr 2013, 21:23
by sszindian
Raring-3.8.4.2
Everything tested appears to work well. Graphics are very nice, could be the best of any puppy so far? Fonts are good and acceptable but I had to use the 108 global font size to make them that way on a 1400x900 screen.
One issue (and this is not just for Rearing but also latest Precise, Racy and Slacko) Occasional WiFi 'drops'. Wheezy is the only puppy that I never ever had a drop with (so far).
While this only affects WiFi users, it is an issue as once you get a drop, you're finished with Internet... it just won't re-config... and you can't shut puppy down without a cold- push and hold in the button (and that's for all the above mentioned puppy's.)
It would be nice if WiFi could restart itself after a drop, (some other distro's do have that feature.)
Just my 2-cents...
>>>---Indian------>
Posted: Fri 12 Apr 2013, 12:53
by anikin
Further to make 3.82 version, I did some research but Google didn't have a clear and simple answer for me. All I could find was that this latest version carried 83 bug fixes, which let me cautiously presume, that I'd be good to go with the upgrade. I compiled the new version from source and put it into devx. But since it was compiled with the "old" make, I recompiled it once again with the new one and put finally into devx. And declared the case solved and closed. I was reading this article:
http://users.wowway.com/~zlinuxman/Kernel.htm and quite unexpectedly an answer popped up:
Step 6: Install kernel-package
This is an unofficial patch: it is not provided by or endorsed by the upstream author or the Debian package maintainer. Also please note that kernel-package version 12.036+nmu1 and earlier releases are
not compatible with "make" version 3.82; so don't install "make" version 3.82 or later. These problems are fixed in version 12.036+nmu2. However, if your kernel source code is version 3.3.x or later, even kernel-package version 12.036+nmu2 will need a patch, particularly if CONFIG_LGUEST is set. The patch is available in the problem log for Debian Bug Report 663474
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=663474.
Version 12.036+nmu3 fixes this problem: no patch is required.
Evidently, ubuntu raring has an old version of make for backward compatibility, which leads me to think: compatibility at the price of 83 bug fixes?
Posted: Sat 13 Apr 2013, 14:38
by OscarTalks
Upup Raring 3.8.4.2
Menu > Shutdown > Exit to prompt gives some acpid messages and a blinking _ but not a prompt.
On my nvidia GeForce FX 5200 machine Nouveau loads with no missing graphics but glxgears shows only 75 FPS
If I boot with nouveau.noaccel=1 the FPS is higher at aound 235 FPS which seems strange.
If I want to use the nvidia proprietary driver on this machine I can not use a Puppy with such a recent kernel I guess. Driver 173.14.36 will not install as it did not install with the 3.7.2 kernel.
SlimBoat browser works OK, but really you need to update JWM to latest snapshot. I think this may be quite important as it may have been a bug relating to Qt apps.
Posted: Sat 13 Apr 2013, 15:08
by pemasu
OscarTalks. I think it is question of need to use glxgears with this command before:
export vblank_mode=0
then glxgears command....
About compiling....I think you need to hunt nvidia patches for recent kernels. I had to do that for ati-amd fglrx modern driver also. I had to patch 3 source files before I was able to compile it for Upup Raring.
Now....I have Upup Raring with preinstalled amd_fglrx_catalyst-13.1-k3.8.4-upup. I unsquashed the main sfs, copied ati inside with already configured xorg.conf file and now I have ready build to use it.
I need it for my hdmi connection to the LCD TV, I get video and audio transferred that way with that proprietary driver.
I will check the menu exit to the prompt. I use always CTRL+ALT+BACKSPACE so I havent noticed anything weird.
Posted: Sat 13 Apr 2013, 17:26
by OscarTalks
Pemasu,
CTRL+ALT+BACKSPACE gave the same result on that machine with that save-file.
If I boot pfix=ram the exit to prompt works OK, but something is not quite right.
On 2 different machines I booted pfix=ram, exit to prompt, typed xorgwizard, probe, selected 1024x768x24, typed xwin, got a desktop at that resolution, BUT then if try to exit to prompt again I get a blinking _ with nothing from the keyboard. Only the power button shuts down with a message. Edit, same on a 3rd machine, after running xorgwizard the exit to prompt no longer works.
The export vblank_mode=0 and glxgears does give a higher FPS at around 470, but I tried Google Earth version 7 and it does not display if Nouveau loads in the normal way, but if I boot with nouveau.noaccel=1 it displays although it is slow and jerky compared to when I run the nvidia driver in other Puppies.
Chromium 28 runs as normal and seems OK.
Posted: Sat 13 Apr 2013, 18:19
by pemasu
OscarTalks. I try to narrow the problem. Are all your tested comps using Nvidia graphics ?
And so...using nouveau driver.
Posted: Sat 13 Apr 2013, 18:29
by OscarTalks
Pemasu,
No, the other two are using the intel driver. I only have one computer with an nvidia card.
Later:- Decided to try it on my old slow Pentium 4 machine which uses the SiS driver, but the boot sequence gets to a white cursor in the centre of a black background and it just freezes like that.
Will keep testing.
Posted: Sun 14 Apr 2013, 06:21
by James C
Might be of interest to some....Nvidia Linux driver 304.* and legacy branch support timeframes
http://nvidia.custhelp.com/app/answers/detail/a_id/3142
Posted: Sun 14 Apr 2013, 18:29
by pemasu
I have now tested the xorgwizard caused hanging. It happens to me frequently with my dual graphics integrated intel...and sometimes with ati radeon also when I switch to use it in BIOS.
Okay...it is not kernel dependent. I have tested with 3.8.7 and with 3.7.10. They all cause the same hanging after you use xorgwizard > probe, start X and exit to prompt again.
Luckily I have acpid listening powerbutton and for me it shuts down the computer. No need to hard shutdown.
Anyway exiting to prompt after using xorgwizard probe property is the clue.
It looks like that xorgwizard with X server: X.Org X Server 1.13.3 is not anymore compatible. There is for example command:
Xorg -configure > /tmp/xorgprobe.txt 2>&1 which give error about screens when you execute it in the prompt.
I suppose that Xorgwizard creates somewhat defective xorg.conf file due to incompatibility. And yes....I can create really weird xorg.conf with xorgwizard.
It has several screens, monitors, devices listed, most of them useless.
Code: Select all
Section "Monitor" #sectionmonitor0
Identifier "Monitor0"
VendorName "Monitor Vendor"
ModelName "Monitor Model"
HorizSync 30-49
VertRefresh 56-61
#UseModes "Modes0" #monitor0usemodes
Option "PreferredMode" "1366x768"
EndSection
Section "Modes"
Identifier "Modes0"
#modes0modeline0
EndSection
Section "Monitor"
Identifier "Monitor1"
VendorName "Monitor Vendor"
ModelName "Monitor Model"
EndSection
Section "Monitor"
Identifier "Monitor2"
VendorName "Monitor Vendor"
ModelName "Monitor Model"
EndSection
Section "Device"
### Available Driver options are:-
### Values: <i>: integer, <f>: float, <bool>: "True"/"False",
### <string>: "String", <freq>: "<f> Hz/kHz/MHz",
### <percent>: "<f>%"
### [arg]: arg optional
#Option "NoAccel" # [<bool>]
#Option "SWcursor" # [<bool>]
#Option "EnablePageFlip" # [<bool>]
#Option "ColorTiling" # [<bool>]
#Option "ColorTiling2D" # [<bool>]
#Option "RenderAccel" # [<bool>]
#Option "SubPixelOrder" # [<str>]
#Option "AccelMethod" # <str>
#Option "EXAVSync" # [<bool>]
#Option "EXAPixmaps" # [<bool>]
#Option "ZaphodHeads" # <str>
#Option "EnablePageFlip" # [<bool>]
#Option "SwapbuffersWait" # [<bool>]
Identifier "Card0"
Driver "radeon" #card0driver
BusID "PCI:1:0:0"
EndSection
Section "Device"
### Available Driver options are:-
### Values: <i>: integer, <f>: float, <bool>: "True"/"False",
### <string>: "String", <freq>: "<f> Hz/kHz/MHz",
### <percent>: "<f>%"
### [arg]: arg optional
#Option "ShadowFB" # [<bool>]
#Option "Rotate" # <str>
#Option "fbdev" # <str>
#Option "debug" # [<bool>]
Identifier "Card1"
Driver "fbdev"
BusID "PCI:1:0:0"
EndSection
Section "Device"
### Available Driver options are:-
### Values: <i>: integer, <f>: float, <bool>: "True"/"False",
### <string>: "String", <freq>: "<f> Hz/kHz/MHz",
### <percent>: "<f>%"
### [arg]: arg optional
#Option "ShadowFB" # [<bool>]
#Option "DefaultRefresh" # [<bool>]
#Option "ModeSetClearScreen" # [<bool>]
Identifier "Card2"
Driver "vesa"
BusID "PCI:1:0:0"
EndSection
Section "Screen"
Identifier "Screen0"
Device "Card0"
Monitor "Monitor0"
DefaultDepth 24
#Option "metamodes" "1366x768_60 +0+0" #METAMODES_0
Subsection "Display"
Depth 24
Modes "1366x768"
EndSubsection
EndSection
Section "Screen"
Identifier "Screen1"
Device "Card1"
Monitor "Monitor1"
DefaultDepth 24
#Option "metamodes" "1366x768_60 +0+0" #METAMODES_1
Subsection "Display"
Depth 24
Modes "1366x768"
EndSubsection
EndSection
Section "Screen"
Identifier "Screen2"
Device "Card2"
Monitor "Monitor2"
DefaultDepth 24
#Option "metamodes" "1366x768_60 +0+0" #METAMODES_2
Subsection "Display"
Depth 24
Modes "1366x768"
EndSubsection
EndSection
I suppose that this kind of defective xorg.conf created by xorgwizard might cause problems.
And it happens with all my tested 2 other kernels. Compiling latest jwm-731 or disabling acpid....causes only that acpid not listening disables my powerbutton shutdown function and I had to hard shutdown. Jwm change to 574 or 731 didnt make any difference.
My conclusion....it is due to X.Org X Server 1.13.3 and xorgwizard incompatibility. But this is just preliminary evaluation of the problem...
I think that Barry Kauler should check this post....when his vacation is over....or anybody who understands more about Xorg - X server and xorgwizard.
By the way...3.7.10 kernel behaves nicely in my laptop. 3.8.X kernels does not like my acpi_call module anymore. If I switch off ati dual graphics discrete card ( I do that because it eat watts and warms the laptop even when it is not in use ), the suspend to ram does not wake up nicely. Sometimes it hangs. I cant cause that hanging with 3.7.10 kernel. It makes me thinking that I could use this version more....posting from 3.7.10 kernel Upup Raring now....Needs more test and usage and prolonged suspend to ram before I am sure...
Posted: Mon 15 Apr 2013, 09:17
by anikin
Hi pemasu,
in all the recent puppies, my eeepc boots right into Intel video driver. It's a bit different with upup raring, as you will see in the attached picture. On the left side is the first run applet that shows upup precise 3.8.3.1, and on the right side, the same applet showing upup raring. It does pick up intel video in the end, but it's not doing so in a "straight" manner. Not a big deal, just an illustration to your thoughts on xorg as probably being buggy.
Posted: Mon 15 Apr 2013, 12:50
by pemasu
Anikin. That is caused just because I forgot to remove Video Wizard. You cant use Video Wizard in my builds. Different kernel. And I cant load my compiled proprietary drivers to ibiblio...so....I have removed it. Now I forgot to do it manually. Should include it to the script.
Thanks.
Posted: Mon 15 Apr 2013, 15:06
by anikin
anikin knows how to notice the obvious.
On the other hand, how else could I have learned about the removed video wizard?