Upup Precise 3.7.2 with kernel 3.7.2, no pae

For talk and support relating specifically to Puppy derivatives
Message
Author
User avatar
charlie6
Posts: 1230
Joined: Mon 30 Jun 2008, 04:03
Location: Saint-Gérard / Walloon part of Belgium

#46 Post by charlie6 »

Hi Keef,
thanks a lot for your post !
i surely have learned something !
would you mind:
1) add:
1.1)

Code: Select all

Section "ServerLayout"
   Identifier     "X.org Configured"
   Screen      0  "Screen0" 0 0
   InputDevice    "Mouse0" "CorePointer"
   InputDevice    "Keyboard0" "CoreKeyboard"
#from kernel 2.6.35 and later the "SendCoreEvents" after InputDevice is no longer required
   InputDevice    "stylus"
   InputDevice    "eraser"  # <<< please do INSERT THIS LINE
EndSection
and modify
1.2)

Code: Select all

Section "InputDevice"
      Identifier    "stylus"
      Driver        "wacom"
      Option        "Device" "/dev/input/wacom"
      Option        "Type" "stylus"
      Option        "USB" "on"
      Option        "Mode" "Relative"
#      Option        "Mode" "Absolute"
EndSection
and add this to get the eraser (i.e. the pens tail tip end...) working:
1.3)

Code: Select all

Section "InputDevice"
      Identifier    "eraser"
      Driver        "wacom"
      Option        "Device" "/dev/input/wacom"
      Option        "Type" "eraser"
      Option        "USB" "on"
      Option        "Mode" "Relative"
#     Option        "Mode" "Absolute"
EndSection 
some cursors arrows behaviour might change ...maybe better ! if not, please disregard.

charlie

User avatar
01micko
Posts: 8741
Joined: Sat 11 Oct 2008, 13:39
Location: qld
Contact:

#47 Post by 01micko »

Just an FYI post...

Regarding the "modesetting" Xorg driver.

In my Slacko tests I did find that this can conflict with some other xorg drivers. It is included in Slacko but it is hidden away in /usr/lib/x with possibly other alternate drivers. There is a problem (bug) with that approach though as xorgwizard does not know about the modesetting driver. Consequently, it is not moved to /usr/lib/xorg/modules/drivers and so is unavailable to xorg.

I am going to add a case to xorgwizard that handles this scenario such modesetting is moved if chosen. Possibly udev rules should be written that on the detection of a cedaview or poulsbo graphics card that the modesetting driver is automatically moved. Jury is out on that one. I am already doing preliminary work on this.


pemasu, sorry for brief off topic here...

@ aarf, who unfortunately can't post here. Good you got modesetting to work in Upup with RoDoN's compile. Regarding Slacko, you must uninstall the Upup version. Then run this command

Code: Select all

mv /usr/lib/x/drivers-alternate/modesetting_drv.so /usr/lib/xorg/modules/drivers/
Then run xorgwizard, choose modesetting driver, that should work. Modesetting is known working with cedarview in Slacko.
Puppy Linux Blog - contact me for access

User avatar
pemasu
Posts: 5474
Joined: Wed 08 Jul 2009, 12:26
Location: Finland

#48 Post by pemasu »

01micko. Thank you of the input. Intel gma 500 (Poulsbo), gma 600, gma 3600 (Cedarview) and 3650 (Cedarview) which uses gma500_gfx.ko and can use xorg modesetting driver have been troublesome PowerVR based graphics.

The graphics support in linux is only usable. No 3D acceleration. But it is better than nothing and intel atom based notebooks are not for power hungry usage anyway.

If you can edit xorgwizard so that gma500_gfx.ko can be used with included xorg modesetting driver, it would be great thing for these intel atom comps with above gma graphics.
Xorgwizard handled xorg driver translocation from alternate to normal location so that the xorg modesetting driver would be in use would be great.

I would include that xorgwizard of course to my builds. Thanks of your initiative. And thanks to aarf who brought this need to my attention. I didnt know of the work done in Slacko. Now I know.

Cheers.

User avatar
pemasu
Posts: 5474
Joined: Wed 08 Jul 2009, 12:26
Location: Finland

#49 Post by pemasu »

I am still quite newbie for compiling ati-amd-fglrx-drivers for petting. But...I have learnt to compile and create them inside extracted ati folder and have learnt to patch the kernel module source before running the installer. So....these might even work.
I created ati-12.10 pet for modern graphics, like mine. Tested the pet with my dual graphics. And then I created ati-12.6-legacy pet.

Cave out. For me suspending does not result working waking up.
There is ati own fixes for that inside: /usr/share/doc/fglrx/examples/etc
You can try to copy the stuff to right places and test them. I havent had time to do so....maybe next time when I boot with ati and watch movies through hdmi from my 42 inch lcd tv. Then I have a reason to try them with suspending.

I created quickpet for this 3.7.2 kernel version. Under drivers sections you find those ati-amd-fglrx pets and also newest nvidia 310.19 pet.

They are in upup repo, so those who want to scan it...find them there. Nvidia 310.19 is there as sfs also. You can download it with Precise sfs downloader also...There are now quite many ways to install stuff...

Test...enjoy...or blame me...
Attachments
quickpet_upup-precise-kernel-3.7.2-0.2.pet
(104 KiB) Downloaded 423 times

User avatar
01micko
Posts: 8741
Joined: Sat 11 Oct 2008, 13:39
Location: qld
Contact:

#50 Post by 01micko »

Further to gma/poulsbo/cedarview issues I have patched xorgwizard to move the modesetting driver from the alternate location to the real location where Xorg can find it, I have also patched xorgwizard-automatic to move the modesetting driver if it detects one of these cards based on device ID, now my database may be incomplete, I don't know. But if your machine is over 6 months old it will almost certainly be listed. Newer ones might be too.

I have not packaged this as a pet as it is only if you have an intel atom infected with the gma 500 (Poulsbo), gma 600, gma 3600 (Cedarview) and 3650 (Cedarview) which uses gma500_gfx.ko.
The package is structured just like a pet to easily merge into the filesystem. The packages contains: a patched /usr/sbin/xorgwizard, a patched /usr/sbin/xorgwizard-automatic and a new script called xcedarview.sh, called from xorgwizard-automatic.

Get the package here:
http://01micko.no-ip.org/testing/xorgwi ... 15.tar.bz2 25.1kb

If you know what you are doing you can do it with no X. Ideally that's how it should be done, but for testing purposes there are 2 things to test.
  • 1.drop to prompt (yau may already be there :P ) and run xorgwizard
    Choose the "modesetting" driver, start x with xwin

    2. From the prompt delete or move /etc/X11/xorg.conf
    If you have ordinary Intel a test will be performed to see if you have a gma device and you will be informed. If you have the device you will then be informed that "modesetting" driver is moved to it's proper location or if it's not there then a "NOT FOUND" message occurs. .
NOTE: you'll also need the modesetting driver in the correct location for the test!! Here it is:
http://01micko.no-ip.org/testing/xf86-v ... T_upup.pet

This is RoDoN's pet just moved. Why is this necessary? Well .. if you have an nvidia and use xorgwizard then it's "modesetting" gets chosen and X crashes unrecoverably! That's why I moved in the first place.

Good Luck!
Puppy Linux Blog - contact me for access

User avatar
RoDoN
Posts: 10
Joined: Fri 09 Sep 2011, 10:34
Location: Russia

gma500

#51 Post by RoDoN »

pemasu, 01micko
Thanks!

For normal work gma500_gfx + modesettting_drv driver poulsbo.ko need to delete or rename to the poulsbo.ko.not
Last edited by RoDoN on Tue 15 Jan 2013, 06:17, edited 1 time in total.

User avatar
pemasu
Posts: 5474
Joined: Wed 08 Jul 2009, 12:26
Location: Finland

#52 Post by pemasu »

01micko. Thanks !!!
I will check them out soon. What do you think of the poulsbo.ko stub driver. It might inhibit the use of gma500_gfx.ko and modesetting driver.
Would it be still reasonable to blacklist it in the build with MODULESCONFIG ?
Post edited because RoDon above answered it already.


Heh. RoDon beat with 2 seconds. Congratulation. You won ice cream. I will fax it to you.
Last edited by pemasu on Tue 15 Jan 2013, 06:18, edited 2 times in total.

User avatar
01micko
Posts: 8741
Joined: Sat 11 Oct 2008, 13:39
Location: qld
Contact:

#53 Post by 01micko »

Yes RoDon is correct, I should have mentioned that. My next slacko will be with out it (only k3.3x and above exhibits this, there is a test for it in my xorgwizard-automatic).
Puppy Linux Blog - contact me for access

User avatar
davids45
Posts: 1326
Joined: Sun 26 Nov 2006, 23:33
Location: Chatswood, NSW

ATI AMD Drivers

#54 Post by davids45 »

G'day pemasu,
I am still quite newbie for compiling ati-amd-fglrx-drivers for petting. But...I have learnt to compile and create them inside extracted ati folder and have learnt to patch the kernel module source before running the installer. So....these might even work.
I created ati-12.10 pet for modern graphics, like mine.
And mine, too (Radeon HD5450 card).

Ran your ati pet and then downloaded the ~50MB driver package with no problem.
I installed first to a Frugal Upup-3.7.2 with a 256MB save-file. A red warning about the save-file remaining (~80MB) and I increased the save file size to 512MB for next boot. But the new driver did load and run well after I did the 'aticonfig --initial' step and restarted X.

But on shutting down to 'reboot' there was a problem - the computer did not shut down and the text displayed seemed to be part of the start up messages and the computer hung. I hit the 'reset' button and re-booted the Frugal-3.7.2 which started up with the ATI driver and a good display. Shutting down now went as expected with no strange message or hanging.

I then installed the ati-fglrx driver on my Full Upup-3.7.2 with the same result. Good install, good initial display, but again hanging on trying to reboot the first time; 'reset' and re-booting to the Full Upup and everything was good.

David S.

User avatar
charlie6
Posts: 1230
Joined: Mon 30 Jun 2008, 04:03
Location: Saint-Gérard / Walloon part of Belgium

wacom stuffs already included in uPup Precise 3.7.2 ...

#55 Post by charlie6 »

Hi,
@keef
thanks again for your post !
I presume it would no longer be necessary to compile the xf86-input-acom driver ... related to the present distro as also puppies with «new» kernel as uPup-54X7 and dPup-Exprimo-3.6.2, and maybe others!
Please consider and read
http://www.murga-linux.com/puppy/viewto ... 882#677882
I suggest we to continue posting about wacom subject in the thread Forum index/House Training/Howto/«How to set up a Wacom Tablet».
Hope this helps
Charlie
Last edited by charlie6 on Tue 15 Jan 2013, 16:49, edited 1 time in total.

LateAdopter
Posts: 361
Joined: Fri 27 May 2011, 17:21
Location: Reading UK

#56 Post by LateAdopter »

Hello Pemasu

Thanks for this, it's working fine on my system.

I compiled TurionPowerControl and then added my usual p-state voltage configuration via rc.local. It is working perfectly, thanks for cpuid and msr.

Amazingly, almost immediately you have posted a fglrx legacy .pet. I have installed that and it is worked with no problem.

I notice that it does not install the GLX stuff which I thought ATI pets normally do. I don't play 3D games, so I don't need it. I expect I could install it from the Ubuntu repo.

Maybe other Puppies come with GLX installed, and I haven't noticed. The r600g has it built-in.

If this is not the expected behaviour, please tell me and I will investigate the log files further.

My main interest is playing pack my MPEG video recordings. Now I can try installing the latest XBMC-XVBA package from the wsnipex PPA on launchpad. Then I can compare it with software decoding using VLC and the Xorg r600g driver.

Thanks again

LateAdopter
Posts: 361
Joined: Fri 27 May 2011, 17:21
Location: Reading UK

#57 Post by LateAdopter »

Hello Pemasu

The problem seems to be that the radeon kernel module is still being loaded, which breaks fglrx acceleration.

I think the radeon module should be blacklisted by blacklist-radeon.conf, and this is not present.

I tried creating blacklist-radeon.conf but it disappeared from modprobe.d after a reboot. So radeon is still loaded.

What is the correct way to disable it?

Thanks.

EDIT: I spotted my mistake, it should have been blacklist-fglrx.conf. I'll try again!

EDIT2: No, that didn't work either!

EDIT3: I renamed radeon.ko which stopped it from loading.

But I still get: AIGLX error: failed to load OpenGL driver.

libgl1-mesa-dri and libgl1-mesa-glx are both shown as installed in the Package Manager

User avatar
pemasu
Posts: 5474
Joined: Wed 08 Jul 2009, 12:26
Location: Finland

#58 Post by pemasu »

LateAdopter and davids45. There was missing files. My way to create packages seems to miss some installer stage. fglrx_dri.so was placed to wrong location. Also some other scripts and files were missing.

I tested the reuploaded ati pet and it should work now. Suspending started to work for me, radeon will be blacklisted, dri has been enabled.
I have reuploaded both ati pets. Oh yeah...ati-12.6 legacy fglrx.ko needed heavy patching. It is possible that the fglrx.ko kernel module is defective due to that heavy patching. Kernel 3.7 might still miss some crucial patch. Time will tell...
I cant test 12.6-legacy pet. I dont have suitable ati card.

I hope the pets are now complete enough.

User avatar
charlie6
Posts: 1230
Joined: Mon 30 Jun 2008, 04:03
Location: Saint-Gérard / Walloon part of Belgium

Firefox18 for your language

#59 Post by charlie6 »

Hi,
here choose the firefox18.tar.bz2 for your language:
http://www.mozilla.org/en-US/firefox/all/
cheers, charlie

LateAdopter
Posts: 361
Joined: Fri 27 May 2011, 17:21
Location: Reading UK

#60 Post by LateAdopter »

Hello Pemasu

I started with a new savefile and installed your updated ATI legacy 12.6 pet.

Everything seems to be working correctly now.

Athlon II X2, 785G (RS880G) chipset.

glxgears runs
hardinfo shows the ATI GL driver
Xorg log shows AIGLX loaded and initialised

Looking good

Many thanks

User avatar
charlie6
Posts: 1230
Joined: Mon 30 Jun 2008, 04:03
Location: Saint-Gérard / Walloon part of Belgium

#61 Post by charlie6 »

Hi,
this might be indicative ...
on a Dell620 laptop running idle in POWERSAVE mode, screenshot combination of the tempicon shows 20°C temperature difference between uPup54X7 and uPup3.7.2.
Attachments
CPUtempk3.7.2vsk3.2.29_idle.jpg
(16.36 KiB) Downloaded 1428 times

User avatar
peebee
Posts: 4370
Joined: Sun 21 Sep 2008, 12:31
Location: Worcestershire, UK
Contact:

New frisbee announced

#62 Post by peebee »

Rerwin has announced a new version of "integrated frisbee" for testing at:

http://murga-linux.com/puppy/viewtopic. ... &start=108

For some reason? the pop-ups display at different parts of the screen in Upup Precise as compared to Precise 5.4.3 or Slacko 5.4? Anybody know why this might be?

Some of the components are already in Upup Precise 3.7.2 from an earlier less developed release - but to be safe all of the new components should be loaded.

Cheers
peebee
ImageLxPup = Puppy + LXDE
Main version used daily: LxPupSc; Assembler of UPups, ScPup & ScPup64, LxPup, LxPupSc & LxPupSc64

User avatar
Billtoo
Posts: 3720
Joined: Tue 07 Apr 2009, 13:47
Location: Ontario Canada

Upup Precise 3.7.2 with kernel 3.7.2, no pae

#63 Post by Billtoo »

I did a manual frugal install to the hard drive.
I installed your amd_fglrx_catalyst-legacy-12.6-k3.7.2-upup.pet
VIDEO REPORT: Upup Precise, version 3.7.2
Chip description:
oem: ATI ATOMBIOS
product: RS780 01.00
Requested by /etc/X11/xorg.conf:
Depth (bits, or planes): 24
Modules requested to be loaded: dbe
Drivers requested to be loaded: fglrx
Probing Xorg startup log file (/var/log/Xorg.0.log):
Driver loaded (and currently in use): fglrx
Loaded modules: amdxmm dbe ddc dri dri2 extmod fb fglrxdrm glesx glx kbd
mouse ramdac record vbe vgahw
Actual rendering on monitor:
Resolution: 1440x900 pixels (380x238 millimeters)
Depth: 24 planes
OpenGL
Vendor ATI Technologies Inc.
Renderer ATI Radeon HD 3200 Graphics
Version 3.3.11653 Compatibility Profile Context
Direct Rendering Yes
# glxgears
19537 frames in 5.0 seconds = 3907.197 FPS
19552 frames in 5.0 seconds = 3910.321 FPS
19574 frames in 5.0 seconds = 3914.723 FPS
Computer
Processor 4x AMD Phenom(tm) II X4 810 Processor
Memory 3374MB (235MB used)
Linksys WUSB54GC v1 802.11g Adapter [Ralink RT73]

It's working well so far.
Attachments
screenshot.jpg
(75.03 KiB) Downloaded 1314 times

User avatar
davids45
Posts: 1326
Joined: Sun 26 Nov 2006, 23:33
Location: Chatswood, NSW

#64 Post by davids45 »

G'day charlie6,
I too saw quite different temperatures reported by the task-bar tempicon between different Pups on my desktop.

But one day, I right-clicked the tempicon to see 'More Info' and found temperatures were being reported for different computer elements - one Pup tempicon was reporting CPU (e.g. 35C), another Pup's tempicon was the graphics card temperature (e.g. 60C).

The 'More Info' shows all the temperatures so you can see which is ending up as the Pup's task bar icon.

Once I found this was happening, the differences between the same part for different Pups were much less.

Maybe there needs to be more icons (temperatures) displayed, maybe colour-coded for the different bits?

David S.

User avatar
Billtoo
Posts: 3720
Joined: Tue 07 Apr 2009, 13:47
Location: Ontario Canada

Upup Precise 3.7.2 with kernel 3.7.2, no pae

#65 Post by Billtoo »

@pemasu
I made and tested pets of the newest Midnight Commander file manager and it's
dependency Slang in Upup 472, mc is working well for me but needs the
devx installed to unpack zip files.

Post Reply