Carolina: Vanguard Edition Release 2 December 2014

[battleshooter]

Tahr's Xorgwizard didn't appear to make a difference in regards to setting the correct input drivers. Must be something else then.

Here's the Xorg pet.

md5sum:dc52d915cb6d00b91566807014bf5a06

Everything is mushed into it. Very large. Very work in progress, use at your own risk, rah-de-rah.

128 MB because everything is in there. Non stripped too I think should have done that before uploading. Oops. Oh well. Pretend I didn't for bug testing reasons

Roughly this pet is a conglomerate of the following packages:

• elfutils-0.160.pet
  libXdmcp-1.1.1.pet
  xcb-proto-1.11.pet
  llvm-3.5.0.src.pet
  xkeyboard-config-2.13.pet
  libepoxy-1.2.pet
  Mesa-10.3.2.pet
  xorg_app.pet
  libevdev-1.3.1.pet
  xorg_drivers.pet
  libffi-3.2.1.pet
  xorg_libraries.pet
  libvdpau-0.8.pet
  util-macros-1.19.0.pet
  xorg_proto.pet
  libXau-1.0.8.pet
  xbitmaps-1.1.1.pet
  xorg-server-1.16.1.pet
  libxcb-1.11.pet
  xcb.pet

Oh darn it. I just remembered I didn't compile the Intel Xorg drivers. I'll be off to compile those and put them here as soon as this is done uploading.

Edit:
xf86-video-intel-2.99.916.pet
pixman-0.32.6.pet (Is needed)
libva-1.4.1.pet
libva-intel-driver-1.4.1.pet

[James C]

Just to add to the display woes.......

On the box with an old Nvidia GT210 card,boots to an incorrect 1600x1200 resolution using "vesa". Any attempt to use Xorgwizard,using "vesa" or the desired "nouveau", results in X refusing to start.

Previous version with k-3.17.0 works fine.

Downloading the above pet to see if it makes any difference.

Code: Select all

# report-video
VIDEO REPORT: tahrpup, version 6.0

Chip description:
VGA compatible controller: NVIDIA Corporation GT218 [GeForce 210] (rev a2)

Requested by /etc/X11/xorg.conf:
  Resolution (widthxheight, in pixels): 1024x768x16
  Depth (bits, or planes):              24
  Modules requested to be loaded:       dbe 

Probing Xorg startup log file (/var/log/Xorg.0.log):
  Driver loaded (and currently in use): nvidia
  Loaded modules: dbe dri2 extmod fb glx kbd mouse ramdac wfb

Actual rendering on monitor:
  Resolution:  1440x900 pixels (411x263 millimeters)
  Depth:       24 planes

...the above also recorded in /tmp/report-video
#

[James C]

For informational purposes.....

Code: Select all

# report-video
Carolina-k3.17, version 1.2 on Tue 16 Dec 2014

Chip description:
 0.0 VGA compatible controller
 nVidia Corporation GT218 [GeForce 210] (rev a2)  
oem: NVIDIA
product: GT218 Board - 08720002 Chip Rev
 
X Server: Xorg
Driver used: nouveau

X.Org version: 1.11.0
  dimensions:    1440x900 pixels (380x238 millimeters)
  depth of root window:    16 planes 

   ...the above also recorded in /tmp/root/ as report-video,
and archived with xorg.conf and Xorg.0.log as report-video-full.gz
# 

[Billtoo]

Tahr's Xorgwizard didn't appear to make a difference in regards to setting the correct input drivers. Must be something else then.

Here's the Xorg pet.

md5sum:dc52d915cb6d00b91566807014bf5a06

Everything is mushed into it. Very large. Very work in progress, use at your own risk, rah-de-rah.

128 MB because everything is in there. Non stripped too I think should have done that before uploading. Oops. Oh well. Pretend I didn't for bug testing reasons

Roughly this pet is a conglomerate of the following packages:

• elfutils-0.160.pet
  libXdmcp-1.1.1.pet
  xcb-proto-1.11.pet
  llvm-3.5.0.src.pet
  xkeyboard-config-2.13.pet
  libepoxy-1.2.pet
  Mesa-10.3.2.pet
  xorg_app.pet
  libevdev-1.3.1.pet
  xorg_drivers.pet
  libffi-3.2.1.pet
  xorg_libraries.pet
  libvdpau-0.8.pet
  util-macros-1.19.0.pet
  xorg_proto.pet
  libXau-1.0.8.pet
  xbitmaps-1.1.1.pet
  xorg-server-1.16.1.pet
  libxcb-1.11.pet
  xcb.pet

Oh darn it. I just remembered I didn't compile the Intel Xorg drivers. I'll be off to compile those and put them here as soon as this is done uploading.

Edit:
xf86-video-intel-2.99.916.pet
pixman-0.32.6.pet (Is needed)

I booted up my Carolina Vanguard usb install on the Acer pc which has the
intel ironlake, downloaded the and installed the 2 pets then shutdown.

Plugged the drive into the ivybridge pc and it started to boot, just at
the point where it goes to the desktop the pc locked up, did a hard
reset.
Booted a second time with "puppy pfix=nox", at the prompt ran xorgwizard
and chose the intel driver, did xwin to get to the desktop, the pc
locked up again.
I'm pretty sure the larger pet had installed okay, I waited long
enough I think, so no luck with the new pets here.

EDIT: I just noticed there was the 3rd pet to install, pixman..., I downloaded that and couldn,t install it with "puppy pfix=nox" petget and booting the flash drive on the ironlake pc locks it up too.
Damn

[James C]

What I get on boot is all I can get.....any attempt to use xorgwizard results in X refusing to start completely.

Just for fun,tried the above main pet, froze up my computer too.

Code: Select all

# report-video
Vanguard-Carolina, version 1.3 on Tue 16 Dec 2014

Chip description:
 0.0 VGA compatible controller
 nVidia Corporation GT218 [GeForce 210] (rev a2)  
oem: NVIDIA
product: GT218 Board - 08720002 Chip Rev
 
X Server: Xorg
Driver used: # vesa

X.Org version: 1.11.0
  dimensions:    1600x1200 pixels (423x318 millimeters)
  depth of root window:    24 planes 

   ...the above also recorded in /tmp/root/ as report-video,
and archived with xorg.conf and Xorg.0.log as report-video-full.gz
# 

[Gobbi]

Nice job @battleshooter
vlina 1.3 is working well on my desktop , it has nice features as usual .
I noticed it's a non-pae version . Shall we have a pae version in the near future , or maybe we could get a hint how to modify the existing version

[Marv]

@battleshooter

Inching forward with the Xorg pets on the Pentium M 855GM as it's forgiving. Downloaded, installed, ran xorgwizard with no restart or reboot yet. hardinfo and Xorg log attached. Now for a restart... Could the kbd and mouse drivers be over-ridden by the template xorg.conf in Carolina?

Glxgears run pre-restart. LibGL error mean anything to you? FPS up about 20%.

Code: Select all

# glxgears
ATTENTION: default value of option vblank_mode overridden by environment.
ATTENTION: default value of option vblank_mode overridden by environment.
libGL error: Version 4 or later of flush extension not found
libGL error: failed to load driver: i915
ATTENTION: default value of option vblank_mode overridden by environment.
ATTENTION: default value of option vblank_mode overridden by environment.
3084 frames in 5.0 seconds = 616.744 FPS
3095 frames in 5.0 seconds = 618.836 FPS
3098 frames in 5.0 seconds = 619.451 FPS
3095 frames in 5.0 seconds = 618.928 FPS

[Marv]

continuing on the Pentium M. Rebooted with the pets installed (after a slight detour to make the savefile larger), both with the self-generated xorg.conf in place from the prev post and xorgwizard auto turned off (line 21 DISTRO_SPECS in initrd.gz) and with that xorg.conf hidden and xorgwizard auto on (the default). Both boots normal with the attached Xorg.0.log. See screenie for only anomalie noted. Hardinfo shows the correct driver and Mesa in all cases.

On to the Bay Trail later today.

[Billtoo]

I installed to a 16gb flash drive with PUI, computer is a Gateway
desktop.
video-info-glx 1.5.3 Tue 16 Dec 2014 on Vanguard-Carolina 1.3 Linux 3.18.0 i686
0.0 VGA compatible controller: nVidia Corporation Device 1082 (rev a1)
oem: NVIDIA
product: GF110 Board - 12630002 Chip Rev

X Server: Xorg Driver: nvidia
X.Org version: 1.11.0
dimensions: 3840x1080 pixels (1016x286 millimeters)
depth of root window: 24 planes

direct rendering: Yes
server glx vendor string: NVIDIA Corporation
server glx version string: 1.4
OpenGL vendor string: NVIDIA Corporation
OpenGL renderer string: GeForce GTX 560 Ti/PCIe/SSE2
OpenGL version string: 4.4.0 NVIDIA 340.65

# glxgears
Running synchronized to the vertical refresh. The framerate should be
approximately the same as the monitor refresh rate.
303 frames in 5.0 seconds = 60.444 FPS
301 frames in 5.0 seconds = 60.000 FPS
301 frames in 5.0 seconds = 60.001 FPS
#

Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz
Core 0: @1600 1: @1600 2: @1600 3: @1600 4: @1600 5: @1600 6: @1600 7: @1600 MHz

2 monitor support but can't install wallpaper,drive icons not there.

Also, compiling works fine on my netbook with the athlon processor but not on
any of my other Vanguard-Carolina 1.3 installs that have an intel processor, this one and the Acer
Revo for sure.
# md5sum devx_vlina_1.3.sfs
0b9ab4a02575e854091053de85ce52c1 devx_vlina_1.3.sfs
#
Pets that I make on the netbook work on the other Vanguard-Carolina 1.3 installs.

[Marv]

I'm posting from the Bay Trail box running with the intel driver and intel rasterizer. The only oddity is the keybinding nag as in my post above. All else is working with a glxgears of 1650 or so. A package of diagnostics attached below. I cheated a bit in that I just used the savefile from the laptop removing .flsynclient, acpid and a few other things not relevant to the desktop. Removed the i915.modeset=0 line from the menu item and it booted right up to the correct resolution and monitor refresh rates. Nothing else at all added or fiddled. I then hid the xorg.conf file and rebooted. Came right up again with the correct resolution. SimpleGTKradio playing now, glxgears cooking at 1600 FPS, huge Carolina repo, on a processor with a 10 watt TDW. I'm pleased

[p310don]

Was about to post my own issues with Carolina Vanguard, only to see word for word my problem from James C:

Just to add to the display woes.......

On the box with an old Nvidia GT210 card,boots to an incorrect 1600x1200 resolution using "vesa". Any attempt to use Xorgwizard,using "vesa" or the desired "nouveau", results in X refusing to start.

Previous version with k-3.17.0 works fine.

Downloading the above pet to see if it makes any difference.

edit: Installing the nvidia driver from the nvidia page works nicely (so far)

[battleshooter]

I'm posting from the Bay Trail box running with the intel driver and intel rasterizer.[ ...] Came right up again with the correct resolution. SimpleGTKradio playing now, glxgears cooking at 1600 FPS, huge Carolina repo, on a processor with a 10 watt TDW. I'm pleased

Sweet, sweet, sweet! Finally some good news. I didn't actually think that would work. So we do have a old Xorg issue.

Also, compiling works fine on my netbook with the athlon processor but not on
any of my other Vanguard-Carolina 1.3 installs that have an intel processor, this one and the Acer
Revo for sure.

Ahh, this issue back to haunt me. Geoff actually pointed it out in testing but I naively hoped it would go away I have an AMD Athlon processor, looks like I didn't compile for Intel. That will be a major fix, pity I'll have to upload devx again, I was hoping to get away with that.

Concerning the Xorg pet (TMP.pet) I put up, try with libva-1.4.1.pet and libva-intel-driver-1.4.1.pet I sneakily put up after you tested it. I think Marv ran his test using them so the full pet worked for him.

On the box with an old Nvidia GT210 card,boots to an incorrect 1600x1200 resolution using "vesa". Any attempt to use Xorgwizard,using "vesa" or the desired "nouveau", results in X refusing to start.

Previous version with k-3.17.0 works fine.

Having a look at this at the moment to take a break from Intel, plus I actually have a machine to test on, but I'm not sure where to start. As you say, Nouvea worked fine on the 3.17 kernel on this machine but displays the same symptoms you have described. I didn't change the kernel options so I assume the kernel drivers changed?

Just for fun,tried the above main pet, froze up my computer too.

Interesting idea of fun, but then again, who am I to judge?

Try with the libva-1.4.1.pet pet I put up, but I think something is missing from Vanguard that was in the k3.17 build vs needing an upgrade to function.

Nice job @battleshooter
vlina 1.3 is working well on my desktop , it has nice features as usual .
I noticed it's a non-pae version . Shall we have a pae version in the near future , or maybe we could get a hint how to modify the existing version

Thanks Gobbi! Yep, next version will be PAE... this build was supposed to be actually but looks like I don't know how to compile a PAE kernel, I don't think the last ones were PAE either. Need to figure what options/patches I'm missing.

[battleshooter]

I still need to look at your reports a little more in-depth Marv, but isn't

Code: Select all

libGL error: Version 4 or later of flush extension not found
libGL error: failed to load driver: i915 

a really bad thing? Doesn't that mean you aren't even using the Intel driver if it failed to load?

Could the kbd and mouse drivers be over-ridden by the template xorg.conf in Carolina?

That's not a bad idea. I tested it by manually setting "evdev" in xorg.conf but then the mouse and keyboard wouldn't move. Hmm. A bit more research and it looks like, from all you guys reports, (Tahr, Quirky Unicorn) you're actually using mouse and kbd in the modern Puppies, so I don't where I got the idea they were phased out from. So the only thing to sort out is the weird keybinding issue. It seems to affect ctrl-alt-backspace too, as that doesn't work.

[Marv]

I still need to look at your reports a little more in-depth Marv, but isn't

Code: Select all

libGL error: Version 4 or later of flush extension not found
libGL error: failed to load driver: i915 

a really bad thing? Doesn't that mean you aren't even using the Intel driver if it failed to load?

Could the kbd and mouse drivers be over-ridden by the template xorg.conf in Carolina?

That's not a bad idea. I tested it by manually setting "evdev" in xorg.conf but then the mouse and keyboard wouldn't move. Hmm. A bit more research and it looks like, from all you guys reports, (Tahr, Quirky Unicorn) you're actually using mouse and kbd in the modern Puppies, so I don't where I got the idea they were phased out from. So the only thing to sort out is the weird keybinding issue. It seems to affect ctrl-alt-backspace too, as that doesn't work.

Good question that. That libGL error only shows up on the 855GM laptop and only when running glxgears. Every other check shows it using the 'intel' driver and the intel Mesa. The FPS in glxgears fits well with that also. I've never seen FPS anywhere near that high with either modesetting or vesa drivers on that laptop. Thus a bit of a puzzle. The Bay Trail box doesn't show that error running glxgears and all ways I can look at it also shows intel driver and intel Mesa with the correct versions and compilations. All can think of is that the intel driver is saying it is falling back to what works with the 855GM due to lack of pipes or ? on that system. As an initial diagnostic stab, I'd give the pet an A on both intel boxes. The proof will be how the newer Xorg plays with other drivers and chipsets.

On the key binding. After I disabled sven, which I usually do, that message no longer appeared.

On the mouse/keyboard drivers, I looked at LxPupTahr, and indeed it uses 'mouse' and 'kbd' on all my systems.

To all a good night

[l0wt3ch]

Looks splendid!

[battleshooter]

Looks splendid!

Thanks l0wt3ch

On the key binding. After I disabled sven, which I usually do, that message no longer appeared.

Oddly enough that comment triggered the thought process for me and I figured out the mystery of the clashing keys, a new Xorg thing, evdev actually. Creates a list of what it guess are keyboard shortcuts matching up to physical keys and sticks it at /usr/share/X11/xkb/keycodes/evdev. Also apparently the option

Code: Select all

Option "DontZap" "false"

in xorg.conf doesn't cut the butter anymore in some cases (in mine anyway), recreating it in .xinitrc with

Code: Select all

setxkbmap -option terminate:ctrl_alt_bksp 

seems to do the trick.

Nothing on the nVidia front at the moment but I'm sure we'll get there.

[Billtoo]

I did a new install and added all the new intel pets, it works!

video-info-glx 1.5.3 Wed 17 Dec 2014 on Vanguard-Carolina 1.3 Linux 3.18.0 i686
2.0 VGA compatible controller: Intel Corporation Ivy Bridge Graphics Controller (rev 09)
oem: Intel(R) Sandybridge/Ivybridge Graphics Chipset Accelerated VGA BIOS
product: Intel(R) Sandybridge/Ivybridge Graphics Controller Hardware Version 0.0

X Server: Xorg Driver: intel
X.Org version: 1.16.1
dimensions: 1920x1080 pixels (507x285 millimeters)
depth of root window: 24 planes

direct rendering: Yes
server glx vendor string: SGI
server glx version string: 1.4
OpenGL vendor string: Intel Open Source Technology Center
OpenGL renderer string: Mesa DRI Intel(R) Ivybridge Desktop x86/MMX/SSE2
OpenGL version string: 3.0 Mesa 10.3.2

# glxgears
Running synchronized to the vertical refresh. The framerate should be
approximately the same as the monitor refresh rate.
303 frames in 5.0 seconds = 60.414 FPS
301 frames in 5.0 seconds = 60.001 FPS
301 frames in 5.0 seconds = 60.001 FPS
#

Intel(R) Core(TM) i5-3550 CPU @ 3.30GHz
Core 0: @1600 1: @1600 2: @1600 3: @1600 MHz

Thanks.

EDIT: New install on another pc with Ivybridge graphics, screenshot2

X Server: Xorg Driver: intel
X.Org version: 1.16.1
dimensions: 1920x1080 pixels (507x285 millimeters)
depth of root window: 24 planes

direct rendering: Yes
server glx vendor string: SGI
server glx version string: 1.4
OpenGL vendor string: Intel Open Source Technology Center
OpenGL renderer string: Mesa DRI Intel(R) Ivybridge Mobile x86/MMX/SSE2
OpenGL version string: 3.0 Mesa 10.3.2

[battleshooter]

Right, got the Nvidia issue.

I seem to have done something to Vesa to disable it from setting up high resolutions and its general compatibility as most people who reported being able to boot in to Carolina 1.3 seem to be doing so with the Vesa driver.

Anyway, using the Xorg, libva and pixman pet I posted earlier I was able to load the nouveau driver with acceleration. After installing the pets, I exited to prompt and typed:

Code: Select all

modprobe nouveau modeset=1

Booted in to X with correct resolution and acceleration.

The pet is starting to look ok, but after all this testing I'll probably recompile again to get a cleaner package. At least that's what I feel about packages I compile with a better sense of what's going on.

I think the only issues left are:

• GCC Intel incompatibility
  Musher's root tty issue
  Yerc's dear touchpad

A bit more testing and then a new release with all the fixes. I really should have called this one RC, I'm not sure what I'm going to call the next one as the main Carolina hasn't moved up a number Maybe something real creative like "Build 2"... How would I fit that in the logo? Hmmm.

Edit: Just read Bill's post:

I did a new install and added all the new intel pets, it works!

Hah! Awesome!!! Smiley faces all round!

I can't believe after all this I should have gone with the Xorg upgrade and that there was actually nothing wrong with the build.

[Marv]

A bit more testing and then a new release with all the fixes.

If it is going in as a fix, after flogging /root/Startup/9acpid more mercilessly, I would move the test for a running acpid to line 10 and leave line 9 as is. Modified version attached for your perusal.

Looked at the 855 GM a bit more wrt the error shown while running glxgears. When setting up the intel framebuffer on that rig, dmesg shows some pipe underruns and 'sets up reporting for them' but then shows the framebuffer as set up and intel. Runs like a scalded Pentium M anyway

Edit: More searching... The pipe underruns on the older processors are a function of kernel (newer than about 3.15) and intel microcode in the bios interaction. Newer microcode is available from intel which might cure it and in principle there is a way for the kernel to read it on boot without burning new code into the bios. I have the newest microcode now but so far the church and state aren't talking. I'll squawk if I get it. Driver runs anyway is the word and is my experience.

Edit: Pulled the CPU frequency governor from the Bay Trail box vlina (unneeded there and left over from the savefile swap while working on the Xorg pets) and the glxgears FPS is up to 2150 FPS. That is both great and matches LxTahrPup on that box pretty much exactly. All good.

[SouthPaws]

I'm not sure what I'm going to call the next one...

Hey battleshooter, what if you just use the date as a sub-title? Then you wouldn't have to change the logo and everyone would be able to tell which is the latest version.