(old)Puppy Linux Discussion Forum

rjbrewer wrote: Same result on all 3 listed intel laptops.

Same thing here with a new desktop that works fine with slacko 533 PAE

Not sure what command to issue in 533 to determine more information that would be of assistance to you.

Thom

Same results on my intel video laptops.

If I boot "pfix=nox,ram" I can run xorgwizard and take the vesa option to get to desktop. But, other desktop navigation issues arise.

Here to help

I did a manual frugal install to a 4gb flash drive.
The only way to avoid a black screen and get to the desktop was to
boot pfix=nox,ram, run xorgwizard and choose vesa + proper resolution.

I installed mesa 8.0.4 from ppm.

# report-video
vesa
Slacko Puppy, version 5.3.5 on Wed 8 Aug 2012

Chip description:
2.0 VGA compatible controller
Intel Corporation Core Processor Integrated Graphics Controller (rev 18)
oem: Intel(R)Ironlake Desktop Graphics Chipset Accelerated VGA BIOS
product: Intel(R)Ironlake Desktop Graphics Controller Hardware Version 0.0

X Server: Xorg
Driver used: vesa

X.Org version: 1.12.3
dimensions: 1280x1024 pixels (339x271 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
# glxgears
3092 frames in 5.0 seconds = 618.388 FPS
3326 frames in 5.0 seconds = 665.160 FPS
3467 frames in 5.0 seconds = 693.200 FPS
#
It's working well enough to run google-earth.
I've just begun testing so that's it so far.

Thanks for all the intel failure reports.

That is what alpha is about. The only intel I tested was my Brookdale and it booted to a 'vesa' desktop at correct res. I do have my eee to test and I will inherit my son's HP lappy on Monday (845G) as we got him a new HP for his birthday, I don't think Puppy will get near that for awhile! It has radeon graphics anyway.

So far, my hypothesis on all this video trouble is that KMS is the cause. There is a "modesetting" xorg driver which I left out of 533 and also left out of this build but it seems that it is necessary for Intels. Also, for now, turning off KMS should produce a 'vesa' desktop. There is also some critical bugs in xorgwizard, one where it is possible that all drivers in /usr/lib/xorg/modules/drivers get moved to /usr/lib/x/drivers-backup!!! Not good at all.

KMS is also likely causing issues with radeon and nouveau. I haven't had any nouveau issues yet but it seems it may need to be turned on for radeon/ATI (AMD) cards so that the ati or radeon or r128 driver gets used. Else 'vesa' gets used unless your card doesn't support KMS (older radeons should work fine, mine does).

Also, all this KMS stuff seems to depend on mesa being present as part of the build, not really surprising seeing that Slackware (and most other distros) ship with mesa out of the box. The only other way around I can see would be to recompile all xorg drivers without mesa support, which would be a mistake in my view.

I'm a bit busy at the moment, but maybe in the next week or so I can get to the bottom of all this nonsense and release a decent alpha.

Those who want to persist with Intel testing, try "pfix=ram i915-modeset=0" either at the "puppy" (boot prompt from cd) or on the grub/grub4dos kernel line.

Nvidia and radeon users can try "nouveau-modeset=0" or "radeon-modeset=1" respectively, or if you already have a usable desktop with an ati/amd card with 'vesa' driver you can edit /etc/modprobe.d/radeon.conf and change the 0 to 1.

Also, search for xf86-video-modesetting and install it (slackware repo) if you have modesetting set at 1 (..and mesa-8.0.4 too)

FOOTNOTE: none of the above applies if you use a proprietary video driver.

Remember this is testing and a few save files may get burned!

____________________________________________________________________

On an unrelated note, gparted fails to work on ntfs partitions, it requires ntfsprogs, which aren't in the repos. Slackware 13.37 package should work ok, untested though so YMMV. It is not supported to invoke gparted for working on the partition you have booted off, and should give you a kernel error.

Once again, thanks for reports.

Finally got around to testing on my P4 with Intel 865 graphics. Boots to a black desktop requiring a hard poweroff.
Trying "pfix=ram i915-modeset=0" now......... no improvement.
Trying pfix=ram,nox ..........chose the "vesa" driver and 1024x768 and got 640x480 but at least it's a desktop.No really usable but a desktop....

Was able to reach a desktop on the 3 laptops using cd
boot and pfix=nox-vesa.
Not one was capable of making a wireless connection.

Hi Mick

Got a working 5.3.5alpha by using the grub4dos edit facility to temporarily add nox to the boot line - and then selecting the vesa driver.

Found that I am unable to blacklist a module - see attached where b43 is both blacklisted and loaded simultaneously.....

Cheers
peebee

Looking good, You Have already corrected the background setting problem. The only other problem that I've noted is when using PPM, clicking on "Examine dependencies" closes the dialogue box without showing the dependencies so a program with dependencies cannot be installed using the PPM.

Thanks to your fine nouveau-unload, the NVIDIA driver was also easily installed.

peebee wrote:Hi Mick

Got a working 5.3.5alpha by using the grub4dos edit facility to temporarily add nox to the boot line - and then selecting the vesa driver.

Found that I am unable to blacklist a module - see attached where b43 is both blacklisted and loaded simultaneously.....

Cheers
peebee

Thanks peebee for persisting.

I wonder if it's really loaded? I did have a similar issue with nouveau being reported as loaded but I knew it wasn't (with lsmod). The nvidia driver still installed fine, if nouveau is loaded it can't install.

Can you please zip (or gzip) your /etc/rc.d/MOSULESCONFIG file to make sure it made the skip list, or just post the 'skiplist' variable.

Cheers

pemasu's version of Bibletime here works, but only from a prompt, it will not work using the bibletime.desktop file, the menu, or clicking on /usr/bin/bibletime. It does execute from a prompt with the following error message.

# bibletime
bibletime: /usr/lib/libcurl.so.4: no version information available (required by /usr/lib/libsword-1.6.2.so)
#

It would be great to have it in your puppy slacko repository that shows up in the PPM if it can work from the menu.

Thanks,
Jim

01micko wrote:I wonder if it's really loaded? I did have a similar issue with nouveau being reported as loaded but I knew it wasn't (with lsmod). The nvidia driver still installed fine, if nouveau is loaded it can't install.

Can you please zip (or gzip) your /etc/rc.d/MOSULESCONFIG file to make sure it made the skip list, or just post the 'skiplist' variable.

Cheers

Hi Mick

I wanted to blacklist b43 and did so in the boot manager - but it will not go away! It is loaded and in use despite being blacklisted - see netwiz screenie attached after blacklisting and rebooting.

MODULESCONFIG is attached - skiplist has b43 appended so it should not be loaded....

Thanks
peebee

01micko,
I found the reason for the blacklisting not working. Script /sbin/pup_event_backend_modprobe, line 103, needs the "-b" option, to use the blacklist. Here is what I ended up with:I will add that to my collection of fixes for woof.

Looking further, I think the same addition should be made to the other 2 similar statments, line 117 and 153, to restore the preference operation. This may explain peebee's problem with module wl, which is related to the preference, ssb:wl.

My interpretation of all this is that until now the -b option was the default, and that now it is required. Otherwise, why have it?

To keep things simple, I attach the corrected module loader, including other woof fixes for firmware loading and modules loaded specifically by udev rules. Please test with it. Thank you.
Richard

I've had a play on my old Toshiba A200 laptop - running from a live CD

pfix=ram gets me to a black screen - maybe a failed x
pfix=ram, nox gets me to the prompt ok, but I can't remember how to start the command line x config wizard
Then made a copy of my slacko save and just let the puppy have it's way (ie update the save file etc - I got a desktop but no mouse control - neither the touchpad nor usb mouse. Maybe no keyboard controls either?

Sorry I can't be more helpful.

rerwin wrote:My interpretation of all this is that until now the -b option was the default, and that now it is required. Otherwise, why have it?

To keep things simple, I attach the corrected module loader, including other woof fixes for firmware loading and modules loaded specifically by udev rules. Please test with it. Thank you.
Richard

Thanks Richard.

I'm sure this is to do with the new "kmod" implementation which replaces "modules-init-tools", mentioned in the first post under the "developer" heading. There were sure to be some bugs turn up with that and I'm glad you found the cause. I'll put your fixes in the next version.

-

grump.. don't worry too much about this alpha, it was to mainly to put it in the wild and see what bugs were flushed out. Fortunately, many have been and should be squashed for the next alpha.

For general users it might be better to wait for either the next version or beta, which will be some weeks yet. Realistically, I can't go beta until Slackware goes final as they will likely be applying patches at least until RC.

Cheers

01micko wrote:

rerwin wrote:Please test with it. Thank you.
Richard

Thanks Richard.
I'm sure this is to do with the new "kmod" implementation which replaces "modules-init-tools", mentioned in the first post under the "developer" heading. There were sure to be some bugs turn up with that and I'm glad you found the cause. I'll put your fixes in the next version.
Cheers

Richard's fix pet works for me - thanks Richard
Cheers
Peter

USB Logitech Receiver not recognized, however unplugging it and replugging it allowed it to be recognized.

Although after using "Desktop Templates for desk icons" and shutting down, another problem occurred. Upon reboot there were no icons, mouse did not work, but did work after unplugging and replugging the receiver. Selecting "Desktop Templates for desk icons" gave an error message that the pinboard is not working.

After another reboot, USB Logitech Receiver is not recognized even after unplugging and replugging.

Cheers,
Jim

Jim, I don't know why your logitech wireless setup isn't loading but I suspect it's something to do with rerwin's fix. Can you install rerwin's pet and reboot please?
I'll have to fix the desktemplates since I have removed slickpet.

----------------------------------------------------------------------------------------------

Not so good news with Intel. I can't get it to work at this stage. I have recompiled several kernels, with/without modesetting, no difference. It does seem that whether modesetting is enabled or not in the kernel is not the issue, but it must be turned off at the command line level (done with the i915.conf rule) for the majority of hardware, radeon included. The string you add to the kernel line is:

i915.modeset=0 (tested and it boots direct to vesa on my eee pc, and brookdale)

Radeons it is already off. Nvidia should not matter.

The current kernel has not got the modesetting kernel modules for X anyway. It seems they are needed for the newest hardware, Ivy-bridge and Sandy-bridge intels, Evergreen radeons and newer. I have what i thought was a reasonably new radeon but it is low end, kernel modesetting does not work with it. There is also a userspace xorg driver called "modesetting". I may have to eliminate it and people who need modesetting can download it.

Still, all this is a waste of time if I can't get the Intel xorg driver to load!!! I get hold of a 945G (I think) tomorrow so will have a go with that. The 915 in my EEE 701 only boots to 'vesa' and seems can only use vesa. When Intel loads there is a "no screens found" error in the xorg log. Same with my Brookdale, the log from it and the EEE are almost identical.

I'm going to install slackware-current on my eee pc right now to see if I can debug.

You reckon testing is fun? Try developing!

01micko. Did you remember that logitech wireless receiver fix in kernel DOTconfig. There was that hid section need for compile that locitech receiver in, not as module. I posted about it in some of your thread, maybe this one. I can dig the needed DOTconfig configuration again.

01micko. Did you remember that logitech wireless receiver fix in kernel DOTconfig. There was that hid section need for compile that locitech receiver in, not as module. I posted about it in some of your thread, maybe this one. I can dig the needed DOTconfig configuration again.

One letter answer ... "Y" . That was the first thing I checked when I saw the report, but I do think it's the bug rerwin fixed.