The Mobile Intel 4, or GM45, Express Chipset is pretty common on newer laptops, among them my Toshiba A305, where its driver is identified "Mesa DRI Mobile Intel(r) GM45 Express Chipset 20080716x86/mmx/sse2". Intel provides code info to everyone, so this one is not a proprietary graphics card. But no puppies from 2.17 to 4.2x are able to provide Xwindows for this chipset.
I installed a fat SUSE 11.1 to see if the problem is Xorg or puppy. SUSE's Xorg installer set up Xwindows with no problem, using, according to display info, a "1.4mesa7.2" driver. Thus the problem in puppy seems to be something removed from the puppy Xwindows implementation, or not updated to.
Can whatever is presently left out from puppy's Xorg be added so we with the Mobile Intel 4 chipset can use puppies? The installer recognizes that the monitor is LCD and 1280x800, but apparently doesn't have that to install and so tries a CRT at 1020x768 application, which might work for an added CRT monitor plugged into a port, but just puts vertical lines on the monitor0 LCD.
VESA provides a tiny square of a corner of a graphic desktop screen in multiple copies all over the monitor. Thus, with this chipset there is no graphic option at all in puppy. To we who are addicted to puppy this is a serious problem.
Can a "fatX-pup", one with the full Xorg, be made? Or can the left out parts be put in a pet, like the win-codecs, with a CL install option maybe?
Thanks whatever is possible.
M'goo
X-Windows and Mobile Intel(r)4 Express Chipset
SUSE doesn't do lspci and Puppy 4.1.2 loaded nox doesn't load far enough to provide any useful information. loaded with xorg it kicks back to edit profile, but not to nox, and with vesa it loops so ctl-alt-backspace only shuts the display down momentarily, so the computer has to be hard-reset. The kernel used is the stock 4.1.2 seamonkey puppy kernel.
Worse, puppy doesn't load enough nox, or backed out from xorgwizard to find my thumb-drive, so I can't out-load anything to send from another system.
I should be able to write the tweak profile, or parts, on paper and type them into an e-mail. Would that help?
m'g
Worse, puppy doesn't load enough nox, or backed out from xorgwizard to find my thumb-drive, so I can't out-load anything to send from another system.
I should be able to write the tweak profile, or parts, on paper and type them into an e-mail. Would that help?
m'g
Hmmm .....
Well, as far as I can tell SlaxerPup (Puppy 412 based) is using a newer kernel that supports newer Intel IGP chipsets. That should work fine. You may want to give it a shot ...
If you insist on your current Puppy 4.1.2 then it's ok to start with nox. At the prompt, do the lspci and look for "Class 0300". That's the display pci id.
Then do a 'uname -a' and look for the kernel version. There's two kernel versions, one is 2.6.21.7 (so-called retro) and 2.6.25.16 (so-called standard).
I have a patched intel-agp module already done (recently) for kernel 2.6.21.7. I have backported the necessary codes from kernel 2.6.30 ..... The intel-agp module for kernel 2.6.25.16 is WIP ... I'm running the retro kernels most of the time ...
Yeah, kernel version is a must-have info .....
Rgds
Well, as far as I can tell SlaxerPup (Puppy 412 based) is using a newer kernel that supports newer Intel IGP chipsets. That should work fine. You may want to give it a shot ...
If you insist on your current Puppy 4.1.2 then it's ok to start with nox. At the prompt, do the lspci and look for "Class 0300". That's the display pci id.
Then do a 'uname -a' and look for the kernel version. There's two kernel versions, one is 2.6.21.7 (so-called retro) and 2.6.25.16 (so-called standard).
I have a patched intel-agp module already done (recently) for kernel 2.6.21.7. I have backported the necessary codes from kernel 2.6.30 ..... The intel-agp module for kernel 2.6.25.16 is WIP ... I'm running the retro kernels most of the time ...
Yeah, kernel version is a must-have info .....
Rgds
The kernel for my P-412 is the stock non-retro 2.6.25.16, the so-called standard. The "class 0300" lines, provided by lspci -v are:
00:02.0 class 0300:8086:2a42 (rev 07)
Subsystem
ff1e
Flags:Bus Master, fast desel, latency 0, IRQ 11
Memory at d0000000 (64-bit nonprefetchable) [size=4M]
Memory at c00000000 (64-bit prefetchable) [size=256M]
I/O ports at 5110
Expansion ROM at <unassigned> <disabled>
Capabilities: [90] message signalled interrupts: 64-bit queue=0/0 enable
Capabilities: [d0] power management version 3
Info for "class 0380" is very similar, though less.
I run P-412 from cd and hd-pup-save, transferring out to usb for solid storage, so I can flush by deleting or changing to another pup-save if anything triggers my paranoia. my cd version is only slightly tweaked, with firefox, openoffice and xzgv added to provide functionality I need. I need Xorg for the keyboard-language options. So far I haven't had good luck with P-42x. All have loaded fine on my older laptops, but have the same problem on the Toshiba A305. I haven't yet had time to sort out and adjust a 4.4x to work for me [I need to rip all the Apple-like features, which put me off, perhaps irrationally, for my having had to work with Apples for a time, which I found like trying to work on rails: You could only go where Apple's rails carried you, unless you got right down to unix-mechanics level to rip up and move the tracks over. Those who teeth on Apple seem to not notice the constrictions]. And so far 412 has done well for me, and been easy to tweak when I've needed.
Something I've just noticed, from trying different Linux distros in this situation is X installs multihead, installing what Xorg used to call xinerama, with the Mobile Intel 4 chipset. In my present Toshiba A305 case the dual display installation is with the built-in display listed lcdv at 1280x800x24 as monitor0 and a second display set at 1024x728 with h and v sync ranges for a crt monitor that can be added through the video port. In the Puppy Xorg setups on this machine, Xorg correctly finds monitor0 an lcd set for 1280x800 and offers 16 layers as default and 24 as an option. Then when X-test fails [it makes all vertical lines, no x-window cross-hatch, then opens a vertical white-space in the middle, to fail] when one kills and goes to tweak xorg.conf the monitor is listed with the crt 1024x768 pixels, horizontal frequency 48.56KHz and vertical refresh 61.00 Hz.
I wonder from this if the problem might be something as simple as that the Mobile Intel 4 Chipset confuses the xorg setup with two sets of monitor instructions and that xorg is set to default to the second instead of the first set of parameters?
On previous laptops I have had to physically attach a second monitor before X would allow xinerama multihead capability to even be setup [it was needed by the oggle video-player].
I tried removing or rewriting the crt monitor identification lines in the xorg.conf file in the puppy install tweak>edit option, but all changes caused the system to change all values to 0.00.
I hope some of this may be of value in sorting this oddity.
I will download a retro-kernel and a Slaxer-Pup and try them, then report how they do.
Thanks
m'goo
00:02.0 class 0300:8086:2a42 (rev 07)
Subsystem
ff1eFlags:Bus Master, fast desel, latency 0, IRQ 11
Memory at d0000000 (64-bit nonprefetchable) [size=4M]
Memory at c00000000 (64-bit prefetchable) [size=256M]
I/O ports at 5110
Expansion ROM at <unassigned> <disabled>
Capabilities: [90] message signalled interrupts: 64-bit queue=0/0 enable
Capabilities: [d0] power management version 3
Info for "class 0380" is very similar, though less.
I run P-412 from cd and hd-pup-save, transferring out to usb for solid storage, so I can flush by deleting or changing to another pup-save if anything triggers my paranoia. my cd version is only slightly tweaked, with firefox, openoffice and xzgv added to provide functionality I need. I need Xorg for the keyboard-language options. So far I haven't had good luck with P-42x. All have loaded fine on my older laptops, but have the same problem on the Toshiba A305. I haven't yet had time to sort out and adjust a 4.4x to work for me [I need to rip all the Apple-like features, which put me off, perhaps irrationally, for my having had to work with Apples for a time, which I found like trying to work on rails: You could only go where Apple's rails carried you, unless you got right down to unix-mechanics level to rip up and move the tracks over. Those who teeth on Apple seem to not notice the constrictions]. And so far 412 has done well for me, and been easy to tweak when I've needed.
Something I've just noticed, from trying different Linux distros in this situation is X installs multihead, installing what Xorg used to call xinerama, with the Mobile Intel 4 chipset. In my present Toshiba A305 case the dual display installation is with the built-in display listed lcdv at 1280x800x24 as monitor0 and a second display set at 1024x728 with h and v sync ranges for a crt monitor that can be added through the video port. In the Puppy Xorg setups on this machine, Xorg correctly finds monitor0 an lcd set for 1280x800 and offers 16 layers as default and 24 as an option. Then when X-test fails [it makes all vertical lines, no x-window cross-hatch, then opens a vertical white-space in the middle, to fail] when one kills and goes to tweak xorg.conf the monitor is listed with the crt 1024x768 pixels, horizontal frequency 48.56KHz and vertical refresh 61.00 Hz.
I wonder from this if the problem might be something as simple as that the Mobile Intel 4 Chipset confuses the xorg setup with two sets of monitor instructions and that xorg is set to default to the second instead of the first set of parameters?
On previous laptops I have had to physically attach a second monitor before X would allow xinerama multihead capability to even be setup [it was needed by the oggle video-player].
I tried removing or rewriting the crt monitor identification lines in the xorg.conf file in the puppy install tweak>edit option, but all changes caused the system to change all values to 0.00.
I hope some of this may be of value in sorting this oddity.
I will download a retro-kernel and a Slaxer-Pup and try them, then report how they do.
Thanks
m'goo
non-update
This is an update to provide some info while I continue to screw around as I have time to.
I downloaded Slaxer and P-4.2.1 with retro kernel. But both stop with a "pup-421[or 412].sfs file not found. dropping out to initial ramdisk console" and "/bin/sh: can't access tty; jobcontrol turned off". This is right after the loading of disk drivers and occurs even with pfix=ram. This suggest drivers not loading, even for the CD. Or a losing of the CD between loading the kernel and booting it, since the CD is where the pup.sfs file is to be found [note, it also fails to find a pup.sfs on a flashdrive, or find and start (blink) the drive].
I've had the pup.sfs file not found hang before, on other computers, usually after I've added and removed too many different systems. I've cured it with repartitioning. I can't rip and replace the NTFS [Vista] in this case, so I'll try putting a fat-distro linux back on, see if puppy can bounce off it to find its CD source-disk and get around the choke point.
I downloaded Slaxer and P-4.2.1 with retro kernel. But both stop with a "pup-421[or 412].sfs file not found. dropping out to initial ramdisk console" and "/bin/sh: can't access tty; jobcontrol turned off". This is right after the loading of disk drivers and occurs even with pfix=ram. This suggest drivers not loading, even for the CD. Or a losing of the CD between loading the kernel and booting it, since the CD is where the pup.sfs file is to be found [note, it also fails to find a pup.sfs on a flashdrive, or find and start (blink) the drive].
I've had the pup.sfs file not found hang before, on other computers, usually after I've added and removed too many different systems. I've cured it with repartitioning. I can't rip and replace the NTFS [Vista] in this case, so I'll try putting a fat-distro linux back on, see if puppy can bounce off it to find its CD source-disk and get around the choke point.
Hmmm .....
I hope you are ready for more messing around. I've completed the intel kernel module updates (for retro & standard) and posted them here.
I've also made a couple of xorg intel driver pets that hopefully works with your hardware. If you're willing to give them a shot, please post some results. Thanks.
Rgds
I hope you are ready for more messing around. I've completed the intel kernel module updates (for retro & standard) and posted them here.
I've also made a couple of xorg intel driver pets that hopefully works with your hardware. If you're willing to give them a shot, please post some results. Thanks.
Rgds