Impressions of 2.6 kernel

[mzilikazi]

Hmmmm. OK, I tried:

Section "InputDevice"
	Identifier  "PS/2 Mouse"
	Driver      "mouse"
	Option      "Protocol" "auto"
Option          "ZAxisMapping"          "4 5"
#	Option      "Device" "/dev/psaux"
	Option      "Device"  "/dev/input/mice" 
	Option      "Emulate3Buttons" "true"
	Option      "Emulate3Timeout" "70"
	Option	    "SendCoreEvents"  "true"
EndSection

Methinks all you need to change is the Protocol and the above should work. IIRC 'auto' doesn't work w/ 2.6.*

[Cuddles]

CrashedAgain,

Here is what I have for my PS/2 Mouse, which is a wheel mouse, connected to my PS/2 port connector, and its just a "generic" wheel mouse - It doesn't even say the manufacturer on the durn mouse, thats how "generic" it is - but I get scroll "up and down" in Konq., Mozilla, Spreadsheets, Word Processors, and mail... Here's my XF86Config-4 "snippet"

Section "InputDevice"
        Identifier  "PS/2 Mouse"
        Driver      "mouse"
        Option      "Protocol" "auto"
Option          "ZAxisMapping"          "4 5"
        Option      "Device" "/dev/psaux"
        Option      "Emulate3Buttons" "true"
        Option      "Emulate3Timeout" "70"
        Option      "SendCoreEvents"  "true"
EndSection

I've never touched a thing, from the hard drive install of kernel 2.6.6 w/ Knoppix v3.4 - so this is "un-tainted" code...

Hope this helps,
Ms. Cuddles

[Durand Hicks]

My impression of the 2.6 kernel is that it doesn't do the suspend to ram as good as the 2.4 kernel. That, the usb bug, and the memory epplet breakage are the chief reasons why I'm not using 2.6. I have a modified 2.4.26 with a modified xfree86 savage server running dri that enables my 2.4 kernel to run with nearly the same speed and responsiveness of the 2.6 kernel. Plus, it doesn't have the exploit bug that I tried to crash it with, so for now, this 2.4 kernel does it for me.

[CrashedAgain]

Got it working. The key is:

Option "Protocol" "IMPS/2"

Thanks for the help.

[jjmac]

The 2.6 is certainly going though a fast development
cycle. The only quirk I've found is with s/w that
relys on the kernels system calls to query the
boot sectors table geometry, via the MBR. But then,
my bios is old, and only deals in 24 bit geometry,
aka, 255 heads.

It hasn't been a problem, but i had to update my
Lilo to the latest version to get it to update
it's table in 2.6.x smoothly... It would always
respond with a fatal warning about geometry
miss-matches .

The whole system runs considerably faster though.
I use the "spirals" screen saver from the
xscreensaver collection... with the 2.4.x kernel
it was quite hypnotic, but with the 2.6.7 it runs
like a rocket. Very happy .

The 2.6.x pre-empts and scheduals (sic) more
efficiently i believe. It is a development kernel
though, in spite of the even minor number. And so,
needs to be followed on that score.

Running knoppix v3.3, that i copied over directly
from a LiveCD to disk, and using the 2.6.7,
from another Linux (deb). There's been marked
improvement in performance.

I'm fascinated to know what you mean though by...

Durand Hicks wrote:
>>
My impression of the 2.6 kernel is that it doesn't do the suspend to ram as good as the 2.4 kernel. That, the usb bug, and the memory epplet breakage are the chief reasons why I'm not using 2.6.
>>

???


>>
I have a modified 2.4.26 with a modified xfree86 savage server running dri that enables my 2.4 kernel to run with nearly the same speed and responsiveness of the 2.6 kernel.
>>

Modifications... do you mean the pre-emptive patches
for 2.4.x. I have those, but possibly not
all. And the 2.4.24 is noticeably slower, though
faster than with out the patches.


>>
Plus, it doesn't have the exploit bug that I tried to crash it with, so for now, this 2.4 kernel does it for me.
>>

Maybe thats been fixed ???


jm