[ltp] Re: How to turn your T42p into a brick with ACPI...

[David A. Desrosiers]

> Dave: some comments

	Its "David" not "Dave, thank you.

> 1. don't make people on the list responsible for your frustrations. 
> That's very asocial when people want to help you understand 
> questions that you asked for help with.

	At no point did I point any blame at any person on this list 
or elsewhere. I'm not sure where you misinterpreted that, or 
incorrectly read that into my posts, but it was clearly not mentioned 
anywhere by me.

> 2. did you remember the kernel parameters pci=noacpi and 
> acpi_sleep=s3_bios ? I'm not sure if everyone needs them, but i do 
> on my 2374-HTG with current BIOS.

	If I want acpi, why would I disable acpi for pci? 

	In fact, after some more testing, it looks like I have to use 
acpi=off in the kernel anyway, otherwise my external usb peripherals 
fail to be recognized and time out (a very common, known problem, 
several thousand users are facing the same issue, and acpi=off solved 
it for them). 

	With acpi=off on the kernel load line, is it possible to still 
use acpi in userland? I doubt it. So now it looks like a choice of 
being able to use my external usb peripherals + suspend, or acpi 
without any external usb peripherals.

	But again, this requires some testing. Perhaps using acpi will 
make the usb devices start working properly, where the thousands of 
others seem to have failed.

> After that by the time linux is loaded, disks are mounted, services 
> are started (I assume you do work on the machine and need a couple 
> of services) and x is ready for login there is no way you are at 
> less that 45 seconds.

	I don't use a graphical login, and I run sawfish, and my 
sessions are managed by gnome-session (the only gnome'y thing I run). 
So if/when I _do_ start X, all of the apps I had running when I shut 
down X, including terminals, browsers (with populated tabs), and other 
things just fire right back up, exactly where they were when I exited.

> [If you're like me, you have are cunning enough to put the machine 
> in the dock, press the power button and go get a coffee.

	I have no need to put my machine in the dock, since I can't 
use it for anything more than another way to get power. None of the 
slots, peripherals, or drives in the dock are seen by Linux at all, 
whether on a cold or "warm" boot, so that's not really an issue. I'm 
unclear that using acpi will allow me to use all of the dock's bits, 
but I'd love to try it... assuming I don't have to disassemble my 
laptop every time it doesn't come back out of suspend.

> Using your method, you now have the pleasure of starting every 
> application you like to use (or if you scripted this, like i do for 
> my normal work environment) you will still need to navigate to the 
> proper documents, and if you have multiple terminals open you will 
> only have one $HISTFILE / user. <-- This is, for me personally, a 
> biggie because i have different terminals on different desktops 
> where i do different things. Suspend / Resume is a lot easier than 
> having differnt histfiles for each ptsX.

	Again, gnome-session handles all of this for me, completely 
transparently. Also, as I mentioned, I use sawfish, and it starts 
within a second or two. I don't use nautilus or icons, docks, wharfs, 
titlebars, windowframes, or any other screen-cluttering eye-candy, so 
my environment starts up very fast, without any delays.

> Using the acpi method with (oh god!) 45 second wait time, i enter my 
> password in xscreensaver and jump right back in where i was. All of 
> my VM's run perfectly without needing to suspend them each seperatly 
> as i used to have to, so there's another place to save time.

	Like I said, I'd be willing to try it, if someone could 
explain to me, in detail, how to get it working without bricking my 
laptop. Clearly there is something very different about the 1.8Ghz 
T42p models and my 2.1Ghz T42p.. that's what I'm gleaning from the 
various posts on the matter from those who have posted their results 
so far.

	Maybe its the BIOS revision, maybe EC, I'm not sure yet, 
because I haven't seen enough data from others on the list to do a 
comparison. The "It works on mine" posts aren't very helpful without 
enough detail to determine the differences, if any.

> 4. The coroner has pronounced this thread dead.

	Well, others still have their opinions and comments, and this 
thread is most-certainly not dead, even if you choose not to 
participate any longer.



David A. Desrosiers
desrod@gnu-designs.com
http://gnu-designs.com