“I have started to port the linux ACPI subsystem because ACPI is becoming more important now and slowly replaces subsystems like the pci routing table, multi-processor table and apm. The Linux ACPI code is based on the os independent intel reference implementation and so the port has been very easy (and fast to do) so far. Currently the ACPI busmanager contains the basic acpi code (about 90% of the code).”
Good luck with that. It seems that ACPI support is a real PITA for people who aren’t Microsoft.
For example, on most of my machines, FreeBSD and DragonFly ACPI works fine, and Linux’s and NetBSD’s does not. The base ACPI code for all those projects comes from the same Intel code. I personally know quite a few people who have the exact opposite experiences as mine.
ACPI is a real sore spot for most OS projects who bother to try.
I’m subscribed to the mailing lists, and from that it seems there’s not much happening any more with Syllable, or at least the basic systems. Glad to see that’s not the whole picture
I’m subscribed to the mailing lists, and from that it seems there’s not much happening any more with Syllable, or at least the basic systems. Glad to see that’s not the whole picture
I’m obviously subscribed to the m-l’s too, but I think you may have joined the wrong ones then! There’s quite some activity in the m-l’s, as seen in this screenshot:
http://img353.imageshack.us/img353/1525/syll3cx.png
And that’s just one of the two active m-l’s!
Ehrr..ok I’m on the same list. But obviously having Gmail’s condensed view of multiple messages and several other high volume mailinglists drowning Syllable out in my inbox might have something to do with it.
Linux’ ACPI implementation doesn’t work for 90% of the laptops out there. My question is: why bother?
You’re copying the wrong ACPI implementation.
My question, on the other hand, is which would you suggest yourself? (it worked on three laptops I had in hands)
I’d suggest reverse engineering the implementation all manufacturers test with.
Either you got very lucky with your laptops, or you never used much functionality of ACPI. Did STR work?
I’d like to know which laptops you tried it on.
Then again, you claim it worked (did you sell the laptops, or did it stop working after a while?) on three laptops you had in hands. But you never said how many you had on which it didn’t work
I’d suggest reverse engineering the implementation all manufacturers test with.
That’d be the Windows one. Good luck reverse-engineering that piece of crap.*
*) No, “piece of crap” is not too harsh. NVIDIA and ATI manage to build implementations of OpenGL compatible with SGI’s reference implementation. The fact that Microsoft’s isn’t compatible with Intel’s reference implementation (when ACPI is an Intel technology!) suggests that their programmers are either too stupid to implement the spec properly, or too arrogant to care.
Of the ~30 laptops that Ubuntu tested for 5.10 (see http://wiki.ubuntu.com/LaptopTestingTeam ), ACPI suspend to RAM works on 25 of them. That doesn’t strike me as too bad.
The problem with ACPI implementations is that everyone is implementing it. The hardware manufacturer and the operating system are both ‘implementing’ the ACPI standard. The problem is that it’s just not a very well designed standard, and it’s entirely possible for everyon to do a reasonable approximation of ‘implementing’ the ACPI standard, and then find that their bits don’t work with each other. Of course, all the hardware manufacturers usually care about is whether their implementation works with Microsoft’s implementation…even if it doesn’t work with Linux’s, or BSD’s, or Syllable’s.
90%?
I’ve owned 3 laptops, acpi worked 100% on them all, everyone I know that runs linux and uses acpi it works for also.
Very well, I believe you all. I must be seeing things then.
But just look at the list of “fixed” DTST’s on the ACPI website. It’s not 10 or 20 or so. It’s more like 1000 of them. Every model needs a new one. Pathetic.
And “patching” your DTST isn’t easy for the casual user. I stand my ground; Linux’s ACPI doesn’t work out-of-the-box for most laptops.
I’m sure that most laptop’s can be fixed. I’ve tried to get mine fixed for over a year, without luck. So please forgive me for not giving much liking to Linux’s crappy ACPI implementation.
But just look at the list of “fixed” DTST’s on the ACPI website. It’s not 10 or 20 or so. It’s more like 1000 of them. Every model needs a new one. Pathetic.
It is pathetic, but also far away from being Linux’s fault. Those DSDT tables do *not* compile cleanly with Intel’s IASL compiler.
You should blame the mainboard manufacturers for not being able to come up with a proper DSDT table — not Linux…
So please forgive me for not giving much liking to Linux’s crappy ACPI implementation.
Again, it’s your laptop’s crappy ACPI implementation.
most likely tested on a test setup from microsoft…
str works perfectly on my toshiba portege 3110ct running debian.