Vista poisons ACPI?
in [Linux Hardware]
|
From: Time Waster on 10 Jun 2008 10:24 Got our new laptop, whoopie! HP dv6700t. Shrunk the Vista disk down while installing FC9. Twiddled, rebooted back to Linux a number of times, everything great. Booted Vista to make sure it could still be done. Next boot to Linux, it froze coming up -- something about an ACPI threshold (?) or somesuch, sorry this will be vague. Retried a couple of times with the same result. Booted with noacpi, and it works again. Anyone know what Vista might have done to the machine, and how to reverse the damage? (Besides of course wiping off Vista completely and using only Linux -- which would be fine with me, but I have to share the machine.) TIA!
From: Aragorn on 10 Jun 2008 10:58 On Tuesday 10 June 2008 16:24, *Time Waster* wrote in /comp.os.linux.hardware:/ > Got our new laptop, whoopie! HP dv6700t. Congratulations. :-) > Shrunk the Vista disk down while installing FC9. Twiddled, rebooted > back to Linux a number of times, everything great. Booted Vista > to make sure it could still be done. Next boot to Linux, it froze > coming up -- something about an ACPI threshold (?) or somesuch, sorry > this will be vague. Retried a couple of times with the same result. > Booted with noacpi, and it works again. > > Anyone know what Vista might have done to the machine, and how > to reverse the damage? All I can say at this stage is something I seem to recall from older BIOS versions that came with a write protection for the CMOS data, which one needed to either enable or disable via the BIOS set-up program. This information is old, so I'm not sure as to whether things have in the meantime changed or not, but considering Microsoft's persistence in using archaic technology, I wouldn't be surprised. The concept was that for ACPI to work in Windows, the CMOS write protection had to be disabled, and thus logic dictates that Windows requires write access to your BIOS set-up data in the CMOS memory, also known as /nvram/ (Non-Volatile RAM). Modern BIOSes however do not seem to have such a write protection anymore, presumably because BIOS and CMOS manufacturers know that Microsoft won't abandon this practice. So presumably, Vista has altered something in the CMOS settings for ACPI, resulting in boot problems for GNU/Linux. And *that* would of course be totally consistent with Microsoft's vantage that a computer should be running only one operating system, and that this operating system should be one that's listed among Microsoft's commercial offerings. <evil grin> > (Besides of course wiping off Vista completely and using only Linux -- > which would be fine with me, but I have to share the machine.) Perhaps you should consider Virtualization, or better yet, persuade your better half - I presume we are talking of a female who co-owns the machine? ;-) - that it's wiser to run a real computer operating system on real computer hardware. :-) -- *Aragorn* (registered GNU/Linux user #223157)
From: Time Waster on 11 Jun 2008 10:33 In article <W72dnZVsuefhkdLVnZ2dneKdnZydnZ2d(a)pipex.net>, Sheridan Hutchinson <Sheridan(a)Shezza.org> wrote: >-=-=-=-=-=- > >Time Waster wrote: ><snip> > >Vista doesn't poison ACPI :) > >The fault is most likely a firmware bug in the BIOS. Is your BIOS >up-to-date? > >ACPI is a very finicky thing and in a lot of cases it's taken a while >for motherboard manufacturers to mature their code. > >Try flashing with a later version and please do let us know how you get on. Thanks for the advice. Haven't tracked down the BIOS version yet -- should've have done that while when i visited the settings. Took long enough for me to find the right function key during the boot as they flash 5 or 6 of them to select for a brief moment. Anyway, with some searching I found that noapic gets the job done, too -- it doesn't require noacpi. And from more googling I see that requiring noapic is fairly common, though i'm not sure what the penalty for using it is. I'm still suspicious since everything worked until i booted Vista once (after the Linux install). Maybe poison is too strong a word. ;)
From: Scott Alfter on 11 Jun 2008 20:45 In article <J8w3k.37067$bs3.11457(a)trnddc07>, Time Waster <noone(a)nowhere.com> wrote: >Shrunk the Vista disk down while installing FC9. Twiddled, rebooted >back to Linux a number of times, everything great. Booted Vista >to make sure it could still be done. Next boot to Linux, it froze >coming up -- something about an ACPI threshold (?) or somesuch, sorry >this will be vague. Retried a couple of times with the same result. >Booted with noacpi, and it works again. Vista isn't your problem; I have it on a machine that dual-boots between that and Gentoo, and it never has any trouble switching between the two. I have an HP notebook (L2000, specifically) that used to need noacpi to boot properly. I'm not sure when I quit using noacpi, but if you're using an older kernel (I'm using 2.6.24-gentoo-r5), you might still need it. I still have reboot=c and pci=assign-busses in my kernel options. You might give those a shot and see if they help. _/_ / v \ Scott Alfter (remove the obvious to send mail) (IIGS( http://alfter.us/ Top-posting! \_^_/ rm -rf /bin/laden >What's the most annoying thing on Usenet?
From: Aragorn on 12 Jun 2008 12:22
On Thursday 12 June 2008 06:37, *Sheridan Hutchinson* wrote in /comp.os.linux.hardware:/ > Time Waster wrote: > <snip> >> Anyway, with some searching I found that noapic gets the job done, >> too -- it doesn't require noacpi. And from more googling I see that >> requiring noapic is fairly common, though i'm not sure what the >> penalty for using it is. > > I should have suggested that, my apologies. > > I'm fairly confident that as long as you're using the latest stock > kernel for your distro and the latest BIOS version that you will no > longer require the noacpi toggle. > > Well the good news is that the plug'n'play won't be affected by the > switch. The bad news is that power management can become an issue for > the machine (probably not relevant if it's no a laptop) and that your > machine will use legacy interrupt tables. [...] Power management may become an issue even if you're not using a laptop, because things like thermal throttling also depend upon ACPI. Additionally, you need ACPI to enable hyperthreading on processors that have this feature. Furthermore, the legacy interrupt tables you're referring to are an issue of disabling the (local) APIC, which is not to be confused with ACPI. The APIC is to be found in the processors themselves, while ACPI is a matter of the BIOS and/or the operating system. ACPI = Advanced Control and Power Interface APIC = Advanced Programmable Interrupt Controller Your mileage may vary... :p -- *Aragorn* (registered GNU/Linux user #223157) |