|
|
Post by joncrndl on Feb 22, 2011 0:25:28 GMT -5
I tried booting a friend's new laptop with knos64 both from a stick and the DVD. It is looping through a message on screen.
ACPI Error: No handler or method for GPE[0x??], disabling event (20101013/evgpe-734)
with variations that replace ?? with characters like hexand what ever follows "event".
btw - this laptop was sick before, tried knos. It failed to find its hard drive. I brought the beast home for a further look.
It came with Win 7 Home Premium 64.
***************
Stay tuned. I just stopped the system and restarted. This time the hard drive came to life. It just came up in windoze safe mode.
Go figure Win7 is back!
|
|
|
|
Post by Kevin McAleavey on Feb 22, 2011 2:17:19 GMT -5
I tried booting a friend's new laptop with knos64 both from a stick and the DVD. It is looping through a message on screen. ACPI Error: No handler or method for GPE[0x??], disabling event (20101013/evgpe-734) with variations that replace ?? with characters like hexand what ever follows "event". btw - this laptop was sick before, tried knos. It failed to find its hard drive. I brought the beast home for a further look. It came with Win 7 Home Premium 64. *************** Stay tuned. I just stopped the system and restarted. This time the hard drive came to life. It just came up in windoze safe mode. Go figure Win7 is back! Yes ... I've seen two other people with similar reports. Different models, but apparently the same motherboard and BIOS problem in the ACPI portions have come in by email. When the machine doesn't boot up, there's no way to get the diagnostics data to figure out what's wrong and so we're hoping one of these fails happens to get caught with a screen capture with a digital camera so we can see the fail code and figure out what's wrong. We had something like this with some Toshibas in beta3 and it turned out to be that BIOS was pointing to a device that was never installed on the motherboard when the machines were built and the manufacturer failed to tell BIOS that it wasn't there on those. The BSD folks fixed that in their acpi code and we've had no complaints on RC1 so far from those folks. None of this of course affects WIndows much since the machines are built to Microsoft's specs, but it can be trouble for any other OS that isn't expecting to see bad BIOS code telling it lies. Heh. But if bootup fails and you can take a picture of the screen just before it kisses the sidewalk, that'd be greatly appreciated and allow us to figure out just where it's going over the edge and fix it. ACPI is a PIA when the manufacturer doesn't follow the rules when they assemble machines.  |
|
|
|
Post by joncrndl on Feb 22, 2011 11:07:44 GMT -5
I have sent pictures of the screen. Good luck!
|
|
|
|
Post by joncrndl on Feb 22, 2011 11:26:23 GMT -5
knoppix 6.2 does load on the HP. Is there some query from Knoppix that would be helpful to you?
|
|
|
|
Post by Kevin McAleavey on Feb 22, 2011 19:28:48 GMT -5
knoppix 6.2 does load on the HP. Is there some query from Knoppix that would be helpful to you? THANK YOU for those screenshots of the failed boot! You've given me a very useful clue there! As for the knoppix stuff, sorry to say "Linux Is Not UniX" (which is what Linux means) will be of absolutely no use because KNOS really IS unix, not a fantastic simulation. Heh. But what you sent had a very valuable clue to me in there. I'm going to email you later once I do a little more research and can find some time to write up a possible test for you to run on your end that may very well solve the problem. It looks like HP made a very serious mistake in their BIOS code and if my guess here is correct, I just might be able to route around their damage for the benefit of all! THANK YOU!!!  |
|
|
|
Post by joncrndl on Feb 23, 2011 0:34:16 GMT -5
No luck. It fails at the same place with the same errors.
What is next?
|
|
|
|
Post by Kevin McAleavey on Feb 23, 2011 1:51:42 GMT -5
No luck. It fails at the same place with the same errors. What is next? Try increasing that value to 128, though I'm pretty sure it won't matter. There's a hole in the BIOS chain on their chip and the way the BSD code works now, we believe the values that we get instead of trying to step over it. I'll go ahead and file a PR with BSD over it and will look at the acpi kernel code myself. I can't modify the kernel code itself because doing so would create a "trust" issue with our customers - it's policy here never to touch the BSD code itself until more folks know they can trust us not to do anything untoward in there. That way, they can fetch that source and build their own if they must. I'll also get after HP and get back to you, but it might take a couple of days to get the answers I need. That 128 value is for branches within the BIOS stack. I can't believe any manufacturer would place that much cruft in BIOS deliberately and so I think they've got a null pointer in there somewhere and I'll need to see about downloading a BIOS image for that machine and checking that myself. Sorry for the headaches ... |
|
|
|
Post by joncrndl on Feb 23, 2011 11:46:26 GMT -5
The olny id for this bios is F.14
Good luck on your end. I will try bumping it up to 128 and get back to you.
|
|
|
|
Post by Kevin McAleavey on Feb 27, 2011 1:49:00 GMT -5
The olny id for this bios is F.14 Good luck on your end. I will try bumping it up to 128 and get back to you. Pity we couldn't play with that one longer, but as you indicated in email it wasn't even working right with Windows. Hope that guy managed to get himself a better machine, from what we went through back and forth that machine was truly bad news all around. |
|
