PDA

View Full Version : Why is suspend and hibernate so much of a problem in linux?



mthakur2006
June 20th, 2007, 01:07 PM
Hi,
Why is suspend and hibernate so much of a problem in linux? is it because of the drivers or something else? on my fiesty, the clean install can suspend and restore succesfully but after installing the updates, it can't.
Why is it so much of a pain?
thanks.

Roger Gundberg
June 20th, 2007, 02:01 PM
A long time ago (under windows) I realized there may be a conflict of some sort with the BIOS. So I went into the BIOS and set all power management to 'disable' and handed all power management off to the software. This worked for me maybe it will work for you?

king0lag
June 20th, 2007, 04:24 PM
Same problem here, at first it would give me a friendly little "doh" but now it doesn't work at all, doesn't give me any reason just pops back on :(

blah blah blah
June 20th, 2007, 04:30 PM
It's a kernel thing.

PriceChild
June 20th, 2007, 04:37 PM
If you use binary drivers, e.g. nvidia or fglrx (ati/radeon/nv/i810 etc. are ok) then you can forget resume.

Basically the binaryness means linux developers don't know how to reinitialise the card and so your resume fails.

Use purely open source drivers and you "should" be ok.

mthakur2006
June 20th, 2007, 04:39 PM
If you use binary drivers, e.g. nvidia or fglrx (ati/radeon/nv/i810 etc. are ok) then you can forget resume.

Basically the binaryness means linux developers don't know how to reinitialise the card and so your resume fails.

Use purely open source drivers and you "should" be ok.

i have opensource intel drivers, a conexant driver but it used to resume before when i had the conexant driver.
the only thing i think would have caused it would be this power management thread: http://ubuntuforums.org/showthread.php?t=419772&highlight=ipw2200+power+save :confused:

prizrak
June 20th, 2007, 05:40 PM
The answer to your question is the following:
Every manufacturer implements their ACPI in their own way that is somewhat different from others. Because of that it may work well on one machine and not on another. Perhaps there was a bug fix that enabled suspend/resume on ASUS machines but that broke it on Lenovo. It's kind of tricky with such crappy standardization.

strabes
June 20th, 2007, 09:13 PM
If you have an ati card follow the guide in the link in my sig

mthakur2006
June 20th, 2007, 09:15 PM
cool, but i have an intel card, the desktop has ati and it works fine, no need for ddrivers or anything

Polygon
June 21st, 2007, 09:56 AM
If you use binary drivers, e.g. nvidia or fglrx (ati/radeon/nv/i810 etc. are ok) then you can forget resume.

Basically the binaryness means linux developers don't know how to reinitialise the card and so your resume fails.

Use purely open source drivers and you "should" be ok.

wrong.

i use a ati card (ati radeon 9800 128mb) and i use the binary fglrx, and both suspend and hibernate work perfectly for me, except for a outstanding bug where alsa screws up and i get no sound when i come out of hibernate/suspend

but other then that it works perfectly.

DoctorMO
June 21st, 2007, 10:56 AM
I wouldn't mind but in the recent US anti trust case, papers show that Microsoft diddled the ACPI specs with each hardware vendor in order to prevent other os makers using ACPI. a very interesting read since it was the text between members of Microsoft and various hardware partners. you can see most of that stuff on Groklaw.

prizrak
June 21st, 2007, 01:58 PM
wrong.

i use a ati card (ati radeon 9800 128mb) and i use the binary fglrx, and both suspend and hibernate work perfectly for me, except for a outstanding bug where alsa screws up and i get no sound when i come out of hibernate/suspend

but other then that it works perfectly.

Yeah on Dapper my nVidia laptop would suspend/resume w/o an issue. The new nVidia driver breaks it tho.

king0lag
June 24th, 2007, 01:45 PM
Same problem here, at first it would give me a friendly little "doh" but now it doesn't work at all, doesn't give me any reason just pops back on :(

I have to take this back, I traced my problem to a faulty AC Adapter for my laptop. Hibernate now works perfectly again! :)

Final Solution: If hibernate is failing, ask your cats if they chewed your wires.

macogw
June 25th, 2007, 07:54 AM
I wouldn't mind but in the recent US anti trust case, papers show that Microsoft diddled the ACPI specs with each hardware vendor in order to prevent other os makers using ACPI. a very interesting read since it was the text between members of Microsoft and various hardware partners. you can see most of that stuff on Groklaw.

As usual, MS breaks standards. What else is new? They still, with all the thousands of programmers and millions of dollars they have, can't make a web browser follow standards. Of course they're trying to kill interoperability at every turn. It's the only thing they have left to keep people buying their crap.

sedd
June 26th, 2007, 11:44 PM
If you use binary drivers, e.g. nvidia or fglrx (ati/radeon/nv/i810 etc. are ok) then you can forget resume.

Basically the binaryness means linux developers don't know how to reinitialise the card and so your resume fails.

Use purely open source drivers and you "should" be ok.


wrong.

i use a ati card (ati radeon 9800 128mb) and i use the binary fglrx, and both suspend and hibernate work perfectly for me, except for a outstanding bug where alsa screws up and i get no sound when i come out of hibernate/suspend

but other then that it works perfectly.

Try this one: suspend/resume works for me with proprietary nvidia drivers, but NOT with opensource nv drivers. Its way more complicated than simplistic "opensource = good, proprietary = bad" FUD.

juxtaposed
June 27th, 2007, 12:03 AM
They still, with all the thousands of programmers and millions of dollars they have, can't make a web browser follow standards.

But since Internet Explorer is the standard (as it is the most used), doesn't it therefore automatically follow standards (if you are the standards, then you follow them...).