First, a big thanks for the help I got so far (reg. kapm-idled thread)...
Today I've experimented some more, and found the following:
- even ALSA 0.90beta11 can't solve the suspend problem. I also noticed
something rather strange, that I didn't realize before (due to
auto-loading of modules):
When I try to load "snd-card-intel8x0" for the first time,
modprobe complains about the card not being found or being
busy.
When I try to load the same module a second time,
everything works "fine"...
But - WHY would linux at first say, the card isn't there, but
find and initialize the card on the second run?
- Due to the problem with ALSA, I went back to try the kernel
sound drivers instead and found the i810_audio to work fine
(and - more importantly - not disturb the suspend... ;)
("historically" I switched to ALSA on a TP600 which was a pain
to get sound working with kernel drivers, but which was fairly
easy to do with ALSA at the time... ...and since then I hadn't
even tried the kernel drivers any more...)
But - the following problems remain:
-> Hibernation doesn't work at all (this used to work on the TP600
and on the TP A21P I had before)...
-> trying to load agpgart (to make DRI work) fails with:
Linux agpgart interface v0.99 (c) Jeff Hartmann
agpgart: Maximum main memory to use for agp memory: 816M
agpgart: Detected an Intel 830M, but could not find the secondary device.
agpgart: no supported devices found.
-> When the system resumes from a suspend, the following message
turns up in dmesg:
APIC error on CPU0: 00(40)
(The system appears to work fine, but I'd still like to know,
what the error is for...?)
-> prism2_pci (linux-wlan-ng-0.1.10; linux-wlan-ng-0.1.13pre[1-4])
WLan throughput is VERY low (~18KB/s, as opposed to ~550KB/s
on a PCMCIA card, linked to the same Wireless access point;
also - Win2K also reaches 550KB/s on the internal prism2
chip)...
Any help on any of the above would be highly appreciated! :)
Benedikt
BEAUTY, n. The power by which a woman charms a lover and terrifies a
husband.
(Ambrose Bierce, The Devil's Dictionary)
[email protected] writes:
> First, a big thanks for the help I got so far (reg. kapm-idled thread)...
>
> Today I've experimented some more, and found the following:
>
>
> - even ALSA 0.90beta11 can't solve the suspend problem. I also noticed
> something rather strange, that I didn't realize before (due to
> auto-loading of modules):
> When I try to load "snd-card-intel8x0" for the first time,
> modprobe complains about the card not being found or being
> busy.
> When I try to load the same module a second time,
> everything works "fine"...
> But - WHY would linux at first say, the card isn't there, but
> find and initialize the card on the second run?
I had been running ALSA 0.9.0beta8a for some time, because whenever I
upgraded beyond that with either 2.2 or 2.4 kernels I would get similar
problems, in particular crashes when trying to suspend my laptop.
Eventually I decided to give ALSA 0.9.0beta11 a go with 2.4.17, and
then my machine would crash just trying to play sounds.
So I cursed for a while and then took another look at the ALSA
documentation. They no longer say to modprobe "snd-card-foo", but just
"snd-foo" for your driver of choice. In addition some other aspects of
the module hierarchy for ALSA seem to have changed, such that I gathered
trying to install a newer ALSA module stack over the old one would leave
some old modules around.
Once I had a from-scratch install of ALSA 0.9.0beta11 over a
from-scratch install of 2.4.18-rc1 and its modules, and modified
modules.conf to load snd-intel8x0 instead of snd-card-intel8x0,
everything started working fine. In fact, I no longer have to have the
apmd suspend script to unload all the ALSA modules before a suspend and
reload them on resume -- the APM support in ALSA seems to have improved
significantly.
On Sat, 23 Feb 2002 23:53:53 +0100 (CET), [email protected] wrote:
> -> Hibernation doesn't work at all (this used to work on the TP600
> and on the TP A21P I had before)...
> -> When the system resumes from a suspend, the following message
> turns up in dmesg:
>
> APIC error on CPU0: 00(40)
This indicates that your A30P has a local-APIC capable P6-class
cpu, and that you're not using the latest 2.4.18-pre or -rc kernel.
The APIC error at resume from suspend is fixed in current 2.4.18-rc,
so a simple kernel upgrade will silence that message.
Your machine survives APM suspend? That's encouraging since I've
had a report that the T20 doesn't if the local APIC is enabled.
What's the difference between suspend and hibernate?
Does the machine survive if you pull the power cord or enter the
BIOS setup screens?
/Mikael
> > -> When the system resumes from a suspend, the following message
> > turns up in dmesg:
> >
> > APIC error on CPU0: 00(40)
>
> This indicates that your A30P has a local-APIC capable P6-class
> cpu, and that you're not using the latest 2.4.18-pre or -rc kernel.
>
> The APIC error at resume from suspend is fixed in current 2.4.18-rc,
> so a simple kernel upgrade will silence that message.
Hmm... OK, I might upgrade to the 2.4.18-rc; unless - any vague clue,
when the "regular" 2.4.18 should be due? Since the message is not
part of a crash or anything, I'd also happily wait for another few
days... ;)
> Your machine survives APM suspend? That's encouraging since I've
> had a report that the T20 doesn't if the local APIC is enabled.
Yes, it does... Consistently...
> > -> Hibernation doesn't work at all (this used to work on the TP600
> > and on the TP A21P I had before)...
> What's the difference between suspend and hibernate?
> Does the machine survive if you pull the power cord or enter the
> BIOS setup screens?
I don't remeber, whether you could even enter the BIOS setup screen
from hibernation, so I can't answer that...
In general - hibernation causes the notebook to dump the entire
system memory contents and system status to a hibernation file
(which is a contiguous hidden file on the Windows partition). Once
the dump is done, the machine powers off completely - AC power cord
and batteries can be safely removed/replaced during the time. When
you switch the system back on, and the hibernation file contains
system data, the RAM and system data is read back, and the system
can resume, from where it is (obviously just all active the
network connections will be gone... On the TP 600E, TP A21P this
worked fine).
The lack of hibernation is not a major problem for me though, as
going to suspend and staying in suspend for 4-8 hours eats less
battery than dumping 1 GB of RAM to disk and rereading it... Also,
suspends for up to about 8 hours conserve battery even compared to
rebooting; so most of the time I just suspend the machine for
transport... I just hibernate or turn off the machine if it's going
to be off for a longer period of time, or if I have to board a
plane with it... ;)
Benedikt
BEAUTY, n. The power by which a woman charms a lover and terrifies a
husband.
(Ambrose Bierce, The Devil's Dictionary)
On Sun, Feb 24, 2002 at 01:22:40PM +0100, Benedikt Heinen wrote:
>
> > > -> Hibernation doesn't work at all (this used to work on the TP600
> > > and on the TP A21P I had before)...
>
> > What's the difference between suspend and hibernate?
> > Does the machine survive if you pull the power cord or enter the
> > BIOS setup screens?
>
> I don't remeber, whether you could even enter the BIOS setup screen
> from hibernation, so I can't answer that...
>
> In general - hibernation causes the notebook to dump the entire
> system memory contents and system status to a hibernation file
> (which is a contiguous hidden file on the Windows partition). Once
> the dump is done, the machine powers off completely - AC power cord
> and batteries can be safely removed/replaced during the time. When
> you switch the system back on, and the hibernation file contains
> system data, the RAM and system data is read back, and the system
> can resume, from where it is (obviously just all active the
> network connections will be gone... On the TP 600E, TP A21P this
> worked fine).
> The lack of hibernation is not a major problem for me though, as
> going to suspend and staying in suspend for 4-8 hours eats less
> battery than dumping 1 GB of RAM to disk and rereading it... Also,
> suspends for up to about 8 hours conserve battery even compared to
> rebooting; so most of the time I just suspend the machine for
> transport... I just hibernate or turn off the machine if it's going
> to be off for a longer period of time, or if I have to board a
> plane with it... ;)
You might want to try out the swsusp patch by Pavel Machek.
It is the beta (or even alpha) stage, but according to Pavels
latest comments it is starting to work quite well.
>From the patch:
+Software Suspend
+CONFIG_SOFTWARE_SUSPEND
+ Enable the possibilty of suspendig machine. It doesn't need APM.
+ You may suspend your machine by either pressing Sysrq-d or with
+ 'swsusp' or 'shutdown -z <time>' (patch for sysvinit needed). It
+ creates an image which is saved in your active swaps. By the next
+ booting the kernel detects the saved image, restores the memory from
+ it and then it continues to run as before you've suspended.
+ If you don't want the previous state to continue use the 'noresume'
+ kernel option. However note that your partitions will be fsck'd and
+ you must re-mkswap your swap partitions/files.
Regards,
Toon.
--
/"\ | Windows XP:
\ / ASCII RIBBON CAMPAIGN | "Sorry Dave...
X AGAINST HTML MAIL | I'm afraid I can't do that."
/ \