Suspend doesn't work proppely - 8.04, Darter Ultra (daru2)

[laserline]

Hello,

I would like to report that Suspend doesn't work propperly in the Ubuntu 8.04 (stable).

I've installed both versions, 32BIT first, then 64BIT and I have the same results.

The installation procedure on both was to wipe my drive (not upgrading from Gutsy 7.10), then I added System76 Repos and installed the latest driver (2.18), then I started the System76 application. Went to "Install Driver" --> "Install". Then, I rebooted.
I didn't choose "Restore System" because I already put the repositores manually.

When I suspend the system, all is good, the power button and indicator led blink. I tried waking the system by pressing P1, P2 buttons, the WIFI/BT button, move the mouse, press all keys on the keyboard, the cdrom eject button - nothing worked.

When I press the power button, the system tries to resume, it looks OK for a second, but then while the screen is still blank, the system Shuts Down (un-cleanly I might add) -- As I said before, this happens on 32BIT and 64BIT of Ubuntu Hardy 8.04 (Stable).

Is there something to do with this ?
Hibernation works OK.

Thanks in advance,
Idan Mashaal

[laserline]

UPDATE:
I've upgraded System76 driver to 2.20 and did a full Restore and reboot - that didn't help.
Instead it installed GnuCash, so I started reading the source to see what it does, and something is weird.
I traced all the things that are supposed to be going on in Ubuntu 8.04 with 'daru2' and for some reason, I don't see any ACPI fixes (although the batter works after installing the driver. (driverscontrol.py line 78 sound.alsa4() )

I also found out that it adds sources, installs GnuCash (and docs), overwrites the driver and GSynaptics.
About GSynaptics, it doesn't fix the SHMConfig "true" in xorg.conf (so why install it?) -- lines 80-82 in base_system.py

In adition to all that, I can't find a thing about ACPI with the daru2...

Is the driver really ready for Ubuntu 8.04 ?

Thanks,
Idan

[AusIV4]

I suspect the driver will get this fixed fairly soon, but in the mean time, can you show us your /etc/default/acip-support file?

[laserline]

Sure, the files is attached.

I just noticed that the battery monitor got screwed also (I even rebooted)

Idan.

[reh4c]

I still have the same battery indicator problem (wacky readings) with 8.04 64-bit. There's a bug open on launchpad, just search for PR200 (same OEM laptop as Daru2/MSI-1221). Needless to say, suspend doesn't work on mine either...same problem as you. The only other issue that I've confirmed is with Cheese and the webcam. Cheese now functions very poorly (i.e. freezing, video in window lags when dragging, etc.) Please let us know if you're having the same problems with that as well.

[laserline]

Hi, just installed Cheese from Ubuntu's repos, and I'm experienceing the same issues as you. Slow FPS, video lagging when dragging.

Taking a picture works.
Shooting a video, makes the app stall and stop responing when pressing "Stop Recording"

The system-76 driver assumes the stock Ubuntu one is ok, but I guess it needs some tweaking.

I really do hope they will be able to fix/resolve issues with Hardy and Daru2. ( I'm kinda disappointed :confused: )

Thanks,
Idan.

[AusIV4]

Line 10 of your acpi-support file reads:
ACPI_SLEEP_MODE=mem
Try setting it to:
ACPI_SLEEP_MODE=standby

[laserline]

I'll try that and post later on.

Isn't that something the driver should handle (or a custom modification) ?
I guess it's not specific to my machine only but to all daru2's out there, as I did a clean installation.

Thanks,
Idan

[laserline]

Hi AusIV4,

I've made the change, rebooted and afterwards tried to suspend my daru2.
The system went to suspend mode (power button and the power led blinking) but waking up, from the power button (the only one that got the system to respond) caused the system to wake up but shutdown - uncleanly.

What log holds these events, maybe the answer is there ??

Thanks,
Idan.

[laserline]

Hello again,

I've found something quite disturbing...Something I overlooked before.
The findings I've found are the same with i.e, ACPI_SLEEP_MODE=mem and ACPI_SLEEP_MODE=standby (I've reverted back to the defaults "mem")

Here goes:

When I try to wake my daru2 (by pressing the power button) it looks alive and after a few seconds shutdown (screen is blank), then I've waited for another 3-4 seconds and the system came back ON (by itself -- the power button and power led were OFF!) all the leds were up and running, fans whistling, HD working, Screen Blank. I've waited something like 30 seconds and there was no response except from the fans working (hard), I tried pressing all the buttons I can find, nothing worked. I pressed the power button, and it ignored me. I pushed and hold the power button for 10 seconds and nothing happend. Then I felt HOT air from my fan (HOT as in the system is doing something, but there was no sign).
I've finally shut it down by taking the battery out :( :(

I had a dual-boot with XP on the system and I never had something like this, I frankly use suspend really often.

I might add that the battery readings are way off... and the battery seems to get lost from the graph. Even when charging, and I know that after 2 hours the battery led should turn off, when Ubuntu 8.04 is running, the battery never gets off, and every charge cycle the Power Manager's graph shows a new cycle (I hope it's not hosing my battery).

Any ideas what can be causing this ??

Thanks,
Idan.

[badbull]

i have the same suspend problems in hardy as i had in gutsy. the suspend to ram don't work.
in gutsy i changed the ACPI_SLEEP_MODE to standby and it worked (but not with blinking lights)
but i think this don't work in hardy. i changed the option but it goes on memory sleep and crash on wakeup, like you described above.

[tuebinger]

I'm having the same problems with the suspend feature. I hope it gets fixed soon, as I'm having to shut down every time I use the computer. And I was just beginning to enjoy the suspend feature working properly under 7.10!:(

[thomasaaron]

We *think* we've got the answer, and it involves rolling a custom kernel.
It will be a few days, but we are hard at work on it.

[maartenlameris]

Is there any progress in the fix for this problem?
Or a alpha version we can test?

[thomasaaron]

Not quite yet. R&D is still working on it. It's a tough one.

[laserline]

Can't we just diff both stable ubuntu kernels and see what has changed and why...

Not talking about diff the whole source, but the appropriate modules/drivers..

Idan.

[thomasaaron]

That behavior is typical.

Right now, suspend (S1 and S3) only work in 64-bit Gutsy. They are defunct in Hardy because of a regression of some type in the Kernel, which we are still trying to locate and fix. (Hibernate does work fine in Hardy, though.)

If you really need suspend, you will have to revert to Gutsy 64-bit and install/run the System76 driver (which will fix the battery monitor problem as well).

If you do not want to revert to Gutsy 64-bit, but you do want the battery monitor fixed, you will need to add...

acpi=noirq

...as a kernel option in your /boot/grub/menu.lst. This will mess up your brightness function keys, though. But it's the best fix we have at the moment.

[Cogito²]

Just to clarify, is hibernate working (in general)? Because it doesn't work on my laptop (Compaq V2000). It stops in the middle of booting up and nothing happens. Suspend also doesn't work, but as I've gathered I shouldn't expect it to...

[greg_g]

Cogito, just so you know, this thread is about a specific laptop, namely the System76 Darter Ultra. If your laptop (a Compaq) is not performing suspend and/or hibernate correctly, it would be best to search http://www.launchpad.net for issues about it, or check this link for more information/hints/tips: https://wiki.ubuntu.com/LaptopTestingTeam/Compaq

But to answer your question, there are many laptops that do perform suspend and hibernate flawlessly. The main factor is if your laptop manufacturer uses components that are not made by companies which hate supporting anything but Windows. </mini_rant>

Best,

Greg

[Cogito²]

Cogito, just so you know, this thread is about a specific laptop, namely the System76 Darter Ultra. If your laptop (a Compaq) is not performing suspend and/or hibernate correctly, it would be best to search http://www.launchpad.net for issues about it, or check this link for more information/hints/tips: https://wiki.ubuntu.com/LaptopTestingTeam/Compaq

But to answer your question, there are many laptops that do perform suspend and hibernate flawlessly. The main factor is if your laptop manufacturer uses components that are not made by companies which hate supporting anything but Windows. </mini_rant>

Best,

Greg

Haha well that explains what "Darter Ultra" means...I wasn't quite sure and upon my scan of this thread I didn't quite figure it out (nor realize it was of importance). I'll look into those links. Thanks for the info.

[digitalbenji]

I have a Darter 1 (white darter). Out of the box, 32 bit Hardy suspend and hibernate worked fine. The last Kernel update (-17) nerfed both suspend and hibernate for me. It looks like someone submitted a bug on this, https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/236272

For now, I will be using the -16 kernel. What version are you having your problem with? You can determine this by typing uname -a

Thanks,

[thomasaaron]

Actually, this post is not for the white Darter (daru1), it is for the newer Darter (DarU2). The source of the problems are radically different.

With the white Darter is probably just a configuration issue. We can do some testing on it.

[laserline]

Hi Tom,

Where are we standing in the development of the fix ?

Idan.

[walkeraj]

I have tried everything at https://bugs.launchpad.net/ubuntu/+source/linux/+bug/226279 and I have had no luck with this problem. I have also installed the 2.6.24-19-image package from hardy-proposed. That also did not work.

The inability to suspend is very frustrating and, frankly, compromises the usability of the notebook somewhat. I'd be very interested to know the status on this problem too.

I was informed at one point that a notebook had been sent to a kernel developer (which is a really good idea). Did anything come of that?

[thomasaaron]

It is still with the developer. They're continuing to work on it. We're not giving up on it.

[MarkID]

Hmmm. A few days ago I tried suspend on my Daru1 and lo and behold, suspend started to work, where it hadn't after the Hardy upgrade. I can even get my wifi back after coming out or suspend. Maybe it was one of the system updates, or maybe it's just magic. (Also, the dimming problem seems to have solved itself.)

[thomasaaron]

Right. But this thread pertains to the DarU2. It is a completely different set of issues.

[maartenlameris]

Is there any progress made on this issue I would like to upgrade to 8.04 but need my suspend to work.

If needed i'm happy to do some testing in this matter.

[arch_o_median]

Hi there! I'm running a gazelle performance gazp3. I also haven't been able to suspend/hibernate since I upgraded to Hardy 8.04. I was able to suspend/hibernate in gutsy 7.10.
When I hibernate via the keyboard Fn->F1 key combination the screen blanks, the disc light blinks and then the system goes completely off as desired. When I try to reactivate it, the screen never displays anything, the fan comes on and runs until I hold down the power button.

[arch_o_median]

My system info is my sig.

[thomasaaron]

Please post that in a separate thread.

Suspend/Hibernate issues are a little different for each of these models. This thread was dedicated to the DarU2, and the fix will not be anything similar to the GazV3.

Also, let us know what processor you are using so we can set up tests in-house.

[walkeraj]

Is there any kind of blog or something we can go to to see the status on this as it develops? While I'm sure the work is ongoing, it would definitely help to know how it's developing so we don't all keep poking our heads in here asking for a timetable.

As a developer, I have learned that a major premise of good UI is some kind of visual indicator of activity, even if it just says "reticulating splines" or somesuch nonsense. You know?

[thomasaaron]

I'm going to talk to Carl about starting an email list or something so that we can provide updates to interested folks.

[walkeraj]

Sweet. I'll friggin' make the whole company fresh gingerbread if you can fix this in the next two months.

[damphoud]

Have you guys been able to figure it out? I have a 1221 as well, and I'm unable to suspend. I don't expect any trouble shooting, as my laptop is not system76.

[maartenlameris]

Is there any progress on this problem?

[Nuno Bettencourt]

Hi, even though i don't own a Darter Ultra, i have a Tsunami Traveller 1275R which is also a rebrand of the MSI-PR200 and I'm having all the difficulties described above running Hardy 64bits kernel 2.6.24-19.

My resume after suspend wakes up the computer and after 2-3 seconds shuts it down. After 3-5 seconds wakes it up again, can ear fans, hdd, see lights coming up, but the screen is blank.

My hibernate mode seems to do it's job.

My battery readings, after enabling acpi=noirq also works, but as said, brightness controls FN+F4 and FN+F5 stopped working. Regarding this, a few days ago i had a Fedora 9 live cd which i tested only for a couple of minutes, and i don't think the problem with AC and battery ocurred. I'll try to post some more info about this.

[gaussian]

I have Darter 2 and I haven't updated to Hardy mostly because of the suspend problems mentioned in this thread. This is going to be slightly off topic, but I went and downloaded "Daily Live CD" images for next version (8.10). Note that this is not even yet in the Alpha stage (or more precisely in the current alpha releases there are no live-CD's).

Some good news: using yesterdays (7/15) daily live-CD's under both 32-bit and 64-bit one of the suspend modes worked perfectly. This means that
$sudo -s
$echo "standby" >/sys/power/state
resulted computer being frozen up like it should be and then returning after hitting the power button. This is the light sleep mode (no blinking lights).

Then the bad news:
$sudo -s
$echo "mem" >/sys/power/state
resulted computer going to "deep sleep" properly, but not returning from there like it seems to do under hardy.

It looks like I am going to stick to Gutsy until 8.10 comes out.

P.S. 64-bit live-cd worked really well otherwise (battery indicator seemed to behave, no screen brightness cycling etc), 32-bit had some bad video/starting X issues. I also could not get wireless to work under either on of them but I am really not worried about that.

[vacula]

I've tried Fedora 9 Live CD few weeks ago - problem with hibernate, battery and touchpad still occured there and I'm pretty sure the root of this all is invalid acpi on ms1221 :(

[walkeraj]

Some good news: using yesterdays (7/15) daily live-CD's under both 32-bit and 64-bit one of the suspend modes worked perfectly. This means that
$sudo -s
$echo "standby" >/sys/power/state
resulted computer being frozen up like it should be and then returning after hitting the power button. This is the light sleep mode (no blinking lights).

Then the bad news:
$sudo -s
$echo "mem" >/sys/power/state
resulted computer going to "deep sleep" properly, but not returning from there like it seems to do under hardy.

This is hardly a change, I'm afraid. As far as I know, regular standby has been working. Unfortunately, this doesn't do a heck of a lot for laptops as most of the board still remains powered.

It looks like I am going to stick to Gutsy until 8.10 comes out.

Unfortunately, I don't think a new version number is going to make any difference unless this problem is fixed in the kernel between now and then. Given the fact that this problem hasn't received a lot of attention from kernel developers and has been ongoing for months with no visible progress whatsoever, I wouldn't hold your breath for 8.10.

Speaking of this ongoing problem, has there been any progress? What happened to that status report and feedback you were going to try and give to us? It's been three weeks and nary a peep.

[williumbillium]

This is hardly a change, I'm afraid. As far as I know, regular standby has been working. Unfortunately, this doesn't do a heck of a lot for laptops as most of the board still remains powered.

I thought the issues with Hardy are the exact opposite of gaussian's ibex testing results. According to the first post in this thread, hibernate (to hard disk) works, but suspend (to ram) does not.

Unfortunately, I don't think a new version number is going to make any difference unless this problem is fixed in the kernel between now and then. Given the fact that this problem hasn't received a lot of attention from kernel developers and has been ongoing for months with no visible progress whatsoever, I wouldn't hold your breath for 8.10.

Speaking of this ongoing problem, has there been any progress? What happened to that status report and feedback you were going to try and give to us? It's been three weeks and nary a peep.

Agreed, I'd like to hear more updates too. Not being able to upgrade is really starting to get frustrating, especially since I bought my Daru2 only a few weeks before Hardy came out :(

Edit: Also, I am subscribed to several bugs in launchpad. Perhaps if more people subscribed to them they'd get more attention. They don't have many subscribers right now...

https://bugs.launchpad.net/ubuntu/+source/hal/+bug/215033
https://bugs.launchpad.net/system76/+bug/225347
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/139701/

Not sure if this one is directly related to the Daru2 issues, perhaps someone knows?

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/226279

[Nuno Bettencourt]

Edit: Also, I am subscribed to several bugs in launchpad. Perhaps if more people subscribed to them they'd get more attention. They don't have many subscribers right now...


Great idea, even though I've got a MSI pr200 clone, I subscribed to those threads that made sense to me and reflected my problems. Maybe this way it will have some more impact.

[gaussian]

I thought the issues with Hardy are the exact opposite of gaussian's ibex testing results. According to the first post in this thread, hibernate (to hard disk) works, but suspend (to ram) does not.



Actually, my very limited testing had nothing to do with hibernate. It was between S1 and S3 sleep. S3 sleep is what we want (power/state=mem), S1 is what we have working (power/state=standby). I claim no expertise on this, see:
http://en.wikipedia.org/wiki/Advanced_Configuration_and_Power_Interface#Global_ states

Has anyone (System 76?) tried going through the procedure explained here: https://wiki.ubuntu.com/DebuggingKernelSuspend ?

[thomasaaron]

Not sure. I'll pass that link on to R&D.

[williumbillium]

Actually, my very limited testing had nothing to do with hibernate. It was between S1 and S3 sleep. S3 sleep is what we want (power/state=mem), S1 is what we have working (power/state=standby). I claim no expertise on this, see:
http://en.wikipedia.org/wiki/Advanced_Configuration_and_Power_Interface#Global_ states

Got it. Looks like I was a little ignorant to the different states. Thanks for the wp link.

[thinman1189]

Not sure if this has been said already or not but I just got my serval performance today and suspend won't work.

[thomasaaron]

Might be a good idea to start a new thread dealing with suspend on servals. The DarU2 is a completely different issue.

Be sure to give more detail in the thread about the symptoms.

[thinman1189]

Okay, I don't have time now but I'll try to suspend again tomorrow and post the error.

[williumbillium]

I noticed that the bug for this issue in the System76 launchpad has priority:undecided. Shouldn't this be upped? This is a showstopper for me.

https://bugs.launchpad.net/system76/+bug/225347

[walkeraj]

Yeah, that's what prompted my comments about attention and progress, frankly. It honestly appears from outside that no one really cares about this bug...:(

[williumbillium]

Sweet. I'll friggin' make the whole company fresh gingerbread if you can fix this in the next two months.

I think it's also worth noting that we're now 1 month into the "fresh gingerbread if you can fix this in the next two months" period. I'm going to up the ante by saying that if this is fixed in the next month, I'll bake brownies for the whole company too :)

[walkeraj]

I think it's also worth noting that we're now 1 month into the "fresh gingerbread if you can fix this in the next two months" period. I'm going to up the ante by saying that if this is fixed in the next month, I'll bake brownies for the whole company too :)

I say draft that up and have it notarized. Let's make this official.

[williumbillium]

Perhaps the Ubuntu kernel team meeting, this tuesday,

http://blog.phunnypharm.org/2008/08/ubuntu-kernel-team-irc-meeting.html

would be a good opportunity to raise this issue with the Ubuntu kernel team?

[walkeraj]

Three weeks now. I'm determined not to let this thread leave the first page until we hear SOMETHING about SOME kind of progress.

[thomasaaron]

I appreciate your tenaciousness. However, the guys in R&D and the devs looking into this one probably do not see this post.

This is a very tough one, and we are still working on it. I promise.

[arch_o_median]

I've been running:

sudo /etc/acpi/hibernate.sh

It works about 70% of the time....

[walkeraj]

I've been running:

sudo /etc/acpi/hibernate.sh

It works about 70% of the time....

That's probably because, like your signature says, you're running a gazelle and not a Darter Ultra 2, which is what this thread is about. I appreciate you trying to be helpful, but, in the future, try and read the thread a bit more before you post.

Edit: This is, in fact, the second time (http://ubuntuforums.org/showpost.php?p=5239497&postcount=29) you've incorrectly commented on this same thread about this same issue.

[gaussian]

An aside: Overall when it comes to hibernating with my Darter, I have found that using s2disk (part of uswsusp or suspend package, depending on the version of Ubuntu) works more reliably than using the standard (pmi) hibernate.

Back to regularly scheduled programming: Suspending (S1) Darter...

[laserline]

Hi Tom,

I found this thread http://ubuntuforums.org/showthread.php?t=810222&highlight=suspend

Apparently, the daru2 isn't alone at all.
Dell, Compaq and other laptops suffer from the same problem.

Is this problem Ubuntu specific or Linux specific ?

I wonder if Fedora / Mandriva have the same problems. Did R&D try installing a diffrent distro ?

Thanks,
Idan.

[thomasaaron]

Apparently, the daru2 isn't alone at all.
Dell, Compaq and other laptops suffer from the same problem.


Is this problem Ubuntu specific or Linux specific ?


I wonder if Fedora / Mandriva have the same problems. Did R&D try installing a diffrent distro ?

Thanks,
Idan.

[thomasaaron]

Apparently, the daru2 isn't alone at all.
Dell, Compaq and other laptops suffer from the same problem.

Honestly, I'm not sure if the problem with Dell and Compaq are the exact same issue. As far as I know, it is pretty unique to the DarU2 and it deals with some conflicts between Ubuntu's tickless kernel and the DarU2's hardware.

Is this problem Ubuntu specific or Linux specific ?

I wonder if Fedora / Mandriva have the same problems. Did R&D try installing a diffrent distro ?

I know it exists under Fedora. I believe some customers have tried other distros without success.

[gaussian]

Honestly, I'm not sure if the problem with Dell and Compaq are the exact same issue. As far as I know, it is pretty unique to the DarU2 and it deals with some conflicts between Ubuntu's tickless kernel and the DarU2's hardware.



I know it exists under Fedora. I believe some customers have tried other distros without success.

Can confirm this under Mandriva (2008.1 Live-CD, 32 bit) and Open-Suse (Live-CD, I think it was 64 bit).

[glacialfury]

As far as I've seen, it's more a Hardy problem; resuming from suspend apparently worked well in Gutsy. Many users across many different hardware sources are having this problem.

Several have fixed it using "quirks". I don't profess to understand it all, but perhaps those more knowledgeable would look into:

http://people.freedesktop.org/~hughsient/quirk/quirk-suspend-index.html

[walkeraj]

Nope. It is not just a hardy problem, it is everywhere. It's the new kernel that came with Hardy. If you downgrade your kernel in Hardy, suspend/hibernate revert to the way they were before (albeit with a host of other problems).

These "quirks", are basically different methods of hardware-specific suspend stored in hal-info, and should already be up-to-date on all of these systems. I'll give a few of them a try and post any successful results here.

[luckytwitch]

So, I'm a new Ubuntu user, but after cruising through these forums and following lots of links, I found something that works for my system (Kubuntu 8.04 and KDE on my Sony Vaio VGN-CR123E). I'm not quite sure what it actually does, but my suspend works now.
Here's the link that I found, https://bugs.launchpad.net/ubuntu/+source/linux/+bug/226279, and the following is the code that I used to make suspend work:echo "SUSPEND_MODULES=ehci-hcd" > /tmp/unload_modules
chmod +x /tmp/unload_modules
sudo mv /tmp/unload_modules /etc/pm/config.d
If anyone can explain what this actually does, it'd be cool, thanks

[walkeraj]

(from wikipedia (http://http://en.wikipedia.org/wiki/Host_controller_interface)):

A host controller interface (HCI) is a register level interface which allows a host controller for USB or FireWire to communicate with the operating system of a personal computer.

On the expansion card or motherboard controller, this involves much custom logic, with digital logic engines in FPGAs plus analog circuitry managing the high speed differential signals. On the software side, it requires a device driver (called a Host Controller Driver, or HCD).

Enhanced Host Controller Interface (EHCI) is a high speed controller standard which is publicly specified.

So, what's happening is that, on your particular laptop (which is totally different from the one being discussed here), either the usb chipset itself or the driver is causing problems with suspend. By adding
SUSPEND_MODULES=ehci-hcd

to /etc/pm/config.d, you are telling linux to unload this module from the kernel before it suspends and reload it when it returns. By having the kernel explicitly unload this module, it ensures that it is not loaded when the actual suspend occurs, and ensures that it is cleanly removed and re-inserted into the running kernel when you come back from suspend. This is a general way to avoid modules that cause problems with suspend. Unfortunately, as I mentioned above, this is a TOTALLY different laptop. Which modules cause these problems (and indeed whether it is a module that causes it or not) is very hardware dependent and will differ greatly between even laptops of the same general make from the same manufacturer.

In the case of this laptop, it definitely appears to be something more than a simple SUSPEND_MODULES directive.

Thanks for trying to help, though.