Vista insomnia- definitely an OS issue

As soon as Vista went gold and began shipping we began hearing from mobile PC users that Vista was causing what Craig Pringle terms "Sleep of Death".  This term fits a number of different problems that notebooks and Tablet PCs exhibit running Vista when Sleep and Hibernate are used.  Sometimes when you put your mobile computer to sleep when you try to wake it up the system comes back and runs normally, except without the display which doesn't wake up.  I have seen this on three different PCs running Vista and lets just state the obvious, it's very hard to use a PC when the screen is not working.  It's impossible to even shut down the PC in this instance too, resulting in a hard kill of the OS.  This doesn't happen every time you put your device to sleep but often enough that it can create a real problem when you are mobile.

Another common problem that many mobile PC users have been reporting since the release of Vista is what Craig aptly terms "Vista Insomnia".  This phenomenon has potentially damaging results as it involves putting the device to sleep and having it turn itself back on at some point.  This is dangerous as it often turns the device back on while it's nestling in the gear bag and it can cause ultra-high temperatures that can actually damage the hardware.  Even the best case in this scenario the user can get to their next appointment to discover their battery has been depleted because the device has turned itself back on.

These two phenomena definitely came into existence with the release of Vista and users with computers bought recently with Vista pre-installed report them too, so it's not a case of older hardware failing to support Vista properly.  Craig has done some investigation and reports that he believes these problems with resuming are hardware driver related and not the fault of Vista.  I have to respectfully disagree because I believe that such an integral function of a mobile PC is fully the responsibility of the OS.  Maybe they happen because of the way that Vista reacts with the hardware drivers but even so the end result is so damaging that Vista has to bear the responsibility for preventing this from happening.  Craig points out that the upcoming Vista SP1 addresses some of these problems and I say that's not good enough.  These problems have been in place for the entire year now since Vista went gold and they should already have been fixed via Windows Update.  Why make millions of users wait for a fix in SP1 when it should be fixed NOW.  It's easy to make an argument that hardware drivers are the responsibility of the OEM and I understand that sentiment but let's face it, if the OS is not working for a large group of users then it is Microsoft's responsibility to fix it, and not next year with SP1.  Get these fixes to the users now and let's put this silly sleep business behind us as we should.

Interestingly, I have only seen the Sleep of Death once on the HP 2710p I have been using for a few months.  I have experienced both types of problems many times on several other Vista running notebooks however, including the Lenovo x61 Tablet PC which has exhibited both types of problems numerous times.

(Craig Pringle via GottaBeMobile)

Comments

[Elijah Buck]

I've had the sleep death issue on my thinkpad R61. Initially I was hard rebooting pretty quickly. Once though I let it sit, and after about 10 minutes, it blue screened. I looked at the crash dump with the windows debugging tools, and it did seem that the nvidia driver was at fault. I don't know what microsoft can do to make these drivers/hardware wake up better, but they definitely need to be working hard with the people who are making the drivers to get this issue resolved.

[Filip Norrgård]

An interesting fact that I discovered one day with the Sleep of Death: I had my external display connected to my notebook when this happened. The external screen worked (to my amazement) and pulled up the display properties to see what the damn problem was. Well, the display properties reported that the notebooks own screen wasn't -- so to say -- connected.

"What?", I thought. So, I pulled up and moved the device manager to diagnose what had happened. Well, what did you know... the notebook's own screen isn't listed in the device manager. And, scanning for any plug and play devices rendered no new devices. So, I pulled up the command prompt (which I could do if blind folded by the SoD):
1) Windows Key + R
2) Enter "cmd" and press Enter
3) Enter "shutdown /r /t 0" and press Enter
4) The result should be a less hard reboot... but still doesn't solve the problem...

I have previously believed that this was due to the year old ATI drivers that I'm forced to run with Vista, since this happened mostly after removing the external monitor and my computer woke from sleep.

[James Kendrick]

What Microsoft can do to make drivers work better is to A) define the reference sufficiently for OEMs to meet the requirements for the new OS and B) not allow OEMs to install Vista on devices with drivers that are known to not meet the requirements.

That said, if Microsoft can release updates in SP1 to help with these problems then they can in fact do something and the sooner the better.

[Kevin White]

I've seen this happen on my home laptop, a Toshiba something-or-other that came with Vista. It would go to sleep, and then wake up without the screen on.

I've also seen it at work, on a brand new Dell Latitude laptop. I had bungled the driver installation - dell makes it annoyingly hard to find out exactly what hardware is in your computer if you weren't the one to buy it - and somehow installed drivers for something the computer didn't actually have. As a result, sleep wouldn't work at all. Windows XP used to nag about this: "A device in your computer is preventing it from going to sleep!". Vista doesn't appear to, not that I've seen.

[Jordan Running]

I don't think I've had either of these problems with Vista on my X61t, but I've had a couple recurring problems with hibernate: 1) I'll tell it to hibernate and it will turn off the screen, but the computer will never turn off. A couple times I've just left it that way overnight and come back in the morning to find that it had rebooted itself. 2) Often when I wake it from hibernate, I'll swipe my finger to log in, then turn away for less than a minute to take care of something, and then turn back to find that it's going back into hibernate. Both drive me bonkers.

[Jake]

I think the only real way that Microsoft could enforce any rules on the driver manufacturers would be to insist on driver signing (as is the case with 64 bit Windows) and then test the code thoroughly themselves. This obviously opens a whole new can of worms and the 64-bit driver signing requirement is already contentious.

I don't do any driver development but I'm familiar with the documentation that Microsoft provide for other programming scenarios. As a result, I would be very surprised if Microsoft hadn't provided driver developers with everything they needed long before Vista went live last November. There has been lots of documentation explaining what programmers needed to do to ensure compatibility with Vista going back at least a couple of years now.

[IWorkThere]

Use your blogging superpowers to get someone to explain project velocity to you and how much effort was put into dragging hardware makers to update their f#cking drivers.

[Michael]

My machine does this most of the time when I switch between AC and battery while in sleep or hibernate. My strategy is to configure the power button for sleep (maybe default? don't know). Whenever the display is gone, I just press the power button shortly to put the machine into sleep mode again and after waking it up for the second time, the display is back on again.

[Paul]

I have the "random restart from sleep" issue on my HP nw8240 w/ XP SP2. Interestingly enough, this started happening after I updated the drivers for the ATI FireGL 5000 video card.

So this is not just a Vista issue.

[Dave Beauvais]

I've had the following happen with my HP 2710p: Three times the system did not enter sleep mode. I close the lid, the drive runs like it's going to sleep, then just does nothing. Every few seconds the drive LED will blink and I'll hear it doing something. If I wait long enough the system reboots and there's a bugcheck entry in the event log indicating a BSOD occurred.

I've also had one instance of trying to wake the machine from sleep mode only to have it hang. No drive activity, no display, just a power light. I let it sit for half an hour before cycling power.

[Daniel]

I have a Dell XPS m1330, with Vista Business (I did my own fresh install), and have had it for the past 3 months. I use the sleep functionality extensively, and have never had any issue at all with it. It's actually worked better than the Powerbook G4.

[Badcam]

And here's me thinking it was something I had been doing. I thought I was imagining things. Now I know that I wasn't. Come on MS, sort this out. Thankfully, I haven't had the "Vista Insomnia" problem as I keep my laptop in it's zero shock case a lot. I'd be gutted if something happened because of the lack of an MS update.

[SteveO]

I have a 2710p the is having the Sleep of Death problem frequently. I can see the screen with an external monitor, but I am usually not at my desk. So far, HP apparently does not know about the issue because: Sent back for another issue (might be related) and had the memory replaced. Called again, they sent someone out to replace the MOBO. Now the web cam never works, and I again reported the sleep issue. It is going back again tomorrow.