Hello Cuppie and Dreifuss,
Great news! That "\forceresetreg" trick has solved the shutdown hang problem on my system too. Special thank goes to Cuppie for sharing this valuable information.
Also I have created a very small app in MFC that uses the same logic as described by cuppie to log the SUSPEND/RESUME events in windows and stores them in a file. For the sack of this test, I have reset the system from the wall switch and now the app is running. I am not doing anything special and just using the system as normal. It normally hangs around 3rd day so I am just waiting. Hopefully this test will revel that magical hang count.
I am really glad that cuppie has opened this case with Intel and hoping that we will get a proper fix this time and soon. Intel boards normally cost extra premium than rest of the market. One of the reasons for choosing this board was my 24x7 operational reliability and stability of the system. Power management issues like these are all too common in different motherboard venders but I am disappointed that Intel is also not an exception in this case.
A little update,
As mentioned before I did a cold reboot (also disconnecting the power from wall) and started my logging app. The system went to sleep mode after 20 min of inactivity fine. But hanged before resuming the first time! It is as usual, the same hangup where the power led is lit and the drives are also spinning but no display. Since the only change to the system I made was the \forceresetreg switch so I removed it. Bizarrely enough but now it seems ok. The log has recorded 3 sleep/wake cycles up to this point.
First sorry for the late reply,
Spoke to Tobias last friday.
He was not able to reproduce problem #2.
He has resumed more than 300 times over 4 or 5 days.
So he closed the call.
We have to find out , how its possible to reproduce this problem.
As I mentioned earlier , maybe it has to do with drivers.
At this moment i've disabled the onboard audio in the bios , so no audio drivers are loaded in xp.
First feelings are good (70+ resumes 1,5 day)
When we can exactly reproduce this , we can re-open the case.
I'm still hoping that also Problem #1 (Register reset) is fixed because bios 0.37 doesn't fix this.
Kind regards ,
I don't know what to say. It is a bad news anyway because the problem is there. However, it is defiantly more complex than what I initially thought. On the first run of my logging app, it has recorded 77 successful sleep/wake cycles over 5+ days. The system gone for sleep 78th time and hanged while resuming. Interestingly, on the second time, I only managed to get 7 cycles and system hanged on the 8th resume.
These are strange results and there is a definite need to further investigate this problem. Is there a possibility that we are seeing this problem due to a bad manufacturing batch? Also can you make sure that Tobias is using WinXp SP 3? Thanks.
Tobias uses XP SP3. With the same board revision (Board AA : E57650-300).
But until now I'm still up 'n running for 5+ days with 300+ resumes without reboot.
Only disabled the onboard audio in the bios.
Is it possible for you to also disable this , to test ?
I've also the parallel and serial ports disabled , and the onboard audio.
I thought it was the onboard audio.
But after almost 7 days and 400+ standby/resumes it hung up.
I think it will be very hard to exaclty reproduce this problem.
We really need to find out what happens , so we can reproduce it.
I was not that lucky this time around as it hanged on 21st cycle! My initial thought was that probably this is a fixed number after which the system hangs but since I started logging, the results are completely random. This is really a very difficult and frustrating situation and apparently there is no clear way to reproduce except the fact that it eventually happens.
One thing that came to my mind the other day was that I have done some changes to keyboard and mouse power management settings to enable S3 sleep mode. Initially when I bought this board about 6 months ago, the system was always going to S1 sleep. With a little help from google, I have found that Windows XP will not let the system go to S3 if there is any USB device present on the system. Because my mouse and Keyboard are USB devices, XP was only letting the system go to S1. So to enable S3 state, I have unchecked the "Allow this device to wake the computer" from keyboard and mouse properties in device manager. The effect of this is that once in sleep I cant wake the system with mouse or keyboard which is ok for me as I have set it up to wake on directed LAN packet. This setup seems to work perfectly until the system hangs on a random wake event. I am not sure if this setup is the cause of problem. What are you doing to enable S3 in XP?
Another interesting observation last night was that when the system hangs, it seems to have partially recovered power. Means that the hard drives were spinning, the power led was not blinking, but the optical USB mouse had no power! The only thing I can think out of that is probably USB subsystem has failed to regain power and caused the system to halt? Can you verify that on your system Cuppie?