Microsoft Hotfixes Windows 7 Memory Leak
A recent BSoD problem has appeared involving a memory leak associated with Windows 7's Power Manager.
Consumers using Windows 7 and Windows Server 2008 R2 may experience a sudden crash of the operating system for no apparent reason. The system failure is accompanied by the infamous "Blue Screen of Death" containing the "STOP: 0x0000000A" error code. Microsoft was alerted to the problem, and has posted an explanation on the Microsoft support website.
According to the company, the actual error reads "STOP: 0x0000000A (parameter1, parameter2, parameter3, parameter4); IRQL_NOT_LESS_OR_EQUAL." Microsoft added that users will observe that the four parameters in the Stop error message may vary, depending on the configuration of the PC. The Redmond company also stated that not all "STOP: 0x0000000A" error messages are caused by this particular issue.
"This issue occurs because Power Manager opens an Advanced local procedure call (ALPC) port," the company explains. "However, Power Manager closes another port instead of closing the ALPC port. Every time that a power request is made, a memory leak occurs. When the leaked memory usage accumulates to a certain level, the computer crashes."
Windows 7 and Windows Server 2008 R2 consumers can obtain a hotfix for the problem by heading here.

does it close the port on another computer? like "its ok dude, I'll save you from a memory leak, I'll take one for the team."
It's an ALPC port, are you thinking of regular data ports or something? It would take an exceptionally large amount of time (or an insanely high amount of power requests) for this to show up in standard use, since each ALPC port takes approximately 4KB of memory. Assuming the port never closes, and assuming new ones are opened at (the default) of one every 30 minutes, it would take 6 days straight for 1MB of memory to leak, and it would take almost 40 years (straight) for the system to bluescreen because of it, under normal circumstances.
I'm not saying this isn't an issue, but is an extreme edge-case scenario really worth making an entire news posting for? It's just another bug fix that would never have shown up in standard testing - I expect this only even manifests itself if you have software installed that polls power status far often than normal is healthy.
You can request the hotfix from http://support.microsoft.com/hotfix/KBHotfix.aspx?kbnum=979444&kbln=en-us.
This kind of hotfix, as it does not cover a security threat or affects many people, will not be on Windows update, but will instead be rolled into the next Service pack release. You need to request it using the above link, and you can then install the .msu manually.
The explanation for that seems very reasonable, since Microsoft found no link relating Windows 7 to the actual problems - Far more people are upgrading from XP (which did not have battery warning systems built in) to Windows 7, on laptops they did not perform adequately using Vista.
People are assuming that because these warnings are only showing up now (as in, yes, a new feature of Windows 7), and didn't on Windows XP, that Windows 7 is the cause of the problems. Instead, it's simply that Windows 7 has the ability to detect these failures by default, whereas Windows XP cannot (without additional software).
Click the link that says "show hotfixes for all platforms and languages", and it'll show the x64 and IA64 versions as well. It only shows 32-bit by default because almost everyone will be using 32-bit browsers even on 64-bit systems, and that's what it uses to try to detect the current platform.
Looks like I'll be needing the hotfix. Thanks so much for the article Toms.
I finally just bought a new card (I have no trouble since) but this bug bring me... a little "chill" when I hear about "IRQL_NOT_LESS_OR_EQUAL" errors.
If it takes so long to crash, explain how Microsoft found this problem. Why would they even be concerned with such a rare event?