Hello IT pros,
If you know a lot about motherboards, it could really be helpful to hear your experiences or theories about this I/O control defect. Even better, though not necessary, it would be great to find out anything you may know about the unresolved instability problem that have plagued certain laptop configurations beginning in 2007. What I'm looking for is any theories, from the general to the scenario-specific -- or hard info -- on the source of this instability, since Intel, the OEMs and Microsoft have been relatively mum.
I have my own theory, but let me get you on the same page with the background and facts ahead of that. Then you can help me ground truth it.
By the way, this is cross-platform. Linux has experienced at least some related problems, but I can't tell you a lot more than that (just an anecdote or two), as it's not my current specialty. If you know Linux, EVEN BETTER! This seems to be a hardware issue, and Linux overlap could nail it down.
The summary of the issue is this: laptops like Lenovo's x6x generation ThinkPads, and some Dells, got BSODs and hangs that all seem to be associated with the Intel NAND cache (Turbo Memory) installation.
More importantly, the ITM system has an AHCI (or RAID) controller service (my terminology may need some help) that Intel calls the "Matrix Storage Manager," which serves a bunch of other masters, not just the NAND.
In Windows, the crashes have been associated with a particular MSM component, called iaNvStor.sys.
Even when the NAND wasn't engaged or installed, MSM has become a necessary AHCI component governing certain I/O system utilities, like an inertial HDD "drop guard" monitor. (That's the Active Protection System of the ThinkPads.)
Somehow, sound cards, especially in Dells, are also caught up in the MSM driver problem. See this thread:
http://forum.notebookreview.com/showthread.php?t=297054
I haven't looked as closely at that sound issue, but superficially, it seems that this happens both when the NAND is present, and when it isn't. It seems that MSM is just using up a lot of system resources.
Some hangs were resolved by turning off the battery-only caching Microsoft calls "Windows Hybrid Hard Disk Power Saving," which is also tied to this ITM system, and my guess is to iaNvStor.sys.
Under Windows, the branded use for this AHCI / NAND system was in ReadyBoost (NAND caching) and ReadyDrive (hybrid HDD I/O).
A lot of BSODs referenced 0x7E and the iaNvStor.sys file, or 0x7B and similar files, or 0xA IRQL_NOT_LESS_OR_EQUAL.
There is a known overlapping, but not directly tied, AHCI problem in Vista, discussed in KB922976. This indicated that if a user switches in BIOS from SATA's PATA Compatability Mode to SATA's AHCI mode, you have to change a registry entry first. In PATA mode, Vista's boot process disables the AHCI drivers from loading, and will BSOD on boot if you have not re-enabled the AHCI drivers in the registry when you change the BIOS settings to AHCI. As you can see, the Intel I/O problem is actually a separate phenomenon.
Those are the widely-reported facts.
From what I've been able to determine, there is another underlying dimension to this, but before I get too far into the weeds, I'd like to hear your experiences with diagnosing these I/O-related processes.
But to give you a tip about how to dig into this issue, I am almost certain at this point in my research that it's almost exclusive to Intel's ICH8-generation mobile I/O Controller Hub (the ICH8-M and ICH8M-E).
If I don't hear any immediate response, I'll try and sweeten the deal a little with the rest of the info. It's more speculative, and I'd prefer not to go out on a limb until I'm more informed by the wealth of experience among this forum's readers.
Many thanks for contributing to the discussion!
akaaka
If you know a lot about motherboards, it could really be helpful to hear your experiences or theories about this I/O control defect. Even better, though not necessary, it would be great to find out anything you may know about the unresolved instability problem that have plagued certain laptop configurations beginning in 2007. What I'm looking for is any theories, from the general to the scenario-specific -- or hard info -- on the source of this instability, since Intel, the OEMs and Microsoft have been relatively mum.
I have my own theory, but let me get you on the same page with the background and facts ahead of that. Then you can help me ground truth it.
By the way, this is cross-platform. Linux has experienced at least some related problems, but I can't tell you a lot more than that (just an anecdote or two), as it's not my current specialty. If you know Linux, EVEN BETTER! This seems to be a hardware issue, and Linux overlap could nail it down.
The summary of the issue is this: laptops like Lenovo's x6x generation ThinkPads, and some Dells, got BSODs and hangs that all seem to be associated with the Intel NAND cache (Turbo Memory) installation.
More importantly, the ITM system has an AHCI (or RAID) controller service (my terminology may need some help) that Intel calls the "Matrix Storage Manager," which serves a bunch of other masters, not just the NAND.
In Windows, the crashes have been associated with a particular MSM component, called iaNvStor.sys.
Even when the NAND wasn't engaged or installed, MSM has become a necessary AHCI component governing certain I/O system utilities, like an inertial HDD "drop guard" monitor. (That's the Active Protection System of the ThinkPads.)
Somehow, sound cards, especially in Dells, are also caught up in the MSM driver problem. See this thread:
http://forum.notebookreview.com/showthread.php?t=297054
I haven't looked as closely at that sound issue, but superficially, it seems that this happens both when the NAND is present, and when it isn't. It seems that MSM is just using up a lot of system resources.
Some hangs were resolved by turning off the battery-only caching Microsoft calls "Windows Hybrid Hard Disk Power Saving," which is also tied to this ITM system, and my guess is to iaNvStor.sys.
Under Windows, the branded use for this AHCI / NAND system was in ReadyBoost (NAND caching) and ReadyDrive (hybrid HDD I/O).
A lot of BSODs referenced 0x7E and the iaNvStor.sys file, or 0x7B and similar files, or 0xA IRQL_NOT_LESS_OR_EQUAL.
There is a known overlapping, but not directly tied, AHCI problem in Vista, discussed in KB922976. This indicated that if a user switches in BIOS from SATA's PATA Compatability Mode to SATA's AHCI mode, you have to change a registry entry first. In PATA mode, Vista's boot process disables the AHCI drivers from loading, and will BSOD on boot if you have not re-enabled the AHCI drivers in the registry when you change the BIOS settings to AHCI. As you can see, the Intel I/O problem is actually a separate phenomenon.
Those are the widely-reported facts.
From what I've been able to determine, there is another underlying dimension to this, but before I get too far into the weeds, I'd like to hear your experiences with diagnosing these I/O-related processes.
But to give you a tip about how to dig into this issue, I am almost certain at this point in my research that it's almost exclusive to Intel's ICH8-generation mobile I/O Controller Hub (the ICH8-M and ICH8M-E).
If I don't hear any immediate response, I'll try and sweeten the deal a little with the rest of the info. It's more speculative, and I'd prefer not to go out on a limb until I'm more informed by the wealth of experience among this forum's readers.
Many thanks for contributing to the discussion!
akaaka
Facebook
Twitter