mrvip replied to MagicAndre1981's topic in Windows Tips 'n' TweaksThanks for the reply, could you point me to the discussion about usb and older nforce chipsets? Also you say PCI(E), do you mean that it should be PCI _Express_ exactly, or PCI would be ok as well? Is there any way to resolve the IRQ conflict (force sata and usb to use different irqs)?
mrvip replied to MagicAndre1981's topic in Windows Tips 'n' TweaksEDIT 6.2.2011: PROBLEM IS FIXED BY INSTALLING THIS UPDATE http://support.microsoft.com/?kbid=976972 ! Hi, it seems to be the best forum to discuss my problem, considering MagicAndre1981 is on it Summary of the problem: high interrupt and DPC count (but no high latency !!!), taking 100% of the second core of my CPU, when using patched 32 bit kernel to enable PAE. System: CPU: Athlon 64 X2 4000+ mb: Gigabyte GA-M56S-S3, nforce 560 chipset OS: Windows 7, 34 bit with patched kernel to enable PAE and access the full 4GB of memory, instead of only 3GB (i have a printer and a scanner that don't have 64 bit drivers, so i have to stick to 32bit) storage: 2 sata hdds in raid0 + another ata hdd driver updates: - only audio driver update listed for my MB/OS combination on Gigabyte's site, downloaded and installed it - only HDMI driver update is listed for my Chipset/OS combination on Nvidia's site, i don't have HDMI audio, but tried to install it just to be sure, and as expected it says that i don't have such hardware. BIOS updates: currently using bios version F5, tried F6C a some time ago, but it didn't have the SATA RAID part of the bios, so i have switched back to F5. I see that there is F6H version available now instead of F6C, i have to try it, but haven't done so yet. The problem shows up after about a day of use and is temporarily cured by reboot. The problem started after i have used a patch that allows the 32 bit kernel of Windows 7 to utilise all 4GB of memory that i have installed, instead of only using 3GB. The patch documentation says that it only enables the PAE in kernel by changing some licensing information, i.e. it doesn't add code or change Microsoft code. When i boot the old kernel (non-PAE), i don't have the problem. Using xpref to analyse the cause showed both storport.sys in funcion RaidpAdapterInterruptRoutine and "Unknown" module in funcion 0x88685bf8 with about the same unusual high number of interrupts. I had both Daemon Tools and Alcohol 120% installed, both install the sptd.sys driver. Following advice on another forum i removed both , but that didn't remove sptd.sys so i had to download the sptd standalone package and use it's uninstall funcion to successfuly remove sptd.sys. After removing it, the problem still appeared, but there was a difference: again one of the drivers with high interrupts was storport.sys/RaidpAdapterInterruptRoutine but the other one now was usbport.sys/USBPORT_InterruptService (instead of the "Unknown" module when the sptd.sys was installed). I have two USB controllers listed in device manager: "standard openhcd usb host controller", this one is showed as the parrent of all usb 1.1 devices, no matter in which port i plug them "standard enhanced pci to usb host controller", this one is showed as the parrent of all usb 2.0 device, again it doesn't matter in which port they are plugged both "standard enhanced pci to usb host controller" and "nvidia nforce serial ata controller" are sharing IRQ 22, which i think is the cause of the problem. the devices with usb 2.0 interface, attached to the system, are a card reader and a canon canoscan 3000 scanner. I think their drivers should not be causing the problem, because they don't handle irqs directly. This article: http://support.microsoft.com/kb/930261 deals with IRQL_NOT_LESS_OR_EQUAL in storport.sys on machines with nvidia chipset using 4gb or more memory, but with Vista as the OS and the solution is to install the latest service pack, obviously Windows 7 has newer drivers than a Vista service pack. It seems that i am close to the solution, but i have yet to connect the dots, so any help would be appreciated. on my todo list: - Disable the usb 2.0 controller when the high interrupt condition occure to see whether it will fix it. It is not a permanent fix, because i need usb2.0 - Try to make the usb 2.0 controller and the sata controller use different irqs. I don't know how to accomplish this - both are onboard devices, so i can't move them around pci slots, and windows 7 doesn't allow the user to change irqs on devices. - try the latest bios (hoping that Gigabyte didn't forget to include the raid support in F6H as they did in F6C) Regards, Velko