http://bugzilla.kernel.org/show_bug.cgi?id=5303
------- Additional Comments From [EMAIL PROTECTED] 2006-02-02 04:04 ------- Int Pending MSR is indeed C001_0055h. ANyone with access to the AMD BKDG should be able to get the fiekd definitions also. This register enables the processor to send a message to the I/O hub that results in the pending interrupt being serviced. The two message types are the IO space message and the HyperTransport INT_PENDING message. HyperTransport INT_PENDING message is defined by the HyperTransport 1.05c specification. If the processor or the I/O hub does not support the INT_PENDING HyperTransport message, the IO space message should be selected by IntPndMsg. A check for a pending interrupt is performed at the end of an IO instruction. If there is a pending interrupt and STPCLK is asserted, the processor executes a byte-size read or write to IO space defined with IORd, IOMsgAddr, and IOMsgData (used only for IO writes) to generate an SMI. The SMI wakes up the processor so the original pending interrupt can be serviced. The SMI handler should not take any action if the SMI is generated by this mechanism. In order to prevent SMI generation with this mechanism in the SMI handler,IntrPndMsgDis bit should be set in the SMI handler before the first IO instruction is executed, and itshould be cleared prior to resuming from SMM. If the processor and the I/O hub support the INT_PENDING HyperTransport message, it should be selected by IntPndMsg bit. The check for a pending interrupt is performed when entering the stop grant state. This MSR is writable by software and can be used to enable both of the above defined methods. SMI method needs a BIOS SMI handler implementation that supports it and the Int_Pending HT message needs a SB that is HT spec 1.05c compliant. Andy, if you want to apply a OS workaround for such problems, one implementation could be to enable int_pending message support in the OS. One way I can think of doing this is to force c2/c3 transitions, enable int_pending messages - SMI method or HT message method and check if there are lost ticks. the method that does not yield lost ticks can be implemented on that platform. I know I make it sound very simplistic where as the implementation might not be that simple. ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642 _______________________________________________ acpi-bugzilla mailing list acpi-bugzilla@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/acpi-bugzilla
