发明名称 ENHANCED DUMP DATA COLLECTION FROM HARDWARE FAIL MODES
摘要 An approach is provided for collecting data for diagnosing a failure of a computer hardware device. After an indication of the failure of the computer hardware device that results in a full system crash is received, an address translation table of a central processing unit (CPU) of the computer hardware device is collected. A format of call stack frames of an operating system (OS) image of the computer hardware device is retrieved. Based on the collected address translation table and the retrieved format of the call stack frames, the call stack frames are retrieved and output to a computer file. Retrieving the call stack frames is performed by a service processor external to system processing units running OS images via a hypervisor of the computer hardware device. The service processor does not require an ambulance logical partition managed by the hypervisor.
申请公布号 US2016299809(A1) 申请公布日期 2016.10.13
申请号 US201615187949 申请日期 2016.06.21
申请人 International Business Machines Corporation 发明人 McCoy Doyle J.
分类号 G06F11/07;G06F9/455;G06F12/10 主分类号 G06F11/07
代理机构 代理人
主权项 1. A method of collecting data for diagnosing a failure of a computer hardware device, the method comprising the steps of: a computer receiving an indication of the failure of the computer hardware device that results in a full system crash; the computer collecting an address translation table of a central processing unit (CPU) of the computer hardware device; the computer retrieving a format of a plurality of call stack frames of a call stack of an operating system (OS) image of the computer hardware device; based on the collected address translation table and the retrieved format of the plurality of call stack frames, a processor of the computer retrieving the plurality of call stack frames, wherein the step of retrieving the plurality of call stack frames is performed by a service processor external to system processing units running a plurality of OS images via a hypervisor of the computer hardware device, and wherein the service processor does not require an ambulance logical partition (LPAR) managed by the hypervisor; and the computer outputting the retrieved plurality of call stack frames to a computer file.
地址 Armonk NY US