发明名称 Fault handling at a transaction level by employing a token and a source-to-destination paradigm in a processor-based system
摘要 A method for detecting errors in a processing device is disclosed. A data source unit of a processing device transmits data and a qualifier synchronously with the data, the qualifier to indicate the data is uncorrectable. At least one intermediate functional unit in the processing device receives the data and the qualifier. The at least one intermediate functional unit detects the data is uncorrectable based on the qualifier. The at least one intermediate functional unit transmits, without using, the data and the qualifier synchronously with the data to a data consumer unit of the processing device. The data consumer unit receives the data and the qualifier. The data consumer unit detects the data is uncorrectable based on the qualifier. The data consumer unit maintains, without using the data and the qualifier.
申请公布号 US9063855(B2) 申请公布日期 2015.06.23
申请号 US201313826843 申请日期 2013.03.14
申请人 Intel Corporation 发明人 McNairy Cameron B.;Agrawal Anil;Mayfield Jenna S.;Gouldey Eric A.;Millican Mark
分类号 G06F11/00;G06F11/07;G06F11/10 主分类号 G06F11/00
代理机构 Lowenstein Sandler LLP 代理人 Lowenstein Sandler LLP
主权项 1. A method, comprising: receiving, by a data consumer unit in a processing device, data and a qualifier synchronously with the data, the qualifier to indicate the data is uncorrectable, the data and the qualifier being received from a data source unit in the processing device; detecting, by the data consumer unit, the data is uncorrectable based on the qualifier; signaling or logging, by the data consumer unit, a high severity error indication in response to detecting the data is uncorrectable; and maintaining, without consuming, by the data consumer unit, the data.
地址 Santa Clara CA US
您可能感兴趣的专利