发明名称 Systems and methods for recovering from uncorrected DRAM bit errors
摘要 Systems, methods, and computer programs are disclosed for recovering from dynamic random access memory (DRAM) defects. One method comprises determining that an uncorrected bit error has occurred for a physical codeword address associated with a dynamic random access memory (DRAM) device coupled to a system on chip (SoC). A kernel page associated with a DRAM page comprising the physical codeword address is identified as a bad page. Recovery from the uncorrected bit error is provided by rebooting a system comprising the SoC and the DRAM device. In response to the rebooting, the identified kernel page is excluded from being allocated for DRAM operation.
申请公布号 US9507675(B2) 申请公布日期 2016.11.29
申请号 US201414253770 申请日期 2014.04.15
申请人 QUALCOMM INCORPORATED 发明人 Chun Dexter Tamio;Li Yanru;Kim Jung Pill;Sriramagiri Deepti Vijayalakshmi
分类号 G06F11/20;G06F11/10;G06F11/14;G06F11/07;G06F11/08;G11C29/00;G11C29/52;G06F3/06;G11C29/08 主分类号 G06F11/20
代理机构 Smith Tempel 代理人 Smith Tempel
主权项 1. A method for recovering from dynamic random access memory (DRAM) defects executed by instructions embodied on a non-transitory computer readable medium, the method comprising: determining that an uncorrected bit error has occurred for a physical codeword address associated with a dynamic random access memory (DRAM) device coupled to a system on chip (SoC); in response to the determination of an uncorrected bit error, branching an operating system to an exception handler running at least one recovery module from one of a static random access memory (SRAM) device and a read only memory (ROM) located on the SoC; determining with the recovery module whether a bit failure occurred at the DRAM device as a result of an error correction code (ECC) error; and responsive to the determination whether the bit failure occurred as a result of ECC error: if the bit failure occurred as a result of ECC error, identifying a kernel page associated with a DRAM page comprising the physical codeword address as a bad page, andrecovering from the uncorrected bit error by rebooting a system comprising the SoC and the DRAM device and, in response to the rebooting, excluding the identified kernel page from being allocated for DRAM operation, orif the bit failure did not occur as a result of ECC error, writing the contents of the DRAM device to a second memory device.
地址 San Diego CA US