Method, system, and product for isolating memory system defects to a particular memory system component

A data processing system for isolating a defect in a memory system to a particular memory system component, said system comprising:

• said memory system including a memory card, a plurality of physical dual inline memory modules (DIMMs) that are attached to said memory card, and a memory controller for controlling said memory card;
• said memory card coupled to said memory controller using a JTAG bus, said memory card including a plurality of empty slots to which additional physical DIMMs can be attached, and said memory card including a plurality of electrical buffers that are part of said memory card;
• testing means for testing said plurality of physical DIMMs first using a physical DIMM test;
• responsive to no error occurring during said physical DIMM test, testing means for testing said memory card using a memory card test;
• responsive to no error occurring during said memory card test, testing means for testing said memory controller next using a memory controller test;
• responsive to said memory controller passing said memory controller test, determining means for determining that said memory system is good;
• said plurality of electrical buffers utilized as a virtual memory controller when testing said plurality of physical DIMMs, and said plurality of electrical buffers utilized as virtual physical DIMMs when testing said memory card;
• coupling said plurality of electrical buffers to a service processor using a JTAG bus, said service processor controlling said plurality of electrical buffers using said JTAG bus;
• said physical DIMM test including:
• scanning means for scanning DIMM test data and address information to said plurality of electrical buffers, which then scan said DIMM test data to all addresses in each one of said plurality of physical DIMMs;
• reading means for reading data from said plurality of physical DIMMs;
• scanning means for scanning, from said plurality of electrical buffers, said data that was read from said plurality of physical DIMMs;
• determining means for determining that said plurality of physical DIMMs passed said physical DIMM test if said data that was read from said plurality of physical DIMMs matches said DIMM test data;
• said memory card test including:
• a unique identifier assigned to each one of said plurality of electrical buffers, each one of said plurality of electrical buffers addressable by the service processor using said unique identifier;
• each one of said plurality of electrical buffers treated as a virtual physical DIMM;
• the service processor testing data bits by writing card test data to each memory location in said virtual physical DIMM, and then reading data from each memory location in said virtual physical DIMM;
• if said data that was read from said virtual physical DIMM matches said card test data, performing means for performing address bit verification on said memory card;
• said address bit verification to test address bits including:
• writing means for writing a known test pattern to a first memory location that falls within a range of said virtual physical DIMM, and then reading means for reading a set of addresses that fall within said range of said virtual physical DIMM;
• determining means for determining whether said known test pattern was read from only said first memory location;
• if said known test pattern was read from only said first memory location:
• said service processor determining that said memory card passed said address bit verification for said first memory location, removing said known test pattern from said first memory location, and performing said address bit verification on a next memory location;
• and
• if all memory locations that fall within said range pass said address bit verification, determining means for determining that said memory card passed said memory card test.

A computer program product that is stored on a tangible computer readable medium in a data processing system for isolating a defect in a memory system to a particular memory system component, said product comprising:

• said memory card coupled to said memory controller using a JTAG bus, said memory card including a plurality of empty slots to which additional physical DIMMs can be attached, and said memory card including a plurality of electrical buffers that are part of said memory card;
• instructions for testing said plurality of physical DIMMs first using a physical DIMM test;
• responsive to no error occurring during said physical DIMM test, instructions for testing said memory card using a memory card test;
• responsive to no error occurring during said memory card test, instructions for testing said memory controller next using a memory controller test;
• responsive to said memory controller passing said memory controller test, instructions for determining that said memory system is good;
• instructions for utilizing said plurality of electrical buffers as a virtual memory controller when testing said plurality of physical DIMMs, and utilizing said plurality of electrical buffers as virtual physical DIMMs when testing said memory card;
• coupling said plurality of electrical buffers to a service processor using a JTAG bus, said service processor controlling said plurality of electrical buffers using said JTAG bus;
• said physical DIMMs test including:
• instructions for scanning DIMMs test data and address information to said plurality of electrical buffers, which then scan said DIMMs test data to all addresses in each one of said plurality of physical DIMMs;
• instructions for reading data from said plurality of physical DIMMs:
• instructions for scanning, from said plurality of electrical buffers, said data that was read from said plurality of physical DIMMs;
• instructions for determining that said plurality of physical DIMMs passed said physical DIMMs test if said data that was read from said plurality of physical DIMMs matches said DIMMs test data;
• said memory card test including:
• instructions for assigning a unique identifier to each one of said plurality of electrical buffers, each one of said plurality of electrical buffers addressable by the service processor using said unique identifier;
• instructions for treating each one of said plurality of electrical buffers as a virtual physical DIMM;
• instructions for testing, by the service processor, data bits by writing card test data to each memory location in said virtual physical DIMM, and then reading data from each memory location in said virtual physical DIMM;
• if said data that was read from said virtual physical DIMM matches said card test data, instructions for performing address bit verification on said memory card;
• said address bit verification to test address bits including:
• instructions for writing a known test pattern to a first memory location that falls within a range of said virtual physical DIMM, and then reading a set of addresses that fall within said range of said virtual physical DIMM;
• instructions for determining whether said known test pattern was read from only said first memory location;
• if said known test pattern was read from only said first memory location:
• said service processor determining that said memory card passed said address bit verification for said first memory location, removing said known test pattern from said first memory location, and performing said address bit verification on a next memory location;
• and
• if all memory locations that fall within said range pass said address bit verification, instructions for determining that said memory card passed said memory card test.