System abstraction layer, processor abstraction layer, and operating system error handling

A computer system comprising:

• a plurality of systems comprising:
• a non volatile memory in which is stored a first error handling routine and a second error handling routine;
• and
• a processor coupled to the non volatile memory to execute the first error handling routine and the second error handling routine and to generate an error log on detecting an error.

A system comprising:

• a processor;
• a system memory;
• and
• a non-volatile memory in which is stored:
• a first error handling routine executed on an error to access state information, store the state information to a log and to attempt to correct the.

A method for cooperative error handling in a computer system comprising:

• detecting an error by a detecting processor;
• executing error handling code of a first layer of software, by the detecting processor, to perform the following:
• saving state information;
• attempting to correct the error;
• on failure to correct the error, executing error handling code of a second layer of software by the detecting processor to perform the following:
• determining severity of error by analyzing state information and the error received from the first layer;
• saving additional state information;
• and
• halting the computer system if necessary;
• and
• on failure to correct the error by the second layer of software, executing error handling code of an operating system by the detecting processor to perform the following:
• checking state information and the error to see if processing can continue;
• halting the computer system if processing unless processing can continue;
• and
• attempting to correct the error.

A method for handling errors in a system comprising:

• detecting an error;
• determining if the error is global or local;
• on the error being global, signaling other processors of the error and allowing one processor to control error handling;
• determining if the error is unrecoverable;
• on the error being unrecoverable, halting the system;
• correcting the error.

A method comprising:

• detecting an error;
• sending a signal to a processor abstraction layer, a system abstraction layer, and an operating system of the error;
• interrupting processing if necessary;
• attempting to correct the error by the processor abstraction layer and informing the system abstraction layer of success or failure in correcting the error;
• upon success by the processor abstraction layer, informing operating system of the correction;
• upon failure by the processor abstraction layer, attempting to correct the error by the system abstraction layer and informing the operating system of success or failure;
• upon failure by the system abstraction layer, attempting to correct the error by the operating system;
• and
• upon failure by the operating system, initiating a system reboot.

A computer readable medium containing computer instructions for instructing a processor to perform a method for cooperatively handling errors comprising:

• attempting to correct an error by a detecting processor;
• on failure, executing firmware code to correct the error;
• and
• on failure, executing operating system code to correct the error.