Method and apparatus for system status monitoring, testing and restoration

A method for restoring channel service, comprising:

• scheduling a plurality of channels to provide controlled access from client applications to a graphics processing unit;
• identifying a failure status of one of the plurality of channels;
• disabling the channel associated with the failure status;
• disabling engines of the graphics processing unit associated with the failure status;
• notifying a client application assigned with the channel that the channel has been disabled;
• restoring the engines associated with the failure status;
• restoring the channel associated with the failure status;
• allowing the client application to destroy and reconstruct command status and state of the channel;
• and
• storing error information for the failure status.

An apparatus for restoring channel throughput, comprising:

• means for scheduling a plurality of channels to provide controlled access from client applications to a graphics processing unit;
• means for identifying a failure status;
• means for disabling a channel of the plurality of channels associated with the failure status;
• means for disabling at least one engine of the graphics processing unit associated with the failure status;
• means for notifying a client application that the channel has been disabled;
• means for restoring the at least one engine associated with the failure status;
• and
• means for restoring the channel associated with the failure status.

A storage media containing a machine-readable program, which when executed by a processor, executes a method for restoring channel throughput, the method comprising:

• scheduling a plurality of channels to provide controlled access from client applications to a graphics processing unit;
• identifying a failure status;
• disabling a channel of the plurality of channels associated with the failure status;
• disabling at least one engine of the graphics processing unit associated with the failure status;
• notifying a client application that the channel has been disabled;
• restoring the at least one engine associated with the failure status;
• and
• restoring the channel associated with the failure status.

A programmed computer, comprising:

• a central processing unit;
• a graphics processing unit coupled to the central processing unit, the graphics processing unit capable of receiving information over a plurality of channels, the graphics processing unit coupled to allow receipt of data and commands from client applications via the plurality of channels;
• a scheduler and a switch configured to provide controlled access from the client applications via the plurality of channels to the graphics processor unit;
• memory coupled to the graphics processing unit and the central processing unit;
• a program at least a portion of which is instantiated in the memory, the software configured to:
• identify a failure status;
• disable a channel of the plurality of channels associated with the failure status;
• disable at least one engine of the graphics processing unit associated with the failure status;
• notify a client application that the channel has been disabled;
• restore the at least one engine;
• and
• restore the channel.