摘要:
A method and apparatus for setting aside a long-latency micro-operation from a reorder buffer is disclosed. In one embodiment, a long-latency micro-operation would conventionally stall a reorder buffer. Therefore a secondary buffer may be used to temporarily store that long-latency micro-operation, and other micro-operations depending from it, until that long-latency micro-operation is ready to execute. These micro-operations may then be reintroduced into the reorder buffer for execution. The use of poisoned bits may be used to ensure correct retirement of register values merged from both pre- and post-execution of the micro-operations which were set aside in the secondary buffer.
摘要:
A method and apparatus for executing a selective recovery after a branch misprediction is disclosed. In one embodiment, the instructions following the mispredicted branch point may be saved for selective re-execution in a buffer. Those instructions that wrote to physical registers between the mispredicted branch point and an exact convergence point, thereby causing false data dependencies, may be followed by corresponding move instructions to eliminate the false data dependencies. The instructions subsequent to the exact convergence point may then be selectively re-executed if subject to the previous false data dependencies.
摘要:
According to one embodiment, a method is disclosed. The method includes detecting a load miss at a central processing unit (CPU), stalling a read only buffer (ROB), speculatively retiring an instruction causing the ROB stall and subsequent instructions, keeping registers that have not been renamed in the ROB upon retirement, and flushing the CPU pipeline upon receiving data from the load miss.
摘要:
A mechanism is provided for cross-allocated block repair in a mounted file system. A set of cross-allocated blocks are identified from a plurality of blocks within an inode of the mounted file system, based on a corresponding bit associated with each cross-allocated block in a duplicated block information bitmap being in a first identified state. The set of cross-allocated blocks are repaired using a user-defined repair process. Then one or more of the set of cross-allocated blocks are deallocated based on results of the user-defined repair process.
摘要:
A method, computer program product, and apparatus for managing a file system is presented. An object in the file system is identified in which one of a first pointer from the object to a first folder in the file system and a second pointer from the first folder to the object is incorrect. A number of folders is generated within a second folder in the file system that represents a path from a root of the file system to the first folder. The first pointer for the object is set to a last folder in the number of folders in which the last folder represents the first folder.
摘要:
A method, computer program product, and apparatus for managing a file system is presented. An object in the file system is identified in which one of a first pointer from the object to a first folder in the file system and a second pointer from the first folder to the object is incorrect. A number of folders is generated within a second folder in the file system that represents a path from a root of the file system to the first folder. The first pointer for the object is set to a last folder in the number of folders in which the last folder represents the first folder.
摘要:
A mechanism is provided for cross-allocated block repair in a mounted file system. A set of cross-allocated blocks are identified from a plurality of blocks within an inode of the mounted file system, based on a corresponding bit associated with each cross-allocated block in a duplicated block information bitmap being in a first identified state. The set of cross-allocated blocks are repaired using a user-defined repair process. Then one or more of the set of cross-allocated blocks are deallocated based on results of the user-defined repair process.