-
1.
公开(公告)号:JP2003108314A
公开(公告)日:2003-04-11
申请号:JP2001286860
申请日:2001-09-20
Applicant: IBM
Inventor: SHIMADA SHOGO , YAMADA HIDEJI , ANDO HARUO , YUKI TOSHIHIKO
Abstract: PROBLEM TO BE SOLVED: To extend a substantial time-out period for writing in an environment where shock and vibration are applied, so that the occurrence of unnecessary time-out is minimized. SOLUTION: After a sector buffer has stored write data and a time delay Td has elapsed, a command complete is issued.
-
公开(公告)号:JPH0991856A
公开(公告)日:1997-04-04
申请号:JP24285895
申请日:1995-09-21
Applicant: IBM
Inventor: YAMADA HIDEJI
Abstract: PROBLEM TO BE SOLVED: To accelerate the entire operating speed by dividing plural procedures for read/write defect recovery by the time required for each execution and selecting the procedure corresponding to the remaining time of a defect recovery operation. SOLUTION: A disk driving device 14 rotates disks 18A and 18B for information recording through a shaft 12 and a cylindrical supporting body 16. Signal converters 20A, 20B, 20C and 20D track data tracks and read and write data by a signal converter driving device 28 respectively through access arms 22A, 22B, 22C and 22D. When a read defect is generated, a control part 30 executes a recovery operation composed of a prescribed series of defect recovery procedures corresponding to the mode of the defect. In this case, the defect recovery procedures are divided into the group of the change of an MR element bias current and an off-track amount or the like, for instance, for which the time required for each execution is short and the group of the change of a floating amount for lowering a disk rotation number and rewrite or the like, for instance, for which the required time is long, selected from either one of the groups and executed.
-
公开(公告)号:JP2001312373A
公开(公告)日:2001-11-09
申请号:JP2000121853
申请日:2000-04-21
Applicant: IBM
Inventor: YAMADA HIDEJI
Abstract: PROBLEM TO BE SOLVED: To improve the performance of a disk drive device by preventing the state of a sector buffer from being occupied, and shortening waiting time at the time of write. SOLUTION: When a held write command is in an execution waiting state, processing is performed for receiving a new write command, checking a hit status expressing a position relation with the LBA of the held command, relocating the order of write onto a recording medium and recording write data onto the medium according to the relocated write order. At such a time, when empty areas in the sector buffer are
-
公开(公告)号:JPH10161818A
公开(公告)日:1998-06-19
申请号:JP30774396
申请日:1996-11-19
Applicant: IBM
Inventor: SUZUKI HIROMASA , ASANO HIDEO , TOBARI ATSUSHI , NISHINO SATOSHI , YAMADA HIDEJI , ANDO HARUO , KOWA TSUGUAKI
Abstract: PROBLEM TO BE SOLVED: To provide a method/device which can assure execution of an ERP (error recovery procedure) up to its last step regardless of the time needed up to the time-out of an instruction and can more surely recover an error. SOLUTION: A disk device carries out an ERP up to its last step or until an error is recovered with no interruption when a reset instruction is received during execution of the ERP. In another mode, the disk device stops execution of the ERP and also holds the number K of the step whose execution is over right before reception of a reset instruction when this instruction is received during execution of the ERP. Then the disk device successively carries out the ERP from the (K+1)th error recovery step when a retry instruction is received.
-
-
-