Summary of the invention
The object of the present invention is to provide a kind of fault-tolerant design method that is used for solid storage device, this method can realize reducing the expense of device fabrication test, can guarantee the acceptance rate of equipment again, can realize prolonging simultaneously the memory device purpose in serviceable life again.
A kind of fault-tolerant design method that is used for solid storage device provided by the invention is provided with at least one redundant flash memory particle in flash array, carry out fault-tolerant design in device fabrication test and/or use, it is characterized in that,
Fault-tolerant design method in the device fabrication test comprises the steps:
(A1) the solid-state storage processor was checked with the flash memory particle working as in the flash array of solid state memory device;
(A2) judge to work as whether break down with the flash memory particle, if, then it is carried out mark, and record trouble information; Change step (A3) over to, otherwise, step (A3) directly entered;
(A3) judge in the flash array when whether complete inspection finishes with the flash memory particle, if, change step (A4) over to, otherwise, the next one in the flash array when checking with the flash memory particle, is changed over to step (A2) then;
(A4) check out all after using the flash memory particle,, then utilize redundant flash memory particle to replace working as of failure rate maximum and use the flash memory particle, the information of replacement is changed in the solid-state storage processor, change step (A6) over to if fault is arranged when using the flash memory particle; Otherwise carry out the operation of the 5th step;
(A5) with the backstage space of redundant flash memory particle, be used to carry out built-in function as the solid-state storage processor;
(A6) finish;
Carrying out the fault-tolerant design method in the use comprises the steps:
(B1) the solid-state storage processor was monitored with the flash memory particle working as in the flash array of solid state memory device;
(B2) the monitoring process in, when host computer system in the solid state memory device when erase operation being arranged with the flash memory granule medium, just the least unit piece to each operation carries out technical operation; Such erase operation, count value just adds up 1;
(B3) judge whether described count value surpasses pre-set threshold, if, enter step (B4), proceed to monitor otherwise change step (B1) over to;
(B4) will carry out orderly data migtation work according to the priority that configures above the task queue that enters data migtation with the flash memory particle of working as of threshold value;
(B5) need working as of migration to use the target of flash memory particle, finish the work of moving of data afterwards on the program backstage according to own passage corresponding redundant flash memory particle selection migration.
The present invention adopts the method that increases standby redundant storage particle on the board design of memory device, when equipment is being produced in the back end test or is being found to have flash memory to break down in using, use the redundant storage particle to replace the fault particle, thereby guarantee that even particle does not pass through packaging and testing, can not influence the acceptance rate that memory device is produced yet; Simultaneously; adopt the method that increases the redundant storage particle by the present invention; come the serviceable life of effective prolongation equipment; when having in the equipment after the flash memory particle reaches certain serviceable life; will be replaced by redundant particle or idle particle; thereby, come the serviceable life of effective prolongation equipment by this replacement operation constantly.
The present invention adopts the redundant particle of memory device can be used for replacing the flash memory particle that unusual condition occurs in the device fabrication test;
The present invention adopts the redundant particle of memory device can be used for replacing the flash memory particle that excessively uses or occur unusual condition in the normal use of memory device;
The definition of excessively using among the present invention can be the threshold value of a plurality of access times, replace the particle that reaches first threshold value at first with redundant particle, redundant particle becomes the particle of normal use, the particle that is replaced becomes standby redundant particle break-off, wait for to replace nextly to reach second threshold value at first and be judged the particle of excessive use again.
The redundant particle of multi-channel flash memory memory device involved in the present invention can be that unit carries out replacement operation with single, also can be that unit carries out replacement operation with the passage.
The present invention exists the generation of bad piece according to existing solid storage medium, and the limited and life problems that causes of erasing times, these two characteristics may cause solid storage device part medium to lose efficacy, therefore take the thought of redundancy backup, reduce the data loss problem of makeing mistakes and losing efficacy and bring because of medium.
Embodiment
Below utilize accompanying drawing and the present invention is described in further detail in conjunction with embodiment.Embodiment described herein is a part of example of the present invention, rather than whole embodiment.Based on embodiments of the invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that is obtained under the creative work prerequisite.
The invention provides a kind of by the particle of storage medium in the solid storage device being carried out the mode of redundancy backup, the next fault-tolerant design method that is about to inefficacy or failed flash particle that substitutes.Thereby greatly reduce the probability and the expense in production test stage that memory device lost efficacy.
As shown in Figure 1, be the functional block diagram of at present general solid storage device.Wherein mainly comprise host computer system 101 and solid storage device 102.Wherein, solid storage device 102 comprises interface module 103, solid-state storage processor 104, and be the Flash array 105 that unit forms with Flash particle 106.Wherein, interface module 103 is mainly used in and realizes and main frame accordant interface agreement, for example SATA; What present to main frame by interface module 103 is a standard storage equipment that has certain logical space.Solid-state storage processor 104 is control cores of whole memory device, move firmware program above, the transmission, Flash management, host logical address of mainly being responsible for control signal between interface module 103 and the flash array 105 and data is to Flash physical address map, wear leveling (thereby logical address is mapped to operation that different physical addresss prevents that single Flash from too being concentrated and inefficacy in advance), bad block management etc.106 is single Flash particle, and a plurality of Flash particles are formed Flash array 105, are the terminal media of entire equipment information stores.
The inventive method is to propose on the basis of above-mentioned solid storage device, at first increases redundant particle in the Flash array, as Fig. 2, Fig. 4, and manages by following process:
The information of redundant block be can write down in the storage of processor, its physical address, number, size etc. comprised.Then, in detecting the Flash array, there is the Flash particle to lose efficacy, firmware program is this spent granules of mark and select a redundant particle to replace then, and the mode of replacement is exactly that the logical address of the Flash particle correspondence that will lose efficacy originally in map addresses is re-assigned on the redundant Flash particle of selection.
As previously mentioned, detect the inefficacy of Flash particle and may occur in the test phase that device fabrication is made, also may occur in the equipment running process explanation respectively below.
As shown in Figure 2, be an embodiment of the present invention, in this embodiment design, be mainly used at the volume production test phase and carry out.In this embodiment, only be provided with a redundant Flash particle to make example, the present invention does not limit the number of redundant particle.
To not need through the Flash particle of test or directly not be welded on the printed circuit board (PCB) of memory device 201 through the Flash wafer of encapsulation, simultaneously, need be on circuit board the positions of a particle of many designs, the position of promptly redundant Flash particle 203.Its purpose is because not process test or encapsulation of the flash memory particle that is welded, when in device detection procedure, finding to have the Flash particle to have fault maybe can not use, just redundant Flash particle is replaced problematic Flash particle by the configuration of storage of processor, thereby, method of testing after finishing by this device fabrication, remedy defective and problem that the Flash particle does not have the process packaging and testing, do not have great changes with the acceptance rate that guarantees entire equipment processing.This mode can be saved the expense and the energy of Flash particle screen selecting.
In addition, the Flash particle of the overwhelming majority, Flash array 202, even without through packaging and testing, its performance also can guarantee, therefore, in the test of most product after production is good, and its redundant Flash particle is not used to replace fault Flash particle.Among the present invention; these redundant Flash particles will be used to prolong service life of equipment; when Flash particle utilization rate reaches some or has after the Flash particle breaks down in equipment uses; allow redundant Flash particle replace Flash particle or the fault Flash particle that reaches the access times threshold value; thereby assurance equipment still can normally use, to reach the purpose that prolongs service life of equipment.
As shown in Figure 3, carry out the process flow diagram of the concrete steps of fault-tolerant design method for above-mentioned situation.
(1) firmware program in the solid-state storage processor when checking with the Flash particle, judges whether it breaks down in the Flash array each;
Working as of indication of the present invention is meant original Flash particle in the Flash array with the Flash particle.
(2) as if current Flash particle non-fault, then next Flash particle is checked, finished until all flash memory particle inspections;
(3) if break down the Flash particle, then it is carried out mark, and with its failure message record, as the fault situation, particle numbering etc.;
(4) check out after all Flash particles,, choose the Flash particle of failure rate maximum and replace, otherwise carry out the operation of the 5th step if there is fault Flash particle then to judge information; The information of replacing is changed in the firmware program of solid-state storage processor, promptly the physical message of the fault Flash particle selected is replaced, use normally with assurance equipment with redundant standby Flash particle;
(5) if non-fault Flash particle, then the Flash particle that this redundancy backup is used is used for carrying out built-in function as the space, backstage of the firmware program in the solid-state storage processor, improving user's actual use experience, and operation such as increase the service life.
As shown in Figure 4, be an embodiment of the present invention, the redundant Flash particle of designing printed circuit board is a plurality of among this embodiment, its application target is consistent with Fig. 2 embodiment, just consider its application target and higher equipment acceptance rate, therefore reserve a plurality of redundant Flash particles, simultaneously, can more prolong the serviceable life of equipment.When this design mainly is used in solid storage device general work situation.
On the pcb board of memory device 401, flash array 402 is reserved the number of redundant Flash particle according to equipment physical channel number, in this example, three passages are arranged in the equipment, therefore be each passage reservation 1 redundant particle, totally 3 redundant particles.Before the Flash particle that occur not to lose efficacy, redundant Flash particle is also available is used as free block for these, because the time ratio of erase block is more of a specified duration, free block is many more in the system, need the probability of erase block just more little, therefore can promote the response speed of read-write operation, promote user's impression.
As shown in Figure 5, carry out the process flow diagram of the concrete steps of fault-tolerant design method for above-mentioned situation.
Following mask body describes this method.
(1) firmware program in the solid-state storage processor is monitored the Flash particle on the equipment, and this operation is to whole Flash particle operations;
(2) in the process that detects, when host computer system had erase operation to the Flash granule medium in the solid state memory device, just the least unit piece to each operation carried out technical operation; Such erase operation, count value just adds up 1;
(3) firmware program in the solid-state storage processor is added up Flash particle overall technology situation and is monitored, when this count value surpasses preset threshold, with regard to execution in step 4, otherwise just proceeds to monitor;
The span of threshold value is 1~N, and wherein N reads and writes number of times (providing this type of parameter by Flash manufacturer) for the maximum that the Flash particle allows.In order to reduce cost, threshold value can be got the value near N usually, as greater than 90%N etc.
(4) the Flash particle that surpasses threshold value enters the task queue of data migtation, carries out orderly data migtation work according to the priority that configures;
(5) need the target of the Flash particle of migration, finish the work of moving of data afterwards on the program backstage according to own passage corresponding redundant flash memory particle selection migration.
In use, when one of them particle access times reaches some, just can be replaced, when so just having a plurality of particles and reaching preset requirement gradually, be replaced by standby particle by one of them redundant Flash particle.Simultaneously, substituted particle is not to go out of use, in order to reach better optimize, equipment is recycled, the threshold value of a plurality of access times can be set, when other particles are used higher access times, can be replaced because of reaching the low particle that uses threshold value to be replaced by these again, like this, can better reach a kind of optimization, strengthen usage ratio of equipment, constantly have the Flash particle that is overused to be replaced, allow other Flash particles that often do not used use a period of time, be replaced up again, therefore, overall equipment can significantly be increased service time, has prolonged the life-span of equipment effectively.Equally, the Flash that can also be designed to whole passage for multichannel flash memory device replaces or is that unit carries out fault-tolerant design with the passage.
Above content be in conjunction with concrete preferred implementation to further describing that the present invention did, can not assert that concrete enforcement of the present invention is confined to these explanations.For the general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, can also make some simple deduction or replace, should be considered as all belonging to that the present invention determines by claims of being submitted to.