CN107301132A - A kind of flash memory garbage reclamation optimization method - Google Patents
A kind of flash memory garbage reclamation optimization method Download PDFInfo
- Publication number
- CN107301132A CN107301132A CN201710434596.XA CN201710434596A CN107301132A CN 107301132 A CN107301132 A CN 107301132A CN 201710434596 A CN201710434596 A CN 201710434596A CN 107301132 A CN107301132 A CN 107301132A
- Authority
- CN
- China
- Prior art keywords
- valid data
- flash memory
- correcting code
- data
- error correcting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/0223—User address space allocation, e.g. contiguous or non contiguous base addressing
- G06F12/023—Free address space management
- G06F12/0253—Garbage collection, i.e. reclamation of unreferenced memory
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/08—Error detection or correction by redundancy in data representation, e.g. by using checking codes
- G06F11/10—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's
- G06F11/1008—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's in individual solid state devices
- G06F11/1012—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's in individual solid state devices using codes or arrangements adapted for a specific type of error
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Quality & Reliability (AREA)
- Techniques For Improving Reliability Of Storages (AREA)
Abstract
The invention discloses a kind of garbage reclamation optimization method of flash memory storage, including：Pair determine reclaim dirty piece in valid data carry out error correcting code error correction threshold test, and judge whether the valid data mistake reaches the error correction threshold value of its error correcting code, the valid data of error correcting code error correction threshold value are not up to mistake, valid data are write direct in flash memory by register；The valid data of error correcting code error correction threshold value have been reached to mistake, first valid data read back controller, then writes in flash memory；Confirm that all valid data migrations of garbage reclamation are completed, by dirty piece of erasing.The invention also discloses the application of the garbage reclamation optimization method, and use the storage device and terminal of this method progress garbage reclamation.The present invention can improve garbage reclamation efficiency on the premise of flash memory storage reliability is ensured, the time required to effectively reducing garbage reclamation, so as to improve the overall I/O performances of system.
Description
Technical field
The invention belongs to solid-state disk technical field of memory, and in particular to a kind of garbage reclamation optimization method of flash memory storage.
Background technology
Garbage reclamation is the removing necessary method of invalid data, whole life of this process in solid-state disk in solid-state disc system
Can continually occur in the life cycle.Garbage reclamation is reclaimed primarily directed to the valid data in dirty piece to be recycled, is removed dirty
Before block, specifically, need to be transferred to valid data therein into other clear positions of flash memory, then could be by dirty piece of erasing.
The method of existing Data Migration, is that first the data of active page are read back controller, is then then written to NAND sudden strains of a muscle
In depositing.That is, the data in dirty piece are required for first reading into device processed, general, in addition it is also necessary to these data are compiled again
Code, then could write flash memory.But this rubbish recovering method needs to take a substantial amount of time, can also choke system I/O, be
The solid-state disc system by storage medium of flash memory is badly in need of the performance bottleneck broken through at present.
In the prior art, CN201180041989 discloses a kind of pre-emptive rubbish recovering method, the core skill of this method
Art is to determine after garbage reclamation amount, judges size of the yield compared to recovery threshold value, rubbish is carried out according to the size of yield
Rubbish reclaimer operation.Although this method can solve to need the problem of quantity of refuse reclaimed overloads, for ensureing data storage
For the demand of reliability, benefited without any.In another prior art CN201510671242, a kind of base is disclosed
In the solid-state disk rubbish recovering method of data lifetime, the core technology of this method is, sets incremental according to application requirement
Time series, according to the expected existence time limit of data to be written, selects different modes to carry out garbage reclamation.The purpose master of this method
It is the writing speed for ensureing solid-state disk, reduces the access expense that data moving belt comes, the data storage for solid-state disk can
Effective solution is not proposed by property.
The content of the invention
For the disadvantages described above or Improvement requirement of prior art, the invention provides a kind of flash memory garbage reclamation optimization method
And its application, it is proposed that a kind of adaptive rubbish recovering method, the optimization method of the technology of the present invention is ensureing solid-state disk data
On the premise of the reliability of storage, the time required to reducing garbage reclamation, system I/O performances are improved.
To achieve the above object, according to one aspect of the present invention there is provided a kind of flash memory garbage reclamation optimization method, its
It is characterised by, including,
Pair determine reclaim dirty piece in valid data carry out error correcting code error correction threshold test；
Judge whether the valid data mistake reaches the error correction threshold value of its error correcting code；
The valid data of error correcting code error correction threshold value are not up to mistake, are led to the valid data using copy-back instructions
Register is crossed to write direct in flash memory；The valid data of error correcting code error correction threshold value have been reached to mistake, first the valid data have been read
Controller is returned, then is write in flash memory.
In technical solution of the present invention, the mistake of valid data to be recycled in dirty piece to be recycled is detected first, determines that it is
The no maximum error correcting capability for reaching error code, that is, carry out error correcting code error correction threshold test.Error correcting code threshold value is not up to for mistake
Instructed in valid data, technical solution of the present invention using copy-back, by register, first valid data are stored in and deposited
Device, then place data into flash memory.It also is located in flash memory due to be recycled dirty piece, by this method, it is not necessary to from flash memory
Middle taking-up valid data are that direct transfer of the valid data in flash memory can be achieved, and greatly save the time of garbage reclamation.It is right
The valid data of error correcting code threshold value have been reached in mistake, the maximum error correction of error correcting code has been reached, it is therefore desirable to will be effective
Data are first read back controller, are re-started after error correcting code coding, then write in flash memory.The method of technical solution of the present invention, passes through
The wrong size in valid data page is detected, whether misjudgment reaches the error correction threshold value of its error correcting code, so as to take two kinds of sides
Formula is reclaimed to the valid data in dirty piece, on the one hand ensure that the accuracy of reclaimed valid data, on the other hand
The time of garbage reclamation is saved.
It is preferred as one of technical solution of the present invention, error correcting code error correction is carried out to each valid data page in dirty piece respectively
Threshold test, with the reliability for the valid data for ensureing to write flash memory.
It is a lasting process that valid data in dirty piece, which are reclaimed, for the valid data specifically reclaimed each time, all
Need to carry out threshold test.Specifically, for each data page to be recycled, error correcting code error correction threshold value is all carried out to it
Whether detection, error correcting code error correction threshold value is reached according to mistake, to determine the recovery for selecting which kind of mode to carry out valid data.Its
In, because the characteristics of copy-back instructs cumulative error, most proceeding by when valid data are reclaimed i.e. to whole
Data page carries out the detection of error correcting code error correction threshold value, judges whether the error in data in valid data page reaches error correcting code error correction
Maximum capacity.When recovery to the valid data that mistake is not reaching to error correcting code error correction threshold value, directly by digital independent
To the register in the same packet of flash memory, then the data read in register are write to the blank page selected.Reached for mistake
The valid data of error correcting code error correction threshold value have been arrived, it is necessary to send into controller and carry out re-encoding process, flash memory could be write.This
Plant way of recycling, it can be ensured that the reliability of the valid data reclaimed every time.
It is preferred as one of technical solution of the present invention, time of the valid data of error correcting code error correction threshold value is not up to mistake
Receipts processing is specifically included,
The register that the mistake is not up in the valid data of error correcting code error correction threshold value write-in flash memory；
Valid data in register are moved on the blank page of flash memory.
It is preferred as one of technical solution of the present invention, having for error correcting code error correction threshold value is not up to by register pair mistake
When effect data are migrated, error correcting code error correction threshold test is carried out to it at any time.
The valid data of error correcting code error correction threshold value are not up to for mistake, technical solution of the present invention, which is mainly, uses copy-
Back instructions carry out the migration of data.The instruction can on the premise of data are not taken out, realize data from a Hash memory pages to
The transfer process of another Hash memory pages.This process is mainly what is realized by register.Register is to be located at flash memory internal
A part, be a kind of high speed depositing element of limited storage capacity, they can be used to temporary instruction, data and ground
Location.It is directly to carry out Data Migration using register in packet because copy-back is operated, can not utilizes and entangle in this process
Error code (ECC, Error Correction Code) completes error correction to data, thus can cause after the completion of Data Migration in data
Mistake adds up, and eventually exceeds the correction capability of error correcting code, and then influence the reliability of whole system.Therefore, returned in each rubbish
Before receipts start to migrate valid data, error correcting code error correction threshold value is carried out to the data for reclaiming each active page in block and examined
Survey, i.e., error detection is carried out to the effective page data for preparing migration first, whether misjudgment reaches ECC maximum error correction energy
Power.
It is preferred as one of technical solution of the present invention, time of the valid data of error correcting code error correction threshold value has been reached to mistake
Receipts processing is specifically included,
The valid data that the mistake has reached error correcting code error correction threshold value are read in into controller；
Valid data feeding encoder in controller is subjected to error correcting code coding；
Send the valid data encoded by error correcting code back to controller；
Judge whether the valid data in dirty piece are all collected into controller, if it is, will be effective in controller
In the free block of data write-in flash memory.
Valid data to having reached error correcting code error correction threshold value, if directly carried out using copy-back instructions to data
Transfer, mistake has reached the error correction threshold value of error correcting code, it is impossible to ensure the data reliability of write-in flash memory.Therefore need to enter it
Row error correcting code is re-encoded, it is ensured that it has self-correction ability, and flash memory is then write again.In particular it is necessary to which mistake is reached
The valid data of error correcting code error correction threshold value read in controller, i.e., by the data buffer area (buffer) of its writing controller, carry out
After error correcting code re-encoding, then the valid data by re-encoding are read back controller, then write flash memory.In addition, in dirty piece
Valid data write-in flash memory is carried out several times, therefore after the completion of a valid data transfer, in order to ensure data record
The reliability of process to be recycled dirty piece, it is necessary to be detected, the valid data in dirty piece of confirmation have all been completed after transfer,
The valid data in the data buffer area (buffer) of controller are write in flash memory again.
It is preferred as one of technical solution of the present invention, after the valid data write-in flash memory, judge effective in dirty piece
Whether data all write flash memory, if so, dirty piece of erasing.
Valid data in confirming dirty piece, which are reclaimed, to be fully completed, it is necessary to to be erased dirty piece before dirty piece is wiped
Detected, confirm that valid data therein are recovered completely, the valid data in only dirty piece are confirmed after all reclaiming,
It can be wiped.
According to another aspect of the present invention, there is provided the flash memory rubbish described in a kind of any one of application claim 1~6
Reclaim the rubbish recovering method of optimization method, it is characterised in that including,
S1 determines to carry out dirty piece of garbage reclamation；
S2 carries out error correcting code error correction threshold test to valid data current in dirty piece；
S3 judges whether valid data mistake current in dirty piece reaches the error correction threshold value of error correcting code, if reaching, goes to step
S6, otherwise, goes to step S4；
S4 perform copy-back orders, by flash memory be grouped in register progress Data Migration；
S5 judges whether valid data migration terminates, if terminating, goes to step S11, otherwise, goes to step S2；
S6 the valid data in dirty piece are read back controller data buffer zone in；
S7 carries out error correcting code coding to data；
Data are sent back in data buffer zone after S8 codings；
S9 judges that valid data are collected into whether data buffer zone is fully completed in dirty piece, if completing, goes to step S10, no
Then, S6 is gone to step；
S10 writes the valid data of collection in the free block of flash memory；
All valid data migrations of S11 garbage reclamations are completed, by dirty piece of erasing.
The rubbish recovering method of technical solution of the present invention, it is determined that after dirty piece of the rubbish to be reclaimed, first to be recycled effective
Data carry out error correcting code error correction threshold test, and the data of error correcting code error correction threshold value are not up to for mistake, can directly be used
Copy-back instructions carry out Data Migration.Like this, the valid data of error correcting code error correction threshold value are not reaching to for mistake
Recovery can be done directly in a flash memory, greatly save the time of garbage reclamation.Error correcting code error correction has been reached for mistake
The data of threshold value, then the data buffer area (buffer) for the controller that needs to be read back and after being re-encoded to it, then write
Flash memory.Specifically, it is exactly the buffer area (buffer) of controller that first valid data to be recycled read back, then by significant figure
Re-encoded according to being taken out being put into encoder.Valid data by re-encoding come back to controller cache area
(buffer), then it can write in blank flash block.On the one hand this method can be instructed by copy-back and improved
The efficiency of garbage reclamation, on the other hand can also be by reaching that the valid data page of error correcting code error correction threshold value is compiled again to mistake
Code is to ensure the reliability of data, so as to ensure the reliability of solid-state disk data storage on the whole, reduces garbage reclamation and is taken
Between, system I/O performances are improved, the reliability of solid-state disk data storage can also be further improved.
According to another aspect of the present invention, there is provided the flash memory rubbish described in a kind of any one of application claim 1~6
Reclaim the garbage retrieving system of optimization method, it is characterised in that including,
Dirty piece of judge module, for determine to be reclaimed dirty piece；
Error correcting code error correction threshold detection module, is examined for carrying out error correcting code error correction threshold value to the currently active data in dirty piece
Survey；
Data Migration module, including the first Data Migration module and the second Data Migration module, first Data Migration
Module, for by the valid data of the error correction threshold value of not up to error correcting code, by flash memory be grouped in register carry out data and move
Move；The second Data Migration module, the valid data of the error correction threshold value for being up to error correcting code, first reads back valid data
Controller, then write in flash memory；
Valid data judge module, for judging whether also there are valid data in dirty piece；
Module is wiped, for wiping dirty piece.
According to another aspect of the present invention there is provided a kind of storage device, wherein a plurality of instruction that is stored with, the instruction
Suitable for being loaded by processor and being performed：
S1 determines to be reclaimed dirty piece；
S2 carries out error correcting code error correction threshold test to the currently active data in dirty piece；
S3 writes direct the valid data of not up to error correcting code error correction threshold value in flash memory by register；It will reach
The valid data of error correcting code error correction threshold value, first read back controller, then writes in flash memory；
S4 judges whether also there are valid data in dirty piece, if so, return to step S2；
S5 wipes dirty piece.
According to another aspect of the present invention, there is provided a kind of terminal, it is characterised in that including：
Processor, is adapted for carrying out each instruction；
Storage device, suitable for storing a plurality of instruction, wherein the instruction is applied to be loaded and performed by processor：
S1 determines to be reclaimed dirty piece；
S2 carries out error correcting code error correction threshold test to the currently active data in dirty piece；
S3 writes direct the valid data of not up to error correcting code error correction threshold value in flash memory by register；It will reach
The valid data of error correcting code error correction threshold value, first read back controller, then writes in flash memory；
S4 judges whether also there are valid data in dirty piece, if so, return to step S2；
S5 wipes dirty piece.
In general, by the contemplated above technical scheme of the present invention compared with prior art, it can obtain down and show
Beneficial effect：
(1) the garbage reclamation optimization method of technical solution of the present invention, can effectively improve the efficiency of garbage reclamation, will be dirty
Before valid data write-in flash memory in block, error correcting code error correction threshold test is first carried out to it, is ensureing flash memory storage reliability
On the premise of, take garbage reclamation copy-back mode to carry out Data Migration, the data of active page in dirty piece are directly existed
Moved in packet by the register in packet in other data block, reduce the biography of the traditional Data Migration of garbage reclamation
Defeated delay, effectively increases the efficiency of garbage reclamation, and then improve the overall I/O performances of system.
(2) the garbage reclamation optimization method of technical solution of the present invention, employs error correcting code error correction threshold test mechanism, is returning
Whether ECC error correction threshold value is detected first is reached to data before receiving valid data, once copy-back operations cause valid data
Mistake is cumulative when reaching ECC maximum error correcting capabilities, that is, changes data record mode, be changed into the traditional Data Migration side of garbage reclamation
Formula.This mechanism ensure that the mistake that data occur is no more than ECC maximum error correcting capability, it is ensured that the reliability of flash memory storage.
Brief description of the drawings
Fig. 1 is the flow chart of the garbage reclamation optimization method of the flash memory storage of the embodiment of the present invention；
Fig. 2 is the garbage reclamation Data Migration structured flowchart of the embodiment of the present invention.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.As long as in addition, technical characteristic involved in each embodiment of invention described below
Not constituting conflict each other can just be mutually combined.
The purpose of technical solution of the present invention is to provide a kind of error correcting code (ECC) error correction threshold test mechanism, rubbish is returned
The Data Migration of receipts is to carry out judgement decision-making using Direct Transfer in flash memory or using traditional Data Migration mode.
In garbage reclamation Data Migration, the mistake occurred first according to the valid data in dirty piece carries out ECC error correction threshold
Value detection, judges whether mistake occurred has reached the maximum capacity of ECC error correction, if it has been reached, then using rubbish
Reclaim in traditional Data Migration mode, the data buffer of controller that the valid data in dirty piece are read back, data are carried out
ECC is encoded, and data are sent back to after data buffer and judge that valid data are collected into whether buffer is fully completed in dirty piece, is completed
The valid data of collection are write to free block other in flash memory afterwards；If the mistake occurred is not up to the maximum energy of ECC error correction
Power, then directly carry out Data Migration using copy-back orders in flash memory.
Data Migration is operated generally by copy-back and completed, and it is a kind of high-level command that solid-state disc system is provided,
The data between same packet (plane) interior different pieces of information block are supported, Data Migration is completed by the register in packet.It is logical
Cross this order, it is possible to reduce data data buffer in flash media and controller transmission in conventional rubbish removal process
Process, reduces transmission delay, improves the efficiency of garbage reclamation, and then lift the performance of whole system.However, due to copy-
Back operations are that Data Migration is carried out in packet, and it can not complete the error correction of data using error control coding (ECC), enter
During row Data Migration, mistake in data can be caused to add up, eventually exceed ECC correction capability, influence whole system
Reliability.On the other hand, as copy-back and write/erase etc. operate the loss to memory cell, the original bit of flash memory
The bit error rate (RBER) can more and more higher.In order to ensure the security of data, it is necessary to using error control coding (ECC) come to data
Carry out error-detection error-correction.But the error correcting capability of error control coding is limited, when the original bit bit error rate (RBER) exceedes certain
During one threshold value, error control coding error correction will can not meet the reliability requirement of data storage.Therefore, how to ensure that data deposit
It is a very good problem to study that the garbage reclamation of flash memory storage, which is optimized, on the premise of storage reliability.Help to reduce
The time required to garbage reclamation, system I/O performances are improved.
When directly carrying out Data Migration in flash memory using copy-back orders, it is possible to moved in all valid data
The cumulative maximum error correcting capability for reaching ECC of mistake that data occur before completing is moved, can not now continue to use copy-back
Order is migrated to data, and needs to carry out Data Migration using the traditional Data Migration mode of garbage reclamation, so adopting
When carrying out Data Migration with copy-back orders, ECC error correction threshold dector needs at any time to detect it, misjudgment
ECC maximum error correcting capability whether is reached, the reliability of flash memory storage has both been ensure that, the efficiency of garbage reclamation is improved again.Entirely
After the valid data migration in portion is completed, dirty piece is wiped.
As shown in figure 1, the garbage reclamation optimization method of the flash memory storage of the embodiment of the present invention comprises the following steps：
(1) garbage reclamation determines to be reclaimed dirty piece.Solid-state disk is after use after a while, and it is many invalid to produce
Data., it is necessary to first determine the position at the dirty piece of place of rubbish in solid-state disk before these invalid datas are removed.Specifically,
It is to determine the position for dirty piece of rubbish for needing to remove in flash memory.
(2) error correcting code (ECC) error correction threshold test is carried out to valid data current in dirty piece.Determine dirty piece of place of rubbish
, it is necessary to further be detected to this dirty piece behind position.In general, the data content in dirty piece is not all to need to remove
, need to dump to the other positions of solid-state disk for some of which valid data.In the embodiment of technical solution of the present invention,
Before transfer storage valid data, valid data have been carried out with the error correction threshold test of error correcting code.The error correction threshold value of error correcting code
It is limited, an error correcting code can only carry out error correction to a range of mistake.Specifically, the reality of technical solution of the present invention
Apply in example and error correcting code error correction threshold test is carried out to valid data, be to confirm the mistake in the valid data, if reach
The error correction scope of its error correcting code.
(3) judge whether valid data mistake current in dirty piece reaches ECC error correction threshold value (ECC maximum error correction energy
Power), if reaching, turn (6), otherwise, turn (4).The error correction threshold value of error correcting code has been reached for the classification in step (2), i.e. mistake
Or the error correction threshold value of error correcting code is not reaching to, in the embodiment of technical solution of the present invention, at different methods
Reason.Garbage reclamation of the prior art and dirty piece of removing, the method for use are that the valid data in dirty piece first read back in flash memory
Controller, then it taken out from controller to feeding encoder and carries out error correcting code coding, it is effective by what is re-encoded
Data are re-fed into controller, and flash memory is then write again.As shown in Figure 2.For whether reaching the error correction threshold value of error correcting code, to effective
The recovery of data is exactly the valid data to being not up to error correcting code error correction threshold value in different ways, in simple terms, and this is had
Effect data are write direct in flash memory by register；Valid data to having reached error correcting code error correction threshold value, first by the significant figure
According to the controller that reads back, then write in flash memory.
(4) perform copy-back orders, by flash memory be grouped in register progress Data Migration.Copy-back orders
It is a kind of high-level command that solid-state disc system is provided, supports the data between same packet (plane) interior different pieces of information block, lead to
The register crossed in packet completes Data Migration.The valid data of error correcting code error correction threshold value, the present invention are not reaching to for mistake
In the method for technical scheme, it is write direct into register to be cleaned dirty piece, sudden strain of a muscle is then write direct by register
Other clear areas in depositing.Because register is inherently located in flash memory, therefore is not reaching to error correcting code error correction threshold for mistake
The valid data of value, can just complete the transfer of valid data, it is not necessary to read and be restored again into a flash memory, effective to improve
Organic efficiency in dirty piece with small data, optimizes whole garbage collection procedure, further, optimizes the storage of flash memory
System.
(5) judge whether valid data migration terminates, if terminating, turn (11), otherwise, turn (2).Utilizing copy-back
, it is necessary to be detected at any time during progress Data Migration.In dirty piece of detection Data Migration is carried out using copy-back
Whether valid data migration completes, if it is not complete, then proceeding Data Migration, until entering line number using copy-back
It is fully completed according to the valid data migration of migration.This part valid data belongs to a part for the valid data of all migrations,
If all migration is completed for valid data in dirty piece, the recovery section in garbage collection procedure just calculates completion, just can be with
Carry out dirty piece of erasing work of next step.
(6) valid data in dirty piece are read back in the data buffer of controller.In technical solution of the present invention, to mistake
The valid data of error correcting code error correction threshold value have been reached, have been reclaimed according to the method described above.Because the mistake of this part valid data
The error correction threshold value of error correcting code is reached, in order to ensure the accuracy of valid data, it is necessary to be re-encoded to it.Controller
Buffer be a data buffer structure being located at outside flash memory, valid data to be recycled need to be introduced into controller
Buffer, then takes out from buffer, into encoder re-encoding.By the valid data of re-encoding, backtracking to control
In device buffer processed, now the error correcting code of valid data have passed through re-encoding, will in the range of its corrigendum of mistake in error correcting code
The valid data write the blank page of flash memory.
(7) error correcting code (ECC) is carried out to data to encode.The significant figure of its error correcting code error correction threshold value has been reached for mistake
According to if being directly deposited into flash memory, it is clear that the accumulation of error rate can cause the accuracy of valid data in whole system to drop
It is low, it is therefore desirable to error correcting code re-encoding process is carried out to it.The process of re-encoding, exactly in order to reduce error rate so that effectively
Number of errors in data is less than the error correction threshold value of error correcting code, optimizes the self-correction ability of valid data.
(8) data are sent back in the buffer of controller after encoding.The valid data re-encoded by error correcting code, pass through original
Road returns to controller buffer, then writes the blank page area of flash memory from controller buffer, is finally completed returning for valid data
Knock off work.
(9) judge that valid data are collected into whether buffer is fully completed in dirty piece, if completing, turn (10), otherwise, turn
(6).For the valid data in dirty piece to be recycled, it is necessary to clearly it is following some.First, the valid data in dirty piece, which are reclaimed, is point
Repeatedly carry out, a part therein is reclaimed every time.Second, entangled for the data currently reclaimed, it is necessary to carry out error correcting code to it
Different modes are selected to be reclaimed after wrong threshold test.3rd, it is necessary to remaining in dirty piece after the completion of a data record
Whether part is detected completed with the valid data recovery confirmed in whole dirty piece.In the embodiment of technical solution of the present invention
It is preferred that after the valid data whole screening in dirty piece is finished, the valid data of error correcting code threshold value are reached for mistake, by again
First it is centrally stored in controller buffer after coding, then batch writes the blank page area of flash memory.In technical solution of the present invention
In another preferred embodiment, the data that controller buffer is collected into can also be respectively written into the blank page area of flash memory.
(10) valid data of collection are write to free block other in flash memory.There are substantial amounts of data in flash memory, for
The valid data reclaimed from dirty piece, are written into a free page area.
(11) all valid data migrations of garbage reclamation are completed, by dirty piece of erasing.By technical solution of the present invention embodiment
Above-mentioned steps, the valid data in dirty piece, which are reclaimed, to be completed, can be by dirty piece of erasing.
(12) this garbage reclamation is completed.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, it is not used to
The limitation present invention, any modifications, equivalent substitutions and improvements made within the spirit and principles of the invention etc., it all should include
Within protection scope of the present invention.
Claims (10)
1. a kind of flash memory garbage reclamation optimization method, it is characterised in that including,
Pair determine reclaim dirty piece in valid data carry out error correcting code error correction threshold test；
Judge whether the valid data mistake reaches the error correction threshold value of its error correcting code；
The valid data of error correcting code error correction threshold value are not up to mistake, using copy-back instructions by the valid data by posting
Storage is write direct in flash memory；The valid data of error correcting code error correction threshold value have been reached to mistake, first the valid data read back control
Device processed, then write in flash memory.
2. a kind of flash memory garbage reclamation optimization method according to claim 1, wherein, respectively to each significant figure in dirty piece
Error correcting code error correction threshold test is carried out according to page, with the reliability for the valid data for ensureing to write flash memory.
3. a kind of flash memory garbage reclamation optimization method according to claim 1 or 2, wherein, error correcting code is not up to mistake
The recycling of the valid data of error correction threshold value is specifically included,
The register that the mistake is not up in the valid data of error correcting code error correction threshold value reading flash memory；
Valid data in register are transferred on the blank page of flash memory.
4. a kind of flash memory garbage reclamation optimization method according to any one of claims 1 to 3, wherein, pass through register pair
When the valid data that mistake is not up to error correcting code error correction threshold value are migrated, error correcting code error correction threshold test is carried out to it at any time.
5. according to a kind of flash memory garbage reclamation optimization method according to any one of claims 1 to 4, wherein, mistake has been reached
Recycling to the valid data of error correcting code error correction threshold value is specifically included,
The valid data that the mistake has reached error correcting code error correction threshold value are read in into controller；
Valid data feeding encoder in controller is subjected to error correcting code coding；
Send the valid data encoded by error correcting code back to controller；
Judge whether the valid data in dirty piece are all collected into controller, if it is, by the valid data in controller
In the free block for writing flash memory.
6. according to a kind of flash memory garbage reclamation optimization method according to any one of claims 1 to 5, wherein, the significant figure
After write-in flash memory, judge whether the valid data in dirty piece all write flash memory, if so, dirty piece of erasing.
7. the rubbish recovering method of the flash memory garbage reclamation optimization method described in application any one of claim 1~6, its feature exists
In, including,
S1 determines to carry out dirty piece of garbage reclamation；
S2 carries out error correcting code error correction threshold test to valid data current in dirty piece；
S3 judges whether valid data mistake current in dirty piece reaches the error correction threshold value of error correcting code, if reaching, goes to step S6, no
Then, S4 is gone to step；
S4 perform copy-back orders, by flash memory be grouped in register progress Data Migration；
S5 judges whether valid data migration terminates, if terminating, goes to step S11, otherwise, goes to step S2；
S6 the valid data in dirty piece are read back controller data buffer zone in；
S7 carries out error correcting code coding to data；
Data are sent back in data buffer zone after S8 codings；
S9 judges that valid data are collected into whether data buffer zone is fully completed in dirty piece, if completing, goes to step S10, otherwise, turns
Step S6；
S10 writes the valid data of collection in the free block of flash memory；
All valid data migrations of S11 garbage reclamations are completed, by dirty piece of erasing.
8. the garbage retrieving system of the flash memory garbage reclamation optimization method described in application any one of claim 1~6, its feature exists
In, including,
Dirty piece of judge module, for determine to be reclaimed dirty piece；
Error correcting code error correction threshold detection module, for carrying out error correcting code error correction threshold test to the currently active data in dirty piece；
Data Migration module, including the first Data Migration module and the second Data Migration module, the first Data Migration module,
For by the valid data of the error correction threshold value of not up to error correcting code, by flash memory be grouped in register carry out Data Migration；Institute
The second Data Migration module is stated, the valid data of the error correction threshold value for being up to error correcting code first read back valid data control
Device, then write in flash memory；
Valid data judge module, for judging whether also there are valid data in dirty piece；
Module is wiped, for wiping dirty piece.
9. a kind of storage device, wherein a plurality of instruction that is stored with, the instruction is applied to be loaded and performed by processor：
S1 determines to be reclaimed dirty piece；
S2 carries out error correcting code error correction threshold test to the currently active data in dirty piece；
S3 by the valid data of the error correction threshold value of not up to error correcting code, by flash memory be grouped in register carry out Data Migration；
The valid data of the error correction threshold value of error correcting code are up to, first valid data read back controller, then writes in flash memory；
S4 judges whether also there are valid data in dirty piece, if so, return to step S2；
S5 wipes dirty piece.
10. a kind of terminal, it is characterised in that including：
Processor, is adapted for carrying out each instruction；
Storage device, suitable for storing a plurality of instruction, wherein the instruction is applied to be loaded and performed by processor：
S1 determines to be reclaimed dirty piece；
S2 carries out error correcting code error correction threshold test to the currently active data in dirty piece；
S3 by the valid data of the error correction threshold value of not up to error correcting code, by flash memory be grouped in register carry out Data Migration；
The valid data of the error correction threshold value of error correcting code are up to, first valid data read back controller, then writes in flash memory；
S4 judges whether also there are valid data in dirty piece, if so, return to step S2；
S5 wipes dirty piece.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710434596.XA CN107301132A (en) | 2017-06-09 | 2017-06-09 | A kind of flash memory garbage reclamation optimization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710434596.XA CN107301132A (en) | 2017-06-09 | 2017-06-09 | A kind of flash memory garbage reclamation optimization method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107301132A true CN107301132A (en) | 2017-10-27 |
Family
ID=60134673
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710434596.XA Pending CN107301132A (en) | 2017-06-09 | 2017-06-09 | A kind of flash memory garbage reclamation optimization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107301132A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108415853A (en) * | 2018-03-15 | 2018-08-17 | 深圳市江波龙电子有限公司 | A kind of method, apparatus and storage device of garbage reclamation |
| CN109388343A (en) * | 2018-09-26 | 2019-02-26 | 深圳市德名利电子有限公司 | A kind of date storage method and memory |
| CN110931072A (en) * | 2019-11-28 | 2020-03-27 | 深信服科技股份有限公司 | Bad track scanning method, device, equipment and storage medium |
| CN111090595A (en) * | 2019-11-19 | 2020-05-01 | 中国航空工业集团公司西安航空计算技术研究所 | NAND FLASH garbage recovery balance optimization method |
| CN111324551A (en) * | 2018-12-13 | 2020-06-23 | 爱思开海力士有限公司 | Data storage device and operation method thereof |
| CN111367697A (en) * | 2018-12-25 | 2020-07-03 | 北京兆易创新科技股份有限公司 | Error processing method and device |
| CN112486415A (en) * | 2020-11-30 | 2021-03-12 | 北京泽石科技有限公司 | Garbage recycling method and device of storage device |
| TWI740454B (en) * | 2020-04-14 | 2021-09-21 | 慧榮科技股份有限公司 | Method and apparatus for accessing to data in response to a power-supply event |
| US11342008B2 (en) | 2020-04-14 | 2022-05-24 | Silicon Motion, Inc. | Method and apparatus for accessing to data in response to power-supply event |
| CN117033268A (en) * | 2023-10-09 | 2023-11-10 | 苏州元脑智能科技有限公司 | SSD controller, data processing method, device and storage medium |
| CN117278176A (en) * | 2023-11-23 | 2023-12-22 | 中科亿海微电子科技（苏州）有限公司 | FPGA configuration file generation device and method with error correction function |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101634938A (en) * | 2009-08-20 | 2010-01-27 | 成都市华为赛门铁克科技有限公司 | Data migration method and data migration device of solid state disk and solid state disk |
| CN101923448A (en) * | 2010-02-08 | 2010-12-22 | 安凯(广州)微电子技术有限公司 | Method for reading and writing conversion layer of NAND flash memory |
| CN102279809A (en) * | 2011-08-10 | 2011-12-14 | 郏惠忠 | Method for redirecting write in and garbage recycling in solid hard disk |
| CN103218274A (en) * | 2013-03-15 | 2013-07-24 | 华为技术有限公司 | Failure accumulation preventing method and solid state disk |
| CN103336744A (en) * | 2013-06-20 | 2013-10-02 | 华中科技大学 | Garbage recovery method for solid-state storage device and system for garbage recovery method |
| CN105022695A (en) * | 2014-04-21 | 2015-11-04 | 群联电子股份有限公司 | Data storage method, memorizer control circuit unit and memorizer storage device |
| US20160092304A1 (en) * | 2014-09-30 | 2016-03-31 | Emc Corporation | Method and system for improving flash storage utilization by predicting bad m-pages |
| CN106502583A (en) * | 2016-10-12 | 2017-03-15 | 记忆科技(深圳)有限公司 | A kind of method for reducing solid state hard disc operating lag |
| CN106547491A (en) * | 2016-11-22 | 2017-03-29 | 郑州云海信息技术有限公司 | The method and system of data-moving in a kind of solid state hard disc |
| CN106776109A (en) * | 2016-12-26 | 2017-05-31 | 湖南国科微电子股份有限公司 | Solid state hard disc read error detection means and reading can not entangle the detection method of error reason |
-
2017
- 2017-06-09 CN CN201710434596.XA patent/CN107301132A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101634938A (en) * | 2009-08-20 | 2010-01-27 | 成都市华为赛门铁克科技有限公司 | Data migration method and data migration device of solid state disk and solid state disk |
| CN101923448A (en) * | 2010-02-08 | 2010-12-22 | 安凯(广州)微电子技术有限公司 | Method for reading and writing conversion layer of NAND flash memory |
| CN102279809A (en) * | 2011-08-10 | 2011-12-14 | 郏惠忠 | Method for redirecting write in and garbage recycling in solid hard disk |
| CN103218274A (en) * | 2013-03-15 | 2013-07-24 | 华为技术有限公司 | Failure accumulation preventing method and solid state disk |
| CN103336744A (en) * | 2013-06-20 | 2013-10-02 | 华中科技大学 | Garbage recovery method for solid-state storage device and system for garbage recovery method |
| CN105022695A (en) * | 2014-04-21 | 2015-11-04 | 群联电子股份有限公司 | Data storage method, memorizer control circuit unit and memorizer storage device |
| US20160092304A1 (en) * | 2014-09-30 | 2016-03-31 | Emc Corporation | Method and system for improving flash storage utilization by predicting bad m-pages |
| CN106502583A (en) * | 2016-10-12 | 2017-03-15 | 记忆科技(深圳)有限公司 | A kind of method for reducing solid state hard disc operating lag |
| CN106547491A (en) * | 2016-11-22 | 2017-03-29 | 郑州云海信息技术有限公司 | The method and system of data-moving in a kind of solid state hard disc |
| CN106776109A (en) * | 2016-12-26 | 2017-05-31 | 湖南国科微电子股份有限公司 | Solid state hard disc read error detection means and reading can not entangle the detection method of error reason |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108415853A (en) * | 2018-03-15 | 2018-08-17 | 深圳市江波龙电子有限公司 | A kind of method, apparatus and storage device of garbage reclamation |
| CN109388343A (en) * | 2018-09-26 | 2019-02-26 | 深圳市德名利电子有限公司 | A kind of date storage method and memory |
| CN111324551B (en) * | 2018-12-13 | 2023-04-07 | 爱思开海力士有限公司 | Data storage device and operation method thereof |
| CN111324551A (en) * | 2018-12-13 | 2020-06-23 | 爱思开海力士有限公司 | Data storage device and operation method thereof |
| CN111367697A (en) * | 2018-12-25 | 2020-07-03 | 北京兆易创新科技股份有限公司 | Error processing method and device |
| CN111090595A (en) * | 2019-11-19 | 2020-05-01 | 中国航空工业集团公司西安航空计算技术研究所 | NAND FLASH garbage recovery balance optimization method |
| CN111090595B (en) * | 2019-11-19 | 2022-12-20 | 中国航空工业集团公司西安航空计算技术研究所 | NAND FLASH garbage recovery balanced optimization method |
| CN110931072B (en) * | 2019-11-28 | 2022-03-22 | 深信服科技股份有限公司 | Bad track scanning method, device, equipment and storage medium |
| CN110931072A (en) * | 2019-11-28 | 2020-03-27 | 深信服科技股份有限公司 | Bad track scanning method, device, equipment and storage medium |
| TWI740454B (en) * | 2020-04-14 | 2021-09-21 | 慧榮科技股份有限公司 | Method and apparatus for accessing to data in response to a power-supply event |
| US11342008B2 (en) | 2020-04-14 | 2022-05-24 | Silicon Motion, Inc. | Method and apparatus for accessing to data in response to power-supply event |
| US11664056B2 (en) | 2020-04-14 | 2023-05-30 | Silicon Motion, Inc. | Method and apparatus for accessing to data in response to power-supply event |
| CN112486415A (en) * | 2020-11-30 | 2021-03-12 | 北京泽石科技有限公司 | Garbage recycling method and device of storage device |
| CN112486415B (en) * | 2020-11-30 | 2024-05-24 | 北京泽石科技有限公司 | Garbage collection method and device for storage device |
| CN117033268A (en) * | 2023-10-09 | 2023-11-10 | 苏州元脑智能科技有限公司 | SSD controller, data processing method, device and storage medium |
| CN117033268B (en) * | 2023-10-09 | 2024-02-09 | 苏州元脑智能科技有限公司 | SSD controller, data processing method, device and storage medium |
| CN117278176A (en) * | 2023-11-23 | 2023-12-22 | 中科亿海微电子科技（苏州）有限公司 | FPGA configuration file generation device and method with error correction function |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107301132A (en) | A kind of flash memory garbage reclamation optimization method | |
| CN102012867B (en) | Data storage system | |
| CN104584135B (en) | Read to clean the flash memory of algorithm using target | |
| CN106598878B (en) | Solid state disk cold and hot data separation method | |
| CN103377010B (en) | The system and method for managing the unreliable memory in data-storage system | |
| JP4812739B2 (en) | Non-volatile data storage device static data area detection method, wear leveling method, data unit merging method and device | |
| CN106776376B (en) | Buffer storage supervisory method, memorizer control circuit unit and storage device | |
| TWI587136B (en) | Flash memory system and managing and collection methods for flash memory with invalid page information thereof | |
| CN103488583A (en) | High-performance reliable solid-state disk realizing method | |
| US8996791B2 (en) | Flash memory device, memory control device, memory control method, and storage system | |
| CN104866429B (en) | Storage management method, memorizer control circuit unit and memory storage apparatus | |
| CN106170773A (en) | On naked core, the selectivity of buffer-type nonvolatile memory returns and copies | |
| CN104484283B (en) | A kind of method for reducing solid state disk write amplification | |
| CN105528301A (en) | NAND Flash memory garbage collection method | |
| CN110471788A (en) | The data structure that asynchronous power loss influences | |
| CN105260325A (en) | Method for collecting garbage blocks in solid state disks | |
| US20130060991A1 (en) | Solid state drive and garbage collection control method thereof | |
| CN107221351B (en) | Optimization processing method of error correcting code in solid-state disk system and application thereof | |
| CN111538618A (en) | Apparatus and techniques for one-time parity check | |
| CN106649132A (en) | Solid-state drive junk recovery method | |
| CN102063266A (en) | Nonvolatile memory controller and method for writing data to nonvolatile memory | |
| CN107391300A (en) | A kind of method and system for improving flash data memory reliability | |
| Chen et al. | Enabling sub-blocks erase management to boost the performance of 3D NAND flash memory | |
| CN103577342A (en) | Method for managing data stored in flash memory and related memory device and controller | |
| CN109710541A (en) | For the optimization method of NAND Flash main control chip Greedy garbage reclamation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171027 |
|
| RJ01 | Rejection of invention patent application after publication |