CN107656875A - Solid state hard disc as system disk shortens the method and system of power-on time - Google Patents
Solid state hard disc as system disk shortens the method and system of power-on time Download PDFInfo
- Publication number
- CN107656875A CN107656875A CN201710833652.7A CN201710833652A CN107656875A CN 107656875 A CN107656875 A CN 107656875A CN 201710833652 A CN201710833652 A CN 201710833652A CN 107656875 A CN107656875 A CN 107656875A
- Authority
- CN
- China
- Prior art keywords
- related data
- management table
- lba
- data management
- reading
- 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.)
- Granted
Links
- 239000007787 solid Substances 0.000 title claims abstract description 49
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000013523 data management Methods 0.000 claims abstract description 83
- 230000015654 memory Effects 0.000 claims abstract description 33
- 230000017702 response to host Effects 0.000 claims abstract description 7
- 230000004044 response Effects 0.000 claims description 11
- 230000001186 cumulative effect Effects 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 8
- 238000007726 management method Methods 0.000 claims description 7
- 238000013507 mapping Methods 0.000 description 28
- 238000011084 recovery Methods 0.000 description 14
- 230000005611 electricity Effects 0.000 description 11
- 230000008859 change Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000005192 partition Methods 0.000 description 3
- 241001269238 Data Species 0.000 description 2
- 230000003321 amplification Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000003199 nucleic acid amplification method Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
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
- G06F12/0292—User address space allocation, e.g. contiguous or non contiguous base addressing using tables or multilevel address translation means
-
- 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
-
- 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
- G06F12/06—Addressing a physical block of locations, e.g. base addressing, module addressing, memory dedication
- G06F12/0646—Configuration or reconfiguration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2212/00—Indexing scheme relating to accessing, addressing or allocation within memory systems or architectures
- G06F2212/10—Providing a specific technical effect
- G06F2212/1016—Performance improvement
- G06F2212/1021—Hit rate improvement
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Techniques For Improving Reliability Of Storages (AREA)
Abstract
The invention discloses the method and system that a kind of solid state hard disc as system disk shortens power-on time, it is related to solid state hard disc field.This method comprises the following steps：Starting record in related data management table has the LBA entry that each power up is read, when each SSD firmwares detect electric, related data management table will be started to be loaded into SSD internal memories, the LBA entry started in related data management table is updated into L2P tables successively, until L2P is rebuild successfully；In power up, if the LBA entry in Host read hit L2P tables, directly in response to Host read；If it is not, L2P is waited to be responded after rebuilding successfully.The present invention can shorten and the fixed SSD upper electroresponse time.
Description
Technical field
The present invention relates to solid state hard disc field, is specifically related to a kind of solid state hard disc as system disk and shortens power-on time
Method and system.
Background technology
No matter in PC or the data center of enterprise-level, SSD (Solid State Drive, solid state hard disc) because
For characteristics such as its superior io throughput (high input and output throughput performance)/command latency (low time delay), just
Become more popular.The firmware (firmware) of solid state hard disc needs to safeguard one from LBA (Logical Block
Address, logical data address, host give the address of the distribution per segment data) arrive storage PPA (Physical Page
Address, physical data address, in particular to the address in solid state hard disc) L2P mapping tables (Logical to Physical
Mapping table, mapping table of the logic to physical address in solid state hard disc), it is used as the address conversion of read-write operation.
As the single-deck capacity of solid state hard disc is just becoming increasing, the capacity that L2P needs to represent also increases therewith, such as
Logical capacity is 2TB solid state hard disc, if taking the page base mapping (page mapping mode) of 4K units, each PPA
Account for 4 byte (byte), then whole section of L2P table has 2GB size.No matter which kind of mapping mode is taken：page mapping
(page mapping)/block mapping (block mapping)/Hybrid mapping (mixed-use developments), it is how electric on solid state hard disc
The times of Shi Houyong quickly recover L2P mapping tables, are just becoming more and more difficult.
At present for the application of solid state hard disc, two classes are roughly divided into, one kind is used as system disk, that is, that generally says be equipped with
The boot disk of operating system；Another is data disks, and the inside does not have the application data of operating system, only main frame storage.
For system disk, no matter apply in PC or enterprise's storage array, response is removed when electric thereon
ADMIN orders are not talked, and mainly hair read command comes bootload partition information (mode such as MBR or GPT) and operation system first
Unite (windows/linux etc.), it is hard for SATA solid-states for personal PC environment under the windows operating systems of main flow
The main start-up course of disk is as follows：
1.BIOS launches into POST (complete machine powering recovery) process, while is powered to solid state hard disc.
2. after SATA physical layers, chain path level initialization, BIOS can be by sending out Identify/Get feature etc.
Ata command initializes solid state hard disc.
3.BIOS reads the information such as boot partition, confirms the information such as operating system position.
4.BIOS read operation systems, control transfers operating system to after loading.
5. operating system initialization, read corresponding operating internal system file.
After 6. operating system itself initializes, after reading the information such as system drive and loading, used into operating system
Family interface.
In step listed above, main frame (herein referring to system BIOS and operating system) required reading in power up
The data taken, are all fixed LBA, and this partial data, for solid state hard disc, belong to " extremely cold " data, that is,
Say, it is considerably less to be written into the number of renewal, such as operating system is in itself, and may be to be read upper electric when.
For such case, SSD firmwares first have to " to learn " which LBA electricity needs, and SSD firmwares need to safeguard one
Bootup reference LBA table (start related data management table), for recording these LBA, and can be according to specific
Corresponding LBA is added and replaced to situation.Meanwhile SSD firmwares will also start related data management table when being necessary according to
Certain algorithm write-in flash memory, it is ensured that can find this table very fast when upper electric.
The also most important, it is exactly it has to be ensured that data integrity.No matter normal powering down recovery process, also right and wrong
Normal powering down recovery process, this table is necessarily all referring to newest and correct LBA positions.Because these data are system core numbers
According to if loss or corrupt data, it is more likely that cause system not start.
The way that upper electricity traditional at present recovers L2P mapping tables is：SSD firmware (firmware) as desired reflect L2P
Penetrate as internal system meta data (metadata, in particular to the key message of SSD firmware internal records), it is regular or indefinite
In the write-in flash (flash memory) of phase, either normal powering down (for SATA solid state hard discs, can be sent out before system power failure
The PM such as Standby immediately/Idle immediately (power management) is ordered；For NVME solid state hard discs,
System can enable CC.SHN shutdown notification to notify solid state hard disc to shut down) or improper power down (
Be exactly that main frame does not have to corresponding off signal) upper electric recovery process, SSD firmwares be according to data write order, according to
It is secondary to read out these L2P mapping pages, for restoration and reconstruction L2P mapping tables, there are L2P mapping tables, SSD could normal response master
Any read write command that machine is sent.
As described above, only whole L2P mapping tables whole recovery (upper electricity is rebuild) is completed, solid state hard disc can just respond
The read write command that main frame is sent.When L2P recovery process becomes very slowly, time of solid state hard disc energy response system order just with
It is elongated.Particularly with improper power down, required recovery L2P map informations are not write into flash memory by possible SSD firmwares in time,
In this case, SSD firmwares need to scan the data of substantial amounts of main frame write-in, corresponding with every section of host data by reading
Metadata, to recover L2P mappings, data integrity issues are removed, the problem of such method is brought mainly there are two：
(1) power-on time becomes longer.For example, in the solid state hard disc of 2TB capacity, solid state hard disc FTL (Flash
Translation Layer, a part for solid state hard disc firmware, be responsible for safeguarding L2P mapping tables) firmware take page map ring
Border, it is assumed that improper power down is, it is necessary to scan 50G capacity, and bottom tape reading width is if 2GB/s, then whole scanning process is then
Need at least 25s.
(2) electric recovery time becomes uncertain on.In the case of especially improper power down, the time point of power down is can not
Expect, so it is also unfixed that solid state hard disc FTL firmwares, which need the data volume that scans, the recovery time brought therewith is also
It is unfixed.
Some manufacturers, in order to solve this problem, the operating frequency that is mapped to flash memory will be write and accelerated, to shorten power-on time,
This refreshing mode, it is generally the case that can be by the most hot data (number that namely main frame newly writes and may often operated
According to) corresponding to mapping frequently write into flash memory, so bring the problem of have it is following some：
A. amplification (WA) influence is write, more metadata are written into flash memory, influence the life-span of solid state hard disc.
B. performance impact.The bandwidth for writing flash memory and being certain to take bottom flash memory will be mapped, then mapping is write into flash memory
This period, the write performance of solid state hard disc will be impacted.
Although c. most hot mapping data are by regular write-in, in upper electric recovery process, main frame operates in fact first
It is not these data, but these so-called cold datas such as boot partition, operating system.So these dsc datas write
Algorithm can accelerate whole section of L2P mapping recovery process, but these mappings have spent many times to recover, can main frame access first and
It is not these data.
Currently existing scheme, no matter increase the frequency of mapping write-in flash memory, or other such as increase the side of similar L2P daily records
Formula, be required for whole section of L2P to be all recovered, could normally respond the read write command that main frame is sent, this method using
L2P is treated as an entirety, is indifferent to which data firstly the need of being read by blind reading L2P tables, SSD firmwares, and algorithm comparison is simple,
But as described above, the problem of bringing therewith is also obvious.How SSD firmwares recognize and first by self-teaching
Fast quick-recovery those LBA, adaptively matching main frame that can really be read first in power up each power up,
It is technical problem urgently to be resolved hurrily to shorten and fix the SSD upper electroresponse time so as to optimize FTL algorithms.
The content of the invention
The invention aims to overcome the shortcomings of above-mentioned background technology, there is provided a kind of solid state hard disc as system disk
Shorten the method and system of power-on time, can shorten and fix the SSD upper electroresponse time.
The present invention provides a kind of method that solid state hard disc as system disk shortens power-on time, comprises the following steps：
Starting record in related data management table has the LBA entry that each power up is read, each SSD firmwares detecting
To it is upper electric when, by start related data management table be loaded into SSD internal memories, will start related data management table in LBA entry
L2P tables are arrived in renewal successively, until L2P is rebuild successfully；
In power up, if the LBA entry in Host read hit L2P tables, directly in response to Host read；
If it is not, L2P is waited to be responded after rebuilding successfully.
On the basis of above-mentioned technical proposal, this method is further comprising the steps of：
In power up, if the number that Host read is hit reaches preset times, SSD firmwares are cumulative to start phase
Close the reading times in data management table.
On the basis of above-mentioned technical proposal, this method is further comprising the steps of：
It is if some LBA entry is not in related data management table is started, reading times are minimum in power up
LBA entry replaces successively.
On the basis of above-mentioned technical proposal, specific mistake that the LBA entry that reading times are minimum replaces successively
Cheng Wei：If the reading times maximum of each LBA entry is preset value, and it is signed number, then every time in power up
All hit；If reading times are less than preset value, then reading times add up；If reading times are equal to preset value, maintain
Preset value is constant；If LBA entry is not hit, reading times subtract one.
On the basis of above-mentioned technical proposal, this method is further comprising the steps of：
Reading order according to related data management table is started establishes a queue to sort：
Reading order is smaller, show it is more early read, start related data management table initialization after, each LBA of the inside
List item is loaded into L2P tables, and L2P table others list items are then initialized to represent the set form of illegal value；SSD firmwares receive
After Host read, L2P tables are searched, if LBA entry is not illegal value, then it represents that the LBA entry is to start related data
Manage table hit, then according to current reading order, as index, go to search the order to have sorted, find the LBA table
；
Reading times are negative, show that LBA entry can be replaced, and during initialization, the reading time of which LBA entry found
Number is negative, and is recorded, when there is Host read not hit startup related data management table, then by these LBA
List item replaces according to the order of reading times；To be the LBA entry of negative according to order from small to large by reading times
A queue is lined up, replaces all replaced since afterbody every time, and update reading order and reading times therein, PPA domains；
Finally, when it is all it is upper it is electronic complete, L2P has also recovered completely, then will start related data management table weight
Newly sorted according to reading order, and be stored in flash memory, in case next time uses.
The present invention also provides the system that a kind of solid state hard disc as system disk shortens power-on time, and the system includes L2P
Table reconstruction unit, response unit, wherein：
L2P table reconstruction units are used for：Starting record in related data management table has the LBA table that each power up is read
, when each SSD firmwares detect electric, related data management table will be started and be loaded into SSD internal memories, related data will be started
LBA entry in management table updates L2P tables successively, until L2P tables are rebuild successfully；
Response unit is used for：In power up, if the LBA entry in Host read hit L2P tables, directly in response to
Host read；If it is not, L2P is waited to be responded after rebuilding successfully.
On the basis of above-mentioned technical proposal, the system also includes reading times summing elements, is used for：In power up,
If the number that Host read is hit reaches preset times, the cumulative reading started in related data management table of SSD firmwares
Take number.
On the basis of above-mentioned technical proposal, the system also includes LBA entry replacement unit, is used for：In power up, such as
Some LBA entry of fruit then replaces the minimum LBA entry of reading times not in related data management table is started successively.
On the basis of above-mentioned technical proposal, specific mistake that the LBA entry that reading times are minimum replaces successively
Cheng Wei：If the reading times maximum of each LBA entry is preset value, and it is signed number, then every time in power up
All hit；If reading times are less than preset value, then reading times add up；If reading times are equal to preset value, maintain
It is constant；If LBA entry is not hit, reading times subtract one.
On the basis of above-mentioned technical proposal, the system also establishes unit including queue, is used for：According to startup related data
The reading order of management table establishes a queue to sort：
Reading order is smaller, show it is more early read, start related data management table initialization after, each LBA of the inside
List item is loaded into L2P tables, and L2P table others list items are then initialized to represent the set form of illegal value；SSD firmwares receive
After Host read, L2P tables are searched, if LBA entry is not illegal value, then it represents that the LBA entry is to start related data
Manage table hit, then according to current reading order, as index, go to search the order to have sorted, find the LBA table
；
Reading times are negative, show that LBA entry can be replaced, and during initialization, the reading time of which LBA entry found
Number is negative, and is recorded, when there is Host read not hit startup related data management table, then by these LBA
List item replaces according to the order of reading times；To be the LBA entry of negative according to order from small to large by reading times
A queue is lined up, replaces all replaced since afterbody every time, and update reading order and reading times therein, PPA domains；
Finally, when it is all it is upper it is electronic complete, L2P has also recovered completely, then will start related data management table weight
Newly sorted according to reading order, and be stored in flash memory, in case next time uses.
Compared with prior art, advantages of the present invention is as follows：
Record has the LBA entry that each power up is read, each SSD in the startup related data management table of the present invention
When firmware detects electric, the present invention is loaded into related data management table is started in SSD internal memories, and will start related data management
LBA entry in table updates L2P tables successively, until L2P is rebuild successfully；In power up, if Host read is hit
LBA entry in L2P tables, directly in response to Host read；If it is not, L2P is waited to be responded after rebuilding successfully.The present invention
The start experience of the Large Copacity solid state hard disc as system disk can be obviously improved, shortens and fixed the SSD upper electroresponse time.
No matter recover for the upper electricity after normal powering down, or the upper electricity after improper power down recovers, and power-on time all greatly contracts
It is short, and keep stable.
Brief description of the drawings
Fig. 1 is the entry format schematic diagram for starting related data management table in the embodiment of the present invention.
Fig. 2 is in the embodiment of the present invention in the case of normal powering down, starts related data and manages table on flash array
Distribution situation schematic diagram.
Fig. 3 is in the embodiment of the present invention under improper power-down conditions, starts related data and manages table on flash array
Distribution situation schematic diagram.
Embodiment
Below in conjunction with the accompanying drawings and specific embodiment the present invention is described in further detail.
The embodiment of the present invention provides a kind of method that solid state hard disc as system disk shortens power-on time, including following step
Suddenly：
Starting record in related data management table has the LBA entry that each power up is read, each SSD firmwares detecting
To it is upper electric when, by start related data management table be loaded into SSD internal memories, will start related data management table in LBA entry
L2P tables are arrived in renewal successively, until L2P is rebuild successfully；
In power up, if the LBA entry in Host read hit L2P tables, directly in response to Host read；
If it is not, L2P is waited to be responded after rebuilding successfully.
In power up, if the number that Host read is hit reaches preset times, SSD firmwares are cumulative to start phase
Close the reading times in data management table.
It is if some LBA entry is not in related data management table is started, reading times are minimum in power up
LBA entry replaces successively, and detailed process is：In power up, if the reading times maximum of each LBA entry is default
Value, and be signed number, then hit every time；If reading times are less than preset value, then reading times add up；If read
Take number to be equal to preset value, then maintain preset value constant；If LBA entry is not hit, reading times subtract one.
Reading order according to related data management table is started establishes a queue to sort：
Reading order is smaller, show it is more early read, start related data management table initialization after, each LBA of the inside
List item is loaded into L2P tables, and L2P table others list items are then initialized to represent the set form of illegal value；SSD firmwares receive
After Host read, L2P tables are searched, if LBA entry is not illegal value, then it represents that the LBA entry is to start related data
Manage table hit, then according to current reading order, as index, go to search the order to have sorted, find the LBA table
；
Reading times are negative, show that LBA entry can be replaced, and during initialization, the reading time of which LBA entry found
Number is negative, and is recorded, when there is Host read not hit startup related data management table, then by these LBA
List item replaces according to the order of reading times；To be the LBA entry of negative according to order from small to large by reading times
A queue is lined up, replaces all replaced since afterbody every time, and update reading order and reading times therein, PPA domains；
Finally, when it is all it is upper it is electronic complete, L2P has also recovered completely, then will start related data management table weight
Newly sorted according to reading order, and be stored in flash memory, in case next time uses.
The embodiment of the present invention also provides the system that a kind of solid state hard disc as system disk shortens power-on time, the system bag
Include L2P tables reconstruction unit, unit is established in response unit, reading times summing elements, LBA entry replacement unit, queue, wherein：
L2P table reconstruction units are used for：Starting record in related data management table has the LBA table that each power up is read
, when each SSD firmwares detect electric, related data management table will be started and be loaded into SSD internal memories, related data will be started
LBA entry in management table updates L2P tables successively, until L2P tables are rebuild successfully.
Response unit is used for：In power up, if the LBA entry in Host read hit L2P tables, directly in response to
Host read；If it is not, L2P is waited to be responded after rebuilding successfully.
Reading times summing elements, are used for：In power up, if the number that Host read is hit reaches default
Number, the cumulative reading times started in related data management table of SSD firmwares.
LBA entry replacement unit, is used for：In power up, if some LBA entry is not starting related data management table
In, then the minimum LBA entry of reading times is replaced successively, detailed process is：In power up, if each LBA entry
Reading times maximum be preset value, and be signed number, then hit every time；If reading times are less than preset value,
So reading times add up；If reading times are equal to preset value, remain unchanged；If LBA entry is not hit, read
Number subtracts one.
Unit is established in queue, is used for：Reading order according to related data management table is started establishes a queue to sort：
Reading order is smaller, show it is more early read, start related data management table initialization after, each LBA of the inside
List item is loaded into L2P tables, and L2P table others list items are then initialized to represent the set form of illegal value；SSD firmwares receive
After Host read, L2P tables are searched, if LBA entry is not illegal value, then it represents that the LBA entry is to start related data
Manage table hit, then according to current reading order, as index, go to search the order to have sorted, find the LBA table
；
Reading times are negative, show that LBA entry can be replaced, and during initialization, the reading time of which LBA entry found
Number is negative, and is recorded, when there is Host read not hit startup related data management table, then by these LBA
List item replaces according to the order of reading times；To be the LBA entry of negative according to order from small to large by reading times
A queue is lined up, replaces all replaced since afterbody every time, and update reading order and reading times therein, PPA domains；
Finally, when it is all it is upper it is electronic complete, L2P has also recovered completely, then will start related data management table weight
Newly sorted according to reading order, and be stored in flash memory, in case next time uses.
The principle of the embodiment of the present invention is illustrated in detail below.
Bootup reference LBA table (starting related data management table) are by SSD firmware maintenances, for remembering
The relevant informations such as the LBA that each power up is read are recorded, and when these LBA are updated, are read how many times, as
Foundation, to decide whether renewal/replacement LBA therein, it is specifically defined as shown in figure 1, wherein every content is as follows：
Byte 0-3：LBA, represents the LBA addresses that a upper electricity is accessed to, and bit wide need to be supported according to current SSD firmwares
Maximum LBA bit wides determine, represented with 4 bytes in this text；
Byte 4-7：PPA, represent the PPA positions that this LBA is be mapped to.This mapping relations must protect with current L2P tables
Hold consistent.Similarly, bit wide is defined by the bit wide that current SSD firmwares define, and is equally represented in this text with 4 bytes；
Byte 8-15：Read Order (reading order), represent the order that this LBA is read in power up；
Byte16：Read Count (reading times), the number that this LBA is read in power up, number is bigger, table
Show and be more required in power up；
Byte17：Next pointer, for managing table as next startup related data in list structure.
Situations such as considering the use of SSD internal memories and SSD capacity, the space of each list item is reasonably distributed, and it is fixed
Make rational size.How many entry number, it can be used for being recorded and replacing with regard to how many LBA, be represented exemplified by 100,
So whole table needs 100*18=1800 byte, if each PPA represents the data space of 4K sizes, then, 100*4K
Can represents：The data of 4M sizes can be recorded, and by very fast recovery in power up.
Power up：
Each SSD firmwares, which detect, establishes the beginning by cable, it is necessary to find startup dependency number before the recovery of whole L2P mapping tables
According to management table, after finding and being loaded into SSD internal memories, each LBA entry in this table is updated arrive L2P tables successively, now one
The L2P tables of breviary version are just rebuild successfully.In power up, if Host read hits these LBA entries, it is possible to straight
Connect response Host read；If it is not, need to wait for responding after the L2P is rebuild successfully.
In power up, if the number that Host read is hit reaches preset value, such as 50 times, SSD firmwares are also
The cumulative reading times started in related data management table are needed, with the increase of upper electric number, some LBA reading times are just
What can be become is increasing, then naturally, these LBA are more needed by system electrification process.
If some LBA in related data management table is started, is not needed the minimum list item of reading times successively
Replace.It is to be noted here that in most cases, as system disk, LBA in the table is all without being replaced.Read secondary
The low LBA of number can be trusted replacement.It is contemplated that operating system change or solid state hard disc application environment acutely change, have
List item that may be very more is not all hit, at this moment, if the reading times of these list items have been added to bigger value,
Erroneous judgement is may result in, concrete condition is as follows：
1. assume that starting related data management table there are 5 list items, have recorded LBA 0-4, its reading times such as following table respectively
It is shown：
| LBA | 0 | 1 | 2 | 3 | 4 |
| Reading times | 255 | 255 | 255 | 10 | 10 |
Example 1：Assuming that electric LBA0-4 is not hit on certain, if LBA0-4 all to be replaced into new LBA, then
Power up after being likely to still needs to read LBA0-4, such as certain user pulls the SSD for itself serving as system disk,
It is inserted on other machines and is used as data disks, disk is put back to again after some data of copy and is re-used as system disk, then can also afterwards
LBA0-4 is hit, so repeatedly, not but not probability is hit in increase, and can cause to start related data management table write-in flash memory
Excessively frequently, influence solid state hard disc writes amplification.
Example 2：Still assume that electric LBA0-4 is not ordered on certain, if simply replace the small LBA3 of reading times and
LBA4, then if all it is later this situation, that is, or else LBA0-4 can be hit, then 255 power up are at most needed,
All list items could be replaced, hence it is evident that do not have acceleration effect.
Certainly there is a lot of other situations that SSD firmwares can be caused to judge by accident, the method that this scheme is taken is：Power up
In, if the reading times maximum of each list item is 10 (or other reasonable values), and it is signed number, then orders every time
In；If reading times are less than 10, then reading times add up；Equal to 10, then maintenance 10 is constant；If the LBA is not hit,
Then reading times subtract one, can be kept to negative value.If a new LBA entry is added into, cumulative, replacement is started counting up from 0
Fall the list item for negative value, so by the increase and decrease of reading times, substantially 5 replacements that can complete LBA, while can prevent from using again
Start the list item being recorded in related data management table in family 5 times caused by other operations to be replaced quickly, detailed process
It is as follows：
1. initial LBA entry is as follows：
2.LBA2, LBA3, LBA4 are not hit, then reading times subtract one：
| LBA | 0 | 1 | 2 | 3 | 4 | ||||||
| Reading times | 5 | 5 | 4 | 4 | 4 |
3.LBA2, LBA3, LBA4 are persistently not hit by 4 times, then reading times are reduced to negative value：
| LBA | 0 | 1 | 2 | 3 | 4 | ||||||
| Reading times | 5 | 5 | -1 | -1 | -1 |
4.LBA5, LBA6, LBA7 are persistently hit, then LBA2, LBA3, LBA4 are replaced.
| LBA | 0 | 1 | 5 | 6 | 7 | ||||||
| Reading times | 5 | 5 | 1 | 1 | 1 |
It is suitable according to the reading for starting related data management table in order to which quick search starts related data management table when upper electric
Sequence establishes a queue to sort.
1. reading order is smaller, indicate that it is more early read, so after related data management table initialization is started, the inside
Each LBA entry be loaded into L2P tables, L2P table others list items are then initialized to represent the set form of illegal value, example
Such as full 0 or complete 1.First, after SSD firmwares receive Host read, L2P tables are searched, if the list item is not one illegal
Value, then it represents that the list item is to start related data management table to hit, then according to current reading order, as index, is gone
The order to have sorted is searched, finds the LBA entry.
2. reading times are negative, show that the list item can be replaced, so in initialization, which list item found
Reading times be negative, and recorded, when thering is Host read not hit to start related data management table, then
These list items are replaced according to the order of reading times.Can by reading times, by these for negative list item according to from
It is small to line up a queue to big order, every time replace will be replaced since afterbody, and update reading order therein and
The domains such as reading times, PPA.
Finally, when it is all it is upper it is electronic complete, L2P has also recovered completely, then needs that related data management will be started
Table sorts according to reading order again, and is stored in flash memory in case next time uses.
Content update and maintenance after electricity on 3.
After the completion of upper electronic work, these LBA entries in related data management table is started can be updated at any time, existing
New position may be rewritten to by main frame write request, it is also possible to by the garbage collection (garbage reclamations inside SSD
Mechanism), the internal module such as wearleveling (Wear-leveling algorithms) be transported to new position as desired, so SSD firmwares
New map updating must be managed in table to related data is started in time.If each main frame write request is all starting correlation
Searched in data management table, in the case where list item is relatively more, possible code overhead is bigger, then can be periodically on backstage
Or before needing this table refreshing write-in flash memory, the LBA of each record is traveled through, L2P is searched, to update what is wherein recorded
PPA positions.If time or long of single renewal, it may be considered that segmentation is handled.
4. refresh write-in flash memory mechanism
Starting related data management table needs to write in flash memory, and SSD firmwares need to quickly find during upper electricity
This position, and load.Here there can be multiple choices, if with a kind of nor flash (non-volatile flash memory) conduct
The solid state hard disc of auxiliary is electrically activated, content can be existed to nor flash fixed positions, upper electricity directly reads nor flash every time
Load can.For no nor flash solid state hard disc, then need that related data management table deposit flash memory will be started
In.
Die is the set of several blocks, and several die are had in a flash chip, can will be in each die0 (block collection
Close page0 (page 0) write-ins of block (block) 0), quilt recently before the block that SSD firmwares first have to find which die0 is power down
The block of write-in, then on page 0, just deposit a startup related data management table, but the table is not necessarily newest, Ke Neng
After write-in starts related data management table, there is the LBA of record renewal again, can be before power down most for normal powering down
The PPA positions of the latter write-in, then a position for starting related data management table, so finding last last page is write,
Startup related data management table is just directly read, as long as shown in Fig. 2 last that part newest table copy is read out, just
There is complete startup related data management table.
Recover for the upper electricity of improper power down, that is to say, that may not have on the position being finally written into that upper electricity is found
Have and start related data management table, in such cases, it is necessary to first by page 0 is deposited on newest 0 piece of the die being written into a table
Read, patch installing is into L2P tables, then by way of scanning is ultimately written the metadata of block, to update L2P tables, thus
There is mapping table of a complete packet containing LBA in all startup related data management tables, it is possible to carry out follow-up action.Fig. 3
Middle background for dot matrix rectangle inframe data be exactly under improper power-down conditions, it is necessary to scan recover start dependency number
According to the region of management table.
Those skilled in the art can carry out various modifications and variations to the embodiment of the present invention, if these modifications and change
Type is within the scope of the claims in the present invention and its equivalent technologies, then these modifications and variations are also in protection scope of the present invention
Within.
The prior art that the content not being described in detail in specification is known to the skilled person.
Claims (10)
1. a kind of method that solid state hard disc as system disk shortens power-on time, it is characterised in that comprise the following steps：
Starting record in related data management table has the LBA entry that each power up is read, and each SSD firmwares detect
When electric, related data management table will be started and be loaded into SSD internal memories, by the LBA entry started in related data management table successively
L2P tables are updated, until L2P is rebuild successfully；
In power up, if the LBA entry in Host read hit L2P tables, directly in response to Host read；If
No, then L2P is waited to be responded after rebuilding successfully.
2. the method that the solid state hard disc as claimed in claim 1 as system disk shortens power-on time, it is characterised in that：The party
Method is further comprising the steps of：
In power up, if the number that Host read is hit reaches preset times, SSD firmwares are cumulative to start dependency number
According to the reading times in management table.
3. the method that the solid state hard disc as claimed in claim 1 as system disk shortens power-on time, it is characterised in that：The party
Method is further comprising the steps of：
In power up, if some LBA entry not start related data management table in, by the LBA that reading times are minimum
List item replaces successively.
4. the method that the solid state hard disc as claimed in claim 3 as system disk shortens power-on time, it is characterised in that：It is described
It is by the detailed process that the minimum LBA entry of reading times replaces successively：In power up, if the reading of each LBA entry
It is preset value to take number maximum, and is signed number, then hits every time；If reading times are less than preset value, then
Reading times add up；If reading times are equal to preset value, maintain preset value constant；If LBA entry is not hit, read
Number is taken to subtract one.
5. the method that the solid state hard disc as claimed in claim 1 as system disk shortens power-on time, it is characterised in that：The party
Method is further comprising the steps of：
Reading order according to related data management table is started establishes a queue to sort：
Reading order is smaller, show it is more early read, start related data management table initialization after, each LBA entry of the inside
L2P tables are loaded into, L2P table others list items are then initialized to represent the set form of illegal value；SSD firmwares receive main frame
After read requests, L2P tables are searched, if LBA entry is not illegal value, then it represents that the LBA entry is to start related data management
Table hit, then according to current reading order, as index, go to search the order to have sorted, find the LBA entry；
Reading times are negative, show that LBA entry can be replaced, and during initialization, the reading times for finding which LBA entry are
Negative, and recorded, when there is Host read not hit startup related data management table, then by these LBA entries
Replaced according to the order of reading times；By reading times, will be lined up for the LBA entry of negative according to order from small to large
One queue, replace all replaced since afterbody every time, and update reading order and reading times therein, PPA domains；
Finally, when it is all it is upper it is electronic complete, L2P has also recovered completely, then will start related data management table and press again
Sorted according to reading order, and be stored in flash memory, in case next time uses.
6. the system that a kind of solid state hard disc as system disk shortens power-on time, it is characterised in that the system includes L2P table weights
Unit, response unit are built, wherein：
L2P table reconstruction units are used for：Starting record in related data management table has the LBA entry that each power up is read,
When each SSD firmwares detect electric, related data management table will be started and be loaded into SSD internal memories, related data pipe will be started
LBA entry in reason table updates L2P tables successively, until L2P tables are rebuild successfully；
Response unit is used for：In power up, if the LBA entry in Host read hit L2P tables, directly in response to main frame
Read requests；If it is not, L2P is waited to be responded after rebuilding successfully.
7. the system that the solid state hard disc as claimed in claim 6 as system disk shortens power-on time, it is characterised in that：This is
System also includes reading times summing elements, is used for：In power up, if the number that Host read is hit reaches default
Number, the cumulative reading times started in related data management table of SSD firmwares.
8. the system that the solid state hard disc as claimed in claim 6 as system disk shortens power-on time, it is characterised in that：This is
System also includes LBA entry replacement unit, is used for：In power up, if some LBA entry is not starting related data management table
In, then the minimum LBA entry of reading times is replaced successively.
9. the system that the solid state hard disc as claimed in claim 8 as system disk shortens power-on time, it is characterised in that：It is described
It is by the detailed process that the minimum LBA entry of reading times replaces successively：In power up, if the reading of each LBA entry
It is preset value to take number maximum, and is signed number, then hits every time；If reading times are less than preset value, then
Reading times add up；If reading times are equal to preset value, remain unchanged；If LBA entry is not hit, reading times
Subtract one.
10. the system that the solid state hard disc as claimed in claim 6 as system disk shortens power-on time, it is characterised in that：Should
System also establishes unit including queue, is used for：Reading order according to related data management table is started establishes a team to sort
Row：
Reading order is smaller, show it is more early read, start related data management table initialization after, each LBA entry of the inside
L2P tables are loaded into, L2P table others list items are then initialized to represent the set form of illegal value；SSD firmwares receive main frame
After read requests, L2P tables are searched, if LBA entry is not illegal value, then it represents that the LBA entry is to start related data management
Table hit, then according to current reading order, as index, go to search the order to have sorted, find the LBA entry；
Reading times are negative, show that LBA entry can be replaced, and during initialization, the reading times for finding which LBA entry are
Negative, and recorded, when there is Host read not hit startup related data management table, then by these LBA entries
Replaced according to the order of reading times；By reading times, will be lined up for the LBA entry of negative according to order from small to large
One queue, replace all replaced since afterbody every time, and update reading order and reading times therein, PPA domains；
Finally, when it is all it is upper it is electronic complete, L2P has also recovered completely, then will start related data management table and press again
Sorted according to reading order, and be stored in flash memory, in case next time uses.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710833652.7A CN107656875B (en) | 2017-09-15 | 2017-09-15 | Method and system for shortening power-on time of solid state disk serving as system disk |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710833652.7A CN107656875B (en) | 2017-09-15 | 2017-09-15 | Method and system for shortening power-on time of solid state disk serving as system disk |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107656875A true CN107656875A (en) | 2018-02-02 |
| CN107656875B CN107656875B (en) | 2020-05-15 |
Family
ID=61129515
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710833652.7A Active CN107656875B (en) | 2017-09-15 | 2017-09-15 | Method and system for shortening power-on time of solid state disk serving as system disk |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107656875B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108710507A (en) * | 2018-02-11 | 2018-10-26 | 深圳忆联信息系统有限公司 | A kind of method of SSD master dormants optimization |
| CN108762674A (en) * | 2018-05-24 | 2018-11-06 | 深圳忆联信息系统有限公司 | Promote the method and device of SSD operating lags |
| CN108804032A (en) * | 2018-05-16 | 2018-11-13 | 山东华芯半导体有限公司 | A kind of adaptive abrasion equilibrium garbage reclamation accelerator and method |
| CN108804026A (en) * | 2018-03-27 | 2018-11-13 | 深圳忆联信息系统有限公司 | A kind of solid state disk scan full hard disk method and solid state disk |
| CN108983935A (en) * | 2018-07-05 | 2018-12-11 | 深圳忆联信息系统有限公司 | A kind of control method and its system of solid state hard disk power-on time |
| CN109213443A (en) * | 2018-08-16 | 2019-01-15 | 北京盛和大地数据科技有限公司 | The method and apparatus that CD is called in advance |
| CN110647295A (en) * | 2019-09-12 | 2020-01-03 | 苏州浪潮智能科技有限公司 | Method, system, device and storage medium for shortening SSD power-on recovery time |
| CN111104345A (en) * | 2019-12-12 | 2020-05-05 | 苏州浪潮智能科技有限公司 | SSD power-on recovery method, system and host |
| CN112052193A (en) * | 2020-09-28 | 2020-12-08 | 成都佰维存储科技有限公司 | Garbage recycling method and device, readable storage medium and electronic equipment |
| CN113535086A (en) * | 2021-07-12 | 2021-10-22 | 成都信息工程大学 | Method for accelerating reconstruction in solid state disk |
| WO2021232427A1 (en) * | 2020-05-22 | 2021-11-25 | Yangtze Memory Technologies Co., Ltd. | Flush method for mapping table of ssd |
| CN114265549A (en) * | 2021-11-21 | 2022-04-01 | 山东云海国创云计算装备产业创新中心有限公司 | NVMe command processing method and device and readable storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101930404A (en) * | 2010-08-27 | 2010-12-29 | 威盛电子股份有限公司 | Memory device and operation method thereof |
| US20130091322A1 (en) * | 2011-10-06 | 2013-04-11 | Mstar Semiconductor, Inc. | Electronic System and Memory Managing Method Thereof |
| US20140089630A1 (en) * | 2012-09-24 | 2014-03-27 | Violin Memory Inc | Virtual addressing |
| CN103970683A (en) * | 2014-05-07 | 2014-08-06 | 山东华芯半导体有限公司 | Method for accelerating starting of USB flash disk |
| CN104040483A (en) * | 2011-12-27 | 2014-09-10 | 英特尔公司 | Optimized cold boot for non-volatile memory |
| CN105740157A (en) * | 2014-11-03 | 2016-07-06 | 慧荣科技股份有限公司 | Data storage device and flash memory control method |
| CN106383792A (en) * | 2016-09-20 | 2017-02-08 | 北京工业大学 | Missing perception-based heterogeneous multi-core cache replacement method |
-
2017
- 2017-09-15 CN CN201710833652.7A patent/CN107656875B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101930404A (en) * | 2010-08-27 | 2010-12-29 | 威盛电子股份有限公司 | Memory device and operation method thereof |
| US20130091322A1 (en) * | 2011-10-06 | 2013-04-11 | Mstar Semiconductor, Inc. | Electronic System and Memory Managing Method Thereof |
| CN104040483A (en) * | 2011-12-27 | 2014-09-10 | 英特尔公司 | Optimized cold boot for non-volatile memory |
| US20140089630A1 (en) * | 2012-09-24 | 2014-03-27 | Violin Memory Inc | Virtual addressing |
| CN103970683A (en) * | 2014-05-07 | 2014-08-06 | 山东华芯半导体有限公司 | Method for accelerating starting of USB flash disk |
| CN105740157A (en) * | 2014-11-03 | 2016-07-06 | 慧荣科技股份有限公司 | Data storage device and flash memory control method |
| CN106383792A (en) * | 2016-09-20 | 2017-02-08 | 北京工业大学 | Missing perception-based heterogeneous multi-core cache replacement method |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108710507A (en) * | 2018-02-11 | 2018-10-26 | 深圳忆联信息系统有限公司 | A kind of method of SSD master dormants optimization |
| CN108710507B (en) * | 2018-02-11 | 2021-01-12 | 深圳忆联信息系统有限公司 | SSD host dormancy optimization method |
| CN108804026A (en) * | 2018-03-27 | 2018-11-13 | 深圳忆联信息系统有限公司 | A kind of solid state disk scan full hard disk method and solid state disk |
| CN108804026B (en) * | 2018-03-27 | 2022-02-18 | 深圳忆联信息系统有限公司 | Solid state disk full-disk scanning method and solid state disk |
| CN108804032A (en) * | 2018-05-16 | 2018-11-13 | 山东华芯半导体有限公司 | A kind of adaptive abrasion equilibrium garbage reclamation accelerator and method |
| CN108804032B (en) * | 2018-05-16 | 2021-05-18 | 山东华芯半导体有限公司 | Self-adaptive wear-balanced garbage recovery accelerating device and method |
| CN108762674A (en) * | 2018-05-24 | 2018-11-06 | 深圳忆联信息系统有限公司 | Promote the method and device of SSD operating lags |
| CN108983935A (en) * | 2018-07-05 | 2018-12-11 | 深圳忆联信息系统有限公司 | A kind of control method and its system of solid state hard disk power-on time |
| CN109213443A (en) * | 2018-08-16 | 2019-01-15 | 北京盛和大地数据科技有限公司 | The method and apparatus that CD is called in advance |
| CN110647295A (en) * | 2019-09-12 | 2020-01-03 | 苏州浪潮智能科技有限公司 | Method, system, device and storage medium for shortening SSD power-on recovery time |
| CN111104345A (en) * | 2019-12-12 | 2020-05-05 | 苏州浪潮智能科技有限公司 | SSD power-on recovery method, system and host |
| CN111104345B (en) * | 2019-12-12 | 2022-11-08 | 苏州浪潮智能科技有限公司 | SSD power-on recovery method, system and host |
| WO2021232427A1 (en) * | 2020-05-22 | 2021-11-25 | Yangtze Memory Technologies Co., Ltd. | Flush method for mapping table of ssd |
| CN112052193A (en) * | 2020-09-28 | 2020-12-08 | 成都佰维存储科技有限公司 | Garbage recycling method and device, readable storage medium and electronic equipment |
| CN113535086A (en) * | 2021-07-12 | 2021-10-22 | 成都信息工程大学 | Method for accelerating reconstruction in solid state disk |
| CN113535086B (en) * | 2021-07-12 | 2023-07-25 | 中飞赛维智能科技股份有限公司 | Acceleration method for reconstruction in solid state disk |
| CN114265549A (en) * | 2021-11-21 | 2022-04-01 | 山东云海国创云计算装备产业创新中心有限公司 | NVMe command processing method and device and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107656875B (en) | 2020-05-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107656875A (en) | Solid state hard disc as system disk shortens the method and system of power-on time | |
| US9164887B2 (en) | Power-failure recovery device and method for flash memory | |
| US8595451B2 (en) | Managing a storage cache utilizing externally assigned cache priority tags | |
| US8291155B2 (en) | Data access method, memory controller and memory storage system | |
| JP5162535B2 (en) | Method and memory system using memory system | |
| US8694984B2 (en) | Memory apparatus and method of updating firmware of the memory apparatus | |
| CN103049349B (en) | Based on Snapshot Method and the system of Bedding storage | |
| US8914608B2 (en) | Data storage device and computing system including the same | |
| US10198318B2 (en) | Storage apparatus having nonvolatile memory device, and nonvolatile memory device | |
| US11126561B2 (en) | Method and system for organizing NAND blocks and placing data to facilitate high-throughput for random writes in a solid state drive | |
| CN109144899B (en) | Method for managing table recovery | |
| CN105243025A (en) | Mapping table forming and loading methods and electronic device | |
| CN103270500A (en) | Transaction log recovery | |
| CN101625897B (en) | Data write-in method, storage system and controller used for quick flash memory | |
| US10552335B2 (en) | Method and electronic device for a mapping table in a solid-state memory | |
| US11119657B2 (en) | Dynamic access in flash system | |
| CN102200946B (en) | Data access method, memory controller and storage system | |
| CN116088770B (en) | Data management method, device, system, electronic equipment and storage medium | |
| CN104915265B (en) | Data back up method and electronic equipment | |
| CN104516792A (en) | Data backup method and device | |
| CN101727397B (en) | Block management and change method, flash memory storage system and controller thereof | |
| KR100987320B1 (en) | Data processing apparatus and Data procssing method, using FAT file system capable of fast file recovery | |
| CN107451069A (en) | The rubbish recovering method of Solid-state disc array based on headspace | |
| CN111506255B (en) | NVM-based solid-state hard disk metadata management method and system | |
| WO2010145967A1 (en) | Memory device for managing the recovery of a non volatile memory |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 430070 Hubei city of Wuhan province Wuhan East Lake New Technology Development Zone Road No. two high Guan Nan Industrial Park No. 2 building 2-3 floor West Applicant after: EXASCEND TECHNOLOGY (WUHAN) CO.,LTD. Address before: 430223 Hubei city of Wuhan province East Lake New Technology Development Zone Road No. two high Guan Nan Industrial Park No. 2 building 2-3 floor West Applicant before: MEMORIGHT (WUHAN) Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Method and system for shortening power on time of solid state hard disk as system disk Effective date of registration: 20210629 Granted publication date: 20200515 Pledgee: Guanggu Branch of Wuhan Rural Commercial Bank Co.,Ltd. Pledgor: EXASCEND TECHNOLOGY (WUHAN) Co.,Ltd. Registration number: Y2021420000059 |
|
| PC01 | Cancellation of the registration of the contract for pledge of patent right | ||
| PC01 | Cancellation of the registration of the contract for pledge of patent right |
Date of cancellation: 20221031 Granted publication date: 20200515 Pledgee: Guanggu Branch of Wuhan Rural Commercial Bank Co.,Ltd. Pledgor: EXASCEND TECHNOLOGY (WUHAN) CO.,LTD. Registration number: Y2021420000059 |
|
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 430000 west of 2-3 / F, No.2 factory building, Guannan Industrial Park, No.1 Gaoxin 2nd Road, Wuhan Donghu New Technology Development Zone, Wuhan City, Hubei Province Patentee after: Zhiyu Technology Co.,Ltd. Address before: 430070 Wuhan, Hubei Wuhan East Lake New Technology Development Zone, high-tech two Road No. 1 South Guan Industrial Park 2 factory 2-3 floor West. Patentee before: EXASCEND TECHNOLOGY (WUHAN) CO.,LTD. |