Patent Application
20220206905
The present disclosure relates generally to memory systems, and more particularly, to enhanced performance of memory systems such as hybrid memory systems having both volatile and non-volatile memory.
Hybrid memory systems are memory systems having a mixture of volatile and non-volatile memory types. Hybrid memory systems utilize non-volatile memory components to securely store volatile system data in the event of system fault or power failure, or upon user requests which may in include certain write or programming operations. Typically, the non-volatile memory used in a hybrid memory system is flash memory. The non-volatile memory components are made up of delimited memory portions, for example blocks, whose lifetimes are limited. Beyond these lifetimes, the memory portions can no longer be reliably used to store data, and any valid data present in them at that time may not be reliably accessed or recovered. Such reduced reliability of memory systems resulting in loss of data can be catastrophic to overall computer system performance or operation.
One type of non-volatile flash memory used in hybrid memory systems is NAND flash. NAND flash devices are available from several vendors and all share a similar architecture. Vendors of flash memory include Samsung™, Micron™, Hynix™, and Toshiba™.
With reference to
The NAND flash device is programmed/written in units of pages. The NAND flash device is erased in units of blocks. If an uncorrectable error occurs in any page of a given block, the entire block is marked invalid.
Correctable and uncorrectable errors can be detected by the use of an error detection and correction algorithm. To insure data integrity, the data in the NAND flash device is typically protected by an error detection and correction algorithm.
One common approach to non-volatile memory management by memory systems, such as hybrid memory systems, is to mark blocks of memory invalid when uncorrectable errors are encountered under various circumstances. In particular, depending on the phase of operation—whether an error recovery can be performed—the error may or may not result in the loss of system data. Once an uncorrectable error occurs, a block is marked invalid.
With reference to
If the number of errors in a segment of (n) symbols exceeds the number that can be corrected (t), the error is uncorrectable, and the original data cannot be recovered.
There is a need for increasing the ability to control the reliability of non-volatile memory systems by determining or detecting when an uncorrectable error may occur and thus provide an early warning about the reliability of a portion of the non-volatile memory, for example a block or a page. Furthermore, there is a need to determine when there are too few non-volatile memory blocks to backup all of the specified or required data. Moreover, there is a need to provide the computer system or the end user a programmable capability to configure and customize certain thresholds for various parameters based on the desired reliability for the overall computer system, a particular application usage, or user level of risk tolerance for data errors.
The accompanying drawings, which are incorporated into and constitute a part of this specification, illustrate one or more examples of embodiments and, together with the description of example embodiments, serve to explain the principles and implementations of the embodiments.
In the drawings:
Example embodiments are described herein in the context of a memory system. Those of ordinary skill in the art will realize that the following description is illustrative only and is not intended to be in any way limiting. Other embodiments will readily suggest themselves to such skilled persons having the benefit of this disclosure. Reference will now be made in detail to implementations of the example embodiments as illustrated in the accompanying drawings. The same reference indicators will be used to the extent possible throughout the drawings and the following description to refer to the same or like items.
In the interest of clarity, not all of the routine features of the implementations described herein are shown and described. It will, of course, be appreciated that in the development of any such actual implementation, numerous implementation-specific decisions must be made in order to achieve the developer's specific goals, such as compliance with application- and business-related constraints, and that these specific goals will vary from one implementation to another and from one developer to another. Moreover, it will be appreciated that such a development effort might be complex and time-consuming, but would nevertheless be a routine undertaking of engineering for those of ordinary skill in the art having the benefit of this disclosure.
In accordance with this disclosure, the components, process steps, and/or data structures described herein may be implemented using various types of operating systems, computing platforms, computer programs, and/or general purpose machines. In addition, those of ordinary skill in the art will recognize that devices of a less general purpose nature, such as hardwired devices, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), or the like, may also be used without departing from the scope and spirit of the inventive concepts disclosed herein. Where a method comprising a series of process steps is implemented by a computer or a machine and those process steps can be stored as a series of instructions readable by the machine, they may be stored on a tangible medium such as a computer memory device (e.g., ROM (Read Only Memory), PROM (Programmable Read Only Memory), EEPROM (Electrically Eraseable Programmable Read Only Memory), FLASH Memory, Jump Drive, and the like), magnetic storage medium (e.g., tape, magnetic disk drive, and the like), optical storage medium (e.g., CD-ROM, DVD-ROM, paper card, paper tape and the like) and other types of program memory.
The term “exemplary” when used herein is merely intended to mean “serving as an example, instance or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.
Described herein are methods and systems for managing and controlling the reliability of the non-volatile memory components of a hybrid memory system (HMS) based on user criteria, application based performance, or other parameters that can be fixed or programmable. In certain embodiments a mechanism for the computer system or an end user to manage and control when a portion of the non-volatile memory system (for example a block) is declared invalid and can no longer be a part of the non-volatile memory subsystem used to store data. In certain embodiments, there is provided the capability for setting one or more thresholds or parameters for the number of invalid blocks that can be reached before the hybrid memory system will issue a warning that a portion of (or all of) the hybrid memory system will be declared unusable. In this manner the end user can manage and control how a block, or any selected portion of the hybrid memory system, or all of the hybrid memory system, is removed from service in accordance with end user or computer system programmable preferences or specifications.
In certain embodiments, HMS 300 comprises a memory module, and may have a printed-circuit board (PCB) form factor. In certain embodiments, HMS 300 has a volatile memory capacity of 512-MB, I-GB, 2-GB, 4-GB, 8-GB, 16-GB, 32-GB, and so on. Other volatile memory capacities are also compatible with certain embodiments described herein. In certain embodiments, HMS 300 has a volatile memory subsystem that is integrated within the circuit 308 such as using internal memory buffer to emulate the volatile memory subsystem 302, and can have various capacities of 16MB, 32MB, 64MB, 128MB, 256MB, 512-MB, I-GB, 2-GB and so on. In certain embodiments, HMS 300 has a volatile memory subsystem that is integrated within the HMS controller 306. In yet other embodiments, HMS 300 has a volatile memory subsystem and circuit 308 that are integrated within the HMS controller 306.
In certain embodiments, HMS 300 has a non-volatile memory capacity of 512-MB, 1-GB, 2-GB, 4-GB, 8-GB, 16-GB, 32-GB, 64-GB, 128-GB, 256-GB and so on, or any other number of memory capacity that is based on the overall system requirement. Other non-volatile memory capacities are also compatible with certain embodiments described herein. In addition, memory systems having widths of 4 bytes, 8 bytes, 16 bytes, 32 bytes, or 32 bits, 64 bits, 128 bits, 256 bits, as well as other widths (in bytes or in bits), are compatible with embodiments described herein. In certain embodiments, the PCB of HMS 300 has an industry-standard form factor such as specified by JEDEC standards. For example, the PCB can have a low profile (LP) form factor with a height of 30 millimeters and a width of 133.35 millimeters. In certain other embodiments, the PCB 20 has a very high profile (VHP) form factor with a height of 50 millimeters or more. In certain other embodiments, the PCB has a very low profile (VLP) form factor with a height of 18.3 millimeters. Other form factors including, but not limited to, small- outline (SO-DIMM), unbuffered (UDIMM), registered (RDIMM), fully-buffered (FBDIMM), miniDIMM, mini-RDIMM, VLP mini-DIMM, micro-DIMM, SRAM DIMM, and Ultra DIMM are also compatible with certain embodiments described herein. For example, in other embodiments, certain non-DIMM form factors are possible such as, for example, single in-line memory module (SIMM), multi-media card (MMC), and small computer system interface (SCSI).
In certain embodiments, HMS 300 is in electrical communication with the host system (not shown). In other embodiments, HMS 300 may communicate with a host system using some other type of communication, such as, for example, optical communication. Examples of host systems include, but are not limited to, blade servers, 1U servers, personal computers (PCs), and other applications in which the performance and reliability of the host is dependent on the performance and reliability of the non-volatile or volatile memory subsystems. In certain embodiments, HMS 300 is a memory module where the interface 310 is standard based DIMM interface as specified by JEDEC. In certain embodiments, HMS 300 is a hybrid memory module that appears to the computer system as a registered memory module (RDIMM). HMS 300 can be in communication with a disk controller of a computer system, for example. An interface 310 that is configured to be in electrical communication with the host system may be provided. For example, the interface 310 can comprise a plurality of edge connections which fit into a corresponding slot connector of the host system and/or be in full compliance with standard based electrical specification such as specified by JEDEC standards. For example, the interface 310 can comprise a standard DDR (e.g. DDR3 or DDR4) DIMM, RDIMM, or LRDIMM edge connector. The interface 310 of certain embodiments provides a conduit for power voltage as well as data, address, and control signals between HMS 300 and the host system. In certain embodiments, HMS 300 include a power source 309 that controls and distributes power to the HMS controller 306, the volatile memory subsystem 302, the circuit 308, and the non-volatile memory subsystem 304. The power source 309 may include energy storage devices, such as super capacitors, to be charged during normal operation drawing power from the host, while when the HMS 300 functions autonomously and independently from the host the power source 309 controls the power distribution to various subsystems or components of HMS 300. Certain functionality of power source 309 can be controlled or programmed by the host or the user in accordance with certain reliability or performance criteria. The power source 309 controls and performs as main power management controller and includes various subsystems such as DC-to-DC conversion in order to accommodate different portions of the HMS 300 that require different power (i.e. voltage and current) requirement.
The volatile memory subsystem 302 comprises one or more volatile memory elements 303 and the non-volatile memory subsystem 304 comprises one or more non-volatile memory elements 305. While the memory subsystems 302 and 304 are shown as physically comprised of discrete memory elements or devices 303, 305, respectively, it will be appreciated that logically, other delimitations and groupings are contemplated. Non-volatile memory subsystem 302 for instance is also comprised of memory portions that can be grouped, in increasing size order, into segments, pages, blocks, and slices of memory. An example of segment size is 242 bytes. An example of an ECC (error correction code) segment (explained below) size is 13 bytes. An example of a page size is 8,192 bytes. An example of a block size is 128 pages, or 1,048,576 bytes. Certain embodiments described herein advantageously provide non-volatile storage via the non-volatile memory subsystem 304 in addition to high-performance (e.g., high speed) storage via the volatile memory subsystem 302. In certain embodiments, the first plurality of volatile memory elements 303 comprises two or more dynamic random-access memory (DRAM) elements. Types of DRAM elements 303 compatible with certain embodiments described herein include, but are not limited to, DDR, DDR2, DDR3, and synchronous DRAM (SDRAM). The volatile memory elements 303 may comprise other types of memory elements such as static random-access memory (SRAM). In addition, volatile memory elements 303 having bit widths of 4, 8, 16, 32, as well as other bit widths, are compatible with certain embodiments described herein. Volatile memory elements 303 compatible with certain embodiments described herein have packaging which includes, but is not limited to, thin small-outline package (TSOP), ball-grid-array (BGA), fine-pitch BGA (FBGA), micro-BGA (1.1,BGA), mini-BGA (mBGA), and chip-scale packaging (CSP).
100231 In certain embodiments, the plurality of non-volatile memory elements 305 comprises one or more flash memory elements. Types of flash memory elements 305 compatible with certain embodiments described herein include, but are not limited to, NOR flash, NAND flash, ONE-NAND flash, and multi-level cell (MLC). In addition, non-volatile memory elements 305 having bit widths of 4, 8, 16, 32, as well as other bit widths, are compatible with certain embodiments described herein. Non-volatile memory elements 305 compatible with certain embodiments described herein have packaging which include, but are not limited to, thin small-outline package (TSOP), ball-grid-array (BGA), fine-pitch BGA (FBGA), micro-BGA (PGA), mini-BGA (mBGA), and chip-scale packaging (CSP).
In terms of storage capacity, in order to account for invalid blocks, HMS 300 preferably contains more non-volatile memory 304 than the amount of volatile memory 302 that it needs to protect. For example, if there is (S) amount of volatile memory within the HMS, then S×F non-volatile memory is provided, wherein F is an over-provisioning factor based on system design requirements. In certain embodiments, HMS 300 contains double, triple or any multiple (integer or non-integer) amount of non-volatile memory to volatile memory. In certain embodiments, the over-provisioning factor F can be hundreds multiples, as an example the non-volatile memory capacity is 500GB while the volatile memory capacity is 1GB, 2GB, or 4GB.
IPT 408 is used by NVM controller to keep track of invalid non-volatile memory portions. These portions may be identified and managed in increments of segments, pages, blocks, or memory slices, as detailed below. Invalid designations of the memory may occur for any of various reasons, and may be functions of programmable or configurable error criteria and thresholds as detailed below. In one example, and not by way of limitation, NVM controller 406 may designate a block of non-volatile memory 304 invalid due to a manufacturing error in the block that prevents access to that block or portions thereof. In another example, operational errors, such as errors encountered during a read or write operation, can lead to declaring memory portions associated with the errors as invalid. In certain embodiments, memory portions are declared invalid when data errors are encountered, or when command response times associated with the portions exceed specified time out limits (“time out errors”). For example, if a read, program, or erase command does not complete within a predetermined length of time, then the corresponding block or blocks may be designated as invalid. In certain embodiments, during a memory program operation, if an error is detected, the portion of memory being programmed is marked invalid in IPT 408 and the data is written to the next valid block. In certain embodiments, during a read operation, if an error is detected, the memory portion, such as segment, page, block, or slice, being read is marked invalid in IPT 408.
Communication controller 402 is operable to receive configuration information from the user or the host system. Communication controller 402 is also operable to deliver HMS 300 status information to the user or host system. This two-way communication is by a messaging protocol conducted through communication interface 410. The messaging protocol is transmitted in certain embodiments over an electrical interface under the management of the communication controller 402.
The configuration information received by communication controller 402 through the messaging protocol and communication interface 410 is used for reading or writing the values of register set 404, which includes threshold registers 412, error registers 414, status registers 416, and control registers 418.
In certain embodiments, values written in threshold registers 412 define the parameters for declaring portions of non-volatile memory 304 invalid. With the ability to set the values in the threshold registers 412, the user or system can control and manage the amount of invalid non-volatile memory that will be tolerated, and the spectrum of actions and remediation measures that can be undertaken with progressive deterioration of memory health. In certain example embodiments, the user or system or application gains the ability to manage and control the amount of invalid non-volatile memory that will be tolerated before declaring that HMS 300, or portions thereof, has or is about to fail or become unreliable or unstable. In certain embodiments, the threshold registers 412 define threshold parameters that are configurable using default values, or are written into the registers, by way of the messaging protocol via communication interface 410 or by way of detection and learning by the NVM controller 406 over certain time period.
The set of threshold parameters can be thresholds that are at the memory segment level, page level, block level, slice level, system level or subsystem level. With reference to Table 1, at each level, three candidate thresholds are possible: a failure threshold (FT), an error threshold (ET), and a warning threshold (WT). The first column in Table 1 relates to the SET threshold, which is a configurable error criterion, detailed further below, that can relate to the number of errors that can occur before a portion of memory, such as a segment, page or block, is declared invalid. The remaining three columns in Table 1 relate to the number of invalid memory portions (segments, pages, or blocks) that can be tolerated before a subsuming memory portion (a page, block or slice) is declared invalid.
Considering the example of a block of the non-volatile memory 304, the number of invalid pages that will be tolerated before declaring a block invalid, or NVM controller 406 issuing warnings to that effect, e.g. to the host, is configurable by setting the thresholds WTB, ETB, and FTB in registers 412 using the communication protocol. In particular, the block failure threshold FTB indicates the maximum number of invalid pages per block that will be tolerated before a block is declared invalid. Thus a block can be considered invalid if more than FTB invalid pages are contained therein. Further, a page in such a determination can be considered invalid based on a configurable error criterion (SET column in Table 1), for example based on the occurrence of a user- or system-settable threshold number of errors per that page, or errors per segment within that page. The configurable error criterion is discussed further below. The block error threshold, ETB, indicates the maximum number of invalid pages for a block to operate reliably in accordance with a specified application or user preference, where ETB:S FTB•If ETB is exceeded, then an error is reported to the system and the system or user may choose that that block be declared invalid or to continue to use the block until the FTB threshold is reached. The block warning threshold WTB, defines the number of invalid pages above which a warning will be reported to the system, with WTB:S ETB:S FTB. The WTB is used to provide the end user with an early indication that the number of invalid pages per block has reached a critical level, e.g. the programmable threshold level, but that the HMS 300 may continue to operate in accordance with a specified application or user preference.
Considering the example of a slice of the non-volatile memory subsystem 304, or the non-volatile memory subsystem in its entirety, the number of invalid blocks therein that will be tolerated before declaring the slice, or the NVM subsystem in its entirety, invalid or no longer suitable for backup purposes, or for issuing warnings to that effect, is configurable by setting the thresholds WTs, ETs, and FTs in registers 412 using the communication protocol. In particular, the slice failure threshold FTs indicates the maximum number of invalid blocks per slice (or entire NVM subsystem) that will be tolerated before declaring the slice, or entire NVM subsystem, unsuitable or unreliable for data storage purposes. Thus a slice can be considered invalid if more than FTs invalid blocks are encountered therein. A block, in tum, can be considered invalid based on a configurable error criterion (SET column in Table 1), for example based on the occurrence of a user- or system-settable threshold number of errors per that block, or, as described above, based on a threshold number of invalid pages per that block.
The slice error threshold, ETs, indicates the maximum number of invalid blocks for a slice to operate reliably in accordance with a specified application or user preference, where ETs≤FTs. If ETs is exceeded, then an error is reported to the system and the system or user may choose that that slice be declared invalid or to continue using that slice until the FTs threshold is reached. The slice warning threshold WTs defines the number of invalid blocks above which a warning will be reported to the system, with WTs≤ETs≤FTs. The WTs is used to provide the end user with an early indication that the number of invalid blocks has reached a critical level, but that the HMS 300 may continue to operate in accordance with a specified application or user preference. Thus if the slice is taken to be large enough relative to the data that is to be stored, the thresholds WTs, ETs, and FTs can signify the points at which reliability of the slice has fallen below the programmed threshold, storing data in the slice will no longer be possible and can produce errors that may or may not be correctable, and failure threshold for the slice has been reached and data storage should no longer be attempted to the slice, respectively. Thus, these early indications can be issued to the user or system by the NVM controller 406, and/or are used by the NVM controller 406 to effectively control and manage the memory space in-use that meets the default system or user specified criteria for storing data in the non-volatile memory 304.
Considering the example of a page of the non-volatile memory 304, thresholds WTP, ETp, and FTp are values, written into threshold registers 412 using the communication protocol, that relate to and define the number of invalid segments per page that can be tolerated. The page failure threshold FTp indicates the maximum number of invalid segments tolerable before a page is declared unusable or invalid. A page (or block or a slice) can be considered invalid based on a configurable error criterion (SET column in Table 1), for example based on the occurrence of a user- or system-settable threshold number of errors per that page. The page error threshold, ETP, indicates the maximum number of invalid segments for a page to operate reliably in accordance with a specified application or user preference, where ETp≤FTp. If ETp is exceeded, then an error is reported to the system and the user or system may choose that the application be terminated or to continue to operate until the threshold FTp is reached. The page warning threshold WTP, defines the number of invalid segments above which a warning will be reported to the system, with WTp≤ETp≤FTP. The WTp is used to provide the end user with an early indication that the number of invalid segments has reached a critical level.
Referring to
Control register 418 stores control information, such as the manner in which data is exchanged between the non-volatile memory 304 and the host or the volatile memory 302, e.g. data write or read from/to non-volatile memory 304, a backup operation initiated, the extent of the backup necessary, and so on.
In certain embodiments, the thresholds in register 418 can be specified by the number of valid non-volatile memory portions with corresponding minimum failure, error, and warning thresholds. Furthermore, the thresholds can also be specified as a percentage value of the overall capacity of the hybrid memory system (HBS) 300, or any other means to indicate a user or system preference.
It will be recalled that, in general, certain error detection and correction algorithms have the following properties:
If the number of errors in a segment of (n) symbols exceeds the number that can be corrected (t), the error is uncorrectable, and the original data cannot be recovered.
As explained above, in certain embodiments, an error criterion can be configurable by the user or system (see first column of Table 1). This error criterion can for example relate to the number of correctable errors that can occur before a portion of memory is declared invalid. In this manner, for example, a symbol error threshold, SET (420), relating to an acceptable number of symbol errors per segment, per page, per block, or per slice, of NVM, can be set. In certain embodiments, SET is set to a value between one (1) and the maximum number of symbol errors that can be detected (2t) or corrected (t). When the number of symbol errors detected in a segment, page, block, or slice, that segment, page, block or slice is declared invalid. At the block level, therefore, and by way of example only, by configuring the symbol error threshold SET, the end user or system can control how many symbol errors are required to mark a block invalid. This ability of a user to control the performance of the HMS 300 reduces the probability of data loss by marking blocks invalid before enough symbol errors have accumulated to create an uncorrectable error. Thus, improving the reliability and performance of the overall computer system utilizing the HMS 300.
There are many possible sets of warning error and failure thresholds based on the number of detected errors. Threshold parameters can relate to: the number of symbol errors per segment that can be tolerated before warning or declaring that a segment or page is unstable or unusable; the number of symbol errors per segment or page that can be tolerated before warning or declaring that a block is unstable or unusable; the number of symbol errors per segment or page or block that can be tolerated before warning or declaring that a non-volatile memory system, a slice of the non-volatile memory system, or hybrid memory system 300 as a whole is unstable or unusable for data storage or backup purposes. Specifically, the thresholds can include: Total number of symbol errors per segment, with warnings and errors displayed on a per segment basis; total number of segments in a page having more than the threshold of symbol errors, with warnings and errors displayed on a per page basis; total number of pages in a block with segments having more than the threshold of symbol errors, with warnings and errors displayed on a per block basis; and total number of blocks having more than the threshold of pages with (E) symbol errors where (E) is also a configurable threshold, with warnings and errors displayed on a per block basis or per HMS system or a slice of HMS system. Warnings and errors may also be triggered by combinations of multiple thresholds levels.
The thresholds may be stored in their associated registers within the register set 404 or may be stored in a preset memory space within the HMS 300 or within the system comprising the HMS 300.
In addition to the communication interface, the HMS 300 may also display error and warning status visually through the use of LEDs or other display mechanisms, e.g. LCD where information is displayed text or graphics. For example, the indications may be displayed as follows:
In the event of an HMS 300 failure, these visual indications make it easier for service personnel to identify a faulty HMS within a host system comprising more than one HMS system, or to identify when a particular warning threshold has been reached.
At the system level, by configuring the failure threshold (FT), error threshold (ET), and warning threshold (WT), the end user can provide a guard band around the number of invalid blocks and thus enhance the overall system reliability and reduce the probability of a catastrophic data loss due to HMS 300 error. The user can tailor or adjust various levels of acceptable tolerance for warning, error, and failure threshold based on the overall computer system implementation, e.g. use in a mission-critical operation. Additionally, the end user can receive a preliminary indication about the health of the HMS or that a certain parameter has reached a critical level.
By configuring the symbol error threshold SET, the end user can manage and control how many symbol errors are required before marking a memory portion such as a segment, page or block invalid and thereby improving the reliability of the overall system and reducing the probability of data loss occurring due HMS errors.
Furthermore, various logical operations that can be performed using these programmable thresholds and the results can be provided as stats or control bits to an application or as part of system diagnostics and health check. Moreover, certain procedures such as multiple read or write operations may be continuously run for a specified number of cycles or time, such as to confirm or to check the number of detected errors from a specific segment, page, block or slice that are produced within a given number of operations.
It will be appreciated that while described in the context of a hybrid memory system that includes both volatile and non-volatile memory, the memory monitoring and diagnostic procedures described herein are applicable to any memory system in which progressive failures in portions of the memory occur whether due to deterioration of system level parameters such as power supply, voltage, current, temperature, or due to deterioration of the memory cells of the non-volatile memory 304 through continuous use, for example read and write operations. Thus, thresholds for tolerating and remediating errors can be programmed into any such system, for example an exclusively non-volatile memory system, to provide control and flexibility over the amount of errors and portions of invalid memory that can be tolerated for a given computer system or intended application.
In accordance with certain embodiments, despite the above thresholds and control, backup and data storage failure can still occur. For instance, system backup or data storage can fail if sufficient non-volatile memory has been declared invalid, or if power loss interrupted the backup or data storage process, or internal code failed to properly execute. It may then be desirable to determine whether any significant data has been lost, and which data has been lost.
After the backup or data storage procedure is performed, backup or data storage flags indicating completion are updated, at 612, and a new pointer in NVM 304 is saved, at 614, to reflect how far the backup or data storage process went (actual end point). At 616, a determination is made of whether the new pointer matches up with the indication of the end point of the backup or data storage from 608. Match-up failure indicates that the backup or data storage was not successfully completed, at 618.
In certain embodiments, HMS controller 306 can conduct a backup or data storage failure analysis process following an incomplete write operation, for example to determine to the extent of the failed write operation. Since typically all memory blocks (or other portions) that make up the portions of NVM subsystem 304 that are intended to receive data must first be erased before data write can commence, remaining erased portions following backup can be used to indicate a failure of the backup to complete. Specifically, flash blocks in particular are sequential (skipping only those marked as invalid); therefore, the first block in the intended write region that is still erased, will point to where the backup fell short of completion. HMS controller 306 determines if and which blocks (or other portions) of memory 304 have remained erased after the write operation to thereby diagnose the outcome and extent of the backup.
With reference to
In certain embodiments, HMS controller 306 is configured with a count of the amount of NVM 304 required to store all of the volatile system data during backup, taking into account additional data capacity needed for error detection and correction as described above. With reference to
As explained above, in the event of system fault or power failure, or any of the aforementioned trigger conditions, backup is initiated, with HMS controller 306 programming or writing the system data from volatile memory subsystem 302 into non-volatile memory subsystem 304 until all of the volatile system data has been programmed. The operation, described with reference to
1. Detect system fault/power failure, trigger condition, user request (802)
2. Set the non-volatile flags (804)
3. Store the volatile system data into the non-volatile memory (806)
4. When the storage operation is complete, set the non-volatile flags (808)
When the HMS powers up after the system fault or power failure has been corrected, the non-volatile flags will be read in order to determine the system state. If the storage operation completed successfully, the MDF will be set to “1” and the MOF will be set to “1.” However, in the event of failure during the storage operation, the MDF will be set to “1” and the MOF will be cleared to “0.” At this point, the host system will be notified of the storage failure. This notification can take one or more of the following forms:
1. An error bit set in an error register.
2. An interrupt generated to the host system.
3. An LED lit to indicate the error condition.
Once the storage error has been identified, the NVM controller 406 can read the NVM 304 to determine how much system data was stored before the failure. The controller reads data from the memory array until a valid block is found that is in the erased state, all bits set to “1” including any bits that may be used by the error detection and correction algorithm. This distinguishes between a block of system data that is all “1” and a block that is erased. The controller can now determine how much valid data is present in the memory array.
It may be the case that the last block containing system data may only be partially programmed. This can be determined by reading the data and detecting/correcting any errors that may be present.
The controller 406 can provide access to the count of blocks (or other portions) containing system data by storing this count in a host system accessible register.
Once the host system has been informed of the storage error and the amount of system data that was stored before the error occurred, the host system can determine the next steps:
While embodiments and applications have been shown and described, it would be apparent to those skilled in the art having the benefit of this disclosure that many more modifications than mentioned above are possible without departing from the inventive concepts disclosed herein. The invention, therefore, is not to be restricted except in the spirit of the appended claims.
| Number | Date | Country | |
|---|---|---|---|
| 61798956 | Mar 2013 | US | |
| 61799271 | Mar 2013 | US | |
| 61799556 | Mar 2013 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 16517210 | Jul 2019 | US |
| Child | 17549251 | US | |
| Parent | 14214652 | Mar 2014 | US |
| Child | 16517210 | US |
