CN102681930B - A kind of chip-scale error logging method - Google Patents
A kind of chip-scale error logging method Download PDFInfo
- Publication number
- CN102681930B CN102681930B CN201210149211.2A CN201210149211A CN102681930B CN 102681930 B CN102681930 B CN 102681930B CN 201210149211 A CN201210149211 A CN 201210149211A CN 102681930 B CN102681930 B CN 102681930B
- Authority
- CN
- China
- Prior art keywords
- mistake
- error
- depositor
- local
- register
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 23
- 230000000712 assembly Effects 0.000 claims abstract description 5
- 238000000429 assembly Methods 0.000 claims abstract description 5
- 238000013507 mapping Methods 0.000 claims description 9
- 230000007246 mechanism Effects 0.000 claims description 6
- 238000002372 labelling Methods 0.000 claims description 3
- 238000013461 design Methods 0.000 abstract description 8
- 238000013517 stratification Methods 0.000 abstract description 5
- 238000012360 testing method Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Debugging And Monitoring (AREA)
Abstract
The present invention proposes a kind of chip-scale error logging method, the chip that profit designs in this way, use the incompatible misregistration of error register collection of stratification tissue, and send out interruption by other assemblies in system and enable the modes such as mistake pin to outside System Reports mistake；Can distinguish and record the mistake of different menace level when misregistration, and can be configured to as required enable or shield certain type of error logging, wherein, the two kinds of error register set in local and the overall situation for misregistration use hierarchical structure to carry out tissue registration's mistake: use the mistake that the incompatible memorization COMS clip of local error register set certain concrete parts internal are corresponding；Use the incompatible error logging collected in each local error set of registers of global error register set, and to outside System Reports.
Description
Technical field
The present invention relates to computer chip design field, be specifically related to a kind of chip-scale error logging method.
Background technology
Along with the high speed development of the application electronizations such as scientific algorithm, commercial service, government function, user is to service
The requirement of the aspects such as the performance of ICT infrastructure, capacity, density, availability, safety such as device, storage and the network equipment is also got over
Come the highest.And chip is as the basic component units of these ICT infrastructure, its importance is self-evident.Grind in chip design
Sending out and throw in sheet production process, chip yield is the principal element affecting R&D cycle, design cost, production cost.Chip
Repeatedly throwing sheet not only increase design, checking, test period, and the throwing sheet production cost of costliness also results in whole chip and grinds
Cost allowance during producing, therefore just rapid build prototype system should carry out chip-scale in early days in chip logic design
With whole-system verification and test, once throw sheet success rate with guarantee.This is accomplished by the chip-scale error logging of a kind of high efficient and flexible
Method, in order to test proof procedure in locating core chip level problem.
Summary of the invention
It is an object of the invention to provide a kind of chip-scale error logging method.
It is an object of the invention to realize in the following manner, local error depositor and global error depositor combine, and adopt
Tissue registration's mistake is carried out: use local error depositor to carry out the mistake that memorization COMS clip certain concrete parts internal are corresponding with hierarchical structure
By mistake；Global error depositor is used to collect the error logging in each local error depositor, and to outside System Reports；
Local error depositor includes 1) local error conditions depositor, 2) local error control depositor, 3) local is wrong
By mistake menace level depositor, 4) local fault log register first, 5) locally subsequent error log register；Wherein:
1) local error conditions depositor, identifies the every kind of mistake occurred in corresponding component, and every kind of mistake uses 1bit
Representing, when there is certain type of mistake, bit corresponding in depositor is set to 1；
2) local error controls depositor, controls whether to record certain type error that corresponding component mistake produces, its bit
Definition and local each one_to_one corresponding of error status register, if certain control bit that local error controls in depositor is set to
Position, then the corresponding mistake detected can be shielded, and does not record and processes；
3) local error menace level depositor, it is provided that certain mistake is mapped to the mechanism of multiple mistake menace level,
Can be according to the definition of the error type in local error menace level depositor-menace level mapping relations during the corresponding mistake of generation
Carry out error reporting, it is assumed that need to support following 3 kinds of wrong menace levels: (1) can right the wrong, the system of referring to can recover and
There is no the loss of information, it is not necessary to the mistake of the participation of software；Including link crc error, it is possible to corrected by link layer retransmission；
(2) recoverable error refers to be corrected by hardware mechanisms, needs the mistake recovered by upper layer software (applications)；(3) fatal error,
Refer to cause specific affairs unreliable, but system remains to properly functioning mistake, including the data portion only affecting affairs
Point the most repairable mistake of ECC, cannot carry out, by hardware or software, the mistake correcting or recover, it may be required system is multiple
Position returns to the mistake of reliable behavior, including cache multidigit labelling PCI-E link failure wrong, permanent；
The menace level that every kind of type of error is corresponding needs to represent with two bit, if 00b correspondence can be righted the wrong, 01b pair
Recoverable error, 10b correspondence fatal error, 11b is answered to retain use；
4) local fault log register first, be used for recording when certain mistake of corresponding component is detected first is corresponding
Information, including message content, mistake address；
5) locally subsequent error log register, is used for recording certain mistake rear supervention in addition to first time of corresponding component
Corresponding information time raw, including error count；
Global error depositor, including global error status register 1), global error control depositor 2), the overall situation first
Fault log register 3), the overall situation subsequent error log register 4), system event status register 5) and system event control
Depositor 6), wherein:
Global error status register 1), identify in chip and whether all parts makes a mistake, the mistake of each parts
State uses 1bit to represent by mistake, and when certain parts makes a mistake, bit corresponding in depositor is set to 1；
Global error controls depositor 2), control whether to record the mistake that certain parts produces, its bit definition and the overall situation are wrong
Each one_to_one corresponding of status register, if certain control bit in global error status register is set, then detects by mistake
The mistake of corresponding component can be shielded, do not record and process；
Overall situation fault log register 3 first) and overall situation subsequent error log register 4) record all parts head respectively
Secondary when making a mistake and follow-up field data when making a mistake；
System event status register 5) menace level corresponding to the mistake that occurs of memorization COMS clip all parts；
System event controls depositor 6) mapping relations of definition menace level-report manner, it is configurable that certain is serious etc.
The mode that the mistake of level is reported to other assemblies of system, including sending out interruption, enabling mistake pin；
Specifically comprise the following steps that
1) certain parts produces a certain class mistake；
2) judge whether the most to have shielded such mistake, if shielded, the most not in " local error control depositor "
Record this mistake, terminate；Otherwise, the corresponding bit in " local error conditions depositor " is set；
3) judging whether this type of mistake first, if occurring first, then " local error log first is deposited in renewal
Device " content；Otherwise, " locally subsequent error log register " content is updated；
4) depositor report mistake is controlled according to the configuration in " local error menace level depositor " to global error；
5) judge whether the most to have shielded this parts relevant error, if shielded in " global error control depositor "
Cover, the most do not record this mistake, terminate；Otherwise, the corresponding bit in " global error status register " is set；
6) judging whether this parts relevant error first, if occurring first, then updating " overall situation mistake day first
Will depositor " content；Otherwise, " overall situation subsequent error log register " content is updated；
7) the corresponding mistake menace level in " system event status register " is updated；
8) according to the menace level-report manner mapping relations of configuration in " system event control depositor ", in sending out
The modes such as disconnected, enable mistake pin are to outside System Reports mistake.So far, whole error logging and reporting process terminate.
The invention has the beneficial effects as follows: this chip-scale error logging side possessing stratification tissue flexibly configurable
Method, compensate for that traditional die error logging method efficiency is low, the deficiency of very flexible, it is possible to shortens chip and designs, verifies, tests
Cycle also can effectively ensure that chip once throws sheet success rate, thus has boundless development prospect and high technology valency
Value.
Accompanying drawing explanation
The error register set schematic diagram that accompanying drawing 1 is stratification tissue；
Accompanying drawing 2 chip-scale error logging schematic flow sheet.
Detailed description of the invention
With reference to the accompanying drawings, to present disclosure with an instantiation describe realize invention described in chip-scale
The process of error logging method.
The present invention proposes a kind of chip-scale error logging method, the chip that profit designs in this way, uses stratification
The error register of tissue gathers misregistration, and sends out interruption by other assemblies in system and enable mistake pin etc.
Mode is to outside System Reports mistake；Can distinguish and record the mistake of different menace level when misregistration, and can be according to need
It is configured to enable or shield certain type of error logging.
The local for misregistration in the present invention and two kinds of error registers of the overall situation gather, and use stratification knot
Gou Lai tissue registration mistake: wherein, uses local error depositor to carry out the mistake that memorization COMS clip certain concrete parts internal are corresponding；Make
The error logging in each local error depositor is collected with global error depositor, and to outside System Reports, such as Fig. 1 institute
Show:
Local error depositor includes that local error conditions depositor, local error control depositor, local error are serious
Level register, local fault log register, locally subsequent error log register first.Wherein:
Local error conditions register identification goes out the every kind of mistake occurred in corresponding component, and every kind of mistake uses 1bit table
Show.When there is certain type of mistake, bit corresponding in depositor is set to 1.
Local error controls whether register controlled records certain type error that corresponding component mistake produces, and its bit is fixed
Justice and local each one_to_one corresponding of error status register.If certain control bit that local error controls in depositor is set to
Position, then the corresponding mistake detected can be shielded, and does not record and processes.
Local error menace level depositor provides the mechanism certain mistake being mapped to multiple mistake menace level, occurs
Can carry out according to the definition of the error type in local error menace level depositor-menace level mapping relations during corresponding mistake
Error reporting.Assume to need to support following 3 kinds of wrong menace levels: can righting the wrong, (system can be recovered and not have information
Lose, it is not necessary to the participation of software.Such as link crc error, can be corrected by link layer retransmission), recoverable error (cannot lead to
Cross hardware mechanisms corrigendum, need the mistake recovered by upper layer software (applications), specific affairs may be caused unreliable, but system remains to
Properly functioning.Mistake as the most repairable in ECC, only affects the data division of affairs), fatal error (hardware or soft cannot be passed through
Part is corrected or recovers, it may be required system reset returns to reliable behavior, as wrong, permanent in cache multidigit labelling
PCI-E link failure etc.).The menace level that then every kind of type of error is corresponding needs to represent with two bit, and can set 00b correspondence can
Right the wrong, 01b correspondence recoverable error, 10b correspondence fatal error, 11b retain use.
Local fault log register first, is used for the corresponding letter recorded when certain mistake of corresponding component is detected first
Breath, such as message content, mistake address etc..
Locally subsequent error log register, is used for recording certain mistake follow-up generation in addition to for the first time of corresponding component
Time corresponding information, such as error count etc..
Global error depositor, controls depositor, overall situation mistake first including global error status register, global error
Log register, overall situation subsequent error log register, system event status register and system event control depositor, its
In:
Global error status register identifies in chip and whether makes a mistake in all parts, the wrong shape of each parts
State uses 1bit to represent.When certain parts makes a mistake, bit corresponding in depositor is set to 1.
Global error controls whether register controlled records the mistake that certain parts produces, its bit definition and global error
Each one_to_one corresponding of status register.If certain control bit in global error status register is set, then detect
The mistake of corresponding component can be shielded, and does not record and processes.
Overall situation fault log register first and overall situation subsequent error log register record all parts respectively and send out first
During raw mistake and follow-up field data when making a mistake.
The menace level of the mistake correspondence that system event status register memorization COMS clip all parts occurs.
System event controls the mapping relations of register definitions menace level-report manner, certain menace level configurable
The mode reported to other assemblies of system of mistake, interrupt as sent out, enable mistake pin etc..
As described in summary of the invention, the error logging process in the present invention refers to accompanying drawing 2, specifically comprises the following steps that
1) certain parts produces a certain class mistake；
2) judge whether the most to have shielded such mistake, if shielded, the most not in " local error control depositor "
Record this mistake, terminate；Otherwise, the corresponding bit in " local error conditions depositor " is set；
3) judging whether this type of mistake first, if occurring first, then " local error log first is deposited in renewal
Device " content；Otherwise, " locally subsequent error log register " content is updated；
4) depositor report mistake is controlled according to the configuration in " local error menace level depositor " to global error；
5) judge whether the most to have shielded this parts relevant error, if shielded in " global error control depositor "
Cover, the most do not record this mistake, terminate；Otherwise, the corresponding bit in " global error status register " is set；
6) judging whether this parts relevant error first, if occurring first, then updating " overall situation mistake day first
Will depositor " content；Otherwise, " overall situation subsequent error log register " content is updated；
7) the corresponding mistake menace level in " system event status register " is updated；
8) according to the menace level-report manner mapping relations of configuration in " system event control depositor ", in sending out
The modes such as disconnected, enable mistake pin are to outside System Reports mistake.So far, whole error logging and reporting process terminate.
In addition to the technical characteristic described in description, it is the known technology of those skilled in the art.
Claims (1)
1. a chip-scale error logging method, it is characterised in that local error depositor and global error depositor combine to be adopted
Tissue registration's mistake is carried out: use local error depositor to carry out the mistake that memorization COMS clip certain concrete parts internal are corresponding with hierarchical structure
By mistake；Global error depositor is used to collect the error logging in each local error depositor, and to outside System Reports；
Local error depositor includes 1) local error conditions depositor, 2) local error control depositor, 3) local error is tight
Weight level register, 4) local fault log register first, 5) locally subsequent error log register；Wherein:
1) local error conditions depositor, identifies the every kind of mistake occurred in corresponding component, and every kind of mistake uses 1bit to represent,
When there is certain type of mistake, bit corresponding in depositor is set to 1；
2) local error controls depositor, controls whether to record certain type error that corresponding component mistake produces, and its bit defines
With each one_to_one corresponding of local error status register, if certain control bit that local error controls in depositor is set,
The corresponding mistake then detected can be shielded, and does not record and processes；
3) local error menace level depositor, it is provided that certain mistake is mapped to the mechanism of multiple mistake menace level, occurs
Can carry out according to the definition of the error type in local error menace level depositor-menace level mapping relations during corresponding mistake
Error reporting, needs to support following 3 kinds of wrong menace levels: (1) can right the wrong, and the system of referring to can be recovered and not have information
Loss, it is not necessary to the mistake of the participation of software；Including link crc error, it is possible to corrected by link layer retransmission；(2) can be extensive
Multiple mistake refers to be corrected by hardware mechanisms, needs the mistake recovered by upper layer software (applications)；(3) fatal error, referring to can
Specific affairs can be caused unreliable, but system remains to properly functioning mistake, including the ECC of the data division only affecting affairs
The most repairable mistake, cannot carry out, by hardware or software, the mistake correcting or recover, it may be required system reset recovers
To the mistake of reliable behavior, including cache multidigit labelling PCI-E link failure wrong, permanent；
The menace level that every kind of type of error is corresponding needs to represent with two bit, if 00b correspondence can be righted the wrong, 01b correspondence can
Recover mistake, 10b correspondence fatal error, 11b retain use；
4) local fault log register first, is used for the corresponding letter recorded when certain mistake of corresponding component is detected first
Breath, including message content, mistake address；
5) locally subsequent error log register, is used for recording follow-up in addition to for the first time of certain mistake of corresponding component when occurring
Corresponding information, including error count；
Global error depositor, controls depositor, overall situation error log first including global error status register, global error
Depositor, overall situation subsequent error log register, system event status register and system event control depositor, wherein:
Global error status register, identifies in chip and whether makes a mistake in all parts, the error condition of each parts
Using 1bit to represent, when certain parts makes a mistake, bit corresponding in depositor is set to 1；
Global error controls depositor, controls whether to record the mistake that certain parts produces, its bit definition and global error state
Each one_to_one corresponding of depositor, if certain control bit in global error status register is set, then detect is corresponding
The mistake of parts can be shielded, and does not record and processes；
Overall situation fault log register first and overall situation subsequent error log register record all parts respectively and mistake occur first
Mistake and follow-up field data when making a mistake；
The menace level of the mistake correspondence that system event status register memorization COMS clip all parts occurs；
System event controls the mapping relations of register definitions menace level-report manner, the mistake of certain menace level configurable
By mistake to the mode of other assemblies of system report, including sending out interruption, enabling mistake pin；
Specifically comprise the following steps that
1) certain parts produces a certain class mistake；
2) judge whether the most to have shielded such mistake in " local error control depositor ", if shielded, the most not record
This mistake, terminates；Otherwise, the corresponding bit in " local error conditions depositor " is set；
3) judging whether this type of mistake first, if occurring first, then updating " local fault log register first "
Content；Otherwise, " locally subsequent error log register " content is updated；
4) depositor report mistake is controlled according to the configuration in " local error menace level depositor " to global error；
5) judge whether the most to have shielded this parts relevant error, if shielded, then in " global error control depositor "
Do not record this mistake, terminate；Otherwise, the corresponding bit in " global error status register " is set；
6) judging whether this parts relevant error first, if occurring first, then " overall situation error log first is posted in renewal
Storage " content；Otherwise, " overall situation subsequent error log register " content is updated；
7) the corresponding mistake menace level in " system event status register " is updated；
8) according to the menace level-report manner mapping relations of configuration in " system event control depositor ", by sending out interruption, making
The mode of energy mistake pin is to outside System Reports mistake, and so far, whole error logging and reporting process terminate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210149211.2A CN102681930B (en) | 2012-05-15 | 2012-05-15 | A kind of chip-scale error logging method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210149211.2A CN102681930B (en) | 2012-05-15 | 2012-05-15 | A kind of chip-scale error logging method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102681930A CN102681930A (en) | 2012-09-19 |
| CN102681930B true CN102681930B (en) | 2016-08-17 |
Family
ID=46813895
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210149211.2A Active CN102681930B (en) | 2012-05-15 | 2012-05-15 | A kind of chip-scale error logging method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102681930B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104077375B (en) * | 2014-06-24 | 2017-09-12 | 华为技术有限公司 | The processing method and node of a kind of wrong catalogue of CC NUMA systems interior joint |
| CN104133751A (en) * | 2014-08-06 | 2014-11-05 | 浪潮(北京)电子信息产业有限公司 | Chip debugging method and chip |
| CN104407952A (en) * | 2014-11-12 | 2015-03-11 | 浪潮(北京)电子信息产业有限公司 | Method and system for debugging through multi-CPU (central processing unit) node controller chip |
| US10922161B2 (en) * | 2018-11-28 | 2021-02-16 | Intel Corporation | Apparatus and method for scalable error detection and reporting |
| CN110399317B (en) * | 2019-07-15 | 2020-12-25 | 西安微电子技术研究所 | Software self-adaptive multifunctional controller of embedded system |
| CN113832663B (en) * | 2021-09-18 | 2022-08-16 | 珠海格力电器股份有限公司 | Control chip fault recording method and device and control chip fault reading method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101599812A (en) * | 2008-06-04 | 2009-12-09 | 富士通株式会社 | Data transmission set |
| CN102073533A (en) * | 2011-01-14 | 2011-05-25 | 中国人民解放军国防科学技术大学 | Multicore architecture supporting dynamic binary translation |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3889391B2 (en) * | 2003-11-06 | 2007-03-07 | ローム株式会社 | Memory device and display device |
-
2012
- 2012-05-15 CN CN201210149211.2A patent/CN102681930B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101599812A (en) * | 2008-06-04 | 2009-12-09 | 富士通株式会社 | Data transmission set |
| CN102073533A (en) * | 2011-01-14 | 2011-05-25 | 中国人民解放军国防科学技术大学 | Multicore architecture supporting dynamic binary translation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102681930A (en) | 2012-09-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102681930B (en) | A kind of chip-scale error logging method | |
| CN101833497B (en) | Computer fault management system based on expert system method | |
| CN102612065B (en) | Quick fault-tolerance detection method for monitoring abnormal event by wireless sensor network | |
| US9495233B2 (en) | Error framework for a microprocesor and system | |
| US20020144193A1 (en) | Method and system for fault isolation methodology for I/O unrecoverable, uncorrectable error | |
| CN103220180B (en) | The processing method that a kind of OpenStack cloud platform is abnormal | |
| CN101960429B (en) | Video media data storage system and related methods | |
| CN102521084A (en) | Data storage and reading method | |
| CN101127675A (en) | Initialization method for main nodes of Ethernet loop network system | |
| CN104731670A (en) | Switch type on-board computer tolerant system facing satellite | |
| CN103761172B (en) | Hardware fault diagnosis system based on neutral net | |
| CN106293984A (en) | A kind of computer glitch automatically processes mode and device | |
| CN106648968A (en) | Data recovery method and device when ECC correction failure occurs on chip | |
| CN104376143B (en) | Soft error screen method based on approximate Logic circuit | |
| CN102404139A (en) | Method for increasing fault tolerance performance of application level of fault tolerance server | |
| CN106227464A (en) | A kind of double-deck redundant storage system and data write, reading and restoration methods | |
| CN103246585A (en) | Storage controller fault detecting method | |
| CN102819480A (en) | Computer and method for monitoring memory thereof | |
| CN106330535A (en) | Train-ground communication data processing method and apparatus | |
| CN107807862A (en) | Detect the method, apparatus and server of hard disk failure point | |
| CN102750194A (en) | Large-scale integrated circuit level error recording and responding method | |
| CN104156276A (en) | RAID method for preventing two disks from being damaged | |
| CN104731677A (en) | High-reliability storage and diagnosis method for external SRAMs (static random access memories) of safety instrument transmitters | |
| CN105893190A (en) | Diagnosis processing method and system for multi-path IO errors | |
| CN103390429B (en) | The online test method of a kind of hard disk and server |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |