OVP Virtual Platform: FU540
This page provides detailed information about the OVP Virtual Platform Model of the
sifive.ovpworld.org FU540 platform.
Licensing
Open Source Apache 2.0
Description
SiFive FU540-C000 SoC module.
On start up or reset, the reset code at 0x1004 will jump to a jump table entry indexed by the MSEL register at address 0x1000 (default initial MSEL value is 0xf which will cause a jump to address 0x10000). Use the msel peripheral's MSEL parameter to change the initial value of this register.
To run a bare metal application use the --program command line option to specify an elf file to be loaded. It must be linked to use addresses corresponding to the implemented memory regions. The --program option will override the initial pc with the ELF file's start address.
To facilitate booting Linux an OVP SmartLoader psuedo-peripheral has been included that provides the functionality of the ZSBL/FSBL. The SmartLoader's dtb parameter should be used to specify the device tree blob file to load, and the bbl elf file should be loaded using the --objfilenoentry command line option.
Reference
SiFive Freedom U540-C000 Manual FU540-C000-v1.0.pdf (https://www.sifive.com/documentation/chips/freedom-u540-c000-manual)
Limitations
Caches and the Cache Controller are not modeled.
The Instruction Tightly Integrated Memory (ITIM) is implemented simply as RAM. Deallocation by writing to the byte immediately following the memory is a NOP.
The L2 Loosely Integrated Memory (L2-LIM) is implemented simply as RAM. It is always available, since the Cache Controller is not modeled.
The L2 Scratchpad memory is not modeled.
The Platform DMA Engine (PDMA) is not modeled.
The Pulse Width Modulator (PWM) is not modeled.
The Inter-Integrated Circuit (I2C) Master Interface is not modeled.
The Serial Peripheral Interfaces (SPI) are not modeled. Instead a Virtio Block MMIO device has been added at reserved address 0x1f000000, using interrupt 54.
The General Purpose Input/Output Controller (GPIO) is modeled with register stubs only.
The One-Time Programmable Memory Interface (OTP) is not modeled.
DDR controller is modeled with register stubs only. DDR memory is modeled as RAM.
The Debug Interface is not modeled.
Location
The FU540 virtual platform is located in an Imperas/OVP installation at the VLNV: sifive.ovpworld.org / module / FU540 / 1.0.
Platform Summary
Table : Components in platform
Platform Simulation Attributes
Table 1: Platform Simulation Attributes
Attribute | Value | Description |
---|
stoponctrlc | stoponctrlc | Stop on control-C |
External Ports for Module FU540
Table 2: External Ports
Port Type | Port Name | Internal Connection |
---|
netport | gi1 | gi1 |
netport | gi2 | gi2 |
netport | gi3 | gi3 |
netport | gi4 | gi4 |
netport | gi5 | gi5 |
netport | gi6 | gi6 |
netport | gi7 | gi7 |
netport | gi8 | gi8 |
netport | gi9 | gi9 |
netport | gi10 | gi10 |
netport | gi11 | gi11 |
netport | gi12 | gi12 |
netport | gi13 | gi13 |
netport | gi14 | gi14 |
netport | gi15 | gi15 |
netport | gi16 | gi16 |
netport | gi17 | gi17 |
netport | gi18 | gi18 |
netport | gi19 | gi19 |
netport | gi20 | gi20 |
netport | gi21 | gi21 |
netport | gi22 | gi22 |
netport | gi23 | gi23 |
netport | gi24 | gi24 |
netport | gi25 | gi25 |
netport | gi26 | gi26 |
netport | gi27 | gi27 |
netport | gi28 | gi28 |
netport | gi29 | gi29 |
netport | gi30 | gi30 |
netport | gi31 | gi31 |
netport | gi32 | gi32 |
netport | gi33 | gi33 |
netport | gi34 | gi34 |
netport | gi35 | gi35 |
netport | gi36 | gi36 |
netport | gi37 | gi37 |
netport | gi38 | gi38 |
netport | gi39 | gi39 |
netport | gi40 | gi40 |
netport | gi41 | gi41 |
netport | gi42 | gi42 |
netport | gi43 | gi43 |
netport | gi44 | gi44 |
netport | gi45 | gi45 |
netport | gi46 | gi46 |
netport | gi47 | gi47 |
netport | gi48 | gi48 |
netport | gi49 | gi49 |
netport | gi50 | gi50 |
netport | gi51 | gi51 |
netport | gi52 | gi52 |
netport | gi53 | gi53 |
Processor [sifive.ovpworld.org/processor/riscv/1.0] instance: E51
Processor model type: 'riscv' variant 'E51' definition
Imperas OVP processor models support multiple variants and details of the variants implemented in this model can be found in:
- the Imperas installation located at ImperasLib/source/sifive.ovpworld.org/processor/riscv/1.0/doc
- the OVP website:
OVP_Model_Specific_Information_sifive_riscv_E51.pdfDescription
RISC-V E51 64-bit processor model
Licensing
This Model is released under the Open Source Apache 2.0
Extensions Enabled by Default
The model has the following architectural extensions enabled, and the corresponding bits in the misa CSR Extensions field will be set upon reset:
misa bit 0: extension A (atomic instructions)
misa bit 2: extension C (compressed instructions)
misa bit 8: RV32I/RV64I/RV128I base integer instruction set
misa bit 12: extension M (integer multiply/divide instructions)
misa bit 20: extension U (User mode)
To specify features that can be dynamically enabled or disabled by writes to the misa register in addition to those listed above, use parameter "add_Extensions_mask". This is a string parameter containing the feature letters to add; for example, value "DV" indicates that double-precision floating point and the Vector Extension can be enabled or disabled by writes to the misa register, if supported on this variant. Parameter "sub_Extensions_mask" can be used to disable dynamic update of features in the same way.
Legacy parameter "misa_Extensions_mask" can also be used. This Uns32-valued parameter specifies all writable bits in the misa Extensions field, replacing any permitted bits defined in the base variant.
Note that any features that are indicated as present in the misa mask but absent in the misa will be ignored. See the next section.
Enabling Other Extensions
The following extensions are supported by the model, but not enabled by default in this variant:
misa bit 3: extension D (double-precision floating point)
misa bit 5: extension F (single-precision floating point)
To add features from this list to the visible set in the misa register, use parameter "add_Extensions". This is a string containing identification letters of features to enable; for example, value "DV" indicates that double-precision floating point and the Vector Extension should be enabled, if they are currently absent and are available on this variant.
Legacy parameter "misa_Extensions" can also be used. This Uns32-valued parameter specifies the reset value for the misa CSR Extensions field, replacing any permitted bits defined in the base variant.
To add features from this list to the implicitly-enabled set (not visible in the misa register), use parameter "add_implicit_Extensions". This is a string parameter in the same format as the "add_Extensions" parameter described above.
Disabling Extensions
The following extensions are enabled by default in the model and can be disabled:
misa bit 12: extension M (integer multiply/divide instructions)
To disable features that are enabled by default, use parameter "sub_Extensions". This is a string containing identification letters of features to disable; for example, value "DF" indicates that double-precision and single-precision floating point extensions should be disabled, if they are enabled by default on this variant.
To remove features from this list from the implicitly-enabled set (not visible in the misa register), use parameter "sub_implicit_Extensions". This is a string parameter in the same format as the "sub_Extensions" parameter described above.
Multicore Features
This is a multicore variant with 1 harts by default. The number of harts may be overridden with the "numHarts" parameter.
mtvec CSR
On this variant, the Machine trap-vector base-address register (mtvec) is writable. It can instead be configured as read-only using parameter "mtvec_is_ro".
Values written to "mtvec" are masked using the value 0x3ffffffffd. A different mask of writable bits may be specified using parameter "mtvec_mask" if required. In addition, when Vectored interrupt mode is enabled, parameter "tvec_align" may be used to specify additional hardware-enforced base address alignment. In this variant, "tvec_align" defaults to 64.
If parameter "mtvec_sext" is True, values written to "mtvec" are sign-extended from the most-significant writable bit. In this variant, "mtvec_sext" is False, indicating that "mtvec" is not sign-extended.
The initial value of "mtvec" is 0x0. A different value may be specified using parameter "mtvec" if required.
Reset
On reset, the model will restart at address 0x0. A different reset address may be specified using parameter "reset_address" or applied using optional input port "reset_addr" if required.
NMI
On an NMI, the model will restart at address 0x0; a different NMI address may be specified using parameter "nmi_address" or applied using optional input port "nmi_addr" if required. The cause reported on an NMI is 0x2 by default; a different cause may be specified using parameter "ecode_nmi" or applied using optional input port "nmi_cause" if required.
If parameter "rnmi_version" is not "none", resumable NMIs are supported, managed by additional CSRs "mnscratch", "mnepc", "mncause" and "mnstatus", following the indicated version of the Resumable NMI extension proposal. In this variant, "rnmi_version" is "0.2.1".
The NMI input is level-sensitive. To instead specify that the NMI input is latched on the rising edge of the NMI signal, set parameter "nmi_is_latched" to True.
WFI
WFI will halt the processor until an interrupt occurs. It can instead be configured as a NOP using parameter "wfi_is_nop". WFI timeout wait is implemented with a time limit of 0 (i.e. WFI causes an Illegal Instruction trap in Supervisor mode when mstatus.TW=1).
cycle CSR
The "cycle" CSR is implemented in this variant. Set parameter "cycle_undefined" to True to instead specify that "cycle" is unimplemented and accesses should cause Illegal Instruction traps.
instret CSR
The "instret" CSR is implemented in this variant. Set parameter "instret_undefined" to True to instead specify that "instret" is unimplemented and accesses should cause Illegal Instruction traps.
hpmcounter CSR
The "hpmcounter" CSRs are implemented in this variant. Set parameter "hpmcounter_undefined" to True to instead specify that "hpmcounter" CSRs are unimplemented and accesses should cause Illegal Instruction traps.
time CSR
The "time" CSR is not implemented in this variant and reads of it will cause Illegal Instruction traps. Set parameter "time_undefined" to False to instead specify that "time" is implemented.
mcycle CSR
The "mcycle" CSR is implemented in this variant. Set parameter "mcycle_undefined" to True to instead specify that "mcycle" is unimplemented and accesses should cause Illegal Instruction traps.
minstret CSR
The "minstret" CSR is implemented in this variant. Set parameter "minstret_undefined" to True to instead specify that "minstret" is unimplemented and accesses should cause Illegal Instruction traps.
mhpmcounter CSR
The "mhpmcounter" CSRs are implemented in this variant. Set parameter "mhpmcounter_undefined" to True to instead specify that "mhpmcounter" CSRs are unimplemented and accesses should cause Illegal Instruction traps.
Unaligned Accesses
Unaligned memory accesses are not supported by this variant. Set parameter "unaligned" to "T" to enable such accesses.
Unaligned memory accesses are not supported for AMO instructions by this variant. Set parameter "unalignedAMO" to "T" to enable such accesses.
Address misaligned exceptions are higher priority than page fault or access fault exceptions on this variant. Set parameter "unaligned_low_pri" to "T" to specify that they are lower priority instead.
PMP
8 PMP entries are implemented by this variant. Use parameter "PMP_registers" to specify a different number of PMP entries; set the parameter to 0 to disable the PMP unit. The PMP grain size (G) is 0, meaning that PMP regions as small as 4 bytes are implemented. Use parameter "PMP_grain" to specify a different grain size if required. Unaligned PMP accesses are not decomposed into separate aligned accesses; use parameter "PMP_decompose" to modify this behavior if required. Parameters to change the write masks for the PMP CSRs are not enabled; use parameter "PMP_maskparams" to modify this behavior if required. Parameters to change the reset values for the PMP CSRs are not enabled; use parameter "PMP_initialparams" to modify this behavior if required
Accesses to unimplemented PMP registers are write-ignored and read as zero on this variant. Set parameter "PMP_undefined" to True to indicate that such accesses should cause Illegal Instruction exceptions instead.
LR/SC Granule
LR/SC instructions are implemented with a 64-byte reservation granule. A different granule size may be specified using parameter "lr_sc_grain".
Compressed Extension
Standard compressed instructions are present in this variant. Legacy compressed extension features may also be configured using parameters described below. Use parameter "commpress_version" to enable more recent compressed extension features if required.
Parameter Zcea_version is used to specify the version of Zcea instructions present. By default, Zcea_version is set to "none" in this variant. Updates to this parameter require a commercial product license.
Parameter Zceb_version is used to specify the version of Zceb instructions present. By default, Zceb_version is set to "none" in this variant. Updates to this parameter require a commercial product license.
Parameter Zcee_version is used to specify the version of Zcee instructions present. By default, Zcee_version is set to "none" in this variant. Updates to this parameter require a commercial product license.
Privileged Architecture
This variant implements the Privileged Architecture with version specified in the References section of this document. Note that parameter "priv_version" can be used to select the required architecture version; see the following sections for detailed information about differences between each supported version.
Legacy Version 1.10
1.10 version of May 7 2017.
Version 20190608
Stable 1.11 version of June 8 2019, with these changes compared to version 1.10:
- mcountinhibit CSR defined;
- pages are never executable in Supervisor mode if page table entry U bit is 1;
- mstatus.TW is writable if any lower-level privilege mode is implemented (previously, it was just if Supervisor mode was implemented);
Version 20211203
1.12 draft version of December 3 2021, with these changes compared to version 20190608:
- mstatush, mseccfg, mseccfgh, menvcfg, menvcfgh, senvcfg, henvcfg, henvcfgh and mconfigptr CSRs defined;
- xret instructions clear mstatus.MPRV when leaving Machine mode if new mode is less privileged than M-mode;
- maximum number of PMP registers increased to 64;
- data endian is now configurable.
Version 1.12
Official 1.12 version, identical to 20211203.
Version master
Unstable master version, currently identical to 1.12.
Unprivileged Architecture
This variant implements the Unprivileged Architecture with version specified in the References section of this document. Note that parameter "user_version" can be used to select the required architecture version; see the following sections for detailed information about differences between each supported version.
Legacy Version 2.2
2.2 version of May 7 2017.
Version 20191213
Stable 20191213-Base-Ratified version of December 13 2019, with these changes compared to version 2.2:
- floating point fmin/fmax instruction behavior modified to comply with IEEE 754-201x.
- numerous other optional behaviors can be separately enabled using Z-prefixed parameters.
Other Extensions
Other extensions that can be configured are described in this section.
Zmmul
Parameter "Zmmul" is 0 on this variant, meaning that all multiply and divide instructions are implemented. if "Zmmul" is set to 1 then multiply instructions are implemented but divide and remainder instructions are not implemented.
Zicsr
Parameter "Zicsr" is 1 on this variant, meaning that standard CSRs and CSR access instructions are implemented. if "Zicsr" is set to 0 then standard CSRs and CSR access instructions are not implemented and an alternative scheme must be provided as a processor extension.
Zifencei
Parameter "Zifencei" is 1 on this variant, meaning that the fence.i instruction is implemented (but treated as a NOP by the model). if "Zifencei" is set to 0 then the fence.i instruction is not implemented.
Zicbom
Parameter "Zicbom" is 0 on this variant, meaning that code block management instructions are undefined. if "Zicbom" is set to 1 then code block management instructions cbo.clean, cbo.flush and cbo.inval are defined.
If Zicbom is present, the cache block size is given by parameter "cmomp_bytes". The instructions may cause traps if used illegally but otherwise are NOPs in this model.
Zicbop
Parameter "Zicbop" is 0 on this variant, meaning that prefetch instructions are undefined. if "Zicbop" is set to 1 then prefetch instructions prefetch.i, prefetch.r and prefetch.w are defined (but behave as NOPs in this model).
Zicboz
Parameter "Zicboz" is 0 on this variant, meaning that the cbo.zero instruction is undefined. if "Zicboz" is set to 1 then the cbo.zero instruction is defined.
If Zicboz is present, the cache block size is given by parameter "cmoz_bytes".
Smstateen
Parameter "Smstateen" is 0 on this variant, meaning that state enable CSRs are undefined. if "Smstateen" is set to 1 then state enable CSRs are defined.
Within the state enable CSRs, only bit 1 (for Zfinx), bit 57 (for xcontext CSR access), bit 62 (for xenvcfg CSR access) and bit 63 (for lower-level state enable CSR access) are currently implemented.
Load-Reserved/Store-Conditional Locking
By default, LR/SC locking is implemented automatically by the model and simulator, with a reservation granule defined by the "lr_sc_grain" parameter. It is also possible to implement locking externally to the model in a platform component, using the "LR_address", "SC_address" and "SC_valid" net ports, as described below.
The "LR_address" output net port is written by the model with the address used by a load-reserved instruction as it executes. This port should be connected as an input to the external lock management component, which should record the address, and also that an LR/SC transaction is active.
The "SC_address" output net port is written by the model with the address used by a store-conditional instruction as it executes. This should be connected as an input to the external lock management component, which should compare the address with the previously-recorded load-reserved address, and determine from this (and other implementation-specific constraints) whether the store should succeed. It should then immediately write the Boolean success/fail code to the "SC_valid" input net port of the model. Finally, it should update state to indicate that an LR/SC transaction is no longer active.
It is also possible to write zero to the "SC_valid" input net port at any time outside the context of a store-conditional instruction, which will mark any active LR/SC transaction as invalid.
Irrespective of whether LR/SC locking is implemented internally or externally, taking any exception or interrupt or executing exception-return instructions (e.g. MRET) will always mark any active LR/SC transaction as invalid.
Parameter "amo_aborts_lr_sc" is used to specify whether AMO operations abort any active LR/SC pair. In this variant, "amo_aborts_lr_sc" is 0.
Active Atomic Operation Indication
The "AMO_active" output net port is written by the model with a code indicating any current atomic memory operation while the instruction is active. The written codes are:
0: no atomic instruction active
1: AMOMIN active
2: AMOMAX active
3: AMOMINU active
4: AMOMAXU active
5: AMOADD active
6: AMOXOR active
7: AMOOR active
8: AMOAND active
9: AMOSWAP active
10: LR active
11: SC active
Interrupts
The "reset" port is an active-high reset input. The processor is halted when "reset" goes high and resumes execution from the reset address specified using the "reset_address" parameter or "reset_addr" port when the signal goes low. The "mcause" register is cleared to zero.
The "nmi" port is an active-high NMI input. The processor resumes execution from the address specified using the "nmi_address" parameter or "nmi_addr" port when the NMI signal goes high. The "mcause" register is cleared to zero.
All other interrupt ports are active high. For each implemented privileged execution level, there are by default input ports for software interrupt, timer interrupt and external interrupt; for example, for Machine mode, these are called "MSWInterrupt", "MTimerInterrupt" and "MExternalInterrupt", respectively. When the N extension is implemented, ports are also present for User mode. Parameter "unimp_int_mask" allows the default behavior to be changed to exclude certain interrupt ports. The parameter value is a mask in the same format as the "mip" CSR; any interrupt corresponding to a non-zero bit in this mask will be removed from the processor and read as zero in "mip", "mie" and "mideleg" CSRs (and Supervisor and User mode equivalents if implemented).
Parameter "external_int_id" can be used to enable extra interrupt ID input ports on each hart. If the parameter is True then when an external interrupt is applied the value on the ID port is sampled and used to fill the Exception Code field in the "mcause" CSR (or the equivalent CSR for other execution levels). For Machine mode, the extra interrupt ID port is called "MExternalInterruptID".
The "deferint" port is an active-high artifact input that, when written to 1, prevents any pending-and-enabled interrupt being taken (normally, such an interrupt would be taken on the next instruction after it becomes pending-and-enabled). The purpose of this signal is to enable alignment with hardware models in step-and-compare usage.
Debug Mode
The model can be configured to implement Debug mode using parameter "debug_mode". This implements features described in Chapter 4 of the RISC-V External Debug Support specification with version specified by parameter "debug_version" (see References). Some aspects of this mode are not defined in the specification because they are implementation-specific; the model provides infrastructure to allow implementation of a Debug Module using a custom harness. Features added are described below.
Parameter "debug_mode" can be used to specify three different behaviors, as follows:
1. If set to value "vector", then operations that would cause entry to Debug mode result in the processor jumping to the address specified by the "debug_address" parameter. It will execute at this address, in Debug mode, until a "dret" instruction causes return to non-Debug mode. Any exception generated during this execution will cause a jump to the address specified by the "dexc_address" parameter.
2. If set to value "interrupt", then operations that would cause entry to Debug mode result in the processor simulation call (e.g. opProcessorSimulate) returning, with a stop reason of OP_SR_INTERRUPT. In this usage scenario, the Debug Module is implemented in the simulation harness.
3. If set to value "halt", then operations that would cause entry to Debug mode result in the processor halting. Depending on the simulation environment, this might cause a return from the simulation call with a stop reason of OP_SR_HALT, or debug mode might be implemented by another platform component which then restarts the debugged processor again.
Debug State Entry
The specification does not define how Debug mode is implemented. In this model, Debug mode is enabled by a Boolean pseudo-register, "DM". When "DM" is True, the processor is in Debug mode. When "DM" is False, mode is defined by "mstatus" in the usual way.
Entry to Debug mode can be performed in any of these ways:
1. By writing True to register "DM" (e.g. using opProcessorRegWrite) followed by simulation of at least one cycle (e.g. using opProcessorSimulate), dcsr cause will be reported as trigger;
2. By writing a 1 then 0 to net "haltreq" (using opNetWrite) followed by simulation of at least one cycle (e.g. using opProcessorSimulate);
3. By writing a 1 to net "resethaltreq" (using opNetWrite) while the "reset" signal undergoes a negedge transition, followed by simulation of at least one cycle (e.g. using opProcessorSimulate);
4. By executing an "ebreak" instruction when Debug mode entry for the current processor mode is enabled by dcsr.ebreakm, dcsr.ebreaks or dcsr.ebreaku.
In all cases, the processor will save required state in "dpc" and "dcsr" and then perform actions described above, depending in the value of the "debug_mode" parameter.
Debug State Exit
Exit from Debug mode can be performed in any of these ways:
1. By writing False to register "DM" (e.g. using opProcessorRegWrite) followed by simulation of at least one cycle (e.g. using opProcessorSimulate);
2. By executing an "dret" instruction when Debug mode.
In both cases, the processor will perform the steps described in section 4.6 (Resume) of the Debug specification.
Debug Registers
When Debug mode is enabled, registers "dcsr", "dpc", "dscratch0" and "dscratch1" are implemented as described in the specification. These may be manipulated externally by a Debug Module using opProcessorRegRead or opProcessorRegWrite; for example, the Debug Module could write "dcsr" to enable "ebreak" instruction behavior as described above, or read and write "dpc" to emulate stepping over an "ebreak" instruction prior to resumption from Debug mode.
Debug Mode Execution
The specification allows execution of code fragments in Debug mode. A Debug Module implementation can cause execution in Debug mode by the following steps:
1. Write the address of a Program Buffer to the program counter using opProcessorPCSet;
2. If "debug_mode" is set to "halt", write 0 to pseudo-register "DMStall" (to leave halted state);
3. If entry to Debug mode was handled by exiting the simulation callback, call opProcessorSimulate or opRootModuleSimulate to resume simulation.
Debug mode will be re-entered in these cases:
1. By execution of an "ebreak" instruction; or:
2. By execution of an instruction that causes an exception.
In both cases, the processor will either jump to the debug exception address, or return control immediately to the harness, with stopReason of OP_SR_INTERRUPT, or perform a halt, depending on the value of the "debug_mode" parameter.
Debug Single Step
When in Debug mode, the processor or harness can cause a single instruction to be executed on return from that mode by setting dcsr.step. After one non-Debug-mode instruction has been executed, control will be returned to the harness. The processor will remain in single-step mode until dcsr.step is cleared.
Debug Event Priorities
The model supports two different models for determining which debug exception occurs when multiple debug events are pending:
1: original mode (when parameter "debug_priority"="original");
2: modified mode, as described in Debug Specification pull request 693 (when parameter "debug_priority"="PR693"). This mode resolves some anomalous behavior of the original specification.
Debug Ports
Port "DM" is an output signal that indicates whether the processor is in Debug mode
Port "haltreq" is a rising-edge-triggered signal that triggers entry to Debug mode (see above).
Port "resethaltreq" is a level-sensitive signal that triggers entry to Debug mode after reset (see above).
Debug Mask
It is possible to enable model debug messages in various categories. This can be done statically using the "debugflags" parameter, or dynamically using the "debugflags" command. Enabled messages are specified using a bitmask value, as follows:
Value 0x002: enable debugging of PMP and virtual memory state;
Value 0x004: enable debugging of interrupt state.
All other bits in the debug bitmask are reserved and must not be set to non-zero values.
Integration Support
This model implements a number of non-architectural pseudo-registers and other features to facilitate integration.
CSR Register External Implementation
If parameter "enable_CSR_bus" is True, an artifact 16-bit bus "CSR" is enabled. Slave callbacks installed on this bus can be used to implement modified CSR behavior (use opBusSlaveNew or icmMapExternalMemory, depending on the client API). A CSR with index 0xABC is mapped on the bus at address 0xABC0; as a concrete example, implementing CSR "time" (number 0xC01) externally requires installation of callbacks at address 0xC010 on the CSR bus.
LR/SC Active Address
Artifact register "LRSCAddress" shows the active LR/SC lock address. The register holds all-ones if there is no LR/SC operation active or if LR/SC locking is implemented externally as described above.
Limitations
Instruction pipelines are not modeled in any way. All instructions are assumed to complete immediately. This means that instruction barrier instructions (e.g. fence.i) are treated as NOPs, with the exception of any Illegal Instruction behavior, which is modeled.
Caches and write buffers are not modeled in any way. All loads, fetches and stores complete immediately and in order, and are fully synchronous. Data barrier instructions (e.g. fence) are treated as NOPs, with the exception of any Illegal Instruction behavior, which is modeled.
Real-world timing effects are not modeled: all instructions are assumed to complete in a single cycle.
Hardware Performance Monitor registers are not implemented and hardwired to zero.
Verification
All instructions have been extensively tested by Imperas, using tests generated specifically for this model and also reference tests from https://github.com/riscv/riscv-tests.
Also reference tests have been used from various sources including:
https://github.com/riscv/riscv-tests
https://github.com/ucb-bar/riscv-torture
The Imperas OVPsim RISC-V models are used in the RISC-V Foundation Compliance Framework as a functional Golden Reference:
https://github.com/riscv/riscv-compliance
where the simulated model is used to provide the reference signatures for compliance testing. The Imperas OVPsim RISC-V models are used as reference in both open source and commercial instruction stream test generators for hardware design verification, for example:
http://valtrix.in/sting from Valtrix
https://github.com/google/riscv-dv from Google
The Imperas OVPsim RISC-V models are also used by commercial and open source RISC-V Core RTL developers as a reference to ensure correct functionality of their IP.
References
The Model details are based upon the following specifications:
RISC-V Instruction Set Manual, Volume I: User-Level ISA (User Architecture Version 20191213)
RISC-V Instruction Set Manual, Volume II: Privileged Architecture (Privileged Architecture Version Ratified-IMFDQC-and-Priv-v1.11)
SiFive E51 Core Complex Manual v1p2
Instance Parameters
Several parameters can be specified when a processor is instanced in a platform. For this processor instance 'E51' it has been instanced with the following parameters:
Table 3: Processor Instance 'E51' Parameters (Configurations)
Parameter | Value | Description |
---|
simulateexceptions | simulateexceptions | Causes the processor simulate exceptions instead of halting |
mips | 1000 | The nominal MIPS for the processor |
Table 4: Processor Instance 'E51' Parameters (Attributes)
Parameter Name | Value | Type |
---|
mhartid | 0 | Uns64 |
local_int_num | 48 | Uns32 |
reset_address | 0x1004 | Uns64 |
variant | E51 | enum |
Memory Map for processor 'E51' bus: 'bus0'
Processor instance 'E51' is connected to bus 'bus0' using master port 'INSTRUCTION'.
Processor instance 'E51' is connected to bus 'bus0' using master port 'DATA'.
Table 5: Memory Map ( 'E51' / 'bus0' [width: 38] )
Lo Address | Hi Address | Instance | Component |
---|
0x0 | 0xFF | safe0addr | ram |
remappable | remappable | emgmt | trap |
0x1000 | 0x1FFF | msel | MSEL |
0x10000 | 0x17FFF | maskROM | ram |
0x1000000 | 0x1001FFF | E51Hart0DTIM | ram |
0x1800000 | 0x1802000 | E51Hart0ITIM | ram |
0x1808000 | 0x180F000 | U54Hart0ITIM | ram |
0x1810000 | 0x1817000 | U54Hart1ITIM | ram |
0x1818000 | 0x181F000 | U54Hart2ITIM | ram |
0x1820000 | 0x1827000 | U54Hart3ITIM | ram |
0x2000000 | 0x200BFFF | clint | CLINT |
0x8000000 | 0x9FFFFFF | l2LIM | ram |
0xC000000 | 0xC3FFFFF | plic | PLIC |
0x10000000 | 0x10000FFF | prci | PRCI |
0x10010000 | 0x1001001B | uart0 | UART |
0x10011000 | 0x1001101B | uart1 | UART |
0x10060000 | 0x10060FFF | gpio | gpio |
0x10090000 | 0x10090FFF | emac | gem |
0x100B0000 | 0x100B3FFF | ddrctl | DDRCTL |
0x100B8000 | 0x100B8007 | ddrctl | DDRCTL |
0x1F000000 | 0x1F0001FF | vbd0 | VirtioBlkMMIO |
0x80000000 | 0xBFFFFFFF | ddr1 | ram |
Net Connections to processor: 'E51'
Table 6: Processor Net Connections ( 'E51' )
Net Port | Net | Instance | Component |
---|
hart0_MTimerInterrupt | MTimerInterrupt0 | clint | CLINT |
hart0_MSWInterrupt | MSWInterrupt0 | clint | CLINT |
hart0_MExternalInterrupt | irqT0 | plic | PLIC |
Processor [sifive.ovpworld.org/processor/riscv/1.0] instance: U54
Processor model type: 'riscv' variant 'U54' definition
Imperas OVP processor models support multiple variants and details of the variants implemented in this model can be found in:
- the Imperas installation located at ImperasLib/source/sifive.ovpworld.org/processor/riscv/1.0/doc
- the OVP website:
OVP_Model_Specific_Information_sifive_riscv_U54.pdfDescription
RISC-V U54 64-bit processor model
Licensing
This Model is released under the Open Source Apache 2.0
Extensions Enabled by Default
The model has the following architectural extensions enabled, and the corresponding bits in the misa CSR Extensions field will be set upon reset:
misa bit 0: extension A (atomic instructions)
misa bit 2: extension C (compressed instructions)
misa bit 3: extension D (double-precision floating point)
misa bit 5: extension F (single-precision floating point)
misa bit 8: RV32I/RV64I/RV128I base integer instruction set
misa bit 12: extension M (integer multiply/divide instructions)
misa bit 18: extension S (Supervisor mode)
misa bit 20: extension U (User mode)
To specify features that can be dynamically enabled or disabled by writes to the misa register in addition to those listed above, use parameter "add_Extensions_mask". This is a string parameter containing the feature letters to add; for example, value "DV" indicates that double-precision floating point and the Vector Extension can be enabled or disabled by writes to the misa register, if supported on this variant. Parameter "sub_Extensions_mask" can be used to disable dynamic update of features in the same way.
Legacy parameter "misa_Extensions_mask" can also be used. This Uns32-valued parameter specifies all writable bits in the misa Extensions field, replacing any permitted bits defined in the base variant.
Note that any features that are indicated as present in the misa mask but absent in the misa will be ignored. See the next section.
Disabling Extensions
The following extensions are enabled by default in the model and can be disabled:
misa bit 0: extension A (atomic instructions)
misa bit 3: extension D (double-precision floating point)
misa bit 5: extension F (single-precision floating point)
misa bit 12: extension M (integer multiply/divide instructions)
To disable features that are enabled by default, use parameter "sub_Extensions". This is a string containing identification letters of features to disable; for example, value "DF" indicates that double-precision and single-precision floating point extensions should be disabled, if they are enabled by default on this variant.
To remove features from this list from the implicitly-enabled set (not visible in the misa register), use parameter "sub_implicit_Extensions". This is a string parameter in the same format as the "sub_Extensions" parameter described above.
Multicore Features
This is a multicore variant with 1 harts by default. The number of harts may be overridden with the "numHarts" parameter.
mtvec CSR
On this variant, the Machine trap-vector base-address register (mtvec) is writable. It can instead be configured as read-only using parameter "mtvec_is_ro".
Values written to "mtvec" are masked using the value 0x3ffffffffd. A different mask of writable bits may be specified using parameter "mtvec_mask" if required. In addition, when Vectored interrupt mode is enabled, parameter "tvec_align" may be used to specify additional hardware-enforced base address alignment. In this variant, "tvec_align" defaults to 64.
If parameter "mtvec_sext" is True, values written to "mtvec" are sign-extended from the most-significant writable bit. In this variant, "mtvec_sext" is False, indicating that "mtvec" is not sign-extended.
The initial value of "mtvec" is 0x0. A different value may be specified using parameter "mtvec" if required.
stvec CSR
Values written to "stvec" are masked using the value 0xfffffffffffffffd. A different mask of writable bits may be specified using parameter "stvec_mask" if required. In addition, when Vectored interrupt mode is enabled, parameter "tvec_align" may be used to specify additional hardware-enforced base address alignment. In this variant, "tvec_align" defaults to 64.
If parameter "stvec_sext" is True, values written to "stvec" are sign-extended from the most-significant writable bit. In this variant, "stvec_sext" is False, indicating that "stvec" is not sign-extended.
Reset
On reset, the model will restart at address 0x0. A different reset address may be specified using parameter "reset_address" or applied using optional input port "reset_addr" if required.
NMI
On an NMI, the model will restart at address 0x0; a different NMI address may be specified using parameter "nmi_address" or applied using optional input port "nmi_addr" if required. The cause reported on an NMI is 0x2 by default; a different cause may be specified using parameter "ecode_nmi" or applied using optional input port "nmi_cause" if required.
If parameter "rnmi_version" is not "none", resumable NMIs are supported, managed by additional CSRs "mnscratch", "mnepc", "mncause" and "mnstatus", following the indicated version of the Resumable NMI extension proposal. In this variant, "rnmi_version" is "0.2.1".
The NMI input is level-sensitive. To instead specify that the NMI input is latched on the rising edge of the NMI signal, set parameter "nmi_is_latched" to True.
WFI
WFI will halt the processor until an interrupt occurs. It can instead be configured as a NOP using parameter "wfi_is_nop". WFI timeout wait is implemented with a time limit of 0 (i.e. WFI causes an Illegal Instruction trap in Supervisor mode when mstatus.TW=1).
cycle CSR
The "cycle" CSR is implemented in this variant. Set parameter "cycle_undefined" to True to instead specify that "cycle" is unimplemented and accesses should cause Illegal Instruction traps.
instret CSR
The "instret" CSR is implemented in this variant. Set parameter "instret_undefined" to True to instead specify that "instret" is unimplemented and accesses should cause Illegal Instruction traps.
hpmcounter CSR
The "hpmcounter" CSRs are implemented in this variant. Set parameter "hpmcounter_undefined" to True to instead specify that "hpmcounter" CSRs are unimplemented and accesses should cause Illegal Instruction traps.
time CSR
The "time" CSR is not implemented in this variant and reads of it will cause Illegal Instruction traps. Set parameter "time_undefined" to False to instead specify that "time" is implemented.
mcycle CSR
The "mcycle" CSR is implemented in this variant. Set parameter "mcycle_undefined" to True to instead specify that "mcycle" is unimplemented and accesses should cause Illegal Instruction traps.
minstret CSR
The "minstret" CSR is implemented in this variant. Set parameter "minstret_undefined" to True to instead specify that "minstret" is unimplemented and accesses should cause Illegal Instruction traps.
mhpmcounter CSR
The "mhpmcounter" CSRs are implemented in this variant. Set parameter "mhpmcounter_undefined" to True to instead specify that "mhpmcounter" CSRs are unimplemented and accesses should cause Illegal Instruction traps.
Virtual Memory
This variant supports address translation modes 0 (bare) and 8 (Sv39). Use parameter "Sv_modes" to specify a bit mask of different implemented modes if required; for example, setting "Sv_modes" to (1<<0)+(1<<8) indicates that mode 0 (bare) and mode 8 (Sv39) are implemented. These indices correspond to writable values in the satp.MODE CSR field.
A 0-bit ASID is implemented. Use parameter "ASID_bits" to specify a different implemented ASID size if required.
TLB behavior is controlled by parameter "ASIDCacheSize". If this parameter is 0, then an unlimited number of TLB entries will be maintained concurrently. If this parameter is non-zero, then only TLB entries for up to "ASIDCacheSize" different ASIDs will be maintained concurrently initially; as new ASIDs are used, TLB entries for less-recently used ASIDs are deleted, which improves model performance in some cases. If the model detects that the TLB entry cache is too small (entry ejections are very frequent), it will increase the cache size automatically. In this variant, "ASIDCacheSize" is 8.
Unaligned Accesses
Unaligned memory accesses are not supported by this variant. Set parameter "unaligned" to "T" to enable such accesses.
Unaligned memory accesses are not supported for AMO instructions by this variant. Set parameter "unalignedAMO" to "T" to enable such accesses.
Address misaligned exceptions are higher priority than page fault or access fault exceptions on this variant. Set parameter "unaligned_low_pri" to "T" to specify that they are lower priority instead.
PMP
8 PMP entries are implemented by this variant. Use parameter "PMP_registers" to specify a different number of PMP entries; set the parameter to 0 to disable the PMP unit. The PMP grain size (G) is 0, meaning that PMP regions as small as 4 bytes are implemented. Use parameter "PMP_grain" to specify a different grain size if required. Unaligned PMP accesses are not decomposed into separate aligned accesses; use parameter "PMP_decompose" to modify this behavior if required. Parameters to change the write masks for the PMP CSRs are not enabled; use parameter "PMP_maskparams" to modify this behavior if required. Parameters to change the reset values for the PMP CSRs are not enabled; use parameter "PMP_initialparams" to modify this behavior if required
Accesses to unimplemented PMP registers are write-ignored and read as zero on this variant. Set parameter "PMP_undefined" to True to indicate that such accesses should cause Illegal Instruction exceptions instead.
LR/SC Granule
LR/SC instructions are implemented with a 64-byte reservation granule. A different granule size may be specified using parameter "lr_sc_grain".
Compressed Extension
Standard compressed instructions are present in this variant. Legacy compressed extension features may also be configured using parameters described below. Use parameter "commpress_version" to enable more recent compressed extension features if required.
Parameter Zcea_version is used to specify the version of Zcea instructions present. By default, Zcea_version is set to "none" in this variant. Updates to this parameter require a commercial product license.
Parameter Zceb_version is used to specify the version of Zceb instructions present. By default, Zceb_version is set to "none" in this variant. Updates to this parameter require a commercial product license.
Parameter Zcee_version is used to specify the version of Zcee instructions present. By default, Zcee_version is set to "none" in this variant. Updates to this parameter require a commercial product license.
Floating Point Features
Half precision floating point is not implemented. Use parameter "Zfh" to enable this if required.
By default, the processor starts with floating-point instructions disabled (mstatus.FS=0). Use parameter "mstatus_FS" to force mstatus.FS to a non-zero value for floating-point to be enabled from the start.
The specification is imprecise regarding the conditions under which mstatus.FS is set to Dirty state (3). Parameter "mstatus_fs_mode" can be used to specify the required behavior in this model, as described below.
If "mstatus_fs_mode" is set to "always_dirty" then the model implements a simplified floating point status view in which mstatus.FS holds values 0 (Off) and 3 (Dirty) only; any write of values 1 (Initial) or 2 (Clean) from privileged code behave as if value 3 was written.
If "mstatus_fs_mode" is set to "write_1" then mstatus.FS will be set to 3 (Dirty) by any explicit write to the fflags, frm or fcsr control registers, or by any executed instruction that writes an FPR, or by any executed floating point compare or conversion to integer/unsigned that signals a floating point exception. Floating point compare or conversion to integer/unsigned instructions that do not signal an exception will not set mstatus.FS.
If "mstatus_fs_mode" is set to "write_any" then mstatus.FS will be set to 3 (Dirty) by any explicit write to the fflags, frm or fcsr control registers, or by any executed instruction that writes an FPR, or by any executed floating point compare or conversion even if those instructions do not signal a floating point exception.
In this variant, "mstatus_fs_mode" is set to "always_dirty".
Privileged Architecture
This variant implements the Privileged Architecture with version specified in the References section of this document. Note that parameter "priv_version" can be used to select the required architecture version; see the following sections for detailed information about differences between each supported version.
Legacy Version 1.10
1.10 version of May 7 2017.
Version 20190608
Stable 1.11 version of June 8 2019, with these changes compared to version 1.10:
- mcountinhibit CSR defined;
- pages are never executable in Supervisor mode if page table entry U bit is 1;
- mstatus.TW is writable if any lower-level privilege mode is implemented (previously, it was just if Supervisor mode was implemented);
Version 20211203
1.12 draft version of December 3 2021, with these changes compared to version 20190608:
- mstatush, mseccfg, mseccfgh, menvcfg, menvcfgh, senvcfg, henvcfg, henvcfgh and mconfigptr CSRs defined;
- xret instructions clear mstatus.MPRV when leaving Machine mode if new mode is less privileged than M-mode;
- maximum number of PMP registers increased to 64;
- data endian is now configurable.
Version 1.12
Official 1.12 version, identical to 20211203.
Version master
Unstable master version, currently identical to 1.12.
Unprivileged Architecture
This variant implements the Unprivileged Architecture with version specified in the References section of this document. Note that parameter "user_version" can be used to select the required architecture version; see the following sections for detailed information about differences between each supported version.
Legacy Version 2.2
2.2 version of May 7 2017.
Version 20191213
Stable 20191213-Base-Ratified version of December 13 2019, with these changes compared to version 2.2:
- floating point fmin/fmax instruction behavior modified to comply with IEEE 754-201x.
- numerous other optional behaviors can be separately enabled using Z-prefixed parameters.
Other Extensions
Other extensions that can be configured are described in this section.
Zmmul
Parameter "Zmmul" is 0 on this variant, meaning that all multiply and divide instructions are implemented. if "Zmmul" is set to 1 then multiply instructions are implemented but divide and remainder instructions are not implemented.
Zicsr
Parameter "Zicsr" is 1 on this variant, meaning that standard CSRs and CSR access instructions are implemented. if "Zicsr" is set to 0 then standard CSRs and CSR access instructions are not implemented and an alternative scheme must be provided as a processor extension.
Zifencei
Parameter "Zifencei" is 1 on this variant, meaning that the fence.i instruction is implemented (but treated as a NOP by the model). if "Zifencei" is set to 0 then the fence.i instruction is not implemented.
Zicbom
Parameter "Zicbom" is 0 on this variant, meaning that code block management instructions are undefined. if "Zicbom" is set to 1 then code block management instructions cbo.clean, cbo.flush and cbo.inval are defined.
If Zicbom is present, the cache block size is given by parameter "cmomp_bytes". The instructions may cause traps if used illegally but otherwise are NOPs in this model.
Zicbop
Parameter "Zicbop" is 0 on this variant, meaning that prefetch instructions are undefined. if "Zicbop" is set to 1 then prefetch instructions prefetch.i, prefetch.r and prefetch.w are defined (but behave as NOPs in this model).
Zicboz
Parameter "Zicboz" is 0 on this variant, meaning that the cbo.zero instruction is undefined. if "Zicboz" is set to 1 then the cbo.zero instruction is defined.
If Zicboz is present, the cache block size is given by parameter "cmoz_bytes".
Svnapot
Parameter "Svnapot_page_mask" is 0x0 on this variant, meaning that NAPOT Translation Contiguity is not implemented. if "Svnapot_page_mask" is non-zero then NAPOT Translation Contiguity is enabled for page sizes indicated by that mask value when page table entry bit 63 is set.
If Svnapot is present, "Svnapot_page_mask" is a mask of page sizes for which contiguous pages can be created. For example, a value of 0x10000 implies that 64KiB contiguous pages are supported.
Svpbmt
Parameter "Svpbmt" is 0 on this variant, meaning that page-based memory types are not implemented. if "Svpbmt" is set to 1 then page-based memory types are indicated by page table entry bits 62:61.
Note that except for their effect on Page Faults, the encoded memory types do not alter the behavior of this model, which always implements strongly-ordered non-cacheable semantics.
Svinval
Parameter "Svinval" is 0 on this variant, meaning that fine-grained address-translation cache invalidation instructions are not implemented. if "Svinval" is set to 1 then fine-grained address-translation cache invalidation instructions sinval.vma, sfence.w.inval and sfence.inval.ir are implemented.
Smstateen
Parameter "Smstateen" is 0 on this variant, meaning that state enable CSRs are undefined. if "Smstateen" is set to 1 then state enable CSRs are defined.
Within the state enable CSRs, only bit 1 (for Zfinx), bit 57 (for xcontext CSR access), bit 62 (for xenvcfg CSR access) and bit 63 (for lower-level state enable CSR access) are currently implemented.
Load-Reserved/Store-Conditional Locking
By default, LR/SC locking is implemented automatically by the model and simulator, with a reservation granule defined by the "lr_sc_grain" parameter. It is also possible to implement locking externally to the model in a platform component, using the "LR_address", "SC_address" and "SC_valid" net ports, as described below.
The "LR_address" output net port is written by the model with the address used by a load-reserved instruction as it executes. This port should be connected as an input to the external lock management component, which should record the address, and also that an LR/SC transaction is active.
The "SC_address" output net port is written by the model with the address used by a store-conditional instruction as it executes. This should be connected as an input to the external lock management component, which should compare the address with the previously-recorded load-reserved address, and determine from this (and other implementation-specific constraints) whether the store should succeed. It should then immediately write the Boolean success/fail code to the "SC_valid" input net port of the model. Finally, it should update state to indicate that an LR/SC transaction is no longer active.
It is also possible to write zero to the "SC_valid" input net port at any time outside the context of a store-conditional instruction, which will mark any active LR/SC transaction as invalid.
Irrespective of whether LR/SC locking is implemented internally or externally, taking any exception or interrupt or executing exception-return instructions (e.g. MRET) will always mark any active LR/SC transaction as invalid.
Parameter "amo_aborts_lr_sc" is used to specify whether AMO operations abort any active LR/SC pair. In this variant, "amo_aborts_lr_sc" is 0.
Active Atomic Operation Indication
The "AMO_active" output net port is written by the model with a code indicating any current atomic memory operation while the instruction is active. The written codes are:
0: no atomic instruction active
1: AMOMIN active
2: AMOMAX active
3: AMOMINU active
4: AMOMAXU active
5: AMOADD active
6: AMOXOR active
7: AMOOR active
8: AMOAND active
9: AMOSWAP active
10: LR active
11: SC active
Interrupts
The "reset" port is an active-high reset input. The processor is halted when "reset" goes high and resumes execution from the reset address specified using the "reset_address" parameter or "reset_addr" port when the signal goes low. The "mcause" register is cleared to zero.
The "nmi" port is an active-high NMI input. The processor resumes execution from the address specified using the "nmi_address" parameter or "nmi_addr" port when the NMI signal goes high. The "mcause" register is cleared to zero.
All other interrupt ports are active high. For each implemented privileged execution level, there are by default input ports for software interrupt, timer interrupt and external interrupt; for example, for Machine mode, these are called "MSWInterrupt", "MTimerInterrupt" and "MExternalInterrupt", respectively. When the N extension is implemented, ports are also present for User mode. Parameter "unimp_int_mask" allows the default behavior to be changed to exclude certain interrupt ports. The parameter value is a mask in the same format as the "mip" CSR; any interrupt corresponding to a non-zero bit in this mask will be removed from the processor and read as zero in "mip", "mie" and "mideleg" CSRs (and Supervisor and User mode equivalents if implemented).
Parameter "external_int_id" can be used to enable extra interrupt ID input ports on each hart. If the parameter is True then when an external interrupt is applied the value on the ID port is sampled and used to fill the Exception Code field in the "mcause" CSR (or the equivalent CSR for other execution levels). For Machine mode, the extra interrupt ID port is called "MExternalInterruptID".
The "deferint" port is an active-high artifact input that, when written to 1, prevents any pending-and-enabled interrupt being taken (normally, such an interrupt would be taken on the next instruction after it becomes pending-and-enabled). The purpose of this signal is to enable alignment with hardware models in step-and-compare usage.
Debug Mode
The model can be configured to implement Debug mode using parameter "debug_mode". This implements features described in Chapter 4 of the RISC-V External Debug Support specification with version specified by parameter "debug_version" (see References). Some aspects of this mode are not defined in the specification because they are implementation-specific; the model provides infrastructure to allow implementation of a Debug Module using a custom harness. Features added are described below.
Parameter "debug_mode" can be used to specify three different behaviors, as follows:
1. If set to value "vector", then operations that would cause entry to Debug mode result in the processor jumping to the address specified by the "debug_address" parameter. It will execute at this address, in Debug mode, until a "dret" instruction causes return to non-Debug mode. Any exception generated during this execution will cause a jump to the address specified by the "dexc_address" parameter.
2. If set to value "interrupt", then operations that would cause entry to Debug mode result in the processor simulation call (e.g. opProcessorSimulate) returning, with a stop reason of OP_SR_INTERRUPT. In this usage scenario, the Debug Module is implemented in the simulation harness.
3. If set to value "halt", then operations that would cause entry to Debug mode result in the processor halting. Depending on the simulation environment, this might cause a return from the simulation call with a stop reason of OP_SR_HALT, or debug mode might be implemented by another platform component which then restarts the debugged processor again.
Debug State Entry
The specification does not define how Debug mode is implemented. In this model, Debug mode is enabled by a Boolean pseudo-register, "DM". When "DM" is True, the processor is in Debug mode. When "DM" is False, mode is defined by "mstatus" in the usual way.
Entry to Debug mode can be performed in any of these ways:
1. By writing True to register "DM" (e.g. using opProcessorRegWrite) followed by simulation of at least one cycle (e.g. using opProcessorSimulate), dcsr cause will be reported as trigger;
2. By writing a 1 then 0 to net "haltreq" (using opNetWrite) followed by simulation of at least one cycle (e.g. using opProcessorSimulate);
3. By writing a 1 to net "resethaltreq" (using opNetWrite) while the "reset" signal undergoes a negedge transition, followed by simulation of at least one cycle (e.g. using opProcessorSimulate);
4. By executing an "ebreak" instruction when Debug mode entry for the current processor mode is enabled by dcsr.ebreakm, dcsr.ebreaks or dcsr.ebreaku.
In all cases, the processor will save required state in "dpc" and "dcsr" and then perform actions described above, depending in the value of the "debug_mode" parameter.
Debug State Exit
Exit from Debug mode can be performed in any of these ways:
1. By writing False to register "DM" (e.g. using opProcessorRegWrite) followed by simulation of at least one cycle (e.g. using opProcessorSimulate);
2. By executing an "dret" instruction when Debug mode.
In both cases, the processor will perform the steps described in section 4.6 (Resume) of the Debug specification.
Debug Registers
When Debug mode is enabled, registers "dcsr", "dpc", "dscratch0" and "dscratch1" are implemented as described in the specification. These may be manipulated externally by a Debug Module using opProcessorRegRead or opProcessorRegWrite; for example, the Debug Module could write "dcsr" to enable "ebreak" instruction behavior as described above, or read and write "dpc" to emulate stepping over an "ebreak" instruction prior to resumption from Debug mode.
Debug Mode Execution
The specification allows execution of code fragments in Debug mode. A Debug Module implementation can cause execution in Debug mode by the following steps:
1. Write the address of a Program Buffer to the program counter using opProcessorPCSet;
2. If "debug_mode" is set to "halt", write 0 to pseudo-register "DMStall" (to leave halted state);
3. If entry to Debug mode was handled by exiting the simulation callback, call opProcessorSimulate or opRootModuleSimulate to resume simulation.
Debug mode will be re-entered in these cases:
1. By execution of an "ebreak" instruction; or:
2. By execution of an instruction that causes an exception.
In both cases, the processor will either jump to the debug exception address, or return control immediately to the harness, with stopReason of OP_SR_INTERRUPT, or perform a halt, depending on the value of the "debug_mode" parameter.
Debug Single Step
When in Debug mode, the processor or harness can cause a single instruction to be executed on return from that mode by setting dcsr.step. After one non-Debug-mode instruction has been executed, control will be returned to the harness. The processor will remain in single-step mode until dcsr.step is cleared.
Debug Event Priorities
The model supports two different models for determining which debug exception occurs when multiple debug events are pending:
1: original mode (when parameter "debug_priority"="original");
2: modified mode, as described in Debug Specification pull request 693 (when parameter "debug_priority"="PR693"). This mode resolves some anomalous behavior of the original specification.
Debug Ports
Port "DM" is an output signal that indicates whether the processor is in Debug mode
Port "haltreq" is a rising-edge-triggered signal that triggers entry to Debug mode (see above).
Port "resethaltreq" is a level-sensitive signal that triggers entry to Debug mode after reset (see above).
Debug Mask
It is possible to enable model debug messages in various categories. This can be done statically using the "debugflags" parameter, or dynamically using the "debugflags" command. Enabled messages are specified using a bitmask value, as follows:
Value 0x002: enable debugging of PMP and virtual memory state;
Value 0x004: enable debugging of interrupt state.
All other bits in the debug bitmask are reserved and must not be set to non-zero values.
Integration Support
This model implements a number of non-architectural pseudo-registers and other features to facilitate integration.
CSR Register External Implementation
If parameter "enable_CSR_bus" is True, an artifact 16-bit bus "CSR" is enabled. Slave callbacks installed on this bus can be used to implement modified CSR behavior (use opBusSlaveNew or icmMapExternalMemory, depending on the client API). A CSR with index 0xABC is mapped on the bus at address 0xABC0; as a concrete example, implementing CSR "time" (number 0xC01) externally requires installation of callbacks at address 0xC010 on the CSR bus.
LR/SC Active Address
Artifact register "LRSCAddress" shows the active LR/SC lock address. The register holds all-ones if there is no LR/SC operation active or if LR/SC locking is implemented externally as described above.
Page Table Walk Introspection
Artifact register "PTWStage" shows the active page table translation stage (0 if no stage active, 1 if HS-stage active, 2 if VS-stage active and 3 if G-stage active). This register is visibly non-zero only in a memory access callback triggered by a page table walk event.
Artifact register "PTWInputAddr" shows the input address of active page table translation. This register is visibly non-zero only in a memory access callback triggered by a page table walk event.
Artifact register "PTWLevel" shows the active level of page table translation (corresponding to index variable "i" in the algorithm described by Virtual Address Translation Process in the RISC-V Privileged Architecture specification). This register is visibly non-zero only in a memory access callback triggered by a page table walk event.
Artifact Register "fflags_i"
If parameter "enable_fflags_i" is True, an 8-bit artifact register "fflags_i" is added to the model. This register shows the floating point flags set by the current instruction (unlike the standard "fflags" CSR, in which the flag bits are sticky).
Limitations
Instruction pipelines are not modeled in any way. All instructions are assumed to complete immediately. This means that instruction barrier instructions (e.g. fence.i) are treated as NOPs, with the exception of any Illegal Instruction behavior, which is modeled.
Caches and write buffers are not modeled in any way. All loads, fetches and stores complete immediately and in order, and are fully synchronous. Data barrier instructions (e.g. fence) are treated as NOPs, with the exception of any Illegal Instruction behavior, which is modeled.
Real-world timing effects are not modeled: all instructions are assumed to complete in a single cycle.
Hardware Performance Monitor registers are not implemented and hardwired to zero.
The TLB is architecturally-accurate but not device accurate. This means that all TLB maintenance and address translation operations are fully implemented but the cache is larger than in the real device.
Verification
All instructions have been extensively tested by Imperas, using tests generated specifically for this model and also reference tests from https://github.com/riscv/riscv-tests.
Also reference tests have been used from various sources including:
https://github.com/riscv/riscv-tests
https://github.com/ucb-bar/riscv-torture
The Imperas OVPsim RISC-V models are used in the RISC-V Foundation Compliance Framework as a functional Golden Reference:
https://github.com/riscv/riscv-compliance
where the simulated model is used to provide the reference signatures for compliance testing. The Imperas OVPsim RISC-V models are used as reference in both open source and commercial instruction stream test generators for hardware design verification, for example:
http://valtrix.in/sting from Valtrix
https://github.com/google/riscv-dv from Google
The Imperas OVPsim RISC-V models are also used by commercial and open source RISC-V Core RTL developers as a reference to ensure correct functionality of their IP.
References
The Model details are based upon the following specifications:
RISC-V Instruction Set Manual, Volume I: User-Level ISA (User Architecture Version 20191213)
RISC-V Instruction Set Manual, Volume II: Privileged Architecture (Privileged Architecture Version Ratified-IMFDQC-and-Priv-v1.11)
SiFive U54-MC Core Complex Manual v1p0
Instance Parameters
Several parameters can be specified when a processor is instanced in a platform. For this processor instance 'U54' it has been instanced with the following parameters:
Table 7: Processor Instance 'U54' Parameters (Configurations)
Parameter | Value | Description |
---|
simulateexceptions | simulateexceptions | Causes the processor simulate exceptions instead of halting |
mips | 1000 | The nominal MIPS for the processor |
Table 8: Processor Instance 'U54' Parameters (Attributes)
Parameter Name | Value | Type |
---|
mhartid | 1 | Uns64 |
local_int_num | 48 | Uns32 |
reset_address | 0x1004 | Uns64 |
numHarts | 4 | Uns32 |
variant | U54 | enum |
Memory Map for processor 'U54' bus: 'bus0'
Processor instance 'U54' is connected to bus 'bus0' using master port 'INSTRUCTION'.
Processor instance 'U54' is connected to bus 'bus0' using master port 'DATA'.
Table 9: Memory Map ( 'U54' / 'bus0' [width: 38] )
Lo Address | Hi Address | Instance | Component |
---|
0x0 | 0xFF | safe0addr | ram |
remappable | remappable | emgmt | trap |
0x1000 | 0x1FFF | msel | MSEL |
0x10000 | 0x17FFF | maskROM | ram |
0x1000000 | 0x1001FFF | E51Hart0DTIM | ram |
0x1800000 | 0x1802000 | E51Hart0ITIM | ram |
0x1808000 | 0x180F000 | U54Hart0ITIM | ram |
0x1810000 | 0x1817000 | U54Hart1ITIM | ram |
0x1818000 | 0x181F000 | U54Hart2ITIM | ram |
0x1820000 | 0x1827000 | U54Hart3ITIM | ram |
0x2000000 | 0x200BFFF | clint | CLINT |
0x8000000 | 0x9FFFFFF | l2LIM | ram |
0xC000000 | 0xC3FFFFF | plic | PLIC |
0x10000000 | 0x10000FFF | prci | PRCI |
0x10010000 | 0x1001001B | uart0 | UART |
0x10011000 | 0x1001101B | uart1 | UART |
0x10060000 | 0x10060FFF | gpio | gpio |
0x10090000 | 0x10090FFF | emac | gem |
0x100B0000 | 0x100B3FFF | ddrctl | DDRCTL |
0x100B8000 | 0x100B8007 | ddrctl | DDRCTL |
0x1F000000 | 0x1F0001FF | vbd0 | VirtioBlkMMIO |
0x80000000 | 0xBFFFFFFF | ddr1 | ram |
Net Connections to processor: 'U54'
Table 10: Processor Net Connections ( 'U54' )
Net Port | Net | Instance | Component |
---|
hart0_MTimerInterrupt | MTimerInterrupt1 | clint | CLINT |
hart0_MSWInterrupt | MSWInterrupt1 | clint | CLINT |
hart0_MExternalInterrupt | irqT1 | plic | PLIC |
hart0_SExternalInterrupt | irqT2 | plic | PLIC |
hart1_MTimerInterrupt | MTimerInterrupt2 | clint | CLINT |
hart1_MSWInterrupt | MSWInterrupt2 | clint | CLINT |
hart1_MExternalInterrupt | irqT3 | plic | PLIC |
hart1_SExternalInterrupt | irqT4 | plic | PLIC |
hart2_MTimerInterrupt | MTimerInterrupt3 | clint | CLINT |
hart2_MSWInterrupt | MSWInterrupt3 | clint | CLINT |
hart2_MExternalInterrupt | irqT5 | plic | PLIC |
hart2_SExternalInterrupt | irqT6 | plic | PLIC |
hart3_MTimerInterrupt | MTimerInterrupt4 | clint | CLINT |
hart3_MSWInterrupt | MSWInterrupt4 | clint | CLINT |
hart3_MExternalInterrupt | irqT7 | plic | PLIC |
hart3_SExternalInterrupt | irqT8 | plic | PLIC |
Peripheral Instances
Peripheral [sifive.ovpworld.org/peripheral/MSEL/1.0] instance: msel
Description
Mode Select reset module. Entered on reset and calls boot code based on MSEL pin state. Override the MSEL parameter to specify the initial value for the MSEL pin state (default 0xf). From application code or debugger write to the MSEL register at offset 0 to change the MSEL pin state.
Limitations
None
Licensing
Open Source Apache 2.0
Reference
SiFive Freedom U540-C000 Manual FU540-C000-v1.0.pdf (https://www.sifive.com/documentation/chips/freedom-u540-c000-manual)
There are no configuration options set for this peripheral instance.
Peripheral [riscv.ovpworld.org/peripheral/CLINT/1.0] instance: clint
Licensing
Open Source Apache 2.0
Description
SiFive-compatabile Risc-V Core Local Interruptor (CLINT).
Use the num_harts parameter to specify the number of harts suported (default 1).
For each supported hart there will be an MTimerInterruptN and MSWInterruptN net port, plus msipN and mtimecmpN registers implemented, where N is a value from 0..num_harts-1.
There is also a single mtime register.
Reference
Various SiFive Core Complex manuals, e.g. SiFive U54 Core Complex Manual (https://sifive.cdn.prismic.io/sifive/a07d1a9a-2cb8-4cf5-bb75-5351888ea2e1_u54_core_complex_manual_21G2.pdf)
Table 11: Configuration options (attributes) set for instance 'clint'
Attributes | Value |
---|
num_harts | 5 |
clockMHz | 1.0 |
Peripheral [sifive.ovpworld.org/peripheral/PLIC/1.0] instance: plic
Licensing
Open Source Apache 2.0
Description
SiFive PLIC Interrupt Controller
Use parameters to configure specific implementation.
Limitations
None
Reference
Various SiFive Core Complex manuals, e.g. SiFive E31 Core Complex Manual 31G2.01.00 (https://sifive.cdn.prismic.io/sifive/c29f9c69-5254-4f9a-9e18-24ea73f34e81_e31_core_complex_manual_21G2.pdf)
Table 12: Configuration options (attributes) set for instance 'plic'
Attributes | Value |
---|
num_targets | 9 |
num_sources | 53 |
Peripheral [sifive.ovpworld.org/peripheral/PRCI/1.0] instance: prci
Description
Power Reset Clocking Interrupt (PRCI) block for SiFive FU540 chip
Limitations
None
Register only model. Reset values based on typical post-ZSBL configuration (1GHz coreclk, 500MHz tlclk).
Licensing
Open Source Apache 2.0
Reference
SiFive Freedom U540-C000 Manual FU540-C000-v1.0.pdf (https://www.sifive.com/documentation/chips/freedom-u540-c000-manual)
There are no configuration options set for this peripheral instance.
Peripheral [sifive.ovpworld.org/peripheral/UART/1.0] instance: uart0
Licensing
Open Source Apache 2.0
Description
Sifive UART
Limitations
When simulatebaud parameter is set to true baud rate delays are modeled for receive only, not transmit. Data always sent immediately.
Reference
SiFive Freedom U540-C000 Manual FU540-C000-v1.0.pdf (https://www.sifive.com/documentation/chips/freedom-u540-c000-manual)
Table 13: Configuration options (attributes) set for instance 'uart0'
Attributes | Value |
---|
refClkFreq | 500000000 |
Peripheral [sifive.ovpworld.org/peripheral/UART/1.0] instance: uart1
Licensing
Open Source Apache 2.0
Description
Sifive UART
Limitations
When simulatebaud parameter is set to true baud rate delays are modeled for receive only, not transmit. Data always sent immediately.
Reference
SiFive Freedom U540-C000 Manual FU540-C000-v1.0.pdf (https://www.sifive.com/documentation/chips/freedom-u540-c000-manual)
Table 14: Configuration options (attributes) set for instance 'uart1'
Attributes | Value |
---|
refClkFreq | 500000000 |
Peripheral [sifive.ovpworld.org/peripheral/gpio/1.0] instance: gpio
Description
SiFive coreip-s51-arty GPIO Registers (gpio)
Licensing
Open Source Apache 2.0
Limitations
This model implements only the registers for generation of input or output data values.
Reference
SiFive Freedom E SDK coreip-s51-arty Board Support Package details.
There are no configuration options set for this peripheral instance.
Peripheral [cadence.ovpworld.org/peripheral/gem/1.0] instance: emac
Description
Model of Cadence Gigabit Ethernet Controller (GEM). For further details please consult README-EMAC.txt
This model is based upon the data and use in the Xilinx Zynq
Basic network Tx/Rx functionality tested using Xilinx Linux Kernel using wget and other similar tools
Tested with Xilinx SDK Example driver.
Licensing
Open Source Apache 2.0
Limitations
This model is based upon the data from the Xilinx Zynq platform, other registers may not be included.
Does not implement: VLAN, pause frames, filtering or timestamps.
Reference
Zynq-7000 TRM (https://www.xilinx.com/support/documentation/user_guides/ug585-Zynq-7000-TRM.pdf)
There are no configuration options set for this peripheral instance.
Peripheral [ovpworld.org/peripheral/trap/1.0] instance: emgmt
Description
Open a port and allocate a region that is defined by parameters.
The region can be configured to act as standard memory or can report read/write accesses.
Licensing
Open Source Apache 2.0
Limitations
This peripheral cannot be used in a hardware description used to generate a TLM platform.
Reference
This is not based upon the operation of a real device but is intended to be used for bring up and development of new virtual platforms.
Table 15: Configuration options (attributes) set for instance 'emgmt'
Attributes | Value |
---|
portAddress | 0x100a0000 |
portSize | 0x1000 |
Peripheral [sifive.ovpworld.org/peripheral/DDRCTL/1.0] instance: ddrctl
Description
DDR Controller Register Block for SiFive FU540 chip
Licensing
Open Source Apache 2.0
Reference
SiFive Freedom U540-C000 Manual FU540-C000-v1.0.pdf (https://www.sifive.com/documentation/chips/freedom-u540-c000-manual)
Limitations
Register only model. Register address space modeled as RAM except for registers that require write masks or reset values.
There are no configuration options set for this peripheral instance.
Peripheral [ovpworld.org/peripheral/VirtioBlkMMIO/1.0] instance: vbd0
Description
VIRTIO version 1 mmio block device This model implements a VIRTIO MMIO block device as described in: http://docs.oasis-open.org/virtio/virtio/v1.0/virtio-v1.0.pdf. Use the VB_DRIVE parameter to specify the disk image file to use. Set the VB_DRIVE_DELTA parameter to 1 to prevent writes to disk during simulation from changing the image file.
Limitations
Only supports the Legacy (Device Version 1) interface. Only little endian guests are supported.
Licensing
Open Source Apache 2.0
Reference
http://docs.oasis-open.org/virtio/virtio/v1.0/virtio-v1.0.pdf
There are no configuration options set for this peripheral instance.
Peripheral [riscv.ovpworld.org/peripheral/SmartLoaderRV64Linux/1.0] instance: smartLoader
Licensing
Open Source Apache 2.0
Description
Psuedo-peripheral to insert boot code for a Riscv 64-bit Linux kernel boot. Loads simulated memory with a device tree blob file and boot code to set regs and jump to a Risc-v Linux Kernel.
Limitations
Only supports little endian
Reference
RISC-V Linux Kernel development
Table 16: Configuration options (attributes) set for instance 'smartLoader'
Attributes | Value |
---|
membase | 0x80000000 |
memsize | 0x40000000 |