URL
https://opencores.org/ocsvn/neorv32/neorv32/trunk
Subversion Repositories neorv32
[/] [neorv32/] [trunk/] [docs/] [datasheet/] [cpu_csr.adoc] - Rev 69
Go to most recent revision | Compare with Previous | Blame | View Log
<<<
:sectnums:
=== Control and Status Registers (CSRs)
The following table shows a summary of all available CSRs. The address field defines the CSR address for
the CSR access instructions. The *[ASM]* name can be used for (inline) assembly code and is directly
understood by the assembler/compiler. The *[C]* names are defined by the NEORV32 core library and can be
used as immediate in plain C code. The *R/W* column shows whether the CSR can be read and/or written.
The NEORV32-specific CSRs are mapped to the official "custom CSRs" CSR address space.
[IMPORTANT]
The CSRs, the CSR-related instructions as well as the complete exception/interrupt processing
system are only available when the `CPU_EXTENSION_RISCV_Zicsr` generic is _true_.
[IMPORTANT]
When trying to write to a read-only CSR (like the `time` CSR) or when trying to access a nonexistent
CSR or when trying to access a machine-mode CSR from less-privileged user-mode an
illegal instruction exception is raised.
[NOTE]
CSR reset value: Please note that most of the CSRs do *NOT* provide a dedicated reset. Hence,
these CSRs are not initialized by a hardware reset and keep an *UNDEFINED* value until they are
explicitly initialized by the software (normally, this is already done by the NEORV32-specific
`crt0.S` start-up code). For more information see section <<_cpu_hardware_reset>>.
**CSR Listing**
The description of each single CSR provides the following summary:
.CSR description
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| _Address_ | _Description_ | _ASM alias_
3+| Reset value: _CSR content after hardware reset_ (also see <<_cpu_hardware_reset>>)
3+| _Detailed description_
|======
.Not Implemented CSRs / CSR Bits
[IMPORTANT]
All CSR bits that are unused / not implemented / not shown are _hardwired to zero_. All CSRs that are not
implemented at all (and are not "disabled" using certain configuration generics) will trigger an exception on
access. The CSR that are implemented within the NEORV32 might cause an exception if they are disabled.
See the according CSR description for more information.
.Debug Mode CSRs
[IMPORTANT]
The _debug mode_ CSRs are not listed here since they are only accessible in debug mode and not during normal CPU operation.
See section <<_cpu_debug_mode_csrs>>.
<<<
// ####################################################################################################################
**CSR Listing Notes**
CSRs with the following notes ...
* `X`: _custom_ - have or are a custom CPU-specific extension (that is allowed by the RISC-V specs)
* `R`: _read-only_ - are read-only (in contrast to the originally specified r/w capability)
* `C`: _constrained_ - have a constrained compatibility, not all specified bits are implemented
.NEORV32 Control and Status Registers (CSRs)
[cols="<6,<11,<16,^3,<25,^3"]
[options="header"]
|=======================
| Address | Name [ASM] | Name [C] | R/W | Function | Note
6+^| **<<_floating_point_csrs>>**
| 0x001 | <<_fflags>> | _CSR_FFLAGS_ | r/w | Floating-point accrued exceptions |
| 0x002 | <<_frm>> | _CSR_FRM_ | r/w | Floating-point dynamic rounding mode |
| 0x003 | <<_fcsr>> | _CSR_FCSR_ | r/w | Floating-point control and status (`frm` + `fflags`) |
6+^| **<<_machine_configuration_csrs>>**
| 0x30a | <<_menvcfg>> | _CSR_MENVCFG_ | r/- | Machine environment configuration register - low word | `R`
| 0x31a | <<_menvcfgh>> | _CSR_MENVCFGH_ | r/- | Machine environment configuration register - low word | `R`
6+^| **<<_machine_trap_setup_csrs>>**
| 0x300 | <<_mstatus>> | _CSR_MSTATUS_ | r/w | Machine status register - low word | `C`
| 0x301 | <<_misa>> | _CSR_MISA_ | r/- | Machine CPU ISA and extensions | `R`
| 0x304 | <<_mie>> | _CSR_MIE_ | r/w | Machine interrupt enable register | `X`
| 0x305 | <<_mtvec>> | _CSR_MTVEC_ | r/w | Machine trap-handler base address (for ALL traps) |
| 0x306 | <<_mcounteren>> | _CSR_MCOUNTEREN_ | r/w | Machine counter-enable register | `C`
| 0x310 | <<_mstatush>> | _CSR_MSTATUSH_ | r/- | Machine status register - high word | `C`
6+^| **<<_machine_trap_handling_csrs>>**
| 0x340 | <<_mscratch>> | _CSR_MSCRATCH_ | r/w | Machine scratch register |
| 0x341 | <<_mepc>> | _CSR_MEPC_ | r/w | Machine exception program counter |
| 0x342 | <<_mcause>> | _CSR_MCAUSE_ | r/w | Machine trap cause | `X`
| 0x343 | <<_mtval>> | _CSR_MTVAL_ | r/- | Machine bad address or instruction | `R`
| 0x344 | <<_mip>> | _CSR_MIP_ | r/w | Machine interrupt pending register | `X`
6+^| **<<_machine_physical_memory_protection_csrs>>**
| 0x3a0 .. 0x3af | <<_pmpcfg, `pmpcfg0`>> .. <<_pmpcfg, `pmpcfg15`>> | _CSR_PMPCFG0_ .. _CSR_PMPCFG15_ | r/w | Physical memory protection config. for region 0..63 | `C`
| 0x3b0 .. 0x3ef | <<_pmpaddr, `pmpaddr0`>> .. <<_pmpaddr, `pmpaddr63`>> | _CSR_PMPADDR0_ .. _CSR_PMPADDR63_ | r/w | Physical memory protection addr. register region 0..63 |
6+^| **<<_machine_counter_and_timer_csrs>>**
| 0xb00 | <<_mcycleh, `mcycle`>> | _CSR_MCYCLE_ | r/w | Machine cycle counter low word |
| 0xb02 | <<_minstreth, `minstret`>> | _CSR_MINSTRET_ | r/w | Machine instruction-retired counter low word |
| 0xb80 | <<_mcycleh>> | _CSR_MCYCLE_ | r/w | Machine cycle counter high word |
| 0xb82 | <<_minstreth>> | _CSR_MINSTRET_ | r/w | Machine instruction-retired counter high word |
| 0xc00 | <<_cycleh, `cycle`>> | _CSR_CYCLE_ | r/- | Cycle counter low word |
| 0xc01 | <<_timeh, `time`>> | _CSR_TIME_ | r/- | System time (from MTIME) low word |
| 0xc02 | <<_instreth, `instret`>> | _CSR_INSTRET_ | r/- | Instruction-retired counter low word |
| 0xc80 | <<_cycleh>> | _CSR_CYCLEH_ | r/- | Cycle counter high word |
| 0xc81 | <<_timeh>> | _CSR_TIMEH_ | r/- | System time (from MTIME) high word |
| 0xc82 | <<_instreth>> | _CSR_INSTRETH_ | r/- | Instruction-retired counter high word |
6+^| **<<_hardware_performance_monitors_hpm_csrs>>**
| 0x323 .. 0x33f | <<_mhpmevent, `mhpmevent3`>> .. <<_mhpmevent, `mhpmevent31`>> | _CSR_MHPMEVENT3_ .. _CSR_MHPMEVENT31_ | r/w | Machine performance-monitoring event selector 3..31 | `X`
| 0xb03 .. 0xb1f | <<_mhpmcounterh, `mhpmcounter3`>> .. <<_mhpmcounterh, `mhpmcounter31`>> | _CSR_MHPMCOUNTER3_ .. _CSR_MHPMCOUNTER31_ | r/w | Machine performance-monitoring counter 3..31 low word |
| 0xb83 .. 0xb9f | <<_mhpmcounterh, `mhpmcounter3h`>> .. <<_mhpmcounterh, `mhpmcounter31h`>> | _CSR_MHPMCOUNTER3H_ .. _CSR_MHPMCOUNTER31H_ | r/w | Machine performance-monitoring counter 3..31 high word |
6+^| **<<_machine_counter_setup_csrs>>**
| 0x320 | <<_mcountinhibit>> | _CSR_MCOUNTINHIBIT_ | r/w | Machine counter-enable register |
6+^| **<<_machine_information_csrs>>**
| 0xf11 | <<_mvendorid>> | _CSR_MVENDORID_ | r/- | Vendor ID |
| 0xf12 | <<_marchid>> | _CSR_MARCHID_ | r/- | Architecture ID |
| 0xf13 | <<_mimpid>> | _CSR_MIMPID_ | r/- | Machine implementation ID / version |
| 0xf14 | <<_mhartid>> | _CSR_MHARTID_ | r/- | Machine thread ID |
| 0xf15 | <<_mconfigptr>> | _CSR_MCONFIGPTR_ | r/- | Machine configuration pointer register |
|=======================
<<<
// ####################################################################################################################
:sectnums:
==== Floating-Point CSRs
These CSRs are available if the `Zfinx` extensions is enabled (`CPU_EXTENSION_RISCV_Zfinx` is _true_).
Otherwise any access to the floating-point CSRs will raise an illegal instruction exception.
:sectnums!:
===== **`fflags`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x001 | **Floating-point accrued exceptions** | `fflags`
3+| Reset value: _UNDEFINED_
3+| The `fflags` CSR is compatible to the RISC-V specifications. It shows the accrued ("accumulated")
exception flags in the lowest 5 bits. This CSR is only available if a floating-point CPU extension is enabled.
See the RISC-V ISA spec for more information.
|======
:sectnums!:
===== **`frm`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x002 | **Floating-point dynamic rounding mode** | `frm`
3+| Reset value: _UNDEFINED_
3+| The `frm` CSR is compatible to the RISC-V specifications and is used to configure the rounding modes using
the lowest 3 bits. This CSR is only available if a floating-point CPU extension is enabled. See the RISC-V
ISA spec for more information.
|======
:sectnums!:
===== **`fcsr`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x003 | **Floating-point control and status register** | `fcsr`
3+| Reset value: _UNDEFINED_
3+| The `fcsr` CSR is compatible to the RISC-V specifications. It provides combined read/write access to the
`fflags` and `frm` CSRs. This CSR is only available if a floating-point CPU extension is enabled. See the
RISC-V ISA spec for more information.
|======
<<<
// ####################################################################################################################
:sectnums:
==== Machine Configuration CSRs
:sectnums!:
===== **`menvcfg`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x30a | **Machine environment configuration register** | `menvcfg`
3+| Reset value: _0x00000000_
3+| The features of this CSR are not implemented yet. The register is read-only. NOTE: This register
only exists if the `U` ISA extensions is enabled.
|======
:sectnums!:
===== **`menvcfgh`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x31a | **Machine environment configuration register - high word** | `menvcfgh`
3+| Reset value: _0x00000000_
3+| The features of this CSR are not implemented yet. The register is read-only. NOTE: This register
only exists if the `U` ISA extensions is enabled.
|======
<<<
// ####################################################################################################################
:sectnums:
==== Machine Trap Setup CSRs
:sectnums!:
===== **`mstatus`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x300 | **Machine status register** | `mstatus`
3+| Reset value: _0x00000000_
3+| The `mstatus` CSR is compatible to the RISC-V specifications. It shows the CPU's current execution state.
The following bits are implemented (all remaining bits are always zero and are read-only).
|======
.Machine status register
[cols="^1,<3,^1,<5"]
[options="header",grid="rows"]
|=======================
| Bit | Name [C] | R/W | Function
| 12:11 | _CSR_MSTATUS_MPP_H_ : _CSR_MSTATUS_MPP_L_ | r/w | Previous machine privilege level, 11 = machine (M) level, 00 = user (U) level
| 7 | _CSR_MSTATUS_MPIE_ | r/w | Previous machine global interrupt enable flag state
| 3 | _CSR_MSTATUS_MIE_ | r/w | Machine global interrupt enable flag
|=======================
When entering an exception/interrupt, the `MIE` flag is copied to `MPIE` and cleared afterwards. When leaving
the exception/interrupt (via the `mret` instruction), `MPIE` is copied back to `MIE`.
:sectnums!:
===== **`misa`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x301 | **ISA and extensions** | `misa`
3+| Reset value: _configuration dependant_
3+| The `misa` CSR gives information about the actual CPU features. The lowest 26 bits show the implemented
CPU extensions. The following bits are implemented (all remaining bits are always zero and are read-only).
|======
[IMPORTANT]
The `misa` CSR is not fully RISC-V-compatible as it is read-only. Hence, implemented CPU
extensions cannot be switch on/off during runtime. For compatibility reasons any write access to this
CSR is simply ignored and will _NOT_ cause an illegal instruction exception.
.Machine ISA and extension register
[cols="^1,<3,^1,<5"]
[options="header",grid="rows"]
|=======================
| Bit | Name [C] | R/W | Function
| 31:30 | _CSR_MISA_MXL_HI_EXT_ : _CSR_MISA_MXL_LO_EXT_ | r/- | 32-bit architecture indicator (always _01_)
| 23 | _CSR_MISA_X_EXT_ | r/- | `X` extension bit is always set to indicate custom non-standard extensions
| 20 | _CSR_MISA_U_EXT_ | r/- | `U` CPU extension (user mode) available, set when _CPU_EXTENSION_RISCV_U_ enabled
| 12 | _CSR_MISA_M_EXT_ | r/- | `M` CPU extension (mul/div) available, set when _CPU_EXTENSION_RISCV_M_ enabled
| 8 | _CSR_MISA_I_EXT_ | r/- | `I` CPU base ISA, cleared when _CPU_EXTENSION_RISCV_E_ enabled
| 4 | _CSR_MISA_E_EXT_ | r/- | `E` CPU extension (embedded) available, set when _CPU_EXTENSION_RISCV_E_ enabled
| 2 | _CSR_MISA_C_EXT_ | r/- | `C` CPU extension (compressed instruction) available, set when _CPU_EXTENSION_RISCV_C_ enabled
| 0 | _CSR_MISA_A_EXT_ | r/- | `A` CPU extension (atomic memory access) available, set when _CPU_EXTENSION_RISCV_A_ enabled
|=======================
[TIP]
Information regarding the implemented RISC-V `Z*` _sub-extensions_ (like `Zicsr` or `Zfinx`) can be found
in the `CPU` <<_system_configuration_information_memory_sysinfo, SYSINFO>> register.
:sectnums!:
===== **`mie`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x304 | **Machine interrupt-enable register** | `mie`
3+| Reset value: _UNDEFINED_
3+| The `mie` CSR is compatible to the RISC-V specifications and features custom extensions for the fast
interrupt channels. It is used to enabled specific interrupts sources. Please note that interrupts also have to be
globally enabled via the `CSR_MSTATUS_MIE` flag of the `mstatus` CSR. The following bits are implemented
(all remaining bits are always zero and are read-only):
|======
.Machine ISA and extension register
[cols="^1,<3,^1,<5"]
[options="header",grid="rows"]
|=======================
| Bit | Name [C] | R/W | Function
| 31:16 | _CSR_MIE_FIRQ15E_ : _CSR_MIE_FIRQ0E_ | r/w | Fast interrupt channel 15..0 enable
| 11 | _CSR_MIE_MEIE_ | r/w | Machine _external_ interrupt enable
| 7 | _CSR_MIE_MTIE_ | r/w | Machine _timer_ interrupt enable (from _MTIME_)
| 3 | _CSR_MIE_MSIE_ | r/w | Machine _software_ interrupt enable
|=======================
:sectnums!:
===== **`mtvec`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x305 | **Machine trap-handler base address** | `mtvec`
3+| Reset value: _UNDEFINED_
3+| The `mtvec` CSR is compatible to the RISC-V specifications. It stores the base address for ALL machine
traps. Thus, it defines the main entry point for exception/interrupt handling regardless of the actual trap
source. The lowest two bits of this register are always zero and cannot be modified (= address mode only).
|======
.Machine trap-handler base address
[cols="^1,^1,<8"]
[options="header",grid="rows"]
|=======================
| Bit | R/W | Function
| 31:2 | r/w | 4-byte aligned base address of trap base handler
| 1:0 | r/- | Always zero
|=======================
:sectnums!:
===== **`mcounteren`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x306 | **Machine counter enable** | `mcounteren`
3+| Reset value: _UNDEFINED_
3+| The `mcounteren` CSR is compatible to the RISC-V specifications. The bits of this CSR define which
counter/timer CSR can be accessed (read) from code running in a less-privileged modes. For example,
if user-level code tries to read from a counter/timer CSR without enabled access, an illegal instruction
exception is raised. NOTE: If the `U` ISA extension is not enabled this CSR does not exist.
|======
.Machine counter enable register
[cols="^1,<3,^1,<5"]
[options="header",grid="rows"]
|=======================
| Bit | Name [C] | R/W | Function
| 31:3 | `0` | r/- | Always zero: user-level code is **not** allowed to read HPM counters
| 2 | _CSR_MCOUNTEREN_IR_ | r/w | User-level code is allowed to read `cycle[h]` CSRs when set
| 1 | _CSR_MCOUNTEREN_TM_ | r/w | User-level code is allowed to read `time[h]` CSRs when set
| 0 | _CSR_MCOUNTEREN_CY_ | r/w | User-level code is allowed to read `instret[h]` CSRs when set
|=======================
:sectnums!:
===== **`mstatush`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x310 | **Machine status register - high word** | `mstatush`
3+| Reset value: _0x00000000_
3+| The `mstatush` CSR is compatible to the RISC-V specifications. In combination with <<_mstatus>> it shows additional
execution state information. The NEORV32 `mstatush` CSR is read-only and all bits are hardwired to zero.
|======
[NOTE]
The NEORV32 `mstatush` CSR is not a physical register. All write access are ignored and all read accesses will always
return zero. However, any access will not raise an illegal instruction exception. The CSR address is implemented
in order to comply with the RISC-V privilege architecture specs.
<<<
// ####################################################################################################################
:sectnums:
==== Machine Trap Handling CSRs
:sectnums!:
===== **`mscratch`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x340 | **Scratch register for machine trap handlers** | `mscratch`
3+| Reset value: _UNDEFINED_
3+| The `mscratch` CSR is compatible to the RISC-V specifications. It is a general purpose scratch register that
can be used by the exception/interrupt handler. The content pf this register after reset is undefined.
|======
:sectnums!:
===== **`mepc`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x341 | **Machine exception program counter** | `mepc`
3+| Reset value: _UNDEFINED_
3+| The `mepc` CSR is compatible to the RISC-V specifications. For exceptions (like an illegal instruction) this
register provides the address of the exception-causing instruction. For Interrupt (like a machine timer
interrupt) this register provides the address of the next not-yet-executed instruction.
|======
:sectnums!:
===== **`mcause`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x342 | **Machine trap cause** | `mcause`
3+| Reset value: _UNDEFINED_
3+| The `mcause` CSR is compatible to the RISC-V specifications. It show the cause ID for a taken exception.
|======
.Machine trap cause register
[cols="^1,^1,<8"]
[options="header",grid="rows"]
|=======================
| Bit | R/W | Function
| 31 | r/w | `1` if the trap is caused by an interrupt (`0` if the trap is caused by an exception)
| 30:5 | r/- | _Reserved_, read as zero
| 4:0 | r/w | Trap ID, see <<_neorv32_trap_listing>>
|=======================
:sectnums!:
===== **`mtval`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x343 | **Machine bad address or instruction** | `mtval`
3+| Reset value: _UNDEFINED_
3+| The `mtval` CSR is compatible to the RISC-V specifications. When a trap is triggered, the CSR shows either
the faulting address (for misaligned/faulting load/stores/fetch) or the faulting instruction itself (for illegal
instructions). For interrupts the CSR is set to zero.
|======
.Machine bad address or instruction register
[cols="^5,^5"]
[options="header",grid="rows"]
|=======================
| Trap cause | `mtval` content
| misaligned instruction fetch address or instruction fetch access fault | address of faulting instruction fetch
| breakpoint | program counter (= address) of faulting instruction itself
| misaligned load address, load access fault, misaligned store address or store access fault | program counter (= address) of faulting instruction itself
| illegal instruction | actual instruction word of faulting instruction
| anything else including interrupts | _0x00000000_ (always zero)
|=======================
[IMPORTAN]
The NEORV32 `mtval` CSR is read-only. However, a write access will _NOT_ raise an illegal instruction exception.
:sectnums!:
===== **`mip`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x344 | **Machine interrupt Pending** | `mip`
3+| Reset value: _0x00000000_
3+| The `mip` CSR is compatible to the RISC-V specifications and also provides custom extensions. It shows currently _pending_ interrupts.
The bits for the standard RISC-V interrupts are read-only. Hence, these interrupts cannot be cleared using the `mip` register and must
be cleared/acknowledged within the according interrupt-generating device.
The upper 16 bits represent the status of the CPU's fast interrupt request lines (FIRQ). Once triggered, these _have to be cleared_ again by setting
the according `mip` bit in the interrupt handler routine to clear the current interrupt request.
|======
.Machine interrupt pending register
[cols="^1,<3,^1,<5"]
[options="header",grid="rows"]
|=======================
| Bit | Name [C] | R/W | Function
| 31:16 | _CSR_MIP_FIRQ15P_ : _CSR_MIP_FIRQ0P_ | r/w | fast interrupt channel 15..0 pending; cleared request by writing 1
| 11 | _CSR_MIP_MEIP_ | r/- | machine _external_ interrupt pending; _cleared by user-defined mechanism_
| 7 | _CSR_MIP_MTIP_ | r/- | machine _timer_ interrupt pending; cleared by incrementing MTIME's time compare register
| 3 | _CSR_MIP_MSIP_ | r/- | machine _software_ interrupt pending; _cleared by user-defined mechanism_
|=======================
<<<
// ####################################################################################################################
:sectnums:
==== Machine Physical Memory Protection CSRs
The available physical memory protection logic is configured via the _PMP_NUM_REGIONS_ and
_PMP_MIN_GRANULARITY_ top entity generics. _PMP_NUM_REGIONS_ defines the number of implemented
protection regions and thus, the availability of the according `pmpcfg*` and `pmpaddr*` CSRs.
[TIP]
If trying to access an PMP-related CSR beyond _PMP_NUM_REGIONS_ **no illegal instruction
exception** is triggered. The according CSRs are read-only (writes are ignored) and always return zero.
[IMPORTANT]
The RISC-V-compatible NEORV32 physical memory protection only implements the _NAPOT_
(naturally aligned power-of-two region) mode with a minimal region granularity of 8 bytes.
:sectnums!:
===== **`pmpcfg`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x3a0 - 0x3af| **Physical memory protection configuration registers** | `pmpcfg0` - `pmpcfg15`
3+| Reset value: _0x00000000_
3+| The `pmpcfg*` CSRs are compatible to the RISC-V specifications. They are used to configure the protected
regions, where each `pmpcfg*` CSR provides configuration bits for four regions. The following bits (for the
first PMP configuration entry) are implemented (all remaining bits are always zero and are read-only):
|======
.Physical memory protection configuration register entry
[cols="^1,^3,^1,<11"]
[options="header",grid="rows"]
|=======================
| Bit | RISC-V name | R/W | Function
| 7 | _L_ | r/w | lock bit, can be set - but not be cleared again (only via CPU reset)
| 6:5 | - | r/- | reserved, read as zero
| 4:3 | _A_ | r/w | mode configuration; only OFF (`00`) and NAPOT (`11`) are supported
| 2 | _X_ | r/w | execute permission
| 1 | _W_ | r/w | write permission
| 0 | _R_ | r/w | read permission
|=======================
:sectnums!:
===== **`pmpaddr`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x3b0 - 0x3ef| **Physical memory protection configuration registers** | `pmpaddr0` - `pmpaddr63`
3+| Reset value: _UNDEFINED_
3+| The `pmpaddr*` CSRs are compatible to the RISC-V specifications. They are used to configure the base
address and the region size.
|======
[NOTE]
When configuring PMP make sure to set `pmpaddr*` before activating the according region via
`pmpcfg*`. When changing the PMP configuration, deactivate the according region via `pmpcfg*`
before modifying `pmpaddr*`.
<<<
// ####################################################################################################################
:sectnums:
==== (Machine) Counter and Timer CSRs
The (machine) counters and timers are implemented when the `Zicntr` ISA extensions is enabled (default)
via the <<_cpu_extension_riscv_zicntr>> generic.
[NOTE]
The <<_cpu_cnt_width>> generic defines the total size of the CPU's <<_cycleh>> and <<_instreth>>
/ <<_mcycleh>> and <<_minstreth>>
counter CSRs (low and high words combined); the time CSRs are not affected by this generic. Note that any
configuration with <<_cpu_cnt_width>> less than 64 is not RISC-V compliant.
.Effective CPU counter width (`[m]cycle` & `[m]instret`)
[IMPORTANT]
If _CPU_CNT_WIDTH_ is less than 64 (the default value) and greater than or equal 32, the according
MSBs of `[m]cycleh` and `[m]instreth` are read-only and always read as zero. This configuration
will also set the _SYSINFO_CPU_ZXSCNT_ flag ("small counters") in the `CPU`
<<_system_configuration_information_memory_sysinfo, SYSINFO>> register. +
+
If _CPU_CNT_WIDTH_ is less than 32 and greater than 0, the `[m]cycleh` and `[m]instreth` CSRs are hardwired to zero
and any write access to them is ignored. Furthermore, the according MSBs of `[m]cycle` and `[m]instret` are read-only
and always read as zero. This configuration will also set the _SYSINFO_CPU_ZXSCNT_ flag ("small counters") in
the `CPU` <<_system_configuration_information_memory_sysinfo, SYSINFO>> register. +
+
If _CPU_CNT_WIDTH_ is 0, the <<_cycleh>> and <<_instreth>> / <<_mcycleh>> and <<_minstreth>> CSRs are hardwired to zero
and any write access to them is ignored.
:sectnums!:
===== **`cycle[h]`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xc00 | **Cycle counter - low word** | `cycle`
| 0xc80 | **Cycle counter - high word** | `cycleh`
3+| Reset value: _UNDEFINED_
3+| The `cycle[h]` CSR is compatible to the RISC-V specifications. It shows the lower/upper 32-bit of the 64-bit cycle
counter. The `cycle[h]` CSR is a read-only shadowed copy of the `mcycle[h]` CSR.
|======
:sectnums!:
===== **`time[h]`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xc01 | **System time - low word** | `time`
| 0xc81 | **System time - high word** | `timeh`
3+| Reset value: _UNDEFINED_
3+| The `time[h]` CSR is compatible to the RISC-V specifications. It shows the lower/upper 32-bit of the 64-bit system
time. The system time is either generated by the processor-internal _MTIME_ system timer unit (if _IO_MTIME_EN_ = _true_) or can be provided by an
external timer unit via the processor's `mtime_i` signal (if _IO_MTIME_EN_ = _false_).
CSR is read-only. Change the system time via the _MTIME_ unit.
|======
:sectnums!:
===== **`instret[h]`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xc02 | **Instructions-retired counter - low word** | `instret`
| 0xc82 | **Instructions-retired counter - high word** | `instreth`
3+| Reset value: _UNDEFINED_
3+| The `instret[h]` CSR is compatible to the RISC-V specifications. It shows the lower/upper 32-bit of the 64-bit retired
instructions counter. The `instret[h]` CSR is a read-only shadowed copy of the `minstret[h]` CSR.
|======
:sectnums!:
===== **`mcycle[h]`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xb00 | **Machine cycle counter - low word** | `mcycle`
| 0xb80 | **Machine cycle counter - high word** | `mcycleh`
3+| Reset value: _UNDEFINED_
3+| The `mcycle[h]` CSR is compatible to the RISC-V specifications. It shows the lower/upper 32-bit of the 64-bit cycle
counter. The `mcycle[h]` CSR can also be written when in machine mode and is copied to the `cycle[h]` CSR.
|======
:sectnums!:
===== **`minstret[h]`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xb02 | **Machine instructions-retired counter - low word** | `minstret`
| 0xb82 | **Machine instructions-retired counter - high word** | `minstreth`
3+| Reset value: _UNDEFINED_
3+| The `minstret[h]` CSR is compatible to the RISC-V specifications. It shows the lower/upper 32-bit of the 64-bit retired
instructions counter. The `minstret[h]` CSR also be written when in machine mode and is copied to the `instret[h]` CSR.
|======
<<<
// ####################################################################################################################
:sectnums:
==== Hardware Performance Monitors (HPM) CSRs
The hardware performance monitor CSRs are implemented when the `Zihpm` ISA extension is enabled via the
<<_cpu_extension_riscv_zihpm>> generic.
The actually implemented hardware performance logic is configured via the <<_hpm_num_cnts>> top entity generic,
which defines the number of implemented performance monitors. Note that always all 28 HPM counter and configuration registers
(`mhpmcounter*[h]` and `mhpmevent*`) are implemented, but only the actually configured ones are real registers and
not hardwired to zero.
[TIP]
If trying to access an HPM-related CSR beyond <<_hpm_num_cnts>> **no illegal instruction exception is
triggered**. The according CSRs are read-only (writes are ignored) and always return zero.
[IMPORTANT]
The HPM system only allows machine-mode access. Hence, `hpmcounter*[h]` CSR are not implemented
and any access (even) from machine mode will raise an exception. Furthermore, the according bits of <<_mcounteren>>
used to configure user-mode access to `hpmcounter*[h]` are hard-wired to zero.
The total counter width of the HPMs can be configured before synthesis via the <<_hpm_cnt_width>> generic (0..64-bit).
[NOTE]
The total LSB-aligned HPM counter size (low word CSR + high word CSR) is defined via the
<<_hpm_num_cnts>> generic (0..64-bit). If <<_hpm_num_cnts>> is less than 64, all unused MSB-aligned
bits are hardwired to zero.
:sectnums!:
===== **`mhpmevent`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x232 -0x33f | **Machine hardware performance monitor event selector** | `mhpmevent3` - `mhpmevent31`
3+| Reset value: _UNDEFINED_
3+| The `mhpmevent*` CSRs are compatible to the RISC-V specifications. The configuration of these CSR define
the architectural events that cause the according `mhpmcounter*[h]` counters to increment. All available events are
listed in the table below. If more than one event is selected, the according counter will increment if any of
the enabled events is observed (logical OR). Note that the counter will only increment by 1 step per clock
cycle even if more than one event is observed. If the CPU is in sleep mode, no HPM counter will increment
at all.
|======
The available hardware performance logic is configured via the _HPM_NUM_CNTS_ top entity generic.
_HPM_NUM_CNTS_ defines the number of implemented performance monitors and thus, the availability of the
according `mhpmcounter*[h]` and `mhpmevent*` CSRs.
.HPM event selector
[cols="^1,<3,^1,<5"]
[options="header",grid="rows"]
|=======================
| Bit | Name [C] | R/W | Event
| 0 | _HPMCNT_EVENT_CY_ | r/w | active clock cycle (not in sleep)
| 1 | - | r/- | _not implemented, always read as zero_
| 2 | _HPMCNT_EVENT_IR_ | r/w | retired instruction
| 3 | _HPMCNT_EVENT_CIR_ | r/w | retired compressed instruction
| 4 | _HPMCNT_EVENT_WAIT_IF_ | r/w | instruction fetch memory wait cycle (if more than 1 cycle memory latency)
| 5 | _HPMCNT_EVENT_WAIT_II_ | r/w | instruction issue pipeline wait cycle (if more than 1 cycle latency), caused by pipelines flushes (like taken branches)
| 6 | _HPMCNT_EVENT_WAIT_MC_ | r/w | multi-cycle ALU operation wait cycle
| 7 | _HPMCNT_EVENT_LOAD_ | r/w | load operation
| 8 | _HPMCNT_EVENT_STORE_ | r/w | store operation
| 9 | _HPMCNT_EVENT_WAIT_LS_ | r/w | load/store memory wait cycle (if more than 1 cycle memory latency)
| 10 | _HPMCNT_EVENT_JUMP_ | r/w | unconditional jump
| 11 | _HPMCNT_EVENT_BRANCH_ | r/w | conditional branch (taken or not taken)
| 12 | _HPMCNT_EVENT_TBRANCH_ | r/w | taken conditional branch
| 13 | _HPMCNT_EVENT_TRAP_ | r/w | entered trap
| 14 | _HPMCNT_EVENT_ILLEGAL_ | r/w | illegal instruction exception
|=======================
:sectnums!:
===== **`mhpmcounter[h]`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xb03 - 0xb1f | **Machine hardware performance monitor - counter low** | `mhpmcounter3` - `mhpmcounter31`
| 0xb83 - 0xb9f | **Machine hardware performance monitor - counter high** | `mhpmcounter3h` - `mhpmcounter31h`
3+| Reset value: _UNDEFINED_
3+| The `mhpmcounter*[h]` CSRs are compatible to the RISC-V specifications. These CSRs provide the lower/upper 32-
bit of arbitrary event counters. The event(s) that trigger an increment of theses counters are selected via the according
`mhpmevent*` CSRs bits.
|======
<<<
// ####################################################################################################################
:sectnums:
==== Machine Counter Setup CSRs
:sectnums!:
===== **`mcountinhibit`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0x320 | **Machine counter-inhibit register** | `mcountinhibit`
3+| Reset value: _UNDEFINED_
3+| The `mcountinhibit` CSR is compatible to the RISC-V specifications. The bits in this register define which
counter/timer CSR are allowed to perform an automatic increment. Automatic update is enabled if the
according bit in `mcountinhibit` is cleared. The following bits are implemented (all remaining bits are
always zero and are read-only).
|======
.Machine counter-inhibit register
[cols="^1,<3,^1,<5"]
[options="header",grid="rows"]
|=======================
| Bit | Name [C] | R/W | Event
| 0 | _CSR_MCOUNTINHIBIT_IR_ | r/w | the `[m]instret[h]` CSRs will auto-increment with each committed instruction when set
| 2 | _CSR_MCOUNTINHIBIT_IR_ | r/w | the `[m]cycle[h]` CSRs will auto-increment with each clock cycle (if CPU is not in sleep state) when set
| 3:31 | _CSR_MCOUNTINHIBIT_HPM3_ _: _CSR_MCOUNTINHIBIT_HPM31_ | r/w | the `mhpmcount*[h]` CSRs will auto-increment according to the configured `mhpmevent*` selector
|=======================
<<<
// ####################################################################################################################
:sectnums:
==== Machine Information CSRs
[NOTE]
All machine information registers can only be accessed in machine mode and are read-only.
:sectnums!:
===== **`mvendorid`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xf11 | **Machine vendor ID** | `mvendorid`
3+| Reset value: _0x00000000_
3+| The `mvendorid` CSR is compatible to the RISC-V specifications. It is read-only and always reads zero.
|======
:sectnums!:
===== **`marchid`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xf12 | **Machine architecture ID** | `marchid`
3+| Reset value: _0x00000013_
3+| The `marchid` CSR is compatible to the RISC-V specifications. It is read-only and shows the NEORV32
official _RISC-V open-source architecture ID_ (decimal: 19, 32-bit hexadecimal: 0x00000013).
|======
:sectnums!:
===== **`mimpid`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xf13 | **Machine implementation ID** | `mimpid`
3+| Reset value: _HW version number_
3+| The `mimpid` CSR is compatible to the RISC-V specifications. It is read-only and shows the version of the
NEORV32 as BCD-coded number (example: `mimpid` = _0x01020312_ → 01.02.03.12 → version 1.2.3.12).
|======
:sectnums!:
===== **`mhartid`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xf14 | **Machine hardware thread ID** | `mhartid`
3+| Reset value: _HW_THREAD_ID_ generic
3+| The `mhartid` CSR is compatible to the RISC-V specifications. It is read-only and shows the core's hart ID,
which is assigned via the CPU's _HW_THREAD_ID_ generic.
|======
:sectnums!:
===== **`mconfigptr`**
[cols="4,27,>7"]
[frame="topbot",grid="none"]
|======
| 0xf15 | **Machine configuration pointer register** | `mconfigptr`
3+| Reset value: `0x00000000`
3+| This register holds a physical address (if not zero) that points to the base address of an architecture configuration structure.
Software can traverse this data structure to discover information about the harts, the platform, and their configuration.
**NOTE: Not assigned yet.**
|======
Go to most recent revision | Compare with Previous | Blame | View Log