diff --git a/src/images/bytefield/menvcfgreg.adoc b/src/images/bytefield/menvcfgreg.adoc index 979634a81..62f1d6dd8 100644 --- a/src/images/bytefield/menvcfgreg.adoc +++ b/src/images/bytefield/menvcfgreg.adoc @@ -3,15 +3,16 @@ (defattrs :plain [:plain {:font-family "M+ 1p Fallback"}]) (def row-height 45) (def row-header-fn nil) -(def boxes-per-row 33) -(draw-column-headers {:height 20 :font-size 18 :labels (reverse ["0" "" "1" "3" "4" "5" "6" "" "7" "" "8" "" "" "" "" "" "31" "32" "" "33" "34" "" "" "" "" "" "60" "" "61" "" "62" "" "63"])}) +(def boxes-per-row 34) +(draw-column-headers {:height 20 :font-size 18 :labels (reverse ["0" "" "1" "3" "4" "5" "6" "" "7" "" "8" "" "" "" "" "" "31" "32" "" "33" "34" "" "" "" "" "59" "" "60" "" "61" "" "62" "" "63"])}) (draw-box "STCE" {:span 2}) (draw-box "PBMTE" {:span 2}) (draw-box "ADUE" {:span 2}) +(draw-box "CDE" {:span 2}) (draw-box (text "WPRI" {:font-weight "bold"}) {:span 7}) (draw-box "PMM" {:span 3}) -(draw-box (text "WPRI" {:font-weight "bold"}) {:span 7}) +(draw-box (text "WPRI" {:font-weight "bold"}) {:span 6}) (draw-box "CBZE" {:span 2}) (draw-box "CBCFE" {:span 2}) (draw-box "CBIE" {:span 2}) @@ -21,9 +22,10 @@ (draw-box "1" {:span 2 :borders {}}) (draw-box "1" {:span 2 :borders {}}) (draw-box "1" {:span 2 :borders {}}) -(draw-box "27" {:span 7 :borders {}}) +(draw-box "1" {:span 2 :borders {}}) +(draw-box "26" {:span 7 :borders {}}) (draw-box "2" {:span 3 :borders {}}) -(draw-box "24" {:span 7 :borders {}}) +(draw-box "24" {:span 6 :borders {}}) (draw-box "1" {:span 2 :borders {}}) (draw-box "1" {:span 2 :borders {}}) (draw-box "2" {:span 2 :borders {}}) diff --git a/src/machine.adoc b/src/machine.adoc index 75661e7db..21e408305 100644 --- a/src/machine.adoc +++ b/src/machine.adoc @@ -2074,6 +2074,8 @@ is read-only zero if `menvcfg`.ADUE is zero. NOTE: The Svade extension requires page-fault exceptions be raised when PTE A/D bits need be set, hence Svade is implemented when ADUE=0. +If the Smcdeleg extension is implemented, the CDE (Counter Delegation Enable) bit controls whether Zicntr and Zihpm counters can be delegated to S-mode. When CDE=1, the Smcdeleg extension is enabled, see <>. When CDE=0, the Smcdeleg and Ssccfg extensions appear to be not implemented. If Smcdeleg is not implemented, CDE is read-only zero. + The definition of the STCE field will be furnished by the forthcoming Sstc extension. Its allocation within `menvcfg` may change prior to the ratification of that extension. diff --git a/src/priv-csrs.adoc b/src/priv-csrs.adoc index 43509db6d..85149f22b 100644 --- a/src/priv-csrs.adoc +++ b/src/priv-csrs.adoc @@ -242,6 +242,10 @@ Supervisor counter enable. |`0x10A` |SRW |`senvcfg` |Supervisor environment configuration register. +4+^|Supervisor Counter Setup + +|`0x120` |SRW |`scountinhibit` |Supervisor counter-inhibit register. + 4+^|Supervisor Trap Handling |`0x140` + diff --git a/src/riscv-privileged.adoc b/src/riscv-privileged.adoc index 3d462806c..43eceb554 100644 --- a/src/riscv-privileged.adoc +++ b/src/riscv-privileged.adoc @@ -86,6 +86,7 @@ include::indirect-csr.adoc[] include::smepmp.adoc[] include::smcntrpmf.adoc[] include::rnmi.adoc[] +include::smcdeleg.adoc[] include::supervisor.adoc[] include::sstc.adoc[] include::sscofpmf.adoc[] diff --git a/src/smcdeleg.adoc b/src/smcdeleg.adoc new file mode 100644 index 000000000..ba890f5d9 --- /dev/null +++ b/src/smcdeleg.adoc @@ -0,0 +1,154 @@ +[[smcdeleg]] +== "Smcdeleg" Counter Delegation Extension, Version 1.0.0 + +In modern “Rich OS” environments, hardware performance monitoring +resources are managed by the kernel, kernel driver, and/or hypervisor. +Counters may be configured with differing scopes, in some cases counting +events system-wide, while in others counting events on behalf of a +single virtual machine or application. In such environments, the latency +of counter writes has a direct impact on overall profiling overhead as a +result of frequent counter writes during: + +. Sample collection, to clear overflow indication, and reload overflowed +counter(s) +. Context switch, between processes, threads, containers, or virtual +machines + +This extension provides a means for M-mode to allow writing select +counters and event selectors from S/HS-mode. The purpose is to avert +transitions to and from M-mode that add latency to these performance +critical supervisor/hypervisor code sections. This extension also +defines one new CSR, scountinhibit. + +For a Machine-level environment, extension *Smcdeleg* (‘Sm’ for +Privileged architecture and Machine-level extension, ‘cdeleg’ for +Counter Delegation) encompasses all added CSRs and all behavior +modifications for a hart, over all privilege levels. For a +Supervisor-level environment, extension *Ssccfg* (‘Ss’ for Privileged +architecture and Supervisor-level extension, ‘ccfg’ for Counter +Configuration) provides access to delegated counters, and to new +supervisor-level state. + +=== Counter Delegation + +The `mcounteren` register allows M-mode to provide the next-lower +privilege mode with read access to select counters. When the Smcdeleg/Ssccfg +extension is enabled (`menvcfg`.CDE=1), it further allows M-mode to delegate select +counters to S-mode. + +The `siselect` (and `vsiselect`) index range 0x40-0x5F is reserved for +delegated counter access. When a counter _i_ is delegated +(`mcounteren`[_i_]=1 and `menvcfg`.CDE=1), the register state associated +with counter _i_ can be read or written via `sireg*`, while `siselect` holds +0x40+__i__. The counter state accessible via alias CSRs is shown in +the table below. + +.Indirect HPM State Mappings +[width="100%",cols="21%,20%,21%,18%,20%",options="header",] +|=== +|*`siselect` value* |*`sireg*` |*`sireg4`* |*`sireg2`* |*`sireg5`* +|0x40 |`cycle`^1^ |`cycleh`^1^ |`cyclecfg`^14^ |`cyclecfgh`^14^ +|0x41 4+^|_See below_ +|0x42 |`instret`^1^ |`instreth`^1^ |`instretcfg`^14^ |`instretcfgh`^14^ +|0x43 |`hpmcounter3`^2^ |`hpmcounter3h`^2^ |`hpmevent3`^2^ |`hpmevent3h`^23^ +|… |… |… |… |… +|0x5F |`hpmcounter31`^2^ |`hpmcounter31h`^2^ |`hpmevent31`^2^ |`hpmevent31h`^23^ +|=== + +^1^ Depends on Zicntr support + +^2^ Depends on Zihpm support + +^3^ Depends on Sscofpmf support + +^4^ Depends on Smcntrpmf support + +[NOTE] +==== +`__hpmevent__i` _represents a subset of the state accessed by the_ `__mhpmevent__i` _register. Likewise, `cyclecfg` and `instretcfg` represent a subset of the state accessed by the `mcyclecfg` and `minstretcfg` registers, respectively. See below for subset details._ +==== + +If extension Smstateen is implemented, refer to extension Smcsrind/Sscsrind (<>) for how setting bit 60 of CSR +`mstateen0` to zero prevents access to registers `siselect`, `sireg*`, +`vsiselect`, and `vsireg*` from privileged modes less privileged than +M-mode, and likewise how setting bit 60 of `hstateen0` to zero prevents +access to `siselect` and `sireg*` (really `vsiselect` and `vsireg*`) from +VS-mode. + +The remaining rules of this section apply only when access to a CSR is +not blocked by `mstateen0`[60] = 0 or `hstateen0`[60] = 0. + +While the privilege mode is M or S and `siselect` holds a value in the +range 0x40-0x5F, illegal instruction exceptions are raised for the +following cases: + +* attempts to access any `sireg*` when `menvcfg`.CDE = 0; +* attempts to access `sireg3` or `sireg6`; +* attempts to access `sireg4` or `sireg5` when XLEN = 64; +* attempts to access `sireg*` when `siselect` = 0x41, or when the counter +selected by `siselect` is not delegated to S-mode (the corresponding bit +in `mcounteren` = 0). + +NOTE: _The memory-mapped `mtime` register is not a performance monitoring +counter to be managed by supervisor software, hence the special +treatment of `siselect` value 0x41 described above._ + +For each `siselect` and `sireg*` combination defined in <>, the table +further indicates the extensions upon which the underlying counter state +depends. If any extension upon which the underlying state depends is not +implemented, an attempt from M or S mode to access the given state +through `sireg*` raises an illegal instruction exception. + +If the hypervisor (H) extension is also implemented, then as specified +by extension Smcsrind/Sscsrind, a virtual instruction exception is +raised for attempts from VS-mode or VU-mode to directly access `vsiselect` +or `vsireg*`, or attempts from VU-mode to access `siselect` or `sireg*`. Furthermore, while `vsiselect` holds a value in the range 0x40-0x5F: + +* An attempt to access any `vsireg*` from M or S mode raises an illegal +instruction exception. +* An attempt from VS-mode to access any `sireg*` (really `vsireg*`) raises +either an illegal instruction exception if `menvcfg`.CDE = 0, or a virtual +instruction exception if `menvcfg`.CDE = 1. + +If Sscofpmf is implemented, `sireg2` and `sireg5` provide access only to a +subset of the event selector registers. Specifically, event selector bit +62 (MINH) is read-only 0 when accessed through `sireg*`. Similarly, if +Smcntrpmf is implemented, `sireg2` and `sireg5` provide access only to a +subset of the counter configuration registers. Counter configuration +register bit 62 (MINH) is read-only 0 when accessed through `sireg*`. + +=== Supervisor Counter Inhibit Register (`scountinhibit`) + +Smcdeleg/Ssccfg defines a new `scountinhibit` register, a masked alias of +`mcountinhibit`. For counters delegated to S-mode, the associated +`mcountinhibit` bits can be accessed via `scountinhibit`. For counters not +delegated to S-mode, the associated bits in `scountinhibit` are read-only +zero. + +When `menvcfg`.CDE=0, attempts to access `scountinhibit` raise an illegal +instruction exception. When the Supervisor Counter Delegation extension +is enabled, attempts to access `scountinhibit` from VS-mode or VU-mode +raise a virtual instruction exception. + +=== Virtualizing `scountovf` + +For implementations that support Smcdeleg/Ssccfg, Sscofpmf, and the H +extension, when `menvcfg`.CDE=1, attempts to access `scountovf` from VS-mode +or VU-mode raise a virtual instruction exception. + +=== Virtualizing Local Counter Overflow Interrupts + +For implementations that support Smcdeleg, Smcofpmf, and Smaia, the +local counter overflow interrupt (LCOFI) bit (bit 13) in each of CSRs +`mvip` and `mvien` is implemented and writable. + +For implementations that support Smcdeleg/Ssccfg, Smcofpmf/Sscofpmf, +Smaia/Ssaia, and the H extension, the LCOFI bit (bit 13) in each of `hvip` +and `hvien` is implemented and writable. + +[NOTE] +==== +_The `hvip` register is defined by the hypervisor (H) extension, while the `mvien` and `hvien` registers are defined by the Smaia/Ssaia extension._ + +_By virtue of implementing `hvip`.LCOFI, it is implicit that the LCOFI bit (bit 13) in each of `vsie` and `vsip` is also implemented._ + +_Requiring support for the LCOFI bits listed above ensures that virtual LCOFIs can be delivered to an OS running in S-mode, and to a guest OS running in VS-mode. It is optional whether the LCOFI bit (bit 13) in each of `mideleg` and `hideleg`, which allows all LCOFIs to be delegated to S-mode and VS-mode, respectively, is implemented and writable._ +==== +