[extension fast track] extra vector crypto instructions, Zvbc32e/Zvkgs

doc/vector-extra/Makefile

@@ -0,0 +1,67 @@
+# Makefile for RISC-V Doc Template
+#
+# This work is licensed under the Creative Commons Attribution-ShareAlike 4.0
+# International License. To view a copy of this license, visit
+# http://creativecommons.org/licenses/by-sa/4.0/ or send a letter to
+# Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.
+#
+# SPDX-License-Identifier: CC-BY-SA-4.0
+#
+# Description:
+#
+# This Makefile is designed to automate the process of building and packaging
+# the Doc Template for RISC-V Extensions.
+
+DOCKER_RUN := docker run --rm -v ${PWD}:/build -w /build \
+riscvintl/riscv-docs-base-container-image:latest
+HEADER_SOURCE := riscv-crypto-spec-vector-extra.adoc
+PDF_RESULT := riscv-crypto-spec-vector-extra.pdf
+SPEC_COMMIT= git-commit.adoc
+ASCIIDOCTOR_PDF := asciidoctor-pdf
+OPTIONS := --trace \
+           -a compress \
+           -a mathematical-format=svg \
+           -a pdf-fontsdir=resources/fonts \
+           -a pdf-style=resources/themes/risc-v_spec-pdf.yml \
+           -a toc \
+           --failure-level=ERROR
+REQUIRES := --require=asciidoctor-bibtex \
+            --require=asciidoctor-diagram \
+            --require=asciidoctor-mathematical
+
+.PHONY: all build clean build-container build-no-container
+
+all: build
+
+cp_bib:
+	@cp ../riscv-crypto-spec.bib ./
+
+$(SPEC_COMMIT):
+	@git rev-parse --abbrev-ref HEAD > ${@}
+	@echo "@" >> ${@}
+	@git log --pretty=format:'%H' -n 1 >> ${@}
+
+build: cp_bib $(SPEC_COMMIT)
+	@echo "Checking if Docker is available..."
+	@if command -v docker &> /dev/null ; then \
+		echo "Docker is available, building inside Docker container..."; \
+		$(MAKE) build-container; \
+	else \
+		echo "Docker is not available, building without Docker..."; \
+		$(MAKE) build-no-container; \
+	fi
+
+build-container:
+	@echo "Starting build inside Docker container..."
+	$(DOCKER_RUN) /bin/sh -c "$(ASCIIDOCTOR_PDF) $(OPTIONS) $(REQUIRES) --out-file=$(PDF_RESULT) $(HEADER_SOURCE)"
+	@echo "Build completed successfully inside Docker container."
+
+build-no-container:
+	@echo "Starting build..."
+	$(ASCIIDOCTOR_PDF) $(OPTIONS) $(REQUIRES) --out-file=$(PDF_RESULT) $(HEADER_SOURCE)
+	@echo "Build completed successfully."
+
+clean:
+	@echo "Cleaning up generated files..."
+	rm -f $(PDF_RESULT)
+	@echo "Cleanup completed."

doc/vector-extra/insns/vclmul-32e.adoc

@@ -0,0 +1,104 @@
+[[insns-vclmul-32e, Vector Carry-less Multiply]]
+= vclmul.[vv,vx]
+
+Synopsis::
+Vector Carry-less Multiply by vector or scalar - returning low half of product.
+
+Mnemonic::
+vclmul.vv vd, vs2, vs1, vm +
+vclmul.vx vd, vs2, rs1, vm
+
+Encoding (Vector-Vector)::
+[wavedrom, , svg]
+....
+{reg:[
+{bits: 7, name: 'OP-V'},
+{bits: 5, name: 'vd'},
+{bits: 3, name: 'OPMVV'},
+{bits: 5, name: 'vs1'},
+{bits: 5, name: 'vs2'},
+{bits: 1, name: 'vm'},
+{bits: 6, name: '001100'},
+]}
+....
+
+Encoding (Vector-Scalar)::
+[wavedrom, , svg]
+....
+{reg:[
+{bits: 7, name: 'OP-V'},
+{bits: 5, name: 'vd'},
+{bits: 3, name: 'OPMVX'},
+{bits: 5, name: 'rs1'},
+{bits: 5, name: 'vs2'},
+{bits: 1, name: 'vm'},
+{bits: 6, name: '001100'},
+]}
+....
+Reserved Encodings::
+* `SEW` is any value other than 32 (`Zvbc32e` only)
+* `SEW` is any value other than 64 (`Zvbc` only)
+* `SEW` is any value other than 32 or 64 (`Zvbc` and `Zvbc32e`)
+
+Arguments::
+
+[%autowidth]
+[%header,cols="4,2,2"]
+|===
+|Register
+|Direction
+|Definition
+
+| Vs1/Rs1 | input  |  multiplier
+| Vs2 | input  |  multiplicand
+| Vd  | output | carry-less product low
+|===
+
+[NOTE]
+====
+`vclmul` instruction was initially defined in `Zvbc` with only `SEW=64-bit` support, this page describes how the specification is extended in `Zvbc32e` to support `SEW=32 bits`.
+====
+
+Description::
+Produces the low half of 128-bit carry-less product.
+
+Each SEW-bit element in the `vs2` vector register is carry-less multiplied by
+either each SEW-bit element in `vs1` (vector-vector), or the SEW-bit value
+from integer register `rs1` (vector-scalar). The result is the least
+significant SEW bits of the carry-less product.
+
+[NOTE]
+====
+The 32-bit carryless multiply instructions can be used for implementing GCM in the absence of the `zvkg` extension.
+In particular for implementation with `ELEN=32` where `Zvkg` cannot be implemented.
+It can also be used to speed-up CRC evaluation.
+====
+
+Operation::
+[source,sail]
+--
+
+
+function clause execute (VCLMUL(vs2, vs1, vd, suffix)) = {
+
+  foreach (i from vstart to vl-1) {
+    let op1 : bits (SEW) = if suffix =="vv" then get_velem(vs1, i)
+                          else zext_or_truncate_to_sew(X(vs1));
+    let op2 : bits (SEW) = get_velem(vs2, i);
+    let product : bits (SEW) = clmul(op1, op2, SEW);
+    set_velem(vd, i, product);
+  }
+  RETIRE_SUCCESS
+}
+
+function clmul(x, y, width) = {
+  let result : bits(width) = zeros();
+  foreach (i from 0 to (width - 1)) {
+    if y[i] == 1 then result = result ^ (x << i);
+  }
+  result
+}
+--
+
+Included in::
+<<zvbc32e>>

doc/vector-extra/insns/vclmulh-32e.adoc

@@ -0,0 +1,99 @@
+[[insns-vclmulh, Vector Carry-less Multiply Return High Half]]
+= vclmulh.[vv,vx]
+
+Synopsis::
+Vector Carry-less Multiply by vector or scalar - returning high half of product.
+
+Mnemonic::
+vclmulh.vv vd, vs2, vs1, vm +
+vclmulh.vx vd, vs2, rs1, vm
+
+Encoding (Vector-Vector)::
+[wavedrom, , svg]
+....
+{reg:[
+{bits: 7, name: 'OP-V'},
+{bits: 5, name: 'vd'},
+{bits: 3, name: 'OPMVV'},
+{bits: 5, name: 'vs1'},
+{bits: 5, name: 'vs2'},
+{bits: 1, name: 'vm'},
+{bits: 6, name: '001101'},
+]}
+....
+
+Encoding (Vector-Scalar)::
+[wavedrom, , svg]
+....
+{reg:[
+{bits: 7, name: 'OP-V'},
+{bits: 5, name: 'vd'},
+{bits: 3, name: 'OPMVX'},
+{bits: 5, name: 'rs1'},
+{bits: 5, name: 'vs2'},
+{bits: 1, name: 'vm'},
+{bits: 6, name: '001101'},
+]}
+....
+Reserved Encodings::
+* `SEW` is any value other than 64 (`Zvbc` only)
+* `SEW` is any value other than 32 (`Zvbc32e` only)
+* `SEW` is any value other than 32 or 64 (`Zvbc32e` and `Zvbc`)
+
+Arguments::
+
+[%autowidth]
+[%header,cols="4,2,2"]
+|===
+|Register
+|Direction
+|Definition
+
+| Vs1 | input  | multiplier
+| Vs2 | input  | multiplicand
+| Vd  | output | carry-less product high
+|===
+
+[NOTE]
+====
+`vclmulh` instruction was initially defined in `Zvbc`, this page describes how the specification is extended in `Zvbc32e` to support `SEW=32 bits`.
+====
+
+Description::
+Produces the high half of 128-bit carry-less product.
+
+Each SEW-bit element in the `vs2` vector register is carry-less multiplied by
+either each SEW-bit element in `vs1` (vector-vector), or the SEW-bit value
+from integer register `rs1` (vector-scalar). The result is the most
+significant SEW bits of the carry-less product.
+
+// This instruction must always be implemented such that its execution latency does not depend
+// on the data being operated upon.
+
+Operation::
+[source,sail]
+--
+function clause execute (VCLMULH(vs2, vs1, vd, suffix)) = {
+
+  foreach (i from vstart to vl-1) {
+    let op1 : bits (SEW) = if suffix =="vv" then get_velem(vs1,i)
+                          else zext_or_truncate_to_sew(X(vs1));
+    let op2 : bits (SEW) = get_velem(vs2, i);
+    let product : bits (SEW) = clmulh(op1, op2, SEW);
+    set_velem(vd, i, product);
+  }
+  RETIRE_SUCCESS
+}
+
+function clmulh(x, y, width) = {
+  let result : bits(width) = 0;
+  foreach (i from 1 to (width - 1)) {
+    if y[i] == 1 then result = result ^ (x >> (width - i));
+  }
+  result
+}
+
+--
+
+Included in::
+<<zvbc32e>>