foundry

#!/usr/bin/env bash
# Boot a Foundry OS image headless on the pinned Rustbelt kernel and check
# the self-test's output on the serial console. Two boot contracts
# (docs/boot-contract.md), picked with CONTRACT=v1|v2 (default v1):
#
#   v1: the kernel runs the shell self-test (B:/scripts/selftest.sh from
#       rootfs/) itself. Our image is disk 0, the kernel's three test disks
#       C:, D:, E: (snapshot: nothing is written back). The expected lines
#       are the kernel repo's own (moved from its scripts/run-headless.sh),
#       so a pass means "same behaviour as the kernel's own image".
#   v2: our /sbin/init runs /scripts/selftest.sh (from root/) and ends the
#       run with its exit status. Only our image is attached (A:, B:, C:
#       are all partitions of it), WITHOUT snapshot, so the caller can
#       check afterwards what was written (make test-v2 boots a copy).
#       The expected lines are all in EXTRA_EXPECT (scripts/expect-v2.txt).
#
# Usage: [CONTRACT=v2] scripts/run-headless.sh IMAGE KERNEL_BOOTIMAGE KERNEL_TARGET OUT_DIR
#   IMAGE             our disk image
#   KERNEL_BOOTIMAGE  out/kernel/bootimage-headless.bin
#   KERNEL_TARGET     out/kernel-target (has the kernel-owned test disks
#                     ext2.img, exfat.img, btrfs.img: drives C:, D:, E:)
#   OUT_DIR           where logs go (out/)
# Env (the Makefile sets these from config.mk; they change with the kernel pin):
#   KERNEL_BANNER   the kernel's first line, e.g. "Rustbelt 0.21.0 (stage 21: ...)"
#   KERNEL_VERSION  e.g. 0.21.0
#   BIN_COUNT       how many programs the image has in B:/bin (`dir bin | wc -l`)
#   ROOT_BIN_COUNT  v2: how many programs are in A:/bin (@ROOT_BIN_COUNT@)
#   EXTRA_EXPECT    optional file: more lines the log must contain, one per
#                   line (Foundry OS's own checks; # starts a comment)
#   QEMU_CPU (default max), QEMU_TIMEOUT (default 120 s).
#
# Only the QEMU this script starts is ever killed: QEMU writes its PID to
# OUT_DIR/qemu-headless.pid and the exit trap kills exactly that PID.
set -euo pipefail
cd "$(dirname "$0")/.."            # repo root: QEMU's guestfwd runs scripts/http-reply.sh
IMAGE=$1 IMG=$2 KT=$3 OUT=$4
CONTRACT=${CONTRACT:-v1}
: "${KERNEL_BANNER:?set KERNEL_BANNER (see config.mk)}" "${KERNEL_VERSION:?}" "${BIN_COUNT:?}"
STAGE=${KERNEL_BANNER#*(}; STAGE=${STAGE%%:*}     # "stage 20"
mkdir -p "$OUT"
if [ "$CONTRACT" = v2 ]; then
    DISKS=(-drive file="$IMAGE",format=raw,if=virtio)
else
    DISKS=(-drive file="$IMAGE",format=raw,if=virtio,snapshot=on)
    for d in ext2 exfat btrfs; do
        DISKS+=(-drive file="$KT/$d.img",format=raw,if=virtio,snapshot=on)
    done
fi
for f in "$IMAGE" "$IMG"; do
    [ -f "$f" ] || { echo "[run-headless] missing $f (run make image / make kernel)"; exit 2; }
done
for d in "${DISKS[@]}"; do
    case $d in file=*) f=${d#file=}; f=${f%%,*}
        [ -f "$f" ] || { echo "[run-headless] missing $f (run make image / make kernel)"; exit 2; } ;; esac
done

PIDFILE=$OUT/qemu-headless.pid
rm -f "$PIDFILE"
cleanup() {
    if [ -s "$PIDFILE" ] && kill -0 "$(cat "$PIDFILE")" 2>/dev/null; then
        kill "$(cat "$PIDFILE")" 2>/dev/null || true
    fi
    [ -n "${CURL_PID:-}" ] && kill "$CURL_PID" 2>/dev/null || true
}
trap cleanup EXIT

# Network: QEMU's user-mode NAT. guestfwd runs a tiny test web server for
# 10.0.2.100:80 (and `cat` as an echo server on :7); hostfwd forwards a free
# host port to the guest's port 80, where the self-test's `httpd -n 1`
# waits, and a background curl below fetches its page.
HTTP_PORT=$(python3 -c 'import socket; s=socket.socket(); s.bind(("127.0.0.1", 0)); print(s.getsockname()[1])')
NETDEV="user,id=net0,guestfwd=tcp:10.0.2.100:80-cmd:sh scripts/http-reply.sh,guestfwd=tcp:10.0.2.100:7-cmd:cat"
NETDEV="$NETDEV,hostfwd=tcp:127.0.0.1:$HTTP_PORT-:80"
CURL_OUT=$OUT/headless-curl.txt
rm -f "$CURL_OUT"
( for _ in $(seq 1 80); do
      sleep 1
      curl -s -m 5 -o "$CURL_OUT" "http://127.0.0.1:$HTTP_PORT/" 2>/dev/null && [ -s "$CURL_OUT" ] && break
  done ) &
CURL_PID=$!

# virtio-gpu, virtio-rng, and a virtio console (hvc0) fed from
# scripts/hvc0-session.txt; its output goes to OUT/headless-hvc0.txt.
HVC_OUT=$OUT/headless-hvc0.txt
rm -f "$HVC_OUT"
VIRTIO_DEVICES=(-vga none -device virtio-vga -device virtio-rng-pci
    -device virtio-serial-pci -device virtconsole,chardev=hvc
    -chardev "file,id=hvc,path=$HVC_OUT,input-path=scripts/hvc0-session.txt")

LOG=$OUT/headless-serial.log
echo "[run-headless] kernel $IMG"
echo "[run-headless] contract $CONTRACT, disk 0 = $IMAGE"
set +e
timeout "${QEMU_TIMEOUT:-120}" qemu-system-x86_64 -smp 4 -cpu "${QEMU_CPU:-max}" \
    -pidfile "$PIDFILE" \
    -drive format=raw,file="$IMG",snapshot=on \
    "${DISKS[@]}" \
    -netdev "$NETDEV" -device virtio-net-pci,netdev=net0 \
    "${VIRTIO_DEVICES[@]}" \
    -device isa-debug-exit,iobase=0xf4,iosize=0x04 \
    -serial stdio -display none -no-reboot | tee "$LOG"
code=${PIPESTATUS[0]}
kill $CURL_PID 2>/dev/null; wait $CURL_PID 2>/dev/null
set -e

# Lines the shell self-test must have printed (same list as the kernel repo at the pinned commit).
if [ "$code" = 33 ] && [ "$CONTRACT" = v2 ]; then
    expect=()   # all in EXTRA_EXPECT
    # Pipe results: `dir bin | wc -l` on B: -> $BIN_COUNT, the todo.txt grep -> 1,
    # `cat C:/seq.txt | wc -l` (downloaded by wget) -> 20000.
    PIPE_RESULTS=("$BIN_COUNT" 1 20000)
    # init's own memory check: frames before the self-test == after it and
    # every process it started (orphans included) were gone.
    FRAMES_RE='^\[init\] memory: ([0-9]+) frames in use before the self-test, \1 after'
elif [ "$code" = 33 ]; then
    PIPE_RESULTS=(1 "$BIN_COUNT" 3 4 20000)
    FRAMES_RE='^\[memory\] ([0-9]+) frames in use before the shell, \1 after'
    expect=(
        "$KERNEL_BANNER x86_64"                       # ver
        "[smp] 4 CPUs online"                          # smp.rs started 3 more cores
        "[fpu] "                                       # XSAVE (-cpu max) or FXSAVE (qemu64)
        "B:  /mnt/disk0   fat32"                       # drives
        "C:  /mnt/disk1   ext2"
        "D:  /mnt/disk2   exfat"
        "E:  /mnt/disk3   btrfs"
        "B:/ $ ls"                                     # cd B: changed the prompt
        "readme.txt  scripts/"                         # ls in columns
        "In the Rustbelt shell this disk is drive B:"  # type B:/readme.txt
        "second line"                                  # >> and <
        "       125  todo.txt"                         # cp + ls -l
        '[trace 18] open("/mnt/disk0/notes/todo.txt", 0o0) = 3'   # run --trace
        "sleep started in the background"              # start
        "[done] pid 19 (start sleep 1), exit code 0"   # wait
        "C:/ $ ls"                                     # the ext2 disk
        "Hello from the ext2 disk!"                    # cat through a symlink
        "big.txt line 12345"                           # double-indirect block
        "written by Rustbelt"                          # mkdir + create on ext2
        "-rw-r--r--        125  todo.txt"              # cp B: -> C:, ls -l
        "lrwxrwxrwx             hello-link -> hello.txt"
        "drwxr-xr-x      <DIR>  made-here/"
        "D:/ $ ls"                                     # the exFAT disk
        "This file has a long name with spaces in it." # quoted name, other case
        "MixedCase.TXT: exFAT keeps the case"          # cp D: -> C:, cat
        "hi from exFAT"                                # md + echo >, other case
        "       125  todo.txt"                         # cp B: -> D:, ls -l D:/
        "     <DIR>  New Folder/"
        "    340000  numbers.txt"
        "E:/ $ ls -l"                                  # the btrfs disk
        "lrwxrwxrwx             hello-link -> hello.txt"
        "Hello from btrfs! (an inline extent"          # inline, through a symlink
        "btrfs never overwrites a block in place"      # compressed inline extent
        " 120000  360000 2160000 big.txt"              # 17 zlib extents
        "btrfs line 119999"
        "Stage 10: a read-only btrfs driver"           # inside subvolume projects
        "sh: new.txt: Read-only file system"           # writes refused
        "rm: hello.txt: Read-only file system"
        "mkdir: newdir: Read-only file system"
        "[exit code 139]"                              # run evil
        "(overflow) killed: stack overflow"            # user stack guard zone
        "4 CPUs online (this program ran on CPU"       # sysinfo
        "  CPU3 ["
        "[net] DHCP lease: 10.0.2.15/24, gateway 10.0.2.2, DNS 10.0.2.3" # net/
        "      inet 10.0.2.15/24  gateway 10.0.2.2  DNS 10.0.2.3  (DHCP)" # ifconfig
        "64 bytes from 10.0.2.2: icmp_seq=3"           # ping the slirp gateway
        "3 packets transmitted, 3 received, 0% packet loss"
        "Hello from the host, via QEMU guestfwd!"      # fetch (TCP + HTTP)
        "fetch: HTTP/1.0 200 OK, saved 108894 bytes to /mnt/disk2/seq.txt" # wget -o D:
        "[httpd] served 1 request, bye"                # httpd, reached via hostfwd
        "[virtio] rng: virtio-rng at 00:04.0"          # stage 13: virtio/
        "[virtio] hvc0: virtio-console at 00:05.0"
        "[virtio] gpu: virtio-gpu 2D at 00:03.0, display 1280x800, framebuffer 1024x768"
        "[virtio] fbcon: text console on the framebuffer, 128x48 characters"
        "crw-rw----      <DEV>  fb0"                   # ls -l /dev (Linux modes since stage 19)
        "crw-------      <DEV>  hvc0"
        "tile: 4 windows on 1024x768, console 61x46"   # the tiling demo
        "console: full screen, 128x48 characters"      # tile -r
        "[forktest] all checks passed"                 # stage 14: processes
        "[forktest args] all checks passed"            # exec'd with argv/envp/auxv
        "[cowtest] all checks passed"
        "[mmaptest] all checks passed"
        "[sigtest] all checks passed"
        "(start spin 1000), exit code 143"             # kill %1 (SIGTERM)
        "15) SIGTERM"                                  # kill -l
        "[threads] all checks passed"                  # stage 15: libc plumbing
        "[echoserver] all checks passed"
        "[ptytest] all checks passed"
        "[nbtest] all checks passed"
        "[fputest] all checks passed"                  # stage 15.1: musl prerequisites
        "[pathtest] all checks passed"
        "rawabi: uname: Rustbelt $KERNEL_VERSION"      # a no-libc gcc binary
        "rawabi: all checks passed"
        "sh: nosuchprogram: command not found"
        "cd: there is no drive Q:"
        "A:/tmp $ pwd"
    )
fi
if [ "$code" = 33 ]; then
    if [ -n "${EXTRA_EXPECT:-}" ]; then
        while IFS= read -r line; do
            case $line in ''|'#'*) continue ;; esac
            line=${line//@KERNEL_VERSION@/$KERNEL_VERSION}
            line=${line//@BIN_COUNT@/$BIN_COUNT}
            line=${line//@ROOT_BIN_COUNT@/${ROOT_BIN_COUNT:-}}
            line=${line//@KERNEL_BANNER@/$KERNEL_BANNER}
            expect+=("$line")
        done < "$EXTRA_EXPECT"
    fi
    missing=0
    for line in "${expect[@]}"; do
        grep -qF -- "$line" "$LOG" || { echo "[run-headless] MISSING: $line"; missing=1; }
    done
    # Pipe results (v1): `... | grep -i EXT | wc -l` -> 1, `dir bin | wc -l` -> $BIN_COUNT (the programs),
    # `cat C:/chunks.txt | wc -l` (copied from E:) -> 3,
    # `type D:/COW.TXT | wc -l` (copied from E:) -> 4, and
    # `cat D:/seq.txt | wc -l` (downloaded by wget) -> 20000. v2: see above.
    for n in "${PIPE_RESULTS[@]}"; do
        grep -qx -- "$n" "$LOG" || { echo "[run-headless] MISSING pipe result: $n"; missing=1; }
    done
    # No memory leaks: once the shell and all ~60 programs it ran have
    # exited, exactly as many frames must be in use as before it started.
    if grep -qE "$FRAMES_RE" "$LOG"; then
        echo "[run-headless] frames in use returned to the baseline"
    else
        echo "[run-headless] MISSING: frames back at the baseline after the shell"; missing=1
    fi
    # The other direction: the host fetched a page from httpd in the guest.
    if grep -qF "Hello from Rustbelt! Rustbelt $KERNEL_VERSION" "$CURL_OUT" 2>/dev/null; then
        echo "[run-headless] host curl got httpd's page: $(head -1 "$CURL_OUT")"
    else
        echo "[run-headless] MISSING: host curl through hostfwd to httpd"; missing=1
    fi
    # random -n 8: 16 hex digits; random -d 6 5: five numbers 1-6.
    grep -qxE -- "[0-9a-f]{16}" "$LOG" || { echo "[run-headless] MISSING: random -n 8 output"; missing=1; }
    grep -qxE -- "[1-6]( [1-6]){4}" "$LOG" || { echo "[run-headless] MISSING: random -d 6 5 output"; missing=1; }
    # hvc0 is a tty now: its output turns "\n" into "\r\n" (ONLCR), so
    # drop the '\r's before comparing whole lines.
    tr -d '\r' < "$HVC_OUT" > "$HVC_OUT.txt" 2>/dev/null && mv "$HVC_OUT.txt" "$HVC_OUT"
    # The shell on hvc0 ran the host's commands and answered over hvc0.
    # Foundry OS's getty greets differently from the kernel repo's own getty
    # (test-reference passes GETTY_GREETING="Rustbelt getty: ...").
    for line in "${GETTY_GREETING:-Foundry OS getty: shell on this terminal}" "Rustbelt $KERNEL_VERSION ($STAGE" \
                "$ echo hi over hvc0" "hi over hvc0" "fb0" "hvc0"; do
        grep -qF -- "$line" "$HVC_OUT" 2>/dev/null || { echo "[run-headless] MISSING on hvc0: $line"; missing=1; }
    done
    grep -qxE -- "[0-9]+ [0-9]+ [0-9]+" "$HVC_OUT" || { echo "[run-headless] MISSING on hvc0: random -d 20 3"; missing=1; }
    if [ "$CONTRACT" = v2 ]; then
        # init adopted and reaped at least one orphan (c/orphan + v2-test.sh make some).
        if grep -qE '^\[init\] self-test finished: exit code 0; init reaped [1-9][0-9]* adopted' "$LOG"; then
            echo "[run-headless] $(grep -E '^\[init\] self-test finished' "$LOG")"
        else
            echo "[run-headless] MISSING: init reaped >= 1 orphan and the self-test exited 0"; missing=1
        fi
    fi
    [ $missing = 0 ] && echo "[run-headless] all ${#expect[@]}+12 expected shell and hvc0 lines found" || code=35
fi

# isa-debug-exit: QEMU exit status = (value << 1) | 1, so Success(0x10) -> 33.
case $code in
    33)  echo "[run-headless] kernel exited OK"; exit 0 ;;
    35)  echo "[run-headless] kernel reported FAILURE"; exit 1 ;;
    124) echo "[run-headless] timed out (kernel hung?)"; exit 124 ;;
    *)   echo "[run-headless] QEMU exited with $code"; exit "$code" ;;
esac