build_library: Change extraction script to xz and rewrite to be simpler

This version writes fewer temporary files and tries cpio multiple times
for concatenated archives again.

Signed-off-by: James Le Cuirot <jlecuirot@microsoft.com>

build_library/extract-initramfs-from-vmlinuz.sh

@@ -7,51 +7,35 @@
 # This will create one or more out-dir/rootfs-N directories that contain the contents of the initramfs.
 
 set -euo pipefail
-# check for unzstd. Will abort the script with an error message if the tool is not present.
-unzstd -V >/dev/null
+
+# check for xzcat. Will abort the script with an error message if the tool is not present.
+xzcat -V >/dev/null
+
 fail() {
     echo "${*}" >&2
     exit 1
 }
 
-# Stolen from extract-vmlinux and modified.
-try_decompress() {
-    local header="${1}"
-    local no_idea="${2}"
-    local tool="${3}"
-    local image="${4}"
-    local tmp="${5}"
-    local output_basename="${6}"
-
-    local pos
-    local tool_filename=$(echo "${tool}" | cut -f1 -d' ')
-    # The obscure use of the "tr" filter is to work around older versions of
-    # "grep" that report the byte offset of the line instead of the pattern.
-
-    # Try to find the header and decompress from here.
-    for pos in $(tr "${header}\n${no_idea}" "\n${no_idea}=" < "${image}" |
-                     grep --text --byte-offset --only-matching "^${no_idea}")
-    do
-        pos=${pos%%:*}
-        # Disable error handling, because we will be potentially
-        # giving the tool garbage or a valid archive with some garbage
-        # appended to it. So let the tool extract the valid archive
-        # and then complain about the garbage at the end, but don't
-        # fail the script because of it.
-        set +e; tail "-c+${pos}" "${image}" | "${tool}" >"${tmp}/out" 2>/dev/null; set -e;
-        if [ -s "${tmp}/out" ]; then
-            mv "${tmp}/out" "${output_basename}-${tool_filename}-at-${pos}"
-        else
-            rm -f "${tmp}/out"
-        fi
-    done
+find_xz_headers() {
+    grep --fixed-strings --text --byte-offset --only-matching $'\xFD\x37\x7A\x58\x5A\x00' "$1" | cut -d: -f1
 }
 
-try_unzstd_decompress() {
-    local image="${1}"
-    local tmp="${2}"
-    local output_basename="${3}"
-    try_decompress '(\265/\375' xxx unzstd "${image}" "${tmp}" "${output_basename}"
+decompress_at() {
+    # Data may not really be a valid xz, so allow for errors.
+    tail "-c+$((${2%:*} + 1))" "$1" | xzcat 2>/dev/null || true
+}
+
+try_extract() {
+    # cpio can do strange things when given garbage, so do a basic check.
+    [[ $(head -c6 "$1") == 070701 ]] || return 0
+
+    # There may be multiple concatenated archives so try cpio till it fails.
+    while cpio --quiet --extract --make-directories --directory="${out}/rootfs-${ROOTFS_IDX}" --nonmatching 'dev/*' 2>/dev/null; do
+        ROOTFS_IDX=$(( ROOTFS_IDX + 1 ))
+    done < "$1"
+
+    # Last cpio attempt may or may not leave an empty directory.
+    rmdir "${out}/rootfs-${ROOTFS_IDX}" 2>/dev/null || ROOTFS_IDX=$(( ROOTFS_IDX + 1 ))
 }
 
 me="${0##*/}"
@@ -65,37 +49,22 @@ if [[ ! -s "${image}" ]]; then
 fi
 mkdir -p "${out}"
 
-tmp=$(mktemp --directory /tmp/eifv-XXXXXX)
-trap "rm -rf ${tmp}" EXIT
-
-tmp_dec="${tmp}/decompress"
-mkdir "${tmp_dec}"
-fr_prefix="${tmp}/first-round"
-
+tmp=$(mktemp --directory eifv-XXXXXX)
+trap 'rm -rf -- "${tmp}"' EXIT
 ROOTFS_IDX=0
-perform_round() {
-    local image="${1}"
-    local tmp_dec="${2}"
-    local round_prefix="${3}"
-    try_unzstd_decompress "${image}" "${tmp_dec}" "${round_prefix}"
-    for rnd in "${round_prefix}"*; do
-        if [[ $(file --brief "${rnd}") =~ 'cpio archive' ]]; then
-            mkdir -p "${out}/rootfs-${ROOTFS_IDX}"
-            # On Linux 6.10, the first rootfs is an extra ghost rootfs of 336K, that has a corrupted CPIO
-            cpio --quiet --extract --make-directories --directory="${out}/rootfs-${ROOTFS_IDX}" --nonmatching 'dev/*' < $rnd || true
-            ROOTFS_IDX=$(( ROOTFS_IDX + 1 ))
-        fi
-    done
-}
 
-shopt -s nullglob
-perform_round "${image}" "${tmp_dec}" "${fr_prefix}"
-for fr in "${fr_prefix}"*; do
-    fr_files="${fr}-files"
-    fr_dec="${fr_files}/decompress"
-    mkdir -p "${fr_dec}"
-    sr_prefix="${fr_files}/second-round"
-    perform_round "${fr}" "${fr_dec}" "${sr_prefix}"
+# arm64 kernels are not compressed, so try decompressing once.
+# Other kernels are compressed, so also try decompressing twice.
+for OFF1 in $(find_xz_headers "${image}")
+do
+    decompress_at "${image}" "${OFF1}" > "${tmp}/initrd.maybe_cpio_or_elf"
+    try_extract "${tmp}/initrd.maybe_cpio_or_elf"
+
+    for OFF2 in $(find_xz_headers "${tmp}/initrd.maybe_cpio_or_elf")
+    do
+        decompress_at "${tmp}/initrd.maybe_cpio_or_elf" "${OFF2}" > "${tmp}/initrd.maybe_cpio"
+        try_extract "${tmp}/initrd.maybe_cpio"
+    done
 done
 
 if [[ ${ROOTFS_IDX} -eq 0 ]]; then