systemd/coccinelle/run-coccinelle.sh
Frantisek Sumsal cd4b16c082 coccinelle: drop a couple of FIXMEs
Turns out Coccinelle can handle compound literals just fine, the parsing
errors were caused by incorrectly parsed macros in code before the
literals, so let's just provide simplified versions for such macros.

The parsing error in `Type *foo[ELEMENTSOF(bar)] = {};` is actually
harmless; it occurs only when creating an array of pointers for a type
that's in an external header and it occurs only on the first parser's
pass, subsequent passes resolve the type correctly.

Also, unset ENABLE_DEBUG_HASHMAP, so Coccinelle doesn't expand the
hashmap debug macros.

As for the remaining FIXMEs, I opened a couple of issues in the
Coccinelle upstream to see if they can be fixed there (or at least
properly analyzed).
2024-01-02 19:12:05 +01:00

86 lines
3.2 KiB
Bash
Executable File

#!/usr/bin/env bash
# SPDX-License-Identifier: LGPL-2.1-or-later
set -e
# Exclude following paths from the Coccinelle transformations
EXCLUDED_PATHS=(
"src/boot/efi/*"
"src/shared/linux/*"
"src/basic/linux/*"
# Symlinked to test-bus-vtable-cc.cc, which causes issues with the IN_SET macro
"src/libsystemd/sd-bus/test-bus-vtable.c"
"src/libsystemd/sd-journal/lookup3.c"
# Ignore man examples, as they redefine some macros we use internally, which makes Coccinelle complain
# and ignore code that tries to use the redefined stuff
"man/*"
)
TOP_DIR="$(git rev-parse --show-toplevel)"
CACHE_DIR="$(dirname "$0")/.coccinelle-cache"
ARGS=()
# Create an array from files tracked by git...
mapfile -t FILES < <(git ls-files ':/*.c')
# ...and filter everything that matches patterns from EXCLUDED_PATHS
for excl in "${EXCLUDED_PATHS[@]}"; do
# shellcheck disable=SC2206
FILES=(${FILES[@]//$excl})
done
case "$1" in
-i)
ARGS+=(--in-place)
shift
;;
esac
if ! parallel -h >/dev/null; then
echo 'Please install GNU parallel (package "parallel")'
exit 1
fi
[[ ${#@} -ne 0 ]] && SCRIPTS=("$@") || SCRIPTS=("$TOP_DIR"/coccinelle/*.cocci)
mkdir -p "$CACHE_DIR"
echo "--x-- Using Coccinelle cache directory: $CACHE_DIR"
echo
for script in "${SCRIPTS[@]}"; do
echo "--x-- Processing $script --x--"
TMPFILE="$(mktemp)"
echo "+ spatch --sp-file $script ${ARGS[*]} ..."
# A couple of notes:
#
# 1) Limit this to 10 files at once, as processing the ASTs is _very_ memory hungry - e.g. with 20 files
# at once one spatch process can take around 2.5 GiB of RAM, which can easily eat up all available RAM
# when paired together with parallel
#
# 2) Make sure spatch can find our includes via -I <dir>, similarly as we do when compiling stuff.
# Also, include the system include path as well, since we're not kernel and we make use of the stdlib
# (and other libraries).
#
# 3) Make sure to include includes from includes (--recursive-includes), but use them only to get type
# definitions (--include-headers-for-types) - otherwise we'd start formatting them as well, which might
# be unwanted, especially for includes we fetch verbatim from third-parties
#
# 4) Explicitly undefine the SD_BOOT symbol, so Coccinelle ignores includes guarded by #if SD_BOOT
#
# 5) Use cache, since generating the full AST is expensive. With cache we can do that only once and then
# reuse the cached ASTs for other rules. This cuts down the time needed to run each rule by ~60%.
parallel --halt now,fail=1 --keep-order --noswap --max-args=10 \
spatch --cache-prefix "$CACHE_DIR" \
-I src \
-I /usr/include \
--recursive-includes \
--include-headers-for-types \
--undefined SD_BOOT \
--undefined ENABLE_DEBUG_HASHMAP \
--macro-file-builtins "coccinelle/parsing_hacks.h" \
--smpl-spacing \
--sp-file "$script" \
"${ARGS[@]}" ::: "${FILES[@]}" \
2>"$TMPFILE" || cat "$TMPFILE"
rm -f "$TMPFILE"
echo -e "--x-- Processed $script --x--\n"
done