path: root/tools/kgit-s2q

#!/bin/bash
# SPDX-License-Identifier: GPL-2.0-only
#
# kgit-s2q
#
# Manage a queue of patches described by a series
#
# Based on:
#    - LTSI helper script to generate a git tree from the quilt LTSI tree
#    - git quilt-import
#    - guilt (helper routines)
#
# Copyright 2013-2016 Bruce Ashfield
#
# Licensed under the GPLv2 only.
#
#

usage()
{
cat <<EOF

 kgit-s2q [--author <author name> ] [-n] [--patches <dir>] [-a] [-i] [-f]
          [--commit-sha <type>] [--current] [--series] [--annotated] [--gen] [-c <count>] [-v] [-h] -- [<patch>]"

    --series:     dump the series file and exit
    --gen:        generate patch headers if missing
    --author:     patch author to use if missing, in the format of: "First Last <foo@bar.com>"
    --patches:    directory containing the series file to use
    --current:    show top currently applied patch
    --annotated:  the series is annotated, use the annotations for auto resume
    --fuzzy:      match autoresume on fuzzy heuristics (i.e. subject only)
    --clean:      remove any patch artifacts that may remain from previous runs (i.e. sentinel files)
    --prev:       reset state to the previous patch (retry last patch, failed or not)
    --commit-sha: type of commit sha to create: "author" or "applied". 'author' uses the date of the
                  patch author, and 'applied' uses the date (i.e. current time) of when the patch is applied.
                  The default is 'applied'
    -c:           number of commits to test when autoresuming
    -a:           autoresume series after last sucessfully applied patch
    -i:           interactively prompt if commit information is missing
    -v:           verbose
    -n:           dry run
    -f:           force the first patch of the series as the starting point
    -h:           help

    --:           marks the end of options on the command line

    <patch>:      patch to use as the starting point (turns autoresume off)

EOF
}

silent()
{
	"$@" >/dev/null 2>/dev/null
}

get_branch()
{
	silent git symbolic-ref HEAD || \
		die "Working on a detached HEAD is unsupported."

	git symbolic-ref HEAD | sed -e 's,^refs/heads/,,'
}

next_patch()
{
	COMMIT_START="$1"

	# find the first "real" commit on the branch, i.e. not a merge commit, and not
	# not a commit that came from a merge
	TOP_COMMIT=`git rev-list -n 1 --no-merges --first-parent $COMMIT_START`
	PARENT=`git show $TOP_COMMIT|grep ommit|grep 'pstream\|herry\|tip'|sed 's/.* \([0-9a-f]\+\).*/\1/'`

	CLEN=`echo $PARENT|wc -c`
	if [ $CLEN -eq 41 ] && [ -n "$ANNOTATED_SERIES" ]; then
		# searching via PARENT ID, which is an annotation found in the patch itself
		if [ -z "$FENCE_POST" ]; then
			for i in `tac $DIR/series | grep -v -e ^# -e ^$`; do
				if [ -z "$FENCE_POST" ]; then
					FENCE_POST=`grep -l $PARENT $DIR/$i`
				fi
			done
		fi
	else
		if [ -n "$VERBOSE" ]; then
			echo "[INFO] No annotated resume point (parent ID) falling back to patch id detection"
		fi

		DS1=`mktemp`
		DS2=`mktemp`
		DS3=`mktemp`
		DS4=`mktemp`
		# get the git patch-id of the top commit
		git show $TOP_COMMIT | git patch-id | cut -f1 -d' ' > $DS1

		# and subject
		git format-patch --stdout $TOP_COMMIT~..$TOP_COMMIT  | \
			git mailinfo -b /dev/null /dev/null | grep Subject: | sed 's/Subject: *//' > $DS4

		# find the patch in the series that matches it
		for i in `tac $DIR/series | grep -v -e ^# -e ^$`; do
			git patch-id < $DIR/$i | cut -f1 -d' ' > $DS2
			cmp -s $DS1 $DS2
			if [ $? = 0 ]; then
				FENCE_POST=$i
				break
			fi
		done

		# return if we found a match
		if [ -n "$FENCE_POST" ]; then
			if [ -n "$VERBOSE" ]; then
				echo "[INFO] Matched autoresume via patch ID detection"
			fi
			rm -f $DS1 $DS2 $DS3 $DS4
			return
		fi

		if [ -n "$VERBOSE" ]; then
			echo "[INFO] Could not autodetect resume point via patch ID falling back to diffstat detection"
		fi

		for i in `tac $DIR/series | grep -v -e ^# -e ^$`; do
			cat $DIR/$i | git mailinfo -b /dev/null /dev/null | grep Subject: | sed 's/Subject: *//' > $DS3
			cmp -s $DS3 $DS4
			if [ $? -eq 0 ]; then
				if [ -n "$VERBOSE" ]; then
					echo "[INFO] Found subject match in $i."
				fi
				MAYBE_START=$i
				break
			fi
		done

		# diffstats
		git show $TOP_COMMIT | diffstat -p0 > $DS1
		for i in `tac $DIR/series | grep -v -e ^# -e ^$`; do
			cat $DIR/$i | diffstat -p0 > $DS2
			cmp -s $DS1 $DS2
			if [ $? = 0 ]; then
				# as a second level check, ensure that if diffstats match, subjects
				# should match as well. saves a lot of false positives.
				if [ $i = "$MAYBE_START" ]; then
					FENCE_POST=$i
					if [ -n "$VERBOSE" ]; then
						echo "[INFO] Matched autoresume via diffstat and subject detection" 
					fi
					rm -f $DS1 $DS2 $DS3 $DS4
					return
				fi
			elif [ $i = "$MAYBE_START" ]; then
					FENCE_POST=$i
					# fuzzy match means that we'll let a subject that maches, but not the diffstat
					# be our fencepost. Otherwise, we keep looking.
					if [ -n "$FUZZY_MATCH" ]; then
						if [ -n "$VERBOSE" ]; then
							echo "[WARNING] Subject matched, but diffstat did not. Fuzzy matching for autoresume"
						fi
						rm -f $DS1 $DS2 $DS3 $DS4
						return
					else
						if [ -n "$VERBOSE" ]; then
							echo "[WARNING] Subject matched, but diffstat did not, not autoresuming"
						fi
					fi
			fi
		done

		# if we are here, it means that we couldn't match the top commit to a patch
		# in the series. The search can continue deeper into history, or we can flip
		# the search. Let's take the last entry in the series, and see if its patch-id
		# is on the branch.

		# get the base commit of the branch
		merge_base=`git merge-base HEAD master`
		# dump all of the real commits on the branch from the head commit to the merge base
		# get their patch IDs
		# git rev-list --no-merges --first-parent $merge_base..HEAD | xargs git show | git patch-id  | cut -f1 -d' ' > $DS2
		git rev-list --no-merges $merge_base..HEAD | xargs git show | git patch-id  | cut -f1 -d' ' > $DS2

		# Dump all of the subjects as well
		echo "" > $DS4
		for j in `git rev-list --no-merges $merge_base..HEAD`; do
			git format-patch --stdout $j~..$j  | \
				git mailinfo -b /dev/null /dev/null | grep Subject: | sed 's/Subject: *//' >> $DS4
		done

		for i in `tac $DIR/series | grep -v -e ^# -e ^$`; do
			# grab the patch ID
			cat $DIR/$i | git patch-id | cut -f1 -d' ' > $DS1
		    
			# if our patch ID is in that list, then the series is completely in
			# the branch and is out of sync. Considering warning.
			grep -qF -f $DS1 $DS2
			if [ $? -eq 0 ]; then
				FENCE_POST=$i
				if [ -n "$VERBOSE" ]; then
					echo "[NOTE] series and branch are out of sync"
					echo "[INFO] found the top of series via patch-id in the branch, autoresuming"
					break
				fi
			else
				# we failed to match by patch-id. In case a patch was fuzzed into the tree, double check via
				# subject.
				cat $DIR/$i | git mailinfo -b /dev/null /dev/null | grep Subject: | sed 's/Subject: *//' > $DS3

				grep -qF -f $DS3 $DS4
				if [ $? -eq 0 ]; then
					FENCE_POST=$i
					if [ -n "$VERBOSE" ]; then
						echo "[NOTE] series and branch are out of sync"
						echo "[INFO] found the top of series via subject match in the branch, autoresuming"
					fi
					break
				fi
			fi
		done

		if [ -z "$FENCE_POST" ]; then
			if [ -n "$VERBOSE" ]; then
				echo "[INFO] Could not autodetect resume point via diffstat"
			fi
		else
			if [ -n "$VERBOSE" ]; then
				echo "[INFO] Matched autoresume via diffstat detection"
			fi
		fi

		rm -f $DS1 $DS2 $DS3 $DS4

		return
	fi
}

create_resolution_script()
{
cat <<EOF
#!/bin/bash

  echo "running: git apply --reject .git/rebase-apply/patch"
  git apply --reject .git/rebase-apply/patch > /dev/null
  echo ""
  echo "[INFO] resolving rejects"
  for f in \`git ls-files -o | grep rej | sed 's/\.rej//'\`; do
      echo "[INFO] resolving: \$f"

      rm -f \$f.porig

      echo "   running: wiggle --replace \$f \$f.rej"
      wiggle --replace \$f \$f.rej > /dev/null
      
      echo "[INFO] checking resolution" 
      echo -------------------- original reject -----------------------------
      cat \$f\.rej
      echo --------------------- wiggled in -----------------------------
      diff -up \$f.porig \$f
      echo ------------------------------------------------------------------

      RJSTAT=\`diffstat \$f\.rej |grep changed\`
      echo "Reject  had: "\$RJSTAT

      RFSTAT=\`diff -up \$f.porig \$f | diffstat |grep changed\`
      echo "Refresh has: "\$RFSTAT

      if [ "\$RJSTAT" = "\$RFSTAT" ]; then
          echo -n "Remove porig and reject now? [Y/n/exit] "
          read RESP
          if [ x\$RESP = x -o x\$RESP = xy ] ; then
                rm \$f\.porig \$f\.rej
          elif [ x\$RESP = xexit ] ; then
                 exit 1
          fi
      fi
  done

  git ls-files -m -o -x "meta*" -x "*.rej" -x "*.porig" | xargs git add
  git am --resolved

  echo "[INFO] refresh patch via:"
  echo "  git format-patch -o <original series dir> HEAD^"
EOF
}

AUTORESUME=""
quilt_author=""
commit_sha_type="applied"
while test $# != 0
do
	case "$1" in
	--author)
		shift
		quilt_author="$1"
		;;
	--commit-sha)
                shift
	        commit_sha_type="$1"
	        ;;
	--patches)
		shift
		QUILT_PATCHES="$1"
		;;
	-c)     shift
		DETECT_COUNT="$1"
		;;
	-a)
		AUTORESUME=1
		;;
	-i)
		INTERACTIVE=1
		;;
	-v)
		VERBOSE=1
		;;
	-f)
		FORCE=1
		;;
        --clean)
                CLEAN=1
                ;;
	--prev)
                PREV=1
                ;;
	--gen)
		;;
	--series)
		DUMP_SERIES=1
		;;
	--current)
		DUMP_NEXT_PATCH=1
		;;
	--annotated)
		ANNOTATED_SERIES=1
		;;
        --fuzzy)
		FUZZY_MATCH=1
		;;
	--help|-h)
		usage
		exit
		;;
	--)
		shift
		break
		;;
	esac

	shift
done

# we change directories ourselves
SUBDIRECTORY_OK=1
. "$(git --exec-path)/git-sh-setup"

# if anything is left on the command line, it is a specified starting patch and
# not a fence post
if [ -n "$1" ]; then
	START=$1
	AUTORESUME=
fi


# find the meta directory
get_meta_dir()
{
    # if there's already a located and logged meta directory, just use that and
    # return. Otherwise, we'll look around to see what we can find.
    if [ -f ".metadir" ]; then
        md=`cat .metadir`
        echo $md
        return
    fi

    # determine the meta directory name
    meta_dir_options=`git ls-files -o --directory`
    for m in $meta_dir_options; do
	if [ -d "$m/cfg" ]; then
	    md=`echo $m | sed 's%/%%'`
	fi
    done

    # if nothing was found, we create the minimal structure
    if [ -z "$md" ]; then
        md=".kernel-meta"
    fi

    mkdir -p "${md}/cfg/scratch"

    # store our results where other scripts can quickly find the answer
    echo "${md}" > .metadir
    # return the directory to he caller
    echo $md

    # return the directory to he caller
    echo $md
}

META=`get_meta_dir`
BRANCH=`get_branch`

if [ -n "${CLEAN}" ]; then
    # clean up and then exit
    rm -f $GIT_DIR/kgit-s2q.last
    rm -f $GIT_DIR/kgit-s2q.previous
    exit 0
fi

# for now, this only pops us back to the previous patch, you cannot run it
# twice if you haven't tried to re-apply patches. If we start to track the
# the entire series, or make the .last file "deeper", then we can support
# multiple runs. This is really just a debug assists to restart a patching
# sequence.
if [ -n "${PREV}" ]; then
    if [ -e $GIT_DIR/kgit-s2q.previous ]; then
	cmp -s $GIT_DIR/kgit-s2q.previous $GIT_DIR/kgit-s2q.last
	if [ $? -eq 0 ]; then
	    # previous and last are the same, remove so we restart
	    rm -f $GIT_DIR/kgit-s2q.last
	else
	    # they are different, roll back to previous so we'll
	    # try the same patch again next time around
	    cp -f $GIT_DIR/kgit-s2q.previous $GIT_DIR/kgit-s2q.last
	fi
    else
	rm -f $GIT_DIR/kgit-s2q.last
    fi
    exit 0
fi

# Quilt Author
if [ -n "$quilt_author" ] ; then
	quilt_author_name=$(expr "z$quilt_author" : 'z\(.*[^ ]\) *<.*') &&
	quilt_author_email=$(expr "z$quilt_author" : '.*<\([^>]*\)') &&
	test '' != "$quilt_author_name" &&
	test '' != "$quilt_author_email" ||
	die "malformed --author parameter"
fi

if [ -n "$QUILT_PATCHES" ]; then
	DIR="$QUILT_PATCHES"
else
	DIR="$META/patches/$BRANCH"
fi

if [ ! -d "$DIR" ]; then
	echo ""
	echo "[ERROR] Can't find patch dir at $DIR"
	usage
fi

mkdir -p $META/cfg/scratch || exit 1
REFRESH=$META/cfg/scratch/refresh

SERIES=$DIR/series
if [ ! -f "$SERIES" ]; then
	echo ""
	echo "[ERROR] Can't find series file at $SERIES"
	usage
	exit 1
fi

if [ -n "$DUMP_SERIES" ]; then
	for i in `cat $SERIES`; do
		echo $DIR/$i
	done
	exit 0
fi

if [ -n "$DUMP_NEXT_PATCH" ]; then
	next_patch HEAD

	if [ -z "$FENCE_POST" ] && [ -n "$DETECT_COUNT" ]; then
		count=0
		commit_string="HEAD"
		while [ -z $FENCE_POST ] && [ $count -lt $DETECT_COUNT ]; do
			commit_string=`echo -n $commit_string^`
			count=`expr $count + 1`

			next_patch "$commit_string"
		done
	fi

	if [ -z "$FENCE_POST" ]; then
		cat $SERIES | head -n 1 | sed "s%^%$DIR/%"
	else
		echo "$FENCE_POST"
	fi

	exit 0
fi

diffstat --help > /dev/null 2>&1
if [ $? != 0 ]; then
	echo It appears you dont have diffstat installed.
	echo Please install it.
	exit 1
fi

if [ -n "$AUTORESUME" ] && [ -z "$FENCE_POST" ]; then

	next_patch HEAD

	if [ -z "$FENCE_POST" ] && [ -n "$DETECT_COUNT" ]; then
		count=0
		commit_string="HEAD"
		while [ -z $FENCE_POST ] && [ $count -lt $DETECT_COUNT ]; do
			commit_string=`echo -n $commit_string^`
			count=`expr $count + 1`

			if [ -n "$VERBOSE" ]; then
				echo "[INFO] detection failed, trying previous commit ($commit_string)"
			fi

			next_patch "$commit_string"
		done
	fi
		

	if [ -z "$FENCE_POST" ]; then
		echo "[INFO] Could not autodetect resume point via annotation, patch ID, diffstat or matching filename."
		echo "       Patch that created current HEAD commit $PARENT is unknown. Starting from the first patch."
	else
		echo "[INFO] Resuming from patch after \"$FENCE_POST\""
	fi
fi

# Temporary directories
tmp_dir="$GIT_DIR"/rebase-apply
tmp_msg="$tmp_dir/msg"
tmp_patch="$tmp_dir/patch"
tmp_info="$tmp_dir/info"

COUNT=`cat $SERIES | grep '^[a-zA-Z0-9_]'|wc -l`
APPLIED=0

# Find the intial commit
commit=$(git rev-parse HEAD)

if [ -z "${START}" ]; then
    if  [ -f "$GIT_DIR/kgit-s2q.last" ]; then
        if [ -z "$FORCE" ]; then
            FENCE_POST=`cat $GIT_DIR/kgit-s2q.last`
            echo "[INFO]: fence post found, starting after: ${FENCE_POST}"
        fi
    fi
fi

for i in `cat $SERIES | grep '^[a-zA-Z0-9_/]'`
do
	APPLIED=$[$APPLIED+1]

	# store the last applied patch so we can resume later
	if [ -f $GIT_DIR/kgit-s2q.last ]; then
	    cp -f $GIT_DIR/kgit-s2q.last $GIT_DIR/kgit-s2q.previous
	fi
	# we really should only copy this if the patch PASSES, but this is
	# currently used as a skip type operation, so if we change when this
	# is written, we also need a --skip OR we would need a --retry flag
	# that indicates we should retry failures, versus resuming after them
	# by default.
	echo $i > $GIT_DIR/kgit-s2q.last

	if [ -n "$START" ]; then
		echo $START | grep -qF $i
		if [ $? -ne 0 ]; then
			# if we don't match, skip to the next patch
			continue
		else
			# if we do match, clear our variables so all future patches
			# will be pushed.
			START=""
			FENCE_POST=""
		fi
	else
		if [ -n "$FENCE_POST" ]; then
			# we compare in the series:
			#     links/kernel-cache/ltsi/patches.dma-mapping/...
			# to $FENCE_POST which will have:
			#     meta/patches/standard/base/links/kernel-cache/ltsi/patches.dma-mapping/...
			# so grep for the series sub-path in $FENCE_POST
			echo $FENCE_POST | grep -qF $i
			if [ $? != 0 ]; then
				continue
			else
				FENCE_POST=""
				continue
			fi
		fi
	fi

	if [ ! -f "$DIR/$i" ];then
		echo "[ERROR]: patch $DIR/$i doesn't exist"
		exit 1
	fi

	if [ -n "$VERBOSE" ]; then
		echo "($APPLIED/$COUNT) `basename $i`"
	fi

	git_am_options="--no-verify"
	if [ "$commit_sha_type" == "author" ]; then
		git_am_options="$git_am_options --committer-date-is-author-date"
	fi
	git am $git_am_options --keep-non-patch $DIR/$i > /dev/null 2>&1
	if [ $? != 0 ];then
		git am $git_am_options --abort > /dev/null 2>&1
		echo "[INFO]: check of $DIR/$i with \"git am\" did not pass, trying reduced context."
		git am $git_am_options -C1 --keep-non-patch $DIR/$i > /dev/null 2>&1
		if [ $? = 0 ]; then
			commit=$(git rev-parse HEAD)
			echo $DIR/$i $commit >> $REFRESH
			continue
		fi
		git am $git_am_options --abort > /dev/null 2>&1
		echo "[INFO]: Context reduced git-am of $DIR/$i with \"git am\" did not work, trying \"apply\"."
	else
		commit=$(git rev-parse HEAD)
		continue
	fi

	# git am failed; take manual processing path.
	if [ ! -d "$tmp_dir" ]; then
		mkdir $tmp_dir || exit 2
	fi

	git mailinfo -b "$tmp_msg" "$tmp_patch" \
	    <"$DIR/$i" >"$tmp_info" || exit 3
	test -s "$tmp_patch" || {
		echo "Patch is empty.  Was it split wrong?"
		exit 1
	}

	#  Parse the author information
	GIT_AUTHOR_NAME=$(sed -ne 's/Author: //p' "$tmp_info")
	GIT_AUTHOR_EMAIL=$(sed -ne 's/Email: //p' "$tmp_info")
	export GIT_AUTHOR_NAME GIT_AUTHOR_EMAIL
	while test -z "$GIT_AUTHOR_EMAIL" && test -z "$GIT_AUTHOR_NAME" ; do
		if [ -n "$quilt_author" ] ; then
			GIT_AUTHOR_NAME="$quilt_author_name";
			GIT_AUTHOR_EMAIL="$quilt_author_email";
		else
			if [ -z "$INTERACTIVE" ]; then
				GIT_AUTHOR_NAME="invalid_git config"
				GIT_AUTHOR_EMAIL="<unknown@unknown>"
			else
				echo "No author found in $patch_name" >&2;
				echo "---"
				cat $tmp_msg
				printf "Author: ";
				read patch_author

				echo "$patch_author"

				patch_author_name=$(expr "z$patch_author" : 'z\(.*[^ ]\) *<.*') &&
				patch_author_email=$(expr "z$patch_author" : '.*<\([^>]*\)') &&
				test '' != "$patch_author_name" &&
				test '' != "$patch_author_email" &&
				GIT_AUTHOR_NAME="$patch_author_name" &&
				GIT_AUTHOR_EMAIL="$patch_author_email"
			fi
		fi
	done
	GIT_AUTHOR_DATE=$(sed -ne 's/Date: //p' "$tmp_info")
	SUBJECT=$(sed -ne 's/Subject: //p' "$tmp_info")
	export GIT_AUTHOR_DATE SUBJECT
	if [ -z "$SUBJECT" ] ; then
		SUBJECT=$(echo $patch_name | sed -e 's/.patch$//')
	fi

	# update-ref is what "git am" does.
	git update-ref ORIG_HEAD HEAD
	git apply --index -C1 "$tmp_patch"
	if [ $? != 0 ]; then

	        # put all the pieces that git am requires to complete
	        # the resolution of the reject. Why not just use git am ?
	        # The reason we are in here is because the patch was likely
	        # missing some requirement elements for git am to apply it.
	        # so rather that re-writing the patch (Which could be in a
	        # read only location), we generate what we need and put 
	        # the parts in place, so it can be resolved and not happen
	        # again.
		echo 1 > $tmp_dir/last
		echo 1 > $tmp_dir/next
		echo t > $tmp_dir/utf8

		echo "GIT_AUTHOR_NAME='$GIT_AUTHOR_NAME'"   > $tmp_dir/author-script
		echo "GIT_AUTHOR_EMAIL='$GIT_AUTHOR_EMAIL'" >> $tmp_dir/author-script
		echo "GIT_AUTHOR_DATE='$GIT_AUTHOR_DATE'"  >> $tmp_dir/author-script

		# also copied from git-am.sh
		git rev-parse --verify -q HEAD > $tmp_dir/abort-safety

		touch $tmp_dir/keep
		touch $tmp_dir/scissors
		touch $tmp_dir/no_inbody_headers
		touch $tmp_dir/quiet
		touch $tmp_dir/threeway
		touch $tmp_dir/applying
		touch $tmp_dir/apply-opt
		touch $tmp_dir/sign

		cp $DIR/$i $tmp_dir/0001

		echo "$SUBJECT"  > $tmp_dir/final-commit
		echo >> $tmp_dir/final-commit
		cat "$tmp_msg" >> $tmp_dir/final-commit

		create_resolution_script > $tmp_dir/resolve_rejects
		chmod +x $tmp_dir/resolve_rejects

		echo "[ERROR]: Application of $DIR/$i failed."
		echo "         Patch needs to be refreshed. Sample resolution script:"
		echo "             .git/rebase-apply/resolve_rejects"
		exit 1
	else
		# git apply worked, take next steps.
		tree=$(git write-tree) &&
		commit=$( (echo "$SUBJECT"; echo; cat "$tmp_msg") | git commit-tree $tree -p $commit) &&
		git update-ref -m "kgit-s2q: $patch_name" HEAD $commit || exit 4

		# also set the ORIG_HEAD, since it is used before applying patches
	        git update-ref ORIG_HEAD HEAD

		echo $DIR/$i $commit >> $REFRESH

		GIT_AUTHOR_NAME=""
		GIT_AUTHOR_EMAIL=""
		GIT_AUTHOR_DATE=""
		SUBJECT=""
	fi
done

# if we've failed to locate $FENCE_POST in the series, and
# not subsequently cleared $FENCE_POST, then something evil
# has happened, and we'll do nothing and silently
# return zero which otherwise would be nasty.
if [ -n "$FENCE_POST" ]; then
	echo Failed to locate $FENCE_POST in the series
	exit 1
fi