Add support for hashed mode to Poudriere

Creates the repo with hash-based filenames to allow use of a CDN

src/etc/poudriere.conf.sample

@@ -365,3 +365,14 @@ DISTFILES_CACHE=/usr/ports/distfiles
 # be fetched.
 # Default: everything
 #PACKAGE_FETCH_WHITELIST="gcc* rust llvm*"
+
+# Have pkg create the repo such that each package is named with the short hash
+# if its file contents in the package filename, with symlinks to the traditional
+# package filenames. The packagesite.yaml file will point to the hashed version
+# of these files. By using hashed pkg filenames, this allows users to lazily
+# cache packages without conflicting with the existing packages, or serving stale
+# packages from a cache. Once the packages are synced the much
+# smaller meta files can then be synced. Allowing a near atomic update of the repo.
+# On caching CDNs this means a need to purge 2-5 files instead of all pkgs that
+# have been updated.
+#PKG_HASH="no"

src/man/poudriere-bulk.8

@@ -28,7 +28,7 @@
 .\"
 .\" Note: The date here should be updated whenever a non-trivial
 .\" change is made to the manual page.
-.Dd September 9, 2021
+.Dd October 31, 2021
 .Dt POUDRIERE-BULK 8
 .Os
 .Sh NAME
@@ -38,7 +38,7 @@
 .Nm
 .Fl a
 .Fl j Ar name
-.Op Fl CcFIikNnRrSTtvw
+.Op Fl CcFHIikNnRrSTtvw
 .Op Fl b Ar name
 .Op Fl B Ar name
 .Op Fl J Ar maxjobs Ns Op Cm \&: Ns Ar prebuildmaxjobs
@@ -48,7 +48,7 @@
 .Nm
 .Fl f Ar file Op Oo Fl f Ar file2 Oc Ar ...
 .Fl j Ar name
-.Op Fl CcFIikNnRrSTtvw
+.Op Fl CcFHIikNnRrSTtvw
 .Op Fl b Ar name
 .Op Fl B Ar name
 .Op Fl J Ar maxjobs Ns Op Cm \&: Ns Ar prebuildmaxjobs
@@ -57,7 +57,7 @@
 .Op Fl z Ar set
 .Nm
 .Fl j Ar name
-.Op Fl CcFIikNnRrSTtvw
+.Op Fl CcFHIikNnRrSTtvw
 .Op Fl b Ar name
 .Op Fl B Ar name
 .Op Fl J Ar maxjobs Ns Op Cm \&: Ns Ar prebuildmaxjobs
@@ -228,6 +228,8 @@ Fetch only from the original
 Skip
 .Fx
 mirrors.
+.It Fl H
+Create a repository where the package filenames contain the short hash of the contents.
 .It Fl I
 Advanced interactive mode.
 .Pp

src/share/poudriere/bulk.sh

@@ -47,6 +47,8 @@ Options:
                    -f file.  Implies -c for -a.
     -c          -- Clean all the previously built binary packages and logs.
     -F          -- Only fetch from original master_site (skip FreeBSD mirrors)
+    -H          -- Create a repository where the package filenames contain the
+                   short hash of the contents.
     -I          -- Advanced Interactive mode. Leaves jail running with ports
                    installed after test.
     -i          -- Interactive mode. Enter jail for interactive testing and
@@ -98,7 +100,7 @@ COMMIT=1
 
 [ $# -eq 0 ] && usage
 
-while getopts "ab:B:CcFf:iIj:J:knNO:p:RrSTtvwz:" FLAG; do
+while getopts "ab:B:CcFf:HiIj:J:knNO:p:RrSTtvwz:" FLAG; do
 	case "${FLAG}" in
 		a)
 			ALL=1
@@ -126,6 +128,9 @@ while getopts "ab:B:CcFf:iIj:J:knNO:p:RrSTtvwz:" FLAG; do
 			    OPTARG="${SAVED_PWD}/${OPTARG}"
 			LISTPKGS="${LISTPKGS:+${LISTPKGS} }${OPTARG}"
 			;;
+		H)
+			PKG_REPO_FLAGS="${PKG_REPO_FLAGS} --hash --symlink"
+			;;
 		I)
 			INTERACTIVE_MODE=2
 			;;

src/share/poudriere/common.sh

@@ -8211,12 +8211,16 @@ sign_pkg() {
 }
 
 build_repo() {
-	local origin pkg_repo_list_files
+	local origin pkg_repo_list_files hashcmd
 
 	msg "Creating pkg repository"
 	if [ ${DRY_RUN} -eq 1 ]; then
 		return 0
 	fi
+	if [ ${PKG_HASH} != "no" ]; then
+		hashcmd="--hash --symlink"
+		PKG_REPO_FLAGS="${PKG_REPO_FLAGS:+${PKG_REPO_FLAGS} }$hashcmd"
+	fi
 	bset status "pkgrepo:"
 	ensure_pkg_installed force_extract || \
 	    err 1 "Unable to extract pkg."
@@ -8231,11 +8235,19 @@ build_repo() {
 		install -m 0400 "${PKG_REPO_META_FILE}" \
 		    ${MASTERMNT}/tmp/pkgmeta
 	fi
+
+	# Remount rw
+	# mount_nullfs does not support mount -u
+	umount ${UMOUNT_NONBUSY} ${MASTERMNT}/packages || \
+	    umount -f ${MASTERMNT}/packages
+	mount_packages
+
 	mkdir -p ${MASTERMNT}/tmp/packages
 	if [ -n "${PKG_REPO_SIGNING_KEY}" ]; then
 		install -m 0400 ${PKG_REPO_SIGNING_KEY} \
 			${MASTERMNT}/tmp/repo.key
 		injail ${PKG_BIN} repo \
+			${PKG_REPO_FLAGS} \
 			${pkg_repo_list_files} \
 			-o /tmp/packages \
 			${PKG_META} \
@@ -8246,12 +8258,14 @@ build_repo() {
 		# using SSH with DNSSEC as older hosts don't support
 		# it.
 		${MASTERMNT}${PKG_BIN} repo \
+		    ${PKG_REPO_FLAGS} \
 		    ${pkg_repo_list_files} \
 		    -o ${MASTERMNT}/tmp/packages ${PKG_META_MASTERMNT} \
 		    ${MASTERMNT}/packages \
 		    ${SIGNING_COMMAND:+signing_command: ${SIGNING_COMMAND}}
 	else
 		JNETNAME="n" injail ${PKG_BIN} repo \
+		    ${PKG_REPO_FLAGS} \
 		    ${pkg_repo_list_files} \
 		    -o /tmp/packages ${PKG_META} /packages \
 		    ${SIGNING_COMMAND:+signing_command: ${SIGNING_COMMAND}}
@@ -8266,6 +8280,11 @@ build_repo() {
 			sign_pkg pubkey "${PACKAGES}/Latest/pkg.${PKG_EXT}"
 		fi
 	fi
+
+	# Remount ro
+	umount ${UMOUNT_NONBUSY} ${MASTERMNT}/packages || \
+	    umount -f ${MASTERMNT}/packages
+	mount_packages -o ro
 }
 
 calculate_size_in_mb() {
@@ -8784,6 +8803,7 @@ INTERACTIVE_MODE=0
 : ${FLAVOR_DEFAULT_ALL:=no}
 : ${NULLFS_PATHS:="/rescue /usr/share /usr/tests /usr/lib32"}
 : ${PACKAGE_FETCH_URL:="pkg+http://pkg.FreeBSD.org/\${ABI}"}
+: ${PKG_HASH:=no}
 
 : ${POUDRIERE_TMPDIR:=$(command mktemp -dt poudriere)}
 : ${SHASH_VAR_PATH_DEFAULT:=${POUDRIERE_TMPDIR}}

src/share/poudriere/include/pkg.sh

@@ -321,6 +321,9 @@ delete_pkg() {
 	[ $# -ne 1 ] && eargs delete_pkg pkg
 	local pkg="$1"
 
+	# If ${pkg} is a symlink, delete the target as well
+	[ -L "${pkg}" ] && unlink $(realpath "${pkg}")
+
 	# Delete the package and the depsfile since this package is being deleted,
 	# which will force it to be recreated
 	unlink "${pkg}"
@@ -339,6 +342,8 @@ delete_pkg_xargs() {
 	# Delete the package and the depsfile since this package is being deleted,
 	# which will force it to be recreated
 	{
+		# If ${pkg} is a symlink, delete the target as well
+		[ -L "${pkg}" ] && echo $(realpath "${pkg}")
 		echo "${pkg}"
 		echo "${pkg_cache_dir}"
 	} >> "${listfile}"

src/share/poudriere/pkgclean.sh

@@ -183,13 +183,16 @@ for file in ${PACKAGES}/All/*; do
 			pkgname="${pkgname%.*}"
 			if ! pkg_get_origin origin "${file}"; then
 				msg_verbose "Found corrupt package: ${file}"
+				[ -L "${file}" ] && echo "$(realpath ${file})" >> ${BADFILES_LIST}
 				echo "${file}" >> ${BADFILES_LIST}
 			elif shash_remove pkgname-forbidden "${pkgname}" \
 			    forbidden; then
 				msg_verbose "Found forbidden package (${forbidden}): ${file}"
+				[ -L "${file}" ] && echo "$(realpath ${file})" >> ${BADFILES_LIST}
 				echo "${file}" >> ${BADFILES_LIST}
 			elif ! pkgbase_is_needed "${pkgname}"; then
 				msg_verbose "Found unwanted package: ${file}"
+				[ -L "${file}" ] && echo "$(realpath ${file})" >> ${BADFILES_LIST}
 				echo "${file}" >> ${BADFILES_LIST}
 			else
 				echo "${file} ${origin}" >> ${FOUND_ORIGINS}
@@ -208,6 +211,7 @@ for file in ${PACKAGES}/All/*; do
 			;&
 		*)
 			msg_verbose "Found incorrect format file: ${file}"
+			[ -L "${file}" ] && echo "$(realpath ${file})" >> ${BADFILES_LIST}
 			echo "${file}" >> ${BADFILES_LIST}
 			;;
 	esac
@@ -257,6 +261,7 @@ END {
 	for pkg in $packages; do
 		pkgversion="${pkg##*-}"
 		pkgversion="${pkgversion%.*}"
+		pkgversion="${pkgversion%~*}"
 
 		if [ -z "${lastpkg}" ]; then
 			lastpkg="${pkg}"