busybox: apply patch for CVE-2023-39810

Backport patch referencing this CVE.

Note that the hardening is not activated by default, it adds defconfig
option to enable it.
Since it introduces a breaking change, it shouldn't be enabled in LTS
release by default.
This patch makes busybox cpio equivalent in this release to what is
currently in master and in kirkstone.
Also note that gnu cpio also does not have this hardening, but the CVE
is created only against busybox.

(From OE-Core rev: 3f2b235526d135094408e3895c01bff7b5b938fb)

Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
This commit is contained in:
Peter Marko 2025-07-12 18:08:27 +02:00 committed by Steve Sakoman
parent ae35f06bb1
commit ef6f8d5f46
2 changed files with 137 additions and 0 deletions

View File

@ -0,0 +1,136 @@
From 9a8796436b9b0641e13480811902ea2ac57881d3 Mon Sep 17 00:00:00 2001
From: Denys Vlasenko <vda.linux@googlemail.com>
Date: Wed, 2 Oct 2024 10:12:05 +0200
Subject: [PATCH] archival: disallow path traversals (CVE-2023-39810)
Create new configure option for archival/libarchive based extractions to
disallow path traversals.
As this is a paranoid option and might introduce backward
incompatibility, default it to no.
Fixes: CVE-2023-39810
Based on the patch by Peter Kaestle <peter.kaestle@nokia.com>
function old new delta
data_extract_all 921 945 +24
strip_unsafe_prefix 101 102 +1
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 2/0 up/down: 25/0) Total: 25 bytes
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
CVE: CVE-2023-39810
Upstream-Status: Backport [https://git.busybox.net/busybox/commit/?id=9a8796436b9b0641e13480811902ea2ac57881d3]
Signed-off-by: Peter Marko <peter.marko@siemens.com>
---
archival/Config.src | 11 +++++++++++
archival/libarchive/data_extract_all.c | 8 ++++++++
archival/libarchive/unsafe_prefix.c | 6 +++++-
scripts/kconfig/lxdialog/check-lxdialog.sh | 2 +-
testsuite/cpio.tests | 23 ++++++++++++++++++++++
5 files changed, 48 insertions(+), 2 deletions(-)
diff --git a/archival/Config.src b/archival/Config.src
index 6f4f30c43..cbcd7217c 100644
--- a/archival/Config.src
+++ b/archival/Config.src
@@ -35,4 +35,15 @@ config FEATURE_LZMA_FAST
This option reduces decompression time by about 25% at the cost of
a 1K bigger binary.
+config FEATURE_PATH_TRAVERSAL_PROTECTION
+ bool "Prevent extraction of filenames with /../ path component"
+ default n
+ help
+ busybox tar and unzip remove "PREFIX/../" (if it exists)
+ from extracted names.
+ This option enables this behavior for all other unpacking applets,
+ such as cpio, ar, rpm.
+ GNU cpio 2.15 has NO such sanity check.
+# try other archivers and document their behavior?
+
endmenu
diff --git a/archival/libarchive/data_extract_all.c b/archival/libarchive/data_extract_all.c
index 049c2c156..8a69711c1 100644
--- a/archival/libarchive/data_extract_all.c
+++ b/archival/libarchive/data_extract_all.c
@@ -65,6 +65,14 @@ void FAST_FUNC data_extract_all(archive_handle_t *archive_handle)
} while (--n != 0);
}
#endif
+#if ENABLE_FEATURE_PATH_TRAVERSAL_PROTECTION
+ /* Strip leading "/" and up to last "/../" path component */
+ dst_name = (char *)strip_unsafe_prefix(dst_name);
+#endif
+// ^^^ This may be a problem if some applets do need to extract absolute names.
+// (Probably will need to invent ARCHIVE_ALLOW_UNSAFE_NAME flag).
+// You might think that rpm needs it, but in my tests rpm's internal cpio
+// archive has names like "./usr/bin/FOO", not "/usr/bin/FOO".
if (archive_handle->ah_flags & ARCHIVE_CREATE_LEADING_DIRS) {
char *slash = strrchr(dst_name, '/');
diff --git a/archival/libarchive/unsafe_prefix.c b/archival/libarchive/unsafe_prefix.c
index 33e487bf9..667081195 100644
--- a/archival/libarchive/unsafe_prefix.c
+++ b/archival/libarchive/unsafe_prefix.c
@@ -14,7 +14,11 @@ const char* FAST_FUNC strip_unsafe_prefix(const char *str)
cp++;
continue;
}
- if (is_prefixed_with(cp, "/../"+1)) {
+ /* We are called lots of times.
+ * is_prefixed_with(cp, "../") is slower than open-coding it,
+ * with minimal code growth (~few bytes).
+ */
+ if (cp[0] == '.' && cp[1] == '.' && cp[2] == '/') {
cp += 3;
continue;
}
diff --git a/scripts/kconfig/lxdialog/check-lxdialog.sh b/scripts/kconfig/lxdialog/check-lxdialog.sh
index 5075ebf2d..910ca1f7c 100755
--- a/scripts/kconfig/lxdialog/check-lxdialog.sh
+++ b/scripts/kconfig/lxdialog/check-lxdialog.sh
@@ -55,7 +55,7 @@ trap "rm -f $tmp" 0 1 2 3 15
check() {
$cc -x c - -o $tmp 2>/dev/null <<'EOF'
#include CURSES_LOC
-main() {}
+int main() { return 0; }
EOF
if [ $? != 0 ]; then
echo " *** Unable to find the ncurses libraries or the" 1>&2
diff --git a/testsuite/cpio.tests b/testsuite/cpio.tests
index 85e746589..a4462c53e 100755
--- a/testsuite/cpio.tests
+++ b/testsuite/cpio.tests
@@ -154,6 +154,29 @@ testing "cpio -R with extract" \
" "" ""
SKIP=
+# Create an archive containing a file with "../dont_write" filename.
+# See that it will not be allowed to unpack.
+# NB: GNU cpio 2.15 DOES NOT do such checks.
+optional FEATURE_PATH_TRAVERSAL_PROTECTION
+rm -rf cpio.testdir
+mkdir -p cpio.testdir/prepare/inner
+echo "file outside of destination was written" > cpio.testdir/prepare/dont_write
+echo "data" > cpio.testdir/prepare/inner/to_extract
+mkdir -p cpio.testdir/extract
+testing "cpio extract file outside of destination" "\
+(cd cpio.testdir/prepare/inner && echo -e '../dont_write\nto_extract' | cpio -o -H newc) | (cd cpio.testdir/extract && cpio -vi 2>&1)
+echo \$?
+ls cpio.testdir/dont_write 2>&1" \
+"\
+cpio: removing leading '../' from member names
+../dont_write
+to_extract
+1 blocks
+0
+ls: cpio.testdir/dont_write: No such file or directory
+" "" ""
+SKIP=
+
# Clean up
rm -rf cpio.testdir cpio.testdir2 2>/dev/null

View File

@ -58,6 +58,7 @@ SRC_URI = "https://busybox.net/downloads/busybox-${PV}.tar.bz2;name=tarball \
file://0001-awk.c-fix-CVE-2023-42366-bug-15874.patch \
file://0001-cut-Fix-s-flag-to-omit-blank-lines.patch \
file://CVE-2022-48174.patch \
file://CVE-2023-39810.patch \
"
SRC_URI:append:libc-musl = " file://musl.cfg "
# TODO http://lists.busybox.net/pipermail/busybox/2023-January/090078.html