The header claimed protocol-level crafting of MSG_IO_ERROR is covered by
msg-io-* crafted-server tests; no such test exists (msg-io-timeout-zero
crafts MSG_IO_TIMEOUT). Say what the test actually locks down.
A peer-supplied MSG_IO_ERROR value was OR'd into the local io_error
without masking, allowing a malicious peer to set arbitrary bits that
propagate to exit codes and get re-forwarded to other peers.
Fix: mask the incoming value with IOERR_VALID_MASK (IOERR_GENERAL |
IOERR_VANISHED | IOERR_DEL_LIMIT) before OR'ing.
Test: testsuite/ki62-io-error-mask_test.py
hashtable_create() and the grow path computed the slot-array byte count as
new_array0(char, size * node_size) in 32-bit int arithmetic; for a large
peer/data-driven size the product wrapped to a tiny value, bypassing my_alloc's
--max-alloc guard (which only saw the already-wrapped count), so the table was
under-allocated while tbl->size kept the huge size -- a later node access then
ran out of bounds (heap overflow; ASan-confirmed). Pass size and node_size as
SEPARATE factors so my_alloc checks both before multiplying; guard each *2
doubling against int overflow BEFORE it happens; make HASH_LOAD_LIMIT divide
before multiplying; promote the HT_NODE index multiply to size_t; and test
size < 16 first so a negative req short-circuits the size-1 (INT_MIN UB).
flist_expand()'s int growth math (used+extra and *=4 / *=2 / += FLIST_LINEAR)
could overflow past INT_MAX on a very large file list; guard each operation
before it overflows and refuse rather than under-size the realloc.
(Leonid Bugaev May-2026 re-audit, KI-11/12/13.)
Backport of the 3.5.0 resolver rewrite (drops openat2/RESOLVE_BENEATH for
a portable held-dirfd-stack walk). secure_relative_open() now walks each
path component from the anchor, opening it O_RDONLY|O_DIRECTORY|O_NOFOLLOW
and keeping a stack of the open dirfds: descending a real subdir pushes
its fd, and an in-tree ".." in a followed symlink target pops back to the
already-pinned parent fd rather than re-resolving ".." -- so an ancestor
renamed mid-walk cannot redirect the climb and the walk can never rise
above the anchor. In-tree symlinks are followed (unlike the per-component
O_NOFOLLOW reject), absolute targets are refused, symlink hops are bounded.
This is the substrate the per-subsystem held-dirfd routing hangs off, and
it follows in-tree symlinks while confining -- which openat2 RESOLVE_BENEATH
could not do without breaking legitimate transfers. NOFOLLOW_HIT_SYMLINK()
(ELOOP/EMLINK/EFTYPE) moves to rsync.h. Behaviour-neutral for the existing
daemon gate; the broadening to all non-chroot receivers is a later commit.
Backport of the 3.5.0 wire-input validation:
- reject an out-of-range file index from the wire in read_ndx();
- reject a zero-length checksum block.
Test: checksum-zero-blocklen.
read_del_stats() in main.c accumulates 5 wire-supplied counts into
the int32 stats.deleted_files field:
stats.deleted_files = read_varint_bounded(..., MAX_WIRE_DEL_STAT, ...);
stats.deleted_files += stats.deleted_dirs = ...;
stats.deleted_files += stats.deleted_symlinks = ...;
stats.deleted_files += stats.deleted_devices = ...;
stats.deleted_files += stats.deleted_specials = ...;
With the previous MAX_WIRE_DEL_STAT = 2^30 (1.07 GB) the worst-case
sum is 5 * 2^30 = 5.37 GB; three maximal values already exceed
INT32_MAX = 2.15 GB on the third "+=", triggering signed integer
overflow (C99 6.5/5 -- undefined behaviour, the compiler may assume
it cannot happen and elide subsequent checks).
The bound was introduced in f0155902 ("defence-in-depth: bound
wire-supplied counts and lengths") with a commit message claiming
"per-summand cap so the total can't overflow", but 2^30 * 5 does
overflow. Lower the per-summand cap to 2^28 (= 268M) so the worst
case is 5 * 2^28 = 1.34 GB < INT32_MAX with margin. 2^28 deletions
per category is still vastly above any plausible real transfer.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Multiple receiver-side fields read from the wire were trusted
without upper-bound checks. A hostile peer could either request
extreme allocations (DoS via --max-alloc) or, on platforms where
read_varint returned a negative value, push ~SIZE_MAX through the
size_t conversion to wrap downstream length checks.
Introduce read_int_bounded(), read_varint_bounded() and
read_varint_size() in io.c so wire-derived integer ranges are
checked at the read site rather than scattered across each
caller, with RERR_PROTOCOL on out-of-range input.
Apply the bounded primitives to:
- sum->count (checksum count -- previously could overflow
(size_t)count * xfer_sum_len on 32-bit with raised max-alloc)
- xattrs: count, name_len, datum_len, plus rel_pos overflow
detect to stop chain wrapping the num accumulator
- acls: ida-entry count
- flist: file mode S_IFMT validation, modtime_nsec range check
- delete-stat counters in main: per-summand cap so the total
can't overflow a signed 32-bit accumulator
Reporters include Joshua Rogers (checksum-count overflow finding).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Put sum2_array into sum_struct to hold an array of sum2 checksums
that are each xfer_sum_len bytes.
- Remove sum2 buf from sum_buf.
- Add macro sum2_at() to access each sum2 array element.
- Throw an error if a sums header has an s2length larger than
xfer_sum_len.
- Size flist checksum data to hold the active size, not the max.
- Add a negotiated hash method to the daemon auth code.
- Use EVP for all openssl digests. This makes it easy to add more
openssl digest methods and avoids deprecation warnings.
- Support a way to re-enable deprecated digests via openssl conf
file and allow a default file to be configured.
- Supply a simple openssl-rsync.cnf file to enable legacy digests.
Some systems apparently put strlcpy() into a separate bsd/strings.h file
without putting the function into a separate library. Thus, configure
finds that the function exists for linking but the build does not have
the declaration (which rsync only supplies if it is also supplying its
own version of the function).
- Make the SIMD ASM code off by default. Use configure --enable-simd-asm
to enable.
- Allow MD5 ASM code to be requested even when OpenSSL is handling MD4
checksums. Use configure --enable-md5-asm to enable.
In 2004, an allocation optimization has been added to the file
list handling code, that preallocates 32k of file_struct pointers
in a file_list. This optimization predates the incremental
recursion feature, for which it is not appropriate anymore. When
copying a tree containing a large number of small directories,
using the incremental recursion, rsync allocates many short
file_lists. Suddenly, the unused file_struct pointers can easily
take 90-95% of the memory allocated by rsync.
- use `grep -E` and `grep -F` (`egrep` and `fgrep` are non-standard)
- use same hashbang style for all test scripts
- use explicit comparisons in test scripts
- remove redundant ; from test scripts
- make test script not executable, just like all the other scripts
- unify codestyle across all test scripts
- make openssl license exception clearer by having it at the top
- use modern links in COPYING. The text now matches:
https://www.gnu.org/licenses/gpl-3.0.txt
- fix typo
- Rename unchanged_file() to quick_check_ok().
- Enhance quick_check_ok() to work with non-regular files.
- Add a get_file_type() function to the generator.
- Use the new functions in the generator code to make the logic simpler.
- Fix a bug where the `--alt-dest` functions were not checking if a
special file fully matched the non-permission mode bits before
deciding if we have found an alt-dest match.
- Enhance the `--info=skip --ignore-existing` output to include extra
info on if the existing file differs in type or passes the standard
quick-check logic.
- Add `--info=skip2` that authorizes rsync to perform a slow checksum
"quick check" when ignoring existing files. This provides the uptodate
and differs info even if we need to checksum a file to get it.
- Use C99 flexible arrays when possible, and fall back on the existing
fname[1] kluges on older compilers.
- Avoid static initialization of a flexible array, which is not really
in the C standard.
- All the memory-allocation macros now auto-check for failure and exit
with a failure message that incudes the caller's file and lineno
info. This includes strdup().
- Added the `--max-alloc=SIZE` option to be able to override the memory
allocator's sanity-check limit. It defaults to 1G (as before).
Fixes bugzilla bug 12769.
Andrew Tridgell