Commit a6f3b8fd authored by Peter Pentchev's avatar Peter Pentchev

Add Daniel Axtens's security and reliability patches.

Closes:		#916960, #916962, #916963, #916964
Reported by:	Salvatore Bonaccorso <carnil@debian.org>
Origin:		https://github.com/libarchive/libarchive/pull/1105
parent 6e369f5f
libarchive (3.3.3-2) UNRELEASED; urgency=medium
* Add Daniel Axtens's security and reliability patches:
- CVE-2018-1000877.patch: Closes: #916964
- CVE-2018-1000878.patch: Closes: #916963
- CVE-2018-1000879.patch: Closes: #916962
- CVE-2018-1000880.patch: Closes: #916960
- all merged upstream in https://github.com/libarchive/libarchive/pull/1105
Thanks to Salvatore Bonaccorso for filing the Debian bugs!
-- Peter Pentchev <roam@debian.org> Fri, 21 Dec 2018 17:55:09 +0200
libarchive (3.3.3-1) unstable; urgency=medium
[ Peter Pentchev ]
......
Description: Avoid a double-free when a window size of 0 is specified
Origin: upstream; https://github.com/libarchive/libarchive/pull/1105/commits/021efa522ad729ff0f5806c4ce53e4a6cc1daa31
Author: Daniel Axtens <dja@axtens.net>
Last-Updated: 2018-12-21
From 021efa522ad729ff0f5806c4ce53e4a6cc1daa31 Mon Sep 17 00:00:00 2001
From: Daniel Axtens <dja@axtens.net>
Date: Tue, 20 Nov 2018 17:56:29 +1100
Subject: [PATCH 1/4] Avoid a double-free when a window size of 0 is specified
new_size can be 0 with a malicious or corrupted RAR archive.
realloc(area, 0) is equivalent to free(area), so the region would
be free()d here and the free()d again in the cleanup function.
Found with a setup running AFL, afl-rb, and qsym.
---
libarchive/archive_read_support_format_rar.c | 5 +++++
1 file changed, 5 insertions(+)
--- a/libarchive/archive_read_support_format_rar.c
+++ b/libarchive/archive_read_support_format_rar.c
@@ -2300,6 +2300,11 @@
new_size = DICTIONARY_MAX_SIZE;
else
new_size = rar_fls((unsigned int)rar->unp_size) << 1;
+ if (new_size == 0) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
+ "Zero window size is invalid.");
+ return (ARCHIVE_FATAL);
+ }
new_window = realloc(rar->lzss.window, new_size);
if (new_window == NULL) {
archive_set_error(&a->archive, ENOMEM,
Description: rar: file split across multi-part archives must match
Origin: upstream; https://github.com/libarchive/libarchive/pull/1105/commits/bfcfe6f04ed20db2504db8a254d1f40a1d84eb28
Author: Daniel Axtens <dja@axtens.net>
Last-Update: 2018-12-21
From bfcfe6f04ed20db2504db8a254d1f40a1d84eb28 Mon Sep 17 00:00:00 2001
From: Daniel Axtens <dja@axtens.net>
Date: Tue, 4 Dec 2018 00:55:22 +1100
Subject: [PATCH 2/4] rar: file split across multi-part archives must match
Fuzzing uncovered some UAF and memory overrun bugs where a file in a
single file archive reported that it was split across multiple
volumes. This was caused by ppmd7 operations calling
rar_br_fillup. This would invoke rar_read_ahead, which would in some
situations invoke archive_read_format_rar_read_header. That would
check the new file name against the old file name, and if they didn't
match up it would free the ppmd7 buffer and allocate a new
one. However, because the ppmd7 decoder wasn't actually done with the
buffer, it would continue to used the freed buffer. Both reads and
writes to the freed region can be observed.
This is quite tricky to solve: once the buffer has been freed it is
too late, as the ppmd7 decoder functions almost universally assume
success - there's no way for ppmd_read to signal error, nor are there
good ways for functions like Range_Normalise to propagate them. So we
can't detect after the fact that we're in an invalid state - e.g. by
checking rar->cursor, we have to prevent ourselves from ever ending up
there. So, when we are in the dangerous part or rar_read_ahead that
assumes a valid split, we set a flag force read_header to either go
down the path for split files or bail. This means that the ppmd7
decoder keeps a valid buffer and just runs out of data.
Found with a combination of AFL, afl-rb and qsym.
---
libarchive/archive_read_support_format_rar.c | 9 +++++++++
1 file changed, 9 insertions(+)
--- a/libarchive/archive_read_support_format_rar.c
+++ b/libarchive/archive_read_support_format_rar.c
@@ -258,6 +258,7 @@
struct data_block_offsets *dbo;
unsigned int cursor;
unsigned int nodes;
+ char filename_must_match;
/* LZSS members */
struct huffman_code maincode;
@@ -1560,6 +1561,12 @@
}
return ret;
}
+ else if (rar->filename_must_match)
+ {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
+ "Mismatch of file parts split across multi-volume archive");
+ return (ARCHIVE_FATAL);
+ }
rar->filename_save = (char*)realloc(rar->filename_save,
filename_size + 1);
@@ -2933,12 +2940,14 @@
else if (*avail == 0 && rar->main_flags & MHD_VOLUME &&
rar->file_flags & FHD_SPLIT_AFTER)
{
+ rar->filename_must_match = 1;
ret = archive_read_format_rar_read_header(a, a->entry);
if (ret == (ARCHIVE_EOF))
{
rar->has_endarc_header = 1;
ret = archive_read_format_rar_read_header(a, a->entry);
}
+ rar->filename_must_match = 0;
if (ret != (ARCHIVE_OK))
return NULL;
return rar_read_ahead(a, min, avail);
Description: Skip 0-length ACL fields
Origin: upstream; https://github.com/libarchive/libarchive/pull/1105/commits/15bf44fd2c1ad0e3fd87048b3fcc90c4dcff1175
Author: Daniel Axtens <dja@axtens.net>
Last-Update: 2018-12-21
From 15bf44fd2c1ad0e3fd87048b3fcc90c4dcff1175 Mon Sep 17 00:00:00 2001
From: Daniel Axtens <dja@axtens.net>
Date: Tue, 4 Dec 2018 14:29:42 +1100
Subject: [PATCH 3/4] Skip 0-length ACL fields
Currently, it is possible to create an archive that crashes bsdtar
with a malformed ACL:
Program received signal SIGSEGV, Segmentation fault.
archive_acl_from_text_l (acl=<optimised out>, text=0x7e2e92 "", want_type=<optimised out>, sc=<optimised out>) at libarchive/archive_acl.c:1726
1726 switch (*s) {
(gdb) p n
$1 = 1
(gdb) p field[n]
$2 = {start = 0x0, end = 0x0}
Stop this by checking that the length is not zero before beginning
the switch statement.
I am pretty sure this is the bug mentioned in the qsym paper [1],
and I was able to replicate it with a qsym + AFL + afl-rb setup.
[1] https://www.usenix.org/conference/usenixsecurity18/presentation/yun
---
libarchive/archive_acl.c | 5 +++++
1 file changed, 5 insertions(+)
--- a/libarchive/archive_acl.c
+++ b/libarchive/archive_acl.c
@@ -1707,6 +1707,11 @@
st = field[n].start + 1;
len = field[n].end - field[n].start;
+ if (len == 0) {
+ ret = ARCHIVE_WARN;
+ continue;
+ }
+
switch (*s) {
case 'u':
if (len == 1 || (len == 4
Description: warc: consume data once read
Origin: upstream; https://github.com/libarchive/libarchive/pull/1105/commits/9c84b7426660c09c18cc349f6d70b5f8168b5680
Author: Daniel Axtens <dja@axtens.net>
Last-Update: 2018-12-21
From 9c84b7426660c09c18cc349f6d70b5f8168b5680 Mon Sep 17 00:00:00 2001
From: Daniel Axtens <dja@axtens.net>
Date: Tue, 4 Dec 2018 16:33:42 +1100
Subject: [PATCH 4/4] warc: consume data once read
The warc decoder only used read ahead, it wouldn't actually consume
data that had previously been printed. This means that if you specify
an invalid content length, it will just reprint the same data over
and over and over again until it hits the desired length.
This means that a WARC resource with e.g.
Content-Length: 666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666665
but only a few hundred bytes of data, causes a quasi-infinite loop.
Consume data in subsequent calls to _warc_read.
Found with an AFL + afl-rb + qsym setup.
---
libarchive/archive_read_support_format_warc.c | 5 +++++
1 file changed, 5 insertions(+)
--- a/libarchive/archive_read_support_format_warc.c
+++ b/libarchive/archive_read_support_format_warc.c
@@ -386,6 +386,11 @@
return (ARCHIVE_EOF);
}
+ if (w->unconsumed) {
+ __archive_read_consume(a, w->unconsumed);
+ w->unconsumed = 0U;
+ }
+
rab = __archive_read_ahead(a, 1U, &nrd);
if (nrd < 0) {
*bsz = 0U;
CVE-2018-1000877.patch
CVE-2018-1000878.patch
CVE-2018-1000879.patch
CVE-2018-1000880.patch
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment