qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Stefan Hajnoczi <stefanha@redhat.com>
To: qemu-devel@nongnu.org
Cc: Kevin Wolf <kwolf@redhat.com>,
	Peter Maydell <peter.maydell@linaro.org>,
	Stefan Hajnoczi <stefanha@redhat.com>
Subject: [Qemu-devel] [PULL for-2.0 09/51] block/cloop: refuse images with bogus offsets (CVE-2014-0144)
Date: Tue,  1 Apr 2014 19:18:47 +0200	[thread overview]
Message-ID: <1396372769-11688-10-git-send-email-stefanha@redhat.com> (raw)
In-Reply-To: <1396372769-11688-1-git-send-email-stefanha@redhat.com>

The offsets[] array allows efficient seeking and tells us the maximum
compressed data size.  If the offsets are bogus the maximum compressed
data size will be unrealistic.

This could cause g_malloc() to abort and bogus offsets mean the image is
broken anyway.  Therefore we should refuse such images.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 block/cloop.c              | 34 +++++++++++++++++++++++++++++-----
 tests/qemu-iotests/075     | 15 +++++++++++++++
 tests/qemu-iotests/075.out |  8 ++++++++
 3 files changed, 52 insertions(+), 5 deletions(-)

diff --git a/block/cloop.c b/block/cloop.c
index 844665e..55a804f 100644
--- a/block/cloop.c
+++ b/block/cloop.c
@@ -124,12 +124,36 @@ static int cloop_open(BlockDriverState *bs, QDict *options, int flags,
     }
 
     for(i=0;i<s->n_blocks;i++) {
+        uint64_t size;
+
         s->offsets[i] = be64_to_cpu(s->offsets[i]);
-        if (i > 0) {
-            uint32_t size = s->offsets[i] - s->offsets[i - 1];
-            if (size > max_compressed_block_size) {
-                max_compressed_block_size = size;
-            }
+        if (i == 0) {
+            continue;
+        }
+
+        if (s->offsets[i] < s->offsets[i - 1]) {
+            error_setg(errp, "offsets not monotonically increasing at "
+                       "index %u, image file is corrupt", i);
+            ret = -EINVAL;
+            goto fail;
+        }
+
+        size = s->offsets[i] - s->offsets[i - 1];
+
+        /* Compressed blocks should be smaller than the uncompressed block size
+         * but maybe compression performed poorly so the compressed block is
+         * actually bigger.  Clamp down on unrealistic values to prevent
+         * ridiculous s->compressed_block allocation.
+         */
+        if (size > 2 * MAX_BLOCK_SIZE) {
+            error_setg(errp, "invalid compressed block size at index %u, "
+                       "image file is corrupt", i);
+            ret = -EINVAL;
+            goto fail;
+        }
+
+        if (size > max_compressed_block_size) {
+            max_compressed_block_size = size;
         }
     }
 
diff --git a/tests/qemu-iotests/075 b/tests/qemu-iotests/075
index 9c00fa8..d74fb33 100755
--- a/tests/qemu-iotests/075
+++ b/tests/qemu-iotests/075
@@ -44,6 +44,7 @@ _supported_os Linux
 
 block_size_offset=128
 n_blocks_offset=132
+offsets_offset=136
 
 echo
 echo "== check that the first sector can be read =="
@@ -80,6 +81,20 @@ _use_sample_img simple-pattern.cloop.bz2
 poke_file "$TEST_IMG" "$n_blocks_offset" "\x04\x00\x00\x01"
 $QEMU_IO -c "read 0 512" $TEST_IMG 2>&1 | _filter_qemu_io | _filter_testdir
 
+echo
+echo "== refuse images with non-monotonically increasing offsets =="
+_use_sample_img simple-pattern.cloop.bz2
+poke_file "$TEST_IMG" "$offsets_offset" "\x00\x00\x00\x00\xff\xff\xff\xff"
+poke_file "$TEST_IMG" $((offsets_offset + 8)) "\x00\x00\x00\x00\xff\xfe\x00\x00"
+$QEMU_IO -c "read 0 512" $TEST_IMG 2>&1 | _filter_qemu_io | _filter_testdir
+
+echo
+echo "== refuse images with invalid compressed block size =="
+_use_sample_img simple-pattern.cloop.bz2
+poke_file "$TEST_IMG" "$offsets_offset" "\x00\x00\x00\x00\x00\x00\x00\x00"
+poke_file "$TEST_IMG" $((offsets_offset + 8)) "\xff\xff\xff\xff\xff\xff\xff\xff"
+$QEMU_IO -c "read 0 512" $TEST_IMG 2>&1 | _filter_qemu_io | _filter_testdir
+
 # success, all done
 echo "*** done"
 rm -f $seq.full
diff --git a/tests/qemu-iotests/075.out b/tests/qemu-iotests/075.out
index 7cdaee1..911cd3b 100644
--- a/tests/qemu-iotests/075.out
+++ b/tests/qemu-iotests/075.out
@@ -23,4 +23,12 @@ no file open, try 'help open'
 == refuse images that require too many offsets ===
 qemu-io: can't open device TEST_DIR/simple-pattern.cloop: image requires too many offsets, try increasing block size
 no file open, try 'help open'
+
+== refuse images with non-monotonically increasing offsets ==
+qemu-io: can't open device TEST_DIR/simple-pattern.cloop: offsets not monotonically increasing at index 1, image file is corrupt
+no file open, try 'help open'
+
+== refuse images with invalid compressed block size ==
+qemu-io: can't open device TEST_DIR/simple-pattern.cloop: invalid compressed block size at index 1, image file is corrupt
+no file open, try 'help open'
 *** done
-- 
1.9.0

  parent reply	other threads:[~2014-04-01 18:04 UTC|newest]

Thread overview: 53+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-04-01 17:18 [Qemu-devel] [PULL for-2.0 00/51] Block patches Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 01/51] qemu-img: Release reference to BlockDriverState Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 02/51] vvfat: Fix :floppy: option to suppress partition table Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 03/51] qcow2: fix two memory leaks in qcow2_open error code path Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 04/51] qemu-iotests: add ./check -cloop support Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 05/51] qemu-iotests: add cloop input validation tests Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 06/51] block/cloop: validate block_size header field (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 07/51] block/cloop: prevent offsets_size integer overflow (CVE-2014-0143) Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 08/51] block/cloop: refuse images with huge offsets arrays (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:18 ` Stefan Hajnoczi [this message]
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 10/51] block/cloop: fix offsets[] size off-by-one Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 11/51] qemu-iotests: Support for bochs format Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 12/51] bochs: Unify header structs and make them QEMU_PACKED Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 13/51] bochs: Use unsigned variables for offsets and sizes (CVE-2014-0147) Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 14/51] bochs: Check catalog_size header field (CVE-2014-0143) Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 15/51] bochs: Check extent_size header field (CVE-2014-0142) Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 16/51] bochs: Fix bitmap offset calculation Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 17/51] vpc/vhd: add bounds check for max_table_entries and block_size (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 18/51] vpc: Validate block size (CVE-2014-0142) Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 19/51] vdi: add bounds checks for blocks_in_image and disk_size header fields (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 20/51] vhdx: Bounds checking for block_size and logical_sector_size (CVE-2014-0148) Stefan Hajnoczi
2014-04-01 17:18 ` [Qemu-devel] [PULL for-2.0 21/51] curl: check data size before memcpy to local buffer. (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 22/51] qcow2: Check header_length (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 23/51] qcow2: Check backing_file_offset (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 24/51] qcow2: Check refcount table size (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 25/51] qcow2: Validate refcount table offset Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 26/51] qcow2: Validate snapshot table offset/size (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 27/51] qcow2: Validate active L1 table offset and size (CVE-2014-0144) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 28/51] qcow2: Fix backing file name length check Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 29/51] qcow2: Don't rely on free_cluster_index in alloc_refcount_block() (CVE-2014-0147) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 30/51] qcow2: Avoid integer overflow in get_refcount (CVE-2014-0143) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 31/51] qcow2: Check new refcount table size on growth Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 32/51] qcow2: Fix types in qcow2_alloc_clusters and alloc_clusters_noref Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 33/51] qcow2: Protect against some integer overflows in bdrv_check Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 34/51] qcow2: Fix new L1 table size check (CVE-2014-0143) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 35/51] dmg: coding style and indentation cleanup Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 36/51] dmg: prevent out-of-bounds array access on terminator Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 37/51] dmg: drop broken bdrv_pread() loop Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 38/51] dmg: use appropriate types when reading chunks Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 39/51] dmg: sanitize chunk length and sectorcount (CVE-2014-0145) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 40/51] dmg: use uint64_t consistently for sectors and lengths Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 41/51] dmg: prevent chunk buffer overflow (CVE-2014-0145) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 42/51] block: vdi bounds check qemu-io tests Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 43/51] block: Limit request size (CVE-2014-0143) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 44/51] qcow2: Fix copy_sectors() with VM state Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 45/51] qcow2: Fix NULL dereference in qcow2_open() error path (CVE-2014-0146) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 46/51] qcow2: Fix L1 allocation size in qcow2_snapshot_load_tmp() (CVE-2014-0145) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 47/51] qcow2: Check maximum L1 size in qcow2_snapshot_load_tmp() (CVE-2014-0143) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 48/51] qcow2: Limit snapshot table size Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 49/51] parallels: Fix catalog size integer overflow (CVE-2014-0143) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 50/51] parallels: Sanity check for s->tracks (CVE-2014-0142) Stefan Hajnoczi
2014-04-01 17:19 ` [Qemu-devel] [PULL for-2.0 51/51] qcow2: link all L2 meta updates in preallocate() Stefan Hajnoczi
2014-04-01 17:39 ` [Qemu-devel] [PULL for-2.0 00/51] Block patches Peter Maydell

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1396372769-11688-10-git-send-email-stefanha@redhat.com \
    --to=stefanha@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).