This is a note to let you know that I've just added the patch titled
libceph: harden msgr2.1 frame segment length checks
to the 6.1-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
libceph-harden-msgr2.1-frame-segment-length-checks.patch
and it can be found in the queue-6.1 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
>From a282a2f10539dce2aa619e71e1817570d557fc97 Mon Sep 17 00:00:00 2001
From: Ilya Dryomov <idryomov@xxxxxxxxx>
Date: Mon, 10 Jul 2023 20:39:29 +0200
Subject: libceph: harden msgr2.1 frame segment length checks
From: Ilya Dryomov <idryomov@xxxxxxxxx>
commit a282a2f10539dce2aa619e71e1817570d557fc97 upstream.
ceph_frame_desc::fd_lens is an int array. decode_preamble() thus
effectively casts u32 -> int but the checks for segment lengths are
written as if on unsigned values. While reading in HELLO or one of the
AUTH frames (before authentication is completed), arithmetic in
head_onwire_len() can get duped by negative ctrl_len and produce
head_len which is less than CEPH_PREAMBLE_LEN but still positive.
This would lead to a buffer overrun in prepare_read_control() as the
preamble gets copied to the newly allocated buffer of size head_len.
Cc: stable@xxxxxxxxxxxxxxx
Fixes: cd1a677cad99 ("libceph, ceph: implement msgr2.1 protocol (crc and secure modes)")
Reported-by: Thelford Williams <thelford@xxxxxxxxxx>
Signed-off-by: Ilya Dryomov <idryomov@xxxxxxxxx>
Reviewed-by: Xiubo Li <xiubli@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
net/ceph/messenger_v2.c | 41 ++++++++++++++++++++++++++---------------
1 file changed, 26 insertions(+), 15 deletions(-)
--- a/net/ceph/messenger_v2.c
+++ b/net/ceph/messenger_v2.c
@@ -392,6 +392,8 @@ static int head_onwire_len(int ctrl_len,
int head_len;
int rem_len;
+ BUG_ON(ctrl_len < 0 || ctrl_len > CEPH_MSG_MAX_CONTROL_LEN);
+
if (secure) {
head_len = CEPH_PREAMBLE_SECURE_LEN;
if (ctrl_len > CEPH_PREAMBLE_INLINE_LEN) {
@@ -410,6 +412,10 @@ static int head_onwire_len(int ctrl_len,
static int __tail_onwire_len(int front_len, int middle_len, int data_len,
bool secure)
{
+ BUG_ON(front_len < 0 || front_len > CEPH_MSG_MAX_FRONT_LEN ||
+ middle_len < 0 || middle_len > CEPH_MSG_MAX_MIDDLE_LEN ||
+ data_len < 0 || data_len > CEPH_MSG_MAX_DATA_LEN);
+
if (!front_len && !middle_len && !data_len)
return 0;
@@ -522,29 +528,34 @@ static int decode_preamble(void *p, stru
desc->fd_aligns[i] = ceph_decode_16(&p);
}
- /*
- * This would fire for FRAME_TAG_WAIT (it has one empty
- * segment), but we should never get it as client.
- */
- if (!desc->fd_lens[desc->fd_seg_cnt - 1]) {
- pr_err("last segment empty\n");
+ if (desc->fd_lens[0] < 0 ||
+ desc->fd_lens[0] > CEPH_MSG_MAX_CONTROL_LEN) {
+ pr_err("bad control segment length %d\n", desc->fd_lens[0]);
return -EINVAL;
}
-
- if (desc->fd_lens[0] > CEPH_MSG_MAX_CONTROL_LEN) {
- pr_err("control segment too big %d\n", desc->fd_lens[0]);
+ if (desc->fd_lens[1] < 0 ||
+ desc->fd_lens[1] > CEPH_MSG_MAX_FRONT_LEN) {
+ pr_err("bad front segment length %d\n", desc->fd_lens[1]);
return -EINVAL;
}
- if (desc->fd_lens[1] > CEPH_MSG_MAX_FRONT_LEN) {
- pr_err("front segment too big %d\n", desc->fd_lens[1]);
+ if (desc->fd_lens[2] < 0 ||
+ desc->fd_lens[2] > CEPH_MSG_MAX_MIDDLE_LEN) {
+ pr_err("bad middle segment length %d\n", desc->fd_lens[2]);
return -EINVAL;
}
- if (desc->fd_lens[2] > CEPH_MSG_MAX_MIDDLE_LEN) {
- pr_err("middle segment too big %d\n", desc->fd_lens[2]);
+ if (desc->fd_lens[3] < 0 ||
+ desc->fd_lens[3] > CEPH_MSG_MAX_DATA_LEN) {
+ pr_err("bad data segment length %d\n", desc->fd_lens[3]);
return -EINVAL;
}
- if (desc->fd_lens[3] > CEPH_MSG_MAX_DATA_LEN) {
- pr_err("data segment too big %d\n", desc->fd_lens[3]);
+
+ /*
+ * This would fire for FRAME_TAG_WAIT (it has one empty
+ * segment), but we should never get it as client.
+ */
+ if (!desc->fd_lens[desc->fd_seg_cnt - 1]) {
+ pr_err("last segment empty, segment count %d\n",
+ desc->fd_seg_cnt);
return -EINVAL;
}
Patches currently in stable-queue which might be from idryomov@xxxxxxxxx are
queue-6.1/ceph-fix-blindly-expanding-the-readahead-windows.patch
queue-6.1/ceph-don-t-let-check_caps-skip-sending-responses-for-revoke-msgs.patch
queue-6.1/libceph-harden-msgr2.1-frame-segment-length-checks.patch
queue-6.1/ceph-add-a-dedicated-private-data-for-netfs-rreq.patch
