ksmbd: Fix ZDI-CAN-18259

This fixes a security problem in ksmbd. It currently has the
ZDI-CAN-18259 ID assigned, but no CVE yet.

Backported from:
8824b7af40
cc4f3b5a6a

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This commit is contained in:
Hauke Mehrtens 2023-01-07 14:41:04 +01:00
parent 2e3d1edf59
commit 76c67fcc66
3 changed files with 100 additions and 1 deletions

View File

@ -2,7 +2,7 @@ include $(TOPDIR)/rules.mk
PKG_NAME:=ksmbd
PKG_VERSION:=3.4.6
PKG_RELEASE:=$(AUTORELEASE)
PKG_RELEASE:=2
PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.gz
PKG_SOURCE_URL:=https://codeload.github.com/cifsd-team/cifsd/tar.gz/$(PKG_VERSION)?

View File

@ -0,0 +1,36 @@
From 8824b7af409f51f1316e92e9887c2fd48c0b26d6 Mon Sep 17 00:00:00 2001
From: William Liu <will@willsroot.io>
Date: Fri, 30 Dec 2022 09:13:35 +0900
Subject: ksmbd: check nt_len to be at least CIFS_ENCPWD_SIZE in
ksmbd_decode_ntlmssp_auth_blob
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
"nt_len - CIFS_ENCPWD_SIZE" is passed directly from
ksmbd_decode_ntlmssp_auth_blob to ksmbd_auth_ntlmv2. Malicious requests
can set nt_len to less than CIFS_ENCPWD_SIZE, which results in a negative
number (or large unsigned value) used for a subsequent memcpy in
ksmbd_auth_ntlvm2 and can cause a panic.
Fixes: e2f3448 ("cifsd: add server-side procedures for SMB3")
Cc: stable@vger.kernel.org
Signed-off-by: William Liu <will@willsroot.io>
Signed-off-by: Hrvoje Mišetić <misetichrvoje@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
---
auth.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
--- a/auth.c
+++ b/auth.c
@@ -583,7 +583,8 @@ int ksmbd_decode_ntlmssp_auth_blob(struc
dn_off = le32_to_cpu(authblob->DomainName.BufferOffset);
dn_len = le16_to_cpu(authblob->DomainName.Length);
- if (blob_len < (u64)dn_off + dn_len || blob_len < (u64)nt_off + nt_len)
+ if (blob_len < (u64)dn_off + dn_len || blob_len < (u64)nt_off + nt_len ||
+ nt_len < CIFS_ENCPWD_SIZE)
return -EINVAL;
#ifdef CONFIG_SMB_INSECURE_SERVER

View File

@ -0,0 +1,63 @@
From cc4f3b5a6ab4693aba94a45cc073188df4d67175 Mon Sep 17 00:00:00 2001
From: Namjae Jeon <linkinjeon@kernel.org>
Date: Mon, 26 Dec 2022 01:28:52 +0900
Subject: ksmbd: fix infinite loop in ksmbd_conn_handler_loop()
If kernel_recvmsg() return -EAGAIN in ksmbd_tcp_readv() and go round
again, It will cause infinite loop issue. And all threads from next
connections would be doing that. This patch add max retry count(2) to
avoid it. kernel_recvmsg() will wait during 7sec timeout and try to
retry two time if -EAGAIN is returned. And add flags of kvmalloc to
__GFP_NOWARN and __GFP_NORETRY to disconnect immediately without
retrying on memory alloation failure.
Fixes: 0626e66 ("cifsd: add server handler for central processing and tranport layers")
Cc: stable@vger.kernel.org
Reported-by: zdi-disclosures@trendmicro.com # ZDI-CAN-18259
Reviewed-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
---
connection.c | 7 +++++--
transport_tcp.c | 5 ++++-
2 files changed, 9 insertions(+), 3 deletions(-)
--- a/connection.c
+++ b/connection.c
@@ -337,9 +337,12 @@ int ksmbd_conn_handler_loop(void *p)
/* 4 for rfc1002 length field */
size = pdu_size + 4;
- conn->request_buf = kvmalloc(size, GFP_KERNEL);
+ conn->request_buf = kvmalloc(size,
+ GFP_KERNEL |
+ __GFP_NOWARN |
+ __GFP_NORETRY);
if (!conn->request_buf)
- continue;
+ break;
memcpy(conn->request_buf, hdr_buf, sizeof(hdr_buf));
if (!ksmbd_smb_request(conn))
--- a/transport_tcp.c
+++ b/transport_tcp.c
@@ -323,6 +323,7 @@ static int ksmbd_tcp_readv(struct tcp_tr
struct msghdr ksmbd_msg;
struct kvec *iov;
struct ksmbd_conn *conn = KSMBD_TRANS(t)->conn;
+ int max_retry = 2;
iov = get_conn_iovec(t, nr_segs);
if (!iov)
@@ -349,9 +350,11 @@ static int ksmbd_tcp_readv(struct tcp_tr
} else if (conn->status == KSMBD_SESS_NEED_RECONNECT) {
total_read = -EAGAIN;
break;
- } else if (length == -ERESTARTSYS || length == -EAGAIN) {
+ } else if ((length == -ERESTARTSYS || length == -EAGAIN) &&
+ max_retry) {
usleep_range(1000, 2000);
length = 0;
+ max_retry--;
continue;
} else if (length <= 0) {
total_read = -EAGAIN;