Thread (15 messages) 15 messages, 6 authors, 2014-07-29

Re: net: socket: NULL ptr deref in sendmsg

From: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Date: 2014-07-26 15:49:00
Also in: lkml
Subsystem: networking [general], the rest · Maintainers: "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds

2014-07-26 2:15 GMT+04:00 Hannes Frederic Sowa [off-list ref]:
Thanks for looking at this! I certainly have overlooked this case.

I wonder, if we allow sendto with valid NULL pointer and positive
msg_namelen to work, why don't we do the same for recvmsg, as in
replacing the VERIFY_WRITE case non-null check with an
access_ok(VERIFY_WRITE, ...) check? We can even get rid of the check
because recvmsg will do it nonetheless when copying the sockaddrs to
user space. So kern_msg->msg_name can always be set to kern_address.
recvmsg's man page are pretty clear about that:

     "If the application does not need to
       know the source address, msg_name can be specified as NULL"

I think we shouldn't change current recvmsg behavior, this could break
some crappy
userpace that doesn't set msg_namelen to zero.

Otherwise I would just set msg_namelen = 0, too, and just not handle
passed in NULL pointers to sockaddrs.
I like that, how about such chage:
diff --git a/net/compat.c b/net/compat.c
index 9a76eaf..bc8aeef 100644
--- a/net/compat.c
+++ b/net/compat.c
@@ -85,7 +85,7 @@ int verify_compat_iovec(struct msghdr *kern_msg,
struct iovec *kern_iov,
 {
        int tot_len;

-       if (kern_msg->msg_namelen) {
+       if (kern_msg->msg_name && kern_msg->msg_namelen) {
                if (mode == VERIFY_READ) {
                        int err = move_addr_to_kernel(kern_msg->msg_name,
                                                      kern_msg->msg_namelen,
@@ -93,10 +93,11 @@ int verify_compat_iovec(struct msghdr *kern_msg,
struct iovec *kern_iov,
                        if (err < 0)
                                return err;
                }
-               if (kern_msg->msg_name)
-                       kern_msg->msg_name = kern_address;
-       } else
+               kern_msg->msg_name = kern_address;
+       } else {
                kern_msg->msg_name = NULL;
+               kern_msg->msg_namelen = 0;
+       }

        tot_len = iov_from_user_compat_to_kern(kern_iov,
                                          (struct compat_iovec __user
*)kern_msg->msg_iov,
diff --git a/net/core/iovec.c b/net/core/iovec.c
index 827dd6b..e1ec45a 100644
--- a/net/core/iovec.c
+++ b/net/core/iovec.c
@@ -39,7 +39,7 @@ int verify_iovec(struct msghdr *m, struct iovec
*iov, struct sockaddr_storage *a
 {
        int size, ct, err;

-       if (m->msg_namelen) {
+       if (m->msg_name && m->msg_namelen) {
                if (mode == VERIFY_READ) {
                        void __user *namep;
                        namep = (void __user __force *) m->msg_name;
@@ -48,10 +48,10 @@ int verify_iovec(struct msghdr *m, struct iovec
*iov, struct sockaddr_storage *a
                        if (err < 0)
                                return err;
                }
-               if (m->msg_name)
-                       m->msg_name = address;
+               m->msg_name = address;
        } else {
                m->msg_name = NULL;
+               m->msg_namelen = 0;
        }

        size = m->msg_iovlen * sizeof(struct iovec);


-- 
Best regards,
Andrey Ryabinin
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help