[PATCH rds linux-next v4 0/2] WARNING in rds_message_alloc_sgs

[PATCH rds linux-next v4 0/2] WARNING in rds_message_alloc_sgs

[Shamir Rabinovitch]

From: shamir rabinovitch <shamir.rabinovitch@oracle.com>

This patch set fix google syzbot rds bug found in linux-next.
The first patch solve the syzbot issue.
The second patch fix issue mentioned by Leon Romanovsky that
drivers should not call WARN_ON as result from user input.

syzbot bug report can be foud here: https://lkml.org/lkml/2018/10/31/28

v1->v2:
- patch 1: make rds_iov_vector fields name more descriptive (Hakon)
- patch 1: fix potential mem leak in rds_rm_size if krealloc fail
  (Hakon)
v2->v3:
- patch 2: harden rds_sendmsg for invalid number of sgs (Gerd)
v3->v4
- Santosh a.b. on both patches + repost to net-dev

shamir rabinovitch (2):
  net/rds: fix warn in rds_message_alloc_sgs
  net/rds: remove user triggered WARN_ON in rds_sendmsg

 net/rds/message.c | 24 +++++++++++----
 net/rds/rdma.c    | 75 +++++++++++++++++++++++------------------------
 net/rds/rds.h     | 23 +++++++++++----
 net/rds/send.c    | 59 ++++++++++++++++++++++++++++++-------
 4 files changed, 120 insertions(+), 61 deletions(-)

-- 
2.19.1

[PATCH rds linux-next v4 1/2] net/rds: fix warn in rds_message_alloc_sgs

[Shamir Rabinovitch]

From: shamir rabinovitch <shamir.rabinovitch@oracle.com>

redundant copy_from_user in rds_sendmsg system call expose rds
to issue where rds_rdma_extra_size walk the rds iovec and and
calculate the number pf pages (sgs) it need to add to the tail of
rds message and later rds_cmsg_rdma_args copy the rds iovec again
and re calculate the same number and get different result causing
WARN_ON in rds_message_alloc_sgs.

fix this by doing the copy_from_user only once per rds_sendmsg
system call.

When issue occur the below dump is seen:

WARNING: CPU: 0 PID: 19789 at net/rds/message.c:316 rds_message_alloc_sgs+0x10c/0x160 net/rds/message.c:316
Kernel panic - not syncing: panic_on_warn set ...
CPU: 0 PID: 19789 Comm: syz-executor827 Not tainted 4.19.0-next-20181030+ #101
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:77 [inline]
 dump_stack+0x244/0x39d lib/dump_stack.c:113
 panic+0x2ad/0x55c kernel/panic.c:188
 __warn.cold.8+0x20/0x45 kernel/panic.c:540
 report_bug+0x254/0x2d0 lib/bug.c:186
 fixup_bug arch/x86/kernel/traps.c:178 [inline]
 do_error_trap+0x11b/0x200 arch/x86/kernel/traps.c:271
 do_invalid_op+0x36/0x40 arch/x86/kernel/traps.c:290
 invalid_op+0x14/0x20 arch/x86/entry/entry_64.S:969
RIP: 0010:rds_message_alloc_sgs+0x10c/0x160 net/rds/message.c:316
Code: c0 74 04 3c 03 7e 6c 44 01 ab 78 01 00 00 e8 2b 9e 35 fa 4c 89 e0 48 83 c4 08 5b 41 5c 41 5d 41 5e 41 5f 5d c3 e8 14 9e 35 fa <0f> 0b 31 ff 44 89 ee e8 18 9f 35 fa 45 85 ed 75 1b e8 fe 9d 35 fa
RSP: 0018:ffff8801c51b7460 EFLAGS: 00010293
RAX: ffff8801bc412080 RBX: ffff8801d7bf4040 RCX: ffffffff8749c9e6
RDX: 0000000000000000 RSI: ffffffff8749ca5c RDI: 0000000000000004
RBP: ffff8801c51b7490 R08: ffff8801bc412080 R09: ffffed003b5c5b67
R10: ffffed003b5c5b67 R11: ffff8801dae2db3b R12: 0000000000000000
R13: 000000000007165c R14: 000000000007165c R15: 0000000000000005
 rds_cmsg_rdma_args+0x82d/0x1510 net/rds/rdma.c:623
 rds_cmsg_send net/rds/send.c:971 [inline]
 rds_sendmsg+0x19a2/0x3180 net/rds/send.c:1273
 sock_sendmsg_nosec net/socket.c:622 [inline]
 sock_sendmsg+0xd5/0x120 net/socket.c:632
 ___sys_sendmsg+0x7fd/0x930 net/socket.c:2117
 __sys_sendmsg+0x11d/0x280 net/socket.c:2155
 __do_sys_sendmsg net/socket.c:2164 [inline]
 __se_sys_sendmsg net/socket.c:2162 [inline]
 __x64_sys_sendmsg+0x78/0xb0 net/socket.c:2162
 do_syscall_64+0x1b9/0x820 arch/x86/entry/common.c:290
 entry_SYSCALL_64_after_hwframe+0x49/0xbe
RIP: 0033:0x44a859
Code: e8 dc e6 ff ff 48 83 c4 18 c3 0f 1f 80 00 00 00 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 0f 83 6b cb fb ff c3 66 2e 0f 1f 84 00 00 00 00
RSP: 002b:00007f1d4710ada8 EFLAGS: 00000297 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 00000000006dcc28 RCX: 000000000044a859
RDX: 0000000000000000 RSI: 0000000020001600 RDI: 0000000000000003
RBP: 00000000006dcc20 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000297 R12: 00000000006dcc2c
R13: 646e732f7665642f R14: 00007f1d4710b9c0 R15: 00000000006dcd2c
Kernel Offset: disabled
Rebooting in 86400 seconds..

Reported-by: syzbot+26de17458aeda9d305d8@syzkaller.appspotmail.com
Acked-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: shamir rabinovitch <shamir.rabinovitch@oracle.com>
---
 net/rds/rdma.c | 63 +++++++++++++++++++++++++-------------------------
 net/rds/rds.h  | 20 ++++++++++++----
 net/rds/send.c | 50 +++++++++++++++++++++++++++++++++------
 3 files changed, 91 insertions(+), 42 deletions(-)

diff --git a/net/rds/rdma.c b/net/rds/rdma.c
index 98237feb607a..e1965d9cbcf8 100644
--- a/net/rds/rdma.c
+++ b/net/rds/rdma.c
@@ -517,9 +517,10 @@ static int rds_rdma_pages(struct rds_iovec iov[], int nr_iovecs)
 	return tot_pages;
 }
 
-int rds_rdma_extra_size(struct rds_rdma_args *args)
+int rds_rdma_extra_size(struct rds_rdma_args *args,
+			struct rds_iov_vector *iov)
 {
-	struct rds_iovec vec;
+	struct rds_iovec *vec;
 	struct rds_iovec __user *local_vec;
 	int tot_pages = 0;
 	unsigned int nr_pages;
@@ -530,13 +531,23 @@ int rds_rdma_extra_size(struct rds_rdma_args *args)
 	if (args->nr_local == 0)
 		return -EINVAL;
 
+	iov->iov = kcalloc(args->nr_local,
+			   sizeof(struct rds_iovec),
+			   GFP_KERNEL);
+	if (!iov->iov)
+		return -ENOMEM;
+
+	vec = &iov->iov[0];
+
+	if (copy_from_user(vec, local_vec, args->nr_local *
+			   sizeof(struct rds_iovec)))
+		return -EFAULT;
+	iov->len = args->nr_local;
+
 	/* figure out the number of pages in the vector */
-	for (i = 0; i < args->nr_local; i++) {
-		if (copy_from_user(&vec, &local_vec[i],
-				   sizeof(struct rds_iovec)))
-			return -EFAULT;
+	for (i = 0; i < args->nr_local; i++, vec++) {
 
-		nr_pages = rds_pages_in_vec(&vec);
+		nr_pages = rds_pages_in_vec(vec);
 		if (nr_pages == 0)
 			return -EINVAL;
 
@@ -558,15 +569,15 @@ int rds_rdma_extra_size(struct rds_rdma_args *args)
  * Extract all arguments and set up the rdma_op
  */
 int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
-			  struct cmsghdr *cmsg)
+		       struct cmsghdr *cmsg,
+		       struct rds_iov_vector *vec)
 {
 	struct rds_rdma_args *args;
 	struct rm_rdma_op *op = &rm->rdma;
 	int nr_pages;
 	unsigned int nr_bytes;
 	struct page **pages = NULL;
-	struct rds_iovec iovstack[UIO_FASTIOV], *iovs = iovstack;
-	int iov_size;
+	struct rds_iovec *iovs;
 	unsigned int i, j;
 	int ret = 0;
 
@@ -586,31 +597,23 @@ int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
 		goto out_ret;
 	}
 
-	/* Check whether to allocate the iovec area */
-	iov_size = args->nr_local * sizeof(struct rds_iovec);
-	if (args->nr_local > UIO_FASTIOV) {
-		iovs = sock_kmalloc(rds_rs_to_sk(rs), iov_size, GFP_KERNEL);
-		if (!iovs) {
-			ret = -ENOMEM;
-			goto out_ret;
-		}
+	if (vec->len != args->nr_local) {
+		ret = -EINVAL;
+		goto out_ret;
 	}
 
-	if (copy_from_user(iovs, (struct rds_iovec __user *)(unsigned long) args->local_vec_addr, iov_size)) {
-		ret = -EFAULT;
-		goto out;
-	}
+	iovs = vec->iov;
 
 	nr_pages = rds_rdma_pages(iovs, args->nr_local);
 	if (nr_pages < 0) {
 		ret = -EINVAL;
-		goto out;
+		goto out_ret;
 	}
 
 	pages = kcalloc(nr_pages, sizeof(struct page *), GFP_KERNEL);
 	if (!pages) {
 		ret = -ENOMEM;
-		goto out;
+		goto out_ret;
 	}
 
 	op->op_write = !!(args->flags & RDS_RDMA_READWRITE);
@@ -623,7 +626,7 @@ int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
 	op->op_sg = rds_message_alloc_sgs(rm, nr_pages);
 	if (!op->op_sg) {
 		ret = -ENOMEM;
-		goto out;
+		goto out_pages;
 	}
 
 	if (op->op_notify || op->op_recverr) {
@@ -635,7 +638,7 @@ int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
 		op->op_notifier = kmalloc(sizeof(struct rds_notifier), GFP_KERNEL);
 		if (!op->op_notifier) {
 			ret = -ENOMEM;
-			goto out;
+			goto out_pages;
 		}
 		op->op_notifier->n_user_token = args->user_token;
 		op->op_notifier->n_status = RDS_RDMA_SUCCESS;
@@ -681,7 +684,7 @@ int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
 		 */
 		ret = rds_pin_pages(iov->addr, nr, pages, !op->op_write);
 		if (ret < 0)
-			goto out;
+			goto out_pages;
 		else
 			ret = 0;
 
@@ -714,13 +717,11 @@ int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
 				nr_bytes,
 				(unsigned int) args->remote_vec.bytes);
 		ret = -EINVAL;
-		goto out;
+		goto out_pages;
 	}
 	op->op_bytes = nr_bytes;
 
-out:
-	if (iovs != iovstack)
-		sock_kfree_s(rds_rs_to_sk(rs), iovs, iov_size);
+out_pages:
 	kfree(pages);
 out_ret:
 	if (ret)
diff --git a/net/rds/rds.h b/net/rds/rds.h
index 6bfaf05b63b2..4d2523100093 100644
--- a/net/rds/rds.h
+++ b/net/rds/rds.h
@@ -386,6 +386,18 @@ static inline void rds_message_zcopy_queue_init(struct rds_msg_zcopy_queue *q)
 	INIT_LIST_HEAD(&q->zcookie_head);
 }
 
+struct rds_iov_vector {
+	struct rds_iovec *iov;
+	int               len;
+};
+
+struct rds_iov_vector_arr {
+	struct rds_iov_vector *vec;
+	int                    len;
+	int                    indx;
+	int                    incr;
+};
+
 struct rds_message {
 	refcount_t		m_refcount;
 	struct list_head	m_sock_item;
@@ -904,13 +916,13 @@ int rds_get_mr(struct rds_sock *rs, char __user *optval, int optlen);
 int rds_get_mr_for_dest(struct rds_sock *rs, char __user *optval, int optlen);
 int rds_free_mr(struct rds_sock *rs, char __user *optval, int optlen);
 void rds_rdma_drop_keys(struct rds_sock *rs);
-int rds_rdma_extra_size(struct rds_rdma_args *args);
-int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
-			  struct cmsghdr *cmsg);
+int rds_rdma_extra_size(struct rds_rdma_args *args,
+			struct rds_iov_vector *iov);
 int rds_cmsg_rdma_dest(struct rds_sock *rs, struct rds_message *rm,
 			  struct cmsghdr *cmsg);
 int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
-			  struct cmsghdr *cmsg);
+			  struct cmsghdr *cmsg,
+			  struct rds_iov_vector *vec);
 int rds_cmsg_rdma_map(struct rds_sock *rs, struct rds_message *rm,
 			  struct cmsghdr *cmsg);
 void rds_rdma_free_op(struct rm_rdma_op *ro);
diff --git a/net/rds/send.c b/net/rds/send.c
index fe785ee819dd..ec2267cbf85f 100644
--- a/net/rds/send.c
+++ b/net/rds/send.c
@@ -876,13 +876,15 @@ static int rds_send_queue_rm(struct rds_sock *rs, struct rds_connection *conn,
  * rds_message is getting to be quite complicated, and we'd like to allocate
  * it all in one go. This figures out how big it needs to be up front.
  */
-static int rds_rm_size(struct msghdr *msg, int num_sgs)
+static int rds_rm_size(struct msghdr *msg, int num_sgs,
+		       struct rds_iov_vector_arr *vct)
 {
 	struct cmsghdr *cmsg;
 	int size = 0;
 	int cmsg_groups = 0;
 	int retval;
 	bool zcopy_cookie = false;
+	struct rds_iov_vector *iov, *tmp_iov;
 
 	for_each_cmsghdr(cmsg, msg) {
 		if (!CMSG_OK(msg, cmsg))
@@ -893,8 +895,24 @@ static int rds_rm_size(struct msghdr *msg, int num_sgs)
 
 		switch (cmsg->cmsg_type) {
 		case RDS_CMSG_RDMA_ARGS:
+			if (vct->indx >= vct->len) {
+				vct->len += vct->incr;
+				tmp_iov =
+					krealloc(vct->vec,
+						 vct->len *
+						 sizeof(struct rds_iov_vector),
+						 GFP_KERNEL);
+				if (!tmp_iov) {
+					vct->len -= vct->incr;
+					return -ENOMEM;
+				}
+				vct->vec = tmp_iov;
+			}
+			iov = &vct->vec[vct->indx];
+			memset(iov, 0, sizeof(struct rds_iov_vector));
+			vct->indx++;
 			cmsg_groups |= 1;
-			retval = rds_rdma_extra_size(CMSG_DATA(cmsg));
+			retval = rds_rdma_extra_size(CMSG_DATA(cmsg), iov);
 			if (retval < 0)
 				return retval;
 			size += retval;
@@ -951,10 +969,11 @@ static int rds_cmsg_zcopy(struct rds_sock *rs, struct rds_message *rm,
 }
 
 static int rds_cmsg_send(struct rds_sock *rs, struct rds_message *rm,
-			 struct msghdr *msg, int *allocated_mr)
+			 struct msghdr *msg, int *allocated_mr,
+			 struct rds_iov_vector_arr *vct)
 {
 	struct cmsghdr *cmsg;
-	int ret = 0;
+	int ret = 0, ind = 0;
 
 	for_each_cmsghdr(cmsg, msg) {
 		if (!CMSG_OK(msg, cmsg))
@@ -968,7 +987,10 @@ static int rds_cmsg_send(struct rds_sock *rs, struct rds_message *rm,
 		 */
 		switch (cmsg->cmsg_type) {
 		case RDS_CMSG_RDMA_ARGS:
-			ret = rds_cmsg_rdma_args(rs, rm, cmsg);
+			if (ind >= vct->indx)
+				return -ENOMEM;
+			ret = rds_cmsg_rdma_args(rs, rm, cmsg, &vct->vec[ind]);
+			ind++;
 			break;
 
 		case RDS_CMSG_RDMA_DEST:
@@ -1084,6 +1106,11 @@ int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
 		      sock_flag(rds_rs_to_sk(rs), SOCK_ZEROCOPY));
 	int num_sgs = ceil(payload_len, PAGE_SIZE);
 	int namelen;
+	struct rds_iov_vector_arr vct = {0};
+	int ind;
+
+	/* expect 1 RDMA CMSG per rds_sendmsg. can still grow if more needed. */
+	vct.incr = 1;
 
 	/* Mirror Linux UDP mirror of BSD error message compatibility */
 	/* XXX: Perhaps MSG_MORE someday */
@@ -1220,7 +1247,7 @@ int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
 		num_sgs = iov_iter_npages(&msg->msg_iter, INT_MAX);
 	}
 	/* size of rm including all sgs */
-	ret = rds_rm_size(msg, num_sgs);
+	ret = rds_rm_size(msg, num_sgs, &vct);
 	if (ret < 0)
 		goto out;
 
@@ -1270,7 +1297,7 @@ int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
 	rm->m_conn_path = cpath;
 
 	/* Parse any control messages the user may have included. */
-	ret = rds_cmsg_send(rs, rm, msg, &allocated_mr);
+	ret = rds_cmsg_send(rs, rm, msg, &allocated_mr, &vct);
 	if (ret) {
 		/* Trigger connection so that its ready for the next retry */
 		if (ret ==  -EAGAIN)
@@ -1348,9 +1375,18 @@ int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
 	if (ret)
 		goto out;
 	rds_message_put(rm);
+
+	for (ind = 0; ind < vct.indx; ind++)
+		kfree(vct.vec[ind].iov);
+	kfree(vct.vec);
+
 	return payload_len;
 
 out:
+	for (ind = 0; ind < vct.indx; ind++)
+		kfree(vct.vec[ind].iov);
+	kfree(vct.vec);
+
 	/* If the user included a RDMA_MAP cmsg, we allocated a MR on the fly.
 	 * If the sendmsg goes through, we keep the MR. If it fails with EAGAIN
 	 * or in any other way, we need to destroy the MR again */
-- 
2.19.1

[PATCH rds linux-next v4 2/2] net/rds: remove user triggered WARN_ON in rds_sendmsg

[Shamir Rabinovitch]

From: shamir rabinovitch <shamir.rabinovitch@oracle.com>

per comment from Leon in rdma mailing list
https://lkml.org/lkml/2018/10/31/312 :

Please don't forget to remove user triggered WARN_ON.
https://lwn.net/Articles/769365/
"Greg Kroah-Hartman raised the problem of core kernel API code that will
use WARN_ON_ONCE() to complain about bad usage; that will not generate
the desired result if WARN_ON_ONCE() is configured to crash the machine.
He was told that the code should just call pr_warn() instead, and that
the called function should return an error in such situations. It was
generally agreed that any WARN_ON() or WARN_ON_ONCE() calls that can be
triggered from user space need to be fixed."

in addition harden rds_sendmsg to detect and overcome issues with
invalid sg count and fail the sendmsg.

Suggested-by: Leon Romanovsky <leon@kernel.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@oracle.com> 
Signed-off-by: shamir rabinovitch <shamir.rabinovitch@oracle.com>
---
 net/rds/message.c | 24 ++++++++++++++++++------
 net/rds/rdma.c    | 12 ++++--------
 net/rds/rds.h     |  3 ++-
 net/rds/send.c    |  9 +++++----
 4 files changed, 29 insertions(+), 19 deletions(-)

diff --git a/net/rds/message.c b/net/rds/message.c
index 4b00b1152a5f..f139420ba1f6 100644
--- a/net/rds/message.c
+++ b/net/rds/message.c
@@ -308,16 +308,27 @@ struct rds_message *rds_message_alloc(unsigned int extra_len, gfp_t gfp)
 /*
  * RDS ops use this to grab SG entries from the rm's sg pool.
  */
-struct scatterlist *rds_message_alloc_sgs(struct rds_message *rm, int nents)
+struct scatterlist *rds_message_alloc_sgs(struct rds_message *rm, int nents,
+					  int *ret)
 {
 	struct scatterlist *sg_first = (struct scatterlist *) &rm[1];
 	struct scatterlist *sg_ret;
 
-	WARN_ON(rm->m_used_sgs + nents > rm->m_total_sgs);
-	WARN_ON(!nents);
+	if (WARN_ON(!ret))
+		return NULL;
 
-	if (rm->m_used_sgs + nents > rm->m_total_sgs)
+	if (nents <= 0) {
+		pr_warn("rds: alloc sgs failed! nents <= 0\n");
+		*ret = -EINVAL;
 		return NULL;
+	}
+
+	if (rm->m_used_sgs + nents > rm->m_total_sgs) {
+		pr_warn("rds: alloc sgs failed! total %d used %d nents %d\n",
+			rm->m_total_sgs, rm->m_used_sgs, nents);
+		*ret = -ENOMEM;
+		return NULL;
+	}
 
 	sg_ret = &sg_first[rm->m_used_sgs];
 	sg_init_table(sg_ret, nents);
@@ -332,6 +343,7 @@ struct rds_message *rds_message_map_pages(unsigned long *page_addrs, unsigned in
 	unsigned int i;
 	int num_sgs = ceil(total_len, PAGE_SIZE);
 	int extra_bytes = num_sgs * sizeof(struct scatterlist);
+	int ret;
 
 	rm = rds_message_alloc(extra_bytes, GFP_NOWAIT);
 	if (!rm)
@@ -340,10 +352,10 @@ struct rds_message *rds_message_map_pages(unsigned long *page_addrs, unsigned in
 	set_bit(RDS_MSG_PAGEVEC, &rm->m_flags);
 	rm->m_inc.i_hdr.h_len = cpu_to_be32(total_len);
 	rm->data.op_nents = ceil(total_len, PAGE_SIZE);
-	rm->data.op_sg = rds_message_alloc_sgs(rm, num_sgs);
+	rm->data.op_sg = rds_message_alloc_sgs(rm, num_sgs, &ret);
 	if (!rm->data.op_sg) {
 		rds_message_put(rm);
-		return ERR_PTR(-ENOMEM);
+		return ERR_PTR(ret);
 	}
 
 	for (i = 0; i < rm->data.op_nents; ++i) {
diff --git a/net/rds/rdma.c b/net/rds/rdma.c
index e1965d9cbcf8..182ab8430594 100644
--- a/net/rds/rdma.c
+++ b/net/rds/rdma.c
@@ -623,11 +623,9 @@ int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
 	op->op_active = 1;
 	op->op_recverr = rs->rs_recverr;
 	WARN_ON(!nr_pages);
-	op->op_sg = rds_message_alloc_sgs(rm, nr_pages);
-	if (!op->op_sg) {
-		ret = -ENOMEM;
+	op->op_sg = rds_message_alloc_sgs(rm, nr_pages, &ret);
+	if (!op->op_sg)
 		goto out_pages;
-	}
 
 	if (op->op_notify || op->op_recverr) {
 		/* We allocate an uninitialized notifier here, because
@@ -839,11 +837,9 @@ int rds_cmsg_atomic(struct rds_sock *rs, struct rds_message *rm,
 	rm->atomic.op_silent = !!(args->flags & RDS_RDMA_SILENT);
 	rm->atomic.op_active = 1;
 	rm->atomic.op_recverr = rs->rs_recverr;
-	rm->atomic.op_sg = rds_message_alloc_sgs(rm, 1);
-	if (!rm->atomic.op_sg) {
-		ret = -ENOMEM;
+	rm->atomic.op_sg = rds_message_alloc_sgs(rm, 1, &ret);
+	if (!rm->atomic.op_sg)
 		goto err;
-	}
 
 	/* verify 8 byte-aligned */
 	if (args->local_addr & 0x7) {
diff --git a/net/rds/rds.h b/net/rds/rds.h
index 4d2523100093..02ec4a3b2799 100644
--- a/net/rds/rds.h
+++ b/net/rds/rds.h
@@ -839,7 +839,8 @@ rds_conn_connecting(struct rds_connection *conn)
 
 /* message.c */
 struct rds_message *rds_message_alloc(unsigned int nents, gfp_t gfp);
-struct scatterlist *rds_message_alloc_sgs(struct rds_message *rm, int nents);
+struct scatterlist *rds_message_alloc_sgs(struct rds_message *rm, int nents,
+					  int *ret);
 int rds_message_copy_from_user(struct rds_message *rm, struct iov_iter *from,
 			       bool zcopy);
 struct rds_message *rds_message_map_pages(unsigned long *page_addrs, unsigned int total_len);
diff --git a/net/rds/send.c b/net/rds/send.c
index ec2267cbf85f..b39b30706210 100644
--- a/net/rds/send.c
+++ b/net/rds/send.c
@@ -886,6 +886,9 @@ static int rds_rm_size(struct msghdr *msg, int num_sgs,
 	bool zcopy_cookie = false;
 	struct rds_iov_vector *iov, *tmp_iov;
 
+	if (num_sgs < 0)
+		return -EINVAL;
+
 	for_each_cmsghdr(cmsg, msg) {
 		if (!CMSG_OK(msg, cmsg))
 			return -EINVAL;
@@ -1259,11 +1262,9 @@ int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
 
 	/* Attach data to the rm */
 	if (payload_len) {
-		rm->data.op_sg = rds_message_alloc_sgs(rm, num_sgs);
-		if (!rm->data.op_sg) {
-			ret = -ENOMEM;
+		rm->data.op_sg = rds_message_alloc_sgs(rm, num_sgs, &ret);
+		if (!rm->data.op_sg)
 			goto out;
-		}
 		ret = rds_message_copy_from_user(rm, &msg->msg_iter, zcopy);
 		if (ret)
 			goto out;
-- 
2.19.1

Re: [PATCH rds linux-next v4 0/2] WARNING in rds_message_alloc_sgs

[David Miller]

From: Shamir Rabinovitch <shamir.rabinovitch@oracle.com>
Date: Sun, 16 Dec 2018 09:01:07 +0200

> From: shamir rabinovitch <shamir.rabinovitch@oracle.com>
> 
> This patch set fix google syzbot rds bug found in linux-next.
> The first patch solve the syzbot issue.
> The second patch fix issue mentioned by Leon Romanovsky that
> drivers should not call WARN_ON as result from user input.
> 
> syzbot bug report can be foud here: https://lkml.org/lkml/2018/10/31/28
> 
> v1->v2:
> - patch 1: make rds_iov_vector fields name more descriptive (Hakon)
> - patch 1: fix potential mem leak in rds_rm_size if krealloc fail
>   (Hakon)
> v2->v3:
> - patch 2: harden rds_sendmsg for invalid number of sgs (Gerd)
> v3->v4
> - Santosh a.b. on both patches + repost to net-dev

Does this bug exist in 'net' too?  It really should target the 'net'
tree if so, and then we can queue these fixes up for -stable too.

Thank you.

Re: [PATCH rds linux-next v4 0/2] WARNING in rds_message_alloc_sgs

[Santosh Shilimkar]

On 12/18/2018 2:57 PM, David Miller wrote:
> From: Shamir Rabinovitch <shamir.rabinovitch@oracle.com>
> Date: Sun, 16 Dec 2018 09:01:07 +0200
> 
>> From: shamir rabinovitch <shamir.rabinovitch@oracle.com>
>>
>> This patch set fix google syzbot rds bug found in linux-next.
>> The first patch solve the syzbot issue.
>> The second patch fix issue mentioned by Leon Romanovsky that
>> drivers should not call WARN_ON as result from user input.
>>
>> syzbot bug report can be foud here: https://lkml.org/lkml/2018/10/31/28
>>
>> v1->v2:
>> - patch 1: make rds_iov_vector fields name more descriptive (Hakon)
>> - patch 1: fix potential mem leak in rds_rm_size if krealloc fail
>>    (Hakon)
>> v2->v3:
>> - patch 2: harden rds_sendmsg for invalid number of sgs (Gerd)
>> v3->v4
>> - Santosh a.b. on both patches + repost to net-dev
> 
> Does this bug exist in 'net' too?  It really should target the 'net'
> tree if so, and then we can queue these fixes up for -stable too.
> 
yes, the bug exist in net too.

Shamir, please check if these apply cleanly to Dave's 'net' tree.

Regards,
Santosh

Re: [PATCH rds linux-next v4 0/2] WARNING in rds_message_alloc_sgs

[Shamir Rabinovitch]

On Tue, Dec 18, 2018 at 03:02:24PM -0800, Santosh Shilimkar wrote:
> On 12/18/2018 2:57 PM, David Miller wrote:
> > From: Shamir Rabinovitch <shamir.rabinovitch@oracle.com>
> > Date: Sun, 16 Dec 2018 09:01:07 +0200
> > 
> > > From: shamir rabinovitch <shamir.rabinovitch@oracle.com>
> > > 
> > > This patch set fix google syzbot rds bug found in linux-next.
> > > The first patch solve the syzbot issue.
> > > The second patch fix issue mentioned by Leon Romanovsky that
> > > drivers should not call WARN_ON as result from user input.
> > > 
> > > syzbot bug report can be foud here: https://lkml.org/lkml/2018/10/31/28
> > > 
> > > v1->v2:
> > > - patch 1: make rds_iov_vector fields name more descriptive (Hakon)
> > > - patch 1: fix potential mem leak in rds_rm_size if krealloc fail
> > >    (Hakon)
> > > v2->v3:
> > > - patch 2: harden rds_sendmsg for invalid number of sgs (Gerd)
> > > v3->v4
> > > - Santosh a.b. on both patches + repost to net-dev
> > 
> > Does this bug exist in 'net' too?  It really should target the 'net'
> > tree if so, and then we can queue these fixes up for -stable too.
> > 
> yes, the bug exist in net too.
> 
> Shamir, please check if these apply cleanly to Dave's 'net' tree.
> 
> Regards,
> Santosh

Santosh, 

The patches apply cleanly on the master branch of 
git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git . 

See below the snippet of the log with the patches applied.

8472aa4edfcb (HEAD -> net/master#rds-syzbot-bug) net/rds: remove user triggered WARN_ON in rds_sendmsg
3f8d6b898c5e net/rds: fix warn in rds_message_alloc_sgs
3061169a47ee (net/master) Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
78abe3d0dfad net/smc: fix TCP fallback socket release
f7db2beb4c2c vxge: ensure data0 is initialized in when fetching firmware version information

Thanks, Shamir

Re: [PATCH rds linux-next v4 0/2] WARNING in rds_message_alloc_sgs

[Santosh Shilimkar]

On 12/19/2018 2:54 AM, Shamir Rabinovitch wrote:
> On Tue, Dec 18, 2018 at 03:02:24PM -0800, Santosh Shilimkar wrote:
[...]

>>
>> Shamir, please check if these apply cleanly to Dave's 'net' tree.
>>

> 
> Santosh,
> 
> The patches apply cleanly on the master branch of
> git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git .
> 
> See below the snippet of the log with the patches applied.
>
> 8472aa4edfcb (HEAD -> net/master#rds-syzbot-bug) net/rds: remove user triggered WARN_ON in rds_sendmsg
> 3f8d6b898c5e net/rds: fix warn in rds_message_alloc_sgs
> 3061169a47ee (net/master) Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
> 78abe3d0dfad net/smc: fix TCP fallback socket release
> f7db2beb4c2c vxge: ensure data0 is initialized in when fetching firmware version information
> 
Great. On patchwork I see the patches marked under review by Dave so
lets wait for his instruction(s).

Regards,
Santosh

Re: [PATCH rds linux-next v4 0/2] WARNING in rds_message_alloc_sgs

[David Miller]

From: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Date: Wed, 19 Dec 2018 10:00:20 -0800

> On 12/19/2018 2:54 AM, Shamir Rabinovitch wrote:
>> On Tue, Dec 18, 2018 at 03:02:24PM -0800, Santosh Shilimkar wrote:
> [...]
> 
>>>
>>> Shamir, please check if these apply cleanly to Dave's 'net' tree.
>>>
> 
>> Santosh,
>> The patches apply cleanly on the master branch of
>> git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git .
>> See below the snippet of the log with the patches applied.
>>
>> 8472aa4edfcb (HEAD -> net/master#rds-syzbot-bug) net/rds: remove user
>> triggered WARN_ON in rds_sendmsg
>> 3f8d6b898c5e net/rds: fix warn in rds_message_alloc_sgs
>> 3061169a47ee (net/master) Merge
>> git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
>> 78abe3d0dfad net/smc: fix TCP fallback socket release
>> f7db2beb4c2c vxge: ensure data0 is initialized in when fetching
>> firmware version information
>> 
> Great. On patchwork I see the patches marked under review by Dave so
> lets wait for his instruction(s).

Since they apply cleanly to 'net' I'll apply them directly and queue up
for -stable.

Thanks for checking.

Re: [PATCH rds linux-next v4 0/2] WARNING in rds_message_alloc_sgs

[Santosh Shilimkar]

On 12/19/2018 10:22 AM, David Miller wrote:
> From: Santosh Shilimkar <santosh.shilimkar@oracle.com>
> Date: Wed, 19 Dec 2018 10:00:20 -0800
> 
>> On 12/19/2018 2:54 AM, Shamir Rabinovitch wrote:

[...]

>> Great. On patchwork I see the patches marked under review by Dave so
>> lets wait for his instruction(s).
> 
> Since they apply cleanly to 'net' I'll apply them directly and queue up
> for -stable.
> 
> Thanks for checking.
> 
Thanks Dave !!

Regards,
Santosh

Re: [PATCH rds linux-next v4 0/2] WARNING in rds_message_alloc_sgs

[David Miller]

From: Shamir Rabinovitch <shamir.rabinovitch@oracle.com>
Date: Sun, 16 Dec 2018 09:01:07 +0200

> From: shamir rabinovitch <shamir.rabinovitch@oracle.com>
> 
> This patch set fix google syzbot rds bug found in linux-next.
> The first patch solve the syzbot issue.
> The second patch fix issue mentioned by Leon Romanovsky that
> drivers should not call WARN_ON as result from user input.
> 
> syzbot bug report can be foud here: https://lkml.org/lkml/2018/10/31/28
> 
> v1->v2:
> - patch 1: make rds_iov_vector fields name more descriptive (Hakon)
> - patch 1: fix potential mem leak in rds_rm_size if krealloc fail
>   (Hakon)
> v2->v3:
> - patch 2: harden rds_sendmsg for invalid number of sgs (Gerd)
> v3->v4
> - Santosh a.b. on both patches + repost to net-dev

Series applied to 'net' and queued up for -stable.

Thanks.