Skip to content

Commit ea01007

Browse files
srabinovdavem330
srabinov
authored and
davem330
committed
net/rds: fix warn in rds_message_alloc_sgs
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> Signed-off-by: David S. Miller <davem@davemloft.net>
1 parent c6f4075 commit ea01007

File tree

3 files changed

+91
-42
lines changed

3 files changed

+91
-42
lines changed

‎net/rds/rdma.c‎

Lines changed: 32 additions & 31 deletions
Original file line numberDiff line numberDiff line change
@@ -517,9 +517,10 @@ static int rds_rdma_pages(struct rds_iovec iov[], int nr_iovecs)
517517
return tot_pages;
518518
}
519519

520-
int rds_rdma_extra_size(struct rds_rdma_args *args)
520+
int rds_rdma_extra_size(struct rds_rdma_args *args,
521+
struct rds_iov_vector *iov)
521522
{
522-
struct rds_iovec vec;
523+
struct rds_iovec *vec;
523524
struct rds_iovec __user *local_vec;
524525
int tot_pages = 0;
525526
unsigned int nr_pages;
@@ -530,13 +531,23 @@ int rds_rdma_extra_size(struct rds_rdma_args *args)
530531
if (args->nr_local == 0)
531532
return -EINVAL;
532533

534+
iov->iov = kcalloc(args->nr_local,
535+
sizeof(struct rds_iovec),
536+
GFP_KERNEL);
537+
if (!iov->iov)
538+
return -ENOMEM;
539+
540+
vec = &iov->iov[0];
541+
542+
if (copy_from_user(vec, local_vec, args->nr_local *
543+
sizeof(struct rds_iovec)))
544+
return -EFAULT;
545+
iov->len = args->nr_local;
546+
533547
/* figure out the number of pages in the vector */
534-
for (i = 0; i < args->nr_local; i++) {
535-
if (copy_from_user(&vec, &local_vec[i],
536-
sizeof(struct rds_iovec)))
537-
return -EFAULT;
548+
for (i = 0; i < args->nr_local; i++, vec++) {
538549

539-
nr_pages = rds_pages_in_vec(&vec);
550+
nr_pages = rds_pages_in_vec(vec);
540551
if (nr_pages == 0)
541552
return -EINVAL;
542553

@@ -558,15 +569,15 @@ int rds_rdma_extra_size(struct rds_rdma_args *args)
558569
* Extract all arguments and set up the rdma_op
559570
*/
560571
int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
561-
struct cmsghdr *cmsg)
572+
struct cmsghdr *cmsg,
573+
struct rds_iov_vector *vec)
562574
{
563575
struct rds_rdma_args *args;
564576
struct rm_rdma_op *op = &rm->rdma;
565577
int nr_pages;
566578
unsigned int nr_bytes;
567579
struct page **pages = NULL;
568-
struct rds_iovec iovstack[UIO_FASTIOV], *iovs = iovstack;
569-
int iov_size;
580+
struct rds_iovec *iovs;
570581
unsigned int i, j;
571582
int ret = 0;
572583

@@ -586,31 +597,23 @@ int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
586597
goto out_ret;
587598
}
588599

589-
/* Check whether to allocate the iovec area */
590-
iov_size = args->nr_local * sizeof(struct rds_iovec);
591-
if (args->nr_local > UIO_FASTIOV) {
592-
iovs = sock_kmalloc(rds_rs_to_sk(rs), iov_size, GFP_KERNEL);
593-
if (!iovs) {
594-
ret = -ENOMEM;
595-
goto out_ret;
596-
}
600+
if (vec->len != args->nr_local) {
601+
ret = -EINVAL;
602+
goto out_ret;
597603
}
598604

599-
if (copy_from_user(iovs, (struct rds_iovec __user *)(unsigned long) args->local_vec_addr, iov_size)) {
600-
ret = -EFAULT;
601-
goto out;
602-
}
605+
iovs = vec->iov;
603606

604607
nr_pages = rds_rdma_pages(iovs, args->nr_local);
605608
if (nr_pages < 0) {
606609
ret = -EINVAL;
607-
goto out;
610+
goto out_ret;
608611
}
609612

610613
pages = kcalloc(nr_pages, sizeof(struct page *), GFP_KERNEL);
611614
if (!pages) {
612615
ret = -ENOMEM;
613-
goto out;
616+
goto out_ret;
614617
}
615618

616619
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,
623626
op->op_sg = rds_message_alloc_sgs(rm, nr_pages);
624627
if (!op->op_sg) {
625628
ret = -ENOMEM;
626-
goto out;
629+
goto out_pages;
627630
}
628631

629632
if (op->op_notify || op->op_recverr) {
@@ -635,7 +638,7 @@ int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
635638
op->op_notifier = kmalloc(sizeof(struct rds_notifier), GFP_KERNEL);
636639
if (!op->op_notifier) {
637640
ret = -ENOMEM;
638-
goto out;
641+
goto out_pages;
639642
}
640643
op->op_notifier->n_user_token = args->user_token;
641644
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,
681684
*/
682685
ret = rds_pin_pages(iov->addr, nr, pages, !op->op_write);
683686
if (ret < 0)
684-
goto out;
687+
goto out_pages;
685688
else
686689
ret = 0;
687690

@@ -714,13 +717,11 @@ int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
714717
nr_bytes,
715718
(unsigned int) args->remote_vec.bytes);
716719
ret = -EINVAL;
717-
goto out;
720+
goto out_pages;
718721
}
719722
op->op_bytes = nr_bytes;
720723

721-
out:
722-
if (iovs != iovstack)
723-
sock_kfree_s(rds_rs_to_sk(rs), iovs, iov_size);
724+
out_pages:
724725
kfree(pages);
725726
out_ret:
726727
if (ret)

‎net/rds/rds.h‎

Lines changed: 16 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -386,6 +386,18 @@ static inline void rds_message_zcopy_queue_init(struct rds_msg_zcopy_queue *q)
386386
INIT_LIST_HEAD(&q->zcookie_head);
387387
}
388388

389+
struct rds_iov_vector {
390+
struct rds_iovec *iov;
391+
int len;
392+
};
393+
394+
struct rds_iov_vector_arr {
395+
struct rds_iov_vector *vec;
396+
int len;
397+
int indx;
398+
int incr;
399+
};
400+
389401
struct rds_message {
390402
refcount_t m_refcount;
391403
struct list_head m_sock_item;
@@ -904,13 +916,13 @@ int rds_get_mr(struct rds_sock *rs, char __user *optval, int optlen);
904916
int rds_get_mr_for_dest(struct rds_sock *rs, char __user *optval, int optlen);
905917
int rds_free_mr(struct rds_sock *rs, char __user *optval, int optlen);
906918
void rds_rdma_drop_keys(struct rds_sock *rs);
907-
int rds_rdma_extra_size(struct rds_rdma_args *args);
908-
int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
909-
struct cmsghdr *cmsg);
919+
int rds_rdma_extra_size(struct rds_rdma_args *args,
920+
struct rds_iov_vector *iov);
910921
int rds_cmsg_rdma_dest(struct rds_sock *rs, struct rds_message *rm,
911922
struct cmsghdr *cmsg);
912923
int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
913-
struct cmsghdr *cmsg);
924+
struct cmsghdr *cmsg,
925+
struct rds_iov_vector *vec);
914926
int rds_cmsg_rdma_map(struct rds_sock *rs, struct rds_message *rm,
915927
struct cmsghdr *cmsg);
916928
void rds_rdma_free_op(struct rm_rdma_op *ro);

‎net/rds/send.c‎

Lines changed: 43 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -876,13 +876,15 @@ static int rds_send_queue_rm(struct rds_sock *rs, struct rds_connection *conn,
876876
* rds_message is getting to be quite complicated, and we'd like to allocate
877877
* it all in one go. This figures out how big it needs to be up front.
878878
*/
879-
static int rds_rm_size(struct msghdr *msg, int num_sgs)
879+
static int rds_rm_size(struct msghdr *msg, int num_sgs,
880+
struct rds_iov_vector_arr *vct)
880881
{
881882
struct cmsghdr *cmsg;
882883
int size = 0;
883884
int cmsg_groups = 0;
884885
int retval;
885886
bool zcopy_cookie = false;
887+
struct rds_iov_vector *iov, *tmp_iov;
886888

887889
for_each_cmsghdr(cmsg, msg) {
888890
if (!CMSG_OK(msg, cmsg))
@@ -893,8 +895,24 @@ static int rds_rm_size(struct msghdr *msg, int num_sgs)
893895

894896
switch (cmsg->cmsg_type) {
895897
case RDS_CMSG_RDMA_ARGS:
898+
if (vct->indx >= vct->len) {
899+
vct->len += vct->incr;
900+
tmp_iov =
901+
krealloc(vct->vec,
902+
vct->len *
903+
sizeof(struct rds_iov_vector),
904+
GFP_KERNEL);
905+
if (!tmp_iov) {
906+
vct->len -= vct->incr;
907+
return -ENOMEM;
908+
}
909+
vct->vec = tmp_iov;
910+
}
911+
iov = &vct->vec[vct->indx];
912+
memset(iov, 0, sizeof(struct rds_iov_vector));
913+
vct->indx++;
896914
cmsg_groups |= 1;
897-
retval = rds_rdma_extra_size(CMSG_DATA(cmsg));
915+
retval = rds_rdma_extra_size(CMSG_DATA(cmsg), iov);
898916
if (retval < 0)
899917
return retval;
900918
size += retval;
@@ -951,10 +969,11 @@ static int rds_cmsg_zcopy(struct rds_sock *rs, struct rds_message *rm,
951969
}
952970

953971
static int rds_cmsg_send(struct rds_sock *rs, struct rds_message *rm,
954-
struct msghdr *msg, int *allocated_mr)
972+
struct msghdr *msg, int *allocated_mr,
973+
struct rds_iov_vector_arr *vct)
955974
{
956975
struct cmsghdr *cmsg;
957-
int ret = 0;
976+
int ret = 0, ind = 0;
958977

959978
for_each_cmsghdr(cmsg, msg) {
960979
if (!CMSG_OK(msg, cmsg))
@@ -968,7 +987,10 @@ static int rds_cmsg_send(struct rds_sock *rs, struct rds_message *rm,
968987
*/
969988
switch (cmsg->cmsg_type) {
970989
case RDS_CMSG_RDMA_ARGS:
971-
ret = rds_cmsg_rdma_args(rs, rm, cmsg);
990+
if (ind >= vct->indx)
991+
return -ENOMEM;
992+
ret = rds_cmsg_rdma_args(rs, rm, cmsg, &vct->vec[ind]);
993+
ind++;
972994
break;
973995

974996
case RDS_CMSG_RDMA_DEST:
@@ -1084,6 +1106,11 @@ int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
10841106
sock_flag(rds_rs_to_sk(rs), SOCK_ZEROCOPY));
10851107
int num_sgs = ceil(payload_len, PAGE_SIZE);
10861108
int namelen;
1109+
struct rds_iov_vector_arr vct = {0};
1110+
int ind;
1111+
1112+
/* expect 1 RDMA CMSG per rds_sendmsg. can still grow if more needed. */
1113+
vct.incr = 1;
10871114

10881115
/* Mirror Linux UDP mirror of BSD error message compatibility */
10891116
/* XXX: Perhaps MSG_MORE someday */
@@ -1220,7 +1247,7 @@ int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
12201247
num_sgs = iov_iter_npages(&msg->msg_iter, INT_MAX);
12211248
}
12221249
/* size of rm including all sgs */
1223-
ret = rds_rm_size(msg, num_sgs);
1250+
ret = rds_rm_size(msg, num_sgs, &vct);
12241251
if (ret < 0)
12251252
goto out;
12261253

@@ -1270,7 +1297,7 @@ int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
12701297
rm->m_conn_path = cpath;
12711298

12721299
/* Parse any control messages the user may have included. */
1273-
ret = rds_cmsg_send(rs, rm, msg, &allocated_mr);
1300+
ret = rds_cmsg_send(rs, rm, msg, &allocated_mr, &vct);
12741301
if (ret) {
12751302
/* Trigger connection so that its ready for the next retry */
12761303
if (ret == -EAGAIN)
@@ -1348,9 +1375,18 @@ int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
13481375
if (ret)
13491376
goto out;
13501377
rds_message_put(rm);
1378+
1379+
for (ind = 0; ind < vct.indx; ind++)
1380+
kfree(vct.vec[ind].iov);
1381+
kfree(vct.vec);
1382+
13511383
return payload_len;
13521384

13531385
out:
1386+
for (ind = 0; ind < vct.indx; ind++)
1387+
kfree(vct.vec[ind].iov);
1388+
kfree(vct.vec);
1389+
13541390
/* If the user included a RDMA_MAP cmsg, we allocated a MR on the fly.
13551391
* If the sendmsg goes through, we keep the MR. If it fails with EAGAIN
13561392
* or in any other way, we need to destroy the MR again */

0 commit comments

Comments
 (0)