RDMA/rtrs: Fix KASAN: stack-out-of-bounds bug

jinpuwang · jgunthorpe · commit 7fbc3c373eef · 2021-01-15T15:25:10.000-04:00
When KASAN is enabled, we notice warning below: [ 483.436975] ================================================================== [ 483.437234] BUG: KASAN: stack-out-of-bounds in _mlx5_ib_post_send+0x188a/0x2560 [mlx5_ib] [ 483.437430] Read of size 4 at addr ffff88a195fd7d30 by task kworker/1:3/6954 [ 483.437731] CPU: 1 PID: 6954 Comm: kworker/1:3 Kdump: loaded Tainted: G O 5.4.82-pserver #5.4.82-1+feature+linux+5.4.y+dbg+20201210.1532+987e7a6~deb10 [ 483.437976] Hardware name: Supermicro Super Server/X11DDW-L, BIOS 3.3 02/21/2020 [ 483.438168] Workqueue: rtrs_server_wq hb_work [rtrs_core] [ 483.438323] Call Trace: [ 483.438486] dump_stack+0x96/0xe0 [ 483.438646] ? _mlx5_ib_post_send+0x188a/0x2560 [mlx5_ib] [ 483.438802] print_address_description.constprop.6+0x1b/0x220 [ 483.438966] ? _mlx5_ib_post_send+0x188a/0x2560 [mlx5_ib] [ 483.439133] ? _mlx5_ib_post_send+0x188a/0x2560 [mlx5_ib] [ 483.439285] __kasan_report.cold.9+0x1a/0x32 [ 483.439444] ? _mlx5_ib_post_send+0x188a/0x2560 [mlx5_ib] [ 483.439597] kasan_report+0x10/0x20 [ 483.439752] _mlx5_ib_post_send+0x188a/0x2560 [mlx5_ib] [ 483.439910] ? update_sd_lb_stats+0xfb1/0xfc0 [ 483.440073] ? set_reg_wr+0x520/0x520 [mlx5_ib] [ 483.440222] ? update_group_capacity+0x340/0x340 [ 483.440377] ? find_busiest_group+0x314/0x870 [ 483.440526] ? update_sd_lb_stats+0xfc0/0xfc0 [ 483.440683] ? __bitmap_and+0x6f/0x100 [ 483.440832] ? __lock_acquire+0xa2/0x2150 [ 483.440979] ? __lock_acquire+0xa2/0x2150 [ 483.441128] ? __lock_acquire+0xa2/0x2150 [ 483.441279] ? debug_lockdep_rcu_enabled+0x23/0x60 [ 483.441430] ? lock_downgrade+0x390/0x390 [ 483.441582] ? __lock_acquire+0xa2/0x2150 [ 483.441729] ? __lock_acquire+0xa2/0x2150 [ 483.441876] ? newidle_balance+0x425/0x8f0 [ 483.442024] ? __lock_acquire+0xa2/0x2150 [ 483.442172] ? debug_lockdep_rcu_enabled+0x23/0x60 [ 483.442330] hb_work+0x15d/0x1d0 [rtrs_core] [ 483.442479] ? schedule_hb+0x50/0x50 [rtrs_core] [ 483.442627] ? lock_downgrade+0x390/0x390 [ 483.442781] ? process_one_work+0x40d/0xa50 [ 483.442931] process_one_work+0x4ee/0xa50 [ 483.443082] ? pwq_dec_nr_in_flight+0x110/0x110 [ 483.443231] ? do_raw_spin_lock+0x119/0x1d0 [ 483.443383] worker_thread+0x65/0x5c0 [ 483.443532] ? process_one_work+0xa50/0xa50 [ 483.451839] kthread+0x1e2/0x200 [ 483.451983] ? kthread_create_on_node+0xc0/0xc0 [ 483.452139] ret_from_fork+0x3a/0x50 The problem is we use wrong type when send wr, hw driver expect the type of IB_WR_RDMA_WRITE_WITH_IMM wr should be ib_rdma_wr, and doing container_of to access member. The fix is simple use ib_rdma_wr instread of ib_send_wr. Fixes: c0894b3 ("RDMA/rtrs: core: lib functions shared between client and server modules") Link: https://lore.kernel.org/r/20201217141915.56989-20-jinpu.wang@cloud.ionos.com Signed-off-by: Jack Wang <jinpu.wang@cloud.ionos.com> Reviewed-by: Gioh Kim <gi-oh.kim@cloud.ionos.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
diff --git a/drivers/infiniband/ulp/rtrs/rtrs.c b/drivers/infiniband/ulp/rtrs/rtrs.c
@@ -182,16 +182,16 @@ int rtrs_post_rdma_write_imm_empty(struct rtrs_con *con, struct ib_cqe *cqe,
 				    u32 imm_data, enum ib_send_flags flags,
 				    struct ib_send_wr *head)
 {
-	struct ib_send_wr wr;
+	struct ib_rdma_wr wr;
 
-	wr = (struct ib_send_wr) {
-		.wr_cqe	= cqe,
-		.send_flags	= flags,
-		.opcode	= IB_WR_RDMA_WRITE_WITH_IMM,
-		.ex.imm_data	= cpu_to_be32(imm_data),
+	wr = (struct ib_rdma_wr) {
+		.wr.wr_cqe	= cqe,
+		.wr.send_flags	= flags,
+		.wr.opcode	= IB_WR_RDMA_WRITE_WITH_IMM,
+		.wr.ex.imm_data	= cpu_to_be32(imm_data),
 	};
 
-	return rtrs_post_send(con->qp, head, &wr);
+	return rtrs_post_send(con->qp, head, &wr.wr);
 }
 EXPORT_SYMBOL_GPL(rtrs_post_rdma_write_imm_empty);
 
