Commit 2ac54150 authored by Zhu Yanjun's avatar Zhu Yanjun Committed by Leon Romanovsky
Browse files

RDMA/rxe: Remove the direct link to net_device 

The similar patch in siw is in the link:
https://git.kernel.org/rdma/rdma/c/16b87037b48889

This problem also occurred in RXE. The following analyze this problem.
In the following Call Traces:
"
BUG: KASAN: slab-use-after-free in dev_get_flags+0x188/0x1d0 net/core/dev.c:8782
Read of size 4 at addr ffff8880554640b0 by task kworker/1:4/5295

CPU: 1 UID: 0 PID: 5295 Comm: kworker/1:4 Not tainted
6.12.0-rc3-syzkaller-00399-g9197b73fd7bb #0
Hardware name: Google Compute Engine/Google Compute Engine,
BIOS Google 09/13/2024
Workqueue: infiniband ib_cache_event_task
Call Trace:
 <TASK>
 __dump_stack lib/dump_stack.c:94 [inline]
 dump_stack_lvl+0x241/0x360 lib/dump_stack.c:120
 print_address_description mm/kasan/report.c:377 [inline]
 print_report+0x169/0x550 mm/kasan/report.c:488
 kasan_report+0x143/0x180 mm/kasan/report.c:601
 dev_get_flags+0x188/0x1d0 net/core/dev.c:8782
 rxe_query_port+0x12d/0x260 drivers/infiniband/sw/rxe/rxe_verbs.c:60
 __ib_query_port drivers/infiniband/core/device.c:2111 [inline]
 ib_query_port+0x168/0x7d0 drivers/infiniband/core/device.c:2143
 ib_cache_update+0x1a9/0xb80 drivers/infiniband/core/cache.c:1494
 ib_cache_event_task+0xf3/0x1e0 drivers/infiniband/core/cache.c:1568
 process_one_work kernel/workqueue.c:3229 [inline]
 process_scheduled_works+0xa65/0x1850 kernel/workqueue.c:3310
 worker_thread+0x870/0xd30 kernel/workqueue.c:3391
 kthread+0x2f2/0x390 kernel/kthread.c:389
 ret_from_fork+0x4d/0x80 arch/x86/kernel/process.c:147
 ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
 </TASK>
"

1). In the link [1],

"
 infiniband syz2: set down
"

This means that on 839.350575, the event ib_cache_event_task was sent andi
queued in ib_wq.

2). In the link [1],

"
 team0 (unregistering): Port device team_slave_0 removed
"

It indicates that before 843.251853, the net device should be freed.

3). In the link [1],

"
 BUG: KASAN: slab-use-after-free in dev_get_flags+0x188/0x1d0
"

This means that on 850.559070, this slab-use-after-free problem occurred.

In all, on 839.350575, the event ib_cache_event_task was sent and queued
in ib_wq,

before 843.251853, the net device veth was freed.

on 850.559070, this event was executed, and the mentioned freed net device
was called. Thus, the above call trace occurred.

[1] https://syzkaller.appspot.com/x/log.txt?x=12e7025f980000



Reported-by: default avatar <syzbot+4b87489410b4efd181bf@syzkaller.appspotmail.com>
Closes: https://syzkaller.appspot.com/bug?extid=4b87489410b4efd181bf


Fixes: 8700e3e7 ("Soft RoCE driver")
Signed-off-by: default avatarZhu Yanjun <yanjun.zhu@linux.dev>
Link: https://patch.msgid.link/20241220222325.2487767-1-yanjun.zhu@linux.dev


Signed-off-by: default avatarLeon Romanovsky <leon@kernel.org>
parent e3debdd4

drivers/infiniband/sw/rxe/rxe.c

+19 −4
Original line number Diff line number Diff line
@@ -40,6 +40,8 @@ void rxe_dealloc(struct ib_device *ib_dev) 
/* initialize rxe device parameters */

static void rxe_init_device_param(struct rxe_dev *rxe)

{

	struct net_device *ndev;



	rxe->max_inline_data			= RXE_MAX_INLINE_DATA;



	rxe->attr.vendor_id			= RXE_VENDOR_ID;
@@ -71,8 +73,15 @@ static void rxe_init_device_param(struct rxe_dev *rxe) 
	rxe->attr.max_fast_reg_page_list_len	= RXE_MAX_FMR_PAGE_LIST_LEN;

	rxe->attr.max_pkeys			= RXE_MAX_PKEYS;

	rxe->attr.local_ca_ack_delay		= RXE_LOCAL_CA_ACK_DELAY;



	ndev = rxe_ib_device_get_netdev(&rxe->ib_dev);

	if (!ndev)

		return;



	addrconf_addr_eui48((unsigned char *)&rxe->attr.sys_image_guid,

			rxe->ndev->dev_addr);

			ndev->dev_addr);



	dev_put(ndev);



	rxe->max_ucontext			= RXE_MAX_UCONTEXT;

}
@@ -109,10 +118,15 @@ static void rxe_init_port_param(struct rxe_port *port) 
static void rxe_init_ports(struct rxe_dev *rxe)

{

	struct rxe_port *port = &rxe->port;

	struct net_device *ndev;



	rxe_init_port_param(port);

	ndev = rxe_ib_device_get_netdev(&rxe->ib_dev);

	if (!ndev)

		return;

	addrconf_addr_eui48((unsigned char *)&port->port_guid,

			    rxe->ndev->dev_addr);

			    ndev->dev_addr);

	dev_put(ndev);

	spin_lock_init(&port->port_lock);

}
@@ -167,12 +181,13 @@ void rxe_set_mtu(struct rxe_dev *rxe, unsigned int ndev_mtu) 
/* called by ifc layer to create new rxe device.

 * The caller should allocate memory for rxe by calling ib_alloc_device.

 */

int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name)

int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name,

			struct net_device *ndev)

{

	rxe_init(rxe);

	rxe_set_mtu(rxe, mtu);



	return rxe_register_device(rxe, ibdev_name);

	return rxe_register_device(rxe, ibdev_name, ndev);

}



static int rxe_newlink(const char *ibdev_name, struct net_device *ndev)

drivers/infiniband/sw/rxe/rxe.h

+2 −1
Original line number Diff line number Diff line
@@ -139,7 +139,8 @@ enum resp_states { 


void rxe_set_mtu(struct rxe_dev *rxe, unsigned int dev_mtu);



int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name);

int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name,

			struct net_device *ndev);



void rxe_rcv(struct sk_buff *skb);

drivers/infiniband/sw/rxe/rxe_mcast.c

+20 −2
Original line number Diff line number Diff line
@@ -31,10 +31,19 @@ 
static int rxe_mcast_add(struct rxe_dev *rxe, union ib_gid *mgid)

{

	unsigned char ll_addr[ETH_ALEN];

	struct net_device *ndev;

	int ret;



	ndev = rxe_ib_device_get_netdev(&rxe->ib_dev);

	if (!ndev)

		return -ENODEV;



	ipv6_eth_mc_map((struct in6_addr *)mgid->raw, ll_addr);



	return dev_mc_add(rxe->ndev, ll_addr);

	ret = dev_mc_add(ndev, ll_addr);

	dev_put(ndev);



	return ret;

}



/**
@@ -47,10 +56,19 @@ static int rxe_mcast_add(struct rxe_dev *rxe, union ib_gid *mgid) 
static int rxe_mcast_del(struct rxe_dev *rxe, union ib_gid *mgid)

{

	unsigned char ll_addr[ETH_ALEN];

	struct net_device *ndev;

	int ret;



	ndev = rxe_ib_device_get_netdev(&rxe->ib_dev);

	if (!ndev)

		return -ENODEV;



	ipv6_eth_mc_map((struct in6_addr *)mgid->raw, ll_addr);



	return dev_mc_del(rxe->ndev, ll_addr);

	ret = dev_mc_del(ndev, ll_addr);

	dev_put(ndev);



	return ret;

}



/**

drivers/infiniband/sw/rxe/rxe_net.c

+20 −4
Original line number Diff line number Diff line
@@ -524,7 +524,16 @@ struct sk_buff *rxe_init_packet(struct rxe_dev *rxe, struct rxe_av *av, 
 */

const char *rxe_parent_name(struct rxe_dev *rxe, unsigned int port_num)

{

	return rxe->ndev->name;

	struct net_device *ndev;

	char *ndev_name;



	ndev = rxe_ib_device_get_netdev(&rxe->ib_dev);

	if (!ndev)

		return NULL;

	ndev_name = ndev->name;

	dev_put(ndev);



	return ndev_name;

}



int rxe_net_add(const char *ibdev_name, struct net_device *ndev)
@@ -536,10 +545,9 @@ int rxe_net_add(const char *ibdev_name, struct net_device *ndev) 
	if (!rxe)

		return -ENOMEM;



	rxe->ndev = ndev;

	ib_mark_name_assigned_by_user(&rxe->ib_dev);



	err = rxe_add(rxe, ndev->mtu, ibdev_name);

	err = rxe_add(rxe, ndev->mtu, ibdev_name, ndev);

	if (err) {

		ib_dealloc_device(&rxe->ib_dev);

		return err;
@@ -587,10 +595,18 @@ void rxe_port_down(struct rxe_dev *rxe) 


void rxe_set_port_state(struct rxe_dev *rxe)

{

	if (netif_running(rxe->ndev) && netif_carrier_ok(rxe->ndev))

	struct net_device *ndev;



	ndev = rxe_ib_device_get_netdev(&rxe->ib_dev);

	if (!ndev)

		return;



	if (netif_running(ndev) && netif_carrier_ok(ndev))

		rxe_port_up(rxe);

	else

		rxe_port_down(rxe);



	dev_put(ndev);

}



static int rxe_notify(struct notifier_block *not_blk,

drivers/infiniband/sw/rxe/rxe_verbs.c

+21 −5
Original line number Diff line number Diff line
@@ -41,6 +41,7 @@ static int rxe_query_port(struct ib_device *ibdev, 
			  u32 port_num, struct ib_port_attr *attr)

{

	struct rxe_dev *rxe = to_rdev(ibdev);

	struct net_device *ndev;

	int err, ret;



	if (port_num != 1) {
@@ -49,6 +50,12 @@ static int rxe_query_port(struct ib_device *ibdev, 
		goto err_out;

	}



	ndev = rxe_ib_device_get_netdev(ibdev);

	if (!ndev) {

		err = -ENODEV;

		goto err_out;

	}



	memcpy(attr, &rxe->port.attr, sizeof(*attr));



	mutex_lock(&rxe->usdev_lock);
@@ -57,13 +64,14 @@ static int rxe_query_port(struct ib_device *ibdev, 


	if (attr->state == IB_PORT_ACTIVE)

		attr->phys_state = IB_PORT_PHYS_STATE_LINK_UP;

	else if (dev_get_flags(rxe->ndev) & IFF_UP)

	else if (dev_get_flags(ndev) & IFF_UP)

		attr->phys_state = IB_PORT_PHYS_STATE_POLLING;

	else

		attr->phys_state = IB_PORT_PHYS_STATE_DISABLED;



	mutex_unlock(&rxe->usdev_lock);



	dev_put(ndev);

	return ret;



err_out:
@@ -1425,9 +1433,16 @@ static const struct attribute_group rxe_attr_group = { 
static int rxe_enable_driver(struct ib_device *ib_dev)

{

	struct rxe_dev *rxe = container_of(ib_dev, struct rxe_dev, ib_dev);

	struct net_device *ndev;



	ndev = rxe_ib_device_get_netdev(ib_dev);

	if (!ndev)

		return -ENODEV;



	rxe_set_port_state(rxe);

	dev_info(&rxe->ib_dev.dev, "added %s\n", netdev_name(rxe->ndev));

	dev_info(&rxe->ib_dev.dev, "added %s\n", netdev_name(ndev));



	dev_put(ndev);

	return 0;

}
@@ -1495,7 +1510,8 @@ static const struct ib_device_ops rxe_dev_ops = { 
	INIT_RDMA_OBJ_SIZE(ib_mw, rxe_mw, ibmw),

};



int rxe_register_device(struct rxe_dev *rxe, const char *ibdev_name)

int rxe_register_device(struct rxe_dev *rxe, const char *ibdev_name,

						struct net_device *ndev)

{

	int err;

	struct ib_device *dev = &rxe->ib_dev;
@@ -1507,13 +1523,13 @@ int rxe_register_device(struct rxe_dev *rxe, const char *ibdev_name) 
	dev->num_comp_vectors = num_possible_cpus();

	dev->local_dma_lkey = 0;

	addrconf_addr_eui48((unsigned char *)&dev->node_guid,

			    rxe->ndev->dev_addr);

			    ndev->dev_addr);



	dev->uverbs_cmd_mask |= BIT_ULL(IB_USER_VERBS_CMD_POST_SEND) |

				BIT_ULL(IB_USER_VERBS_CMD_REQ_NOTIFY_CQ);



	ib_set_device_ops(dev, &rxe_dev_ops);

	err = ib_device_set_netdev(&rxe->ib_dev, rxe->ndev, 1);

	err = ib_device_set_netdev(&rxe->ib_dev, ndev, 1);

	if (err)

		return err;