kernel.git/net/can, branch linux-6.18.y

can: bcm: fix locking for bcm_op runtime updates

2026-03-12T11:09:51Z

[ Upstream commit c35636e91e392e1540949bbc67932167cb48bc3a ] Commit c2aba69d0c36 ("can: bcm: add locking for bcm_op runtime updates") added a locking for some variables that can be modified at runtime when updating the sending bcm_op with a new TX_SETUP command in bcm_tx_setup(). Usually the RX_SETUP only handles and filters incoming traffic with one exception: When the RX_RTR_FRAME flag is set a predefined CAN frame is sent when a specific RTR frame is received. Therefore the rx bcm_op uses bcm_can_tx() which uses the bcm_tx_lock that was only initialized in bcm_tx_setup(). Add the missing spin_lock_init() when allocating the bcm_op in bcm_rx_setup() to handle the RTR case properly. Fixes: c2aba69d0c36 ("can: bcm: add locking for bcm_op runtime updates") Reported-by: syzbot+5b11eccc403dd1cea9f8@syzkaller.appspotmail.com Closes: https://lore.kernel.org/linux-can/699466e4.a70a0220.2c38d7.00ff.GAE@google.com/ Signed-off-by: Oliver Hartkopp Link: https://patch.msgid.link/20260218-bcm_spin_lock_init-v1-1-592634c8a5b5@hartkopp.net Signed-off-by: Marc Kleine-Budde Signed-off-by: Sasha Levin

net: can: j1939: j1939_xtp_rx_rts_session_active(): deactivate session upon receiving the second rts

2026-01-23T10:21:24Z

commit 1809c82aa073a11b7d335ae932d81ce51a588a4a upstream. Since j1939_session_deactivate_activate_next() in j1939_tp_rxtimer() is called only when the timer is enabled, we need to call j1939_session_deactivate_activate_next() if we cancelled the timer. Otherwise, refcount for j1939_session leaks, which will later appear as | unregister_netdevice: waiting for vcan0 to become free. Usage count = 2. problem. Reported-by: syzbot Closes: https://syzkaller.appspot.com/bug?extid=881d65229ca4f9ae8c84 Signed-off-by: Tetsuo Handa Tested-by: Oleksij Rempel Acked-by: Oleksij Rempel Fixes: 9d71dd0c7009 ("can: add support of SAE J1939 protocol") Link: https://patch.msgid.link/b1212653-8fa1-44e1-be9d-12f950fb3a07@I-love.SAKURA.ne.jp Cc: stable@vger.kernel.org Signed-off-by: Marc Kleine-Budde Signed-off-by: Greg Kroah-Hartman

can: j1939: make j1939_session_activate() fail if device is no longer registered

2026-01-17T15:35:33Z

[ Upstream commit 5d5602236f5db19e8b337a2cd87a90ace5ea776d ] syzbot is still reporting unregister_netdevice: waiting for vcan0 to become free. Usage count = 2 even after commit 93a27b5891b8 ("can: j1939: add missing calls in NETDEV_UNREGISTER notification handler") was added. A debug printk() patch found that j1939_session_activate() can succeed even after j1939_cancel_active_session() from j1939_netdev_notify(NETDEV_UNREGISTER) has completed. Since j1939_cancel_active_session() is processed with the session list lock held, checking ndev->reg_state in j1939_session_activate() with the session list lock held can reliably close the race window. Reported-by: syzbot Closes: https://syzkaller.appspot.com/bug?extid=881d65229ca4f9ae8c84 Signed-off-by: Tetsuo Handa Acked-by: Oleksij Rempel Link: https://patch.msgid.link/b9653191-d479-4c8b-8536-1326d028db5c@I-love.SAKURA.ne.jp Signed-off-by: Marc Kleine-Budde Signed-off-by: Sasha Levin

can: j1939: make j1939_sk_bind() fail if device is no longer registered

2026-01-02T11:56:48Z

[ Upstream commit 46cea215dc9444ec32a76b1b6a9cb809e17b64d5 ] There is a theoretical race window in j1939_sk_netdev_event_unregister() where two j1939_sk_bind() calls jump in between read_unlock_bh() and lock_sock(). The assumption jsk->priv == priv can fail if the first j1939_sk_bind() call once made jsk->priv == NULL due to failed j1939_local_ecu_get() call and the second j1939_sk_bind() call again made jsk->priv != NULL due to successful j1939_local_ecu_get() call. Since the socket lock is held by both j1939_sk_netdev_event_unregister() and j1939_sk_bind(), checking ndev->reg_state with the socket lock held can reliably make the second j1939_sk_bind() call fail (and close this race window). Fixes: 7fcbe5b2c6a4 ("can: j1939: implement NETDEV_UNREGISTER notification handler") Signed-off-by: Tetsuo Handa Acked-by: Oleksij Rempel Link: https://patch.msgid.link/5732921e-247e-4957-a364-da74bd7031d7@I-love.SAKURA.ne.jp Signed-off-by: Marc Kleine-Budde Signed-off-by: Sasha Levin

can: j1939: add missing calls in NETDEV_UNREGISTER notification handler

2025-10-13T19:26:31Z

Currently NETDEV_UNREGISTER event handler is not calling j1939_cancel_active_session() and j1939_sk_queue_drop_all(). This will result in these calls being skipped when j1939_sk_release() is called. And I guess that the reason syzbot is still reporting unregister_netdevice: waiting for vcan0 to become free. Usage count = 2 is caused by lack of these calls. Calling j1939_cancel_active_session(priv, sk) from j1939_sk_release() can be covered by calling j1939_cancel_active_session(priv, NULL) from j1939_netdev_notify(). Calling j1939_sk_queue_drop_all() from j1939_sk_release() can be covered by calling j1939_sk_netdev_event_netdown() from j1939_netdev_notify(). Therefore, we can reuse j1939_cancel_active_session(priv, NULL) and j1939_sk_netdev_event_netdown(priv) for NETDEV_UNREGISTER event handler. Fixes: 7fcbe5b2c6a4 ("can: j1939: implement NETDEV_UNREGISTER notification handler") Signed-off-by: Tetsuo Handa Tested-by: Oleksij Rempel Acked-by: Oleksij Rempel Link: https://patch.msgid.link/3ad3c7f8-5a74-4b07-a193-cb0725823558@I-love.SAKURA.ne.jp Signed-off-by: Marc Kleine-Budde

can: annotate mtu accesses with READ_ONCE()

2025-09-23T08:04:58Z

As hinted in commit 501a90c94510 ("inet: protect against too small mtu values."), net_device->mtu is vulnerable to race conditions if it is written and read without holding the RTNL. At the moment, all the writes are done while the interface is down, either in the devices' probe() function or in can_changelink(). So there are no such issues yet. But upcoming changes will allow to modify the MTU while the CAN XL devices are up. In preparation to the introduction of CAN XL, annotate all the net_device->mtu accesses which are not yet guarded by the RTNL with a READ_ONCE(). Note that all the write accesses are already either guarded by the RTNL or are already annotated and thus need no changes. Signed-off-by: Vincent Mailhol Link: https://patch.msgid.link/20250923-can-fix-mtu-v3-1-581bde113f52@kernel.org Signed-off-by: Marc Kleine-Budde

can: raw: reorder struct raw_sock's members to optimise packing

2025-09-19T18:56:35Z

struct raw_sock has several holes. Reorder the fields to save 8 bytes. Statistics before: $ pahole --class_name=raw_sock net/can/raw.o struct raw_sock { struct sock sk __attribute__((__aligned__(8))); /* 0 776 */ /* XXX last struct has 1 bit hole */ /* --- cacheline 12 boundary (768 bytes) was 8 bytes ago --- */ int ifindex; /* 776 4 */ /* XXX 4 bytes hole, try to pack */ struct net_device * dev; /* 784 8 */ netdevice_tracker dev_tracker; /* 792 0 */ struct list_head notifier; /* 792 16 */ unsigned int bound:1; /* 808: 0 4 */ unsigned int loopback:1; /* 808: 1 4 */ unsigned int recv_own_msgs:1; /* 808: 2 4 */ unsigned int fd_frames:1; /* 808: 3 4 */ unsigned int xl_frames:1; /* 808: 4 4 */ unsigned int join_filters:1; /* 808: 5 4 */ /* XXX 2 bits hole, try to pack */ /* Bitfield combined with next fields */ struct can_raw_vcid_options raw_vcid_opts; /* 809 4 */ /* XXX 3 bytes hole, try to pack */ canid_t tx_vcid_shifted; /* 816 4 */ canid_t rx_vcid_shifted; /* 820 4 */ canid_t rx_vcid_mask_shifted; /* 824 4 */ int count; /* 828 4 */ /* --- cacheline 13 boundary (832 bytes) --- */ struct can_filter dfilter; /* 832 8 */ struct can_filter * filter; /* 840 8 */ can_err_mask_t err_mask; /* 848 4 */ /* XXX 4 bytes hole, try to pack */ struct uniqframe * uniq; /* 856 8 */ /* size: 864, cachelines: 14, members: 20 */ /* sum members: 852, holes: 3, sum holes: 11 */ /* sum bitfield members: 6 bits, bit holes: 1, sum bit holes: 2 bits */ /* member types with bit holes: 1, total: 1 */ /* forced alignments: 1 */ /* last cacheline: 32 bytes */ } __attribute__((__aligned__(8))); ...and after: $ pahole --class_name=raw_sock net/can/raw.o struct raw_sock { struct sock sk __attribute__((__aligned__(8))); /* 0 776 */ /* XXX last struct has 1 bit hole */ /* --- cacheline 12 boundary (768 bytes) was 8 bytes ago --- */ struct net_device * dev; /* 776 8 */ netdevice_tracker dev_tracker; /* 784 0 */ struct list_head notifier; /* 784 16 */ int ifindex; /* 800 4 */ unsigned int bound:1; /* 804: 0 4 */ unsigned int loopback:1; /* 804: 1 4 */ unsigned int recv_own_msgs:1; /* 804: 2 4 */ unsigned int fd_frames:1; /* 804: 3 4 */ unsigned int xl_frames:1; /* 804: 4 4 */ unsigned int join_filters:1; /* 804: 5 4 */ /* XXX 2 bits hole, try to pack */ /* Bitfield combined with next fields */ struct can_raw_vcid_options raw_vcid_opts; /* 805 4 */ /* XXX 3 bytes hole, try to pack */ canid_t tx_vcid_shifted; /* 812 4 */ canid_t rx_vcid_shifted; /* 816 4 */ canid_t rx_vcid_mask_shifted; /* 820 4 */ can_err_mask_t err_mask; /* 824 4 */ int count; /* 828 4 */ /* --- cacheline 13 boundary (832 bytes) --- */ struct can_filter dfilter; /* 832 8 */ struct can_filter * filter; /* 840 8 */ struct uniqframe * uniq; /* 848 8 */ /* size: 856, cachelines: 14, members: 20 */ /* sum members: 852, holes: 1, sum holes: 3 */ /* sum bitfield members: 6 bits, bit holes: 1, sum bit holes: 2 bits */ /* member types with bit holes: 1, total: 1 */ /* forced alignments: 1 */ /* last cacheline: 24 bytes */ } __attribute__((__aligned__(8))); Acked-by: Oliver Hartkopp Signed-off-by: Vincent Mailhol Link: https://patch.msgid.link/20250917-can-raw-repack-v2-3-395e8b3a4437@kernel.org Signed-off-by: Marc Kleine-Budde

can: raw: use bitfields to store flags in struct raw_sock

2025-09-19T18:56:34Z

The bound, loopback, recv_own_msgs, fd_frames, xl_frames and join_filters fields of struct raw_sock just need to store one bit of information. Declare all those members as a bitfields of type unsigned int and width one bit. Add a temporary variable to raw_setsockopt() and raw_getsockopt() to make the conversion between the stored bits and the socket interface. This reduces the size of struct raw_sock by sixteen bytes. Statistics before: $ pahole --class_name=raw_sock net/can/raw.o struct raw_sock { struct sock sk __attribute__((__aligned__(8))); /* 0 776 */ /* XXX last struct has 1 bit hole */ /* --- cacheline 12 boundary (768 bytes) was 8 bytes ago --- */ int bound; /* 776 4 */ int ifindex; /* 780 4 */ struct net_device * dev; /* 784 8 */ netdevice_tracker dev_tracker; /* 792 0 */ struct list_head notifier; /* 792 16 */ int loopback; /* 808 4 */ int recv_own_msgs; /* 812 4 */ int fd_frames; /* 816 4 */ int xl_frames; /* 820 4 */ struct can_raw_vcid_options raw_vcid_opts; /* 824 4 */ canid_t tx_vcid_shifted; /* 828 4 */ /* --- cacheline 13 boundary (832 bytes) --- */ canid_t rx_vcid_shifted; /* 832 4 */ canid_t rx_vcid_mask_shifted; /* 836 4 */ int join_filters; /* 840 4 */ int count; /* 844 4 */ struct can_filter dfilter; /* 848 8 */ struct can_filter * filter; /* 856 8 */ can_err_mask_t err_mask; /* 864 4 */ /* XXX 4 bytes hole, try to pack */ struct uniqframe * uniq; /* 872 8 */ /* size: 880, cachelines: 14, members: 20 */ /* sum members: 876, holes: 1, sum holes: 4 */ /* member types with bit holes: 1, total: 1 */ /* forced alignments: 1 */ /* last cacheline: 48 bytes */ } __attribute__((__aligned__(8))); ...and after: $ pahole --class_name=raw_sock net/can/raw.o struct raw_sock { struct sock sk __attribute__((__aligned__(8))); /* 0 776 */ /* XXX last struct has 1 bit hole */ /* --- cacheline 12 boundary (768 bytes) was 8 bytes ago --- */ int ifindex; /* 776 4 */ /* XXX 4 bytes hole, try to pack */ struct net_device * dev; /* 784 8 */ netdevice_tracker dev_tracker; /* 792 0 */ struct list_head notifier; /* 792 16 */ unsigned int bound:1; /* 808: 0 4 */ unsigned int loopback:1; /* 808: 1 4 */ unsigned int recv_own_msgs:1; /* 808: 2 4 */ unsigned int fd_frames:1; /* 808: 3 4 */ unsigned int xl_frames:1; /* 808: 4 4 */ unsigned int join_filters:1; /* 808: 5 4 */ /* XXX 2 bits hole, try to pack */ /* Bitfield combined with next fields */ struct can_raw_vcid_options raw_vcid_opts; /* 809 4 */ /* XXX 3 bytes hole, try to pack */ canid_t tx_vcid_shifted; /* 816 4 */ canid_t rx_vcid_shifted; /* 820 4 */ canid_t rx_vcid_mask_shifted; /* 824 4 */ int count; /* 828 4 */ /* --- cacheline 13 boundary (832 bytes) --- */ struct can_filter dfilter; /* 832 8 */ struct can_filter * filter; /* 840 8 */ can_err_mask_t err_mask; /* 848 4 */ /* XXX 4 bytes hole, try to pack */ struct uniqframe * uniq; /* 856 8 */ /* size: 864, cachelines: 14, members: 20 */ /* sum members: 852, holes: 3, sum holes: 11 */ /* sum bitfield members: 6 bits, bit holes: 1, sum bit holes: 2 bits */ /* member types with bit holes: 1, total: 1 */ /* forced alignments: 1 */ /* last cacheline: 32 bytes */ } __attribute__((__aligned__(8))); Acked-by: Oliver Hartkopp Signed-off-by: Vincent Mailhol Link: https://patch.msgid.link/20250917-can-raw-repack-v2-2-395e8b3a4437@kernel.org Signed-off-by: Marc Kleine-Budde

can: raw: reorder struct uniqframe's members to optimise packing

2025-09-19T18:56:34Z

struct uniqframe has one hole. Reorder the fields to save 8 bytes. Statistics before: $ pahole --class_name=uniqframe net/can/raw.o struct uniqframe { int skbcnt; /* 0 4 */ /* XXX 4 bytes hole, try to pack */ const struct sk_buff * skb; /* 8 8 */ unsigned int join_rx_count; /* 16 4 */ /* size: 24, cachelines: 1, members: 3 */ /* sum members: 16, holes: 1, sum holes: 4 */ /* padding: 4 */ /* last cacheline: 24 bytes */ }; ...and after: $ pahole --class_name=uniqframe net/can/raw.o struct uniqframe { const struct sk_buff * skb; /* 0 8 */ int skbcnt; /* 8 4 */ unsigned int join_rx_count; /* 12 4 */ /* size: 16, cachelines: 1, members: 3 */ /* last cacheline: 16 bytes */ }; Acked-by: Oliver Hartkopp Signed-off-by: Vincent Mailhol Link: https://patch.msgid.link/20250917-can-raw-repack-v2-1-395e8b3a4437@kernel.org Signed-off-by: Marc Kleine-Budde

can: j1939: j1939_local_ecu_get(): undo increment when j1939_local_ecu_get() fails

2025-09-10T15:12:05Z

Since j1939_sk_bind() and j1939_sk_release() call j1939_local_ecu_put() when J1939_SOCK_BOUND was already set, but the error handling path for j1939_sk_bind() will not set J1939_SOCK_BOUND when j1939_local_ecu_get() fails, j1939_local_ecu_get() needs to undo priv->ents[sa].nusers++ when j1939_local_ecu_get() returns an error. Fixes: 9d71dd0c7009 ("can: add support of SAE J1939 protocol") Signed-off-by: Tetsuo Handa Tested-by: Oleksij Rempel Acked-by: Oleksij Rempel Link: https://patch.msgid.link/e7f80046-4ff7-4ce2-8ad8-7c3c678a42c9@I-love.SAKURA.ne.jp Signed-off-by: Marc Kleine-Budde