kernel/linux.git/drivers/net/usb/usbnet.c, branch linux-6.19.y

qmi_wwan: allow max_mtu above hard_mtu to control rx_urb_size

2026-03-19T15:15:25+00:00

commit 55f854dd5bdd8e19b936a00ef1f8d776ac32c7b0 upstream. Commit c7159e960f14 ("usbnet: limit max_mtu based on device's hard_mtu") capped net->max_mtu to the device's hard_mtu in usbnet_probe(). While this correctly prevents oversized packets on standard USB network devices, it breaks the qmi_wwan driver. qmi_wwan relies on userspace (e.g. ModemManager) setting a large MTU on the wwan0 interface to configure rx_urb_size via usbnet_change_mtu(). QMI modems negotiate USB transfer sizes of 16,383 or 32,767 bytes, and the USB receive buffers must be sized accordingly. With max_mtu capped to hard_mtu (~1500 bytes), userspace can no longer raise the MTU, the receive buffers remain small, and download speeds drop from >300 Mbps to ~0.8 Mbps. Introduce a FLAG_NOMAXMTU driver flag that allows individual usbnet drivers to opt out of the max_mtu cap. Set this flag in qmi_wwan's driver_info structures to restore the previous behavior for QMI devices, while keeping the safety fix in place for all other usbnet drivers. Fixes: c7159e960f14 ("usbnet: limit max_mtu based on device's hard_mtu") Cc: stable@vger.kernel.org Link: https://lore.kernel.org/lkml/CAPh3n803k8JcBPV5qEzUB-oKzWkAs-D5CU7z=Vd_nLRCr5ZqQg@mail.gmail.com/ Reported-by: Koen Vandeputte Tested-by: Daniele Palmas Signed-off-by: Laurent Vivier Link: https://patch.msgid.link/20260304134338.1785002-1-lvivier@redhat.com Signed-off-by: Jakub Kicinski Signed-off-by: Greg Kroah-Hartman

usbnet: limit max_mtu based on device's hard_mtu

2026-01-21T02:58:54+00:00

The usbnet driver initializes net->max_mtu to ETH_MAX_MTU before calling the device's bind() callback. When the bind() callback sets dev->hard_mtu based the device's actual capability (from CDC Ethernet's wMaxSegmentSize descriptor), max_mtu is never updated to reflect this hardware limitation). This allows userspace (DHCP or IPv6 RA) to configure MTU larger than the device can handle, leading to silent packet drops when the backend sends packet exceeding the device's buffer size. Fix this by limiting net->max_mtu to the device's hard_mtu after the bind callback returns. See https://gitlab.com/qemu-project/qemu/-/issues/3268 and https://bugs.passt.top/attachment.cgi?bugid=189 Fixes: f77f0aee4da4 ("net: use core MTU range checking in USB NIC drivers") Signed-off-by: Laurent Vivier Link: https://bugs.passt.top/show_bug.cgi?id=189 Reviewed-by: Stefano Brivio Link: https://patch.msgid.link/20260119075518.2774373-1-lvivier@redhat.com Signed-off-by: Jakub Kicinski

usbnet: fix crash due to missing BQL accounting after resume

2026-01-16T03:40:59+00:00

In commit 7ff14c52049e ("usbnet: Add support for Byte Queue Limits (BQL)"), it was missed that usbnet_resume() may enqueue SKBs using __skb_queue_tail() without reporting them to BQL. As a result, the next call to netdev_completed_queue() triggers a BUG_ON() in dql_completed(), since the SKBs queued during resume were never accounted for. This patch fixes the issue by adding a corresponding netdev_sent_queue() call in usbnet_resume() when SKBs are queued after suspend. Because dev->txq.lock is held at this point, no concurrent calls to netdev_sent_queue() from usbnet_start_xmit() can occur. The crash can be reproduced by generating network traffic (e.g. iperf3 -c ... -t 0), suspending the system, and then waking it up (e.g. rtcwake -m mem -s 5). When testing USB2 Android tethering (cdc_ncm), the system crashed within three suspend/resume cycles without this patch. With the patch applied, no crashes were observed after 90 cycles. Testing with an AX88179 USB Ethernet adapter also showed no crashes. Fixes: 7ff14c52049e ("usbnet: Add support for Byte Queue Limits (BQL)") Reported-by: Bard Liao Tested-by: Bard Liao Tested-by: Simon Schippers Signed-off-by: Simon Schippers Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/20260113075139.6735-1-simon.schippers@tu-dortmund.de Signed-off-by: Jakub Kicinski

usbnet: avoid a possible crash in dql_completed()

2025-12-29T18:23:34+00:00

syzbot reported a crash [1] in dql_completed() after recent usbnet BQL adoption. The reason for the crash is that netdev_reset_queue() is called too soon. It should be called after cancel_work_sync(&dev->bh_work) to make sure no more TX completion can happen. [1] kernel BUG at lib/dynamic_queue_limits.c:99 ! Oops: invalid opcode: 0000 [#1] SMP KASAN PTI CPU: 1 UID: 0 PID: 5197 Comm: udevd Tainted: G L syzkaller #0 PREEMPT(full) Tainted: [L]=SOFTLOCKUP Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/25/2025 RIP: 0010:dql_completed+0xbe1/0xbf0 lib/dynamic_queue_limits.c:99 Call Trace: netdev_tx_completed_queue include/linux/netdevice.h:3864 [inline] netdev_completed_queue include/linux/netdevice.h:3894 [inline] usbnet_bh+0x793/0x1020 drivers/net/usb/usbnet.c:1601 process_one_work kernel/workqueue.c:3257 [inline] process_scheduled_works+0xad1/0x1770 kernel/workqueue.c:3340 bh_worker+0x2b1/0x600 kernel/workqueue.c:3611 tasklet_action+0xc/0x70 kernel/softirq.c:952 handle_softirqs+0x27d/0x850 kernel/softirq.c:622 __do_softirq kernel/softirq.c:656 [inline] invoke_softirq kernel/softirq.c:496 [inline] __irq_exit_rcu+0xca/0x1f0 kernel/softirq.c:723 irq_exit_rcu+0x9/0x30 kernel/softirq.c:739 Fixes: 7ff14c52049e ("usbnet: Add support for Byte Queue Limits (BQL)") Reported-by: syzbot+5b55e49f8bbd84631a9c@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/6945644f.a70a0220.207337.0113.GAE@google.com/T/#u Signed-off-by: Eric Dumazet Cc: Simon Schippers Link: https://patch.msgid.link/20251219144459.692715-1-edumazet@google.com Signed-off-by: Paolo Abeni

net: usb: usbnet: adhere to style

2025-11-14T01:47:38+00:00

This satisfies the coding style. Signed-off-by: Oliver Neukum Link: https://patch.msgid.link/20251112102610.281565-1-oneukum@suse.com Signed-off-by: Jakub Kicinski

usbnet: Add support for Byte Queue Limits (BQL)

2025-11-11T01:28:38+00:00

In the current implementation, usbnet uses a fixed tx_qlen of: USB2: 60 * 1518 bytes = 91.08 KB USB3: 60 * 5 * 1518 bytes = 454.80 KB Such large transmit queues can be problematic, especially for cellular modems. For example, with a typical celluar link speed of 10 Mbit/s, a fully occupied USB3 transmit queue results in: 454.80 KB / (10 Mbit/s / 8 bit/byte) = 363.84 ms of additional latency. This patch adds support for Byte Queue Limits (BQL) [1] to dynamically manage the transmit queue size and reduce latency without sacrificing throughput. Testing was performed on various devices using the usbnet driver for packet transmission: - DELOCK 66045: USB3 to 2.5 GbE adapter (ax88179_178a) - DELOCK 61969: USB2 to 1 GbE adapter (asix) - Quectel RM520: 5G modem (qmi_wwan) - USB2 Android tethering (cdc_ncm) No performance degradation was observed for iperf3 TCP or UDP traffic, while latency for a prioritized ping application was significantly reduced. For example, using the USB3 to 2.5 GbE adapter, which was fully utilized by iperf3 UDP traffic, the prioritized ping was improved from 1.6 ms to 0.6 ms. With the same setup but with a 100 Mbit/s Ethernet connection, the prioritized ping was improved from 35 ms to 5 ms. [1] https://lwn.net/Articles/469652/ Signed-off-by: Simon Schippers Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/20251106175615.26948-1-simon.schippers@tu-dortmund.de Signed-off-by: Jakub Kicinski

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2025-10-31T13:46:03+00:00

Cross-merge networking fixes after downstream PR (net-6.18-rc4). No conflicts, adjacent changes: drivers/net/ethernet/stmicro/stmmac/stmmac_main.c ded9813d17d3 ("net: stmmac: Consider Tx VLAN offload tag length for maxSDU") 26ab9830beab ("net: stmmac: replace has_xxxx with core_type") Signed-off-by: Jakub Kicinski

net: usb: usbnet: coding style for functions

2025-10-25T02:02:29+00:00

Functions are not to have blanks between names and parameter lists. Remove them. Signed-off-by: Oliver Neukum Link: https://patch.msgid.link/20251023100136.909118-1-oneukum@suse.com Signed-off-by: Jakub Kicinski

usbnet: Prevents free active kevent

2025-10-24T00:27:44+00:00

The root cause of this issue are: 1. When probing the usbnet device, executing usbnet_link_change(dev, 0, 0); put the kevent work in global workqueue. However, the kevent has not yet been scheduled when the usbnet device is unregistered. Therefore, executing free_netdev() results in the "free active object (kevent)" error reported here. 2. Another factor is that when calling usbnet_disconnect()->unregister_netdev(), if the usbnet device is up, ndo_stop() is executed to cancel the kevent. However, because the device is not up, ndo_stop() is not executed. The solution to this problem is to cancel the kevent before executing free_netdev(). Fixes: a69e617e533e ("usbnet: Fix linkwatch use-after-free on disconnect") Reported-by: Sam Sun Closes: https://syzkaller.appspot.com/bug?extid=8bfd7bcc98f7300afb84 Signed-off-by: Lizhi Xu Link: https://patch.msgid.link/20251022024007.1831898-1-lizhi.xu@windriver.com Signed-off-by: Jakub Kicinski

usbnet: Fix using smp_processor_id() in preemptible code warnings

2025-10-14T09:55:53+00:00

Syzbot reported the following warning: BUG: using smp_processor_id() in preemptible [00000000] code: dhcpcd/2879 caller is usbnet_skb_return+0x74/0x490 drivers/net/usb/usbnet.c:331 CPU: 1 UID: 0 PID: 2879 Comm: dhcpcd Not tainted 6.15.0-rc4-syzkaller-00098-g615dca38c2ea #0 PREEMPT(voluntary) Call Trace: __dump_stack lib/dump_stack.c:94 [inline] dump_stack_lvl+0x16c/0x1f0 lib/dump_stack.c:120 check_preemption_disabled+0xd0/0xe0 lib/smp_processor_id.c:49 usbnet_skb_return+0x74/0x490 drivers/net/usb/usbnet.c:331 usbnet_resume_rx+0x4b/0x170 drivers/net/usb/usbnet.c:708 usbnet_change_mtu+0x1be/0x220 drivers/net/usb/usbnet.c:417 __dev_set_mtu net/core/dev.c:9443 [inline] netif_set_mtu_ext+0x369/0x5c0 net/core/dev.c:9496 netif_set_mtu+0xb0/0x160 net/core/dev.c:9520 dev_set_mtu+0xae/0x170 net/core/dev_api.c:247 dev_ifsioc+0xa31/0x18d0 net/core/dev_ioctl.c:572 dev_ioctl+0x223/0x10e0 net/core/dev_ioctl.c:821 sock_do_ioctl+0x19d/0x280 net/socket.c:1204 sock_ioctl+0x42f/0x6a0 net/socket.c:1311 vfs_ioctl fs/ioctl.c:51 [inline] __do_sys_ioctl fs/ioctl.c:906 [inline] __se_sys_ioctl fs/ioctl.c:892 [inline] __x64_sys_ioctl+0x190/0x200 fs/ioctl.c:892 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline] do_syscall_64+0xcd/0x260 arch/x86/entry/syscall_64.c:94 entry_SYSCALL_64_after_hwframe+0x77/0x7f For historical and portability reasons, the netif_rx() is usually run in the softirq or interrupt context, this commit therefore add local_bh_disable/enable() protection in the usbnet_resume_rx(). Fixes: 43daa96b166c ("usbnet: Stop RX Q on MTU change") Link: https://syzkaller.appspot.com/bug?id=81f55dfa587ee544baaaa5a359a060512228c1e1 Suggested-by: Jakub Kicinski Signed-off-by: Zqiang Link: https://patch.msgid.link/20251011070518.7095-1-qiang.zhang@linux.dev Signed-off-by: Paolo Abeni