From 0c73772cd2b8cc108d5f5334de89ad648d89b9ec Mon Sep 17 00:00:00 2001 From: Seungjin Bae Date: Thu, 23 Oct 2025 12:27:09 -0400 Subject: [PATCH 01/47] can: kvaser_usb: leaf: Fix potential infinite loop in command parsers The `kvaser_usb_leaf_wait_cmd()` and `kvaser_usb_leaf_read_bulk_callback` functions contain logic to zero-length commands. These commands are used to align data to the USB endpoint's wMaxPacketSize boundary. The driver attempts to skip these placeholders by aligning the buffer position `pos` to the next packet boundary using `round_up()` function. However, if zero-length command is found exactly on a packet boundary (i.e., `pos` is a multiple of wMaxPacketSize, including 0), `round_up` function will return the unchanged value of `pos`. This prevents `pos` to be increased, causing an infinite loop in the parsing logic. This patch fixes this in the function by using `pos + 1` instead. This ensures that even if `pos` is on a boundary, the calculation is based on `pos + 1`, forcing `round_up()` to always return the next aligned boundary. Fixes: 7259124eac7d ("can: kvaser_usb: Split driver into kvaser_usb_core.c and kvaser_usb_leaf.c") Signed-off-by: Seungjin Bae Reviewed-by: Jimmy Assarsson Tested-by: Jimmy Assarsson Link: https://patch.msgid.link/20251023162709.348240-1-eeodqql09@gmail.com Signed-off-by: Marc Kleine-Budde --- drivers/net/can/usb/kvaser_usb/kvaser_usb_leaf.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/can/usb/kvaser_usb/kvaser_usb_leaf.c b/drivers/net/can/usb/kvaser_usb/kvaser_usb_leaf.c index c29828a94ad0..1167d38344f1 100644 --- a/drivers/net/can/usb/kvaser_usb/kvaser_usb_leaf.c +++ b/drivers/net/can/usb/kvaser_usb/kvaser_usb_leaf.c @@ -685,7 +685,7 @@ static int kvaser_usb_leaf_wait_cmd(const struct kvaser_usb *dev, u8 id, * for further details. */ if (tmp->len == 0) { - pos = round_up(pos, + pos = round_up(pos + 1, le16_to_cpu (dev->bulk_in->wMaxPacketSize)); continue; @@ -1732,7 +1732,7 @@ static void kvaser_usb_leaf_read_bulk_callback(struct kvaser_usb *dev, * number of events in case of a heavy rx load on the bus. */ if (cmd->len == 0) { - pos = round_up(pos, le16_to_cpu + pos = round_up(pos + 1, le16_to_cpu (dev->bulk_in->wMaxPacketSize)); continue; } From 30db4451c7f6aabcada029b15859a76962ec0cf8 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Thomas=20M=C3=BChlbacher?= Date: Sat, 15 Nov 2025 15:34:56 +0000 Subject: [PATCH 02/47] can: sja1000: fix max irq loop handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Reading the interrupt register `SJA1000_IR` causes all of its bits to be reset. If we ever reach the condition of handling more than `SJA1000_MAX_IRQ` IRQs, we will have read the register and reset all its bits but without actually handling the interrupt inside of the loop body. This may, among other issues, cause us to never `netif_wake_queue()` again after a transmission interrupt. Fixes: 429da1cc841b ("can: Driver for the SJA1000 CAN controller") Cc: stable@vger.kernel.org Signed-off-by: Thomas Mühlbacher Acked-by: Oliver Hartkopp Link: https://patch.msgid.link/20251115153437.11419-1-tmuehlbacher@posteo.net Signed-off-by: Marc Kleine-Budde --- drivers/net/can/sja1000/sja1000.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/can/sja1000/sja1000.c b/drivers/net/can/sja1000/sja1000.c index 4d245857ef1c..83476af8adb5 100644 --- a/drivers/net/can/sja1000/sja1000.c +++ b/drivers/net/can/sja1000/sja1000.c @@ -548,8 +548,8 @@ irqreturn_t sja1000_interrupt(int irq, void *dev_id) if (priv->read_reg(priv, SJA1000_IER) == IRQ_OFF) goto out; - while ((isrc = priv->read_reg(priv, SJA1000_IR)) && - (n < SJA1000_MAX_IRQ)) { + while ((n < SJA1000_MAX_IRQ) && + (isrc = priv->read_reg(priv, SJA1000_IR))) { status = priv->read_reg(priv, SJA1000_SR); /* check for absent controller due to hw unplug */ From 516a0cd1c03fa266bb67dd87940a209fd4e53ce7 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Sat, 8 Nov 2025 10:01:01 +0100 Subject: [PATCH 03/47] can: gs_usb: gs_usb_xmit_callback(): fix handling of failed transmitted URBs The driver lacks the cleanup of failed transfers of URBs. This reduces the number of available URBs per error by 1. This leads to reduced performance and ultimately to a complete stop of the transmission. If the sending of a bulk URB fails do proper cleanup: - increase netdev stats - mark the echo_sbk as free - free the driver's context and do accounting - wake the send queue Closes: https://github.com/candle-usb/candleLight_fw/issues/187 Fixes: d08e973a77d1 ("can: gs_usb: Added support for the GS_USB CAN devices") Link: https://patch.msgid.link/20251114-gs_usb-fix-usb-callbacks-v1-1-a29b42eacada@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/usb/gs_usb.c | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) diff --git a/drivers/net/can/usb/gs_usb.c b/drivers/net/can/usb/gs_usb.c index 69b8d6da651b..fa9bab8c89ae 100644 --- a/drivers/net/can/usb/gs_usb.c +++ b/drivers/net/can/usb/gs_usb.c @@ -750,8 +750,21 @@ static void gs_usb_xmit_callback(struct urb *urb) struct gs_can *dev = txc->dev; struct net_device *netdev = dev->netdev; - if (urb->status) - netdev_info(netdev, "usb xmit fail %u\n", txc->echo_id); + if (!urb->status) + return; + + if (urb->status != -ESHUTDOWN && net_ratelimit()) + netdev_info(netdev, "failed to xmit URB %u: %pe\n", + txc->echo_id, ERR_PTR(urb->status)); + + netdev->stats.tx_dropped++; + netdev->stats.tx_errors++; + + can_free_echo_skb(netdev, txc->echo_id, NULL); + gs_free_tx_context(txc); + atomic_dec(&dev->active_tx_urbs); + + netif_wake_queue(netdev); } static netdev_tx_t gs_can_start_xmit(struct sk_buff *skb, From 6fe9f3279f7d2518439a7962c5870c6e9ecbadcf Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Sat, 8 Nov 2025 10:01:02 +0100 Subject: [PATCH 04/47] can: gs_usb: gs_usb_receive_bulk_callback(): check actual_length before accessing header The driver expects to receive a struct gs_host_frame in gs_usb_receive_bulk_callback(). Use struct_group to describe the header of the struct gs_host_frame and check that we have at least received the header before accessing any members of it. To resubmit the URB, do not dereference the pointer chain "dev->parent->hf_size_rx" but use "parent->hf_size_rx" instead. Since "urb->context" contains "parent", it is always defined, while "dev" is not defined if the URB it too short. Fixes: d08e973a77d1 ("can: gs_usb: Added support for the GS_USB CAN devices") Link: https://patch.msgid.link/20251114-gs_usb-fix-usb-callbacks-v1-2-a29b42eacada@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/usb/gs_usb.c | 26 +++++++++++++++++++------- 1 file changed, 19 insertions(+), 7 deletions(-) diff --git a/drivers/net/can/usb/gs_usb.c b/drivers/net/can/usb/gs_usb.c index fa9bab8c89ae..51f8d694104d 100644 --- a/drivers/net/can/usb/gs_usb.c +++ b/drivers/net/can/usb/gs_usb.c @@ -262,13 +262,15 @@ struct canfd_quirk { } __packed; struct gs_host_frame { - u32 echo_id; - __le32 can_id; + struct_group(header, + u32 echo_id; + __le32 can_id; - u8 can_dlc; - u8 channel; - u8 flags; - u8 reserved; + u8 can_dlc; + u8 channel; + u8 flags; + u8 reserved; + ); union { DECLARE_FLEX_ARRAY(struct classic_can, classic_can); @@ -576,6 +578,7 @@ static void gs_usb_receive_bulk_callback(struct urb *urb) int rc; struct net_device_stats *stats; struct gs_host_frame *hf = urb->transfer_buffer; + unsigned int minimum_length; struct gs_tx_context *txc; struct can_frame *cf; struct canfd_frame *cfd; @@ -594,6 +597,15 @@ static void gs_usb_receive_bulk_callback(struct urb *urb) return; } + minimum_length = sizeof(hf->header); + if (urb->actual_length < minimum_length) { + dev_err_ratelimited(&parent->udev->dev, + "short read (actual_length=%u, minimum_length=%u)\n", + urb->actual_length, minimum_length); + + goto resubmit_urb; + } + /* device reports out of range channel id */ if (hf->channel >= parent->channel_cnt) goto device_detach; @@ -687,7 +699,7 @@ static void gs_usb_receive_bulk_callback(struct urb *urb) resubmit_urb: usb_fill_bulk_urb(urb, parent->udev, parent->pipe_in, - hf, dev->parent->hf_size_rx, + hf, parent->hf_size_rx, gs_usb_receive_bulk_callback, parent); rc = usb_submit_urb(urb, GFP_ATOMIC); From 395d988f93861101ec89d0dd9e3b876ae9392a5b Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Sat, 8 Nov 2025 10:01:03 +0100 Subject: [PATCH 05/47] can: gs_usb: gs_usb_receive_bulk_callback(): check actual_length before accessing data The URB received in gs_usb_receive_bulk_callback() contains a struct gs_host_frame. The length of the data after the header depends on the gs_host_frame hf::flags and the active device features (e.g. time stamping). Introduce a new function gs_usb_get_minimum_length() and check that we have at least received the required amount of data before accessing it. Only copy the data to that skb that has actually been received. Fixes: d08e973a77d1 ("can: gs_usb: Added support for the GS_USB CAN devices") Link: https://patch.msgid.link/20251114-gs_usb-fix-usb-callbacks-v1-3-a29b42eacada@pengutronix.de [mkl: rename gs_usb_get_minimum_length() -> +gs_usb_get_minimum_rx_length()] Signed-off-by: Marc Kleine-Budde --- drivers/net/can/usb/gs_usb.c | 59 +++++++++++++++++++++++++++++++++--- 1 file changed, 54 insertions(+), 5 deletions(-) diff --git a/drivers/net/can/usb/gs_usb.c b/drivers/net/can/usb/gs_usb.c index 51f8d694104d..8d8a610f9144 100644 --- a/drivers/net/can/usb/gs_usb.c +++ b/drivers/net/can/usb/gs_usb.c @@ -261,6 +261,11 @@ struct canfd_quirk { u8 quirk; } __packed; +/* struct gs_host_frame::echo_id == GS_HOST_FRAME_ECHO_ID_RX indicates + * a regular RX'ed CAN frame + */ +#define GS_HOST_FRAME_ECHO_ID_RX 0xffffffff + struct gs_host_frame { struct_group(header, u32 echo_id; @@ -570,6 +575,37 @@ gs_usb_get_echo_skb(struct gs_can *dev, struct sk_buff *skb, return len; } +static unsigned int +gs_usb_get_minimum_rx_length(const struct gs_can *dev, const struct gs_host_frame *hf, + unsigned int *data_length_p) +{ + unsigned int minimum_length, data_length = 0; + + if (hf->flags & GS_CAN_FLAG_FD) { + if (hf->echo_id == GS_HOST_FRAME_ECHO_ID_RX) + data_length = can_fd_dlc2len(hf->can_dlc); + + if (dev->feature & GS_CAN_FEATURE_HW_TIMESTAMP) + /* timestamp follows data field of max size */ + minimum_length = struct_size(hf, canfd_ts, 1); + else + minimum_length = sizeof(hf->header) + data_length; + } else { + if (hf->echo_id == GS_HOST_FRAME_ECHO_ID_RX && + !(hf->can_id & cpu_to_le32(CAN_RTR_FLAG))) + data_length = can_cc_dlc2len(hf->can_dlc); + + if (dev->feature & GS_CAN_FEATURE_HW_TIMESTAMP) + /* timestamp follows data field of max size */ + minimum_length = struct_size(hf, classic_can_ts, 1); + else + minimum_length = sizeof(hf->header) + data_length; + } + + *data_length_p = data_length; + return minimum_length; +} + static void gs_usb_receive_bulk_callback(struct urb *urb) { struct gs_usb *parent = urb->context; @@ -578,7 +614,7 @@ static void gs_usb_receive_bulk_callback(struct urb *urb) int rc; struct net_device_stats *stats; struct gs_host_frame *hf = urb->transfer_buffer; - unsigned int minimum_length; + unsigned int minimum_length, data_length; struct gs_tx_context *txc; struct can_frame *cf; struct canfd_frame *cfd; @@ -621,20 +657,33 @@ static void gs_usb_receive_bulk_callback(struct urb *urb) if (!netif_running(netdev)) goto resubmit_urb; - if (hf->echo_id == -1) { /* normal rx */ + minimum_length = gs_usb_get_minimum_rx_length(dev, hf, &data_length); + if (urb->actual_length < minimum_length) { + stats->rx_errors++; + stats->rx_length_errors++; + + if (net_ratelimit()) + netdev_err(netdev, + "short read (actual_length=%u, minimum_length=%u)\n", + urb->actual_length, minimum_length); + + goto resubmit_urb; + } + + if (hf->echo_id == GS_HOST_FRAME_ECHO_ID_RX) { /* normal rx */ if (hf->flags & GS_CAN_FLAG_FD) { skb = alloc_canfd_skb(netdev, &cfd); if (!skb) return; cfd->can_id = le32_to_cpu(hf->can_id); - cfd->len = can_fd_dlc2len(hf->can_dlc); + cfd->len = data_length; if (hf->flags & GS_CAN_FLAG_BRS) cfd->flags |= CANFD_BRS; if (hf->flags & GS_CAN_FLAG_ESI) cfd->flags |= CANFD_ESI; - memcpy(cfd->data, hf->canfd->data, cfd->len); + memcpy(cfd->data, hf->canfd->data, data_length); } else { skb = alloc_can_skb(netdev, &cf); if (!skb) @@ -643,7 +692,7 @@ static void gs_usb_receive_bulk_callback(struct urb *urb) cf->can_id = le32_to_cpu(hf->can_id); can_frame_set_cc_len(cf, hf->can_dlc, dev->can.ctrlmode); - memcpy(cf->data, hf->classic_can->data, 8); + memcpy(cf->data, hf->classic_can->data, data_length); /* ERROR frames tell us information about the controller */ if (le32_to_cpu(hf->can_id) & CAN_ERR_FLAG) From 76544beea7cfe5bcce6d60f53811657b88ec8be1 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Sun, 16 Nov 2025 16:55:26 +0100 Subject: [PATCH 06/47] can: sun4i_can: sun4i_can_interrupt(): fix max irq loop handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Reading the interrupt register `SUN4I_REG_INT_ADDR` causes all of its bits to be reset. If we ever reach the condition of handling more than `SUN4I_CAN_MAX_IRQ` IRQs, we will have read the register and reset all its bits but without actually handling the interrupt inside of the loop body. This may, among other issues, cause us to never `netif_wake_queue()` again after a transmission interrupt. Fixes: 0738eff14d81 ("can: Allwinner A10/A20 CAN Controller support - Kernel module") Cc: stable@vger.kernel.org Co-developed-by: Thomas Mühlbacher Signed-off-by: Thomas Mühlbacher Acked-by: Jernej Skrabec Link: https://patch.msgid.link/20251116-sun4i-fix-loop-v1-1-3d76d3f81950@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/sun4i_can.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/can/sun4i_can.c b/drivers/net/can/sun4i_can.c index 53bfd873de9b..0a7ba0942839 100644 --- a/drivers/net/can/sun4i_can.c +++ b/drivers/net/can/sun4i_can.c @@ -657,8 +657,8 @@ static irqreturn_t sun4i_can_interrupt(int irq, void *dev_id) u8 isrc, status; int n = 0; - while ((isrc = readl(priv->base + SUN4I_REG_INT_ADDR)) && - (n < SUN4I_CAN_MAX_IRQ)) { + while ((n < SUN4I_CAN_MAX_IRQ) && + (isrc = readl(priv->base + SUN4I_REG_INT_ADDR))) { n++; status = readl(priv->base + SUN4I_REG_STA_ADDR); From 4015b979767125cf8a2233a145a3b3af78bfd8fb Mon Sep 17 00:00:00 2001 From: Chris Lu Date: Wed, 12 Nov 2025 15:53:34 +0800 Subject: [PATCH 07/47] Bluetooth: btusb: mediatek: Fix kernel crash when releasing mtk iso interface When performing reset tests and encountering abnormal card drop issues that lead to a kernel crash, it is necessary to perform a null check before releasing resources to avoid attempting to release a null pointer. <4>[ 29.158070] Hardware name: Google Quigon sku196612/196613 board (DT) <4>[ 29.158076] Workqueue: hci0 hci_cmd_sync_work [bluetooth] <4>[ 29.158154] pstate: 20400009 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) <4>[ 29.158162] pc : klist_remove+0x90/0x158 <4>[ 29.158174] lr : klist_remove+0x88/0x158 <4>[ 29.158180] sp : ffffffc0846b3c00 <4>[ 29.158185] pmr_save: 000000e0 <4>[ 29.158188] x29: ffffffc0846b3c30 x28: ffffff80cd31f880 x27: ffffff80c1bdc058 <4>[ 29.158199] x26: dead000000000100 x25: ffffffdbdc624ea3 x24: ffffff80c1bdc4c0 <4>[ 29.158209] x23: ffffffdbdc62a3e6 x22: ffffff80c6c07000 x21: ffffffdbdc829290 <4>[ 29.158219] x20: 0000000000000000 x19: ffffff80cd3e0648 x18: 000000031ec97781 <4>[ 29.158229] x17: ffffff80c1bdc4a8 x16: ffffffdc10576548 x15: ffffff80c1180428 <4>[ 29.158238] x14: 0000000000000000 x13: 000000000000e380 x12: 0000000000000018 <4>[ 29.158248] x11: ffffff80c2a7fd10 x10: 0000000000000000 x9 : 0000000100000000 <4>[ 29.158257] x8 : 0000000000000000 x7 : 7f7f7f7f7f7f7f7f x6 : 2d7223ff6364626d <4>[ 29.158266] x5 : 0000008000000000 x4 : 0000000000000020 x3 : 2e7325006465636e <4>[ 29.158275] x2 : ffffffdc11afeff8 x1 : 0000000000000000 x0 : ffffffdc11be4d0c <4>[ 29.158285] Call trace: <4>[ 29.158290] klist_remove+0x90/0x158 <4>[ 29.158298] device_release_driver_internal+0x20c/0x268 <4>[ 29.158308] device_release_driver+0x1c/0x30 <4>[ 29.158316] usb_driver_release_interface+0x70/0x88 <4>[ 29.158325] btusb_mtk_release_iso_intf+0x68/0xd8 [btusb (HASH:e8b6 5)] <4>[ 29.158347] btusb_mtk_reset+0x5c/0x480 [btusb (HASH:e8b6 5)] <4>[ 29.158361] hci_cmd_sync_work+0x10c/0x188 [bluetooth (HASH:a4fa 6)] <4>[ 29.158430] process_scheduled_works+0x258/0x4e8 <4>[ 29.158441] worker_thread+0x300/0x428 <4>[ 29.158448] kthread+0x108/0x1d0 <4>[ 29.158455] ret_from_fork+0x10/0x20 <0>[ 29.158467] Code: 91343000 940139d1 f9400268 927ff914 (f9401297) <4>[ 29.158474] ---[ end trace 0000000000000000 ]--- <0>[ 29.167129] Kernel panic - not syncing: Oops: Fatal exception <2>[ 29.167144] SMP: stopping secondary CPUs <4>[ 29.167158] ------------[ cut here ]------------ Fixes: ceac1cb0259d ("Bluetooth: btusb: mediatek: add ISO data transmission functions") Signed-off-by: Chris Lu Signed-off-by: Luiz Augusto von Dentz --- drivers/bluetooth/btusb.c | 34 +++++++++++++++++++++++++------- include/net/bluetooth/hci_core.h | 1 - 2 files changed, 27 insertions(+), 8 deletions(-) diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c index a722446ec73d..202a845e0236 100644 --- a/drivers/bluetooth/btusb.c +++ b/drivers/bluetooth/btusb.c @@ -2711,9 +2711,16 @@ static int btusb_recv_event_realtek(struct hci_dev *hdev, struct sk_buff *skb) static void btusb_mtk_claim_iso_intf(struct btusb_data *data) { - struct btmtk_data *btmtk_data = hci_get_priv(data->hdev); + struct btmtk_data *btmtk_data; int err; + if (!data->hdev) + return; + + btmtk_data = hci_get_priv(data->hdev); + if (!btmtk_data) + return; + /* * The function usb_driver_claim_interface() is documented to need * locks held if it's not called from a probe routine. The code here @@ -2735,17 +2742,30 @@ static void btusb_mtk_claim_iso_intf(struct btusb_data *data) static void btusb_mtk_release_iso_intf(struct hci_dev *hdev) { - struct btmtk_data *btmtk_data = hci_get_priv(hdev); + struct btmtk_data *btmtk_data; + + if (!hdev) + return; + + btmtk_data = hci_get_priv(hdev); + if (!btmtk_data) + return; if (test_bit(BTMTK_ISOPKT_OVER_INTR, &btmtk_data->flags)) { usb_kill_anchored_urbs(&btmtk_data->isopkt_anchor); clear_bit(BTMTK_ISOPKT_RUNNING, &btmtk_data->flags); - dev_kfree_skb_irq(btmtk_data->isopkt_skb); - btmtk_data->isopkt_skb = NULL; - usb_set_intfdata(btmtk_data->isopkt_intf, NULL); - usb_driver_release_interface(&btusb_driver, - btmtk_data->isopkt_intf); + if (btmtk_data->isopkt_skb) { + dev_kfree_skb_irq(btmtk_data->isopkt_skb); + btmtk_data->isopkt_skb = NULL; + } + + if (btmtk_data->isopkt_intf) { + usb_set_intfdata(btmtk_data->isopkt_intf, NULL); + usb_driver_release_interface(&btusb_driver, + btmtk_data->isopkt_intf); + btmtk_data->isopkt_intf = NULL; + } } clear_bit(BTMTK_ISOPKT_OVER_INTR, &btmtk_data->flags); diff --git a/include/net/bluetooth/hci_core.h b/include/net/bluetooth/hci_core.h index b8100dbfe5d7..32b1c08c8bba 100644 --- a/include/net/bluetooth/hci_core.h +++ b/include/net/bluetooth/hci_core.h @@ -749,7 +749,6 @@ struct hci_conn { __u8 remote_cap; __u8 remote_auth; - __u8 remote_id; unsigned int sent; From 275ddfeb3fdc274050c2173ffd985b1e80a9aa37 Mon Sep 17 00:00:00 2001 From: Luiz Augusto von Dentz Date: Thu, 13 Nov 2025 09:49:27 -0500 Subject: [PATCH 08/47] Bluetooth: hci_core: Fix triggering cmd_timer for HCI_OP_NOP HCI_OP_NOP means no command was actually sent so there is no point in triggering cmd_timer which may cause a hdev->reset in the process since it is assumed that the controller is stuck processing a command. Fixes: e2d471b7806b ("Bluetooth: ISO: Fix not using SID from adv report") Signed-off-by: Luiz Augusto von Dentz --- net/bluetooth/hci_core.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c index 3418d7b964a1..1920e3d62bda 100644 --- a/net/bluetooth/hci_core.c +++ b/net/bluetooth/hci_core.c @@ -4121,7 +4121,7 @@ static void hci_rx_work(struct work_struct *work) } } -static void hci_send_cmd_sync(struct hci_dev *hdev, struct sk_buff *skb) +static int hci_send_cmd_sync(struct hci_dev *hdev, struct sk_buff *skb) { int err; @@ -4133,16 +4133,19 @@ static void hci_send_cmd_sync(struct hci_dev *hdev, struct sk_buff *skb) if (!hdev->sent_cmd) { skb_queue_head(&hdev->cmd_q, skb); queue_work(hdev->workqueue, &hdev->cmd_work); - return; + return -EINVAL; } if (hci_skb_opcode(skb) != HCI_OP_NOP) { err = hci_send_frame(hdev, skb); if (err < 0) { hci_cmd_sync_cancel_sync(hdev, -err); - return; + return err; } atomic_dec(&hdev->cmd_cnt); + } else { + err = -ENODATA; + kfree_skb(skb); } if (hdev->req_status == HCI_REQ_PEND && @@ -4150,12 +4153,15 @@ static void hci_send_cmd_sync(struct hci_dev *hdev, struct sk_buff *skb) kfree_skb(hdev->req_skb); hdev->req_skb = skb_clone(hdev->sent_cmd, GFP_KERNEL); } + + return err; } static void hci_cmd_work(struct work_struct *work) { struct hci_dev *hdev = container_of(work, struct hci_dev, cmd_work); struct sk_buff *skb; + int err; BT_DBG("%s cmd_cnt %d cmd queued %d", hdev->name, atomic_read(&hdev->cmd_cnt), skb_queue_len(&hdev->cmd_q)); @@ -4166,7 +4172,9 @@ static void hci_cmd_work(struct work_struct *work) if (!skb) return; - hci_send_cmd_sync(hdev, skb); + err = hci_send_cmd_sync(hdev, skb); + if (err) + return; rcu_read_lock(); if (test_bit(HCI_RESET, &hdev->flags) || From 89bb613511cc21ed5ba6bddc1c9b9ae9c0dad392 Mon Sep 17 00:00:00 2001 From: Edward Adam Davis Date: Sun, 16 Nov 2025 17:04:43 +0800 Subject: [PATCH 09/47] Bluetooth: hci_sock: Prevent race in socket write iter and sock bind There is a potential race condition between sock bind and socket write iter. bind may free the same cmd via mgmt_pending before write iter sends the cmd, just as syzbot reported in UAF[1]. Here we use hci_dev_lock to synchronize the two, thereby avoiding the UAF mentioned in [1]. [1] syzbot reported: BUG: KASAN: slab-use-after-free in mgmt_pending_remove+0x3b/0x210 net/bluetooth/mgmt_util.c:316 Read of size 8 at addr ffff888077164818 by task syz.0.17/5989 Call Trace: mgmt_pending_remove+0x3b/0x210 net/bluetooth/mgmt_util.c:316 set_link_security+0x5c2/0x710 net/bluetooth/mgmt.c:1918 hci_mgmt_cmd+0x9c9/0xef0 net/bluetooth/hci_sock.c:1719 hci_sock_sendmsg+0x6ca/0xef0 net/bluetooth/hci_sock.c:1839 sock_sendmsg_nosec net/socket.c:727 [inline] __sock_sendmsg+0x21c/0x270 net/socket.c:742 sock_write_iter+0x279/0x360 net/socket.c:1195 Allocated by task 5989: mgmt_pending_add+0x35/0x140 net/bluetooth/mgmt_util.c:296 set_link_security+0x557/0x710 net/bluetooth/mgmt.c:1910 hci_mgmt_cmd+0x9c9/0xef0 net/bluetooth/hci_sock.c:1719 hci_sock_sendmsg+0x6ca/0xef0 net/bluetooth/hci_sock.c:1839 sock_sendmsg_nosec net/socket.c:727 [inline] __sock_sendmsg+0x21c/0x270 net/socket.c:742 sock_write_iter+0x279/0x360 net/socket.c:1195 Freed by task 5991: mgmt_pending_free net/bluetooth/mgmt_util.c:311 [inline] mgmt_pending_foreach+0x30d/0x380 net/bluetooth/mgmt_util.c:257 mgmt_index_removed+0x112/0x2f0 net/bluetooth/mgmt.c:9477 hci_sock_bind+0xbe9/0x1000 net/bluetooth/hci_sock.c:1314 Fixes: 6fe26f694c82 ("Bluetooth: MGMT: Protect mgmt_pending list with its own lock") Reported-by: syzbot+9aa47cd4633a3cf92a80@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=9aa47cd4633a3cf92a80 Tested-by: syzbot+9aa47cd4633a3cf92a80@syzkaller.appspotmail.com Signed-off-by: Edward Adam Davis Signed-off-by: Luiz Augusto von Dentz --- net/bluetooth/hci_sock.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/bluetooth/hci_sock.c b/net/bluetooth/hci_sock.c index fc866759910d..ad19022ae127 100644 --- a/net/bluetooth/hci_sock.c +++ b/net/bluetooth/hci_sock.c @@ -1311,7 +1311,9 @@ static int hci_sock_bind(struct socket *sock, struct sockaddr *addr, goto done; } + hci_dev_lock(hdev); mgmt_index_removed(hdev); + hci_dev_unlock(hdev); err = hci_dev_open(hdev->id); if (err) { From 79a2d4678ba90bdba577dc3af88cc900d6dcd5ee Mon Sep 17 00:00:00 2001 From: Pauli Virtanen Date: Sat, 15 Nov 2025 18:43:55 +0200 Subject: [PATCH 10/47] Bluetooth: hci_core: lookup hci_conn on RX path on protocol side The hdev lock/lookup/unlock/use pattern in the packet RX path doesn't ensure hci_conn* is not concurrently modified/deleted. This locking appears to be leftover from before conn_hash started using RCU commit bf4c63252490b ("Bluetooth: convert conn hash to RCU") and not clear if it had purpose since then. Currently, there are code paths that delete hci_conn* from elsewhere than the ordered hdev->workqueue where the RX work runs in. E.g. commit 5af1f84ed13a ("Bluetooth: hci_sync: Fix UAF on hci_abort_conn_sync") introduced some of these, and there probably were a few others before it. It's better to do the locking so that even if these run concurrently no UAF is possible. Move the lookup of hci_conn and associated socket-specific conn to protocol recv handlers, and do them within a single critical section to cover hci_conn* usage and lookup. syzkaller has reported a crash that appears to be this issue: [Task hdev->workqueue] [Task 2] hci_disconnect_all_sync l2cap_recv_acldata(hcon) hci_conn_get(hcon) hci_abort_conn_sync(hcon) hci_dev_lock hci_dev_lock hci_conn_del(hcon) v-------------------------------- hci_dev_unlock hci_conn_put(hcon) conn = hcon->l2cap_data (UAF) Fixes: 5af1f84ed13a ("Bluetooth: hci_sync: Fix UAF on hci_abort_conn_sync") Reported-by: syzbot+d32d77220b92eddd89ad@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=d32d77220b92eddd89ad Signed-off-by: Pauli Virtanen Signed-off-by: Luiz Augusto von Dentz --- include/net/bluetooth/hci_core.h | 20 ++++++--- net/bluetooth/hci_core.c | 73 +++++++++++--------------------- net/bluetooth/iso.c | 30 ++++++++++--- net/bluetooth/l2cap_core.c | 23 +++++++--- net/bluetooth/sco.c | 35 +++++++++++---- 5 files changed, 108 insertions(+), 73 deletions(-) diff --git a/include/net/bluetooth/hci_core.h b/include/net/bluetooth/hci_core.h index 32b1c08c8bba..0cb87687837f 100644 --- a/include/net/bluetooth/hci_core.h +++ b/include/net/bluetooth/hci_core.h @@ -856,11 +856,12 @@ extern struct mutex hci_cb_list_lock; /* ----- HCI interface to upper protocols ----- */ int l2cap_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr); int l2cap_disconn_ind(struct hci_conn *hcon); -void l2cap_recv_acldata(struct hci_conn *hcon, struct sk_buff *skb, u16 flags); +int l2cap_recv_acldata(struct hci_dev *hdev, u16 handle, struct sk_buff *skb, + u16 flags); #if IS_ENABLED(CONFIG_BT_BREDR) int sco_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, __u8 *flags); -void sco_recv_scodata(struct hci_conn *hcon, struct sk_buff *skb); +int sco_recv_scodata(struct hci_dev *hdev, u16 handle, struct sk_buff *skb); #else static inline int sco_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, __u8 *flags) @@ -868,23 +869,30 @@ static inline int sco_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, return 0; } -static inline void sco_recv_scodata(struct hci_conn *hcon, struct sk_buff *skb) +static inline int sco_recv_scodata(struct hci_dev *hdev, u16 handle, + struct sk_buff *skb) { + kfree_skb(skb); + return -ENOENT; } #endif #if IS_ENABLED(CONFIG_BT_LE) int iso_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, __u8 *flags); -void iso_recv(struct hci_conn *hcon, struct sk_buff *skb, u16 flags); +int iso_recv(struct hci_dev *hdev, u16 handle, struct sk_buff *skb, + u16 flags); #else static inline int iso_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, __u8 *flags) { return 0; } -static inline void iso_recv(struct hci_conn *hcon, struct sk_buff *skb, - u16 flags) + +static inline int iso_recv(struct hci_dev *hdev, u16 handle, + struct sk_buff *skb, u16 flags) { + kfree_skb(skb); + return -ENOENT; } #endif diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c index 1920e3d62bda..8ccec73dce45 100644 --- a/net/bluetooth/hci_core.c +++ b/net/bluetooth/hci_core.c @@ -3832,13 +3832,14 @@ static void hci_tx_work(struct work_struct *work) static void hci_acldata_packet(struct hci_dev *hdev, struct sk_buff *skb) { struct hci_acl_hdr *hdr; - struct hci_conn *conn; __u16 handle, flags; + int err; hdr = skb_pull_data(skb, sizeof(*hdr)); if (!hdr) { bt_dev_err(hdev, "ACL packet too small"); - goto drop; + kfree_skb(skb); + return; } handle = __le16_to_cpu(hdr->handle); @@ -3850,36 +3851,27 @@ static void hci_acldata_packet(struct hci_dev *hdev, struct sk_buff *skb) hdev->stat.acl_rx++; - hci_dev_lock(hdev); - conn = hci_conn_hash_lookup_handle(hdev, handle); - hci_dev_unlock(hdev); - - if (conn) { - hci_conn_enter_active_mode(conn, BT_POWER_FORCE_ACTIVE_OFF); - - /* Send to upper protocol */ - l2cap_recv_acldata(conn, skb, flags); - return; - } else { + err = l2cap_recv_acldata(hdev, handle, skb, flags); + if (err == -ENOENT) bt_dev_err(hdev, "ACL packet for unknown connection handle %d", handle); - } - -drop: - kfree_skb(skb); + else if (err) + bt_dev_dbg(hdev, "ACL packet recv for handle %d failed: %d", + handle, err); } /* SCO data packet */ static void hci_scodata_packet(struct hci_dev *hdev, struct sk_buff *skb) { struct hci_sco_hdr *hdr; - struct hci_conn *conn; __u16 handle, flags; + int err; hdr = skb_pull_data(skb, sizeof(*hdr)); if (!hdr) { bt_dev_err(hdev, "SCO packet too small"); - goto drop; + kfree_skb(skb); + return; } handle = __le16_to_cpu(hdr->handle); @@ -3891,34 +3883,28 @@ static void hci_scodata_packet(struct hci_dev *hdev, struct sk_buff *skb) hdev->stat.sco_rx++; - hci_dev_lock(hdev); - conn = hci_conn_hash_lookup_handle(hdev, handle); - hci_dev_unlock(hdev); + hci_skb_pkt_status(skb) = flags & 0x03; - if (conn) { - /* Send to upper protocol */ - hci_skb_pkt_status(skb) = flags & 0x03; - sco_recv_scodata(conn, skb); - return; - } else { + err = sco_recv_scodata(hdev, handle, skb); + if (err == -ENOENT) bt_dev_err_ratelimited(hdev, "SCO packet for unknown connection handle %d", handle); - } - -drop: - kfree_skb(skb); + else if (err) + bt_dev_dbg(hdev, "SCO packet recv for handle %d failed: %d", + handle, err); } static void hci_isodata_packet(struct hci_dev *hdev, struct sk_buff *skb) { struct hci_iso_hdr *hdr; - struct hci_conn *conn; __u16 handle, flags; + int err; hdr = skb_pull_data(skb, sizeof(*hdr)); if (!hdr) { bt_dev_err(hdev, "ISO packet too small"); - goto drop; + kfree_skb(skb); + return; } handle = __le16_to_cpu(hdr->handle); @@ -3928,22 +3914,13 @@ static void hci_isodata_packet(struct hci_dev *hdev, struct sk_buff *skb) bt_dev_dbg(hdev, "len %d handle 0x%4.4x flags 0x%4.4x", skb->len, handle, flags); - hci_dev_lock(hdev); - conn = hci_conn_hash_lookup_handle(hdev, handle); - hci_dev_unlock(hdev); - - if (!conn) { + err = iso_recv(hdev, handle, skb, flags); + if (err == -ENOENT) bt_dev_err(hdev, "ISO packet for unknown connection handle %d", handle); - goto drop; - } - - /* Send to upper protocol */ - iso_recv(conn, skb, flags); - return; - -drop: - kfree_skb(skb); + else if (err) + bt_dev_dbg(hdev, "ISO packet recv for handle %d failed: %d", + handle, err); } static bool hci_req_is_complete(struct hci_dev *hdev) diff --git a/net/bluetooth/iso.c b/net/bluetooth/iso.c index 3d98cb6291da..616c2fef91d2 100644 --- a/net/bluetooth/iso.c +++ b/net/bluetooth/iso.c @@ -2314,14 +2314,31 @@ static void iso_disconn_cfm(struct hci_conn *hcon, __u8 reason) iso_conn_del(hcon, bt_to_errno(reason)); } -void iso_recv(struct hci_conn *hcon, struct sk_buff *skb, u16 flags) +int iso_recv(struct hci_dev *hdev, u16 handle, struct sk_buff *skb, u16 flags) { - struct iso_conn *conn = hcon->iso_data; + struct hci_conn *hcon; + struct iso_conn *conn; struct skb_shared_hwtstamps *hwts; __u16 pb, ts, len, sn; - if (!conn) - goto drop; + hci_dev_lock(hdev); + + hcon = hci_conn_hash_lookup_handle(hdev, handle); + if (!hcon) { + hci_dev_unlock(hdev); + kfree_skb(skb); + return -ENOENT; + } + + conn = iso_conn_hold_unless_zero(hcon->iso_data); + hcon = NULL; + + hci_dev_unlock(hdev); + + if (!conn) { + kfree_skb(skb); + return -EINVAL; + } pb = hci_iso_flags_pb(flags); ts = hci_iso_flags_ts(flags); @@ -2377,7 +2394,7 @@ void iso_recv(struct hci_conn *hcon, struct sk_buff *skb, u16 flags) hci_skb_pkt_status(skb) = flags & 0x03; hci_skb_pkt_seqnum(skb) = sn; iso_recv_frame(conn, skb); - return; + goto done; } if (pb == ISO_SINGLE) { @@ -2455,6 +2472,9 @@ void iso_recv(struct hci_conn *hcon, struct sk_buff *skb, u16 flags) drop: kfree_skb(skb); +done: + iso_conn_put(conn); + return 0; } static struct hci_cb iso_cb = { diff --git a/net/bluetooth/l2cap_core.c b/net/bluetooth/l2cap_core.c index 35c57657bcf4..07b493331fd7 100644 --- a/net/bluetooth/l2cap_core.c +++ b/net/bluetooth/l2cap_core.c @@ -7510,13 +7510,24 @@ struct l2cap_conn *l2cap_conn_hold_unless_zero(struct l2cap_conn *c) return c; } -void l2cap_recv_acldata(struct hci_conn *hcon, struct sk_buff *skb, u16 flags) +int l2cap_recv_acldata(struct hci_dev *hdev, u16 handle, + struct sk_buff *skb, u16 flags) { + struct hci_conn *hcon; struct l2cap_conn *conn; int len; - /* Lock hdev to access l2cap_data to avoid race with l2cap_conn_del */ - hci_dev_lock(hcon->hdev); + /* Lock hdev for hci_conn, and race on l2cap_data vs. l2cap_conn_del */ + hci_dev_lock(hdev); + + hcon = hci_conn_hash_lookup_handle(hdev, handle); + if (!hcon) { + hci_dev_unlock(hdev); + kfree_skb(skb); + return -ENOENT; + } + + hci_conn_enter_active_mode(hcon, BT_POWER_FORCE_ACTIVE_OFF); conn = hcon->l2cap_data; @@ -7524,12 +7535,13 @@ void l2cap_recv_acldata(struct hci_conn *hcon, struct sk_buff *skb, u16 flags) conn = l2cap_conn_add(hcon); conn = l2cap_conn_hold_unless_zero(conn); + hcon = NULL; - hci_dev_unlock(hcon->hdev); + hci_dev_unlock(hdev); if (!conn) { kfree_skb(skb); - return; + return -EINVAL; } BT_DBG("conn %p len %u flags 0x%x", conn, skb->len, flags); @@ -7643,6 +7655,7 @@ drop: unlock: mutex_unlock(&conn->lock); l2cap_conn_put(conn); + return 0; } static struct hci_cb l2cap_cb = { diff --git a/net/bluetooth/sco.c b/net/bluetooth/sco.c index ab0cf442d57b..298c2a9ab4df 100644 --- a/net/bluetooth/sco.c +++ b/net/bluetooth/sco.c @@ -1458,22 +1458,39 @@ static void sco_disconn_cfm(struct hci_conn *hcon, __u8 reason) sco_conn_del(hcon, bt_to_errno(reason)); } -void sco_recv_scodata(struct hci_conn *hcon, struct sk_buff *skb) +int sco_recv_scodata(struct hci_dev *hdev, u16 handle, struct sk_buff *skb) { - struct sco_conn *conn = hcon->sco_data; + struct hci_conn *hcon; + struct sco_conn *conn; - if (!conn) - goto drop; + hci_dev_lock(hdev); + + hcon = hci_conn_hash_lookup_handle(hdev, handle); + if (!hcon) { + hci_dev_unlock(hdev); + kfree_skb(skb); + return -ENOENT; + } + + conn = sco_conn_hold_unless_zero(hcon->sco_data); + hcon = NULL; + + hci_dev_unlock(hdev); + + if (!conn) { + kfree_skb(skb); + return -EINVAL; + } BT_DBG("conn %p len %u", conn, skb->len); - if (skb->len) { + if (skb->len) sco_recv_frame(conn, skb); - return; - } + else + kfree_skb(skb); -drop: - kfree_skb(skb); + sco_conn_put(conn); + return 0; } static struct hci_cb sco_cb = { From c884a0b27b4586e607431d86a1aa0bb4fb39169c Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Thu, 20 Nov 2025 08:12:28 -0800 Subject: [PATCH 11/47] Bluetooth: btusb: mediatek: Avoid btusb_mtk_claim_iso_intf() NULL deref In btusb_mtk_setup(), we set `btmtk_data->isopkt_intf` to: usb_ifnum_to_if(data->udev, MTK_ISO_IFNUM) That function can return NULL in some cases. Even when it returns NULL, though, we still go on to call btusb_mtk_claim_iso_intf(). As of commit e9087e828827 ("Bluetooth: btusb: mediatek: Add locks for usb_driver_claim_interface()"), calling btusb_mtk_claim_iso_intf() when `btmtk_data->isopkt_intf` is NULL will cause a crash because we'll end up passing a bad pointer to device_lock(). Prior to that commit we'd pass the NULL pointer directly to usb_driver_claim_interface() which would detect it and return an error, which was handled. Resolve the crash in btusb_mtk_claim_iso_intf() by adding a NULL check at the start of the function. This makes the code handle a NULL `btmtk_data->isopkt_intf` the same way it did before the problematic commit (just with a slight change to the error message printed). Reported-by: IncogCyberpunk Closes: http://lore.kernel.org/r/a380d061-479e-4713-bddd-1d6571ca7e86@leemhuis.info Fixes: e9087e828827 ("Bluetooth: btusb: mediatek: Add locks for usb_driver_claim_interface()") Cc: stable@vger.kernel.org Tested-by: IncogCyberpunk Signed-off-by: Douglas Anderson Signed-off-by: Luiz Augusto von Dentz --- drivers/bluetooth/btusb.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c index 202a845e0236..fa683bb7f0b4 100644 --- a/drivers/bluetooth/btusb.c +++ b/drivers/bluetooth/btusb.c @@ -2721,6 +2721,11 @@ static void btusb_mtk_claim_iso_intf(struct btusb_data *data) if (!btmtk_data) return; + if (!btmtk_data->isopkt_intf) { + bt_dev_err(data->hdev, "Can't claim NULL iso interface"); + return; + } + /* * The function usb_driver_claim_interface() is documented to need * locks held if it's not called from a probe routine. The code here From 545d7827b2cd5de5eb85580cebeda6b35b3ff443 Mon Sep 17 00:00:00 2001 From: Luiz Augusto von Dentz Date: Mon, 17 Nov 2025 13:45:13 -0500 Subject: [PATCH 12/47] Bluetooth: SMP: Fix not generating mackey and ltk when repairing The change eed467b517e8 ("Bluetooth: fix passkey uninitialized when used") introduced a goto that bypasses the creation of temporary mackey and ltk which are later used by the likes of DHKey Check step. Later ffee202a78c2 ("Bluetooth: Always request for user confirmation for Just Works (LE SC)") which means confirm_hint is always set in case JUST_WORKS so the branch checking for an existing LTK becomes pointless as confirm_hint will always be set, so this just merge both cases of malicious or legitimate devices to be confirmed before continuing with the pairing procedure. Link: https://github.com/bluez/bluez/issues/1622 Fixes: eed467b517e8 ("Bluetooth: fix passkey uninitialized when used") Signed-off-by: Luiz Augusto von Dentz --- net/bluetooth/smp.c | 31 +++++++------------------------ 1 file changed, 7 insertions(+), 24 deletions(-) diff --git a/net/bluetooth/smp.c b/net/bluetooth/smp.c index 45512b2ba951..3a1ce04a7a53 100644 --- a/net/bluetooth/smp.c +++ b/net/bluetooth/smp.c @@ -2136,7 +2136,7 @@ static u8 smp_cmd_pairing_random(struct l2cap_conn *conn, struct sk_buff *skb) struct smp_chan *smp = chan->data; struct hci_conn *hcon = conn->hcon; u8 *pkax, *pkbx, *na, *nb, confirm_hint; - u32 passkey; + u32 passkey = 0; int err; bt_dev_dbg(hcon->hdev, "conn %p", conn); @@ -2188,24 +2188,6 @@ static u8 smp_cmd_pairing_random(struct l2cap_conn *conn, struct sk_buff *skb) smp_send_cmd(conn, SMP_CMD_PAIRING_RANDOM, sizeof(smp->prnd), smp->prnd); SMP_ALLOW_CMD(smp, SMP_CMD_DHKEY_CHECK); - - /* Only Just-Works pairing requires extra checks */ - if (smp->method != JUST_WORKS) - goto mackey_and_ltk; - - /* If there already exists long term key in local host, leave - * the decision to user space since the remote device could - * be legitimate or malicious. - */ - if (hci_find_ltk(hcon->hdev, &hcon->dst, hcon->dst_type, - hcon->role)) { - /* Set passkey to 0. The value can be any number since - * it'll be ignored anyway. - */ - passkey = 0; - confirm_hint = 1; - goto confirm; - } } mackey_and_ltk: @@ -2226,11 +2208,12 @@ mackey_and_ltk: if (err) return SMP_UNSPECIFIED; - confirm_hint = 0; - -confirm: - if (smp->method == JUST_WORKS) - confirm_hint = 1; + /* Always require user confirmation for Just-Works pairing to prevent + * impersonation attacks, or in case of a legitimate device that is + * repairing use the confirmation as acknowledgment to proceed with the + * creation of new keys. + */ + confirm_hint = smp->method == JUST_WORKS ? 1 : 0; err = mgmt_user_confirm_request(hcon->hdev, &hcon->dst, hcon->type, hcon->dst_type, passkey, confirm_hint); From 4b4749b7b4b3badabd0fa26f454782842db6ae53 Mon Sep 17 00:00:00 2001 From: Sayooj K Karun Date: Wed, 19 Nov 2025 14:27:47 +0530 Subject: [PATCH 13/47] net: atm: fix incorrect cleanup function call in error path In atm_init(), if atmsvc_init() fails, the code jumps to out_atmpvc_exit label which incorrectly calls atmsvc_exit() instead of atmpvc_exit(). This results in calling the wrong cleanup function and failing to properly clean up atmpvc_init(). Fix this by calling atmpvc_exit() in the out_atmpvc_exit error path. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Sayooj K Karun Link: https://patch.msgid.link/20251119085747.67139-1-sayooj@aerlync.com Signed-off-by: Jakub Kicinski --- net/atm/common.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/atm/common.c b/net/atm/common.c index 881c7f259dbd..c4edc1111bf0 100644 --- a/net/atm/common.c +++ b/net/atm/common.c @@ -881,7 +881,7 @@ out_atmproc_exit: out_atmsvc_exit: atmsvc_exit(); out_atmpvc_exit: - atmsvc_exit(); + atmpvc_exit(); out_unregister_vcc_proto: proto_unregister(&vcc_proto); goto out; From a14602fcae17a3f1cb8a8521bedf31728f9e7e39 Mon Sep 17 00:00:00 2001 From: Jesper Dangaard Brouer Date: Wed, 19 Nov 2025 17:28:36 +0100 Subject: [PATCH 14/47] veth: reduce XDP no_direct return section to fix race As explain in commit fa349e396e48 ("veth: Fix race with AF_XDP exposing old or uninitialized descriptors") for veth there is a chance after napi_complete_done() that another CPU can manage start another NAPI instance running veth_pool(). For NAPI this is correctly handled as the napi_schedule_prep() check will prevent multiple instances from getting scheduled, but for the remaining code in veth_pool() this can run concurrent with the newly started NAPI instance. The problem/race is that xdp_clear_return_frame_no_direct() isn't designed to be nested. Prior to commit 401cb7dae813 ("net: Reference bpf_redirect_info via task_struct on PREEMPT_RT.") the temporary BPF net context bpf_redirect_info was stored per CPU, where this wasn't an issue. Since this commit the BPF context is stored in 'current' task_struct. When running veth in threaded-NAPI mode, then the kthread becomes the storage area. Now a race exists between two concurrent veth_pool() function calls one exiting NAPI and one running new NAPI, both using the same BPF net context. Race is when another CPU gets within the xdp_set_return_frame_no_direct() section before exiting veth_pool() calls the clear-function xdp_clear_return_frame_no_direct(). Fixes: 401cb7dae8130 ("net: Reference bpf_redirect_info via task_struct on PREEMPT_RT.") Signed-off-by: Jesper Dangaard Brouer Link: https://patch.msgid.link/176356963888.337072.4805242001928705046.stgit@firesoul Signed-off-by: Jakub Kicinski --- drivers/net/veth.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/drivers/net/veth.c b/drivers/net/veth.c index 35dd89aff4a9..cc502bf022d5 100644 --- a/drivers/net/veth.c +++ b/drivers/net/veth.c @@ -975,6 +975,9 @@ static int veth_poll(struct napi_struct *napi, int budget) if (stats.xdp_redirect > 0) xdp_do_flush(); + if (stats.xdp_tx > 0) + veth_xdp_flush(rq, &bq); + xdp_clear_return_frame_no_direct(); if (done < budget && napi_complete_done(napi, done)) { /* Write rx_notify_masked before reading ptr_ring */ @@ -987,10 +990,6 @@ static int veth_poll(struct napi_struct *napi, int budget) } } - if (stats.xdp_tx > 0) - veth_xdp_flush(rq, &bq); - xdp_clear_return_frame_no_direct(); - /* Release backpressure per NAPI poll */ smp_rmb(); /* Paired with netif_tx_stop_queue set_bit */ if (peer_txq && netif_tx_queue_stopped(peer_txq)) { From ec3803b5917b6ff2f86ea965d0985c95d8a85119 Mon Sep 17 00:00:00 2001 From: Daniel Golle Date: Thu, 20 Nov 2025 14:17:13 +0000 Subject: [PATCH 15/47] net: phy: mxl-gpy: fix bogus error on USXGMII and integrated PHY As the interface mode doesn't need to be updated on PHYs connected with USXGMII and integrated PHYs, gpy_update_interface() should just return 0 in these cases rather than -EINVAL which has wrongly been introduced by commit 7a495dde27ebc ("net: phy: mxl-gpy: Change gpy_update_interface() function return type"), as this breaks support for those PHYs. Fixes: 7a495dde27ebc ("net: phy: mxl-gpy: Change gpy_update_interface() function return type") Signed-off-by: Daniel Golle Reviewed-by: Maxime Chevallier Reviewed-by: Russell King (Oracle) Link: https://patch.msgid.link/f744f721a1fcc5e2e936428c62ff2c7d94d2a293.1763648168.git.daniel@makrotopia.org Signed-off-by: Jakub Kicinski --- drivers/net/phy/mxl-gpy.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/phy/mxl-gpy.c b/drivers/net/phy/mxl-gpy.c index 0c8dc16ee7bd..221b315203d0 100644 --- a/drivers/net/phy/mxl-gpy.c +++ b/drivers/net/phy/mxl-gpy.c @@ -540,7 +540,7 @@ static int gpy_update_interface(struct phy_device *phydev) /* Interface mode is fixed for USXGMII and integrated PHY */ if (phydev->interface == PHY_INTERFACE_MODE_USXGMII || phydev->interface == PHY_INTERFACE_MODE_INTERNAL) - return -EINVAL; + return 0; /* Automatically switch SERDES interface between SGMII and 2500-BaseX * according to speed. Disable ANEG in 2500-BaseX mode. From 4fe5a00ec70717a7f1002d8913ec6143582b3c8e Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 21 Nov 2025 15:41:00 +0000 Subject: [PATCH 16/47] net: sched: fix TCF_LAYER_TRANSPORT handling in tcf_get_base_ptr() syzbot reported that tcf_get_base_ptr() can be called while transport header is not set [1]. Instead of returning a dangling pointer, return NULL. Fix tcf_get_base_ptr() callers to handle this NULL value. [1] WARNING: CPU: 1 PID: 6019 at ./include/linux/skbuff.h:3071 skb_transport_header include/linux/skbuff.h:3071 [inline] WARNING: CPU: 1 PID: 6019 at ./include/linux/skbuff.h:3071 tcf_get_base_ptr include/net/pkt_cls.h:539 [inline] WARNING: CPU: 1 PID: 6019 at ./include/linux/skbuff.h:3071 em_nbyte_match+0x2d8/0x3f0 net/sched/em_nbyte.c:43 Modules linked in: CPU: 1 UID: 0 PID: 6019 Comm: syz.0.17 Not tainted syzkaller #0 PREEMPT(full) Call Trace: tcf_em_match net/sched/ematch.c:494 [inline] __tcf_em_tree_match+0x1ac/0x770 net/sched/ematch.c:520 tcf_em_tree_match include/net/pkt_cls.h:512 [inline] basic_classify+0x115/0x2d0 net/sched/cls_basic.c:50 tc_classify include/net/tc_wrapper.h:197 [inline] __tcf_classify net/sched/cls_api.c:1764 [inline] tcf_classify+0x4cf/0x1140 net/sched/cls_api.c:1860 multiq_classify net/sched/sch_multiq.c:39 [inline] multiq_enqueue+0xfd/0x4c0 net/sched/sch_multiq.c:66 dev_qdisc_enqueue+0x4e/0x260 net/core/dev.c:4118 __dev_xmit_skb net/core/dev.c:4214 [inline] __dev_queue_xmit+0xe83/0x3b50 net/core/dev.c:4729 packet_snd net/packet/af_packet.c:3076 [inline] packet_sendmsg+0x3e33/0x5080 net/packet/af_packet.c:3108 sock_sendmsg_nosec net/socket.c:727 [inline] __sock_sendmsg+0x21c/0x270 net/socket.c:742 ____sys_sendmsg+0x505/0x830 net/socket.c:2630 Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Reported-by: syzbot+f3a497f02c389d86ef16@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/6920855a.a70a0220.2ea503.0058.GAE@google.com/T/#u Signed-off-by: Eric Dumazet Reviewed-by: Jamal Hadi Salim Link: https://patch.msgid.link/20251121154100.1616228-1-edumazet@google.com Signed-off-by: Jakub Kicinski --- include/net/pkt_cls.h | 2 ++ net/sched/em_cmp.c | 5 ++++- net/sched/em_nbyte.c | 2 ++ net/sched/em_text.c | 11 +++++++++-- 4 files changed, 17 insertions(+), 3 deletions(-) diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h index c64fd896b1f9..99ac747b7906 100644 --- a/include/net/pkt_cls.h +++ b/include/net/pkt_cls.h @@ -536,6 +536,8 @@ static inline unsigned char * tcf_get_base_ptr(struct sk_buff *skb, int layer) case TCF_LAYER_NETWORK: return skb_network_header(skb); case TCF_LAYER_TRANSPORT: + if (!skb_transport_header_was_set(skb)) + break; return skb_transport_header(skb); } diff --git a/net/sched/em_cmp.c b/net/sched/em_cmp.c index 64b637f18bc7..48c1bce74f49 100644 --- a/net/sched/em_cmp.c +++ b/net/sched/em_cmp.c @@ -22,9 +22,12 @@ static int em_cmp_match(struct sk_buff *skb, struct tcf_ematch *em, struct tcf_pkt_info *info) { struct tcf_em_cmp *cmp = (struct tcf_em_cmp *) em->data; - unsigned char *ptr = tcf_get_base_ptr(skb, cmp->layer) + cmp->off; + unsigned char *ptr = tcf_get_base_ptr(skb, cmp->layer); u32 val = 0; + if (!ptr) + return 0; + ptr += cmp->off; if (!tcf_valid_offset(skb, ptr, cmp->align)) return 0; diff --git a/net/sched/em_nbyte.c b/net/sched/em_nbyte.c index 4f9f21a05d5e..c65ffa5fff94 100644 --- a/net/sched/em_nbyte.c +++ b/net/sched/em_nbyte.c @@ -42,6 +42,8 @@ static int em_nbyte_match(struct sk_buff *skb, struct tcf_ematch *em, struct nbyte_data *nbyte = (struct nbyte_data *) em->data; unsigned char *ptr = tcf_get_base_ptr(skb, nbyte->hdr.layer); + if (!ptr) + return 0; ptr += nbyte->hdr.off; if (!tcf_valid_offset(skb, ptr, nbyte->hdr.len)) diff --git a/net/sched/em_text.c b/net/sched/em_text.c index 6b3d0af72c39..692e2be1793e 100644 --- a/net/sched/em_text.c +++ b/net/sched/em_text.c @@ -29,12 +29,19 @@ static int em_text_match(struct sk_buff *skb, struct tcf_ematch *m, struct tcf_pkt_info *info) { struct text_match *tm = EM_TEXT_PRIV(m); + unsigned char *ptr; int from, to; - from = tcf_get_base_ptr(skb, tm->from_layer) - skb->data; + ptr = tcf_get_base_ptr(skb, tm->from_layer); + if (!ptr) + return 0; + from = ptr - skb->data; from += tm->from_offset; - to = tcf_get_base_ptr(skb, tm->to_layer) - skb->data; + ptr = tcf_get_base_ptr(skb, tm->to_layer); + if (!ptr) + return 0; + to = ptr - skb->data; to += tm->to_offset; return skb_find_text(skb, from, to, tm->config) != UINT_MAX; From 7526183cfdbe352c51c285762f0e15b7c428ea06 Mon Sep 17 00:00:00 2001 From: Kai-Heng Feng Date: Thu, 20 Nov 2025 12:15:33 +0800 Subject: [PATCH 17/47] net: aquantia: Add missing descriptor cache invalidation on ATL2 ATL2 hardware was missing descriptor cache invalidation in hw_stop(), causing SMMU translation faults during device shutdown and module removal: [ 70.355743] arm-smmu-v3 arm-smmu-v3.5.auto: event 0x10 received: [ 70.361893] arm-smmu-v3 arm-smmu-v3.5.auto: 0x0002060000000010 [ 70.367948] arm-smmu-v3 arm-smmu-v3.5.auto: 0x0000020000000000 [ 70.374002] arm-smmu-v3 arm-smmu-v3.5.auto: 0x00000000ff9bc000 [ 70.380055] arm-smmu-v3 arm-smmu-v3.5.auto: 0x0000000000000000 [ 70.386109] arm-smmu-v3 arm-smmu-v3.5.auto: event: F_TRANSLATION client: 0001:06:00.0 sid: 0x20600 ssid: 0x0 iova: 0xff9bc000 ipa: 0x0 [ 70.398531] arm-smmu-v3 arm-smmu-v3.5.auto: unpriv data write s1 "Input address caused fault" stag: 0x0 Commit 7a1bb49461b1 ("net: aquantia: fix potential IOMMU fault after driver unbind") and commit ed4d81c4b3f2 ("net: aquantia: when cleaning hw cache it should be toggled") fixed cache invalidation for ATL B0, but ATL2 was left with only interrupt disabling. This allowed hardware to write to cached descriptors after DMA memory was unmapped, triggering SMMU faults. Once cache invalidation is applied to ATL2, the translation fault can't be observed anymore. Add shared aq_hw_invalidate_descriptor_cache() helper and use it in both ATL B0 and ATL2 hw_stop() implementations for consistent behavior. Fixes: e54dcf4bba3e ("net: atlantic: basic A2 init/deinit hw_ops") Tested-by: Carol Soto Signed-off-by: Kai-Heng Feng Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251120041537.62184-1-kaihengf@nvidia.com Signed-off-by: Paolo Abeni --- .../ethernet/aquantia/atlantic/aq_hw_utils.c | 22 +++++++++++++++++++ .../ethernet/aquantia/atlantic/aq_hw_utils.h | 1 + .../aquantia/atlantic/hw_atl/hw_atl_b0.c | 19 +--------------- .../aquantia/atlantic/hw_atl2/hw_atl2.c | 2 +- 4 files changed, 25 insertions(+), 19 deletions(-) diff --git a/drivers/net/ethernet/aquantia/atlantic/aq_hw_utils.c b/drivers/net/ethernet/aquantia/atlantic/aq_hw_utils.c index 1921741f7311..18b08277d2e1 100644 --- a/drivers/net/ethernet/aquantia/atlantic/aq_hw_utils.c +++ b/drivers/net/ethernet/aquantia/atlantic/aq_hw_utils.c @@ -15,6 +15,7 @@ #include "aq_hw.h" #include "aq_nic.h" +#include "hw_atl/hw_atl_llh.h" void aq_hw_write_reg_bit(struct aq_hw_s *aq_hw, u32 addr, u32 msk, u32 shift, u32 val) @@ -81,6 +82,27 @@ void aq_hw_write_reg64(struct aq_hw_s *hw, u32 reg, u64 value) lo_hi_writeq(value, hw->mmio + reg); } +int aq_hw_invalidate_descriptor_cache(struct aq_hw_s *hw) +{ + int err; + u32 val; + + /* Invalidate Descriptor Cache to prevent writing to the cached + * descriptors and to the data pointer of those descriptors + */ + hw_atl_rdm_rx_dma_desc_cache_init_tgl(hw); + + err = aq_hw_err_from_flags(hw); + if (err) + goto err_exit; + + readx_poll_timeout_atomic(hw_atl_rdm_rx_dma_desc_cache_init_done_get, + hw, val, val == 1, 1000U, 10000U); + +err_exit: + return err; +} + int aq_hw_err_from_flags(struct aq_hw_s *hw) { int err = 0; diff --git a/drivers/net/ethernet/aquantia/atlantic/aq_hw_utils.h b/drivers/net/ethernet/aquantia/atlantic/aq_hw_utils.h index ffa6e4067c21..d89c63d88e4a 100644 --- a/drivers/net/ethernet/aquantia/atlantic/aq_hw_utils.h +++ b/drivers/net/ethernet/aquantia/atlantic/aq_hw_utils.h @@ -35,6 +35,7 @@ u32 aq_hw_read_reg(struct aq_hw_s *hw, u32 reg); void aq_hw_write_reg(struct aq_hw_s *hw, u32 reg, u32 value); u64 aq_hw_read_reg64(struct aq_hw_s *hw, u32 reg); void aq_hw_write_reg64(struct aq_hw_s *hw, u32 reg, u64 value); +int aq_hw_invalidate_descriptor_cache(struct aq_hw_s *hw); int aq_hw_err_from_flags(struct aq_hw_s *hw); int aq_hw_num_tcs(struct aq_hw_s *hw); int aq_hw_q_per_tc(struct aq_hw_s *hw); diff --git a/drivers/net/ethernet/aquantia/atlantic/hw_atl/hw_atl_b0.c b/drivers/net/ethernet/aquantia/atlantic/hw_atl/hw_atl_b0.c index 493432d036b9..c7895bfb2ecf 100644 --- a/drivers/net/ethernet/aquantia/atlantic/hw_atl/hw_atl_b0.c +++ b/drivers/net/ethernet/aquantia/atlantic/hw_atl/hw_atl_b0.c @@ -1198,26 +1198,9 @@ static int hw_atl_b0_hw_interrupt_moderation_set(struct aq_hw_s *self) static int hw_atl_b0_hw_stop(struct aq_hw_s *self) { - int err; - u32 val; - hw_atl_b0_hw_irq_disable(self, HW_ATL_B0_INT_MASK); - /* Invalidate Descriptor Cache to prevent writing to the cached - * descriptors and to the data pointer of those descriptors - */ - hw_atl_rdm_rx_dma_desc_cache_init_tgl(self); - - err = aq_hw_err_from_flags(self); - - if (err) - goto err_exit; - - readx_poll_timeout_atomic(hw_atl_rdm_rx_dma_desc_cache_init_done_get, - self, val, val == 1, 1000U, 10000U); - -err_exit: - return err; + return aq_hw_invalidate_descriptor_cache(self); } int hw_atl_b0_hw_ring_tx_stop(struct aq_hw_s *self, struct aq_ring_s *ring) diff --git a/drivers/net/ethernet/aquantia/atlantic/hw_atl2/hw_atl2.c b/drivers/net/ethernet/aquantia/atlantic/hw_atl2/hw_atl2.c index b0ed572e88c6..0ce9caae8799 100644 --- a/drivers/net/ethernet/aquantia/atlantic/hw_atl2/hw_atl2.c +++ b/drivers/net/ethernet/aquantia/atlantic/hw_atl2/hw_atl2.c @@ -759,7 +759,7 @@ static int hw_atl2_hw_stop(struct aq_hw_s *self) { hw_atl_b0_hw_irq_disable(self, HW_ATL2_INT_MASK); - return 0; + return aq_hw_invalidate_descriptor_cache(self); } static struct aq_stats_s *hw_atl2_utils_get_hw_stats(struct aq_hw_s *self) From 7b3c09e1667977edee11de94a85e2593a7c15e87 Mon Sep 17 00:00:00 2001 From: "Bastien Curutchet (Schneider Electric)" Date: Thu, 20 Nov 2025 10:12:00 +0100 Subject: [PATCH 18/47] net: dsa: microchip: common: Fix checks on irq_find_mapping() irq_find_mapping() returns a positive IRQ number or 0 if no IRQ is found but it never returns a negative value. However, on each irq_find_mapping() call, we verify that the returned value isn't negative. Fix the irq_find_mapping() checks to enter error paths when 0 is returned. Return -EINVAL in such cases. CC: stable@vger.kernel.org Fixes: c9cd961c0d43 ("net: dsa: microchip: lan937x: add interrupt support for port phy link") Reviewed-by: Andrew Lunn Signed-off-by: Bastien Curutchet (Schneider Electric) Link: https://patch.msgid.link/20251120-ksz-fix-v6-1-891f80ae7f8f@bootlin.com Signed-off-by: Paolo Abeni --- drivers/net/dsa/microchip/ksz_common.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/net/dsa/microchip/ksz_common.c b/drivers/net/dsa/microchip/ksz_common.c index 933ae8dc6337..e684568a4eda 100644 --- a/drivers/net/dsa/microchip/ksz_common.c +++ b/drivers/net/dsa/microchip/ksz_common.c @@ -2587,8 +2587,8 @@ static int ksz_irq_phy_setup(struct ksz_device *dev) irq = irq_find_mapping(dev->ports[port].pirq.domain, PORT_SRC_PHY_INT); - if (irq < 0) { - ret = irq; + if (!irq) { + ret = -EINVAL; goto out; } ds->user_mii_bus->irq[phy] = irq; @@ -2952,8 +2952,8 @@ static int ksz_pirq_setup(struct ksz_device *dev, u8 p) snprintf(pirq->name, sizeof(pirq->name), "port_irq-%d", p); pirq->irq_num = irq_find_mapping(dev->girq.domain, p); - if (pirq->irq_num < 0) - return pirq->irq_num; + if (!pirq->irq_num) + return -EINVAL; return ksz_irq_common_setup(dev, pirq); } From 9e059305be41a5bd27e03458d8333cf30d70be34 Mon Sep 17 00:00:00 2001 From: "Bastien Curutchet (Schneider Electric)" Date: Thu, 20 Nov 2025 10:12:01 +0100 Subject: [PATCH 19/47] net: dsa: microchip: ptp: Fix checks on irq_find_mapping() irq_find_mapping() returns a positive IRQ number or 0 if no IRQ is found but it never returns a negative value. However, during the PTP IRQ setup, we verify that its returned value isn't negative. Fix the irq_find_mapping() check to enter the error path when 0 is returned. Return -EINVAL in such case. Cc: stable@vger.kernel.org Fixes: cc13ab18b201 ("net: dsa: microchip: ptp: enable interrupt for timestamping") Reviewed-by: Andrew Lunn Signed-off-by: Bastien Curutchet (Schneider Electric) Link: https://patch.msgid.link/20251120-ksz-fix-v6-2-891f80ae7f8f@bootlin.com Signed-off-by: Paolo Abeni --- drivers/net/dsa/microchip/ksz_ptp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/dsa/microchip/ksz_ptp.c b/drivers/net/dsa/microchip/ksz_ptp.c index 35fc21b1ee48..c8bfbe5e2157 100644 --- a/drivers/net/dsa/microchip/ksz_ptp.c +++ b/drivers/net/dsa/microchip/ksz_ptp.c @@ -1139,8 +1139,8 @@ int ksz_ptp_irq_setup(struct dsa_switch *ds, u8 p) irq_create_mapping(ptpirq->domain, irq); ptpirq->irq_num = irq_find_mapping(port->pirq.domain, PORT_SRC_PTP_INT); - if (ptpirq->irq_num < 0) { - ret = ptpirq->irq_num; + if (!ptpirq->irq_num) { + ret = -EINVAL; goto out; } From 25b62cc5b22c45face094ae3e8717258e46d1d19 Mon Sep 17 00:00:00 2001 From: "Bastien Curutchet (Schneider Electric)" Date: Thu, 20 Nov 2025 10:12:02 +0100 Subject: [PATCH 20/47] net: dsa: microchip: Don't free uninitialized ksz_irq If something goes wrong at setup, ksz_irq_free() can be called on uninitialized ksz_irq (for example when ksz_ptp_irq_setup() fails). It leads to freeing uninitialized IRQ numbers and/or domains. Use dsa_switch_for_each_user_port_continue_reverse() in the error path to iterate only over the fully initialized ports. Cc: stable@vger.kernel.org Fixes: cc13ab18b201 ("net: dsa: microchip: ptp: enable interrupt for timestamping") Signed-off-by: Bastien Curutchet (Schneider Electric) Link: https://patch.msgid.link/20251120-ksz-fix-v6-3-891f80ae7f8f@bootlin.com Signed-off-by: Paolo Abeni --- drivers/net/dsa/microchip/ksz_common.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/dsa/microchip/ksz_common.c b/drivers/net/dsa/microchip/ksz_common.c index e684568a4eda..a927055423f3 100644 --- a/drivers/net/dsa/microchip/ksz_common.c +++ b/drivers/net/dsa/microchip/ksz_common.c @@ -3082,7 +3082,7 @@ out_ptpirq: ksz_ptp_irq_free(ds, dp->index); out_pirq: if (dev->irq > 0) - dsa_switch_for_each_user_port(dp, dev->ds) + dsa_switch_for_each_user_port_continue_reverse(dp, dev->ds) ksz_irq_free(&dev->ports[dp->index].pirq); out_girq: if (dev->irq > 0) From 0f80e21bf6229637e193248fbd284c0ec44bc0fd Mon Sep 17 00:00:00 2001 From: "Bastien Curutchet (Schneider Electric)" Date: Thu, 20 Nov 2025 10:12:03 +0100 Subject: [PATCH 21/47] net: dsa: microchip: Free previously initialized ports on init failures If a port interrupt setup fails after at least one port has already been successfully initialized, the gotos miss some resource releasing: - the already initialized PTP IRQs aren't released - the already initialized port IRQs aren't released if the failure occurs in ksz_pirq_setup(). Merge 'out_girq' and 'out_ptpirq' into a single 'port_release' label. Behind this label, use the reverse loop to release all IRQ resources for all initialized ports. Jump in the middle of the reverse loop if an error occurs in ksz_ptp_irq_setup() to only release the port IRQ of the current iteration. Cc: stable@vger.kernel.org Fixes: c9cd961c0d43 ("net: dsa: microchip: lan937x: add interrupt support for port phy link") Signed-off-by: Bastien Curutchet (Schneider Electric) Link: https://patch.msgid.link/20251120-ksz-fix-v6-4-891f80ae7f8f@bootlin.com Signed-off-by: Paolo Abeni --- drivers/net/dsa/microchip/ksz_common.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) diff --git a/drivers/net/dsa/microchip/ksz_common.c b/drivers/net/dsa/microchip/ksz_common.c index a927055423f3..0c10351fe5eb 100644 --- a/drivers/net/dsa/microchip/ksz_common.c +++ b/drivers/net/dsa/microchip/ksz_common.c @@ -3038,12 +3038,12 @@ static int ksz_setup(struct dsa_switch *ds) dsa_switch_for_each_user_port(dp, dev->ds) { ret = ksz_pirq_setup(dev, dp->index); if (ret) - goto out_girq; + goto port_release; if (dev->info->ptp_capable) { ret = ksz_ptp_irq_setup(ds, dp->index); if (ret) - goto out_pirq; + goto pirq_release; } } } @@ -3053,7 +3053,7 @@ static int ksz_setup(struct dsa_switch *ds) if (ret) { dev_err(dev->dev, "Failed to register PTP clock: %d\n", ret); - goto out_ptpirq; + goto port_release; } } @@ -3076,17 +3076,16 @@ static int ksz_setup(struct dsa_switch *ds) out_ptp_clock_unregister: if (dev->info->ptp_capable) ksz_ptp_clock_unregister(ds); -out_ptpirq: - if (dev->irq > 0 && dev->info->ptp_capable) - dsa_switch_for_each_user_port(dp, dev->ds) - ksz_ptp_irq_free(ds, dp->index); -out_pirq: - if (dev->irq > 0) - dsa_switch_for_each_user_port_continue_reverse(dp, dev->ds) +port_release: + if (dev->irq > 0) { + dsa_switch_for_each_user_port_continue_reverse(dp, dev->ds) { + if (dev->info->ptp_capable) + ksz_ptp_irq_free(ds, dp->index); +pirq_release: ksz_irq_free(&dev->ports[dp->index].pirq); -out_girq: - if (dev->irq > 0) + } ksz_irq_free(&dev->girq); + } return ret; } From d0b8fec8ae50525b57139393d0bb1f446e82ff7e Mon Sep 17 00:00:00 2001 From: "Bastien Curutchet (Schneider Electric)" Date: Thu, 20 Nov 2025 10:12:04 +0100 Subject: [PATCH 22/47] net: dsa: microchip: Fix symetry in ksz_ptp_msg_irq_{setup/free}() The IRQ numbers created through irq_create_mapping() are only assigned to ptpmsg_irq[n].num at the end of the IRQ setup. So if an error occurs between their creation and their assignment (for instance during the request_threaded_irq() step), we enter the error path and fail to release the newly created virtual IRQs because they aren't yet assigned to ptpmsg_irq[n].num. Move the mapping creation to ksz_ptp_msg_irq_setup() to ensure symetry with what's released by ksz_ptp_msg_irq_free(). In the error path, move the irq_dispose_mapping to the out_ptp_msg label so it will be called only on created IRQs. Cc: stable@vger.kernel.org Fixes: cc13ab18b201 ("net: dsa: microchip: ptp: enable interrupt for timestamping") Reviewed-by: Andrew Lunn Signed-off-by: Bastien Curutchet (Schneider Electric) Link: https://patch.msgid.link/20251120-ksz-fix-v6-5-891f80ae7f8f@bootlin.com Signed-off-by: Paolo Abeni --- drivers/net/dsa/microchip/ksz_ptp.c | 18 +++++++----------- 1 file changed, 7 insertions(+), 11 deletions(-) diff --git a/drivers/net/dsa/microchip/ksz_ptp.c b/drivers/net/dsa/microchip/ksz_ptp.c index c8bfbe5e2157..997e4a76d0a6 100644 --- a/drivers/net/dsa/microchip/ksz_ptp.c +++ b/drivers/net/dsa/microchip/ksz_ptp.c @@ -1093,19 +1093,19 @@ static int ksz_ptp_msg_irq_setup(struct ksz_port *port, u8 n) static const char * const name[] = {"pdresp-msg", "xdreq-msg", "sync-msg"}; const struct ksz_dev_ops *ops = port->ksz_dev->dev_ops; + struct ksz_irq *ptpirq = &port->ptpirq; struct ksz_ptp_irq *ptpmsg_irq; ptpmsg_irq = &port->ptpmsg_irq[n]; + ptpmsg_irq->num = irq_create_mapping(ptpirq->domain, n); + if (!ptpmsg_irq->num) + return -EINVAL; ptpmsg_irq->port = port; ptpmsg_irq->ts_reg = ops->get_port_addr(port->num, ts_reg[n]); strscpy(ptpmsg_irq->name, name[n]); - ptpmsg_irq->num = irq_find_mapping(port->ptpirq.domain, n); - if (ptpmsg_irq->num < 0) - return ptpmsg_irq->num; - return request_threaded_irq(ptpmsg_irq->num, NULL, ksz_ptp_msg_thread_fn, IRQF_ONESHOT, ptpmsg_irq->name, ptpmsg_irq); @@ -1135,9 +1135,6 @@ int ksz_ptp_irq_setup(struct dsa_switch *ds, u8 p) if (!ptpirq->domain) return -ENOMEM; - for (irq = 0; irq < ptpirq->nirqs; irq++) - irq_create_mapping(ptpirq->domain, irq); - ptpirq->irq_num = irq_find_mapping(port->pirq.domain, PORT_SRC_PTP_INT); if (!ptpirq->irq_num) { ret = -EINVAL; @@ -1159,12 +1156,11 @@ int ksz_ptp_irq_setup(struct dsa_switch *ds, u8 p) out_ptp_msg: free_irq(ptpirq->irq_num, ptpirq); - while (irq--) + while (irq--) { free_irq(port->ptpmsg_irq[irq].num, &port->ptpmsg_irq[irq]); -out: - for (irq = 0; irq < ptpirq->nirqs; irq++) irq_dispose_mapping(port->ptpmsg_irq[irq].num); - + } +out: irq_domain_remove(ptpirq->domain); return ret; From 82fca3d8a4a34667f01ec2351a607135249c9cff Mon Sep 17 00:00:00 2001 From: Gui-Dong Han Date: Thu, 20 Nov 2025 20:06:57 +0800 Subject: [PATCH 23/47] atm/fore200e: Fix possible data race in fore200e_open() Protect access to fore200e->available_cell_rate with rate_mtx lock in the error handling path of fore200e_open() to prevent a data race. The field fore200e->available_cell_rate is a shared resource used to track available bandwidth. It is concurrently accessed by fore200e_open(), fore200e_close(), and fore200e_change_qos(). In fore200e_open(), the lock rate_mtx is correctly held when subtracting vcc->qos.txtp.max_pcr from available_cell_rate to reserve bandwidth. However, if the subsequent call to fore200e_activate_vcin() fails, the function restores the reserved bandwidth by adding back to available_cell_rate without holding the lock. This introduces a race condition because available_cell_rate is a global device resource shared across all VCCs. If the error path in fore200e_open() executes concurrently with operations like fore200e_close() or fore200e_change_qos() on other VCCs, a read-modify-write race occurs. Specifically, the error path reads the rate without the lock. If another CPU acquires the lock and modifies the rate (e.g., releasing bandwidth in fore200e_close()) between this read and the subsequent write, the error path will overwrite the concurrent update with a stale value. This results in incorrect bandwidth accounting. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable@vger.kernel.org Signed-off-by: Gui-Dong Han Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251120120657.2462194-1-hanguidong02@gmail.com Signed-off-by: Paolo Abeni --- drivers/atm/fore200e.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/atm/fore200e.c b/drivers/atm/fore200e.c index 4fea1149e003..f62e38571440 100644 --- a/drivers/atm/fore200e.c +++ b/drivers/atm/fore200e.c @@ -1374,7 +1374,9 @@ fore200e_open(struct atm_vcc *vcc) vcc->dev_data = NULL; + mutex_lock(&fore200e->rate_mtx); fore200e->available_cell_rate += vcc->qos.txtp.max_pcr; + mutex_unlock(&fore200e->rate_mtx); kfree(fore200e_vcc); return -EINVAL; From 081156ce13f8fa4e97b5148dc54d8c0ddf02117b Mon Sep 17 00:00:00 2001 From: Daniel Golle Date: Thu, 20 Nov 2025 15:02:19 +0000 Subject: [PATCH 24/47] net: phy: mxl-gpy: fix link properties on USXGMII and internal PHYs gpy_update_interface() returns early in case the PHY is internal or connected via USXGMII. In this case the gigabit master/slave property as well as MDI/MDI-X status also won't be read which seems wrong. Always read those properties by moving the logic to retrieve them to gpy_read_status(). Fixes: fd8825cd8c6fc ("net: phy: mxl-gpy: Add PHY Auto/MDI/MDI-X set driver for GPY211 chips") Fixes: 311abcdddc00a ("net: phy: add support to get Master-Slave configuration") Suggested-by: "Russell King (Oracle)" Signed-off-by: Daniel Golle Link: https://patch.msgid.link/71fccf3f56742116eb18cc070d2a9810479ea7f9.1763650701.git.daniel@makrotopia.org Signed-off-by: Paolo Abeni --- drivers/net/phy/mxl-gpy.c | 18 +++++++++++------- 1 file changed, 11 insertions(+), 7 deletions(-) diff --git a/drivers/net/phy/mxl-gpy.c b/drivers/net/phy/mxl-gpy.c index 221b315203d0..2a873f791733 100644 --- a/drivers/net/phy/mxl-gpy.c +++ b/drivers/net/phy/mxl-gpy.c @@ -578,13 +578,7 @@ static int gpy_update_interface(struct phy_device *phydev) break; } - if (phydev->speed == SPEED_2500 || phydev->speed == SPEED_1000) { - ret = genphy_read_master_slave(phydev); - if (ret < 0) - return ret; - } - - return gpy_update_mdix(phydev); + return 0; } static int gpy_read_status(struct phy_device *phydev) @@ -639,6 +633,16 @@ static int gpy_read_status(struct phy_device *phydev) ret = gpy_update_interface(phydev); if (ret < 0) return ret; + + if (phydev->speed == SPEED_2500 || phydev->speed == SPEED_1000) { + ret = genphy_read_master_slave(phydev); + if (ret < 0) + return ret; + } + + ret = gpy_update_mdix(phydev); + if (ret < 0) + return ret; } return 0; From 9780f535f8e0f20b4632b5a173ead71aa8f095d2 Mon Sep 17 00:00:00 2001 From: Horatiu Vultur Date: Fri, 21 Nov 2025 07:14:11 +0100 Subject: [PATCH 25/47] net: lan966x: Fix the initialization of taprio To initialize the taprio block in lan966x, it is required to configure the register REVISIT_DLY. The purpose of this register is to set the delay before revisit the next gate and the value of this register depends on the system clock. The problem is that the we calculated wrong the value of the system clock period in picoseconds. The actual system clock is ~165.617754MHZ and this correspond to a period of 6038 pico seconds and not 15125 as currently set. Fixes: e462b2717380b4 ("net: lan966x: Add offload support for taprio") Signed-off-by: Horatiu Vultur Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251121061411.810571-1-horatiu.vultur@microchip.com Signed-off-by: Paolo Abeni --- drivers/net/ethernet/microchip/lan966x/lan966x_ptp.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/microchip/lan966x/lan966x_ptp.c b/drivers/net/ethernet/microchip/lan966x/lan966x_ptp.c index b4377b8613c3..8c40db90ee8f 100644 --- a/drivers/net/ethernet/microchip/lan966x/lan966x_ptp.c +++ b/drivers/net/ethernet/microchip/lan966x/lan966x_ptp.c @@ -1,11 +1,14 @@ // SPDX-License-Identifier: GPL-2.0+ #include +#include #include "lan966x_main.h" #include "vcap_api.h" #include "vcap_api_client.h" +#define LAN9X66_CLOCK_RATE 165617754 + #define LAN966X_MAX_PTP_ID 512 /* Represents 1ppm adjustment in 2^59 format with 6.037735849ns as reference @@ -1126,5 +1129,5 @@ void lan966x_ptp_rxtstamp(struct lan966x *lan966x, struct sk_buff *skb, u32 lan966x_ptp_get_period_ps(void) { /* This represents the system clock period in picoseconds */ - return 15125; + return PICO / LAN9X66_CLOCK_RATE; } From d2099d9f16dbfa1c5266d4230ff7860047bb0b68 Mon Sep 17 00:00:00 2001 From: Danielle Costantino Date: Mon, 24 Nov 2025 10:00:43 -0800 Subject: [PATCH 26/47] net/mlx5e: Fix validation logic in rate limiting The rate limiting validation condition currently checks the output variable max_bw_value[i] instead of the input value maxrate->tc_maxrate[i]. This causes the validation to compare an uninitialized or stale value rather than the actual requested rate. The condition should check the input rate to properly validate against the upper limit: } else if (maxrate->tc_maxrate[i] <= upper_limit_gbps) { This aligns with the pattern used in the first branch, which correctly checks maxrate->tc_maxrate[i] against upper_limit_mbps. The current implementation can lead to unreliable validation behavior: - For rates between 25.5 Gbps and 255 Gbps, if max_bw_value[i] is 0 from initialization, the GBPS path may be taken regardless of whether the actual rate is within bounds - When processing multiple TCs (i > 0), max_bw_value[i] contains the value computed for the previous TC, affecting the validation logic - The overflow check for rates exceeding 255 Gbps may not trigger consistently depending on previous array values This patch ensures the validation correctly examines the requested rate value for proper bounds checking. Fixes: 43b27d1bd88a ("net/mlx5e: Fix wraparound in rate limiting for values above 255 Gbps") Signed-off-by: Danielle Costantino Reviewed-by: Gal Pressman Link: https://patch.msgid.link/20251124180043.2314428-1-dcostantino@meta.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c index 9b93da4d52f6..cf8f14ce4cd5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c @@ -627,7 +627,7 @@ static int mlx5e_dcbnl_ieee_setmaxrate(struct net_device *netdev, MLX5E_100MB); max_bw_value[i] = max_bw_value[i] ? max_bw_value[i] : 1; max_bw_unit[i] = MLX5_100_MBPS_UNIT; - } else if (max_bw_value[i] <= upper_limit_gbps) { + } else if (maxrate->tc_maxrate[i] <= upper_limit_gbps) { max_bw_value[i] = div_u64(maxrate->tc_maxrate[i], MLX5E_1GB); max_bw_unit[i] = MLX5_GBPS_UNIT; From 0ae9cfc454ea5ead5f3ddbdfe2e70270d8e2c8ef Mon Sep 17 00:00:00 2001 From: "Nikola Z. Ivanov" Date: Sat, 22 Nov 2025 02:20:27 +0200 Subject: [PATCH 27/47] team: Move team device type change at the end of team_port_add Attempting to add a port device that is already up will expectedly fail, but not before modifying the team device header_ops. In the case of the syzbot reproducer the gre0 device is already in state UP when it attempts to add it as a port device of team0, this fails but before that header_ops->create of team0 is changed from eth_header to ipgre_header in the call to team_dev_type_check_change. Later when we end up in ipgre_header() struct ip_tunnel* points to nonsense as the private data of the device still holds a struct team. Example sequence of iproute2 commands to reproduce the hang/BUG(): ip link add dev team0 type team ip link add dev gre0 type gre ip link set dev gre0 up ip link set dev gre0 master team0 ip link set dev team0 up ping -I team0 1.1.1.1 Move team_dev_type_check_change down where all other checks have passed as it changes the dev type with no way to restore it in case one of the checks that follow it fail. Also make sure to preserve the origial mtu assignment: - If port_dev is not the same type as dev, dev takes mtu from port_dev - If port_dev is the same type as dev, port_dev takes mtu from dev This is done by adding a conditional before the call to dev_set_mtu to prevent it from assigning port_dev->mtu = dev->mtu and instead letting team_dev_type_check_change assign dev->mtu = port_dev->mtu. The conditional is needed because the patch moves the call to team_dev_type_check_change past dev_set_mtu. Testing: - team device driver in-tree selftests - Add/remove various devices as slaves of team device - syzbot Reported-by: syzbot+a2a3b519de727b0f7903@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=a2a3b519de727b0f7903 Fixes: 1d76efe1577b ("team: add support for non-ethernet devices") Signed-off-by: Nikola Z. Ivanov Reviewed-by: Jiri Pirko Link: https://patch.msgid.link/20251122002027.695151-1-zlatistiv@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/team/team_core.c | 23 +++++++++++++++-------- 1 file changed, 15 insertions(+), 8 deletions(-) diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c index 17f07eb0ee52..25562b17debe 100644 --- a/drivers/net/team/team_core.c +++ b/drivers/net/team/team_core.c @@ -1191,10 +1191,6 @@ static int team_port_add(struct team *team, struct net_device *port_dev, return -EPERM; } - err = team_dev_type_check_change(dev, port_dev); - if (err) - return err; - if (port_dev->flags & IFF_UP) { NL_SET_ERR_MSG(extack, "Device is up. Set it down before adding it as a team port"); netdev_err(dev, "Device %s is up. Set it down before adding it as a team port\n", @@ -1212,10 +1208,16 @@ static int team_port_add(struct team *team, struct net_device *port_dev, INIT_LIST_HEAD(&port->qom_list); port->orig.mtu = port_dev->mtu; - err = dev_set_mtu(port_dev, dev->mtu); - if (err) { - netdev_dbg(dev, "Error %d calling dev_set_mtu\n", err); - goto err_set_mtu; + /* + * MTU assignment will be handled in team_dev_type_check_change + * if dev and port_dev are of different types + */ + if (dev->type == port_dev->type) { + err = dev_set_mtu(port_dev, dev->mtu); + if (err) { + netdev_dbg(dev, "Error %d calling dev_set_mtu\n", err); + goto err_set_mtu; + } } memcpy(port->orig.dev_addr, port_dev->dev_addr, port_dev->addr_len); @@ -1290,6 +1292,10 @@ static int team_port_add(struct team *team, struct net_device *port_dev, } } + err = team_dev_type_check_change(dev, port_dev); + if (err) + goto err_set_dev_type; + if (dev->flags & IFF_UP) { netif_addr_lock_bh(dev); dev_uc_sync_multiple(port_dev, dev); @@ -1308,6 +1314,7 @@ static int team_port_add(struct team *team, struct net_device *port_dev, return 0; +err_set_dev_type: err_set_slave_promisc: __team_option_inst_del_port(team, port); From f5bce28f6b9125502abec4a67d68eabcd24b3b17 Mon Sep 17 00:00:00 2001 From: Alexey Kodanev Date: Fri, 21 Nov 2025 12:38:34 +0000 Subject: [PATCH 28/47] net: sxgbe: fix potential NULL dereference in sxgbe_rx() Currently, when skb is null, the driver prints an error and then dereferences skb on the next line. To fix this, let's add a 'break' after the error message to switch to sxgbe_rx_refill(), which is similar to the approach taken by the other drivers in this particular case, e.g. calxeda with xgmac_rx(). Found during a code review. Fixes: 1edb9ca69e8a ("net: sxgbe: add basic framework for Samsung 10Gb ethernet driver") Signed-off-by: Alexey Kodanev Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251121123834.97748-1-aleksei.kodanev@bell-sw.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c b/drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c index 75bad561b352..849c5a6c2af1 100644 --- a/drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c +++ b/drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c @@ -1521,8 +1521,10 @@ static int sxgbe_rx(struct sxgbe_priv_data *priv, int limit) skb = priv->rxq[qnum]->rx_skbuff[entry]; - if (unlikely(!skb)) + if (unlikely(!skb)) { netdev_err(priv->dev, "rx descriptor is not consistent\n"); + break; + } prefetch(skb->data - NET_IP_ALIGN); priv->rxq[qnum]->rx_skbuff[entry] = NULL; From ae1737e7339b513f8c2fc21b500a0fc215d155c3 Mon Sep 17 00:00:00 2001 From: Heiner Kallweit Date: Sat, 22 Nov 2025 15:23:02 +0100 Subject: [PATCH 29/47] r8169: fix RTL8127 hang on suspend/shutdown There have been reports that RTL8127 hangs on suspend and shutdown, partially disappearing from lspci until power-cycling. According to Realtek disabling PLL's when switching to D3 should be avoided on that chip version. Fix this by aligning disabling PLL's with the vendor drivers, what in addition results in PLL's not being disabled when switching to D3hot on other chip versions. Fixes: f24f7b2f3af9 ("r8169: add support for RTL8127A") Tested-by: Fabio Baltieri Cc: stable@vger.kernel.org Signed-off-by: Heiner Kallweit Link: https://patch.msgid.link/d7faae7e-66bc-404a-a432-3a496600575f@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/realtek/r8169_main.c | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-) diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c index d18734fe12e4..853aabedb128 100644 --- a/drivers/net/ethernet/realtek/r8169_main.c +++ b/drivers/net/ethernet/realtek/r8169_main.c @@ -1514,11 +1514,20 @@ static enum rtl_dash_type rtl_get_dash_type(struct rtl8169_private *tp) static void rtl_set_d3_pll_down(struct rtl8169_private *tp, bool enable) { - if (tp->mac_version >= RTL_GIGA_MAC_VER_25 && - tp->mac_version != RTL_GIGA_MAC_VER_28 && - tp->mac_version != RTL_GIGA_MAC_VER_31 && - tp->mac_version != RTL_GIGA_MAC_VER_38) - r8169_mod_reg8_cond(tp, PMCH, D3_NO_PLL_DOWN, !enable); + switch (tp->mac_version) { + case RTL_GIGA_MAC_VER_02 ... RTL_GIGA_MAC_VER_24: + case RTL_GIGA_MAC_VER_28: + case RTL_GIGA_MAC_VER_31: + case RTL_GIGA_MAC_VER_38: + break; + case RTL_GIGA_MAC_VER_80: + r8169_mod_reg8_cond(tp, PMCH, D3_NO_PLL_DOWN, true); + break; + default: + r8169_mod_reg8_cond(tp, PMCH, D3HOT_NO_PLL_DOWN, true); + r8169_mod_reg8_cond(tp, PMCH, D3COLD_NO_PLL_DOWN, !enable); + break; + } } static void rtl_reset_packet_filter(struct rtl8169_private *tp) From 0ebc27a4c67d44e5ce88d21cdad8201862b78837 Mon Sep 17 00:00:00 2001 From: Fernando Fernandez Mancera Date: Mon, 24 Nov 2025 18:14:09 +0100 Subject: [PATCH 30/47] xsk: avoid data corruption on cq descriptor number Since commit 30f241fcf52a ("xsk: Fix immature cq descriptor production"), the descriptor number is stored in skb control block and xsk_cq_submit_addr_locked() relies on it to put the umem addrs onto pool's completion queue. skb control block shouldn't be used for this purpose as after transmit xsk doesn't have control over it and other subsystems could use it. This leads to the following kernel panic due to a NULL pointer dereference. BUG: kernel NULL pointer dereference, address: 0000000000000000 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 0 P4D 0 Oops: Oops: 0000 [#1] SMP NOPTI CPU: 2 UID: 1 PID: 927 Comm: p4xsk.bin Not tainted 6.16.12+deb14-cloud-amd64 #1 PREEMPT(lazy) Debian 6.16.12-1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.17.0-debian-1.17.0-1 04/01/2014 RIP: 0010:xsk_destruct_skb+0xd0/0x180 [...] Call Trace: ? napi_complete_done+0x7a/0x1a0 ip_rcv_core+0x1bb/0x340 ip_rcv+0x30/0x1f0 __netif_receive_skb_one_core+0x85/0xa0 process_backlog+0x87/0x130 __napi_poll+0x28/0x180 net_rx_action+0x339/0x420 handle_softirqs+0xdc/0x320 ? handle_edge_irq+0x90/0x1e0 do_softirq.part.0+0x3b/0x60 __local_bh_enable_ip+0x60/0x70 __dev_direct_xmit+0x14e/0x1f0 __xsk_generic_xmit+0x482/0xb70 ? __remove_hrtimer+0x41/0xa0 ? __xsk_generic_xmit+0x51/0xb70 ? _raw_spin_unlock_irqrestore+0xe/0x40 xsk_sendmsg+0xda/0x1c0 __sys_sendto+0x1ee/0x200 __x64_sys_sendto+0x24/0x30 do_syscall_64+0x84/0x2f0 ? __pfx_pollwake+0x10/0x10 ? __rseq_handle_notify_resume+0xad/0x4c0 ? restore_fpregs_from_fpstate+0x3c/0x90 ? switch_fpu_return+0x5b/0xe0 ? do_syscall_64+0x204/0x2f0 ? do_syscall_64+0x204/0x2f0 ? do_syscall_64+0x204/0x2f0 entry_SYSCALL_64_after_hwframe+0x76/0x7e [...] Kernel panic - not syncing: Fatal exception in interrupt Kernel Offset: 0x1c000000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff) Instead use the skb destructor_arg pointer along with pointer tagging. As pointers are always aligned to 8B, use the bottom bit to indicate whether this a single address or an allocated struct containing several addresses. Fixes: 30f241fcf52a ("xsk: Fix immature cq descriptor production") Closes: https://lore.kernel.org/netdev/0435b904-f44f-48f8-afb0-68868474bf1c@nop.hu/ Suggested-by: Jakub Kicinski Signed-off-by: Fernando Fernandez Mancera Reviewed-by: Maciej Fijalkowski Reviewed-by: Jason Xing Link: https://patch.msgid.link/20251124171409.3845-1-fmancera@suse.de Signed-off-by: Jakub Kicinski --- net/xdp/xsk.c | 143 +++++++++++++++++++++++++++++++------------------- 1 file changed, 88 insertions(+), 55 deletions(-) diff --git a/net/xdp/xsk.c b/net/xdp/xsk.c index 7b0c68a70888..69bbcca8ac75 100644 --- a/net/xdp/xsk.c +++ b/net/xdp/xsk.c @@ -36,20 +36,13 @@ #define TX_BATCH_SIZE 32 #define MAX_PER_SOCKET_BUDGET 32 -struct xsk_addr_node { - u64 addr; - struct list_head addr_node; -}; - -struct xsk_addr_head { +struct xsk_addrs { u32 num_descs; - struct list_head addrs_list; + u64 addrs[MAX_SKB_FRAGS + 1]; }; static struct kmem_cache *xsk_tx_generic_cache; -#define XSKCB(skb) ((struct xsk_addr_head *)((skb)->cb)) - void xsk_set_rx_need_wakeup(struct xsk_buff_pool *pool) { if (pool->cached_need_wakeup & XDP_WAKEUP_RX) @@ -558,29 +551,68 @@ static int xsk_cq_reserve_locked(struct xsk_buff_pool *pool) return ret; } +static bool xsk_skb_destructor_is_addr(struct sk_buff *skb) +{ + return (uintptr_t)skb_shinfo(skb)->destructor_arg & 0x1UL; +} + +static u64 xsk_skb_destructor_get_addr(struct sk_buff *skb) +{ + return (u64)((uintptr_t)skb_shinfo(skb)->destructor_arg & ~0x1UL); +} + +static void xsk_skb_destructor_set_addr(struct sk_buff *skb, u64 addr) +{ + skb_shinfo(skb)->destructor_arg = (void *)((uintptr_t)addr | 0x1UL); +} + +static void xsk_inc_num_desc(struct sk_buff *skb) +{ + struct xsk_addrs *xsk_addr; + + if (!xsk_skb_destructor_is_addr(skb)) { + xsk_addr = (struct xsk_addrs *)skb_shinfo(skb)->destructor_arg; + xsk_addr->num_descs++; + } +} + +static u32 xsk_get_num_desc(struct sk_buff *skb) +{ + struct xsk_addrs *xsk_addr; + + if (xsk_skb_destructor_is_addr(skb)) + return 1; + + xsk_addr = (struct xsk_addrs *)skb_shinfo(skb)->destructor_arg; + + return xsk_addr->num_descs; +} + static void xsk_cq_submit_addr_locked(struct xsk_buff_pool *pool, struct sk_buff *skb) { - struct xsk_addr_node *pos, *tmp; + u32 num_descs = xsk_get_num_desc(skb); + struct xsk_addrs *xsk_addr; u32 descs_processed = 0; unsigned long flags; - u32 idx; + u32 idx, i; spin_lock_irqsave(&pool->cq_lock, flags); idx = xskq_get_prod(pool->cq); - xskq_prod_write_addr(pool->cq, idx, - (u64)(uintptr_t)skb_shinfo(skb)->destructor_arg); - descs_processed++; + if (unlikely(num_descs > 1)) { + xsk_addr = (struct xsk_addrs *)skb_shinfo(skb)->destructor_arg; - if (unlikely(XSKCB(skb)->num_descs > 1)) { - list_for_each_entry_safe(pos, tmp, &XSKCB(skb)->addrs_list, addr_node) { + for (i = 0; i < num_descs; i++) { xskq_prod_write_addr(pool->cq, idx + descs_processed, - pos->addr); + xsk_addr->addrs[i]); descs_processed++; - list_del(&pos->addr_node); - kmem_cache_free(xsk_tx_generic_cache, pos); } + kmem_cache_free(xsk_tx_generic_cache, xsk_addr); + } else { + xskq_prod_write_addr(pool->cq, idx, + xsk_skb_destructor_get_addr(skb)); + descs_processed++; } xskq_prod_submit_n(pool->cq, descs_processed); spin_unlock_irqrestore(&pool->cq_lock, flags); @@ -595,16 +627,6 @@ static void xsk_cq_cancel_locked(struct xsk_buff_pool *pool, u32 n) spin_unlock_irqrestore(&pool->cq_lock, flags); } -static void xsk_inc_num_desc(struct sk_buff *skb) -{ - XSKCB(skb)->num_descs++; -} - -static u32 xsk_get_num_desc(struct sk_buff *skb) -{ - return XSKCB(skb)->num_descs; -} - static void xsk_destruct_skb(struct sk_buff *skb) { struct xsk_tx_metadata_compl *compl = &skb_shinfo(skb)->xsk_meta; @@ -621,27 +643,22 @@ static void xsk_destruct_skb(struct sk_buff *skb) static void xsk_skb_init_misc(struct sk_buff *skb, struct xdp_sock *xs, u64 addr) { - BUILD_BUG_ON(sizeof(struct xsk_addr_head) > sizeof(skb->cb)); - INIT_LIST_HEAD(&XSKCB(skb)->addrs_list); skb->dev = xs->dev; skb->priority = READ_ONCE(xs->sk.sk_priority); skb->mark = READ_ONCE(xs->sk.sk_mark); - XSKCB(skb)->num_descs = 0; skb->destructor = xsk_destruct_skb; - skb_shinfo(skb)->destructor_arg = (void *)(uintptr_t)addr; + xsk_skb_destructor_set_addr(skb, addr); } static void xsk_consume_skb(struct sk_buff *skb) { struct xdp_sock *xs = xdp_sk(skb->sk); u32 num_descs = xsk_get_num_desc(skb); - struct xsk_addr_node *pos, *tmp; + struct xsk_addrs *xsk_addr; if (unlikely(num_descs > 1)) { - list_for_each_entry_safe(pos, tmp, &XSKCB(skb)->addrs_list, addr_node) { - list_del(&pos->addr_node); - kmem_cache_free(xsk_tx_generic_cache, pos); - } + xsk_addr = (struct xsk_addrs *)skb_shinfo(skb)->destructor_arg; + kmem_cache_free(xsk_tx_generic_cache, xsk_addr); } skb->destructor = sock_wfree; @@ -701,7 +718,6 @@ static struct sk_buff *xsk_build_skb_zerocopy(struct xdp_sock *xs, { struct xsk_buff_pool *pool = xs->pool; u32 hr, len, ts, offset, copy, copied; - struct xsk_addr_node *xsk_addr; struct sk_buff *skb = xs->skb; struct page *page; void *buffer; @@ -727,16 +743,26 @@ static struct sk_buff *xsk_build_skb_zerocopy(struct xdp_sock *xs, return ERR_PTR(err); } } else { - xsk_addr = kmem_cache_zalloc(xsk_tx_generic_cache, GFP_KERNEL); - if (!xsk_addr) - return ERR_PTR(-ENOMEM); + struct xsk_addrs *xsk_addr; + + if (xsk_skb_destructor_is_addr(skb)) { + xsk_addr = kmem_cache_zalloc(xsk_tx_generic_cache, + GFP_KERNEL); + if (!xsk_addr) + return ERR_PTR(-ENOMEM); + + xsk_addr->num_descs = 1; + xsk_addr->addrs[0] = xsk_skb_destructor_get_addr(skb); + skb_shinfo(skb)->destructor_arg = (void *)xsk_addr; + } else { + xsk_addr = (struct xsk_addrs *)skb_shinfo(skb)->destructor_arg; + } /* in case of -EOVERFLOW that could happen below, * xsk_consume_skb() will release this node as whole skb * would be dropped, which implies freeing all list elements */ - xsk_addr->addr = desc->addr; - list_add_tail(&xsk_addr->addr_node, &XSKCB(skb)->addrs_list); + xsk_addr->addrs[xsk_addr->num_descs] = desc->addr; } len = desc->len; @@ -813,10 +839,25 @@ static struct sk_buff *xsk_build_skb(struct xdp_sock *xs, } } else { int nr_frags = skb_shinfo(skb)->nr_frags; - struct xsk_addr_node *xsk_addr; + struct xsk_addrs *xsk_addr; struct page *page; u8 *vaddr; + if (xsk_skb_destructor_is_addr(skb)) { + xsk_addr = kmem_cache_zalloc(xsk_tx_generic_cache, + GFP_KERNEL); + if (!xsk_addr) { + err = -ENOMEM; + goto free_err; + } + + xsk_addr->num_descs = 1; + xsk_addr->addrs[0] = xsk_skb_destructor_get_addr(skb); + skb_shinfo(skb)->destructor_arg = (void *)xsk_addr; + } else { + xsk_addr = (struct xsk_addrs *)skb_shinfo(skb)->destructor_arg; + } + if (unlikely(nr_frags == (MAX_SKB_FRAGS - 1) && xp_mb_desc(desc))) { err = -EOVERFLOW; goto free_err; @@ -828,13 +869,6 @@ static struct sk_buff *xsk_build_skb(struct xdp_sock *xs, goto free_err; } - xsk_addr = kmem_cache_zalloc(xsk_tx_generic_cache, GFP_KERNEL); - if (!xsk_addr) { - __free_page(page); - err = -ENOMEM; - goto free_err; - } - vaddr = kmap_local_page(page); memcpy(vaddr, buffer, len); kunmap_local(vaddr); @@ -842,8 +876,7 @@ static struct sk_buff *xsk_build_skb(struct xdp_sock *xs, skb_add_rx_frag(skb, nr_frags, page, 0, len, PAGE_SIZE); refcount_add(PAGE_SIZE, &xs->sk.sk_wmem_alloc); - xsk_addr->addr = desc->addr; - list_add_tail(&xsk_addr->addr_node, &XSKCB(skb)->addrs_list); + xsk_addr->addrs[xsk_addr->num_descs] = desc->addr; } } @@ -1904,7 +1937,7 @@ static int __init xsk_init(void) goto out_pernet; xsk_tx_generic_cache = kmem_cache_create("xsk_generic_xmit_cache", - sizeof(struct xsk_addr_node), + sizeof(struct xsk_addrs), 0, SLAB_HWCACHE_ALIGN, NULL); if (!xsk_tx_generic_cache) { err = -ENOMEM; From 6d849ff573722afcf5508d2800017bdd40f27eb9 Mon Sep 17 00:00:00 2001 From: Biju Das Date: Tue, 18 Nov 2025 12:39:25 +0000 Subject: [PATCH 31/47] can: rcar_canfd: Fix CAN-FD mode as default The commit 5cff263606a1 ("can: rcar_canfd: Fix controller mode setting") has aligned with the flow mentioned in the hardware manual for all SoCs except R-Car Gen3 and RZ/G2L SoCs. On R-Car Gen4 and RZ/G3E SoCs, due to the wrong logic in the commit[1] sets the default mode to FD-Only mode instead of CAN-FD mode. This patch sets the CAN-FD mode as the default for all SoCs by dropping the rcar_canfd_set_mode() as some SoC requires mode setting in global reset mode, and the rest of the SoCs in channel reset mode and update the rcar_canfd_reset_controller() to take care of these constraints. Moreover, the RZ/G3E and R-Car Gen4 SoCs support 3 modes compared to 2 modes on the R-Car Gen3. Use inverted logic in rcar_canfd_reset_controller() to simplify the code later to support FD-only mode. [1] commit 45721c406dcf ("can: rcar_canfd: Add support for r8a779a0 SoC") Fixes: 5cff263606a1 ("can: rcar_canfd: Fix controller mode setting") Cc: stable@vger.kernel.org Signed-off-by: Biju Das Link: https://patch.msgid.link/20251118123926.193445-1-biju.das.jz@bp.renesas.com Signed-off-by: Marc Kleine-Budde --- drivers/net/can/rcar/rcar_canfd.c | 53 ++++++++++++++++++------------- 1 file changed, 31 insertions(+), 22 deletions(-) diff --git a/drivers/net/can/rcar/rcar_canfd.c b/drivers/net/can/rcar/rcar_canfd.c index 45d36adb51b7..4c0d7d26df9f 100644 --- a/drivers/net/can/rcar/rcar_canfd.c +++ b/drivers/net/can/rcar/rcar_canfd.c @@ -709,6 +709,11 @@ static void rcar_canfd_set_bit_reg(void __iomem *addr, u32 val) rcar_canfd_update(val, val, addr); } +static void rcar_canfd_clear_bit_reg(void __iomem *addr, u32 val) +{ + rcar_canfd_update(val, 0, addr); +} + static void rcar_canfd_update_bit_reg(void __iomem *addr, u32 mask, u32 val) { rcar_canfd_update(mask, val, addr); @@ -755,25 +760,6 @@ static void rcar_canfd_set_rnc(struct rcar_canfd_global *gpriv, unsigned int ch, rcar_canfd_set_bit(gpriv->base, RCANFD_GAFLCFG(w), rnc); } -static void rcar_canfd_set_mode(struct rcar_canfd_global *gpriv) -{ - if (gpriv->info->ch_interface_mode) { - u32 ch, val = gpriv->fdmode ? RCANFD_GEN4_FDCFG_FDOE - : RCANFD_GEN4_FDCFG_CLOE; - - for_each_set_bit(ch, &gpriv->channels_mask, - gpriv->info->max_channels) - rcar_canfd_set_bit_reg(&gpriv->fcbase[ch].cfdcfg, val); - } else { - if (gpriv->fdmode) - rcar_canfd_set_bit(gpriv->base, RCANFD_GRMCFG, - RCANFD_GRMCFG_RCMC); - else - rcar_canfd_clear_bit(gpriv->base, RCANFD_GRMCFG, - RCANFD_GRMCFG_RCMC); - } -} - static int rcar_canfd_reset_controller(struct rcar_canfd_global *gpriv) { struct device *dev = &gpriv->pdev->dev; @@ -806,6 +792,16 @@ static int rcar_canfd_reset_controller(struct rcar_canfd_global *gpriv) /* Reset Global error flags */ rcar_canfd_write(gpriv->base, RCANFD_GERFL, 0x0); + /* Set the controller into appropriate mode */ + if (!gpriv->info->ch_interface_mode) { + if (gpriv->fdmode) + rcar_canfd_set_bit(gpriv->base, RCANFD_GRMCFG, + RCANFD_GRMCFG_RCMC); + else + rcar_canfd_clear_bit(gpriv->base, RCANFD_GRMCFG, + RCANFD_GRMCFG_RCMC); + } + /* Transition all Channels to reset mode */ for_each_set_bit(ch, &gpriv->channels_mask, gpriv->info->max_channels) { rcar_canfd_clear_bit(gpriv->base, @@ -823,10 +819,23 @@ static int rcar_canfd_reset_controller(struct rcar_canfd_global *gpriv) dev_dbg(dev, "channel %u reset failed\n", ch); return err; } - } - /* Set the controller into appropriate mode */ - rcar_canfd_set_mode(gpriv); + /* Set the controller into appropriate mode */ + if (gpriv->info->ch_interface_mode) { + /* Do not set CLOE and FDOE simultaneously */ + if (!gpriv->fdmode) { + rcar_canfd_clear_bit_reg(&gpriv->fcbase[ch].cfdcfg, + RCANFD_GEN4_FDCFG_FDOE); + rcar_canfd_set_bit_reg(&gpriv->fcbase[ch].cfdcfg, + RCANFD_GEN4_FDCFG_CLOE); + } else { + rcar_canfd_clear_bit_reg(&gpriv->fcbase[ch].cfdcfg, + RCANFD_GEN4_FDCFG_FDOE); + rcar_canfd_clear_bit_reg(&gpriv->fcbase[ch].cfdcfg, + RCANFD_GEN4_FDCFG_CLOE); + } + } + } return 0; } From 0c922106d7a58d106c6a5c52a741ae101cfaf088 Mon Sep 17 00:00:00 2001 From: Shaurya Rane Date: Wed, 26 Nov 2025 14:27:18 +0530 Subject: [PATCH 32/47] net/sched: em_canid: fix uninit-value in em_canid_match Use pskb_may_pull() to ensure a complete CAN frame is present in the linear data buffer before reading the CAN ID. A simple skb->len check is insufficient because it only verifies the total data length but does not guarantee the data is present in skb->data (it could be in fragments). pskb_may_pull() both validates the length and pulls fragmented data into the linear buffer if necessary, making it safe to directly access skb->data. Reported-by: syzbot+5d8269a1e099279152bc@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=5d8269a1e099279152bc Fixes: f057bbb6f9ed ("net: em_canid: Ematch rule to match CAN frames according to their identifiers") Signed-off-by: Shaurya Rane Link: https://patch.msgid.link/20251126085718.50808-1-ssranevjti@gmail.com Signed-off-by: Marc Kleine-Budde --- net/sched/em_canid.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/net/sched/em_canid.c b/net/sched/em_canid.c index 5337bc462755..2d27f91d8441 100644 --- a/net/sched/em_canid.c +++ b/net/sched/em_canid.c @@ -99,6 +99,9 @@ static int em_canid_match(struct sk_buff *skb, struct tcf_ematch *m, int i; const struct can_filter *lp; + if (!pskb_may_pull(skb, CAN_MTU)) + return 0; + can_id = em_canid_get_id(skb); if (can_id & CAN_EFF_FLAG) { From 779bcdd4b9ae6566f309043c53c946e8ac0015fd Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Thu, 20 Nov 2025 10:29:50 +0800 Subject: [PATCH 33/47] vhost: rewind next_avail_head while discarding descriptors When discarding descriptors with IN_ORDER, we should rewind next_avail_head otherwise it would run out of sync with last_avail_idx. This would cause driver to report "id X is not a head". Fixing this by returning the number of descriptors that is used for each buffer via vhost_get_vq_desc_n() so caller can use the value while discarding descriptors. Fixes: 67a873df0c41 ("vhost: basic in order support") Cc: stable@vger.kernel.org Signed-off-by: Jason Wang Acked-by: Michael S. Tsirkin Link: https://patch.msgid.link/20251120022950.10117-1-jasowang@redhat.com Signed-off-by: Jakub Kicinski --- drivers/vhost/net.c | 53 ++++++++++++++++++------------ drivers/vhost/vhost.c | 76 +++++++++++++++++++++++++++++++++++-------- drivers/vhost/vhost.h | 10 +++++- 3 files changed, 103 insertions(+), 36 deletions(-) diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c index 35ded4330431..8f7f50acb6d6 100644 --- a/drivers/vhost/net.c +++ b/drivers/vhost/net.c @@ -592,14 +592,15 @@ static void vhost_net_busy_poll(struct vhost_net *net, static int vhost_net_tx_get_vq_desc(struct vhost_net *net, struct vhost_net_virtqueue *tnvq, unsigned int *out_num, unsigned int *in_num, - struct msghdr *msghdr, bool *busyloop_intr) + struct msghdr *msghdr, bool *busyloop_intr, + unsigned int *ndesc) { struct vhost_net_virtqueue *rnvq = &net->vqs[VHOST_NET_VQ_RX]; struct vhost_virtqueue *rvq = &rnvq->vq; struct vhost_virtqueue *tvq = &tnvq->vq; - int r = vhost_get_vq_desc(tvq, tvq->iov, ARRAY_SIZE(tvq->iov), - out_num, in_num, NULL, NULL); + int r = vhost_get_vq_desc_n(tvq, tvq->iov, ARRAY_SIZE(tvq->iov), + out_num, in_num, NULL, NULL, ndesc); if (r == tvq->num && tvq->busyloop_timeout) { /* Flush batched packets first */ @@ -610,8 +611,8 @@ static int vhost_net_tx_get_vq_desc(struct vhost_net *net, vhost_net_busy_poll(net, rvq, tvq, busyloop_intr, false); - r = vhost_get_vq_desc(tvq, tvq->iov, ARRAY_SIZE(tvq->iov), - out_num, in_num, NULL, NULL); + r = vhost_get_vq_desc_n(tvq, tvq->iov, ARRAY_SIZE(tvq->iov), + out_num, in_num, NULL, NULL, ndesc); } return r; @@ -642,12 +643,14 @@ static int get_tx_bufs(struct vhost_net *net, struct vhost_net_virtqueue *nvq, struct msghdr *msg, unsigned int *out, unsigned int *in, - size_t *len, bool *busyloop_intr) + size_t *len, bool *busyloop_intr, + unsigned int *ndesc) { struct vhost_virtqueue *vq = &nvq->vq; int ret; - ret = vhost_net_tx_get_vq_desc(net, nvq, out, in, msg, busyloop_intr); + ret = vhost_net_tx_get_vq_desc(net, nvq, out, in, msg, + busyloop_intr, ndesc); if (ret < 0 || ret == vq->num) return ret; @@ -766,6 +769,7 @@ static void handle_tx_copy(struct vhost_net *net, struct socket *sock) int sent_pkts = 0; bool sock_can_batch = (sock->sk->sk_sndbuf == INT_MAX); bool in_order = vhost_has_feature(vq, VIRTIO_F_IN_ORDER); + unsigned int ndesc = 0; do { bool busyloop_intr = false; @@ -774,7 +778,7 @@ static void handle_tx_copy(struct vhost_net *net, struct socket *sock) vhost_tx_batch(net, nvq, sock, &msg); head = get_tx_bufs(net, nvq, &msg, &out, &in, &len, - &busyloop_intr); + &busyloop_intr, &ndesc); /* On error, stop handling until the next kick. */ if (unlikely(head < 0)) break; @@ -806,7 +810,7 @@ static void handle_tx_copy(struct vhost_net *net, struct socket *sock) goto done; } else if (unlikely(err != -ENOSPC)) { vhost_tx_batch(net, nvq, sock, &msg); - vhost_discard_vq_desc(vq, 1); + vhost_discard_vq_desc(vq, 1, ndesc); vhost_net_enable_vq(net, vq); break; } @@ -829,7 +833,7 @@ static void handle_tx_copy(struct vhost_net *net, struct socket *sock) err = sock->ops->sendmsg(sock, &msg, len); if (unlikely(err < 0)) { if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) { - vhost_discard_vq_desc(vq, 1); + vhost_discard_vq_desc(vq, 1, ndesc); vhost_net_enable_vq(net, vq); break; } @@ -868,6 +872,7 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock) int err; struct vhost_net_ubuf_ref *ubufs; struct ubuf_info_msgzc *ubuf; + unsigned int ndesc = 0; bool zcopy_used; int sent_pkts = 0; @@ -879,7 +884,7 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock) busyloop_intr = false; head = get_tx_bufs(net, nvq, &msg, &out, &in, &len, - &busyloop_intr); + &busyloop_intr, &ndesc); /* On error, stop handling until the next kick. */ if (unlikely(head < 0)) break; @@ -941,7 +946,7 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock) vq->heads[ubuf->desc].len = VHOST_DMA_DONE_LEN; } if (retry) { - vhost_discard_vq_desc(vq, 1); + vhost_discard_vq_desc(vq, 1, ndesc); vhost_net_enable_vq(net, vq); break; } @@ -1045,11 +1050,12 @@ static int get_rx_bufs(struct vhost_net_virtqueue *nvq, unsigned *iovcount, struct vhost_log *log, unsigned *log_num, - unsigned int quota) + unsigned int quota, + unsigned int *ndesc) { struct vhost_virtqueue *vq = &nvq->vq; bool in_order = vhost_has_feature(vq, VIRTIO_F_IN_ORDER); - unsigned int out, in; + unsigned int out, in, desc_num, n = 0; int seg = 0; int headcount = 0; unsigned d; @@ -1064,9 +1070,9 @@ static int get_rx_bufs(struct vhost_net_virtqueue *nvq, r = -ENOBUFS; goto err; } - r = vhost_get_vq_desc(vq, vq->iov + seg, - ARRAY_SIZE(vq->iov) - seg, &out, - &in, log, log_num); + r = vhost_get_vq_desc_n(vq, vq->iov + seg, + ARRAY_SIZE(vq->iov) - seg, &out, + &in, log, log_num, &desc_num); if (unlikely(r < 0)) goto err; @@ -1093,6 +1099,7 @@ static int get_rx_bufs(struct vhost_net_virtqueue *nvq, ++headcount; datalen -= len; seg += in; + n += desc_num; } *iovcount = seg; @@ -1113,9 +1120,11 @@ static int get_rx_bufs(struct vhost_net_virtqueue *nvq, nheads[0] = headcount; } + *ndesc = n; + return headcount; err: - vhost_discard_vq_desc(vq, headcount); + vhost_discard_vq_desc(vq, headcount, n); return r; } @@ -1151,6 +1160,7 @@ static void handle_rx(struct vhost_net *net) struct iov_iter fixup; __virtio16 num_buffers; int recv_pkts = 0; + unsigned int ndesc; mutex_lock_nested(&vq->mutex, VHOST_NET_VQ_RX); sock = vhost_vq_get_backend(vq); @@ -1182,7 +1192,8 @@ static void handle_rx(struct vhost_net *net) headcount = get_rx_bufs(nvq, vq->heads + count, vq->nheads + count, vhost_len, &in, vq_log, &log, - likely(mergeable) ? UIO_MAXIOV : 1); + likely(mergeable) ? UIO_MAXIOV : 1, + &ndesc); /* On error, stop handling until the next kick. */ if (unlikely(headcount < 0)) goto out; @@ -1228,7 +1239,7 @@ static void handle_rx(struct vhost_net *net) if (unlikely(err != sock_len)) { pr_debug("Discarded rx packet: " " len %d, expected %zd\n", err, sock_len); - vhost_discard_vq_desc(vq, headcount); + vhost_discard_vq_desc(vq, headcount, ndesc); continue; } /* Supply virtio_net_hdr if VHOST_NET_F_VIRTIO_NET_HDR */ @@ -1252,7 +1263,7 @@ static void handle_rx(struct vhost_net *net) copy_to_iter(&num_buffers, sizeof num_buffers, &fixup) != sizeof num_buffers) { vq_err(vq, "Failed num_buffers write"); - vhost_discard_vq_desc(vq, headcount); + vhost_discard_vq_desc(vq, headcount, ndesc); goto out; } nvq->done_idx += headcount; diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c index 8570fdf2e14a..a78226b37739 100644 --- a/drivers/vhost/vhost.c +++ b/drivers/vhost/vhost.c @@ -2792,18 +2792,34 @@ static int get_indirect(struct vhost_virtqueue *vq, return 0; } -/* This looks in the virtqueue and for the first available buffer, and converts - * it to an iovec for convenient access. Since descriptors consist of some - * number of output then some number of input descriptors, it's actually two - * iovecs, but we pack them into one and note how many of each there were. +/** + * vhost_get_vq_desc_n - Fetch the next available descriptor chain and build iovecs + * @vq: target virtqueue + * @iov: array that receives the scatter/gather segments + * @iov_size: capacity of @iov in elements + * @out_num: the number of output segments + * @in_num: the number of input segments + * @log: optional array to record addr/len for each writable segment; NULL if unused + * @log_num: optional output; number of entries written to @log when provided + * @ndesc: optional output; number of descriptors consumed from the available ring + * (useful for rollback via vhost_discard_vq_desc) * - * This function returns the descriptor number found, or vq->num (which is - * never a valid descriptor number) if none was found. A negative code is - * returned on error. */ -int vhost_get_vq_desc(struct vhost_virtqueue *vq, - struct iovec iov[], unsigned int iov_size, - unsigned int *out_num, unsigned int *in_num, - struct vhost_log *log, unsigned int *log_num) + * Extracts one available descriptor chain from @vq and translates guest addresses + * into host iovecs. + * + * On success, advances @vq->last_avail_idx by 1 and @vq->next_avail_head by the + * number of descriptors consumed (also stored via @ndesc when non-NULL). + * + * Return: + * - head index in [0, @vq->num) on success; + * - @vq->num if no descriptor is currently available; + * - negative errno on failure + */ +int vhost_get_vq_desc_n(struct vhost_virtqueue *vq, + struct iovec iov[], unsigned int iov_size, + unsigned int *out_num, unsigned int *in_num, + struct vhost_log *log, unsigned int *log_num, + unsigned int *ndesc) { bool in_order = vhost_has_feature(vq, VIRTIO_F_IN_ORDER); struct vring_desc desc; @@ -2921,17 +2937,49 @@ int vhost_get_vq_desc(struct vhost_virtqueue *vq, vq->last_avail_idx++; vq->next_avail_head += c; + if (ndesc) + *ndesc = c; + /* Assume notifications from guest are disabled at this point, * if they aren't we would need to update avail_event index. */ BUG_ON(!(vq->used_flags & VRING_USED_F_NO_NOTIFY)); return head; } +EXPORT_SYMBOL_GPL(vhost_get_vq_desc_n); + +/* This looks in the virtqueue and for the first available buffer, and converts + * it to an iovec for convenient access. Since descriptors consist of some + * number of output then some number of input descriptors, it's actually two + * iovecs, but we pack them into one and note how many of each there were. + * + * This function returns the descriptor number found, or vq->num (which is + * never a valid descriptor number) if none was found. A negative code is + * returned on error. + */ +int vhost_get_vq_desc(struct vhost_virtqueue *vq, + struct iovec iov[], unsigned int iov_size, + unsigned int *out_num, unsigned int *in_num, + struct vhost_log *log, unsigned int *log_num) +{ + return vhost_get_vq_desc_n(vq, iov, iov_size, out_num, in_num, + log, log_num, NULL); +} EXPORT_SYMBOL_GPL(vhost_get_vq_desc); -/* Reverse the effect of vhost_get_vq_desc. Useful for error handling. */ -void vhost_discard_vq_desc(struct vhost_virtqueue *vq, int n) +/** + * vhost_discard_vq_desc - Reverse the effect of vhost_get_vq_desc_n() + * @vq: target virtqueue + * @nbufs: number of buffers to roll back + * @ndesc: number of descriptors to roll back + * + * Rewinds the internal consumer cursors after a failed attempt to use buffers + * returned by vhost_get_vq_desc_n(). + */ +void vhost_discard_vq_desc(struct vhost_virtqueue *vq, int nbufs, + unsigned int ndesc) { - vq->last_avail_idx -= n; + vq->next_avail_head -= ndesc; + vq->last_avail_idx -= nbufs; } EXPORT_SYMBOL_GPL(vhost_discard_vq_desc); diff --git a/drivers/vhost/vhost.h b/drivers/vhost/vhost.h index 621a6d9a8791..b49f08e4a1b4 100644 --- a/drivers/vhost/vhost.h +++ b/drivers/vhost/vhost.h @@ -230,7 +230,15 @@ int vhost_get_vq_desc(struct vhost_virtqueue *, struct iovec iov[], unsigned int iov_size, unsigned int *out_num, unsigned int *in_num, struct vhost_log *log, unsigned int *log_num); -void vhost_discard_vq_desc(struct vhost_virtqueue *, int n); + +int vhost_get_vq_desc_n(struct vhost_virtqueue *vq, + struct iovec iov[], unsigned int iov_size, + unsigned int *out_num, unsigned int *in_num, + struct vhost_log *log, unsigned int *log_num, + unsigned int *ndesc); + +void vhost_discard_vq_desc(struct vhost_virtqueue *, int nbuf, + unsigned int ndesc); bool vhost_vq_work_queue(struct vhost_virtqueue *vq, struct vhost_work *work); bool vhost_vq_has_work(struct vhost_virtqueue *vq); From 4fcb8ab4a09b1855dbfd7062605dd13abd64c086 Mon Sep 17 00:00:00 2001 From: Slark Xiao Date: Tue, 25 Nov 2025 15:09:00 +0800 Subject: [PATCH 34/47] net: wwan: mhi: Keep modem name match with Foxconn T99W640 Correct it since M.2 device T99W640 has updated from T99W515. We need to align it with MHI side otherwise this modem can't get the network. Fixes: ae5a34264354 ("bus: mhi: host: pci_generic: Fix the modem name of Foxconn T99W640") Signed-off-by: Slark Xiao Reviewed-by: Loic Poulain Link: https://patch.msgid.link/20251125070900.33324-1-slark_xiao@163.com Signed-off-by: Jakub Kicinski --- drivers/net/wwan/mhi_wwan_mbim.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/wwan/mhi_wwan_mbim.c b/drivers/net/wwan/mhi_wwan_mbim.c index c814fbd756a1..f8bc9a39bfa3 100644 --- a/drivers/net/wwan/mhi_wwan_mbim.c +++ b/drivers/net/wwan/mhi_wwan_mbim.c @@ -98,7 +98,7 @@ static struct mhi_mbim_link *mhi_mbim_get_link_rcu(struct mhi_mbim_context *mbim static int mhi_mbim_get_link_mux_id(struct mhi_controller *cntrl) { if (strcmp(cntrl->name, "foxconn-dw5934e") == 0 || - strcmp(cntrl->name, "foxconn-t99w515") == 0) + strcmp(cntrl->name, "foxconn-t99w640") == 0) return WDS_BIND_MUX_DATA_PORT_MUX_ID; return 0; From b9ba6338bc6e0e97dee30fa705f40950ed556729 Mon Sep 17 00:00:00 2001 From: Alexandra Winter Date: Tue, 25 Nov 2025 09:58:29 +0100 Subject: [PATCH 35/47] s390/net: list Aswin Karuvally as maintainer Thank you Aswin for taking this responsibility. Signed-off-by: Alexandra Winter Acked-by: Aswin Karuvally Link: https://patch.msgid.link/20251125085829.3679506-1-wintera@linux.ibm.com Signed-off-by: Jakub Kicinski --- MAINTAINERS | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index 83fd8480d0d6..a8a13dda94fc 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -22654,7 +22654,7 @@ F: arch/s390/mm S390 NETWORK DRIVERS M: Alexandra Winter -R: Aswin Karuvally +M: Aswin Karuvally L: linux-s390@vger.kernel.org L: netdev@vger.kernel.org S: Supported From da62abaaa268357b1aa66b372ace562189a05df1 Mon Sep 17 00:00:00 2001 From: Vladimir Oltean Date: Sat, 22 Nov 2025 13:13:24 +0200 Subject: [PATCH 36/47] net: dsa: sja1105: fix SGMII linking at 10M or 100M but not passing traffic When using the SGMII PCS as a fixed-link chip-to-chip connection, it is easy to miss the fact that traffic passes only at 1G, since that's what any normal such connection would use. When using the SGMII PCS connected towards an on-board PHY or an SFP module, it is immediately noticeable that when the link resolves to a speed other than 1G, traffic from the MAC fails to pass: TX counters increase, but nothing gets decoded by the other end, and no local RX counters increase either. Artificially lowering a fixed-link rate to speed = <100> makes us able to see the same issue as in the case of having an SGMII PHY. Some debugging shows that the XPCS configuration is A-OK, but that the MAC Configuration Table entry for the port has the SPEED bits still set to 1000Mbps, due to a special condition in the driver. Deleting that condition, and letting the resolved link speed be programmed directly into the MAC speed field, results in a functional link at all 3 speeds. This piece of evidence, based on testing on both generations with SGMII support (SJA1105S and SJA1110A) directly contradicts the statement from the blamed commit that "the MAC is fixed at 1 Gbps and we need to configure the PCS only (if even that)". Worse, that statement is not backed by any documentation, and no one from NXP knows what it might refer to. I am unable to recall sufficient context regarding my testing from March 2020 to understand what led me to draw such a braindead and factually incorrect conclusion. Yet, there is nothing of value regarding forcing the MAC speed, either for SGMII or 2500Base-X (introduced at a later stage), so remove all such logic. Fixes: ffe10e679cec ("net: dsa: sja1105: Add support for the SGMII port") Signed-off-by: Vladimir Oltean Link: https://patch.msgid.link/20251122111324.136761-1-vladimir.oltean@nxp.com Signed-off-by: Jakub Kicinski --- drivers/net/dsa/sja1105/sja1105_main.c | 7 ------- 1 file changed, 7 deletions(-) diff --git a/drivers/net/dsa/sja1105/sja1105_main.c b/drivers/net/dsa/sja1105/sja1105_main.c index f674c400f05b..aa2145cf29a6 100644 --- a/drivers/net/dsa/sja1105/sja1105_main.c +++ b/drivers/net/dsa/sja1105/sja1105_main.c @@ -1302,14 +1302,7 @@ static int sja1105_set_port_speed(struct sja1105_private *priv, int port, * table, since this will be used for the clocking setup, and we no * longer need to store it in the static config (already told hardware * we want auto during upload phase). - * Actually for the SGMII port, the MAC is fixed at 1 Gbps and - * we need to configure the PCS only (if even that). */ - if (priv->phy_mode[port] == PHY_INTERFACE_MODE_SGMII) - speed = priv->info->port_speed[SJA1105_SPEED_1000MBPS]; - else if (priv->phy_mode[port] == PHY_INTERFACE_MODE_2500BASEX) - speed = priv->info->port_speed[SJA1105_SPEED_2500MBPS]; - mac[port].speed = speed; return 0; From 27fd02860164bfa78cec2640dfad630d832e302c Mon Sep 17 00:00:00 2001 From: Paolo Abeni Date: Tue, 25 Nov 2025 17:59:11 +0100 Subject: [PATCH 37/47] mptcp: clear scheduled subflows on retransmit When __mptcp_retrans() kicks-in, it schedules one or more subflows for retransmission, but such subflows could be actually left alone if there is no more data to retransmit and/or in case of concurrent fallback. Scheduled subflows could be processed much later in time, i.e. when new data will be transmitted, leading to bad subflow selection. Explicitly clear all scheduled subflows before leaving the retransmission function. Fixes: ee2708aedad0 ("mptcp: use get_retrans wrapper") Cc: stable@vger.kernel.org Reported-by: Filip Pokryvka Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts (NGI0) Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20251125-net-mptcp-clear-sched-rtx-v1-1-1cea4ad2165f@kernel.org Signed-off-by: Jakub Kicinski --- net/mptcp/protocol.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 33b5dce431c2..8abb425d8b5f 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -2665,7 +2665,7 @@ static void __mptcp_retrans(struct sock *sk) } if (!mptcp_send_head(sk)) - return; + goto clear_scheduled; goto reset_timer; } @@ -2696,7 +2696,7 @@ static void __mptcp_retrans(struct sock *sk) if (__mptcp_check_fallback(msk)) { spin_unlock_bh(&msk->fallback_lock); release_sock(ssk); - return; + goto clear_scheduled; } while (info.sent < info.limit) { @@ -2728,6 +2728,15 @@ reset_timer: if (!mptcp_rtx_timer_pending(sk)) mptcp_reset_rtx_timer(sk); + +clear_scheduled: + /* If no rtx data was available or in case of fallback, there + * could be left-over scheduled subflows; clear them all + * or later xmit could use bad ones + */ + mptcp_for_each_subflow(msk, subflow) + if (READ_ONCE(subflow->scheduled)) + mptcp_subflow_set_scheduled(subflow, false); } /* schedule the timeout timer for the relevant event: either close timeout From 6d66e093e0740d39a36ef742c60eec247df26f41 Mon Sep 17 00:00:00 2001 From: Mohsin Bashir Date: Tue, 25 Nov 2025 13:17:04 -0800 Subject: [PATCH 38/47] eth: fbnic: Fix counter roll-over issue Fix a potential counter roll-over issue in fbnic_mbx_alloc_rx_msgs() when calculating descriptor slots. The issue occurs when head - tail results in a large positive value (unsigned) and the compiler interprets head - tail - 1 as a signed value. Since FBNIC_IPC_MBX_DESC_LEN is a power of two, use a masking operation, which is a common way of avoiding this problem when dealing with these sort of ring space calculations. Fixes: da3cde08209e ("eth: fbnic: Add FW communication mechanism") Signed-off-by: Mohsin Bashir Link: https://patch.msgid.link/20251125211704.3222413-1-mohsin.bashr@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/meta/fbnic/fbnic_fw.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/meta/fbnic/fbnic_fw.c b/drivers/net/ethernet/meta/fbnic/fbnic_fw.c index c87cb9ed09e7..fcd9912e7ad3 100644 --- a/drivers/net/ethernet/meta/fbnic/fbnic_fw.c +++ b/drivers/net/ethernet/meta/fbnic/fbnic_fw.c @@ -201,7 +201,7 @@ static int fbnic_mbx_alloc_rx_msgs(struct fbnic_dev *fbd) return -ENODEV; /* Fill all but 1 unused descriptors in the Rx queue. */ - count = (head - tail - 1) % FBNIC_IPC_MBX_DESC_LEN; + count = (head - tail - 1) & (FBNIC_IPC_MBX_DESC_LEN - 1); while (!err && count--) { struct fbnic_tlv_msg *msg; From 1cd1c472343b06d6d32038636ce51bfa2251e3cf Mon Sep 17 00:00:00 2001 From: Jon Kohler Date: Tue, 25 Nov 2025 15:27:53 -0700 Subject: [PATCH 39/47] virtio-net: avoid unnecessary checksum calculation on guest RX Commit a2fb4bc4e2a6 ("net: implement virtio helpers to handle UDP GSO tunneling.") inadvertently altered checksum offload behavior for guests not using UDP GSO tunneling. Before, tun_put_user called tun_vnet_hdr_from_skb, which passed has_data_valid = true to virtio_net_hdr_from_skb. After, tun_put_user began calling tun_vnet_hdr_tnl_from_skb instead, which passes has_data_valid = false into both call sites. This caused virtio hdr flags to not include VIRTIO_NET_HDR_F_DATA_VALID for SKBs where skb->ip_summed == CHECKSUM_UNNECESSARY. As a result, guests are forced to recalculate checksums unnecessarily. Restore the previous behavior by ensuring has_data_valid = true is passed in the !tnl_gso_type case, but only from tun side, as virtio_net_hdr_tnl_from_skb() is used also by the virtio_net driver, which in turn must not use VIRTIO_NET_HDR_F_DATA_VALID on tx. cc: stable@vger.kernel.org Fixes: a2fb4bc4e2a6 ("net: implement virtio helpers to handle UDP GSO tunneling.") Signed-off-by: Jon Kohler Acked-by: Michael S. Tsirkin Acked-by: Jason Wang Link: https://patch.msgid.link/20251125222754.1737443-1-jon@nutanix.com Signed-off-by: Jakub Kicinski --- drivers/net/tun_vnet.h | 2 +- drivers/net/virtio_net.c | 3 ++- include/linux/virtio_net.h | 7 ++++--- 3 files changed, 7 insertions(+), 5 deletions(-) diff --git a/drivers/net/tun_vnet.h b/drivers/net/tun_vnet.h index 81662328b2c7..a5f93b6c4482 100644 --- a/drivers/net/tun_vnet.h +++ b/drivers/net/tun_vnet.h @@ -244,7 +244,7 @@ tun_vnet_hdr_tnl_from_skb(unsigned int flags, if (virtio_net_hdr_tnl_from_skb(skb, tnl_hdr, has_tnl_offload, tun_vnet_is_little_endian(flags), - vlan_hlen)) { + vlan_hlen, true)) { struct virtio_net_hdr_v1 *hdr = &tnl_hdr->hash_hdr.hdr; struct skb_shared_info *sinfo = skb_shinfo(skb); diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index 0369dda5ed60..8e04adb57f52 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -3339,7 +3339,8 @@ static int xmit_skb(struct send_queue *sq, struct sk_buff *skb, bool orphan) hdr = &skb_vnet_common_hdr(skb)->tnl_hdr; if (virtio_net_hdr_tnl_from_skb(skb, hdr, vi->tx_tnl, - virtio_is_little_endian(vi->vdev), 0)) + virtio_is_little_endian(vi->vdev), 0, + false)) return -EPROTO; if (vi->mergeable_rx_bufs) diff --git a/include/linux/virtio_net.h b/include/linux/virtio_net.h index b673c31569f3..75dabb763c65 100644 --- a/include/linux/virtio_net.h +++ b/include/linux/virtio_net.h @@ -384,7 +384,8 @@ virtio_net_hdr_tnl_from_skb(const struct sk_buff *skb, struct virtio_net_hdr_v1_hash_tunnel *vhdr, bool tnl_hdr_negotiated, bool little_endian, - int vlan_hlen) + int vlan_hlen, + bool has_data_valid) { struct virtio_net_hdr *hdr = (struct virtio_net_hdr *)vhdr; unsigned int inner_nh, outer_th; @@ -394,8 +395,8 @@ virtio_net_hdr_tnl_from_skb(const struct sk_buff *skb, tnl_gso_type = skb_shinfo(skb)->gso_type & (SKB_GSO_UDP_TUNNEL | SKB_GSO_UDP_TUNNEL_CSUM); if (!tnl_gso_type) - return virtio_net_hdr_from_skb(skb, hdr, little_endian, false, - vlan_hlen); + return virtio_net_hdr_from_skb(skb, hdr, little_endian, + has_data_valid, vlan_hlen); /* Tunnel support not negotiated but skb ask for it. */ if (!tnl_hdr_negotiated) From 384c1a4e2722b43f26b55cc041562ff8b229b536 Mon Sep 17 00:00:00 2001 From: Jon Kohler Date: Tue, 25 Nov 2025 18:57:48 -0700 Subject: [PATCH 40/47] MAINTAINERS: separate VIRTIO NET DRIVER and add netdev Changes to virtio network stack should be cc'd to netdev DL, separate it into its own group to add netdev in addition to virtualization DL. Signed-off-by: Jon Kohler Acked-by: Michael S. Tsirkin Link: https://patch.msgid.link/20251126015750.2200267-1-jon@nutanix.com Signed-off-by: Jakub Kicinski --- MAINTAINERS | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index a8a13dda94fc..8a05e461ee62 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -27121,7 +27121,7 @@ S: Maintained F: drivers/char/virtio_console.c F: include/uapi/linux/virtio_console.h -VIRTIO CORE AND NET DRIVERS +VIRTIO CORE M: "Michael S. Tsirkin" M: Jason Wang R: Xuan Zhuo @@ -27134,7 +27134,6 @@ F: Documentation/devicetree/bindings/virtio/ F: Documentation/driver-api/virtio/ F: drivers/block/virtio_blk.c F: drivers/crypto/virtio/ -F: drivers/net/virtio_net.c F: drivers/vdpa/ F: drivers/virtio/ F: include/linux/vdpa.h @@ -27143,7 +27142,6 @@ F: include/linux/vringh.h F: include/uapi/linux/virtio_*.h F: net/vmw_vsock/virtio* F: tools/virtio/ -F: tools/testing/selftests/drivers/net/virtio_net/ VIRTIO CRYPTO DRIVER M: Gonglei @@ -27255,6 +27253,19 @@ W: https://virtio-mem.gitlab.io/ F: drivers/virtio/virtio_mem.c F: include/uapi/linux/virtio_mem.h +VIRTIO NET DRIVER +M: "Michael S. Tsirkin" +M: Jason Wang +R: Xuan Zhuo +R: Eugenio Pérez +L: netdev@vger.kernel.org +L: virtualization@lists.linux.dev +S: Maintained +F: drivers/net/virtio_net.c +F: include/linux/virtio_net.h +F: include/uapi/linux/virtio_net.h +F: tools/testing/selftests/drivers/net/virtio_net/ + VIRTIO PMEM DRIVER M: Pankaj Gupta L: virtualization@lists.linux.dev From 5ffcb7b890f61541201461580bb6622ace405aec Mon Sep 17 00:00:00 2001 From: Jiefeng Zhang Date: Wed, 26 Nov 2025 11:22:49 +0800 Subject: [PATCH 41/47] net: atlantic: fix fragment overflow handling in RX path The atlantic driver can receive packets with more than MAX_SKB_FRAGS (17) fragments when handling large multi-descriptor packets. This causes an out-of-bounds write in skb_add_rx_frag_netmem() leading to kernel panic. The issue occurs because the driver doesn't check the total number of fragments before calling skb_add_rx_frag(). When a packet requires more than MAX_SKB_FRAGS fragments, the fragment index exceeds the array bounds. Fix by assuming there will be an extra frag if buff->len > AQ_CFG_RX_HDR_SIZE, then all fragments are accounted for. And reusing the existing check to prevent the overflow earlier in the code path. This crash occurred in production with an Aquantia AQC113 10G NIC. Stack trace from production environment: ``` RIP: 0010:skb_add_rx_frag_netmem+0x29/0xd0 Code: 90 f3 0f 1e fa 0f 1f 44 00 00 48 89 f8 41 89 ca 48 89 d7 48 63 ce 8b 90 c0 00 00 00 48 c1 e1 04 48 01 ca 48 03 90 c8 00 00 00 <48> 89 7a 30 44 89 52 3c 44 89 42 38 40 f6 c7 01 75 74 48 89 fa 83 RSP: 0018:ffffa9bec02a8d50 EFLAGS: 00010287 RAX: ffff925b22e80a00 RBX: ffff925ad38d2700 RCX: fffffffe0a0c8000 RDX: ffff9258ea95bac0 RSI: ffff925ae0a0c800 RDI: 0000000000037a40 RBP: 0000000000000024 R08: 0000000000000000 R09: 0000000000000021 R10: 0000000000000848 R11: 0000000000000000 R12: ffffa9bec02a8e24 R13: ffff925ad8615570 R14: 0000000000000000 R15: ffff925b22e80a00 FS: 0000000000000000(0000) GS:ffff925e47880000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffff9258ea95baf0 CR3: 0000000166022004 CR4: 0000000000f72ef0 PKRU: 55555554 Call Trace: aq_ring_rx_clean+0x175/0xe60 [atlantic] ? aq_ring_rx_clean+0x14d/0xe60 [atlantic] ? aq_ring_tx_clean+0xdf/0x190 [atlantic] ? kmem_cache_free+0x348/0x450 ? aq_vec_poll+0x81/0x1d0 [atlantic] ? __napi_poll+0x28/0x1c0 ? net_rx_action+0x337/0x420 ``` Fixes: 6aecbba12b5c ("net: atlantic: add check for MAX_SKB_FRAGS") Changes in v4: - Add Fixes: tag to satisfy patch validation requirements. Changes in v3: - Fix by assuming there will be an extra frag if buff->len > AQ_CFG_RX_HDR_SIZE, then all fragments are accounted for. Signed-off-by: Jiefeng Zhang Link: https://patch.msgid.link/20251126032249.69358-1-jiefeng.z.zhang@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/aquantia/atlantic/aq_ring.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/net/ethernet/aquantia/atlantic/aq_ring.c b/drivers/net/ethernet/aquantia/atlantic/aq_ring.c index f21de0c21e52..d23d23bed39f 100644 --- a/drivers/net/ethernet/aquantia/atlantic/aq_ring.c +++ b/drivers/net/ethernet/aquantia/atlantic/aq_ring.c @@ -547,6 +547,11 @@ static int __aq_ring_rx_clean(struct aq_ring_s *self, struct napi_struct *napi, if (!buff->is_eop) { unsigned int frag_cnt = 0U; + + /* There will be an extra fragment */ + if (buff->len > AQ_CFG_RX_HDR_SIZE) + frag_cnt++; + buff_ = buff; do { bool is_rsc_completed = true; From b3e528a5811bbc8246dbdb962f0812dc9b721681 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Tue, 25 Nov 2025 14:48:54 +0800 Subject: [PATCH 42/47] net: mctp: unconditionally set skb->dev on dst output On transmit, we are currently relying on skb->dev being set by mctp_local_output() when we first set up the skb destination fields. However, forwarded skbs do not use the local_output path, so will retain their incoming netdev as their ->dev on tx. This does not work when we're forwarding between interfaces. Set skb->dev unconditionally in the transmit path, to allow for proper forwarding. We keep the skb->dev initialisation in mctp_local_output(), as we use it for fragmentation. Fixes: 269936db5eb3 ("net: mctp: separate routing database from routing operations") Suggested-by: Vince Chang Signed-off-by: Jeremy Kerr Link: https://patch.msgid.link/20251125-dev-forward-v1-1-54ecffcd0616@codeconstruct.com.au Signed-off-by: Paolo Abeni --- net/mctp/route.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/mctp/route.c b/net/mctp/route.c index 4d314e062ba9..2ac4011a953f 100644 --- a/net/mctp/route.c +++ b/net/mctp/route.c @@ -623,6 +623,7 @@ static int mctp_dst_output(struct mctp_dst *dst, struct sk_buff *skb) skb->protocol = htons(ETH_P_MCTP); skb->pkt_type = PACKET_OUTGOING; + skb->dev = dst->dev->dev; if (skb->len > dst->mtu) { kfree_skb(skb); From 50caa744689e505414673c20359b04aa918439e3 Mon Sep 17 00:00:00 2001 From: Wei Fang Date: Tue, 25 Nov 2025 16:52:07 +0800 Subject: [PATCH 43/47] net: fec: cancel perout_timer when PEROUT is disabled The PEROUT allows the user to set a specified future time to output the periodic signal. If the future time is far from the current time, the FEC driver will use hrtimer to configure PEROUT one second before the future time. However, the hrtimer will not be canceled if the PEROUT is disabled before the hrtimer expires. So the PEROUT will be configured when the hrtimer expires, which is not as expected. Therefore, cancel the hrtimer in fec_ptp_pps_disable() to fix this issue. Fixes: 350749b909bf ("net: fec: Add support for periodic output signal of PPS") Signed-off-by: Wei Fang Link: https://patch.msgid.link/20251125085210.1094306-2-wei.fang@nxp.com Signed-off-by: Paolo Abeni --- drivers/net/ethernet/freescale/fec_ptp.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/ethernet/freescale/fec_ptp.c b/drivers/net/ethernet/freescale/fec_ptp.c index fa88b47d526c..7a5367ea9410 100644 --- a/drivers/net/ethernet/freescale/fec_ptp.c +++ b/drivers/net/ethernet/freescale/fec_ptp.c @@ -497,6 +497,8 @@ static int fec_ptp_pps_disable(struct fec_enet_private *fep, uint channel) { unsigned long flags; + hrtimer_cancel(&fep->perout_timer); + spin_lock_irqsave(&fep->tmreg_lock, flags); writel(0, fep->hwp + FEC_TCSR(channel)); spin_unlock_irqrestore(&fep->tmreg_lock, flags); From e97faa0c20ea8840f45569ba434e30538fff8fc9 Mon Sep 17 00:00:00 2001 From: Wei Fang Date: Tue, 25 Nov 2025 16:52:08 +0800 Subject: [PATCH 44/47] net: fec: do not update PEROUT if it is enabled If the previously set PEROUT is already active, updating it will cause the new PEROUT to start immediately instead of at the specified time. This is because fep->reload_period is updated whithout check whether the PEROUT is enabled, and the old PEROUT is not disabled. Therefore, the pulse period will be updated immediately in the pulse interrupt handler fec_pps_interrupt(). Currently, the driver does not support directly updating PEROUT and it will make the logic be more complicated. To fix the current issue, add a check before enabling the PEROUT, the driver will return an error if PEROUT is enabled. If users wants to update a new PEROUT, they should disable the old PEROUT first. Fixes: 350749b909bf ("net: fec: Add support for periodic output signal of PPS") Signed-off-by: Wei Fang Link: https://patch.msgid.link/20251125085210.1094306-3-wei.fang@nxp.com Signed-off-by: Paolo Abeni --- drivers/net/ethernet/freescale/fec.h | 1 + drivers/net/ethernet/freescale/fec_ptp.c | 43 ++++++++++++++++++------ 2 files changed, 34 insertions(+), 10 deletions(-) diff --git a/drivers/net/ethernet/freescale/fec.h b/drivers/net/ethernet/freescale/fec.h index 41e0d85d15da..abf1ef8e76c6 100644 --- a/drivers/net/ethernet/freescale/fec.h +++ b/drivers/net/ethernet/freescale/fec.h @@ -687,6 +687,7 @@ struct fec_enet_private { unsigned int reload_period; int pps_enable; unsigned int next_counter; + bool perout_enable; struct hrtimer perout_timer; u64 perout_stime; diff --git a/drivers/net/ethernet/freescale/fec_ptp.c b/drivers/net/ethernet/freescale/fec_ptp.c index 7a5367ea9410..f31b1626c12f 100644 --- a/drivers/net/ethernet/freescale/fec_ptp.c +++ b/drivers/net/ethernet/freescale/fec_ptp.c @@ -243,6 +243,7 @@ static int fec_ptp_pps_perout(struct fec_enet_private *fep) * the FEC_TCCR register in time and missed the start time. */ if (fep->perout_stime < curr_time + 100 * NSEC_PER_MSEC) { + fep->perout_enable = false; dev_err(&fep->pdev->dev, "Current time is too close to the start time!\n"); spin_unlock_irqrestore(&fep->tmreg_lock, flags); return -1; @@ -500,6 +501,7 @@ static int fec_ptp_pps_disable(struct fec_enet_private *fep, uint channel) hrtimer_cancel(&fep->perout_timer); spin_lock_irqsave(&fep->tmreg_lock, flags); + fep->perout_enable = false; writel(0, fep->hwp + FEC_TCSR(channel)); spin_unlock_irqrestore(&fep->tmreg_lock, flags); @@ -531,6 +533,8 @@ static int fec_ptp_enable(struct ptp_clock_info *ptp, return ret; } else if (rq->type == PTP_CLK_REQ_PEROUT) { + u32 reload_period; + /* Reject requests with unsupported flags */ if (rq->perout.flags) return -EOPNOTSUPP; @@ -550,12 +554,14 @@ static int fec_ptp_enable(struct ptp_clock_info *ptp, return -EOPNOTSUPP; } - fep->reload_period = div_u64(period_ns, 2); - if (on && fep->reload_period) { + reload_period = div_u64(period_ns, 2); + if (on && reload_period) { + u64 perout_stime; + /* Convert 1588 timestamp to ns*/ start_time.tv_sec = rq->perout.start.sec; start_time.tv_nsec = rq->perout.start.nsec; - fep->perout_stime = timespec64_to_ns(&start_time); + perout_stime = timespec64_to_ns(&start_time); mutex_lock(&fep->ptp_clk_mutex); if (!fep->ptp_clk_on) { @@ -564,18 +570,35 @@ static int fec_ptp_enable(struct ptp_clock_info *ptp, return -EOPNOTSUPP; } spin_lock_irqsave(&fep->tmreg_lock, flags); + + if (fep->perout_enable) { + dev_err(&fep->pdev->dev, + "PEROUT has been enabled\n"); + ret = -EBUSY; + goto unlock; + } + /* Read current timestamp */ curr_time = timecounter_read(&fep->tc); + if (perout_stime <= curr_time) { + dev_err(&fep->pdev->dev, + "Start time must be greater than current time\n"); + ret = -EINVAL; + goto unlock; + } + + /* Calculate time difference */ + delta = perout_stime - curr_time; + fep->reload_period = reload_period; + fep->perout_stime = perout_stime; + fep->perout_enable = true; + +unlock: spin_unlock_irqrestore(&fep->tmreg_lock, flags); mutex_unlock(&fep->ptp_clk_mutex); - /* Calculate time difference */ - delta = fep->perout_stime - curr_time; - - if (fep->perout_stime <= curr_time) { - dev_err(&fep->pdev->dev, "Start time must larger than current time!\n"); - return -EINVAL; - } + if (ret) + return ret; /* Because the timer counter of FEC only has 31-bits, correspondingly, * the time comparison register FEC_TCCR also only low 31 bits can be From c0a1f3d7e128e8d1b6c0fe09c68eac5ebcf677c8 Mon Sep 17 00:00:00 2001 From: Wei Fang Date: Tue, 25 Nov 2025 16:52:09 +0800 Subject: [PATCH 45/47] net: fec: do not allow enabling PPS and PEROUT simultaneously In the current driver, PPS and PEROUT use the same channel to generate the events, so they cannot be enabled at the same time. Otherwise, the later configuration will overwrite the earlier configuration. Therefore, when configuring PPS, the driver will check whether PEROUT is enabled. Similarly, when configuring PEROUT, the driver will check whether PPS is enabled. Fixes: 350749b909bf ("net: fec: Add support for periodic output signal of PPS") Signed-off-by: Wei Fang Link: https://patch.msgid.link/20251125085210.1094306-4-wei.fang@nxp.com Signed-off-by: Paolo Abeni --- drivers/net/ethernet/freescale/fec_ptp.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/drivers/net/ethernet/freescale/fec_ptp.c b/drivers/net/ethernet/freescale/fec_ptp.c index f31b1626c12f..ed5d59abeb53 100644 --- a/drivers/net/ethernet/freescale/fec_ptp.c +++ b/drivers/net/ethernet/freescale/fec_ptp.c @@ -128,6 +128,12 @@ static int fec_ptp_enable_pps(struct fec_enet_private *fep, uint enable) spin_lock_irqsave(&fep->tmreg_lock, flags); + if (fep->perout_enable) { + spin_unlock_irqrestore(&fep->tmreg_lock, flags); + dev_err(&fep->pdev->dev, "PEROUT is running"); + return -EBUSY; + } + if (fep->pps_enable == enable) { spin_unlock_irqrestore(&fep->tmreg_lock, flags); return 0; @@ -571,6 +577,12 @@ static int fec_ptp_enable(struct ptp_clock_info *ptp, } spin_lock_irqsave(&fep->tmreg_lock, flags); + if (fep->pps_enable) { + dev_err(&fep->pdev->dev, "PPS is running"); + ret = -EBUSY; + goto unlock; + } + if (fep->perout_enable) { dev_err(&fep->pdev->dev, "PEROUT has been enabled\n"); From 9a060d0fac9e75524f72864adec6d8cdb70a5bca Mon Sep 17 00:00:00 2001 From: Wei Fang Date: Tue, 25 Nov 2025 16:52:10 +0800 Subject: [PATCH 46/47] net: fec: do not register PPS event for PEROUT There are currently two situations that can trigger the PTP interrupt, one is the PPS event, the other is the PEROUT event. However, the irq handler fec_pps_interrupt() does not check the irq event type and directly registers a PPS event into the system, but the event may be a PEROUT event. This is incorrect because PEROUT is an output signal, while PPS is the input of the kernel PPS system. Therefore, add a check for the event type, if pps_enable is true, it means that the current event is a PPS event, and then the PPS event is registered. Fixes: 350749b909bf ("net: fec: Add support for periodic output signal of PPS") Signed-off-by: Wei Fang Link: https://patch.msgid.link/20251125085210.1094306-5-wei.fang@nxp.com Signed-off-by: Paolo Abeni --- drivers/net/ethernet/freescale/fec_ptp.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/freescale/fec_ptp.c b/drivers/net/ethernet/freescale/fec_ptp.c index ed5d59abeb53..4b7bad9a485d 100644 --- a/drivers/net/ethernet/freescale/fec_ptp.c +++ b/drivers/net/ethernet/freescale/fec_ptp.c @@ -718,8 +718,11 @@ static irqreturn_t fec_pps_interrupt(int irq, void *dev_id) fep->next_counter = (fep->next_counter + fep->reload_period) & fep->cc.mask; - event.type = PTP_CLOCK_PPS; - ptp_clock_event(fep->ptp_clock, &event); + if (fep->pps_enable) { + event.type = PTP_CLOCK_PPS; + ptp_clock_event(fep->ptp_clock, &event); + } + return IRQ_HANDLED; } From f07f4ea53e22429c84b20832fa098b5ecc0d4e35 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Tue, 25 Nov 2025 19:53:29 +0000 Subject: [PATCH 47/47] mptcp: Initialise rcv_mss before calling tcp_send_active_reset() in mptcp_do_fastclose(). syzbot reported divide-by-zero in __tcp_select_window() by MPTCP socket. [0] We had a similar issue for the bare TCP and fixed in commit 499350a5a6e7 ("tcp: initialize rcv_mss to TCP_MIN_MSS instead of 0"). Let's apply the same fix to mptcp_do_fastclose(). [0]: Oops: divide error: 0000 [#1] SMP KASAN PTI CPU: 0 UID: 0 PID: 6068 Comm: syz.0.17 Not tainted syzkaller #0 PREEMPT(full) Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/25/2025 RIP: 0010:__tcp_select_window+0x824/0x1320 net/ipv4/tcp_output.c:3336 Code: ff ff ff 44 89 f1 d3 e0 89 c1 f7 d1 41 01 cc 41 21 c4 e9 a9 00 00 00 e8 ca 49 01 f8 e9 9c 00 00 00 e8 c0 49 01 f8 44 89 e0 99 7c 24 1c 41 29 d4 48 bb 00 00 00 00 00 fc ff df e9 80 00 00 00 RSP: 0018:ffffc90003017640 EFLAGS: 00010293 RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffff88807b469e40 RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 RBP: ffffc90003017730 R08: ffff888033268143 R09: 1ffff1100664d028 R10: dffffc0000000000 R11: ffffed100664d029 R12: 0000000000000000 R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000 FS: 000055557faa0500(0000) GS:ffff888126135000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f64a1912ff8 CR3: 0000000072122000 CR4: 00000000003526f0 Call Trace: tcp_select_window net/ipv4/tcp_output.c:281 [inline] __tcp_transmit_skb+0xbc7/0x3aa0 net/ipv4/tcp_output.c:1568 tcp_transmit_skb net/ipv4/tcp_output.c:1649 [inline] tcp_send_active_reset+0x2d1/0x5b0 net/ipv4/tcp_output.c:3836 mptcp_do_fastclose+0x27e/0x380 net/mptcp/protocol.c:2793 mptcp_disconnect+0x238/0x710 net/mptcp/protocol.c:3253 mptcp_sendmsg_fastopen+0x2f8/0x580 net/mptcp/protocol.c:1776 mptcp_sendmsg+0x1774/0x1980 net/mptcp/protocol.c:1855 sock_sendmsg_nosec net/socket.c:727 [inline] __sock_sendmsg+0xe5/0x270 net/socket.c:742 __sys_sendto+0x3bd/0x520 net/socket.c:2244 __do_sys_sendto net/socket.c:2251 [inline] __se_sys_sendto net/socket.c:2247 [inline] __x64_sys_sendto+0xde/0x100 net/socket.c:2247 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline] do_syscall_64+0xfa/0xfa0 arch/x86/entry/syscall_64.c:94 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0033:0x7f66e998f749 Code: ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 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 73 01 c3 48 c7 c1 a8 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007ffff9acedb8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c RAX: ffffffffffffffda RBX: 00007f66e9be5fa0 RCX: 00007f66e998f749 RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000003 RBP: 00007ffff9acee10 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000001 R13: 00007f66e9be5fa0 R14: 00007f66e9be5fa0 R15: 0000000000000006 Fixes: ae155060247b ("mptcp: fix duplicate reset on fastclose") Reported-by: syzbot+3a92d359bc2ec6255a33@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/69260882.a70a0220.d98e3.00b4.GAE@google.com/ Signed-off-by: Kuniyuki Iwashima Reviewed-by: Matthieu Baerts (NGI0) Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20251125195331.309558-1-kuniyu@google.com Signed-off-by: Paolo Abeni --- net/mptcp/protocol.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 8abb425d8b5f..1e413426deee 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -2798,6 +2798,12 @@ static void mptcp_do_fastclose(struct sock *sk) goto unlock; subflow->send_fastclose = 1; + + /* Initialize rcv_mss to TCP_MIN_MSS to avoid division by 0 + * issue in __tcp_select_window(), see tcp_disconnect(). + */ + inet_csk(ssk)->icsk_ack.rcv_mss = TCP_MIN_MSS; + tcp_send_active_reset(ssk, ssk->sk_allocation, SK_RST_REASON_TCP_ABORT_ON_CLOSE); unlock: