From 2a27f6a8fb5722223d526843040f747e9b0e8060 Mon Sep 17 00:00:00 2001 From: Celeste Liu Date: Tue, 30 Sep 2025 19:34:28 +0800 Subject: [PATCH 01/50] can: gs_usb: increase max interface to U8_MAX This issue was found by Runcheng Lu when develop HSCanT USB to CAN FD converter[1]. The original developers may have only 3 interfaces device to test so they write 3 here and wait for future change. During the HSCanT development, we actually used 4 interfaces, so the limitation of 3 is not enough now. But just increase one is not future-proofed. Since the channel index type in gs_host_frame is u8, just make canch[] become a flexible array with a u8 index, so it naturally constraint by U8_MAX and avoid statically allocate 256 pointer for every gs_usb device. [1]: https://github.com/cherry-embedded/HSCanT-hardware Fixes: d08e973a77d1 ("can: gs_usb: Added support for the GS_USB CAN devices") Reported-by: Runcheng Lu Cc: stable@vger.kernel.org Reviewed-by: Vincent Mailhol Signed-off-by: Celeste Liu Link: https://patch.msgid.link/20250930-gs-usb-max-if-v5-1-863330bf6666@coelacanthus.name Signed-off-by: Marc Kleine-Budde --- drivers/net/can/usb/gs_usb.c | 22 ++++++++++------------ 1 file changed, 10 insertions(+), 12 deletions(-) diff --git a/drivers/net/can/usb/gs_usb.c b/drivers/net/can/usb/gs_usb.c index c9482d6e947b..9fb4cbbd6d6d 100644 --- a/drivers/net/can/usb/gs_usb.c +++ b/drivers/net/can/usb/gs_usb.c @@ -289,11 +289,6 @@ struct gs_host_frame { #define GS_MAX_RX_URBS 30 #define GS_NAPI_WEIGHT 32 -/* Maximum number of interfaces the driver supports per device. - * Current hardware only supports 3 interfaces. The future may vary. - */ -#define GS_MAX_INTF 3 - struct gs_tx_context { struct gs_can *dev; unsigned int echo_id; @@ -324,7 +319,6 @@ struct gs_can { /* usb interface struct */ struct gs_usb { - struct gs_can *canch[GS_MAX_INTF]; struct usb_anchor rx_submitted; struct usb_device *udev; @@ -336,9 +330,11 @@ struct gs_usb { unsigned int hf_size_rx; u8 active_channels; + u8 channel_cnt; unsigned int pipe_in; unsigned int pipe_out; + struct gs_can *canch[] __counted_by(channel_cnt); }; /* 'allocate' a tx context. @@ -599,7 +595,7 @@ static void gs_usb_receive_bulk_callback(struct urb *urb) } /* device reports out of range channel id */ - if (hf->channel >= GS_MAX_INTF) + if (hf->channel >= parent->channel_cnt) goto device_detach; dev = parent->canch[hf->channel]; @@ -699,7 +695,7 @@ resubmit_urb: /* USB failure take down all interfaces */ if (rc == -ENODEV) { device_detach: - for (rc = 0; rc < GS_MAX_INTF; rc++) { + for (rc = 0; rc < parent->channel_cnt; rc++) { if (parent->canch[rc]) netif_device_detach(parent->canch[rc]->netdev); } @@ -1460,17 +1456,19 @@ static int gs_usb_probe(struct usb_interface *intf, icount = dconf.icount + 1; dev_info(&intf->dev, "Configuring for %u interfaces\n", icount); - if (icount > GS_MAX_INTF) { + if (icount > type_max(parent->channel_cnt)) { dev_err(&intf->dev, "Driver cannot handle more that %u CAN interfaces\n", - GS_MAX_INTF); + type_max(parent->channel_cnt)); return -EINVAL; } - parent = kzalloc(sizeof(*parent), GFP_KERNEL); + parent = kzalloc(struct_size(parent, canch, icount), GFP_KERNEL); if (!parent) return -ENOMEM; + parent->channel_cnt = icount; + init_usb_anchor(&parent->rx_submitted); usb_set_intfdata(intf, parent); @@ -1531,7 +1529,7 @@ static void gs_usb_disconnect(struct usb_interface *intf) return; } - for (i = 0; i < GS_MAX_INTF; i++) + for (i = 0; i < parent->channel_cnt; i++) if (parent->canch[i]) gs_destroy_candev(parent->canch[i]); From a12f0bc764da3781da2019c60826f47a6d7ed64f Mon Sep 17 00:00:00 2001 From: Celeste Liu Date: Tue, 30 Sep 2025 14:53:39 +0800 Subject: [PATCH 02/50] can: gs_usb: gs_make_candev(): populate net_device->dev_port The gs_usb driver supports USB devices with more than 1 CAN channel. In old kernel before 3.15, it uses net_device->dev_id to distinguish different channel in userspace, which was done in commit acff76fa45b4 ("can: gs_usb: gs_make_candev(): set netdev->dev_id"). But since 3.15, the correct way is populating net_device->dev_port. And according to documentation, if network device support multiple interface, lack of net_device->dev_port SHALL be treated as a bug. Fixes: acff76fa45b4 ("can: gs_usb: gs_make_candev(): set netdev->dev_id") Cc: stable@vger.kernel.org Signed-off-by: Celeste Liu Link: https://patch.msgid.link/20250930-gs-usb-populate-net_device-dev_port-v1-1-68a065de6937@coelacanthus.name Signed-off-by: Marc Kleine-Budde --- drivers/net/can/usb/gs_usb.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/can/usb/gs_usb.c b/drivers/net/can/usb/gs_usb.c index 9fb4cbbd6d6d..69b8d6da651b 100644 --- a/drivers/net/can/usb/gs_usb.c +++ b/drivers/net/can/usb/gs_usb.c @@ -1245,6 +1245,7 @@ static struct gs_can *gs_make_candev(unsigned int channel, netdev->flags |= IFF_ECHO; /* we support full roundtrip echo */ netdev->dev_id = channel; + netdev->dev_port = channel; /* dev setup */ strcpy(dev->bt_const.name, KBUILD_MODNAME); From ba569fb07a7e9e9b71e9282e27e993ba859295c2 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Wed, 6 Aug 2025 17:46:32 +0200 Subject: [PATCH 03/50] can: m_can: m_can_plat_remove(): add missing pm_runtime_disable() Commit 227619c3ff7c ("can: m_can: move runtime PM enable/disable to m_can_platform") moved the PM runtime enable from the m_can core driver into the m_can_platform. That patch forgot to move the pm_runtime_disable() to m_can_plat_remove(), so that unloading the m_can_platform driver causes an "Unbalanced pm_runtime_enable!" error message. Add the missing pm_runtime_disable() to m_can_plat_remove() to fix the problem. Cc: Patrik Flykt Fixes: 227619c3ff7c ("can: m_can: move runtime PM enable/disable to m_can_platform") Reviewed-by: Markus Schneider-Pargmann Link: https://patch.msgid.link/20250929-m_can-fix-state-handling-v4-1-682b49b49d9a@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/m_can/m_can_platform.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/can/m_can/m_can_platform.c b/drivers/net/can/m_can/m_can_platform.c index b832566efda0..057eaa7b8b4b 100644 --- a/drivers/net/can/m_can/m_can_platform.c +++ b/drivers/net/can/m_can/m_can_platform.c @@ -180,7 +180,7 @@ static void m_can_plat_remove(struct platform_device *pdev) struct m_can_classdev *mcan_class = &priv->cdev; m_can_class_unregister(mcan_class); - + pm_runtime_disable(mcan_class->dev); m_can_class_free_dev(mcan_class->net); } From 3d9db29b45f970d81acf61cf91a65442efbeb997 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Wed, 6 Aug 2025 16:56:15 +0200 Subject: [PATCH 04/50] can: m_can: m_can_handle_state_errors(): fix CAN state transition to Error Active The CAN Error State is determined by the receive and transmit error counters. The CAN error counters decrease when reception/transmission is successful, so that a status transition back to the Error Active status is possible. This transition is not handled by m_can_handle_state_errors(). Add the missing detection of the Error Active state to m_can_handle_state_errors() and extend the handling of this state in m_can_handle_state_change(). Fixes: e0d1f4816f2a ("can: m_can: add Bosch M_CAN controller support") Fixes: cd0d83eab2e0 ("can: m_can: m_can_handle_state_change(): fix state change") Reviewed-by: Markus Schneider-Pargmann Link: https://patch.msgid.link/20250929-m_can-fix-state-handling-v4-2-682b49b49d9a@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/m_can/m_can.c | 55 +++++++++++++++++++++-------------- 1 file changed, 33 insertions(+), 22 deletions(-) diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c index e1d725979685..ac864183a536 100644 --- a/drivers/net/can/m_can/m_can.c +++ b/drivers/net/can/m_can/m_can.c @@ -812,6 +812,9 @@ static int m_can_handle_state_change(struct net_device *dev, u32 timestamp = 0; switch (new_state) { + case CAN_STATE_ERROR_ACTIVE: + cdev->can.state = CAN_STATE_ERROR_ACTIVE; + break; case CAN_STATE_ERROR_WARNING: /* error warning state */ cdev->can.can_stats.error_warning++; @@ -841,6 +844,12 @@ static int m_can_handle_state_change(struct net_device *dev, __m_can_get_berr_counter(dev, &bec); switch (new_state) { + case CAN_STATE_ERROR_ACTIVE: + cf->can_id |= CAN_ERR_CRTL | CAN_ERR_CNT; + cf->data[1] = CAN_ERR_CRTL_ACTIVE; + cf->data[6] = bec.txerr; + cf->data[7] = bec.rxerr; + break; case CAN_STATE_ERROR_WARNING: /* error warning state */ cf->can_id |= CAN_ERR_CRTL | CAN_ERR_CNT; @@ -877,30 +886,33 @@ static int m_can_handle_state_change(struct net_device *dev, return 1; } -static int m_can_handle_state_errors(struct net_device *dev, u32 psr) +static enum can_state +m_can_state_get_by_psr(struct m_can_classdev *cdev) +{ + u32 reg_psr; + + reg_psr = m_can_read(cdev, M_CAN_PSR); + + if (reg_psr & PSR_BO) + return CAN_STATE_BUS_OFF; + if (reg_psr & PSR_EP) + return CAN_STATE_ERROR_PASSIVE; + if (reg_psr & PSR_EW) + return CAN_STATE_ERROR_WARNING; + + return CAN_STATE_ERROR_ACTIVE; +} + +static int m_can_handle_state_errors(struct net_device *dev) { struct m_can_classdev *cdev = netdev_priv(dev); - int work_done = 0; + enum can_state new_state; - if (psr & PSR_EW && cdev->can.state != CAN_STATE_ERROR_WARNING) { - netdev_dbg(dev, "entered error warning state\n"); - work_done += m_can_handle_state_change(dev, - CAN_STATE_ERROR_WARNING); - } + new_state = m_can_state_get_by_psr(cdev); + if (new_state == cdev->can.state) + return 0; - if (psr & PSR_EP && cdev->can.state != CAN_STATE_ERROR_PASSIVE) { - netdev_dbg(dev, "entered error passive state\n"); - work_done += m_can_handle_state_change(dev, - CAN_STATE_ERROR_PASSIVE); - } - - if (psr & PSR_BO && cdev->can.state != CAN_STATE_BUS_OFF) { - netdev_dbg(dev, "entered error bus off state\n"); - work_done += m_can_handle_state_change(dev, - CAN_STATE_BUS_OFF); - } - - return work_done; + return m_can_handle_state_change(dev, new_state); } static void m_can_handle_other_err(struct net_device *dev, u32 irqstatus) @@ -1031,8 +1043,7 @@ static int m_can_rx_handler(struct net_device *dev, int quota, u32 irqstatus) } if (irqstatus & IR_ERR_STATE) - work_done += m_can_handle_state_errors(dev, - m_can_read(cdev, M_CAN_PSR)); + work_done += m_can_handle_state_errors(dev); if (irqstatus & IR_ERR_BUS_30X) work_done += m_can_handle_bus_errors(dev, irqstatus, From 4942c42fe1849e6d68dfb5b36ccba344a9fac016 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Wed, 6 Aug 2025 18:24:12 +0200 Subject: [PATCH 05/50] can: m_can: m_can_chip_config(): bring up interface in correct state In some SoCs (observed on the STM32MP15) the M_CAN IP core keeps the CAN state and CAN error counters over an internal reset cycle. An external reset is not always possible, due to the shared reset with the other CAN core. This caused the core not always be in Error Active state when bringing up the controller. Instead of always setting the CAN state to Error Active in m_can_chip_config(), fix this by reading and decoding the Protocol Status Regitser (PSR) and set the CAN state accordingly. Fixes: e0d1f4816f2a ("can: m_can: add Bosch M_CAN controller support") Reviewed-by: Markus Schneider-Pargmann Link: https://patch.msgid.link/20250929-m_can-fix-state-handling-v4-3-682b49b49d9a@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/m_can/m_can.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c index ac864183a536..b6db5b57241c 100644 --- a/drivers/net/can/m_can/m_can.c +++ b/drivers/net/can/m_can/m_can.c @@ -1617,7 +1617,7 @@ static int m_can_start(struct net_device *dev) netdev_queue_set_dql_min_limit(netdev_get_tx_queue(cdev->net, 0), cdev->tx_max_coalesced_frames); - cdev->can.state = CAN_STATE_ERROR_ACTIVE; + cdev->can.state = m_can_state_get_by_psr(cdev); m_can_enable_all_interrupts(cdev); From a9e30a22d6f23a2684c248871cad4c3061181639 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Tue, 12 Aug 2025 16:58:31 +0200 Subject: [PATCH 06/50] can: m_can: fix CAN state in system PM A suspend/resume cycle on a down interface results in the interface coming up in Error Active state. A suspend/resume cycle on an Up interface will always result in Error Active state, regardless of the actual CAN state. During suspend, only set running interfaces to CAN_STATE_SLEEPING. During resume only touch the CAN state of running interfaces. For wakeup sources, set the CAN state depending on the Protocol Status Regitser (PSR), for non wakeup source interfaces m_can_start() will do the same. Fixes: e0d1f4816f2a ("can: m_can: add Bosch M_CAN controller support") Reviewed-by: Markus Schneider-Pargmann Link: https://patch.msgid.link/20250929-m_can-fix-state-handling-v4-4-682b49b49d9a@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/m_can/m_can.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c index b6db5b57241c..f2576e577058 100644 --- a/drivers/net/can/m_can/m_can.c +++ b/drivers/net/can/m_can/m_can.c @@ -2503,12 +2503,11 @@ int m_can_class_suspend(struct device *dev) } m_can_clk_stop(cdev); + cdev->can.state = CAN_STATE_SLEEPING; } pinctrl_pm_select_sleep_state(dev); - cdev->can.state = CAN_STATE_SLEEPING; - return ret; } EXPORT_SYMBOL_GPL(m_can_class_suspend); @@ -2521,8 +2520,6 @@ int m_can_class_resume(struct device *dev) pinctrl_pm_select_default_state(dev); - cdev->can.state = CAN_STATE_ERROR_ACTIVE; - if (netif_running(ndev)) { ret = m_can_clk_start(cdev); if (ret) @@ -2540,6 +2537,8 @@ int m_can_class_resume(struct device *dev) if (cdev->ops->init) ret = cdev->ops->init(cdev); + cdev->can.state = m_can_state_get_by_psr(cdev); + m_can_write(cdev, M_CAN_IE, cdev->active_interrupts); } else { ret = m_can_start(ndev); From 49836ff2f37dd6d52bfe3153c0bcbd96025a6100 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Thu, 9 Oct 2025 08:25:36 +0200 Subject: [PATCH 07/50] can: m_can: replace Dong Aisheng's old email address Dong Aisheng's old Freescale email is not valid anymore and bounces, replace it by the new NXP one. Reviewed-by: Dong Aisheng Link: https://patch.msgid.link/20251009-m_can-update-email-address-v1-1-30a268587f69@pengutronix.de Signed-off-by: Marc Kleine-Budde --- .mailmap | 1 + drivers/net/can/m_can/m_can.c | 4 ++-- drivers/net/can/m_can/m_can_platform.c | 4 ++-- 3 files changed, 5 insertions(+), 4 deletions(-) diff --git a/.mailmap b/.mailmap index d30f9363a4c9..8160c62f11e9 100644 --- a/.mailmap +++ b/.mailmap @@ -227,6 +227,7 @@ Dmitry Safonov <0x7f454c46@gmail.com> Dmitry Safonov <0x7f454c46@gmail.com> Dmitry Safonov <0x7f454c46@gmail.com> Domen Puncer +Dong Aisheng Douglas Gilbert Drew Fustini diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c index f2576e577058..ad4f577c1ef7 100644 --- a/drivers/net/can/m_can/m_can.c +++ b/drivers/net/can/m_can/m_can.c @@ -1,7 +1,7 @@ // SPDX-License-Identifier: GPL-2.0 // CAN bus driver for Bosch M_CAN controller // Copyright (C) 2014 Freescale Semiconductor, Inc. -// Dong Aisheng +// Dong Aisheng // Copyright (C) 2018-19 Texas Instruments Incorporated - http://www.ti.com/ /* Bosch M_CAN user manual can be obtained from: @@ -2556,7 +2556,7 @@ int m_can_class_resume(struct device *dev) } EXPORT_SYMBOL_GPL(m_can_class_resume); -MODULE_AUTHOR("Dong Aisheng "); +MODULE_AUTHOR("Dong Aisheng "); MODULE_AUTHOR("Dan Murphy "); MODULE_LICENSE("GPL v2"); MODULE_DESCRIPTION("CAN bus driver for Bosch M_CAN controller"); diff --git a/drivers/net/can/m_can/m_can_platform.c b/drivers/net/can/m_can/m_can_platform.c index 057eaa7b8b4b..4a412add2b8d 100644 --- a/drivers/net/can/m_can/m_can_platform.c +++ b/drivers/net/can/m_can/m_can_platform.c @@ -1,7 +1,7 @@ // SPDX-License-Identifier: GPL-2.0 // IOMapped CAN bus driver for Bosch M_CAN controller // Copyright (C) 2014 Freescale Semiconductor, Inc. -// Dong Aisheng +// Dong Aisheng // // Copyright (C) 2018-19 Texas Instruments Incorporated - http://www.ti.com/ @@ -236,7 +236,7 @@ static struct platform_driver m_can_plat_driver = { module_platform_driver(m_can_plat_driver); -MODULE_AUTHOR("Dong Aisheng "); +MODULE_AUTHOR("Dong Aisheng "); MODULE_AUTHOR("Dan Murphy "); MODULE_LICENSE("GPL v2"); MODULE_DESCRIPTION("M_CAN driver for IO Mapped Bosch controllers"); From f3426ac54c42c3260096ddc50b5470eb179fb06a Mon Sep 17 00:00:00 2001 From: Ivan Vecera Date: Wed, 8 Oct 2025 16:14:18 +0200 Subject: [PATCH 08/50] dpll: zl3073x: Increase maximum size of flash utility Newer firmware bundles contain a flash utility whose size exceeds the currently allowed limit. Increase the maximum allowed size to accommodate the newer utility version. Without this patch: # devlink dev flash i2c/1-0070 file fw_nosplit_v3.hex Failed to load firmware Flashing failed Error: zl3073x: FW load failed: [utility] component is too big (11000 bytes) Fixes: ca017409da694 ("dpll: zl3073x: Add firmware loading functionality") Suggested-by: Prathosh Satish Signed-off-by: Ivan Vecera Reviewed-by: Vadim Fedorenko Link: https://patch.msgid.link/20251008141418.841053-1-ivecera@redhat.com Signed-off-by: Paolo Abeni --- drivers/dpll/zl3073x/fw.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/dpll/zl3073x/fw.c b/drivers/dpll/zl3073x/fw.c index d5418ff74886..def37fe8d9b0 100644 --- a/drivers/dpll/zl3073x/fw.c +++ b/drivers/dpll/zl3073x/fw.c @@ -37,7 +37,7 @@ struct zl3073x_fw_component_info { static const struct zl3073x_fw_component_info component_info[] = { [ZL_FW_COMPONENT_UTIL] = { .name = "utility", - .max_size = 0x2300, + .max_size = 0x4000, .load_addr = 0x20000000, .flash_type = ZL3073X_FLASH_TYPE_NONE, }, From d5d790ba1558dbb8d179054f514476e2ee970b8e Mon Sep 17 00:00:00 2001 From: Bhanu Seshu Kumar Valluri Date: Thu, 9 Oct 2025 11:00:09 +0530 Subject: [PATCH 09/50] net: usb: lan78xx: Fix lost EEPROM write timeout error(-ETIMEDOUT) in lan78xx_write_raw_eeprom The function lan78xx_write_raw_eeprom failed to properly propagate EEPROM write timeout errors (-ETIMEDOUT). In the timeout fallthrough path, it first attempted to restore the pin configuration for LED outputs and then returned only the status of that restore operation, discarding the original timeout error saved in ret. As a result, callers could mistakenly treat EEPROM write operation as successful even though the EEPROM write had actually timed out with no or partial data write. To fix this, handle errors in restoring the LED pin configuration separately. If the restore succeeds, return any prior EEPROM write timeout error saved in ret to the caller. Suggested-by: Oleksij Rempel Fixes: 8b1b2ca83b20 ("net: usb: lan78xx: Improve error handling in EEPROM and OTP operations") cc: stable@vger.kernel.org Signed-off-by: Bhanu Seshu Kumar Valluri Reviewed-by: Simon Horman Signed-off-by: David S. Miller --- drivers/net/usb/lan78xx.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) diff --git a/drivers/net/usb/lan78xx.c b/drivers/net/usb/lan78xx.c index 42d35cc6b421..28195d9a8d6b 100644 --- a/drivers/net/usb/lan78xx.c +++ b/drivers/net/usb/lan78xx.c @@ -1175,10 +1175,13 @@ static int lan78xx_write_raw_eeprom(struct lan78xx_net *dev, u32 offset, } write_raw_eeprom_done: - if (dev->chipid == ID_REV_CHIP_ID_7800_) - return lan78xx_write_reg(dev, HW_CFG, saved); - - return 0; + if (dev->chipid == ID_REV_CHIP_ID_7800_) { + int rc = lan78xx_write_reg(dev, HW_CFG, saved); + /* If USB fails, there is nothing to do */ + if (rc < 0) + return rc; + } + return ret; } static int lan78xx_read_raw_otp(struct lan78xx_net *dev, u32 offset, From 3abc0e55ea1fa2250e52bc860e8f24b2b9a2093a Mon Sep 17 00:00:00 2001 From: Rex Lu Date: Thu, 9 Oct 2025 08:29:34 +0200 Subject: [PATCH 10/50] net: mtk: wed: add dma mask limitation and GFP_DMA32 for device with more than 4GB DRAM Limit tx/rx buffer address to 32-bit address space for board with more than 4GB DRAM. Fixes: 804775dfc2885 ("net: ethernet: mtk_eth_soc: add support for Wireless Ethernet Dispatch (WED)") Fixes: 6757d345dd7db ("net: ethernet: mtk_wed: introduce hw_rro support for MT7988") Tested-by: Daniel Pawlik Tested-by: Matteo Croce Signed-off-by: Rex Lu Co-developed-by: Lorenzo Bianconi Signed-off-by: Lorenzo Bianconi Reviewed-by: Simon Horman Signed-off-by: David S. Miller --- drivers/net/ethernet/mediatek/mtk_wed.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mediatek/mtk_wed.c b/drivers/net/ethernet/mediatek/mtk_wed.c index 3dbb113b792c..1ed1f88dd7f8 100644 --- a/drivers/net/ethernet/mediatek/mtk_wed.c +++ b/drivers/net/ethernet/mediatek/mtk_wed.c @@ -677,7 +677,7 @@ mtk_wed_tx_buffer_alloc(struct mtk_wed_device *dev) void *buf; int s; - page = __dev_alloc_page(GFP_KERNEL); + page = __dev_alloc_page(GFP_KERNEL | GFP_DMA32); if (!page) return -ENOMEM; @@ -800,7 +800,7 @@ mtk_wed_hwrro_buffer_alloc(struct mtk_wed_device *dev) struct page *page; int s; - page = __dev_alloc_page(GFP_KERNEL); + page = __dev_alloc_page(GFP_KERNEL | GFP_DMA32); if (!page) return -ENOMEM; @@ -2426,6 +2426,10 @@ mtk_wed_attach(struct mtk_wed_device *dev) dev->version = hw->version; dev->hw->pcie_base = mtk_wed_get_pcie_base(dev); + ret = dma_set_mask_and_coherent(hw->dev, DMA_BIT_MASK(32)); + if (ret) + goto out; + if (hw->eth->dma_dev == hw->eth->dev && of_dma_is_coherent(hw->eth->dev->of_node)) mtk_eth_set_dma_device(hw->eth, hw->dev); From 65946eac6d888d50ae527c4e5c237dbe5cc3a2f2 Mon Sep 17 00:00:00 2001 From: Yeounsu Moon Date: Fri, 10 Oct 2025 00:57:16 +0900 Subject: [PATCH 11/50] net: dlink: handle dma_map_single() failure properly There is no error handling for `dma_map_single()` failures. Add error handling by checking `dma_mapping_error()` and freeing the `skb` using `dev_kfree_skb()` (process context) when it fails. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Yeounsu Moon Tested-on: D-Link DGE-550T Rev-A3 Suggested-by: Simon Horman Reviewed-by: Simon Horman Signed-off-by: David S. Miller --- drivers/net/ethernet/dlink/dl2k.c | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-) diff --git a/drivers/net/ethernet/dlink/dl2k.c b/drivers/net/ethernet/dlink/dl2k.c index 1996d2e4e3e2..7077d705e471 100644 --- a/drivers/net/ethernet/dlink/dl2k.c +++ b/drivers/net/ethernet/dlink/dl2k.c @@ -508,25 +508,34 @@ static int alloc_list(struct net_device *dev) for (i = 0; i < RX_RING_SIZE; i++) { /* Allocated fixed size of skbuff */ struct sk_buff *skb; + dma_addr_t addr; skb = netdev_alloc_skb_ip_align(dev, np->rx_buf_sz); np->rx_skbuff[i] = skb; - if (!skb) { - free_list(dev); - return -ENOMEM; - } + if (!skb) + goto err_free_list; + + addr = dma_map_single(&np->pdev->dev, skb->data, + np->rx_buf_sz, DMA_FROM_DEVICE); + if (dma_mapping_error(&np->pdev->dev, addr)) + goto err_kfree_skb; np->rx_ring[i].next_desc = cpu_to_le64(np->rx_ring_dma + ((i + 1) % RX_RING_SIZE) * sizeof(struct netdev_desc)); /* Rubicon now supports 40 bits of addressing space. */ - np->rx_ring[i].fraginfo = - cpu_to_le64(dma_map_single(&np->pdev->dev, skb->data, - np->rx_buf_sz, DMA_FROM_DEVICE)); + np->rx_ring[i].fraginfo = cpu_to_le64(addr); np->rx_ring[i].fraginfo |= cpu_to_le64((u64)np->rx_buf_sz << 48); } return 0; + +err_kfree_skb: + dev_kfree_skb(np->rx_skbuff[i]); + np->rx_skbuff[i] = NULL; +err_free_list: + free_list(dev); + return -ENOMEM; } static void rio_hw_init(struct net_device *dev) From 68a052239fc4b351e961f698b824f7654a346091 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Thu, 9 Oct 2025 13:56:29 -0700 Subject: [PATCH 12/50] selftests: drv-net: update remaining Python init files Convert remaining __init__ files similar to what we did in commit b615879dbfea ("selftests: drv-net: make linters happy with our imports") Signed-off-by: Jakub Kicinski Signed-off-by: Stanislav Fomichev Reviewed-by: Simon Horman Signed-off-by: David S. Miller --- .../drivers/net/hw/lib/py/__init__.py | 40 ++++++++++++++----- .../selftests/drivers/net/lib/py/__init__.py | 4 +- .../testing/selftests/net/lib/py/__init__.py | 29 ++++++++++++-- 3 files changed, 57 insertions(+), 16 deletions(-) diff --git a/tools/testing/selftests/drivers/net/hw/lib/py/__init__.py b/tools/testing/selftests/drivers/net/hw/lib/py/__init__.py index 0ceb297e7757..fb010a48a5a1 100644 --- a/tools/testing/selftests/drivers/net/hw/lib/py/__init__.py +++ b/tools/testing/selftests/drivers/net/hw/lib/py/__init__.py @@ -1,5 +1,13 @@ # SPDX-License-Identifier: GPL-2.0 +""" +Driver test environment (hardware-only tests). +NetDrvEnv and NetDrvEpEnv are the main environment classes. +Former is for local host only tests, latter creates / connects +to a remote endpoint. See NIPA wiki for more information about +running and writing driver tests. +""" + import sys from pathlib import Path @@ -8,26 +16,36 @@ KSFT_DIR = (Path(__file__).parent / "../../../../..").resolve() try: sys.path.append(KSFT_DIR.as_posix()) - from net.lib.py import * - from drivers.net.lib.py import * - # Import one by one to avoid pylint false positives + from net.lib.py import NetNS, NetNSEnter, NetdevSimDev from net.lib.py import EthtoolFamily, NetdevFamily, NetshaperFamily, \ NlError, RtnlFamily, DevlinkFamily, PSPFamily from net.lib.py import CmdExitFailure - from net.lib.py import bkg, cmd, defer, ethtool, fd_read_timeout, ip, \ - rand_port, tool, wait_port_listen - from net.lib.py import fd_read_timeout + from net.lib.py import bkg, cmd, bpftool, bpftrace, defer, ethtool, \ + fd_read_timeout, ip, rand_port, wait_port_listen, wait_file from net.lib.py import KsftSkipEx, KsftFailEx, KsftXfailEx from net.lib.py import ksft_disruptive, ksft_exit, ksft_pr, ksft_run, \ ksft_setup from net.lib.py import ksft_eq, ksft_ge, ksft_in, ksft_is, ksft_lt, \ ksft_ne, ksft_not_in, ksft_raises, ksft_true, ksft_gt, ksft_not_none - from net.lib.py import NetNSEnter - from drivers.net.lib.py import GenerateTraffic + from drivers.net.lib.py import GenerateTraffic, Remote from drivers.net.lib.py import NetDrvEnv, NetDrvEpEnv + + __all__ = ["NetNS", "NetNSEnter", "NetdevSimDev", + "EthtoolFamily", "NetdevFamily", "NetshaperFamily", + "NlError", "RtnlFamily", "DevlinkFamily", "PSPFamily", + "CmdExitFailure", + "bkg", "cmd", "bpftool", "bpftrace", "defer", "ethtool", + "fd_read_timeout", "ip", "rand_port", + "wait_port_listen", "wait_file", + "KsftSkipEx", "KsftFailEx", "KsftXfailEx", + "ksft_disruptive", "ksft_exit", "ksft_pr", "ksft_run", + "ksft_setup", + "ksft_eq", "ksft_ge", "ksft_in", "ksft_is", "ksft_lt", + "ksft_ne", "ksft_not_in", "ksft_raises", "ksft_true", "ksft_gt", + "ksft_not_none", "ksft_not_none", + "NetDrvEnv", "NetDrvEpEnv", "GenerateTraffic", "Remote"] except ModuleNotFoundError as e: - ksft_pr("Failed importing `net` library from kernel sources") - ksft_pr(str(e)) - ktap_result(True, comment="SKIP") + print("Failed importing `net` library from kernel sources") + print(str(e)) sys.exit(4) diff --git a/tools/testing/selftests/drivers/net/lib/py/__init__.py b/tools/testing/selftests/drivers/net/lib/py/__init__.py index e6c070f32f51..b0c6300150fb 100644 --- a/tools/testing/selftests/drivers/net/lib/py/__init__.py +++ b/tools/testing/selftests/drivers/net/lib/py/__init__.py @@ -22,7 +22,7 @@ try: NlError, RtnlFamily, DevlinkFamily, PSPFamily from net.lib.py import CmdExitFailure from net.lib.py import bkg, cmd, bpftool, bpftrace, defer, ethtool, \ - fd_read_timeout, ip, rand_port, tool, wait_port_listen, wait_file + fd_read_timeout, ip, rand_port, wait_port_listen, wait_file from net.lib.py import KsftSkipEx, KsftFailEx, KsftXfailEx from net.lib.py import ksft_disruptive, ksft_exit, ksft_pr, ksft_run, \ ksft_setup @@ -34,7 +34,7 @@ try: "NlError", "RtnlFamily", "DevlinkFamily", "PSPFamily", "CmdExitFailure", "bkg", "cmd", "bpftool", "bpftrace", "defer", "ethtool", - "fd_read_timeout", "ip", "rand_port", "tool", + "fd_read_timeout", "ip", "rand_port", "wait_port_listen", "wait_file", "KsftSkipEx", "KsftFailEx", "KsftXfailEx", "ksft_disruptive", "ksft_exit", "ksft_pr", "ksft_run", diff --git a/tools/testing/selftests/net/lib/py/__init__.py b/tools/testing/selftests/net/lib/py/__init__.py index 997b85cc216a..97b7cf2b20eb 100644 --- a/tools/testing/selftests/net/lib/py/__init__.py +++ b/tools/testing/selftests/net/lib/py/__init__.py @@ -1,9 +1,32 @@ # SPDX-License-Identifier: GPL-2.0 +""" +Python selftest helpers for netdev. +""" + from .consts import KSRC -from .ksft import * +from .ksft import KsftFailEx, KsftSkipEx, KsftXfailEx, ksft_pr, ksft_eq, \ + ksft_ne, ksft_true, ksft_not_none, ksft_in, ksft_not_in, ksft_is, \ + ksft_ge, ksft_gt, ksft_lt, ksft_raises, ksft_busy_wait, \ + ktap_result, ksft_disruptive, ksft_setup, ksft_run, ksft_exit from .netns import NetNS, NetNSEnter -from .nsim import * -from .utils import * +from .nsim import NetdevSim, NetdevSimDev +from .utils import CmdExitFailure, fd_read_timeout, cmd, bkg, defer, \ + bpftool, ip, ethtool, bpftrace, rand_port, wait_port_listen, wait_file from .ynl import NlError, YnlFamily, EthtoolFamily, NetdevFamily, RtnlFamily, RtnlAddrFamily from .ynl import NetshaperFamily, DevlinkFamily, PSPFamily + +__all__ = ["KSRC", + "KsftFailEx", "KsftSkipEx", "KsftXfailEx", "ksft_pr", "ksft_eq", + "ksft_ne", "ksft_true", "ksft_not_none", "ksft_in", "ksft_not_in", + "ksft_is", "ksft_ge", "ksft_gt", "ksft_lt", "ksft_raises", + "ksft_busy_wait", "ktap_result", "ksft_disruptive", "ksft_setup", + "ksft_run", "ksft_exit", + "NetNS", "NetNSEnter", + "CmdExitFailure", "fd_read_timeout", "cmd", "bkg", "defer", + "bpftool", "ip", "ethtool", "bpftrace", "rand_port", + "wait_port_listen", "wait_file", + "NetdevSim", "NetdevSimDev", + "NetshaperFamily", "DevlinkFamily", "PSPFamily", "NlError", + "YnlFamily", "EthtoolFamily", "NetdevFamily", "RtnlFamily", + "RtnlAddrFamily"] From 0b4b77eff5f8cd9be062783a1c1e198d46d0a753 Mon Sep 17 00:00:00 2001 From: Nicolas Dichtel Date: Fri, 10 Oct 2025 16:18:59 +0200 Subject: [PATCH 13/50] doc: fix seg6_flowlabel path This sysctl is not per interface; it's global per netns. Fixes: 292ecd9f5a94 ("doc: move seg6_flowlabel to seg6-sysctl.rst") Reported-by: Philippe Guibert Signed-off-by: Nicolas Dichtel Reviewed-by: Simon Horman Signed-off-by: David S. Miller --- Documentation/networking/seg6-sysctl.rst | 3 +++ 1 file changed, 3 insertions(+) diff --git a/Documentation/networking/seg6-sysctl.rst b/Documentation/networking/seg6-sysctl.rst index 07c20e470baf..1b6af4779be1 100644 --- a/Documentation/networking/seg6-sysctl.rst +++ b/Documentation/networking/seg6-sysctl.rst @@ -25,6 +25,9 @@ seg6_require_hmac - INTEGER Default is 0. +/proc/sys/net/ipv6/seg6_* variables: +==================================== + seg6_flowlabel - INTEGER Controls the behaviour of computing the flowlabel of outer IPv6 header in case of SR T.encaps From c282993ccd97ad627d213645dc485086de034647 Mon Sep 17 00:00:00 2001 From: Vincent Mailhol Date: Mon, 13 Oct 2025 19:10:22 +0900 Subject: [PATCH 14/50] can: remove false statement about 1:1 mapping between DLC and length The CAN-FD section of can.rst still states that there is a 1:1 mapping between the Classical CAN DLC and its length. This is only true for the DLC values up to 8. Beyond that point, the length remains at 8. For reference, the mapping between the CAN DLC and the length is given in below table [1]: DLC value CBFF and CEFF FBFF and FEFF [decimal] [byte] [byte] ---------------------------------------------- 0 0 0 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6 7 7 7 8 8 8 9 8 12 10 8 16 11 8 20 12 8 24 13 8 32 14 8 48 15 8 64 Remove the erroneous statement. Instead just state that the length of a Classical CAN frame ranges from 0 to 8. [1] ISO 11898-1:2024, Table 5 -- DLC: coding of the four LSB Signed-off-by: Vincent Mailhol Link: https://patch.msgid.link/20251013-can-fd-doc-v2-1-5d53bdc8f2ad@kernel.org Signed-off-by: Marc Kleine-Budde --- Documentation/networking/can.rst | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/Documentation/networking/can.rst b/Documentation/networking/can.rst index 7650c4b5be5f..ccd321d29a8a 100644 --- a/Documentation/networking/can.rst +++ b/Documentation/networking/can.rst @@ -1398,10 +1398,9 @@ second bit timing has to be specified in order to enable the CAN FD bitrate. Additionally CAN FD capable CAN controllers support up to 64 bytes of payload. The representation of this length in can_frame.len and canfd_frame.len for userspace applications and inside the Linux network -layer is a plain value from 0 .. 64 instead of the CAN 'data length code'. -The data length code was a 1:1 mapping to the payload length in the Classical -CAN frames anyway. The payload length to the bus-relevant DLC mapping is -only performed inside the CAN drivers, preferably with the helper +layer is a plain value from 0 .. 64 instead of the Classical CAN length +which ranges from 0 to 8. The payload length to the bus-relevant DLC mapping +is only performed inside the CAN drivers, preferably with the helper functions can_fd_dlc2len() and can_fd_len2dlc(). The CAN netdevice driver capabilities can be distinguished by the network From b5746b3e8ea4a8a4df776e0864322028d4f5e4b1 Mon Sep 17 00:00:00 2001 From: Vincent Mailhol Date: Mon, 13 Oct 2025 19:10:23 +0900 Subject: [PATCH 15/50] can: add Transmitter Delay Compensation (TDC) documentation Back in 2021, support for CAN TDC was added to the kernel in series [1] and in iproute2 in series [2]. However, the documentation was never updated. Add a new sub-section under CAN-FD driver support to document how to configure the TDC using the "ip tool". [1] add the netlink interface for CAN-FD Transmitter Delay Compensation (TDC) Link: https://lore.kernel.org/all/20210918095637.20108-1-mailhol.vincent@wanadoo.fr/ [2] iplink_can: cleaning, fixes and adding TDC support Link: https://lore.kernel.org/all/20211103164428.692722-1-mailhol.vincent@wanadoo.fr/ Signed-off-by: Vincent Mailhol Link: https://patch.msgid.link/20251013-can-fd-doc-v2-2-5d53bdc8f2ad@kernel.org Signed-off-by: Marc Kleine-Budde --- Documentation/networking/can.rst | 64 ++++++++++++++++++++++++++++++++ 1 file changed, 64 insertions(+) diff --git a/Documentation/networking/can.rst b/Documentation/networking/can.rst index ccd321d29a8a..194e305ae973 100644 --- a/Documentation/networking/can.rst +++ b/Documentation/networking/can.rst @@ -1464,6 +1464,70 @@ Example when 'fd-non-iso on' is added on this switchable CAN FD adapter:: can state ERROR-ACTIVE (berr-counter tx 0 rx 0) restart-ms 0 +Transmitter Delay Compensation +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +At high bit rates, the propagation delay from the TX pin to the RX pin of +the transceiver might become greater than the actual bit time causing +measurement errors: the RX pin would still be measuring the previous bit. + +The Transmitter Delay Compensation (thereafter, TDC) resolves this problem +by introducing a Secondary Sample Point (SSP) equal to the distance, in +minimum time quantum, from the start of the bit time on the TX pin to the +actual measurement on the RX pin. The SSP is calculated as the sum of two +configurable values: the TDC Value (TDCV) and the TDC offset (TDCO). + +TDC, if supported by the device, can be configured together with CAN-FD +using the ip tool's "tdc-mode" argument as follow: + +**omitted** + When no "tdc-mode" option is provided, the kernel will automatically + decide whether TDC should be turned on, in which case it will + calculate a default TDCO and use the TDCV as measured by the + device. This is the recommended method to use TDC. + +**"tdc-mode off"** + TDC is explicitly disabled. + +**"tdc-mode auto"** + The user must provide the "tdco" argument. The TDCV will be + automatically calculated by the device. This option is only + available if the device supports the TDC-AUTO CAN controller mode. + +**"tdc-mode manual"** + The user must provide both the "tdco" and "tdcv" arguments. This + option is only available if the device supports the TDC-MANUAL CAN + controller mode. + +Note that some devices may offer an additional parameter: "tdcf" (TDC Filter +window). If supported by your device, this can be added as an optional +argument to either "tdc-mode auto" or "tdc-mode manual". + +Example configuring a 500 kbit/s arbitration bitrate, a 5 Mbit/s data +bitrate, a TDCO of 15 minimum time quantum and a TDCV automatically measured +by the device:: + + $ ip link set can0 up type can bitrate 500000 \ + fd on dbitrate 4000000 \ + tdc-mode auto tdco 15 + $ ip -details link show can0 + 5: can0: mtu 72 qdisc pfifo_fast state UP \ + mode DEFAULT group default qlen 10 + link/can promiscuity 0 allmulti 0 minmtu 72 maxmtu 72 + can state ERROR-ACTIVE restart-ms 0 + bitrate 500000 sample-point 0.875 + tq 12 prop-seg 69 phase-seg1 70 phase-seg2 20 sjw 10 brp 1 + ES582.1/ES584.1: tseg1 2..256 tseg2 2..128 sjw 1..128 brp 1..512 \ + brp_inc 1 + dbitrate 4000000 dsample-point 0.750 + dtq 12 dprop-seg 7 dphase-seg1 7 dphase-seg2 5 dsjw 2 dbrp 1 + tdco 15 tdcf 0 + ES582.1/ES584.1: dtseg1 2..32 dtseg2 1..16 dsjw 1..8 dbrp 1..32 \ + dbrp_inc 1 + tdco 0..127 tdcf 0..127 + clock 80000000 + + Supported CAN Hardware ---------------------- From 93a27b5891b8194a8c083c9a80d2141d4bf47ba8 Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Sat, 27 Sep 2025 21:11:16 +0900 Subject: [PATCH 16/50] can: j1939: add missing calls in NETDEV_UNREGISTER notification handler Currently NETDEV_UNREGISTER event handler is not calling j1939_cancel_active_session() and j1939_sk_queue_drop_all(). This will result in these calls being skipped when j1939_sk_release() is called. And I guess that the reason syzbot is still reporting unregister_netdevice: waiting for vcan0 to become free. Usage count = 2 is caused by lack of these calls. Calling j1939_cancel_active_session(priv, sk) from j1939_sk_release() can be covered by calling j1939_cancel_active_session(priv, NULL) from j1939_netdev_notify(). Calling j1939_sk_queue_drop_all() from j1939_sk_release() can be covered by calling j1939_sk_netdev_event_netdown() from j1939_netdev_notify(). Therefore, we can reuse j1939_cancel_active_session(priv, NULL) and j1939_sk_netdev_event_netdown(priv) for NETDEV_UNREGISTER event handler. Fixes: 7fcbe5b2c6a4 ("can: j1939: implement NETDEV_UNREGISTER notification handler") Signed-off-by: Tetsuo Handa Tested-by: Oleksij Rempel Acked-by: Oleksij Rempel Link: https://patch.msgid.link/3ad3c7f8-5a74-4b07-a193-cb0725823558@I-love.SAKURA.ne.jp Signed-off-by: Marc Kleine-Budde --- net/can/j1939/main.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/can/j1939/main.c b/net/can/j1939/main.c index 3706a872ecaf..a93af55df5fd 100644 --- a/net/can/j1939/main.c +++ b/net/can/j1939/main.c @@ -378,6 +378,8 @@ static int j1939_netdev_notify(struct notifier_block *nb, j1939_ecu_unmap_all(priv); break; case NETDEV_UNREGISTER: + j1939_cancel_active_session(priv, NULL); + j1939_sk_netdev_event_netdown(priv); j1939_sk_netdev_event_unregister(priv); break; } From fcb8b32a68fd40b0440cb9468cf6f6ab9de9f3c5 Mon Sep 17 00:00:00 2001 From: Ivan Vecera Date: Wed, 8 Oct 2025 16:14:45 +0200 Subject: [PATCH 17/50] dpll: zl3073x: Handle missing or corrupted flash configuration If the internal flash contains missing or corrupted configuration, basic communication over the bus still functions, but the device is not capable of normal operation (for example, using mailboxes). This condition is indicated in the info register by the ready bit. If this bit is cleared, the probe procedure times out while fetching the device state. Handle this case by checking the ready bit value in zl3073x_dev_start() and skipping DPLL device and pin registration if it is cleared. Do not report this condition as an error, allowing the devlink device to be registered and enabling the user to flash the correct configuration. Prior this patch: [ 31.112299] zl3073x-i2c 1-0070: Failed to fetch input state: -ETIMEDOUT [ 31.116332] zl3073x-i2c 1-0070: error -ETIMEDOUT: Failed to start device [ 31.136881] zl3073x-i2c 1-0070: probe with driver zl3073x-i2c failed with error -110 After this patch: [ 41.011438] zl3073x-i2c 1-0070: FW not fully ready - missing or corrupted config Fixes: 75a71ecc24125 ("dpll: zl3073x: Register DPLL devices and pins") Signed-off-by: Ivan Vecera Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251008141445.841113-1-ivecera@redhat.com Signed-off-by: Jakub Kicinski --- drivers/dpll/zl3073x/core.c | 21 +++++++++++++++++++++ drivers/dpll/zl3073x/regs.h | 3 +++ 2 files changed, 24 insertions(+) diff --git a/drivers/dpll/zl3073x/core.c b/drivers/dpll/zl3073x/core.c index 092e7027948a..e42e527813cf 100644 --- a/drivers/dpll/zl3073x/core.c +++ b/drivers/dpll/zl3073x/core.c @@ -1038,8 +1038,29 @@ zl3073x_dev_phase_meas_setup(struct zl3073x_dev *zldev) int zl3073x_dev_start(struct zl3073x_dev *zldev, bool full) { struct zl3073x_dpll *zldpll; + u8 info; int rc; + rc = zl3073x_read_u8(zldev, ZL_REG_INFO, &info); + if (rc) { + dev_err(zldev->dev, "Failed to read device status info\n"); + return rc; + } + + if (!FIELD_GET(ZL_INFO_READY, info)) { + /* The ready bit indicates that the firmware was successfully + * configured and is ready for normal operation. If it is + * cleared then the configuration stored in flash is wrong + * or missing. In this situation the driver will expose + * only devlink interface to give an opportunity to flash + * the correct config. + */ + dev_info(zldev->dev, + "FW not fully ready - missing or corrupted config\n"); + + return 0; + } + if (full) { /* Fetch device state */ rc = zl3073x_dev_state_fetch(zldev); diff --git a/drivers/dpll/zl3073x/regs.h b/drivers/dpll/zl3073x/regs.h index 19a25325bd9c..d837bee72b17 100644 --- a/drivers/dpll/zl3073x/regs.h +++ b/drivers/dpll/zl3073x/regs.h @@ -67,6 +67,9 @@ * Register Page 0, General **************************/ +#define ZL_REG_INFO ZL_REG(0, 0x00, 1) +#define ZL_INFO_READY BIT(7) + #define ZL_REG_ID ZL_REG(0, 0x01, 2) #define ZL_REG_REVISION ZL_REG(0, 0x03, 2) #define ZL_REG_FW_VER ZL_REG(0, 0x05, 2) From 25718fdcbdd2dadd15fc8b684df59b43970b91ed Mon Sep 17 00:00:00 2001 From: Sebastian Andrzej Siewior Date: Thu, 9 Oct 2025 11:43:38 +0200 Subject: [PATCH 18/50] net: gro_cells: Use nested-BH locking for gro_cell The gro_cell data structure is per-CPU variable and relies on disabled BH for its locking. Without per-CPU locking in local_bh_disable() on PREEMPT_RT this data structure requires explicit locking. Add a local_lock_t to the data structure and use local_lock_nested_bh() for locking. This change adds only lockdep coverage and does not alter the functional behaviour for !PREEMPT_RT. Reported-by: syzbot+8715dd783e9b0bef43b1@syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/68c6c3b1.050a0220.2ff435.0382.GAE@google.com/ Fixes: 3253cb49cbad ("softirq: Allow to drop the softirq-BKL lock on PREEMPT_RT") Signed-off-by: Sebastian Andrzej Siewior Link: https://patch.msgid.link/20251009094338.j1jyKfjR@linutronix.de Signed-off-by: Jakub Kicinski --- net/core/gro_cells.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/net/core/gro_cells.c b/net/core/gro_cells.c index ff8e5b64bf6b..b43911562f4d 100644 --- a/net/core/gro_cells.c +++ b/net/core/gro_cells.c @@ -8,11 +8,13 @@ struct gro_cell { struct sk_buff_head napi_skbs; struct napi_struct napi; + local_lock_t bh_lock; }; int gro_cells_receive(struct gro_cells *gcells, struct sk_buff *skb) { struct net_device *dev = skb->dev; + bool have_bh_lock = false; struct gro_cell *cell; int res; @@ -25,6 +27,8 @@ int gro_cells_receive(struct gro_cells *gcells, struct sk_buff *skb) goto unlock; } + local_lock_nested_bh(&gcells->cells->bh_lock); + have_bh_lock = true; cell = this_cpu_ptr(gcells->cells); if (skb_queue_len(&cell->napi_skbs) > READ_ONCE(net_hotdata.max_backlog)) { @@ -39,6 +43,9 @@ drop: if (skb_queue_len(&cell->napi_skbs) == 1) napi_schedule(&cell->napi); + if (have_bh_lock) + local_unlock_nested_bh(&gcells->cells->bh_lock); + res = NET_RX_SUCCESS; unlock: @@ -54,6 +61,7 @@ static int gro_cell_poll(struct napi_struct *napi, int budget) struct sk_buff *skb; int work_done = 0; + __local_lock_nested_bh(&cell->bh_lock); while (work_done < budget) { skb = __skb_dequeue(&cell->napi_skbs); if (!skb) @@ -64,6 +72,7 @@ static int gro_cell_poll(struct napi_struct *napi, int budget) if (work_done < budget) napi_complete_done(napi, work_done); + __local_unlock_nested_bh(&cell->bh_lock); return work_done; } @@ -79,6 +88,7 @@ int gro_cells_init(struct gro_cells *gcells, struct net_device *dev) struct gro_cell *cell = per_cpu_ptr(gcells->cells, i); __skb_queue_head_init(&cell->napi_skbs); + local_lock_init(&cell->bh_lock); set_bit(NAPI_STATE_NO_BUSY_POLL, &cell->napi.state); From 70f92ab97042f243e1c8da1c457ff56b9b3e49f1 Mon Sep 17 00:00:00 2001 From: Linmao Li Date: Thu, 9 Oct 2025 20:25:49 +0800 Subject: [PATCH 19/50] r8169: fix packet truncation after S4 resume on RTL8168H/RTL8111H After resume from S4 (hibernate), RTL8168H/RTL8111H truncates incoming packets. Packet captures show messages like "IP truncated-ip - 146 bytes missing!". The issue is caused by RxConfig not being properly re-initialized after resume. Re-initializing the RxConfig register before the chip re-initialization sequence avoids the truncation and restores correct packet reception. This follows the same pattern as commit ef9da46ddef0 ("r8169: fix data corruption issue on RTL8402"). Fixes: 6e1d0b898818 ("r8169:add support for RTL8168H and RTL8107E") Signed-off-by: Linmao Li Reviewed-by: Jacob Keller Reviewed-by: Heiner Kallweit Link: https://patch.msgid.link/20251009122549.3955845-1-lilinmao@kylinos.cn Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/realtek/r8169_main.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c index 8903ae90afcb..d18734fe12e4 100644 --- a/drivers/net/ethernet/realtek/r8169_main.c +++ b/drivers/net/ethernet/realtek/r8169_main.c @@ -4994,8 +4994,9 @@ static int rtl8169_resume(struct device *device) if (!device_may_wakeup(tp_to_dev(tp))) clk_prepare_enable(tp->clk); - /* Reportedly at least Asus X453MA truncates packets otherwise */ - if (tp->mac_version == RTL_GIGA_MAC_VER_37) + /* Some chip versions may truncate packets without this initialization */ + if (tp->mac_version == RTL_GIGA_MAC_VER_37 || + tp->mac_version == RTL_GIGA_MAC_VER_46) rtl_init_rxcfg(tp); return rtl8169_runtime_resume(device); From e4d0c909bf8328d986bf3aadba0c33a72b5ae30d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Kamil=20Hor=C3=A1k=20-=202N?= Date: Thu, 9 Oct 2025 15:06:56 +0200 Subject: [PATCH 20/50] net: phy: bcm54811: Fix GMII/MII/MII-Lite selection MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The Broadcom bcm54811 is hardware-strapped to select among RGMII and GMII/MII/MII-Lite modes. However, the corresponding bit, RGMII Enable in Miscellaneous Control Register must be also set to select desired RGMII or MII(-lite)/GMII mode. Fixes: 3117a11fff5af9e7 ("net: phy: bcm54811: PHY initialization") Signed-off-by: Kamil Horák - 2N Reviewed-by: Florian Fainelli Link: https://patch.msgid.link/20251009130656.1308237-2-kamilh@axis.com Signed-off-by: Jakub Kicinski --- drivers/net/phy/broadcom.c | 20 +++++++++++++++++++- include/linux/brcmphy.h | 1 + 2 files changed, 20 insertions(+), 1 deletion(-) diff --git a/drivers/net/phy/broadcom.c b/drivers/net/phy/broadcom.c index 3459a0e9d8b9..cb306f9e80cc 100644 --- a/drivers/net/phy/broadcom.c +++ b/drivers/net/phy/broadcom.c @@ -405,7 +405,7 @@ static int bcm5481x_set_brrmode(struct phy_device *phydev, bool on) static int bcm54811_config_init(struct phy_device *phydev) { struct bcm54xx_phy_priv *priv = phydev->priv; - int err, reg, exp_sync_ethernet; + int err, reg, exp_sync_ethernet, aux_rgmii_en; /* Enable CLK125 MUX on LED4 if ref clock is enabled. */ if (!(phydev->dev_flags & PHY_BRCM_RX_REFCLK_UNUSED)) { @@ -434,6 +434,24 @@ static int bcm54811_config_init(struct phy_device *phydev) if (err < 0) return err; + /* Enable RGMII if configured */ + if (phy_interface_is_rgmii(phydev)) + aux_rgmii_en = MII_BCM54XX_AUXCTL_SHDWSEL_MISC_RGMII_EN | + MII_BCM54XX_AUXCTL_SHDWSEL_MISC_RGMII_SKEW_EN; + else + aux_rgmii_en = 0; + + /* Also writing Reserved bits 6:5 because the documentation requires + * them to be written to 0b11 + */ + err = bcm54xx_auxctl_write(phydev, + MII_BCM54XX_AUXCTL_SHDWSEL_MISC, + MII_BCM54XX_AUXCTL_MISC_WREN | + aux_rgmii_en | + MII_BCM54XX_AUXCTL_SHDWSEL_MISC_RSVD); + if (err < 0) + return err; + return bcm5481x_set_brrmode(phydev, priv->brr_mode); } diff --git a/include/linux/brcmphy.h b/include/linux/brcmphy.h index 15c35655f482..115a964f3006 100644 --- a/include/linux/brcmphy.h +++ b/include/linux/brcmphy.h @@ -137,6 +137,7 @@ #define MII_BCM54XX_AUXCTL_SHDWSEL_MISC 0x07 #define MII_BCM54XX_AUXCTL_SHDWSEL_MISC_WIRESPEED_EN 0x0010 +#define MII_BCM54XX_AUXCTL_SHDWSEL_MISC_RSVD 0x0060 #define MII_BCM54XX_AUXCTL_SHDWSEL_MISC_RGMII_EN 0x0080 #define MII_BCM54XX_AUXCTL_SHDWSEL_MISC_RGMII_SKEW_EN 0x0100 #define MII_BCM54XX_AUXCTL_MISC_FORCE_AMDIX 0x0200 From 21f4d45eba0b2dcae5dbc9e5e0ad08735c993f16 Mon Sep 17 00:00:00 2001 From: Dmitry Safonov Date: Thu, 9 Oct 2025 16:02:19 +0100 Subject: [PATCH 21/50] net/ip6_tunnel: Prevent perpetual tunnel growth Similarly to ipv4 tunnel, ipv6 version updates dev->needed_headroom, too. While ipv4 tunnel headroom adjustment growth was limited in commit 5ae1e9922bbd ("net: ip_tunnel: prevent perpetual headroom growth"), ipv6 tunnel yet increases the headroom without any ceiling. Reflect ipv4 tunnel headroom adjustment limit on ipv6 version. Credits to Francesco Ruggeri, who was originally debugging this issue and wrote local Arista-specific patch and a reproducer. Fixes: 8eb30be0352d ("ipv6: Create ip6_tnl_xmit") Cc: Florian Westphal Cc: Francesco Ruggeri Signed-off-by: Dmitry Safonov Link: https://patch.msgid.link/20251009-ip6_tunnel-headroom-v2-1-8e4dbd8f7e35@arista.com Signed-off-by: Jakub Kicinski --- include/net/ip_tunnels.h | 15 +++++++++++++++ net/ipv4/ip_tunnel.c | 14 -------------- net/ipv6/ip6_tunnel.c | 3 +-- 3 files changed, 16 insertions(+), 16 deletions(-) diff --git a/include/net/ip_tunnels.h b/include/net/ip_tunnels.h index 4314a97702ea..ecae35512b9b 100644 --- a/include/net/ip_tunnels.h +++ b/include/net/ip_tunnels.h @@ -611,6 +611,21 @@ struct metadata_dst *iptunnel_metadata_reply(struct metadata_dst *md, int skb_tunnel_check_pmtu(struct sk_buff *skb, struct dst_entry *encap_dst, int headroom, bool reply); +static inline void ip_tunnel_adj_headroom(struct net_device *dev, + unsigned int headroom) +{ + /* we must cap headroom to some upperlimit, else pskb_expand_head + * will overflow header offsets in skb_headers_offset_update(). + */ + const unsigned int max_allowed = 512; + + if (headroom > max_allowed) + headroom = max_allowed; + + if (headroom > READ_ONCE(dev->needed_headroom)) + WRITE_ONCE(dev->needed_headroom, headroom); +} + int iptunnel_handle_offloads(struct sk_buff *skb, int gso_type_mask); static inline int iptunnel_pull_offloads(struct sk_buff *skb) diff --git a/net/ipv4/ip_tunnel.c b/net/ipv4/ip_tunnel.c index aaeb5d16f0c9..158a30ae7c5f 100644 --- a/net/ipv4/ip_tunnel.c +++ b/net/ipv4/ip_tunnel.c @@ -568,20 +568,6 @@ static int tnl_update_pmtu(struct net_device *dev, struct sk_buff *skb, return 0; } -static void ip_tunnel_adj_headroom(struct net_device *dev, unsigned int headroom) -{ - /* we must cap headroom to some upperlimit, else pskb_expand_head - * will overflow header offsets in skb_headers_offset_update(). - */ - static const unsigned int max_allowed = 512; - - if (headroom > max_allowed) - headroom = max_allowed; - - if (headroom > READ_ONCE(dev->needed_headroom)) - WRITE_ONCE(dev->needed_headroom, headroom); -} - void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, u8 proto, int tunnel_hlen) { diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c index 3262e81223df..6405072050e0 100644 --- a/net/ipv6/ip6_tunnel.c +++ b/net/ipv6/ip6_tunnel.c @@ -1257,8 +1257,7 @@ route_lookup: */ max_headroom = LL_RESERVED_SPACE(tdev) + sizeof(struct ipv6hdr) + dst->header_len + t->hlen; - if (max_headroom > READ_ONCE(dev->needed_headroom)) - WRITE_ONCE(dev->needed_headroom, max_headroom); + ip_tunnel_adj_headroom(dev, max_headroom); err = ip6_tnl_encap(skb, t, &proto, fl6); if (err) From a3f8c0a273120fd2638f03403e786c3de2382e72 Mon Sep 17 00:00:00 2001 From: Milena Olech Date: Thu, 9 Oct 2025 17:03:46 -0700 Subject: [PATCH 22/50] idpf: cleanup remaining SKBs in PTP flows When the driver requests Tx timestamp value, one of the first steps is to clone SKB using skb_get. It increases the reference counter for that SKB to prevent unexpected freeing by another component. However, there may be a case where the index is requested, SKB is assigned and never consumed by PTP flows - for example due to reset during running PTP apps. Add a check in release timestamping function to verify if the SKB assigned to Tx timestamp latch was freed, and release remaining SKBs. Fixes: 4901e83a94ef ("idpf: add Tx timestamp capabilities negotiation") Signed-off-by: Milena Olech Signed-off-by: Anton Nadezhdin Reviewed-by: Aleksandr Loktionov Tested-by: Samuel Salin Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20251009-jk-iwl-net-2025-10-01-v3-1-ef32a425b92a@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/idpf/idpf_ptp.c | 3 +++ drivers/net/ethernet/intel/idpf/idpf_virtchnl_ptp.c | 1 + 2 files changed, 4 insertions(+) diff --git a/drivers/net/ethernet/intel/idpf/idpf_ptp.c b/drivers/net/ethernet/intel/idpf/idpf_ptp.c index 142823af1f9e..3e1052d070cf 100644 --- a/drivers/net/ethernet/intel/idpf/idpf_ptp.c +++ b/drivers/net/ethernet/intel/idpf/idpf_ptp.c @@ -863,6 +863,9 @@ static void idpf_ptp_release_vport_tstamp(struct idpf_vport *vport) u64_stats_inc(&vport->tstamp_stats.flushed); list_del(&ptp_tx_tstamp->list_member); + if (ptp_tx_tstamp->skb) + consume_skb(ptp_tx_tstamp->skb); + kfree(ptp_tx_tstamp); } u64_stats_update_end(&vport->tstamp_stats.stats_sync); diff --git a/drivers/net/ethernet/intel/idpf/idpf_virtchnl_ptp.c b/drivers/net/ethernet/intel/idpf/idpf_virtchnl_ptp.c index 8a2e0f8c5e36..61cedb6f2854 100644 --- a/drivers/net/ethernet/intel/idpf/idpf_virtchnl_ptp.c +++ b/drivers/net/ethernet/intel/idpf/idpf_virtchnl_ptp.c @@ -517,6 +517,7 @@ idpf_ptp_get_tstamp_value(struct idpf_vport *vport, shhwtstamps.hwtstamp = ns_to_ktime(tstamp); skb_tstamp_tx(ptp_tx_tstamp->skb, &shhwtstamps); consume_skb(ptp_tx_tstamp->skb); + ptp_tx_tstamp->skb = NULL; list_add(&ptp_tx_tstamp->list_member, &tx_tstamp_caps->latches_free); From 53f0eb62b4d23d40686f2dd51776b8220f2887bb Mon Sep 17 00:00:00 2001 From: Jedrzej Jagielski Date: Thu, 9 Oct 2025 17:03:47 -0700 Subject: [PATCH 23/50] ixgbevf: fix getting link speed data for E610 devices E610 adapters no longer use the VFLINKS register to read PF's link speed and linkup state. As a result VF driver cannot get actual link state and it incorrectly reports 10G which is the default option. It leads to a situation where even 1G adapters print 10G as actual link speed. The same happens when PF driver set speed different than 10G. Add new mailbox operation to let the VF driver request a PF driver to provide actual link data. Update the mailbox api to v1.6. Incorporate both ways of getting link status within the legacy ixgbe_check_mac_link_vf() function. Fixes: 4c44b450c69b ("ixgbevf: Add support for Intel(R) E610 device") Co-developed-by: Andrzej Wilczynski Signed-off-by: Andrzej Wilczynski Reviewed-by: Przemek Kitszel Reviewed-by: Aleksandr Loktionov Cc: stable@vger.kernel.org Signed-off-by: Jedrzej Jagielski Tested-by: Rafal Romanowski Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20251009-jk-iwl-net-2025-10-01-v3-2-ef32a425b92a@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ixgbevf/defines.h | 1 + .../net/ethernet/intel/ixgbevf/ixgbevf_main.c | 6 +- drivers/net/ethernet/intel/ixgbevf/mbx.h | 4 + drivers/net/ethernet/intel/ixgbevf/vf.c | 137 ++++++++++++++---- 4 files changed, 116 insertions(+), 32 deletions(-) diff --git a/drivers/net/ethernet/intel/ixgbevf/defines.h b/drivers/net/ethernet/intel/ixgbevf/defines.h index a9bc96f6399d..e177d1d58696 100644 --- a/drivers/net/ethernet/intel/ixgbevf/defines.h +++ b/drivers/net/ethernet/intel/ixgbevf/defines.h @@ -28,6 +28,7 @@ /* Link speed */ typedef u32 ixgbe_link_speed; +#define IXGBE_LINK_SPEED_UNKNOWN 0 #define IXGBE_LINK_SPEED_1GB_FULL 0x0020 #define IXGBE_LINK_SPEED_10GB_FULL 0x0080 #define IXGBE_LINK_SPEED_100_FULL 0x0008 diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c index 28e25641b167..92671638b428 100644 --- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c +++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c @@ -2275,6 +2275,7 @@ static void ixgbevf_negotiate_api(struct ixgbevf_adapter *adapter) { struct ixgbe_hw *hw = &adapter->hw; static const int api[] = { + ixgbe_mbox_api_16, ixgbe_mbox_api_15, ixgbe_mbox_api_14, ixgbe_mbox_api_13, @@ -2294,7 +2295,8 @@ static void ixgbevf_negotiate_api(struct ixgbevf_adapter *adapter) idx++; } - if (hw->api_version >= ixgbe_mbox_api_15) { + /* Following is not supported by API 1.6, it is specific for 1.5 */ + if (hw->api_version == ixgbe_mbox_api_15) { hw->mbx.ops.init_params(hw); memcpy(&hw->mbx.ops, &ixgbevf_mbx_ops, sizeof(struct ixgbe_mbx_operations)); @@ -2651,6 +2653,7 @@ static void ixgbevf_set_num_queues(struct ixgbevf_adapter *adapter) case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: case ixgbe_mbox_api_15: + case ixgbe_mbox_api_16: if (adapter->xdp_prog && hw->mac.max_tx_queues == rss) rss = rss > 3 ? 2 : 1; @@ -4645,6 +4648,7 @@ static int ixgbevf_probe(struct pci_dev *pdev, const struct pci_device_id *ent) case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: case ixgbe_mbox_api_15: + case ixgbe_mbox_api_16: netdev->max_mtu = IXGBE_MAX_JUMBO_FRAME_SIZE - (ETH_HLEN + ETH_FCS_LEN); break; diff --git a/drivers/net/ethernet/intel/ixgbevf/mbx.h b/drivers/net/ethernet/intel/ixgbevf/mbx.h index 835bbcc5cc8e..c1494fd1f67b 100644 --- a/drivers/net/ethernet/intel/ixgbevf/mbx.h +++ b/drivers/net/ethernet/intel/ixgbevf/mbx.h @@ -66,6 +66,7 @@ enum ixgbe_pfvf_api_rev { ixgbe_mbox_api_13, /* API version 1.3, linux/freebsd VF driver */ ixgbe_mbox_api_14, /* API version 1.4, linux/freebsd VF driver */ ixgbe_mbox_api_15, /* API version 1.5, linux/freebsd VF driver */ + ixgbe_mbox_api_16, /* API version 1.6, linux/freebsd VF driver */ /* This value should always be last */ ixgbe_mbox_api_unknown, /* indicates that API version is not known */ }; @@ -102,6 +103,9 @@ enum ixgbe_pfvf_api_rev { #define IXGBE_VF_GET_LINK_STATE 0x10 /* get vf link state */ +/* mailbox API, version 1.6 VF requests */ +#define IXGBE_VF_GET_PF_LINK_STATE 0x11 /* request PF to send link info */ + /* length of permanent address message returned from PF */ #define IXGBE_VF_PERMADDR_MSG_LEN 4 /* word in permanent address message with the current multicast type */ diff --git a/drivers/net/ethernet/intel/ixgbevf/vf.c b/drivers/net/ethernet/intel/ixgbevf/vf.c index dcaef34b88b6..f05246fb5a74 100644 --- a/drivers/net/ethernet/intel/ixgbevf/vf.c +++ b/drivers/net/ethernet/intel/ixgbevf/vf.c @@ -313,6 +313,7 @@ int ixgbevf_get_reta_locked(struct ixgbe_hw *hw, u32 *reta, int num_rx_queues) * is not supported for this device type. */ switch (hw->api_version) { + case ixgbe_mbox_api_16: case ixgbe_mbox_api_15: case ixgbe_mbox_api_14: case ixgbe_mbox_api_13: @@ -382,6 +383,7 @@ int ixgbevf_get_rss_key_locked(struct ixgbe_hw *hw, u8 *rss_key) * or if the operation is not supported for this device type. */ switch (hw->api_version) { + case ixgbe_mbox_api_16: case ixgbe_mbox_api_15: case ixgbe_mbox_api_14: case ixgbe_mbox_api_13: @@ -552,6 +554,7 @@ static s32 ixgbevf_update_xcast_mode(struct ixgbe_hw *hw, int xcast_mode) case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: case ixgbe_mbox_api_15: + case ixgbe_mbox_api_16: break; default: return -EOPNOTSUPP; @@ -624,6 +627,48 @@ static s32 ixgbevf_hv_get_link_state_vf(struct ixgbe_hw *hw, bool *link_state) return -EOPNOTSUPP; } +/** + * ixgbevf_get_pf_link_state - Get PF's link status + * @hw: pointer to the HW structure + * @speed: link speed + * @link_up: indicate if link is up/down + * + * Ask PF to provide link_up state and speed of the link. + * + * Return: IXGBE_ERR_MBX in the case of mailbox error, + * -EOPNOTSUPP if the op is not supported or 0 on success. + */ +static int ixgbevf_get_pf_link_state(struct ixgbe_hw *hw, ixgbe_link_speed *speed, + bool *link_up) +{ + u32 msgbuf[3] = {}; + int err; + + switch (hw->api_version) { + case ixgbe_mbox_api_16: + break; + default: + return -EOPNOTSUPP; + } + + msgbuf[0] = IXGBE_VF_GET_PF_LINK_STATE; + + err = ixgbevf_write_msg_read_ack(hw, msgbuf, msgbuf, + ARRAY_SIZE(msgbuf)); + if (err || (msgbuf[0] & IXGBE_VT_MSGTYPE_FAILURE)) { + err = IXGBE_ERR_MBX; + *speed = IXGBE_LINK_SPEED_UNKNOWN; + /* No need to set @link_up to false as it will be done by + * ixgbe_check_mac_link_vf(). + */ + } else { + *speed = msgbuf[1]; + *link_up = msgbuf[2]; + } + + return err; +} + /** * ixgbevf_set_vfta_vf - Set/Unset VLAN filter table address * @hw: pointer to the HW structure @@ -658,6 +703,58 @@ mbx_err: return err; } +/** + * ixgbe_read_vflinks - Read VFLINKS register + * @hw: pointer to the HW structure + * @speed: link speed + * @link_up: indicate if link is up/down + * + * Get linkup status and link speed from the VFLINKS register. + */ +static void ixgbe_read_vflinks(struct ixgbe_hw *hw, ixgbe_link_speed *speed, + bool *link_up) +{ + u32 vflinks = IXGBE_READ_REG(hw, IXGBE_VFLINKS); + + /* if link status is down no point in checking to see if PF is up */ + if (!(vflinks & IXGBE_LINKS_UP)) { + *link_up = false; + return; + } + + /* for SFP+ modules and DA cables on 82599 it can take up to 500usecs + * before the link status is correct + */ + if (hw->mac.type == ixgbe_mac_82599_vf) { + for (int i = 0; i < 5; i++) { + udelay(100); + vflinks = IXGBE_READ_REG(hw, IXGBE_VFLINKS); + + if (!(vflinks & IXGBE_LINKS_UP)) { + *link_up = false; + return; + } + } + } + + /* We reached this point so there's link */ + *link_up = true; + + switch (vflinks & IXGBE_LINKS_SPEED_82599) { + case IXGBE_LINKS_SPEED_10G_82599: + *speed = IXGBE_LINK_SPEED_10GB_FULL; + break; + case IXGBE_LINKS_SPEED_1G_82599: + *speed = IXGBE_LINK_SPEED_1GB_FULL; + break; + case IXGBE_LINKS_SPEED_100_82599: + *speed = IXGBE_LINK_SPEED_100_FULL; + break; + default: + *speed = IXGBE_LINK_SPEED_UNKNOWN; + } +} + /** * ixgbevf_hv_set_vfta_vf - * Hyper-V variant - just a stub. * @hw: unused @@ -705,7 +802,6 @@ static s32 ixgbevf_check_mac_link_vf(struct ixgbe_hw *hw, struct ixgbe_mbx_info *mbx = &hw->mbx; struct ixgbe_mac_info *mac = &hw->mac; s32 ret_val = 0; - u32 links_reg; u32 in_msg = 0; /* If we were hit with a reset drop the link */ @@ -715,36 +811,14 @@ static s32 ixgbevf_check_mac_link_vf(struct ixgbe_hw *hw, if (!mac->get_link_status) goto out; - /* if link status is down no point in checking to see if pf is up */ - links_reg = IXGBE_READ_REG(hw, IXGBE_VFLINKS); - if (!(links_reg & IXGBE_LINKS_UP)) - goto out; - - /* for SFP+ modules and DA cables on 82599 it can take up to 500usecs - * before the link status is correct - */ - if (mac->type == ixgbe_mac_82599_vf) { - int i; - - for (i = 0; i < 5; i++) { - udelay(100); - links_reg = IXGBE_READ_REG(hw, IXGBE_VFLINKS); - - if (!(links_reg & IXGBE_LINKS_UP)) - goto out; - } - } - - switch (links_reg & IXGBE_LINKS_SPEED_82599) { - case IXGBE_LINKS_SPEED_10G_82599: - *speed = IXGBE_LINK_SPEED_10GB_FULL; - break; - case IXGBE_LINKS_SPEED_1G_82599: - *speed = IXGBE_LINK_SPEED_1GB_FULL; - break; - case IXGBE_LINKS_SPEED_100_82599: - *speed = IXGBE_LINK_SPEED_100_FULL; - break; + if (hw->mac.type == ixgbe_mac_e610_vf) { + ret_val = ixgbevf_get_pf_link_state(hw, speed, link_up); + if (ret_val) + goto out; + } else { + ixgbe_read_vflinks(hw, speed, link_up); + if (*link_up == false) + goto out; } /* if the read failed it could just be a mailbox collision, best wait @@ -951,6 +1025,7 @@ int ixgbevf_get_queues(struct ixgbe_hw *hw, unsigned int *num_tcs, case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: case ixgbe_mbox_api_15: + case ixgbe_mbox_api_16: break; default: return 0; From f7f97cbc03a470ce405d48dedb7f135713caa0fa Mon Sep 17 00:00:00 2001 From: Jedrzej Jagielski Date: Thu, 9 Oct 2025 17:03:48 -0700 Subject: [PATCH 24/50] ixgbe: handle IXGBE_VF_GET_PF_LINK_STATE mailbox operation Update supported API version and provide handler for IXGBE_VF_GET_PF_LINK_STATE cmd. Simply put stored values of link speed and link_up from adapter context. Reviewed-by: Przemek Kitszel Reviewed-by: Aleksandr Loktionov Signed-off-by: Jedrzej Jagielski Link: https://lore.kernel.org/stable/20250828095227.1857066-3-jedrzej.jagielski%40intel.com Tested-by: Rafal Romanowski Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20251009-jk-iwl-net-2025-10-01-v3-3-ef32a425b92a@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h | 5 +++ .../net/ethernet/intel/ixgbe/ixgbe_sriov.c | 42 +++++++++++++++++++ 2 files changed, 47 insertions(+) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h index 4af149b63a39..f7256a339c99 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h @@ -50,6 +50,8 @@ enum ixgbe_pfvf_api_rev { ixgbe_mbox_api_12, /* API version 1.2, linux/freebsd VF driver */ ixgbe_mbox_api_13, /* API version 1.3, linux/freebsd VF driver */ ixgbe_mbox_api_14, /* API version 1.4, linux/freebsd VF driver */ + ixgbe_mbox_api_15, /* API version 1.5, linux/freebsd VF driver */ + ixgbe_mbox_api_16, /* API version 1.6, linux/freebsd VF driver */ /* This value should always be last */ ixgbe_mbox_api_unknown, /* indicates that API version is not known */ }; @@ -86,6 +88,9 @@ enum ixgbe_pfvf_api_rev { #define IXGBE_VF_GET_LINK_STATE 0x10 /* get vf link state */ +/* mailbox API, version 1.6 VF requests */ +#define IXGBE_VF_GET_PF_LINK_STATE 0x11 /* request PF to send link info */ + /* length of permanent address message returned from PF */ #define IXGBE_VF_PERMADDR_MSG_LEN 4 /* word in permanent address message with the current multicast type */ diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c index 32ac1e020d91..b09271d61a4e 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c @@ -510,6 +510,7 @@ static int ixgbe_set_vf_lpe(struct ixgbe_adapter *adapter, u32 max_frame, u32 vf case ixgbe_mbox_api_12: case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: + case ixgbe_mbox_api_16: /* Version 1.1 supports jumbo frames on VFs if PF has * jumbo frames enabled which means legacy VFs are * disabled @@ -1046,6 +1047,7 @@ static int ixgbe_negotiate_vf_api(struct ixgbe_adapter *adapter, case ixgbe_mbox_api_12: case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: + case ixgbe_mbox_api_16: adapter->vfinfo[vf].vf_api = api; return 0; default: @@ -1072,6 +1074,7 @@ static int ixgbe_get_vf_queues(struct ixgbe_adapter *adapter, case ixgbe_mbox_api_12: case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: + case ixgbe_mbox_api_16: break; default: return -1; @@ -1112,6 +1115,7 @@ static int ixgbe_get_vf_reta(struct ixgbe_adapter *adapter, u32 *msgbuf, u32 vf) /* verify the PF is supporting the correct API */ switch (adapter->vfinfo[vf].vf_api) { + case ixgbe_mbox_api_16: case ixgbe_mbox_api_14: case ixgbe_mbox_api_13: case ixgbe_mbox_api_12: @@ -1145,6 +1149,7 @@ static int ixgbe_get_vf_rss_key(struct ixgbe_adapter *adapter, /* verify the PF is supporting the correct API */ switch (adapter->vfinfo[vf].vf_api) { + case ixgbe_mbox_api_16: case ixgbe_mbox_api_14: case ixgbe_mbox_api_13: case ixgbe_mbox_api_12: @@ -1174,6 +1179,7 @@ static int ixgbe_update_vf_xcast_mode(struct ixgbe_adapter *adapter, fallthrough; case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: + case ixgbe_mbox_api_16: break; default: return -EOPNOTSUPP; @@ -1244,6 +1250,7 @@ static int ixgbe_get_vf_link_state(struct ixgbe_adapter *adapter, case ixgbe_mbox_api_12: case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: + case ixgbe_mbox_api_16: break; default: return -EOPNOTSUPP; @@ -1254,6 +1261,38 @@ static int ixgbe_get_vf_link_state(struct ixgbe_adapter *adapter, return 0; } +/** + * ixgbe_send_vf_link_status - send link status data to VF + * @adapter: pointer to adapter struct + * @msgbuf: pointer to message buffers + * @vf: VF identifier + * + * Reply for IXGBE_VF_GET_PF_LINK_STATE mbox command sending link status data. + * + * Return: 0 on success or -EOPNOTSUPP when operation is not supported. + */ +static int ixgbe_send_vf_link_status(struct ixgbe_adapter *adapter, + u32 *msgbuf, u32 vf) +{ + struct ixgbe_hw *hw = &adapter->hw; + + switch (adapter->vfinfo[vf].vf_api) { + case ixgbe_mbox_api_16: + if (hw->mac.type != ixgbe_mac_e610) + return -EOPNOTSUPP; + break; + default: + return -EOPNOTSUPP; + } + /* Simply provide stored values as watchdog & link status events take + * care of its freshness. + */ + msgbuf[1] = adapter->link_speed; + msgbuf[2] = adapter->link_up; + + return 0; +} + static int ixgbe_rcv_msg_from_vf(struct ixgbe_adapter *adapter, u32 vf) { u32 mbx_size = IXGBE_VFMAILBOX_SIZE; @@ -1328,6 +1367,9 @@ static int ixgbe_rcv_msg_from_vf(struct ixgbe_adapter *adapter, u32 vf) case IXGBE_VF_IPSEC_DEL: retval = ixgbe_ipsec_vf_del_sa(adapter, msgbuf, vf); break; + case IXGBE_VF_GET_PF_LINK_STATE: + retval = ixgbe_send_vf_link_status(adapter, msgbuf, vf); + break; default: e_err(drv, "Unhandled Msg %8.8x\n", msgbuf[0]); retval = -EIO; From a7075f501bd33c93570af759b6f4302ef0175168 Mon Sep 17 00:00:00 2001 From: Jedrzej Jagielski Date: Thu, 9 Oct 2025 17:03:49 -0700 Subject: [PATCH 25/50] ixgbevf: fix mailbox API compatibility by negotiating supported features There was backward compatibility in the terms of mailbox API. Various drivers from various OSes supporting 10G adapters from Intel portfolio could easily negotiate mailbox API. This convention has been broken since introducing API 1.4. Commit 0062e7cc955e ("ixgbevf: add VF IPsec offload code") added support for IPSec which is specific only for the kernel ixgbe driver. None of the rest of the Intel 10G PF/VF drivers supports it. And actually lack of support was not included in the IPSec implementation - there were no such code paths. No possibility to negotiate support for the feature was introduced along with introduction of the feature itself. Commit 339f28964147 ("ixgbevf: Add support for new mailbox communication between PF and VF") increasing API version to 1.5 did the same - it introduced code supported specifically by the PF ESX driver. It altered API version for the VF driver in the same time not touching the version defined for the PF ixgbe driver. It led to additional discrepancies, as the code provided within API 1.6 cannot be supported for Linux ixgbe driver as it causes crashes. The issue was noticed some time ago and mitigated by Jake within the commit d0725312adf5 ("ixgbevf: stop attempting IPSEC offload on Mailbox API 1.5"). As a result we have regression for IPsec support and after increasing API to version 1.6 ixgbevf driver stopped to support ESX MBX. To fix this mess add new mailbox op asking PF driver about supported features. Basing on a response determine whether to set support for IPSec and ESX-specific enhanced mailbox. New mailbox op, for compatibility purposes, must be added within new API revision, as API version of OOT PF & VF drivers is already increased to 1.6 and doesn't incorporate features negotiate op. Features negotiation mechanism gives possibility to be extended with new features when needed in the future. Reported-by: Jacob Keller Closes: https://lore.kernel.org/intel-wired-lan/20241101-jk-ixgbevf-mailbox-v1-5-fixes-v1-0-f556dc9a66ed@intel.com/ Fixes: 0062e7cc955e ("ixgbevf: add VF IPsec offload code") Fixes: 339f28964147 ("ixgbevf: Add support for new mailbox communication between PF and VF") Reviewed-by: Jacob Keller Reviewed-by: Przemek Kitszel Reviewed-by: Aleksandr Loktionov Cc: stable@vger.kernel.org Signed-off-by: Jedrzej Jagielski Tested-by: Rafal Romanowski Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20251009-jk-iwl-net-2025-10-01-v3-4-ef32a425b92a@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ixgbevf/ipsec.c | 10 +++++ drivers/net/ethernet/intel/ixgbevf/ixgbevf.h | 7 +++ .../net/ethernet/intel/ixgbevf/ixgbevf_main.c | 32 ++++++++++++- drivers/net/ethernet/intel/ixgbevf/mbx.h | 4 ++ drivers/net/ethernet/intel/ixgbevf/vf.c | 45 ++++++++++++++++++- drivers/net/ethernet/intel/ixgbevf/vf.h | 1 + 6 files changed, 96 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/intel/ixgbevf/ipsec.c b/drivers/net/ethernet/intel/ixgbevf/ipsec.c index 65580b9cb06f..fce35924ff8b 100644 --- a/drivers/net/ethernet/intel/ixgbevf/ipsec.c +++ b/drivers/net/ethernet/intel/ixgbevf/ipsec.c @@ -273,6 +273,9 @@ static int ixgbevf_ipsec_add_sa(struct net_device *dev, adapter = netdev_priv(dev); ipsec = adapter->ipsec; + if (!(adapter->pf_features & IXGBEVF_PF_SUP_IPSEC)) + return -EOPNOTSUPP; + if (xs->id.proto != IPPROTO_ESP && xs->id.proto != IPPROTO_AH) { NL_SET_ERR_MSG_MOD(extack, "Unsupported protocol for IPsec offload"); return -EINVAL; @@ -405,6 +408,9 @@ static void ixgbevf_ipsec_del_sa(struct net_device *dev, adapter = netdev_priv(dev); ipsec = adapter->ipsec; + if (!(adapter->pf_features & IXGBEVF_PF_SUP_IPSEC)) + return; + if (xs->xso.dir == XFRM_DEV_OFFLOAD_IN) { sa_idx = xs->xso.offload_handle - IXGBE_IPSEC_BASE_RX_INDEX; @@ -612,6 +618,10 @@ void ixgbevf_init_ipsec_offload(struct ixgbevf_adapter *adapter) size_t size; switch (adapter->hw.api_version) { + case ixgbe_mbox_api_17: + if (!(adapter->pf_features & IXGBEVF_PF_SUP_IPSEC)) + return; + break; case ixgbe_mbox_api_14: break; default: diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h index 3a379e6a3a2a..039187607e98 100644 --- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h +++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h @@ -363,6 +363,13 @@ struct ixgbevf_adapter { struct ixgbe_hw hw; u16 msg_enable; + u32 pf_features; +#define IXGBEVF_PF_SUP_IPSEC BIT(0) +#define IXGBEVF_PF_SUP_ESX_MBX BIT(1) + +#define IXGBEVF_SUPPORTED_FEATURES (IXGBEVF_PF_SUP_IPSEC | \ + IXGBEVF_PF_SUP_ESX_MBX) + struct ixgbevf_hw_stats stats; unsigned long state; diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c index 92671638b428..d5ce20f47def 100644 --- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c +++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c @@ -2271,10 +2271,35 @@ static void ixgbevf_init_last_counter_stats(struct ixgbevf_adapter *adapter) adapter->stats.base_vfmprc = adapter->stats.last_vfmprc; } +/** + * ixgbevf_set_features - Set features supported by PF + * @adapter: pointer to the adapter struct + * + * Negotiate with PF supported features and then set pf_features accordingly. + */ +static void ixgbevf_set_features(struct ixgbevf_adapter *adapter) +{ + u32 *pf_features = &adapter->pf_features; + struct ixgbe_hw *hw = &adapter->hw; + int err; + + err = hw->mac.ops.negotiate_features(hw, pf_features); + if (err && err != -EOPNOTSUPP) + netdev_dbg(adapter->netdev, + "PF feature negotiation failed.\n"); + + /* Address also pre API 1.7 cases */ + if (hw->api_version == ixgbe_mbox_api_14) + *pf_features |= IXGBEVF_PF_SUP_IPSEC; + else if (hw->api_version == ixgbe_mbox_api_15) + *pf_features |= IXGBEVF_PF_SUP_ESX_MBX; +} + static void ixgbevf_negotiate_api(struct ixgbevf_adapter *adapter) { struct ixgbe_hw *hw = &adapter->hw; static const int api[] = { + ixgbe_mbox_api_17, ixgbe_mbox_api_16, ixgbe_mbox_api_15, ixgbe_mbox_api_14, @@ -2295,8 +2320,9 @@ static void ixgbevf_negotiate_api(struct ixgbevf_adapter *adapter) idx++; } - /* Following is not supported by API 1.6, it is specific for 1.5 */ - if (hw->api_version == ixgbe_mbox_api_15) { + ixgbevf_set_features(adapter); + + if (adapter->pf_features & IXGBEVF_PF_SUP_ESX_MBX) { hw->mbx.ops.init_params(hw); memcpy(&hw->mbx.ops, &ixgbevf_mbx_ops, sizeof(struct ixgbe_mbx_operations)); @@ -2654,6 +2680,7 @@ static void ixgbevf_set_num_queues(struct ixgbevf_adapter *adapter) case ixgbe_mbox_api_14: case ixgbe_mbox_api_15: case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: if (adapter->xdp_prog && hw->mac.max_tx_queues == rss) rss = rss > 3 ? 2 : 1; @@ -4649,6 +4676,7 @@ static int ixgbevf_probe(struct pci_dev *pdev, const struct pci_device_id *ent) case ixgbe_mbox_api_14: case ixgbe_mbox_api_15: case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: netdev->max_mtu = IXGBE_MAX_JUMBO_FRAME_SIZE - (ETH_HLEN + ETH_FCS_LEN); break; diff --git a/drivers/net/ethernet/intel/ixgbevf/mbx.h b/drivers/net/ethernet/intel/ixgbevf/mbx.h index c1494fd1f67b..a8ed23ee66aa 100644 --- a/drivers/net/ethernet/intel/ixgbevf/mbx.h +++ b/drivers/net/ethernet/intel/ixgbevf/mbx.h @@ -67,6 +67,7 @@ enum ixgbe_pfvf_api_rev { ixgbe_mbox_api_14, /* API version 1.4, linux/freebsd VF driver */ ixgbe_mbox_api_15, /* API version 1.5, linux/freebsd VF driver */ ixgbe_mbox_api_16, /* API version 1.6, linux/freebsd VF driver */ + ixgbe_mbox_api_17, /* API version 1.7, linux/freebsd VF driver */ /* This value should always be last */ ixgbe_mbox_api_unknown, /* indicates that API version is not known */ }; @@ -106,6 +107,9 @@ enum ixgbe_pfvf_api_rev { /* mailbox API, version 1.6 VF requests */ #define IXGBE_VF_GET_PF_LINK_STATE 0x11 /* request PF to send link info */ +/* mailbox API, version 1.7 VF requests */ +#define IXGBE_VF_FEATURES_NEGOTIATE 0x12 /* get features supported by PF*/ + /* length of permanent address message returned from PF */ #define IXGBE_VF_PERMADDR_MSG_LEN 4 /* word in permanent address message with the current multicast type */ diff --git a/drivers/net/ethernet/intel/ixgbevf/vf.c b/drivers/net/ethernet/intel/ixgbevf/vf.c index f05246fb5a74..74d320879513 100644 --- a/drivers/net/ethernet/intel/ixgbevf/vf.c +++ b/drivers/net/ethernet/intel/ixgbevf/vf.c @@ -313,6 +313,7 @@ int ixgbevf_get_reta_locked(struct ixgbe_hw *hw, u32 *reta, int num_rx_queues) * is not supported for this device type. */ switch (hw->api_version) { + case ixgbe_mbox_api_17: case ixgbe_mbox_api_16: case ixgbe_mbox_api_15: case ixgbe_mbox_api_14: @@ -383,6 +384,7 @@ int ixgbevf_get_rss_key_locked(struct ixgbe_hw *hw, u8 *rss_key) * or if the operation is not supported for this device type. */ switch (hw->api_version) { + case ixgbe_mbox_api_17: case ixgbe_mbox_api_16: case ixgbe_mbox_api_15: case ixgbe_mbox_api_14: @@ -555,6 +557,7 @@ static s32 ixgbevf_update_xcast_mode(struct ixgbe_hw *hw, int xcast_mode) case ixgbe_mbox_api_14: case ixgbe_mbox_api_15: case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: break; default: return -EOPNOTSUPP; @@ -646,6 +649,7 @@ static int ixgbevf_get_pf_link_state(struct ixgbe_hw *hw, ixgbe_link_speed *spee switch (hw->api_version) { case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: break; default: return -EOPNOTSUPP; @@ -669,6 +673,42 @@ static int ixgbevf_get_pf_link_state(struct ixgbe_hw *hw, ixgbe_link_speed *spee return err; } +/** + * ixgbevf_negotiate_features_vf - negotiate supported features with PF driver + * @hw: pointer to the HW structure + * @pf_features: bitmask of features supported by PF + * + * Return: IXGBE_ERR_MBX in the case of mailbox error, + * -EOPNOTSUPP if the op is not supported or 0 on success. + */ +static int ixgbevf_negotiate_features_vf(struct ixgbe_hw *hw, u32 *pf_features) +{ + u32 msgbuf[2] = {}; + int err; + + switch (hw->api_version) { + case ixgbe_mbox_api_17: + break; + default: + return -EOPNOTSUPP; + } + + msgbuf[0] = IXGBE_VF_FEATURES_NEGOTIATE; + msgbuf[1] = IXGBEVF_SUPPORTED_FEATURES; + + err = ixgbevf_write_msg_read_ack(hw, msgbuf, msgbuf, + ARRAY_SIZE(msgbuf)); + + if (err || (msgbuf[0] & IXGBE_VT_MSGTYPE_FAILURE)) { + err = IXGBE_ERR_MBX; + *pf_features = 0x0; + } else { + *pf_features = msgbuf[1]; + } + + return err; +} + /** * ixgbevf_set_vfta_vf - Set/Unset VLAN filter table address * @hw: pointer to the HW structure @@ -799,6 +839,7 @@ static s32 ixgbevf_check_mac_link_vf(struct ixgbe_hw *hw, bool *link_up, bool autoneg_wait_to_complete) { + struct ixgbevf_adapter *adapter = hw->back; struct ixgbe_mbx_info *mbx = &hw->mbx; struct ixgbe_mac_info *mac = &hw->mac; s32 ret_val = 0; @@ -825,7 +866,7 @@ static s32 ixgbevf_check_mac_link_vf(struct ixgbe_hw *hw, * until we are called again and don't report an error */ if (mbx->ops.read(hw, &in_msg, 1)) { - if (hw->api_version >= ixgbe_mbox_api_15) + if (adapter->pf_features & IXGBEVF_PF_SUP_ESX_MBX) mac->get_link_status = false; goto out; } @@ -1026,6 +1067,7 @@ int ixgbevf_get_queues(struct ixgbe_hw *hw, unsigned int *num_tcs, case ixgbe_mbox_api_14: case ixgbe_mbox_api_15: case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: break; default: return 0; @@ -1080,6 +1122,7 @@ static const struct ixgbe_mac_operations ixgbevf_mac_ops = { .setup_link = ixgbevf_setup_mac_link_vf, .check_link = ixgbevf_check_mac_link_vf, .negotiate_api_version = ixgbevf_negotiate_api_version_vf, + .negotiate_features = ixgbevf_negotiate_features_vf, .set_rar = ixgbevf_set_rar_vf, .update_mc_addr_list = ixgbevf_update_mc_addr_list_vf, .update_xcast_mode = ixgbevf_update_xcast_mode, diff --git a/drivers/net/ethernet/intel/ixgbevf/vf.h b/drivers/net/ethernet/intel/ixgbevf/vf.h index 2d791bc26ae4..4f19b8900c29 100644 --- a/drivers/net/ethernet/intel/ixgbevf/vf.h +++ b/drivers/net/ethernet/intel/ixgbevf/vf.h @@ -26,6 +26,7 @@ struct ixgbe_mac_operations { s32 (*stop_adapter)(struct ixgbe_hw *); s32 (*get_bus_info)(struct ixgbe_hw *); s32 (*negotiate_api_version)(struct ixgbe_hw *hw, int api); + int (*negotiate_features)(struct ixgbe_hw *hw, u32 *pf_features); /* Link */ s32 (*setup_link)(struct ixgbe_hw *, ixgbe_link_speed, bool, bool); From 823be089f9c8ab136ba382b516aedd3f7ac854bd Mon Sep 17 00:00:00 2001 From: Jedrzej Jagielski Date: Thu, 9 Oct 2025 17:03:50 -0700 Subject: [PATCH 26/50] ixgbe: handle IXGBE_VF_FEATURES_NEGOTIATE mbox cmd Send to VF information about features supported by the PF driver. Increase API version to 1.7. Reviewed-by: Przemek Kitszel Reviewed-by: Aleksandr Loktionov Signed-off-by: Jedrzej Jagielski Tested-by: Rafal Romanowski Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20251009-jk-iwl-net-2025-10-01-v3-5-ef32a425b92a@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h | 10 +++++ .../net/ethernet/intel/ixgbe/ixgbe_sriov.c | 37 +++++++++++++++++++ 2 files changed, 47 insertions(+) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h index f7256a339c99..0334ed4b8fa3 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h @@ -52,6 +52,7 @@ enum ixgbe_pfvf_api_rev { ixgbe_mbox_api_14, /* API version 1.4, linux/freebsd VF driver */ ixgbe_mbox_api_15, /* API version 1.5, linux/freebsd VF driver */ ixgbe_mbox_api_16, /* API version 1.6, linux/freebsd VF driver */ + ixgbe_mbox_api_17, /* API version 1.7, linux/freebsd VF driver */ /* This value should always be last */ ixgbe_mbox_api_unknown, /* indicates that API version is not known */ }; @@ -91,6 +92,9 @@ enum ixgbe_pfvf_api_rev { /* mailbox API, version 1.6 VF requests */ #define IXGBE_VF_GET_PF_LINK_STATE 0x11 /* request PF to send link info */ +/* mailbox API, version 1.7 VF requests */ +#define IXGBE_VF_FEATURES_NEGOTIATE 0x12 /* get features supported by PF */ + /* length of permanent address message returned from PF */ #define IXGBE_VF_PERMADDR_MSG_LEN 4 /* word in permanent address message with the current multicast type */ @@ -101,6 +105,12 @@ enum ixgbe_pfvf_api_rev { #define IXGBE_VF_MBX_INIT_TIMEOUT 2000 /* number of retries on mailbox */ #define IXGBE_VF_MBX_INIT_DELAY 500 /* microseconds between retries */ +/* features negotiated between PF/VF */ +#define IXGBEVF_PF_SUP_IPSEC BIT(0) +#define IXGBEVF_PF_SUP_ESX_MBX BIT(1) + +#define IXGBE_SUPPORTED_FEATURES IXGBEVF_PF_SUP_IPSEC + struct ixgbe_hw; int ixgbe_read_mbx(struct ixgbe_hw *, u32 *, u16, u16); diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c index b09271d61a4e..ee133d6749b3 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c @@ -511,6 +511,7 @@ static int ixgbe_set_vf_lpe(struct ixgbe_adapter *adapter, u32 max_frame, u32 vf case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: /* Version 1.1 supports jumbo frames on VFs if PF has * jumbo frames enabled which means legacy VFs are * disabled @@ -1048,6 +1049,7 @@ static int ixgbe_negotiate_vf_api(struct ixgbe_adapter *adapter, case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: adapter->vfinfo[vf].vf_api = api; return 0; default: @@ -1075,6 +1077,7 @@ static int ixgbe_get_vf_queues(struct ixgbe_adapter *adapter, case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: break; default: return -1; @@ -1115,6 +1118,7 @@ static int ixgbe_get_vf_reta(struct ixgbe_adapter *adapter, u32 *msgbuf, u32 vf) /* verify the PF is supporting the correct API */ switch (adapter->vfinfo[vf].vf_api) { + case ixgbe_mbox_api_17: case ixgbe_mbox_api_16: case ixgbe_mbox_api_14: case ixgbe_mbox_api_13: @@ -1149,6 +1153,7 @@ static int ixgbe_get_vf_rss_key(struct ixgbe_adapter *adapter, /* verify the PF is supporting the correct API */ switch (adapter->vfinfo[vf].vf_api) { + case ixgbe_mbox_api_17: case ixgbe_mbox_api_16: case ixgbe_mbox_api_14: case ixgbe_mbox_api_13: @@ -1180,6 +1185,7 @@ static int ixgbe_update_vf_xcast_mode(struct ixgbe_adapter *adapter, case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: break; default: return -EOPNOTSUPP; @@ -1251,6 +1257,7 @@ static int ixgbe_get_vf_link_state(struct ixgbe_adapter *adapter, case ixgbe_mbox_api_13: case ixgbe_mbox_api_14: case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: break; default: return -EOPNOTSUPP; @@ -1278,6 +1285,7 @@ static int ixgbe_send_vf_link_status(struct ixgbe_adapter *adapter, switch (adapter->vfinfo[vf].vf_api) { case ixgbe_mbox_api_16: + case ixgbe_mbox_api_17: if (hw->mac.type != ixgbe_mac_e610) return -EOPNOTSUPP; break; @@ -1293,6 +1301,32 @@ static int ixgbe_send_vf_link_status(struct ixgbe_adapter *adapter, return 0; } +/** + * ixgbe_negotiate_vf_features - negotiate supported features with VF driver + * @adapter: pointer to adapter struct + * @msgbuf: pointer to message buffers + * @vf: VF identifier + * + * Return: 0 on success or -EOPNOTSUPP when operation is not supported. + */ +static int ixgbe_negotiate_vf_features(struct ixgbe_adapter *adapter, + u32 *msgbuf, u32 vf) +{ + u32 features = msgbuf[1]; + + switch (adapter->vfinfo[vf].vf_api) { + case ixgbe_mbox_api_17: + break; + default: + return -EOPNOTSUPP; + } + + features &= IXGBE_SUPPORTED_FEATURES; + msgbuf[1] = features; + + return 0; +} + static int ixgbe_rcv_msg_from_vf(struct ixgbe_adapter *adapter, u32 vf) { u32 mbx_size = IXGBE_VFMAILBOX_SIZE; @@ -1370,6 +1404,9 @@ static int ixgbe_rcv_msg_from_vf(struct ixgbe_adapter *adapter, u32 vf) case IXGBE_VF_GET_PF_LINK_STATE: retval = ixgbe_send_vf_link_status(adapter, msgbuf, vf); break; + case IXGBE_VF_FEATURES_NEGOTIATE: + retval = ixgbe_negotiate_vf_features(adapter, msgbuf, vf); + break; default: e_err(drv, "Unhandled Msg %8.8x\n", msgbuf[0]); retval = -EIO; From 5feef67b646d8f5064bac288e22204ffba2b9a4a Mon Sep 17 00:00:00 2001 From: Koichiro Den Date: Thu, 9 Oct 2025 17:03:51 -0700 Subject: [PATCH 27/50] ixgbe: fix too early devlink_free() in ixgbe_remove() Since ixgbe_adapter is embedded in devlink, calling devlink_free() prematurely in the ixgbe_remove() path can lead to UAF. Move devlink_free() to the end. KASAN report: BUG: KASAN: use-after-free in ixgbe_reset_interrupt_capability+0x140/0x180 [ixgbe] Read of size 8 at addr ffff0000adf813e0 by task bash/2095 CPU: 1 UID: 0 PID: 2095 Comm: bash Tainted: G S 6.17.0-rc2-tnguy.net-queue+ #1 PREEMPT(full) [...] Call trace: show_stack+0x30/0x90 (C) dump_stack_lvl+0x9c/0xd0 print_address_description.constprop.0+0x90/0x310 print_report+0x104/0x1f0 kasan_report+0x88/0x180 __asan_report_load8_noabort+0x20/0x30 ixgbe_reset_interrupt_capability+0x140/0x180 [ixgbe] ixgbe_clear_interrupt_scheme+0xf8/0x130 [ixgbe] ixgbe_remove+0x2d0/0x8c0 [ixgbe] pci_device_remove+0xa0/0x220 device_remove+0xb8/0x170 device_release_driver_internal+0x318/0x490 device_driver_detach+0x40/0x68 unbind_store+0xec/0x118 drv_attr_store+0x64/0xb8 sysfs_kf_write+0xcc/0x138 kernfs_fop_write_iter+0x294/0x440 new_sync_write+0x1fc/0x588 vfs_write+0x480/0x6a0 ksys_write+0xf0/0x1e0 __arm64_sys_write+0x70/0xc0 invoke_syscall.constprop.0+0xcc/0x280 el0_svc_common.constprop.0+0xa8/0x248 do_el0_svc+0x44/0x68 el0_svc+0x54/0x160 el0t_64_sync_handler+0xa0/0xe8 el0t_64_sync+0x1b0/0x1b8 Fixes: a0285236ab93 ("ixgbe: add initial devlink support") Signed-off-by: Koichiro Den Tested-by: Rinitha S Reviewed-by: Jedrzej Jagielski Reviewed-by: Aleksandr Loktionov Reviewed-by: Paul Menzel Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20251009-jk-iwl-net-2025-10-01-v3-6-ef32a425b92a@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index 90d4e57b1c93..ca1ccc630001 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -12101,7 +12101,6 @@ static void ixgbe_remove(struct pci_dev *pdev) devl_port_unregister(&adapter->devlink_port); devl_unlock(adapter->devlink); - devlink_free(adapter->devlink); ixgbe_stop_ipsec_offload(adapter); ixgbe_clear_interrupt_scheme(adapter); @@ -12137,6 +12136,8 @@ static void ixgbe_remove(struct pci_dev *pdev) if (disable_dev) pci_disable_device(pdev); + + devlink_free(adapter->devlink); } /** From 2c67301584f2671e320236df6bbe75ae09feb4d0 Mon Sep 17 00:00:00 2001 From: Marek Vasut Date: Sat, 11 Oct 2025 13:02:49 +0200 Subject: [PATCH 28/50] net: phy: realtek: Avoid PHYCR2 access if PHYCR2 not present The driver is currently checking for PHYCR2 register presence in rtl8211f_config_init(), but it does so after accessing PHYCR2 to disable EEE. This was introduced in commit bfc17c165835 ("net: phy: realtek: disable PHY-mode EEE"). Move the PHYCR2 presence test before the EEE disablement and simplify the code. Fixes: bfc17c165835 ("net: phy: realtek: disable PHY-mode EEE") Signed-off-by: Marek Vasut Reviewed-by: Maxime Chevallier Reviewed-by: Russell King (Oracle) Link: https://patch.msgid.link/20251011110309.12664-1-marek.vasut@mailbox.org Signed-off-by: Jakub Kicinski --- drivers/net/phy/realtek/realtek_main.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) diff --git a/drivers/net/phy/realtek/realtek_main.c b/drivers/net/phy/realtek/realtek_main.c index 82d8e1335215..a724b21b4fe7 100644 --- a/drivers/net/phy/realtek/realtek_main.c +++ b/drivers/net/phy/realtek/realtek_main.c @@ -633,26 +633,25 @@ static int rtl8211f_config_init(struct phy_device *phydev) str_enabled_disabled(val_rxdly)); } + if (!priv->has_phycr2) + return 0; + /* Disable PHY-mode EEE so LPI is passed to the MAC */ ret = phy_modify_paged(phydev, RTL8211F_PHYCR_PAGE, RTL8211F_PHYCR2, RTL8211F_PHYCR2_PHY_EEE_ENABLE, 0); if (ret) return ret; - if (priv->has_phycr2) { - ret = phy_modify_paged(phydev, RTL8211F_PHYCR_PAGE, - RTL8211F_PHYCR2, RTL8211F_CLKOUT_EN, - priv->phycr2); - if (ret < 0) { - dev_err(dev, "clkout configuration failed: %pe\n", - ERR_PTR(ret)); - return ret; - } - - return genphy_soft_reset(phydev); + ret = phy_modify_paged(phydev, RTL8211F_PHYCR_PAGE, + RTL8211F_PHYCR2, RTL8211F_CLKOUT_EN, + priv->phycr2); + if (ret < 0) { + dev_err(dev, "clkout configuration failed: %pe\n", + ERR_PTR(ret)); + return ret; } - return 0; + return genphy_soft_reset(phydev); } static int rtl821x_suspend(struct phy_device *phydev) From 2616222e423398bb374ffcb5d23dea4ba2c3e524 Mon Sep 17 00:00:00 2001 From: Raju Rangoju Date: Fri, 10 Oct 2025 12:21:42 +0530 Subject: [PATCH 29/50] amd-xgbe: Avoid spurious link down messages during interface toggle During interface toggle operations (ifdown/ifup), the driver currently resets the local helper variable 'phy_link' to -1. This causes the link state machine to incorrectly interpret the state as a link change event, resulting in spurious "Link is down" messages being logged when the interface is brought back up. Preserve the phy_link state across interface toggles to avoid treating the -1 sentinel value as a legitimate link state transition. Fixes: 88131a812b16 ("amd-xgbe: Perform phy connect/disconnect at dev open/stop") Signed-off-by: Raju Rangoju Reviewed-by: Dawid Osuchowski Link: https://patch.msgid.link/20251010065142.1189310-1-Raju.Rangoju@amd.com Signed-off-by: Paolo Abeni --- drivers/net/ethernet/amd/xgbe/xgbe-drv.c | 1 - drivers/net/ethernet/amd/xgbe/xgbe-mdio.c | 1 + 2 files changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-drv.c b/drivers/net/ethernet/amd/xgbe/xgbe-drv.c index f0989aa01855..4dc631af7933 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-drv.c +++ b/drivers/net/ethernet/amd/xgbe/xgbe-drv.c @@ -1080,7 +1080,6 @@ static void xgbe_free_rx_data(struct xgbe_prv_data *pdata) static int xgbe_phy_reset(struct xgbe_prv_data *pdata) { - pdata->phy_link = -1; pdata->phy_speed = SPEED_UNKNOWN; return pdata->phy_if.phy_reset(pdata); diff --git a/drivers/net/ethernet/amd/xgbe/xgbe-mdio.c b/drivers/net/ethernet/amd/xgbe/xgbe-mdio.c index 1a37ec45e650..7675bb98f029 100644 --- a/drivers/net/ethernet/amd/xgbe/xgbe-mdio.c +++ b/drivers/net/ethernet/amd/xgbe/xgbe-mdio.c @@ -1555,6 +1555,7 @@ static int xgbe_phy_init(struct xgbe_prv_data *pdata) pdata->phy.duplex = DUPLEX_FULL; } + pdata->phy_link = 0; pdata->phy.link = 0; pdata->phy.pause_autoneg = pdata->pause_autoneg; From c5705a2a4aa35350e504b72a94b5c71c3754833c Mon Sep 17 00:00:00 2001 From: Harshit Mogalapalli Date: Fri, 10 Oct 2025 13:42:39 -0700 Subject: [PATCH 30/50] Octeontx2-af: Fix missing error code in cgx_probe() When CGX fails mapping to NIX, set the error code to -ENODEV, currently err is zero and that is treated as success path. Reported-by: Dan Carpenter Closes: https://lore.kernel.org/all/aLAdlCg2_Yv7Y-3h@stanley.mountain/ Fixes: d280233fc866 ("Octeontx2-af: Fix NIX X2P calibration failures") Signed-off-by: Harshit Mogalapalli Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251010204239.94237-1-harshit.m.mogalapalli@oracle.com Signed-off-by: Paolo Abeni --- drivers/net/ethernet/marvell/octeontx2/af/cgx.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c index d374a4454836..ec0e11c77cbf 100644 --- a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c +++ b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c @@ -1981,6 +1981,7 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id) !is_cgx_mapped_to_nix(pdev->subsystem_device, cgx->cgx_id)) { dev_notice(dev, "CGX %d not mapped to NIX, skipping probe\n", cgx->cgx_id); + err = -ENODEV; goto err_release_regions; } From 327cd4b68b4398b6c24f10eb2b2533ffbfc10185 Mon Sep 17 00:00:00 2001 From: Zqiang Date: Sat, 11 Oct 2025 15:05:18 +0800 Subject: [PATCH 31/50] usbnet: Fix using smp_processor_id() in preemptible code warnings Syzbot reported the following warning: BUG: using smp_processor_id() in preemptible [00000000] code: dhcpcd/2879 caller is usbnet_skb_return+0x74/0x490 drivers/net/usb/usbnet.c:331 CPU: 1 UID: 0 PID: 2879 Comm: dhcpcd Not tainted 6.15.0-rc4-syzkaller-00098-g615dca38c2ea #0 PREEMPT(voluntary) Call Trace: __dump_stack lib/dump_stack.c:94 [inline] dump_stack_lvl+0x16c/0x1f0 lib/dump_stack.c:120 check_preemption_disabled+0xd0/0xe0 lib/smp_processor_id.c:49 usbnet_skb_return+0x74/0x490 drivers/net/usb/usbnet.c:331 usbnet_resume_rx+0x4b/0x170 drivers/net/usb/usbnet.c:708 usbnet_change_mtu+0x1be/0x220 drivers/net/usb/usbnet.c:417 __dev_set_mtu net/core/dev.c:9443 [inline] netif_set_mtu_ext+0x369/0x5c0 net/core/dev.c:9496 netif_set_mtu+0xb0/0x160 net/core/dev.c:9520 dev_set_mtu+0xae/0x170 net/core/dev_api.c:247 dev_ifsioc+0xa31/0x18d0 net/core/dev_ioctl.c:572 dev_ioctl+0x223/0x10e0 net/core/dev_ioctl.c:821 sock_do_ioctl+0x19d/0x280 net/socket.c:1204 sock_ioctl+0x42f/0x6a0 net/socket.c:1311 vfs_ioctl fs/ioctl.c:51 [inline] __do_sys_ioctl fs/ioctl.c:906 [inline] __se_sys_ioctl fs/ioctl.c:892 [inline] __x64_sys_ioctl+0x190/0x200 fs/ioctl.c:892 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline] do_syscall_64+0xcd/0x260 arch/x86/entry/syscall_64.c:94 entry_SYSCALL_64_after_hwframe+0x77/0x7f For historical and portability reasons, the netif_rx() is usually run in the softirq or interrupt context, this commit therefore add local_bh_disable/enable() protection in the usbnet_resume_rx(). Fixes: 43daa96b166c ("usbnet: Stop RX Q on MTU change") Link: https://syzkaller.appspot.com/bug?id=81f55dfa587ee544baaaa5a359a060512228c1e1 Suggested-by: Jakub Kicinski Signed-off-by: Zqiang Link: https://patch.msgid.link/20251011070518.7095-1-qiang.zhang@linux.dev Signed-off-by: Paolo Abeni --- drivers/net/usb/usbnet.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c index 511c4154cf74..bf01f2728531 100644 --- a/drivers/net/usb/usbnet.c +++ b/drivers/net/usb/usbnet.c @@ -702,6 +702,7 @@ void usbnet_resume_rx(struct usbnet *dev) struct sk_buff *skb; int num = 0; + local_bh_disable(); clear_bit(EVENT_RX_PAUSED, &dev->flags); while ((skb = skb_dequeue(&dev->rxq_pause)) != NULL) { @@ -710,6 +711,7 @@ void usbnet_resume_rx(struct usbnet *dev) } queue_work(system_bh_wq, &dev->bh_work); + local_bh_enable(); netif_dbg(dev, rx_status, dev->net, "paused rx queue disabled, %d skbs requeued\n", num); From 75527d61d60d493d1eb064f335071a20ca581f54 Mon Sep 17 00:00:00 2001 From: Yi Cong Date: Sat, 11 Oct 2025 16:24:15 +0800 Subject: [PATCH 32/50] r8152: add error handling in rtl8152_driver_init rtl8152_driver_init() is missing the error handling. When rtl8152_driver registration fails, rtl8152_cfgselector_driver should be deregistered. Fixes: ec51fbd1b8a2 ("r8152: add USB device driver for config selection") Cc: stable@vger.kernel.org Signed-off-by: Yi Cong Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251011082415.580740-1-yicongsrfy@163.com [pabeni@redhat.com: clarified the commit message] Signed-off-by: Paolo Abeni --- drivers/net/usb/r8152.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c index 44cba7acfe7d..a22d4bb2cf3b 100644 --- a/drivers/net/usb/r8152.c +++ b/drivers/net/usb/r8152.c @@ -10122,7 +10122,12 @@ static int __init rtl8152_driver_init(void) ret = usb_register_device_driver(&rtl8152_cfgselector_driver, THIS_MODULE); if (ret) return ret; - return usb_register(&rtl8152_driver); + + ret = usb_register(&rtl8152_driver); + if (ret) + usb_deregister_device_driver(&rtl8152_cfgselector_driver); + + return ret; } static void __exit rtl8152_driver_exit(void) From 295ce1eb36ae47dc862d6c8a1012618a25516208 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Sat, 11 Oct 2025 11:57:42 +0000 Subject: [PATCH 33/50] tcp: fix tcp_tso_should_defer() vs large RTT Neal reported that using neper tcp_stream with TCP_TX_DELAY set to 50ms would often lead to flows stuck in a small cwnd mode, regardless of the congestion control. While tcp_stream sets TCP_TX_DELAY too late after the connect(), it highlighted two kernel bugs. The following heuristic in tcp_tso_should_defer() seems wrong for large RTT: delta = tp->tcp_clock_cache - head->tstamp; /* If next ACK is likely to come too late (half srtt), do not defer */ if ((s64)(delta - (u64)NSEC_PER_USEC * (tp->srtt_us >> 4)) < 0) goto send_now; If next ACK is expected to come in more than 1 ms, we should not defer because we prefer a smooth ACK clocking. While blamed commit was a step in the good direction, it was not generic enough. Another patch fixing TCP_TX_DELAY for established flows will be proposed when net-next reopens. Fixes: 50c8339e9299 ("tcp: tso: restore IW10 after TSO autosizing") Reported-by: Neal Cardwell Signed-off-by: Eric Dumazet Reviewed-by: Neal Cardwell Tested-by: Neal Cardwell Link: https://patch.msgid.link/20251011115742.1245771-1-edumazet@google.com [pabeni@redhat.com: fixed whitespace issue] Signed-off-by: Paolo Abeni --- net/ipv4/tcp_output.c | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-) diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index bb3576ac0ad7..b94efb3050d2 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -2369,7 +2369,8 @@ static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb, u32 max_segs) { const struct inet_connection_sock *icsk = inet_csk(sk); - u32 send_win, cong_win, limit, in_flight; + u32 send_win, cong_win, limit, in_flight, threshold; + u64 srtt_in_ns, expected_ack, how_far_is_the_ack; struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *head; int win_divisor; @@ -2431,9 +2432,19 @@ static bool tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb, head = tcp_rtx_queue_head(sk); if (!head) goto send_now; - delta = tp->tcp_clock_cache - head->tstamp; - /* If next ACK is likely to come too late (half srtt), do not defer */ - if ((s64)(delta - (u64)NSEC_PER_USEC * (tp->srtt_us >> 4)) < 0) + + srtt_in_ns = (u64)(NSEC_PER_USEC >> 3) * tp->srtt_us; + /* When is the ACK expected ? */ + expected_ack = head->tstamp + srtt_in_ns; + /* How far from now is the ACK expected ? */ + how_far_is_the_ack = expected_ack - tp->tcp_clock_cache; + + /* If next ACK is likely to come too late, + * ie in more than min(1ms, half srtt), do not defer. + */ + threshold = min(srtt_in_ns >> 1, NSEC_PER_MSEC); + + if ((s64)(how_far_is_the_ack - threshold) > 0) goto send_now; /* Ok, it looks like it is advisable to defer. From bd5afca115f181c85f992d42a57cd497bc823ccb Mon Sep 17 00:00:00 2001 From: Lorenzo Bianconi Date: Sun, 12 Oct 2025 11:19:44 +0200 Subject: [PATCH 34/50] net: airoha: Take into account out-of-order tx completions in airoha_dev_xmit() Completion napi can free out-of-order tx descriptors if hw QoS is enabled and packets with different priority are queued to same DMA ring. Take into account possible out-of-order reports checking if the tx queue is full using circular buffer head/tail pointer instead of the number of queued packets. Fixes: 23020f0493270 ("net: airoha: Introduce ethernet support for EN7581 SoC") Suggested-by: Simon Horman Signed-off-by: Lorenzo Bianconi Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251012-airoha-tx-busy-queue-v2-1-a600b08bab2d@kernel.org Signed-off-by: Paolo Abeni --- drivers/net/ethernet/airoha/airoha_eth.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c index 833dd911980b..433a646e9831 100644 --- a/drivers/net/ethernet/airoha/airoha_eth.c +++ b/drivers/net/ethernet/airoha/airoha_eth.c @@ -1873,6 +1873,20 @@ static u32 airoha_get_dsa_tag(struct sk_buff *skb, struct net_device *dev) #endif } +static bool airoha_dev_tx_queue_busy(struct airoha_queue *q, u32 nr_frags) +{ + u32 tail = q->tail <= q->head ? q->tail + q->ndesc : q->tail; + u32 index = q->head + nr_frags; + + /* completion napi can free out-of-order tx descriptors if hw QoS is + * enabled and packets with different priorities are queued to the same + * DMA ring. Take into account possible out-of-order reports checking + * if the tx queue is full using circular buffer head/tail pointers + * instead of the number of queued packets. + */ + return index >= tail; +} + static netdev_tx_t airoha_dev_xmit(struct sk_buff *skb, struct net_device *dev) { @@ -1926,7 +1940,7 @@ static netdev_tx_t airoha_dev_xmit(struct sk_buff *skb, txq = netdev_get_tx_queue(dev, qid); nr_frags = 1 + skb_shinfo(skb)->nr_frags; - if (q->queued + nr_frags > q->ndesc) { + if (airoha_dev_tx_queue_busy(q, nr_frags)) { /* not enough space in the queue */ netif_tx_stop_queue(txq); spin_unlock_bh(&q->lock); From 4f86eb0a38bc719ba966f155071a6f0594327f34 Mon Sep 17 00:00:00 2001 From: Wang Liang Date: Mon, 13 Oct 2025 16:00:39 +0800 Subject: [PATCH 35/50] selftests: net: check jq command is supported The jq command is used in vlan_bridge_binding.sh, if it is not supported, the test will spam the following log. # ./vlan_bridge_binding.sh: line 51: jq: command not found # ./vlan_bridge_binding.sh: line 51: jq: command not found # ./vlan_bridge_binding.sh: line 51: jq: command not found # ./vlan_bridge_binding.sh: line 51: jq: command not found # ./vlan_bridge_binding.sh: line 51: jq: command not found # TEST: Test bridge_binding on->off when lower down [FAIL] # Got operstate of , expected 0 The rtnetlink.sh has the same problem. It makes sense to check if jq is installed before running these tests. After this patch, the vlan_bridge_binding.sh skipped if jq is not supported: # timeout set to 3600 # selftests: net: vlan_bridge_binding.sh # TEST: jq not installed [SKIP] Fixes: dca12e9ab760 ("selftests: net: Add a VLAN bridge binding selftest") Fixes: 6a414fd77f61 ("selftests: rtnetlink: Add an address proto test") Signed-off-by: Wang Liang Reviewed-by: Hangbin Liu Link: https://patch.msgid.link/20251013080039.3035898-1-wangliang74@huawei.com Signed-off-by: Paolo Abeni --- tools/testing/selftests/net/rtnetlink.sh | 2 ++ tools/testing/selftests/net/vlan_bridge_binding.sh | 2 ++ 2 files changed, 4 insertions(+) diff --git a/tools/testing/selftests/net/rtnetlink.sh b/tools/testing/selftests/net/rtnetlink.sh index dbf77513f617..163a084d525d 100755 --- a/tools/testing/selftests/net/rtnetlink.sh +++ b/tools/testing/selftests/net/rtnetlink.sh @@ -1466,6 +1466,8 @@ usage: ${0##*/} OPTS EOF } +require_command jq + #check for needed privileges if [ "$(id -u)" -ne 0 ];then end_test "SKIP: Need root privileges" diff --git a/tools/testing/selftests/net/vlan_bridge_binding.sh b/tools/testing/selftests/net/vlan_bridge_binding.sh index db481af9b6b3..e8c02c64e03a 100755 --- a/tools/testing/selftests/net/vlan_bridge_binding.sh +++ b/tools/testing/selftests/net/vlan_bridge_binding.sh @@ -249,6 +249,8 @@ test_binding_toggle_off_when_upper_down() do_test_binding_off : "on->off when upper down" } +require_command jq + trap defer_scopes_cleanup EXIT setup_prepare tests_run From df5a1f4aeb6ff5e7c5ac47d16a347f03509dd441 Mon Sep 17 00:00:00 2001 From: Jonas Gorski Date: Mon, 13 Oct 2025 20:03:47 +0200 Subject: [PATCH 36/50] MAINTAINERS: add myself as maintainer for b53 I wrote the original OpenWrt driver that Florian used as the base for the dsa driver, I might as well take responsibility for it. Signed-off-by: Jonas Gorski Reviewed-by: Florian Fainelli Reviewed-by: Andrew Lunn Reviewed-by: Vladimir Oltean Link: https://patch.msgid.link/20251013180347.133246-1-jonas.gorski@gmail.com Signed-off-by: Jakub Kicinski --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index 3a27901781c2..8a213950e37e 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4804,6 +4804,7 @@ F: drivers/net/ethernet/broadcom/b44.* BROADCOM B53/SF2 ETHERNET SWITCH DRIVER M: Florian Fainelli +M: Jonas Gorski L: netdev@vger.kernel.org L: openwrt-devel@lists.openwrt.org (subscribers-only) S: Supported From 7f0fddd817ba6daebea1445ae9fab4b6d2294fa8 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Mon, 13 Oct 2025 20:50:52 +0200 Subject: [PATCH 37/50] net: core: fix lockdep splat on device unregister Since blamed commit, unregister_netdevice_many_notify() takes the netdev mutex if the device needs it. If the device list is too long, this will lock more device mutexes than lockdep can handle: unshare -n \ bash -c 'for i in $(seq 1 100);do ip link add foo$i type dummy;done' BUG: MAX_LOCK_DEPTH too low! turning off the locking correctness validator. depth: 48 max: 48! 48 locks held by kworker/u16:1/69: #0: ..148 ((wq_completion)netns){+.+.}-{0:0}, at: process_one_work #1: ..d40 (net_cleanup_work){+.+.}-{0:0}, at: process_one_work #2: ..bd0 (pernet_ops_rwsem){++++}-{4:4}, at: cleanup_net #3: ..aa8 (rtnl_mutex){+.+.}-{4:4}, at: default_device_exit_batch #4: ..cb0 (&dev_instance_lock_key#3){+.+.}-{4:4}, at: unregister_netdevice_many_notify [..] Add a helper to close and then unlock a list of net_devices. Devices that are not up have to be skipped - netif_close_many always removes them from the list without any other actions taken, so they'd remain in locked state. Close devices whenever we've used up half of the tracking slots or we processed entire list without hitting the limit. Fixes: 7e4d784f5810 ("net: hold netdev instance lock during rtnetlink operations") Signed-off-by: Florian Westphal Link: https://patch.msgid.link/20251013185052.14021-1-fw@strlen.de Signed-off-by: Jakub Kicinski --- net/core/dev.c | 40 +++++++++++++++++++++++++++++++++++----- 1 file changed, 35 insertions(+), 5 deletions(-) diff --git a/net/core/dev.c b/net/core/dev.c index a64cef2c537e..2acfa44927da 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -12176,6 +12176,35 @@ static void dev_memory_provider_uninstall(struct net_device *dev) } } +/* devices must be UP and netdev_lock()'d */ +static void netif_close_many_and_unlock(struct list_head *close_head) +{ + struct net_device *dev, *tmp; + + netif_close_many(close_head, false); + + /* ... now unlock them */ + list_for_each_entry_safe(dev, tmp, close_head, close_list) { + netdev_unlock(dev); + list_del_init(&dev->close_list); + } +} + +static void netif_close_many_and_unlock_cond(struct list_head *close_head) +{ +#ifdef CONFIG_LOCKDEP + /* We can only track up to MAX_LOCK_DEPTH locks per task. + * + * Reserve half the available slots for additional locks possibly + * taken by notifiers and (soft)irqs. + */ + unsigned int limit = MAX_LOCK_DEPTH / 2; + + if (lockdep_depth(current) > limit) + netif_close_many_and_unlock(close_head); +#endif +} + void unregister_netdevice_many_notify(struct list_head *head, u32 portid, const struct nlmsghdr *nlh) { @@ -12208,17 +12237,18 @@ void unregister_netdevice_many_notify(struct list_head *head, /* If device is running, close it first. Start with ops locked... */ list_for_each_entry(dev, head, unreg_list) { + if (!(dev->flags & IFF_UP)) + continue; if (netdev_need_ops_lock(dev)) { list_add_tail(&dev->close_list, &close_head); netdev_lock(dev); } + netif_close_many_and_unlock_cond(&close_head); } - netif_close_many(&close_head, true); - /* ... now unlock them and go over the rest. */ + netif_close_many_and_unlock(&close_head); + /* ... now go over the rest. */ list_for_each_entry(dev, head, unreg_list) { - if (netdev_need_ops_lock(dev)) - netdev_unlock(dev); - else + if (!netdev_need_ops_lock(dev)) list_add_tail(&dev->close_list, &close_head); } netif_close_many(&close_head, true); From bfdd74166a639930baaba27a8d729edaacd46907 Mon Sep 17 00:00:00 2001 From: Tim Hostetler Date: Tue, 14 Oct 2025 00:47:39 +0000 Subject: [PATCH 38/50] gve: Check valid ts bit on RX descriptor before hw timestamping The device returns a valid bit in the LSB of the low timestamp byte in the completion descriptor that the driver should check before setting the SKB's hardware timestamp. If the timestamp is not valid, do not hardware timestamp the SKB. Cc: stable@vger.kernel.org Fixes: b2c7aeb49056 ("gve: Implement ndo_hwtstamp_get/set for RX timestamping") Reviewed-by: Joshua Washington Signed-off-by: Tim Hostetler Signed-off-by: Harshitha Ramamurthy Reviewed-by: Simon Horman Reviewed-by: Willem de Bruijn Reviewed-by: Vadim Fedorenko Link: https://patch.msgid.link/20251014004740.2775957-1-hramamurthy@google.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/google/gve/gve.h | 2 ++ drivers/net/ethernet/google/gve/gve_desc_dqo.h | 3 ++- drivers/net/ethernet/google/gve/gve_rx_dqo.c | 16 +++++++++++----- 3 files changed, 15 insertions(+), 6 deletions(-) diff --git a/drivers/net/ethernet/google/gve/gve.h b/drivers/net/ethernet/google/gve/gve.h index bceaf9b05cb4..4cc6dcbfd367 100644 --- a/drivers/net/ethernet/google/gve/gve.h +++ b/drivers/net/ethernet/google/gve/gve.h @@ -100,6 +100,8 @@ */ #define GVE_DQO_QPL_ONDEMAND_ALLOC_THRESHOLD 96 +#define GVE_DQO_RX_HWTSTAMP_VALID 0x1 + /* Each slot in the desc ring has a 1:1 mapping to a slot in the data ring */ struct gve_rx_desc_queue { struct gve_rx_desc *desc_ring; /* the descriptor ring */ diff --git a/drivers/net/ethernet/google/gve/gve_desc_dqo.h b/drivers/net/ethernet/google/gve/gve_desc_dqo.h index d17da841b5a0..f7786b03c744 100644 --- a/drivers/net/ethernet/google/gve/gve_desc_dqo.h +++ b/drivers/net/ethernet/google/gve/gve_desc_dqo.h @@ -236,7 +236,8 @@ struct gve_rx_compl_desc_dqo { u8 status_error1; - __le16 reserved5; + u8 reserved5; + u8 ts_sub_nsecs_low; __le16 buf_id; /* Buffer ID which was sent on the buffer queue. */ union { diff --git a/drivers/net/ethernet/google/gve/gve_rx_dqo.c b/drivers/net/ethernet/google/gve/gve_rx_dqo.c index 55393b784317..1aff3bbb8cfc 100644 --- a/drivers/net/ethernet/google/gve/gve_rx_dqo.c +++ b/drivers/net/ethernet/google/gve/gve_rx_dqo.c @@ -456,14 +456,20 @@ static void gve_rx_skb_hash(struct sk_buff *skb, * Note that this means if the time delta between packet reception and the last * clock read is greater than ~2 seconds, this will provide invalid results. */ -static void gve_rx_skb_hwtstamp(struct gve_rx_ring *rx, u32 hwts) +static void gve_rx_skb_hwtstamp(struct gve_rx_ring *rx, + const struct gve_rx_compl_desc_dqo *desc) { u64 last_read = READ_ONCE(rx->gve->last_sync_nic_counter); struct sk_buff *skb = rx->ctx.skb_head; - u32 low = (u32)last_read; - s32 diff = hwts - low; + u32 ts, low; + s32 diff; - skb_hwtstamps(skb)->hwtstamp = ns_to_ktime(last_read + diff); + if (desc->ts_sub_nsecs_low & GVE_DQO_RX_HWTSTAMP_VALID) { + ts = le32_to_cpu(desc->ts); + low = (u32)last_read; + diff = ts - low; + skb_hwtstamps(skb)->hwtstamp = ns_to_ktime(last_read + diff); + } } static void gve_rx_free_skb(struct napi_struct *napi, struct gve_rx_ring *rx) @@ -944,7 +950,7 @@ static int gve_rx_complete_skb(struct gve_rx_ring *rx, struct napi_struct *napi, gve_rx_skb_csum(rx->ctx.skb_head, desc, ptype); if (rx->gve->ts_config.rx_filter == HWTSTAMP_FILTER_ALL) - gve_rx_skb_hwtstamp(rx, le32_to_cpu(desc->ts)); + gve_rx_skb_hwtstamp(rx, desc); /* RSC packets must set gso_size otherwise the TCP stack will complain * that packets are larger than MTU. From bc384963bc18e4f21cf8615b57cbbc9c5e0d309a Mon Sep 17 00:00:00 2001 From: Justin Iurman Date: Tue, 14 Oct 2025 19:06:50 +0200 Subject: [PATCH 39/50] MAINTAINERS: new entry for IPv6 IOAM Create a maintainer entry for IPv6 IOAM. Add myself as I authored most if not all of the IPv6 IOAM code in the kernel and actively participate in the related IETF groups. Signed-off-by: Justin Iurman Link: https://patch.msgid.link/20251014170650.27679-1-justin.iurman@uliege.be Signed-off-by: Jakub Kicinski --- MAINTAINERS | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index 8a213950e37e..eba5e091a086 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -18012,6 +18012,16 @@ X: net/rfkill/ X: net/wireless/ X: tools/testing/selftests/net/can/ +NETWORKING [IOAM] +M: Justin Iurman +S: Maintained +F: Documentation/networking/ioam6* +F: include/linux/ioam6* +F: include/net/ioam6* +F: include/uapi/linux/ioam6* +F: net/ipv6/ioam6* +F: tools/testing/selftests/net/ioam6* + NETWORKING [IPSEC] M: Steffen Klassert M: Herbert Xu From 0c3f2e62815a43628e748b1e4ad97a1c46cce703 Mon Sep 17 00:00:00 2001 From: Alexey Simakov Date: Tue, 14 Oct 2025 19:47:38 +0300 Subject: [PATCH 40/50] tg3: prevent use of uninitialized remote_adv and local_adv variables Some execution paths that jump to the fiber_setup_done label could leave the remote_adv and local_adv variables uninitialized and then use it. Initialize this variables at the point of definition to avoid this. Fixes: 85730a631f0c ("tg3: Add SGMII phy support for 5719/5718 serdes") Co-developed-by: Alexandr Sapozhnikov Signed-off-by: Alexandr Sapozhnikov Signed-off-by: Alexey Simakov Reviewed-by: Pavan Chebbi Link: https://patch.msgid.link/20251014164736.5890-1-bigalex934@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/broadcom/tg3.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/net/ethernet/broadcom/tg3.c b/drivers/net/ethernet/broadcom/tg3.c index 7f00ec7fd7b9..d78cafdb2094 100644 --- a/drivers/net/ethernet/broadcom/tg3.c +++ b/drivers/net/ethernet/broadcom/tg3.c @@ -5803,7 +5803,7 @@ static int tg3_setup_fiber_mii_phy(struct tg3 *tp, bool force_reset) u32 current_speed = SPEED_UNKNOWN; u8 current_duplex = DUPLEX_UNKNOWN; bool current_link_up = false; - u32 local_adv, remote_adv, sgsr; + u32 local_adv = 0, remote_adv = 0, sgsr; if ((tg3_asic_rev(tp) == ASIC_REV_5719 || tg3_asic_rev(tp) == ASIC_REV_5720) && @@ -5944,9 +5944,6 @@ static int tg3_setup_fiber_mii_phy(struct tg3 *tp, bool force_reset) else current_duplex = DUPLEX_HALF; - local_adv = 0; - remote_adv = 0; - if (bmcr & BMCR_ANENABLE) { u32 common; From ce5af41e3234425a40974696682163edfd21128c Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Tue, 14 Oct 2025 11:16:56 +0200 Subject: [PATCH 41/50] tls: trim encrypted message to match the plaintext on short splice During tls_sw_sendmsg_locked, we pre-allocate the encrypted message for the size we're expecting to send during the current iteration, but we may end up sending less, for example when splicing: if we're getting the data from small fragments of memory, we may fill up all the slots in the skmsg with less data than expected. In this case, we need to trim the encrypted message to only the length we actually need, to avoid pushing uninitialized bytes down the underlying TCP socket. Fixes: fe1e81d4f73b ("tls/sw: Support MSG_SPLICE_PAGES") Reported-by: Jann Horn Signed-off-by: Sabrina Dubroca Link: https://patch.msgid.link/66a0ae99c9efc15f88e9e56c1f58f902f442ce86.1760432043.git.sd@queasysnail.net Signed-off-by: Jakub Kicinski --- net/tls/tls_sw.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index daac9fd4be7e..36ca3011ab87 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -1112,8 +1112,11 @@ alloc_encrypted: goto send_end; tls_ctx->pending_open_record_frags = true; - if (sk_msg_full(msg_pl)) + if (sk_msg_full(msg_pl)) { full_record = true; + sk_msg_trim(sk, msg_en, + msg_pl->sg.size + prot->overhead_size); + } if (full_record || eor) goto copied; From b014a4e066c555185b7c367efacdc33f16695495 Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Tue, 14 Oct 2025 11:16:57 +0200 Subject: [PATCH 42/50] tls: wait for async encrypt in case of error during latter iterations of sendmsg If we hit an error during the main loop of tls_sw_sendmsg_locked (eg failed allocation), we jump to send_end and immediately return. Previous iterations may have queued async encryption requests that are still pending. We should wait for those before returning, as we could otherwise be reading from memory that userspace believes we're not using anymore, which would be a sort of use-after-free. This is similar to what tls_sw_recvmsg already does: failures during the main loop jump to the "wait for async" code, not straight to the unlock/return. Fixes: a42055e8d2c3 ("net/tls: Add support for async encryption of records for performance") Reported-by: Jann Horn Signed-off-by: Sabrina Dubroca Link: https://patch.msgid.link/c793efe9673b87f808d84fdefc0f732217030c52.1760432043.git.sd@queasysnail.net Signed-off-by: Jakub Kicinski --- net/tls/tls_sw.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index 36ca3011ab87..1478d515badc 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -1054,7 +1054,7 @@ static int tls_sw_sendmsg_locked(struct sock *sk, struct msghdr *msg, if (ret == -EINPROGRESS) num_async++; else if (ret != -EAGAIN) - goto send_end; + goto end; } } @@ -1226,8 +1226,9 @@ trim_sgl: goto alloc_encrypted; } +send_end: if (!num_async) { - goto send_end; + goto end; } else if (num_zc || eor) { int err; @@ -1245,7 +1246,7 @@ trim_sgl: tls_tx_records(sk, msg->msg_flags); } -send_end: +end: ret = sk_stream_error(sk, msg->msg_flags, ret); return copied > 0 ? copied : ret; } From b6fe4c29bb51cf239ecf48eacf72b924565cb619 Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Tue, 14 Oct 2025 11:16:58 +0200 Subject: [PATCH 43/50] tls: always set record_type in tls_process_cmsg When userspace wants to send a non-DATA record (via the TLS_SET_RECORD_TYPE cmsg), we need to send any pending data from a previous MSG_MORE send() as a separate DATA record. If that DATA record is encrypted asynchronously, tls_handle_open_record will return -EINPROGRESS. This is currently treated as an error by tls_process_cmsg, and it will skip setting record_type to the correct value, but the caller (tls_sw_sendmsg_locked) handles that return value correctly and proceeds with sending the new message with an incorrect record_type (DATA instead of whatever was requested in the cmsg). Always set record_type before handling the open record. If tls_handle_open_record returns an error, record_type will be ignored. If it succeeds, whether with synchronous crypto (returning 0) or asynchronous (returning -EINPROGRESS), the caller will proceed correctly. Fixes: a42055e8d2c3 ("net/tls: Add support for async encryption of records for performance") Reported-by: Jann Horn Signed-off-by: Sabrina Dubroca Link: https://patch.msgid.link/0457252e578a10a94e40c72ba6288b3a64f31662.1760432043.git.sd@queasysnail.net Signed-off-by: Jakub Kicinski --- net/tls/tls_main.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c index a3ccb3135e51..39a2ab47fe72 100644 --- a/net/tls/tls_main.c +++ b/net/tls/tls_main.c @@ -255,12 +255,9 @@ int tls_process_cmsg(struct sock *sk, struct msghdr *msg, if (msg->msg_flags & MSG_MORE) return -EINVAL; - rc = tls_handle_open_record(sk, msg->msg_flags); - if (rc) - return rc; - *record_type = *(unsigned char *)CMSG_DATA(cmsg); - rc = 0; + + rc = tls_handle_open_record(sk, msg->msg_flags); break; default: return -EINVAL; From b8a6ff84abbcbbc445463de58704686011edc8e1 Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Tue, 14 Oct 2025 11:16:59 +0200 Subject: [PATCH 44/50] tls: wait for pending async decryptions if tls_strp_msg_hold fails Async decryption calls tls_strp_msg_hold to create a clone of the input skb to hold references to the memory it uses. If we fail to allocate that clone, proceeding with async decryption can lead to various issues (UAF on the skb, writing into userspace memory after the recv() call has returned). In this case, wait for all pending decryption requests. Fixes: 84c61fe1a75b ("tls: rx: do not use the standard strparser") Reported-by: Jann Horn Signed-off-by: Sabrina Dubroca Link: https://patch.msgid.link/b9fe61dcc07dab15da9b35cf4c7d86382a98caf2.1760432043.git.sd@queasysnail.net Signed-off-by: Jakub Kicinski --- net/tls/tls_sw.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index 1478d515badc..e3d852091e7a 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -1641,8 +1641,10 @@ static int tls_decrypt_sg(struct sock *sk, struct iov_iter *out_iov, if (unlikely(darg->async)) { err = tls_strp_msg_hold(&ctx->strp, &ctx->async_hold); - if (err) - __skb_queue_tail(&ctx->async_hold, darg->skb); + if (err) { + err = tls_decrypt_async_wait(ctx); + darg->async = false; + } return err; } From 7f846c65ca11e63d2409868ff039081f80e42ae4 Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Tue, 14 Oct 2025 11:17:00 +0200 Subject: [PATCH 45/50] tls: don't rely on tx_work during send() With async crypto, we rely on tx_work to actually transmit records once encryption completes. But while send() is running, both the tx_lock and socket lock are held, so tx_work_handler cannot process the queue of encrypted records, and simply reschedules itself. During a large send(), this could last a long time, and use a lot of memory. Transmit any pending encrypted records before restarting the main loop of tls_sw_sendmsg_locked. Fixes: a42055e8d2c3 ("net/tls: Add support for async encryption of records for performance") Reported-by: Jann Horn Signed-off-by: Sabrina Dubroca Link: https://patch.msgid.link/8396631478f70454b44afb98352237d33f48d34d.1760432043.git.sd@queasysnail.net Signed-off-by: Jakub Kicinski --- net/tls/tls_sw.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index e3d852091e7a..d17135369980 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -1152,6 +1152,13 @@ alloc_encrypted: } else if (ret != -EAGAIN) goto send_end; } + + /* Transmit if any encryptions have completed */ + if (test_and_clear_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask)) { + cancel_delayed_work(&ctx->tx_work.work); + tls_tx_records(sk, msg->msg_flags); + } + continue; rollback_iter: copied -= try_to_copy; @@ -1207,6 +1214,12 @@ copied: goto send_end; } } + + /* Transmit if any encryptions have completed */ + if (test_and_clear_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask)) { + cancel_delayed_work(&ctx->tx_work.work); + tls_tx_records(sk, msg->msg_flags); + } } continue; From f95fce1e953b8a7af3fbad84aaffe92804196e2d Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Tue, 14 Oct 2025 11:17:01 +0200 Subject: [PATCH 46/50] selftests: net: tls: add tests for cmsg vs MSG_MORE We don't have a test to check that MSG_MORE won't let us merge records of different types across sendmsg calls. Add new tests that check: - MSG_MORE is only allowed for DATA records - a pending DATA record gets closed and pushed before a non-DATA record is processed Signed-off-by: Sabrina Dubroca Link: https://patch.msgid.link/b34feeadefe8a997f068d5ed5617afd0072df3c0.1760432043.git.sd@queasysnail.net Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/tls.c | 34 +++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+) diff --git a/tools/testing/selftests/net/tls.c b/tools/testing/selftests/net/tls.c index e788b84551ca..77e4e30b46cc 100644 --- a/tools/testing/selftests/net/tls.c +++ b/tools/testing/selftests/net/tls.c @@ -564,6 +564,40 @@ TEST_F(tls, msg_more) EXPECT_EQ(memcmp(buf, test_str, send_len), 0); } +TEST_F(tls, cmsg_msg_more) +{ + char *test_str = "test_read"; + char record_type = 100; + int send_len = 10; + + /* we don't allow MSG_MORE with non-DATA records */ + EXPECT_EQ(tls_send_cmsg(self->fd, record_type, test_str, send_len, + MSG_MORE), -1); + EXPECT_EQ(errno, EINVAL); +} + +TEST_F(tls, msg_more_then_cmsg) +{ + char *test_str = "test_read"; + char record_type = 100; + int send_len = 10; + char buf[10 * 2]; + int ret; + + EXPECT_EQ(send(self->fd, test_str, send_len, MSG_MORE), send_len); + EXPECT_EQ(recv(self->cfd, buf, send_len, MSG_DONTWAIT), -1); + + ret = tls_send_cmsg(self->fd, record_type, test_str, send_len, 0); + EXPECT_EQ(ret, send_len); + + /* initial DATA record didn't get merged with the non-DATA record */ + EXPECT_EQ(recv(self->cfd, buf, send_len * 2, 0), send_len); + + EXPECT_EQ(tls_recv_cmsg(_metadata, self->cfd, record_type, + buf, sizeof(buf), MSG_WAITALL), + send_len); +} + TEST_F(tls, msg_more_unsent) { char const *test_str = "test_read"; From 3667e9b442b95b021189db793b9156552f918e99 Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Tue, 14 Oct 2025 11:17:02 +0200 Subject: [PATCH 47/50] selftests: tls: add test for short splice due to full skmsg We don't have a test triggering a partial splice caused by a full skmsg. Add one, based on a program by Jann Horn. Use MAX_FRAGS=48 to make sure the skmsg will be full for any allowed value of CONFIG_MAX_SKB_FRAGS (17..45). Signed-off-by: Sabrina Dubroca Link: https://patch.msgid.link/1d129a15f526ea3602f3a2b368aa0b6f7e0d35d5.1760432043.git.sd@queasysnail.net Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/tls.c | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) diff --git a/tools/testing/selftests/net/tls.c b/tools/testing/selftests/net/tls.c index 77e4e30b46cc..5c6d8215021c 100644 --- a/tools/testing/selftests/net/tls.c +++ b/tools/testing/selftests/net/tls.c @@ -946,6 +946,37 @@ TEST_F(tls, peek_and_splice) EXPECT_EQ(memcmp(mem_send, mem_recv, send_len), 0); } +#define MAX_FRAGS 48 +TEST_F(tls, splice_short) +{ + struct iovec sendchar_iov; + char read_buf[0x10000]; + char sendbuf[0x100]; + char sendchar = 'S'; + int pipefds[2]; + int i; + + sendchar_iov.iov_base = &sendchar; + sendchar_iov.iov_len = 1; + + memset(sendbuf, 's', sizeof(sendbuf)); + + ASSERT_GE(pipe2(pipefds, O_NONBLOCK), 0); + ASSERT_GE(fcntl(pipefds[0], F_SETPIPE_SZ, (MAX_FRAGS + 1) * 0x1000), 0); + + for (i = 0; i < MAX_FRAGS; i++) + ASSERT_GE(vmsplice(pipefds[1], &sendchar_iov, 1, 0), 0); + + ASSERT_EQ(write(pipefds[1], sendbuf, sizeof(sendbuf)), sizeof(sendbuf)); + + EXPECT_EQ(splice(pipefds[0], NULL, self->fd, NULL, MAX_FRAGS + 0x1000, 0), + MAX_FRAGS + sizeof(sendbuf)); + EXPECT_EQ(recv(self->cfd, read_buf, sizeof(read_buf), 0), MAX_FRAGS + sizeof(sendbuf)); + EXPECT_EQ(recv(self->cfd, read_buf, sizeof(read_buf), MSG_DONTWAIT), -1); + EXPECT_EQ(errno, EAGAIN); +} +#undef MAX_FRAGS + TEST_F(tls, recvmsg_single) { char const *test_str = "test_recvmsg_single"; From 1a8fed52f7be14e45785e8e54d0d0b50fc17dbd8 Mon Sep 17 00:00:00 2001 From: Breno Leitao Date: Tue, 14 Oct 2025 02:17:25 -0700 Subject: [PATCH 48/50] netdevsim: set the carrier when the device goes up Bringing a linked netdevsim device down and then up causes communication failure because both interfaces lack carrier. Basically a ifdown/ifup on the interface make the link broken. Commit 3762ec05a9fbda ("netdevsim: add NAPI support") added supported for NAPI, calling netif_carrier_off() in nsim_stop(). This patch re-enables the carrier symmetrically on nsim_open(), in case the device is linked and the peer is up. Signed-off-by: Breno Leitao Fixes: 3762ec05a9fbda ("netdevsim: add NAPI support") Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/20251014-netdevsim_fix-v2-1-53b40590dae1@debian.org Signed-off-by: Jakub Kicinski --- drivers/net/netdevsim/netdev.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c index ebc3833e95b4..fa1d97885caa 100644 --- a/drivers/net/netdevsim/netdev.c +++ b/drivers/net/netdevsim/netdev.c @@ -545,6 +545,7 @@ static void nsim_enable_napi(struct netdevsim *ns) static int nsim_open(struct net_device *dev) { struct netdevsim *ns = netdev_priv(dev); + struct netdevsim *peer; int err; netdev_assert_locked(dev); @@ -555,6 +556,12 @@ static int nsim_open(struct net_device *dev) nsim_enable_napi(ns); + peer = rtnl_dereference(ns->peer); + if (peer && netif_running(peer->netdev)) { + netif_carrier_on(dev); + netif_carrier_on(peer->netdev); + } + return 0; } From 8d93ff40d49d70e05c82a74beae31f883fe0eaf8 Mon Sep 17 00:00:00 2001 From: I Viswanath Date: Mon, 13 Oct 2025 23:46:48 +0530 Subject: [PATCH 49/50] net: usb: lan78xx: fix use of improperly initialized dev->chipid in lan78xx_reset dev->chipid is used in lan78xx_init_mac_address before it's initialized: lan78xx_reset() { lan78xx_init_mac_address() lan78xx_read_eeprom() lan78xx_read_raw_eeprom() <- dev->chipid is used here dev->chipid = ... <- dev->chipid is initialized correctly here } Reorder initialization so that dev->chipid is set before calling lan78xx_init_mac_address(). Fixes: a0db7d10b76e ("lan78xx: Add to handle mux control per chip id") Signed-off-by: I Viswanath Reviewed-by: Vadim Fedorenko Reviewed-by: Khalid Aziz Link: https://patch.msgid.link/20251013181648.35153-1-viswanathiyyappan@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/usb/lan78xx.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/net/usb/lan78xx.c b/drivers/net/usb/lan78xx.c index 28195d9a8d6b..00397a807393 100644 --- a/drivers/net/usb/lan78xx.c +++ b/drivers/net/usb/lan78xx.c @@ -3250,10 +3250,6 @@ static int lan78xx_reset(struct lan78xx_net *dev) } } while (buf & HW_CFG_LRST_); - ret = lan78xx_init_mac_address(dev); - if (ret < 0) - return ret; - /* save DEVID for later usage */ ret = lan78xx_read_reg(dev, ID_REV, &buf); if (ret < 0) @@ -3262,6 +3258,10 @@ static int lan78xx_reset(struct lan78xx_net *dev) dev->chipid = (buf & ID_REV_CHIP_ID_MASK_) >> 16; dev->chiprev = buf & ID_REV_CHIP_REV_MASK_; + ret = lan78xx_init_mac_address(dev); + if (ret < 0) + return ret; + /* Respond to the IN token with a NAK */ ret = lan78xx_read_reg(dev, USB_CFG0, &buf); if (ret < 0) From 6de1dec1c166c7f7324ce52ccfdf43e2fa743b19 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 15 Oct 2025 05:27:15 +0000 Subject: [PATCH 50/50] udp: do not use skb_release_head_state() before skb_attempt_defer_free() Michal reported and bisected an issue after recent adoption of skb_attempt_defer_free() in UDP. The issue here is that skb_release_head_state() is called twice per skb, one time from skb_consume_udp(), then a second time from skb_defer_free_flush() and napi_consume_skb(). As Sabrina suggested, remove skb_release_head_state() call from skb_consume_udp(). Add a DEBUG_NET_WARN_ON_ONCE(skb_nfct(skb)) in skb_attempt_defer_free() Many thanks to Michal, Sabrina, Paolo and Florian for their help. Fixes: 6471658dc66c ("udp: use skb_attempt_defer_free()") Reported-and-bisected-by: Michal Kubecek Closes: https://lore.kernel.org/netdev/gpjh4lrotyephiqpuldtxxizrsg6job7cvhiqrw72saz2ubs3h@g6fgbvexgl3r/ Signed-off-by: Eric Dumazet Tested-by: Michal Kubecek Cc: Sabrina Dubroca Cc: Florian Westphal Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20251015052715.4140493-1-edumazet@google.com Signed-off-by: Paolo Abeni --- net/core/skbuff.c | 1 + net/ipv4/udp.c | 2 -- 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/net/core/skbuff.c b/net/core/skbuff.c index bc12790017b0..6be01454f262 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -7200,6 +7200,7 @@ nodefer: kfree_skb_napi_cache(skb); DEBUG_NET_WARN_ON_ONCE(skb_dst(skb)); DEBUG_NET_WARN_ON_ONCE(skb->destructor); + DEBUG_NET_WARN_ON_ONCE(skb_nfct(skb)); sdn = per_cpu_ptr(net_hotdata.skb_defer_nodes, cpu) + numa_node_id(); diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 95241093b7f0..30dfbf73729d 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -1851,8 +1851,6 @@ void skb_consume_udp(struct sock *sk, struct sk_buff *skb, int len) sk_peek_offset_bwd(sk, len); if (!skb_shared(skb)) { - if (unlikely(udp_skb_has_head_state(skb))) - skb_release_head_state(skb); skb_attempt_defer_free(skb); return; }