From aed237f645caaa7ae183dc70856c318058003358 Mon Sep 17 00:00:00 2001 From: Roxana Nicolescu Date: Tue, 21 Oct 2025 14:28:53 +0200 Subject: [PATCH 1/7] firmware: arm_scpi: Ensure scpi_info is not assigned if the probe fails jira VULN-70088 cve CVE-2022-50087 commit-author Sudeep Holla commit 689640efc0a2c4e07e6f88affe6d42cd40cc3f85 When scpi probe fails, at any point, we need to ensure that the scpi_info is not set and will remain NULL until the probe succeeds. If it is not taken care, then it could result use-after-free as the value is exported via get_scpi_ops() and could refer to a memory allocated via devm_kzalloc() but freed when the probe fails. Link: https://lore.kernel.org/r/20220701160310.148344-1-sudeep.holla@arm.com Cc: stable@vger.kernel.org # 4.19+ Reported-by: huhai Reviewed-by: Jackie Liu Signed-off-by: Sudeep Holla (cherry picked from commit 689640efc0a2c4e07e6f88affe6d42cd40cc3f85) Signed-off-by: Roxana Nicolescu --- drivers/firmware/arm_scpi.c | 61 +++++++++++++++++++++---------------- 1 file changed, 35 insertions(+), 26 deletions(-) diff --git a/drivers/firmware/arm_scpi.c b/drivers/firmware/arm_scpi.c index ddf0b9ff9e15c..435d0e2658a42 100644 --- a/drivers/firmware/arm_scpi.c +++ b/drivers/firmware/arm_scpi.c @@ -815,7 +815,7 @@ static int scpi_init_versions(struct scpi_drvinfo *info) info->firmware_version = le32_to_cpu(caps.platform_version); } /* Ignore error if not implemented */ - if (scpi_info->is_legacy && ret == -EOPNOTSUPP) + if (info->is_legacy && ret == -EOPNOTSUPP) return 0; return ret; @@ -913,13 +913,14 @@ static int scpi_probe(struct platform_device *pdev) struct resource res; struct device *dev = &pdev->dev; struct device_node *np = dev->of_node; + struct scpi_drvinfo *scpi_drvinfo; - scpi_info = devm_kzalloc(dev, sizeof(*scpi_info), GFP_KERNEL); - if (!scpi_info) + scpi_drvinfo = devm_kzalloc(dev, sizeof(*scpi_drvinfo), GFP_KERNEL); + if (!scpi_drvinfo) return -ENOMEM; if (of_match_device(legacy_scpi_of_match, &pdev->dev)) - scpi_info->is_legacy = true; + scpi_drvinfo->is_legacy = true; count = of_count_phandle_with_args(np, "mboxes", "#mbox-cells"); if (count < 0) { @@ -927,19 +928,19 @@ static int scpi_probe(struct platform_device *pdev) return -ENODEV; } - scpi_info->channels = devm_kcalloc(dev, count, sizeof(struct scpi_chan), - GFP_KERNEL); - if (!scpi_info->channels) + scpi_drvinfo->channels = + devm_kcalloc(dev, count, sizeof(struct scpi_chan), GFP_KERNEL); + if (!scpi_drvinfo->channels) return -ENOMEM; - ret = devm_add_action(dev, scpi_free_channels, scpi_info); + ret = devm_add_action(dev, scpi_free_channels, scpi_drvinfo); if (ret) return ret; - for (; scpi_info->num_chans < count; scpi_info->num_chans++) { + for (; scpi_drvinfo->num_chans < count; scpi_drvinfo->num_chans++) { resource_size_t size; - int idx = scpi_info->num_chans; - struct scpi_chan *pchan = scpi_info->channels + idx; + int idx = scpi_drvinfo->num_chans; + struct scpi_chan *pchan = scpi_drvinfo->channels + idx; struct mbox_client *cl = &pchan->cl; struct device_node *shmem = of_parse_phandle(np, "shmem", idx); @@ -986,45 +987,53 @@ static int scpi_probe(struct platform_device *pdev) return ret; } - scpi_info->commands = scpi_std_commands; + scpi_drvinfo->commands = scpi_std_commands; - platform_set_drvdata(pdev, scpi_info); + platform_set_drvdata(pdev, scpi_drvinfo); - if (scpi_info->is_legacy) { + if (scpi_drvinfo->is_legacy) { /* Replace with legacy variants */ scpi_ops.clk_set_val = legacy_scpi_clk_set_val; - scpi_info->commands = scpi_legacy_commands; + scpi_drvinfo->commands = scpi_legacy_commands; /* Fill priority bitmap */ for (idx = 0; idx < ARRAY_SIZE(legacy_hpriority_cmds); idx++) set_bit(legacy_hpriority_cmds[idx], - scpi_info->cmd_priority); + scpi_drvinfo->cmd_priority); } - ret = scpi_init_versions(scpi_info); + scpi_info = scpi_drvinfo; + + ret = scpi_init_versions(scpi_drvinfo); if (ret) { dev_err(dev, "incorrect or no SCP firmware found\n"); + scpi_info = NULL; return ret; } - if (scpi_info->is_legacy && !scpi_info->protocol_version && - !scpi_info->firmware_version) + if (scpi_drvinfo->is_legacy && !scpi_drvinfo->protocol_version && + !scpi_drvinfo->firmware_version) dev_info(dev, "SCP Protocol legacy pre-1.0 firmware\n"); else dev_info(dev, "SCP Protocol %lu.%lu Firmware %lu.%lu.%lu version\n", FIELD_GET(PROTO_REV_MAJOR_MASK, - scpi_info->protocol_version), + scpi_drvinfo->protocol_version), FIELD_GET(PROTO_REV_MINOR_MASK, - scpi_info->protocol_version), + scpi_drvinfo->protocol_version), FIELD_GET(FW_REV_MAJOR_MASK, - scpi_info->firmware_version), + scpi_drvinfo->firmware_version), FIELD_GET(FW_REV_MINOR_MASK, - scpi_info->firmware_version), + scpi_drvinfo->firmware_version), FIELD_GET(FW_REV_PATCH_MASK, - scpi_info->firmware_version)); - scpi_info->scpi_ops = &scpi_ops; + scpi_drvinfo->firmware_version)); + + scpi_drvinfo->scpi_ops = &scpi_ops; - return devm_of_platform_populate(dev); + ret = devm_of_platform_populate(dev); + if (ret) + scpi_info = NULL; + + return ret; } static const struct of_device_id scpi_of_match[] = { From 851bb9db07fde503e3c9f80c51ee855bbe3132dd Mon Sep 17 00:00:00 2001 From: Roxana Nicolescu Date: Tue, 21 Oct 2025 14:29:00 +0200 Subject: [PATCH 2/7] net: usb: smsc75xx: Limit packet length to skb->len jira VULN-67490 cve CVE-2023-53125 commit-author Szymon Heidrich commit d8b228318935044dafe3a5bc07ee71a1f1424b8d Packet length retrieved from skb data may be larger than the actual socket buffer length (up to 9026 bytes). In such case the cloned skb passed up the network stack will leak kernel memory contents. Fixes: d0cad871703b ("smsc75xx: SMSC LAN75xx USB gigabit ethernet adapter driver") Signed-off-by: Szymon Heidrich Signed-off-by: David S. Miller (cherry picked from commit d8b228318935044dafe3a5bc07ee71a1f1424b8d) Signed-off-by: Roxana Nicolescu --- drivers/net/usb/smsc75xx.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/usb/smsc75xx.c b/drivers/net/usb/smsc75xx.c index 76f7af1613139..705bd31b18787 100644 --- a/drivers/net/usb/smsc75xx.c +++ b/drivers/net/usb/smsc75xx.c @@ -2211,7 +2211,8 @@ static int smsc75xx_rx_fixup(struct usbnet *dev, struct sk_buff *skb) dev->net->stats.rx_frame_errors++; } else { /* MAX_SINGLE_PACKET_SIZE + 4(CRC) + 2(COE) + 4(Vlan) */ - if (unlikely(size > (MAX_SINGLE_PACKET_SIZE + ETH_HLEN + 12))) { + if (unlikely(size > (MAX_SINGLE_PACKET_SIZE + ETH_HLEN + 12) || + size > skb->len)) { netif_dbg(dev, rx_err, dev->net, "size err rx_cmd_a=0x%08x\n", rx_cmd_a); From efc8e7271d83e4e4c94b4ce8d050e8f051bc3d9d Mon Sep 17 00:00:00 2001 From: Roxana Nicolescu Date: Wed, 22 Oct 2025 12:45:04 +0200 Subject: [PATCH 3/7] net: usb: smsc75xx: Move packet length check to prevent kernel panic in skb_pull jira VULN-67490 cve-bf CVE-2023-53125 commit-author Szymon Heidrich commit 43ffe6caccc7a1bb9d7442fbab521efbf6c1378c Packet length check needs to be located after size and align_count calculation to prevent kernel panic in skb_pull() in case rx_cmd_a & RX_CMD_A_RED evaluates to true. Fixes: d8b228318935 ("net: usb: smsc75xx: Limit packet length to skb->len") Signed-off-by: Szymon Heidrich Link: https://lore.kernel.org/r/20230316110540.77531-1-szymon.heidrich@gmail.com Signed-off-by: Jakub Kicinski (cherry picked from commit 43ffe6caccc7a1bb9d7442fbab521efbf6c1378c) Signed-off-by: Roxana Nicolescu --- drivers/net/usb/smsc75xx.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/net/usb/smsc75xx.c b/drivers/net/usb/smsc75xx.c index 705bd31b18787..7c3e866514199 100644 --- a/drivers/net/usb/smsc75xx.c +++ b/drivers/net/usb/smsc75xx.c @@ -2199,6 +2199,13 @@ static int smsc75xx_rx_fixup(struct usbnet *dev, struct sk_buff *skb) size = (rx_cmd_a & RX_CMD_A_LEN) - RXW_PADDING; align_count = (4 - ((size + RXW_PADDING) % 4)) % 4; + if (unlikely(size > skb->len)) { + netif_dbg(dev, rx_err, dev->net, + "size err rx_cmd_a=0x%08x\n", + rx_cmd_a); + return 0; + } + if (unlikely(rx_cmd_a & RX_CMD_A_RED)) { netif_dbg(dev, rx_err, dev->net, "Error rx_cmd_a=0x%08x\n", rx_cmd_a); @@ -2211,8 +2218,7 @@ static int smsc75xx_rx_fixup(struct usbnet *dev, struct sk_buff *skb) dev->net->stats.rx_frame_errors++; } else { /* MAX_SINGLE_PACKET_SIZE + 4(CRC) + 2(COE) + 4(Vlan) */ - if (unlikely(size > (MAX_SINGLE_PACKET_SIZE + ETH_HLEN + 12) || - size > skb->len)) { + if (unlikely(size > (MAX_SINGLE_PACKET_SIZE + ETH_HLEN + 12))) { netif_dbg(dev, rx_err, dev->net, "size err rx_cmd_a=0x%08x\n", rx_cmd_a); From 8eb0e45cb5194338cce52d3ad3cd79c510637637 Mon Sep 17 00:00:00 2001 From: Roxana Nicolescu Date: Tue, 21 Oct 2025 14:29:10 +0200 Subject: [PATCH 4/7] net_sched: hfsc: Fix a potential UAF in hfsc_dequeue() too jira VULN-67737 cve CVE-2025-37823 commit-author Cong Wang commit 6ccbda44e2cc3d26fd22af54c650d6d5d801addf Similarly to the previous patch, we need to safe guard hfsc_dequeue() too. But for this one, we don't have a reliable reproducer. Fixes: 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 ("Linux-2.6.12-rc2") Reported-by: Gerrard Tai Signed-off-by: Cong Wang Reviewed-by: Jamal Hadi Salim Link: https://patch.msgid.link/20250417184732.943057-3-xiyou.wangcong@gmail.com Signed-off-by: Jakub Kicinski (cherry picked from commit 6ccbda44e2cc3d26fd22af54c650d6d5d801addf) Signed-off-by: Roxana Nicolescu --- net/sched/sch_hfsc.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 84baf86e4a04c..71ca8e0fd6fab 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1649,10 +1649,16 @@ hfsc_dequeue(struct Qdisc *sch) if (cl->qdisc->q.qlen != 0) { /* update ed */ next_len = qdisc_peek_len(cl->qdisc); - if (realtime) - update_ed(cl, next_len); - else - update_d(cl, next_len); + /* Check queue length again since some qdisc implementations + * (e.g., netem/codel) might empty the queue during the peek + * operation. + */ + if (cl->qdisc->q.qlen != 0) { + if (realtime) + update_ed(cl, next_len); + else + update_d(cl, next_len); + } } else { /* the class becomes passive */ eltree_remove(cl); From a22d3872336d727d1d09727a8cf1a6959859f4b7 Mon Sep 17 00:00:00 2001 From: Roxana Nicolescu Date: Tue, 21 Oct 2025 14:29:18 +0200 Subject: [PATCH 5/7] i40e: fix MMIO write access to an invalid page in i40e_clear_hw jira VULN-72065 cve CVE-2025-38200 commit-author Kyungwook Boo commit 015bac5daca978448f2671478c553ce1f300c21e When the device sends a specific input, an integer underflow can occur, leading to MMIO write access to an invalid page. Prevent the integer underflow by changing the type of related variables. Signed-off-by: Kyungwook Boo Link: https://lore.kernel.org/lkml/ffc91764-1142-4ba2-91b6-8c773f6f7095@gmail.com/T/ Reviewed-by: Przemek Kitszel Reviewed-by: Simon Horman Reviewed-by: Aleksandr Loktionov Tested-by: Rinitha S (A Contingent worker at Intel) Signed-off-by: Tony Nguyen (cherry picked from commit 015bac5daca978448f2671478c553ce1f300c21e) Signed-off-by: Roxana Nicolescu --- drivers/net/ethernet/intel/i40e/i40e_common.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/intel/i40e/i40e_common.c b/drivers/net/ethernet/intel/i40e/i40e_common.c index de6ca62957427..f1045ec5b9af1 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_common.c +++ b/drivers/net/ethernet/intel/i40e/i40e_common.c @@ -1070,10 +1070,11 @@ int i40e_pf_reset(struct i40e_hw *hw) void i40e_clear_hw(struct i40e_hw *hw) { u32 num_queues, base_queue; - u32 num_pf_int; - u32 num_vf_int; + s32 num_pf_int; + s32 num_vf_int; u32 num_vfs; - u32 i, j; + s32 i; + u32 j; u32 val; u32 eol = 0x7ff; From 8b5c04f568ed0b9439ec1ed9a7e398c8db163325 Mon Sep 17 00:00:00 2001 From: Roxana Nicolescu Date: Tue, 21 Oct 2025 14:29:25 +0200 Subject: [PATCH 6/7] idpf: convert control queue mutex to a spinlock jira VULN-136700 cve CVE-2025-38392 commit-author Ahmed Zaki commit b2beb5bb2cd90d7939e470ed4da468683f41baa3 With VIRTCHNL2_CAP_MACFILTER enabled, the following warning is generated on module load: [ 324.701677] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:578 [ 324.701684] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1582, name: NetworkManager [ 324.701689] preempt_count: 201, expected: 0 [ 324.701693] RCU nest depth: 0, expected: 0 [ 324.701697] 2 locks held by NetworkManager/1582: [ 324.701702] #0: ffffffff9f7be770 (rtnl_mutex){....}-{3:3}, at: rtnl_newlink+0x791/0x21e0 [ 324.701730] #1: ff1100216c380368 (_xmit_ETHER){....}-{2:2}, at: __dev_open+0x3f0/0x870 [ 324.701749] Preemption disabled at: [ 324.701752] [] __dev_open+0x3dd/0x870 [ 324.701765] CPU: 30 UID: 0 PID: 1582 Comm: NetworkManager Not tainted 6.15.0-rc5+ #2 PREEMPT(voluntary) [ 324.701771] Hardware name: Intel Corporation M50FCP2SBSTD/M50FCP2SBSTD, BIOS SE5C741.86B.01.01.0001.2211140926 11/14/2022 [ 324.701774] Call Trace: [ 324.701777] [ 324.701779] dump_stack_lvl+0x5d/0x80 [ 324.701788] ? __dev_open+0x3dd/0x870 [ 324.701793] __might_resched.cold+0x1ef/0x23d <..> [ 324.701818] __mutex_lock+0x113/0x1b80 <..> [ 324.701917] idpf_ctlq_clean_sq+0xad/0x4b0 [idpf] [ 324.701935] ? kasan_save_track+0x14/0x30 [ 324.701941] idpf_mb_clean+0x143/0x380 [idpf] <..> [ 324.701991] idpf_send_mb_msg+0x111/0x720 [idpf] [ 324.702009] idpf_vc_xn_exec+0x4cc/0x990 [idpf] [ 324.702021] ? rcu_is_watching+0x12/0xc0 [ 324.702035] idpf_add_del_mac_filters+0x3ed/0xb50 [idpf] <..> [ 324.702122] __hw_addr_sync_dev+0x1cf/0x300 [ 324.702126] ? find_held_lock+0x32/0x90 [ 324.702134] idpf_set_rx_mode+0x317/0x390 [idpf] [ 324.702152] __dev_open+0x3f8/0x870 [ 324.702159] ? __pfx___dev_open+0x10/0x10 [ 324.702174] __dev_change_flags+0x443/0x650 <..> [ 324.702208] netif_change_flags+0x80/0x160 [ 324.702218] do_setlink.isra.0+0x16a0/0x3960 <..> [ 324.702349] rtnl_newlink+0x12fd/0x21e0 The sequence is as follows: rtnl_newlink()-> __dev_change_flags()-> __dev_open()-> dev_set_rx_mode() - > # disables BH and grabs "dev->addr_list_lock" idpf_set_rx_mode() -> # proceed only if VIRTCHNL2_CAP_MACFILTER is ON __dev_uc_sync() -> idpf_add_mac_filter -> idpf_add_del_mac_filters -> idpf_send_mb_msg() -> idpf_mb_clean() -> idpf_ctlq_clean_sq() # mutex_lock(cq_lock) Fix by converting cq_lock to a spinlock. All operations under the new lock are safe except freeing the DMA memory, which may use vunmap(). Fix by requesting a contiguous physical memory for the DMA mapping. Fixes: a251eee62133 ("idpf: add SRIOV support and other ndo_ops") Reviewed-by: Aleksandr Loktionov Signed-off-by: Ahmed Zaki Reviewed-by: Simon Horman Tested-by: Samuel Salin Signed-off-by: Tony Nguyen (cherry picked from commit b2beb5bb2cd90d7939e470ed4da468683f41baa3) Signed-off-by: Roxana Nicolescu --- .../net/ethernet/intel/idpf/idpf_controlq.c | 23 +++++++++---------- .../ethernet/intel/idpf/idpf_controlq_api.h | 2 +- drivers/net/ethernet/intel/idpf/idpf_lib.c | 12 ++++++---- 3 files changed, 20 insertions(+), 17 deletions(-) diff --git a/drivers/net/ethernet/intel/idpf/idpf_controlq.c b/drivers/net/ethernet/intel/idpf/idpf_controlq.c index c7f43d2fcd136..1595b6b332fe9 100644 --- a/drivers/net/ethernet/intel/idpf/idpf_controlq.c +++ b/drivers/net/ethernet/intel/idpf/idpf_controlq.c @@ -96,7 +96,7 @@ static void idpf_ctlq_init_rxq_bufs(struct idpf_ctlq_info *cq) */ static void idpf_ctlq_shutdown(struct idpf_hw *hw, struct idpf_ctlq_info *cq) { - mutex_lock(&cq->cq_lock); + spin_lock(&cq->cq_lock); /* free ring buffers and the ring itself */ idpf_ctlq_dealloc_ring_res(hw, cq); @@ -104,8 +104,7 @@ static void idpf_ctlq_shutdown(struct idpf_hw *hw, struct idpf_ctlq_info *cq) /* Set ring_size to 0 to indicate uninitialized queue */ cq->ring_size = 0; - mutex_unlock(&cq->cq_lock); - mutex_destroy(&cq->cq_lock); + spin_unlock(&cq->cq_lock); } /** @@ -173,7 +172,7 @@ int idpf_ctlq_add(struct idpf_hw *hw, idpf_ctlq_init_regs(hw, cq, is_rxq); - mutex_init(&cq->cq_lock); + spin_lock_init(&cq->cq_lock); list_add(&cq->cq_list, &hw->cq_list_head); @@ -272,7 +271,7 @@ int idpf_ctlq_send(struct idpf_hw *hw, struct idpf_ctlq_info *cq, int err = 0; int i; - mutex_lock(&cq->cq_lock); + spin_lock(&cq->cq_lock); /* Ensure there are enough descriptors to send all messages */ num_desc_avail = IDPF_CTLQ_DESC_UNUSED(cq); @@ -332,7 +331,7 @@ int idpf_ctlq_send(struct idpf_hw *hw, struct idpf_ctlq_info *cq, wr32(hw, cq->reg.tail, cq->next_to_use); err_unlock: - mutex_unlock(&cq->cq_lock); + spin_unlock(&cq->cq_lock); return err; } @@ -364,7 +363,7 @@ int idpf_ctlq_clean_sq(struct idpf_ctlq_info *cq, u16 *clean_count, if (*clean_count > cq->ring_size) return -EBADR; - mutex_lock(&cq->cq_lock); + spin_lock(&cq->cq_lock); ntc = cq->next_to_clean; @@ -394,7 +393,7 @@ int idpf_ctlq_clean_sq(struct idpf_ctlq_info *cq, u16 *clean_count, cq->next_to_clean = ntc; - mutex_unlock(&cq->cq_lock); + spin_unlock(&cq->cq_lock); /* Return number of descriptors actually cleaned */ *clean_count = i; @@ -432,7 +431,7 @@ int idpf_ctlq_post_rx_buffs(struct idpf_hw *hw, struct idpf_ctlq_info *cq, if (*buff_count > 0) buffs_avail = true; - mutex_lock(&cq->cq_lock); + spin_lock(&cq->cq_lock); if (tbp >= cq->ring_size) tbp = 0; @@ -519,7 +518,7 @@ int idpf_ctlq_post_rx_buffs(struct idpf_hw *hw, struct idpf_ctlq_info *cq, wr32(hw, cq->reg.tail, cq->next_to_post); } - mutex_unlock(&cq->cq_lock); + spin_unlock(&cq->cq_lock); /* return the number of buffers that were not posted */ *buff_count = *buff_count - i; @@ -552,7 +551,7 @@ int idpf_ctlq_recv(struct idpf_ctlq_info *cq, u16 *num_q_msg, return -EBADR; /* take the lock before we start messing with the ring */ - mutex_lock(&cq->cq_lock); + spin_lock(&cq->cq_lock); ntc = cq->next_to_clean; @@ -611,7 +610,7 @@ int idpf_ctlq_recv(struct idpf_ctlq_info *cq, u16 *num_q_msg, cq->next_to_clean = ntc; - mutex_unlock(&cq->cq_lock); + spin_unlock(&cq->cq_lock); *num_q_msg = i; if (*num_q_msg == 0) diff --git a/drivers/net/ethernet/intel/idpf/idpf_controlq_api.h b/drivers/net/ethernet/intel/idpf/idpf_controlq_api.h index 8dee098bbfb0c..385e21ebbb2d0 100644 --- a/drivers/net/ethernet/intel/idpf/idpf_controlq_api.h +++ b/drivers/net/ethernet/intel/idpf/idpf_controlq_api.h @@ -94,7 +94,7 @@ struct idpf_ctlq_info { enum idpf_ctlq_type cq_type; int q_id; - struct mutex cq_lock; /* control queue lock */ + spinlock_t cq_lock; /* control queue lock */ /* used for interrupt processing */ u16 next_to_use; u16 next_to_clean; diff --git a/drivers/net/ethernet/intel/idpf/idpf_lib.c b/drivers/net/ethernet/intel/idpf/idpf_lib.c index 0241e498cc20f..999ff7c4a54b9 100644 --- a/drivers/net/ethernet/intel/idpf/idpf_lib.c +++ b/drivers/net/ethernet/intel/idpf/idpf_lib.c @@ -2329,8 +2329,12 @@ void *idpf_alloc_dma_mem(struct idpf_hw *hw, struct idpf_dma_mem *mem, u64 size) struct idpf_adapter *adapter = hw->back; size_t sz = ALIGN(size, 4096); - mem->va = dma_alloc_coherent(&adapter->pdev->dev, sz, - &mem->pa, GFP_KERNEL); + /* The control queue resources are freed under a spinlock, contiguous + * pages will avoid IOMMU remapping and the use vmap (and vunmap in + * dma_free_*() path. + */ + mem->va = dma_alloc_attrs(&adapter->pdev->dev, sz, &mem->pa, + GFP_KERNEL, DMA_ATTR_FORCE_CONTIGUOUS); mem->size = sz; return mem->va; @@ -2345,8 +2349,8 @@ void idpf_free_dma_mem(struct idpf_hw *hw, struct idpf_dma_mem *mem) { struct idpf_adapter *adapter = hw->back; - dma_free_coherent(&adapter->pdev->dev, mem->size, - mem->va, mem->pa); + dma_free_attrs(&adapter->pdev->dev, mem->size, + mem->va, mem->pa, DMA_ATTR_FORCE_CONTIGUOUS); mem->size = 0; mem->va = NULL; mem->pa = 0; From 2a15422fed934f073e84849ee3e7b2d8cf9dd004 Mon Sep 17 00:00:00 2001 From: Roxana Nicolescu Date: Tue, 21 Oct 2025 14:29:35 +0200 Subject: [PATCH 7/7] netfilter: nf_conntrack: fix crash due to removal of uninitialised entry jira VULN-89210 cve CVE-2025-38472 commit-author Florian Westphal commit 2d72afb340657f03f7261e9243b44457a9228ac7 A crash in conntrack was reported while trying to unlink the conntrack entry from the hash bucket list: [exception RIP: __nf_ct_delete_from_lists+172] [..] #7 [ff539b5a2b043aa0] nf_ct_delete at ffffffffc124d421 [nf_conntrack] #8 [ff539b5a2b043ad0] nf_ct_gc_expired at ffffffffc124d999 [nf_conntrack] #9 [ff539b5a2b043ae0] __nf_conntrack_find_get at ffffffffc124efbc [nf_conntrack] [..] The nf_conn struct is marked as allocated from slab but appears to be in a partially initialised state: ct hlist pointer is garbage; looks like the ct hash value (hence crash). ct->status is equal to IPS_CONFIRMED|IPS_DYING, which is expected ct->timeout is 30000 (=30s), which is unexpected. Everything else looks like normal udp conntrack entry. If we ignore ct->status and pretend its 0, the entry matches those that are newly allocated but not yet inserted into the hash: - ct hlist pointers are overloaded and store/cache the raw tuple hash - ct->timeout matches the relative time expected for a new udp flow rather than the absolute 'jiffies' value. If it were not for the presence of IPS_CONFIRMED, __nf_conntrack_find_get() would have skipped the entry. Theory is that we did hit following race: cpu x cpu y cpu z found entry E found entry E E is expired nf_ct_delete() return E to rcu slab init_conntrack E is re-inited, ct->status set to 0 reply tuplehash hnnode.pprev stores hash value. cpu y found E right before it was deleted on cpu x. E is now re-inited on cpu z. cpu y was preempted before checking for expiry and/or confirm bit. ->refcnt set to 1 E now owned by skb ->timeout set to 30000 If cpu y were to resume now, it would observe E as expired but would skip E due to missing CONFIRMED bit. nf_conntrack_confirm gets called sets: ct->status |= CONFIRMED This is wrong: E is not yet added to hashtable. cpu y resumes, it observes E as expired but CONFIRMED: nf_ct_expired() -> yes (ct->timeout is 30s) confirmed bit set. cpu y will try to delete E from the hashtable: nf_ct_delete() -> set DYING bit __nf_ct_delete_from_lists Even this scenario doesn't guarantee a crash: cpu z still holds the table bucket lock(s) so y blocks: wait for spinlock held by z CONFIRMED is set but there is no guarantee ct will be added to hash: "chaintoolong" or "clash resolution" logic both skip the insert step. reply hnnode.pprev still stores the hash value. unlocks spinlock return NF_DROP In case CPU z does insert the entry into the hashtable, cpu y will unlink E again right away but no crash occurs. Without 'cpu y' race, 'garbage' hlist is of no consequence: ct refcnt remains at 1, eventually skb will be free'd and E gets destroyed via: nf_conntrack_put -> nf_conntrack_destroy -> nf_ct_destroy. To resolve this, move the IPS_CONFIRMED assignment after the table insertion but before the unlock. Pablo points out that the confirm-bit-store could be reordered to happen before hlist add resp. the timeout fixup, so switch to set_bit and before_atomic memory barrier to prevent this. It doesn't matter if other CPUs can observe a newly inserted entry right before the CONFIRMED bit was set: Such event cannot be distinguished from above "E is the old incarnation" case: the entry will be skipped. Also change nf_ct_should_gc() to first check the confirmed bit. The gc sequence is: 1. Check if entry has expired, if not skip to next entry 2. Obtain a reference to the expired entry. 3. Call nf_ct_should_gc() to double-check step 1. nf_ct_should_gc() is thus called only for entries that already failed an expiry check. After this patch, once the confirmed bit check passes ct->timeout has been altered to reflect the absolute 'best before' date instead of a relative time. Step 3 will therefore not remove the entry. Without this change to nf_ct_should_gc() we could still get this sequence: 1. Check if entry has expired. 2. Obtain a reference. 3. Call nf_ct_should_gc() to double-check step 1: 4 - entry is still observed as expired 5 - meanwhile, ct->timeout is corrected to absolute value on other CPU and confirm bit gets set 6 - confirm bit is seen 7 - valid entry is removed again First do check 6), then 4) so the gc expiry check always picks up either confirmed bit unset (entry gets skipped) or expiry re-check failure for re-inited conntrack objects. This change cannot be backported to releases before 5.19. Without commit 8a75a2c17410 ("netfilter: conntrack: remove unconfirmed list") |= IPS_CONFIRMED line cannot be moved without further changes. Cc: Razvan Cojocaru Link: https://lore.kernel.org/netfilter-devel/20250627142758.25664-1-fw@strlen.de/ Link: https://lore.kernel.org/netfilter-devel/4239da15-83ff-4ca4-939d-faef283471bb@gmail.com/ Fixes: 1397af5bfd7d ("netfilter: conntrack: remove the percpu dying list") Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso (cherry picked from commit 2d72afb340657f03f7261e9243b44457a9228ac7) Signed-off-by: Roxana Nicolescu --- include/net/netfilter/nf_conntrack.h | 15 +++++++++++++-- net/netfilter/nf_conntrack_core.c | 26 ++++++++++++++++++++------ 2 files changed, 33 insertions(+), 8 deletions(-) diff --git a/include/net/netfilter/nf_conntrack.h b/include/net/netfilter/nf_conntrack.h index 702f15b1ebe43..75dd8d92f3d73 100644 --- a/include/net/netfilter/nf_conntrack.h +++ b/include/net/netfilter/nf_conntrack.h @@ -300,8 +300,19 @@ static inline bool nf_ct_is_expired(const struct nf_conn *ct) /* use after obtaining a reference count */ static inline bool nf_ct_should_gc(const struct nf_conn *ct) { - return nf_ct_is_expired(ct) && nf_ct_is_confirmed(ct) && - !nf_ct_is_dying(ct); + if (!nf_ct_is_confirmed(ct)) + return false; + + /* load ct->timeout after is_confirmed() test. + * Pairs with __nf_conntrack_confirm() which: + * 1. Increases ct->timeout value + * 2. Inserts ct into rcu hlist + * 3. Sets the confirmed bit + * 4. Unlocks the hlist lock + */ + smp_acquire__after_ctrl_dep(); + + return nf_ct_is_expired(ct) && !nf_ct_is_dying(ct); } #define NF_CT_DAY (86400 * HZ) diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index 75de47106342a..bcb283728ad8d 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -1087,6 +1087,12 @@ static int nf_ct_resolve_clash_harder(struct sk_buff *skb, u32 repl_idx) hlist_nulls_add_head_rcu(&loser_ct->tuplehash[IP_CT_DIR_REPLY].hnnode, &nf_conntrack_hash[repl_idx]); + /* confirmed bit must be set after hlist add, not before: + * loser_ct can still be visible to other cpu due to + * SLAB_TYPESAFE_BY_RCU. + */ + smp_mb__before_atomic(); + set_bit(IPS_CONFIRMED_BIT, &loser_ct->status); NF_CT_STAT_INC(net, clash_resolve); return NF_ACCEPT; @@ -1224,8 +1230,6 @@ __nf_conntrack_confirm(struct sk_buff *skb) * user context, else we insert an already 'dead' hash, blocking * further use of that particular connection -JM. */ - ct->status |= IPS_CONFIRMED; - if (unlikely(nf_ct_is_dying(ct))) { NF_CT_STAT_INC(net, insert_failed); goto dying; @@ -1257,7 +1261,7 @@ __nf_conntrack_confirm(struct sk_buff *skb) } } - /* Timer relative to confirmation time, not original + /* Timeout is relative to confirmation time, not original setting time, otherwise we'd get timer wrap in weird delay cases. */ ct->timeout += nfct_time_stamp; @@ -1265,11 +1269,21 @@ __nf_conntrack_confirm(struct sk_buff *skb) __nf_conntrack_insert_prepare(ct); /* Since the lookup is lockless, hash insertion must be done after - * starting the timer and setting the CONFIRMED bit. The RCU barriers - * guarantee that no other CPU can find the conntrack before the above - * stores are visible. + * setting ct->timeout. The RCU barriers guarantee that no other CPU + * can find the conntrack before the above stores are visible. */ __nf_conntrack_hash_insert(ct, hash, reply_hash); + + /* IPS_CONFIRMED unset means 'ct not (yet) in hash', conntrack lookups + * skip entries that lack this bit. This happens when a CPU is looking + * at a stale entry that is being recycled due to SLAB_TYPESAFE_BY_RCU + * or when another CPU encounters this entry right after the insertion + * but before the set-confirm-bit below. This bit must not be set until + * after __nf_conntrack_hash_insert(). + */ + smp_mb__before_atomic(); + set_bit(IPS_CONFIRMED_BIT, &ct->status); + nf_conntrack_double_unlock(hash, reply_hash); local_bh_enable();