- Kernel - mailweb.openeuler.org

[PATCH] net: usb: qmi_wwan: initialize MAC header offset in qmimux_rx_fixup
by Jiacheng Yu 06 Jan '26

06 Jan '26

From: Qendrim Maxhuni <qendrim.maxhuni(a)garderos.com> [ Upstream commit e120f46768d98151ece8756ebd688b0e43dc8b29 ] Raw IP packets have no MAC header, leaving skb->mac_header uninitialized. This can trigger kernel panics on ARM64 when xfrm or other subsystems access the offset due to strict alignment checks. Initialize the MAC header to prevent such crashes. This can trigger kernel panics on ARM when running IPsec over the qmimux0 interface. Example trace: Internal error: Oops: 000000009600004f [#1] SMP CPU: 0 UID: 0 PID: 0 Comm: swapper/0 Not tainted 6.12.34-gbe78e49cb433 #1 Hardware name: LS1028A RDB Board (DT) pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--) pc : xfrm_input+0xde8/0x1318 lr : xfrm_input+0x61c/0x1318 sp : ffff800080003b20 Call trace: xfrm_input+0xde8/0x1318 xfrm6_rcv+0x38/0x44 xfrm6_esp_rcv+0x48/0xa8 ip6_protocol_deliver_rcu+0x94/0x4b0 ip6_input_finish+0x44/0x70 ip6_input+0x44/0xc0 ipv6_rcv+0x6c/0x114 __netif_receive_skb_one_core+0x5c/0x8c __netif_receive_skb+0x18/0x60 process_backlog+0x78/0x17c __napi_poll+0x38/0x180 net_rx_action+0x168/0x2f0 Fixes: c6adf77953bc ("net: usb: qmi_wwan: add qmap mux protocol support") Signed-off-by: Qendrim Maxhuni <qendrim.maxhuni(a)garderos.com> Link: https://patch.msgid.link/20251029075744.105113-1-qendrim.maxhuni@garderos.c… Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- drivers/net/usb/qmi_wwan.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/net/usb/qmi_wwan.c b/drivers/net/usb/qmi_wwan.c index eba755b584a4..73df808978b2 100644 --- a/drivers/net/usb/qmi_wwan.c +++ b/drivers/net/usb/qmi_wwan.c @@ -192,6 +192,12 @@ static int qmimux_rx_fixup(struct usbnet *dev, struct sk_buff *skb) if (!skbn) return 0; + /* Raw IP packets don't have a MAC header, but other subsystems + * (like xfrm) may still access MAC header offsets, so they must + * be initialized. + */ + skb_reset_mac_header(skbn); + switch (skb->data[offset + qmimux_hdr_sz] & 0xf0) { case 0x40: skbn->protocol = htons(ETH_P_IP); -- 2.43.0

1 0

[PATCH OLK-5.10] i40e: fix input validation logic for action_meta
by Yao Kai 06 Jan '26

06 Jan '26

From: Lukasz Czapnik <lukasz.czapnik(a)intel.com> stable inclusion from stable-v5.10.245 commit 28465770ca3b694286ff9ed6dfd558413f57d98f category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/7944 CVE: CVE-2025-39970 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit 9739d5830497812b0bdeaee356ddefbe60830b88 upstream. Fix condition to check 'greater or equal' to prevent OOB dereference. Fixes: e284fc280473 ("i40e: Add and delete cloud filter") Cc: stable(a)vger.kernel.org Signed-off-by: Lukasz Czapnik <lukasz.czapnik(a)intel.com> Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov(a)intel.com> Signed-off-by: Przemek Kitszel <przemyslaw.kitszel(a)intel.com> Reviewed-by: Simon Horman <horms(a)kernel.org> Tested-by: Rafal Romanowski <rafal.romanowski(a)intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen(a)intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Yao Kai <yaokai34(a)huawei.com> --- drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c index 83a8130b6850..56649d757fc7 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c +++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c @@ -3424,7 +3424,7 @@ static int i40e_validate_cloud_filter(struct i40e_vf *vf, /* action_meta is TC number here to which the filter is applied */ if (!tc_filter->action_meta || - tc_filter->action_meta > vf->num_tc) { + tc_filter->action_meta >= vf->num_tc) { dev_info(&pf->pdev->dev, "VF %d: Invalid TC number %u\n", vf->vf_id, tc_filter->action_meta); goto err; -- 2.43.0

2 5

[PATCH OLK-5.10] wifi: mt76: do not run mt76u_status_worker if the device is not running
by Jiacheng Yu 06 Jan '26

06 Jan '26

From: Lorenzo Bianconi <lorenzo(a)kernel.org> stable inclusion from stable-v5.15.86 commit 69346de0eb956fb92949b9473de4647d9c34a54f category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/12848 CVE: CVE-2022-50735 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit bd5dac7ced5a7c9faa4dc468ac9560c3256df845 ] Fix the following NULL pointer dereference avoiding to run mt76u_status_worker thread if the device is not running yet. KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007] CPU: 0 PID: 98 Comm: kworker/u2:2 Not tainted 5.14.0+ #78 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014 Workqueue: mt76 mt76u_tx_status_data RIP: 0010:mt76x02_mac_fill_tx_status.isra.0+0x82c/0x9e0 Code: c5 48 b8 00 00 00 00 00 fc ff df 80 3c 02 00 0f 85 94 01 00 00 48 b8 00 00 00 00 00 fc ff df 4d 8b 34 24 4c 89 f2 48 c1 ea 03 <0f> b6 04 02 84 c0 74 08 3c 03 0f 8e 89 01 00 00 41 8b 16 41 0f b7 RSP: 0018:ffffc900005af988 EFLAGS: 00010246 RAX: dffffc0000000000 RBX: ffffc900005afae8 RCX: 0000000000000000 RDX: 0000000000000000 RSI: ffffffff832fc661 RDI: ffffc900005afc2a RBP: ffffc900005afae0 R08: 0000000000000001 R09: fffff520000b5f3c R10: 0000000000000003 R11: fffff520000b5f3b R12: ffff88810b6132d8 R13: 000000000000ffff R14: 0000000000000000 R15: ffffc900005afc28 FS: 0000000000000000(0000) GS:ffff88811aa00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fa0eda6a000 CR3: 0000000118f17000 CR4: 0000000000750ef0 PKRU: 55555554 Call Trace: mt76x02_send_tx_status+0x1d2/0xeb0 mt76x02_tx_status_data+0x8e/0xd0 mt76u_tx_status_data+0xe1/0x240 process_one_work+0x92b/0x1460 worker_thread+0x95/0xe00 kthread+0x3a1/0x480 ret_from_fork+0x1f/0x30 Modules linked in: --[ end trace 8df5d20fc5040f65 ]-- RIP: 0010:mt76x02_mac_fill_tx_status.isra.0+0x82c/0x9e0 Code: c5 48 b8 00 00 00 00 00 fc ff df 80 3c 02 00 0f 85 94 01 00 00 48 b8 00 00 00 00 00 fc ff df 4d 8b 34 24 4c 89 f2 48 c1 ea 03 <0f> b6 04 02 84 c0 74 08 3c 03 0f 8e 89 01 00 00 41 8b 16 41 0f b7 RSP: 0018:ffffc900005af988 EFLAGS: 00010246 RAX: dffffc0000000000 RBX: ffffc900005afae8 RCX: 0000000000000000 RDX: 0000000000000000 RSI: ffffffff832fc661 RDI: ffffc900005afc2a RBP: ffffc900005afae0 R08: 0000000000000001 R09: fffff520000b5f3c R10: 0000000000000003 R11: fffff520000b5f3b R12: ffff88810b6132d8 R13: 000000000000ffff R14: 0000000000000000 R15: ffffc900005afc28 FS: 0000000000000000(0000) GS:ffff88811aa00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fa0eda6a000 CR3: 0000000118f17000 CR4: 0000000000750ef0 PKRU: 55555554 Moreover move stat_work schedule out of the for loop. Reported-by: Dokyung Song <dokyungs(a)yonsei.ac.kr> Co-developed-by: Deren Wu <deren.wu(a)mediatek.com> Signed-off-by: Deren Wu <deren.wu(a)mediatek.com> Signed-off-by: Lorenzo Bianconi <lorenzo(a)kernel.org> Signed-off-by: Felix Fietkau <nbd(a)nbd.name> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Conflicts: drivers/net/wireless/mediatek/mt76/usb.c [Context conflicts.] Signed-off-by: Jiacheng Yu <yujiacheng3(a)huawei.com> --- drivers/net/wireless/mediatek/mt76/usb.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) diff --git a/drivers/net/wireless/mediatek/mt76/usb.c b/drivers/net/wireless/mediatek/mt76/usb.c index f1ae9ff835b2..7426fc0894b1 100644 --- a/drivers/net/wireless/mediatek/mt76/usb.c +++ b/drivers/net/wireless/mediatek/mt76/usb.c @@ -800,6 +800,9 @@ static void mt76u_tx_worker(struct mt76_worker *w) bool wake; int i; + if (!test_bit(MT76_STATE_RUNNING, &dev->phy.state)) + return; + for (i = 0; i < IEEE80211_NUM_ACS; i++) { q = dev->q_tx[i]; @@ -821,13 +824,13 @@ static void mt76u_tx_worker(struct mt76_worker *w) wake_up(&dev->tx_wait); mt76_txq_schedule(&dev->phy, i); - - if (dev->drv->tx_status_data && - !test_and_set_bit(MT76_READING_STATS, &dev->phy.state)) - queue_work(dev->wq, &dev->usb.stat_work); if (wake) ieee80211_wake_queue(dev->hw, i); } + + if (dev->drv->tx_status_data && + !test_and_set_bit(MT76_READING_STATS, &dev->phy.state)) + queue_work(dev->wq, &dev->usb.stat_work); } static void mt76u_tx_status_data(struct work_struct *work) -- 2.43.0

2 1

[PATCH OLK-5.10] RDMA/bnxt_re: Prevent handling any completions after qp destroy
by Jiacheng Yu 06 Jan '26

06 Jan '26

From: Kashyap Desai <kashyap.desai(a)broadcom.com> stable inclusion from stable-v5.15.124 commit b79a0e71d6e8692e0b6da05f8aaa7d69191cf7e7 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/12854 CVE: CVE-2023-54048 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit b5bbc6551297447d3cca55cf907079e206e9cd82 ] HW may generate completions that indicates QP is destroyed. Driver should not be scheduling any more completion handlers for this QP, after the QP is destroyed. Since CQs are active during the QP destroy, driver may still schedule completion handlers. This can cause a race where the destroy_cq and poll_cq running simultaneously. Snippet of kernel panic while doing bnxt_re driver load unload in loop. This indicates a poll after the CQ is freed. [77786.481636] Call Trace: [77786.481640] <TASK> [77786.481644] bnxt_re_poll_cq+0x14a/0x620 [bnxt_re] [77786.481658] ? kvm_clock_read+0x14/0x30 [77786.481693] __ib_process_cq+0x57/0x190 [ib_core] [77786.481728] ib_cq_poll_work+0x26/0x80 [ib_core] [77786.481761] process_one_work+0x1e5/0x3f0 [77786.481768] worker_thread+0x50/0x3a0 [77786.481785] ? __pfx_worker_thread+0x10/0x10 [77786.481790] kthread+0xe2/0x110 [77786.481794] ? __pfx_kthread+0x10/0x10 [77786.481797] ret_from_fork+0x2c/0x50 To avoid this, complete all completion handlers before returning the destroy QP. If free_cq is called soon after destroy_qp, IB stack will cancel the CQ work before invoking the destroy_cq verb and this will prevent any race mentioned. Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver") Signed-off-by: Kashyap Desai <kashyap.desai(a)broadcom.com> Signed-off-by: Selvin Xavier <selvin.xavier(a)broadcom.com> Link: https://lore.kernel.org/r/1689322969-25402-2-git-send-email-selvin.xavier@b… Signed-off-by: Leon Romanovsky <leon(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Conflicts: drivers/infiniband/hw/bnxt_re/ib_verbs.c [context conflicts.] Signed-off-by: Jiacheng Yu <yujiacheng3(a)huawei.com> --- drivers/infiniband/hw/bnxt_re/ib_verbs.c | 12 ++++++++++++ drivers/infiniband/hw/bnxt_re/qplib_fp.c | 18 ++++++++++++++++++ drivers/infiniband/hw/bnxt_re/qplib_fp.h | 1 + 3 files changed, 31 insertions(+) diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index a0d7777acb6d..867f7fadbb25 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -788,7 +788,10 @@ static int bnxt_re_destroy_gsi_sqp(struct bnxt_re_qp *qp) int bnxt_re_destroy_qp(struct ib_qp *ib_qp, struct ib_udata *udata) { struct bnxt_re_qp *qp = container_of(ib_qp, struct bnxt_re_qp, ib_qp); + struct bnxt_qplib_qp *qplib_qp = &qp->qplib_qp; struct bnxt_re_dev *rdev = qp->rdev; + struct bnxt_qplib_nq *scq_nq = NULL; + struct bnxt_qplib_nq *rcq_nq = NULL; unsigned int flags; int rc; @@ -822,6 +825,15 @@ int bnxt_re_destroy_qp(struct ib_qp *ib_qp, struct ib_udata *udata) ib_umem_release(qp->rumem); ib_umem_release(qp->sumem); + /* Flush all the entries of notification queue associated with + * given qp. + */ + scq_nq = qplib_qp->scq->nq; + rcq_nq = qplib_qp->rcq->nq; + bnxt_re_synchronize_nq(scq_nq); + if (scq_nq != rcq_nq) + bnxt_re_synchronize_nq(rcq_nq); + kfree(qp); return 0; sh_fail: diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index fc560e846b47..5d4ce0a6dc69 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -386,6 +386,24 @@ static void bnxt_qplib_service_nq(struct tasklet_struct *t) spin_unlock_bh(&hwq->lock); } +/* bnxt_re_synchronize_nq - self polling notification queue. + * @nq - notification queue pointer + * + * This function will start polling entries of a given notification queue + * for all pending entries. + * This function is useful to synchronize notification entries while resources + * are going away. + */ + +void bnxt_re_synchronize_nq(struct bnxt_qplib_nq *nq) +{ + int budget = nq->budget; + + nq->budget = nq->hwq.max_elements; + bnxt_qplib_service_nq(&nq->nq_tasklet); + nq->budget = budget; +} + static irqreturn_t bnxt_qplib_nq_irq(int irq, void *dev_instance) { struct bnxt_qplib_nq *nq = dev_instance; diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.h b/drivers/infiniband/hw/bnxt_re/qplib_fp.h index 667f93d90045..98f156a876ee 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.h @@ -546,6 +546,7 @@ int bnxt_qplib_process_flush_list(struct bnxt_qplib_cq *cq, struct bnxt_qplib_cqe *cqe, int num_cqes); void bnxt_qplib_flush_cqn_wq(struct bnxt_qplib_qp *qp); +void bnxt_re_synchronize_nq(struct bnxt_qplib_nq *nq); static inline void *bnxt_qplib_get_swqe(struct bnxt_qplib_q *que, u32 *swq_idx) { -- 2.43.0

2 1

[PATCH OLK-5.10] net: usb: qmi_wwan: initialize MAC header offset in qmimux_rx_fixup
by Jiacheng Yu 06 Jan '26

06 Jan '26

From: Qendrim Maxhuni <qendrim.maxhuni(a)garderos.com> stable inclusion from stable-v5.10.247 commit 0aabccdcec1f4a36f95829ea2263f845bbc77223 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/11602 CVE: CVE-2025-68192 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit e120f46768d98151ece8756ebd688b0e43dc8b29 ] Raw IP packets have no MAC header, leaving skb->mac_header uninitialized. This can trigger kernel panics on ARM64 when xfrm or other subsystems access the offset due to strict alignment checks. Initialize the MAC header to prevent such crashes. This can trigger kernel panics on ARM when running IPsec over the qmimux0 interface. Example trace: Internal error: Oops: 000000009600004f [#1] SMP CPU: 0 UID: 0 PID: 0 Comm: swapper/0 Not tainted 6.12.34-gbe78e49cb433 #1 Hardware name: LS1028A RDB Board (DT) pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--) pc : xfrm_input+0xde8/0x1318 lr : xfrm_input+0x61c/0x1318 sp : ffff800080003b20 Call trace: xfrm_input+0xde8/0x1318 xfrm6_rcv+0x38/0x44 xfrm6_esp_rcv+0x48/0xa8 ip6_protocol_deliver_rcu+0x94/0x4b0 ip6_input_finish+0x44/0x70 ip6_input+0x44/0xc0 ipv6_rcv+0x6c/0x114 __netif_receive_skb_one_core+0x5c/0x8c __netif_receive_skb+0x18/0x60 process_backlog+0x78/0x17c __napi_poll+0x38/0x180 net_rx_action+0x168/0x2f0 Fixes: c6adf77953bc ("net: usb: qmi_wwan: add qmap mux protocol support") Signed-off-by: Qendrim Maxhuni <qendrim.maxhuni(a)garderos.com> Link: https://patch.msgid.link/20251029075744.105113-1-qendrim.maxhuni@garderos.c… Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Jiacheng Yu <yujiacheng3(a)huawei.com> --- drivers/net/usb/qmi_wwan.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/net/usb/qmi_wwan.c b/drivers/net/usb/qmi_wwan.c index 306ff69a10f0..0aab7a1b9886 100644 --- a/drivers/net/usb/qmi_wwan.c +++ b/drivers/net/usb/qmi_wwan.c @@ -207,6 +207,12 @@ static int qmimux_rx_fixup(struct usbnet *dev, struct sk_buff *skb) return 0; skbn->dev = net; + /* Raw IP packets don't have a MAC header, but other subsystems + * (like xfrm) may still access MAC header offsets, so they must + * be initialized. + */ + skb_reset_mac_header(skbn); + switch (skb->data[offset + qmimux_hdr_sz] & 0xf0) { case 0x40: skbn->protocol = htons(ETH_P_IP); -- 2.43.0

2 1

[PATCH OLK-6.6 00/10] EEVDF null pointer
by Zicheng Qu 06 Jan '26

06 Jan '26

Fernand Sieber (1): sched/fair: Forfeit vruntime on yield Ingo Molnar (1): sched/balancing: Change comment formatting to not overlap Git conflict marker lines Peter Zijlstra (4): sched/fair: Avoid re-setting virtual deadline on 'migrations' sched/eevdf: Remove min_vruntime_copy sched/core: Add comment explaining force-idle vruntime snapshots sched/eevdf: Fix min_vruntime vs avg_vruntime Zicheng Qu (4): sched: Fix kabi breakage of struct sched_entity for on_rq and rel_deadline sched: Fix kabi breakage of struct cfs_rq for min_vruntime_copy Revert "sched/fair: Only increment deadline once on yield" sched: Fix kabi breakage of struct cfs_rq for zero_vruntime include/linux/sched.h | 4 +- kernel/sched/debug.c | 8 +- kernel/sched/fair.c | 277 ++++++++++++++++++++++++++++++---------- kernel/sched/features.h | 4 + kernel/sched/sched.h | 6 +- 5 files changed, 223 insertions(+), 76 deletions(-) -- 2.34.1

2 11

[PATCH OLK-6.6] mlx5: Fix default values in create CQ
by Jiacheng Yu 06 Jan '26

06 Jan '26

From: Akiva Goldberger <agoldberger(a)nvidia.com> stable inclusion from stable-v6.17.9 commit 08469f5393a1a39f26a6e2eb2e8c33187665c1f4 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/11579 CVE: CVE-2025-68209 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit e5eba42f01340f73888dfe560be2806057c25913 ] Currently, CQs without a completion function are assigned the mlx5_add_cq_to_tasklet function by default. This is problematic since only user CQs created through the mlx5_ib driver are intended to use this function. Additionally, all CQs that will use doorbells instead of polling for completions must call mlx5_cq_arm. However, the default CQ creation flow leaves a valid value in the CQ's arm_db field, allowing FW to send interrupts to polling-only CQs in certain corner cases. These two factors would allow a polling-only kernel CQ to be triggered by an EQ interrupt and call a completion function intended only for user CQs, causing a null pointer exception. Some areas in the driver have prevented this issue with one-off fixes but did not address the root cause. This patch fixes the described issue by adding defaults to the create CQ flow. It adds a default dummy completion function to protect against null pointer exceptions, and it sets an invalid command sequence number by default in kernel CQs to prevent the FW from sending an interrupt to the CQ until it is armed. User CQs are responsible for their own initialization values. Callers of mlx5_core_create_cq are responsible for changing the completion function and arming the CQ per their needs. Fixes: cdd04f4d4d71 ("net/mlx5: Add support to create SQ and CQ for ASO") Signed-off-by: Akiva Goldberger <agoldberger(a)nvidia.com> Reviewed-by: Moshe Shemesh <moshe(a)nvidia.com> Signed-off-by: Tariq Toukan <tariqt(a)nvidia.com> Acked-by: Leon Romanovsky <leon(a)kernel.org> Link: https://patch.msgid.link/1762681743-1084694-1-git-send-email-tariqt@nvidia.… Signed-off-by: Paolo Abeni <pabeni(a)redhat.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Conflicts: drivers/net/ethernet/mellanox/mlx5/core/cq.c drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c drivers/net/ethernet/mellanox/mlx5/core/steering/hws/send.c drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_send.c drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c [cq.c, conn.c context conflicts. send.c 2ca62599aa0bbc0e61595531614e0989ba6b3194 not committed. dr_send.c e03cf321882badadc77dfc428f465173523d5f79 not committed] Signed-off-by: Jiacheng Yu <yujiacheng3(a)huawei.com> --- drivers/infiniband/hw/mlx5/cq.c | 11 ++++--- drivers/net/ethernet/mellanox/mlx5/core/cq.c | 23 ++++++++++++-- .../net/ethernet/mellanox/mlx5/core/en_main.c | 1 - .../ethernet/mellanox/mlx5/core/fpga/conn.c | 17 +++++------ .../mellanox/mlx5/core/steering/dr_send.c | 30 +++++-------------- drivers/vdpa/mlx5/net/mlx5_vnet.c | 6 ++-- include/linux/mlx5/cq.h | 1 + 7 files changed, 46 insertions(+), 43 deletions(-) diff --git a/drivers/infiniband/hw/mlx5/cq.c b/drivers/infiniband/hw/mlx5/cq.c index ee9acd58c512..0e2177323154 100644 --- a/drivers/infiniband/hw/mlx5/cq.c +++ b/drivers/infiniband/hw/mlx5/cq.c @@ -1011,15 +1011,18 @@ int mlx5_ib_create_cq(struct ib_cq *ibcq, const struct ib_cq_init_attr *attr, if (cq->create_flags & IB_UVERBS_CQ_FLAGS_IGNORE_OVERRUN) MLX5_SET(cqc, cqc, oi, 1); + if (udata) { + cq->mcq.comp = mlx5_add_cq_to_tasklet; + cq->mcq.tasklet_ctx.comp = mlx5_ib_cq_comp; + } else { + cq->mcq.comp = mlx5_ib_cq_comp; + } + err = mlx5_core_create_cq(dev->mdev, &cq->mcq, cqb, inlen, out, sizeof(out)); if (err) goto err_cqb; mlx5_ib_dbg(dev, "cqn 0x%x\n", cq->mcq.cqn); - if (udata) - cq->mcq.tasklet_ctx.comp = mlx5_ib_cq_comp; - else - cq->mcq.comp = mlx5_ib_cq_comp; cq->mcq.event = mlx5_ib_cq_event; INIT_LIST_HEAD(&cq->wc_list); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cq.c b/drivers/net/ethernet/mellanox/mlx5/core/cq.c index 4caa1b6f40ba..d21c14bb3ecf 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cq.c @@ -66,8 +66,8 @@ void mlx5_cq_tasklet_cb(struct tasklet_struct *t) tasklet_schedule(&ctx->task); } -static void mlx5_add_cq_to_tasklet(struct mlx5_core_cq *cq, - struct mlx5_eqe *eqe) +void mlx5_add_cq_to_tasklet(struct mlx5_core_cq *cq, + struct mlx5_eqe *eqe) { unsigned long flags; struct mlx5_eq_tasklet *tasklet_ctx = cq->tasklet_ctx.priv; @@ -84,7 +84,15 @@ static void mlx5_add_cq_to_tasklet(struct mlx5_core_cq *cq, } spin_unlock_irqrestore(&tasklet_ctx->lock, flags); } +EXPORT_SYMBOL(mlx5_add_cq_to_tasklet); +static void mlx5_core_cq_dummy_cb(struct mlx5_core_cq *cq, struct mlx5_eqe *eqe) +{ + mlx5_core_err(cq->eq->core.dev, + "CQ default completion callback, CQ #%u\n", cq->cqn); +} + +#define MLX5_CQ_INIT_CMD_SN cpu_to_be32(2 << 28) /* Callers must verify outbox status in case of err */ int mlx5_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, u32 *in, int inlen, u32 *out, int outlen) @@ -110,10 +118,19 @@ int mlx5_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, cq->arm_sn = 0; cq->eq = eq; cq->uid = MLX5_GET(create_cq_in, in, uid); + + /* Kernel CQs must set the arm_db address prior to calling + * this function, allowing for the proper value to be + * initialized. User CQs are responsible for their own + * initialization since they do not use the arm_db field. + */ + if (cq->arm_db) + *cq->arm_db = MLX5_CQ_INIT_CMD_SN; + refcount_set(&cq->refcount, 1); init_completion(&cq->free); if (!cq->comp) - cq->comp = mlx5_add_cq_to_tasklet; + cq->comp = mlx5_core_cq_dummy_cb; /* assuming CQ will be deleted before the EQ */ cq->tasklet_ctx.priv = &eq->tasklet_ctx; INIT_LIST_HEAD(&cq->tasklet_ctx.list); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index ec1b582cb9d5..08f771b60014 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -1936,7 +1936,6 @@ static int mlx5e_alloc_cq_common(struct mlx5e_priv *priv, mcq->set_ci_db = cq->wq_ctrl.db.db; mcq->arm_db = cq->wq_ctrl.db.db + 1; *mcq->set_ci_db = 0; - *mcq->arm_db = 0; mcq->vector = param->eq_ix; mcq->comp = mlx5e_completion_event; mcq->event = mlx5e_cq_error_event; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c index c4de6bf8d1b6..4db17fa397be 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c @@ -421,6 +421,14 @@ static int mlx5_fpga_conn_create_cq(struct mlx5_fpga_conn *conn, int cq_size) __be64 *pas; u32 i; + conn->cq.mcq.cqe_sz = 64; + conn->cq.mcq.set_ci_db = conn->cq.wq_ctrl.db.db; + conn->cq.mcq.arm_db = conn->cq.wq_ctrl.db.db + 1; + *conn->cq.mcq.set_ci_db = 0; + conn->cq.mcq.vector = 0; + conn->cq.mcq.comp = mlx5_fpga_conn_cq_complete; + conn->cq.mcq.uar = fdev->conn_res.uar; + cq_size = roundup_pow_of_two(cq_size); MLX5_SET(cqc, temp_cqc, log_cq_size, ilog2(cq_size)); @@ -468,16 +476,7 @@ static int mlx5_fpga_conn_create_cq(struct mlx5_fpga_conn *conn, int cq_size) if (err) goto err_cqwq; - conn->cq.mcq.cqe_sz = 64; - conn->cq.mcq.set_ci_db = conn->cq.wq_ctrl.db.db; - conn->cq.mcq.arm_db = conn->cq.wq_ctrl.db.db + 1; - *conn->cq.mcq.set_ci_db = 0; - *conn->cq.mcq.arm_db = 0; - conn->cq.mcq.vector = 0; - conn->cq.mcq.comp = mlx5_fpga_conn_cq_complete; - conn->cq.mcq.uar = fdev->conn_res.uar; tasklet_setup(&conn->cq.tasklet, mlx5_fpga_conn_cq_tasklet); - mlx5_fpga_dbg(fdev, "Created CQ #0x%x\n", conn->cq.mcq.cqn); goto out; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c index f57c84e5128b..30401ed45b17 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c @@ -1049,12 +1049,6 @@ static int dr_prepare_qp_to_rts(struct mlx5dr_domain *dmn) return 0; } -static void dr_cq_complete(struct mlx5_core_cq *mcq, - struct mlx5_eqe *eqe) -{ - pr_err("CQ completion CQ: #%u\n", mcq->cqn); -} - static struct mlx5dr_cq *dr_create_cq(struct mlx5_core_dev *mdev, struct mlx5_uars_page *uar, size_t ncqe) @@ -1089,6 +1083,14 @@ static struct mlx5dr_cq *dr_create_cq(struct mlx5_core_dev *mdev, cqe->op_own = MLX5_CQE_INVALID << 4 | MLX5_CQE_OWNER_MASK; } + cq->mcq.cqe_sz = 64; + cq->mcq.set_ci_db = cq->wq_ctrl.db.db; + cq->mcq.arm_db = cq->wq_ctrl.db.db + 1; + *cq->mcq.set_ci_db = 0; + cq->mcq.vector = 0; + cq->mcq.uar = uar; + cq->mdev = mdev; + inlen = MLX5_ST_SZ_BYTES(create_cq_in) + sizeof(u64) * cq->wq_ctrl.buf.npages; in = kvzalloc(inlen, GFP_KERNEL); @@ -1112,28 +1114,12 @@ static struct mlx5dr_cq *dr_create_cq(struct mlx5_core_dev *mdev, pas = (__be64 *)MLX5_ADDR_OF(create_cq_in, in, pas); mlx5_fill_page_frag_array(&cq->wq_ctrl.buf, pas); - cq->mcq.comp = dr_cq_complete; - err = mlx5_core_create_cq(mdev, &cq->mcq, in, inlen, out, sizeof(out)); kvfree(in); if (err) goto err_cqwq; - cq->mcq.cqe_sz = 64; - cq->mcq.set_ci_db = cq->wq_ctrl.db.db; - cq->mcq.arm_db = cq->wq_ctrl.db.db + 1; - *cq->mcq.set_ci_db = 0; - - /* set no-zero value, in order to avoid the HW to run db-recovery on - * CQ that used in polling mode. - */ - *cq->mcq.arm_db = cpu_to_be32(2 << 28); - - cq->mcq.vector = 0; - cq->mcq.uar = uar; - cq->mdev = mdev; - return cq; err_cqwq: diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c index e538e7e0874f..743ced205fca 100644 --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c @@ -553,6 +553,8 @@ static int cq_create(struct mlx5_vdpa_net *ndev, u16 idx, u32 num_ent) vcq->mcq.set_ci_db = vcq->db.db; vcq->mcq.arm_db = vcq->db.db + 1; vcq->mcq.cqe_sz = 64; + vcq->mcq.comp = mlx5_vdpa_cq_comp; + vcq->cqe = num_ent; err = cq_frag_buf_alloc(ndev, &vcq->buf, num_ent); if (err) @@ -592,10 +594,6 @@ static int cq_create(struct mlx5_vdpa_net *ndev, u16 idx, u32 num_ent) if (err) goto err_vec; - vcq->mcq.comp = mlx5_vdpa_cq_comp; - vcq->cqe = num_ent; - vcq->mcq.set_ci_db = vcq->db.db; - vcq->mcq.arm_db = vcq->db.db + 1; mlx5_cq_arm(&mvq->cq.mcq, MLX5_CQ_DB_REQ_NOT, uar_page, mvq->cq.mcq.cons_index); kfree(in); return 0; diff --git a/include/linux/mlx5/cq.h b/include/linux/mlx5/cq.h index cb15308b5cb0..804432fd0f27 100644 --- a/include/linux/mlx5/cq.h +++ b/include/linux/mlx5/cq.h @@ -183,6 +183,7 @@ static inline void mlx5_cq_put(struct mlx5_core_cq *cq) complete(&cq->free); } +void mlx5_add_cq_to_tasklet(struct mlx5_core_cq *cq, struct mlx5_eqe *eqe); int mlx5_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, u32 *in, int inlen, u32 *out, int outlen); int mlx5_core_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, -- 2.43.0

2 1

[PATCH OLK-6.6] nvme-fc: use lock accessing port_state and rport state
by Lin Yujun 06 Jan '26

06 Jan '26

From: Daniel Wagner <wagi(a)kernel.org> stable inclusion from stable-v6.6.117 commit de3d91af47bc015031e7721b100a29989f6498a5 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/12600 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 891cdbb162ccdb079cd5228ae43bdeebce8597ad ] nvme_fc_unregister_remote removes the remote port on a lport object at any point in time when there is no active association. This races with with the reconnect logic, because nvme_fc_create_association is not taking a lock to check the port_state and atomically increase the active count on the rport. Reported-by: Shinichiro Kawasaki <shinichiro.kawasaki(a)wdc.com> Closes: https://lore.kernel.org/all/u4ttvhnn7lark5w3sgrbuy2rxupcvosp4qmvj46nwzgeo5a… Signed-off-by: Daniel Wagner <wagi(a)kernel.org> Reviewed-by: Hannes Reinecke <hare(a)suse.de> Signed-off-by: Keith Busch <kbusch(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Lin Yujun <linyujun809(a)h-partners.com> --- drivers/nvme/host/fc.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c index 3e0da2422b33..bf9ab0725764 100644 --- a/drivers/nvme/host/fc.c +++ b/drivers/nvme/host/fc.c @@ -3022,15 +3022,21 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl) unsigned long flags; int ret; ++ctrl->ctrl.nr_reconnects; - if (ctrl->rport->remoteport.port_state != FC_OBJSTATE_ONLINE) + spin_lock_irqsave(&ctrl->rport->lock, flags); + if (ctrl->rport->remoteport.port_state != FC_OBJSTATE_ONLINE) { + spin_unlock_irqrestore(&ctrl->rport->lock, flags); return -ENODEV; + } - if (nvme_fc_ctlr_active_on_rport(ctrl)) + if (nvme_fc_ctlr_active_on_rport(ctrl)) { + spin_unlock_irqrestore(&ctrl->rport->lock, flags); return -ENOTUNIQ; + } + spin_unlock_irqrestore(&ctrl->rport->lock, flags); dev_info(ctrl->ctrl.device, "NVME-FC{%d}: create association : host wwpn 0x%016llx " " rport wwpn 0x%016llx: NQN \"%s\"\n", ctrl->cnum, ctrl->lport->localport.port_name, -- 2.34.1

2 1

[PATCH OLK-5.10] nvme-fc: use lock accessing port_state and rport state
by Lin Yujun 06 Jan '26

06 Jan '26

From: Daniel Wagner <wagi(a)kernel.org> stable inclusion from stable-v5.10.247 commit de3d91af47bc015031e7721b100a29989f6498a5 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/12600 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 891cdbb162ccdb079cd5228ae43bdeebce8597ad ] nvme_fc_unregister_remote removes the remote port on a lport object at any point in time when there is no active association. This races with with the reconnect logic, because nvme_fc_create_association is not taking a lock to check the port_state and atomically increase the active count on the rport. Reported-by: Shinichiro Kawasaki <shinichiro.kawasaki(a)wdc.com> Closes: https://lore.kernel.org/all/u4ttvhnn7lark5w3sgrbuy2rxupcvosp4qmvj46nwzgeo5a… Signed-off-by: Daniel Wagner <wagi(a)kernel.org> Reviewed-by: Hannes Reinecke <hare(a)suse.de> Signed-off-by: Keith Busch <kbusch(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Lin Yujun <linyujun809(a)h-partners.com> --- drivers/nvme/host/fc.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c index 55a329f20148..04bee22ec5ce 100644 --- a/drivers/nvme/host/fc.c +++ b/drivers/nvme/host/fc.c @@ -3040,15 +3040,21 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl) int ret; bool changed; ++ctrl->ctrl.nr_reconnects; - if (ctrl->rport->remoteport.port_state != FC_OBJSTATE_ONLINE) + spin_lock_irqsave(&ctrl->rport->lock, flags); + if (ctrl->rport->remoteport.port_state != FC_OBJSTATE_ONLINE) { + spin_unlock_irqrestore(&ctrl->rport->lock, flags); return -ENODEV; + } - if (nvme_fc_ctlr_active_on_rport(ctrl)) + if (nvme_fc_ctlr_active_on_rport(ctrl)) { + spin_unlock_irqrestore(&ctrl->rport->lock, flags); return -ENOTUNIQ; + } + spin_unlock_irqrestore(&ctrl->rport->lock, flags); dev_info(ctrl->ctrl.device, "NVME-FC{%d}: create association : host wwpn 0x%016llx " " rport wwpn 0x%016llx: NQN \"%s\"\n", ctrl->cnum, ctrl->lport->localport.port_name, -- 2.34.1

2 1

[PATCH v2 openEuler-1.0-LTS 0/4] add mcs support for migrate pages
by Wupeng Ma 06 Jan '26

06 Jan '26

Backport the following patches for OLK-5.10: - e14f9d96acb7 mm/memory-failure: support disabling soft offline for HugeTLB pages - 2f0c01a690d1 mm/memory-failure: userspace controls soft-offlining pages - 9bb3fd60f91e arm64: mm: Add copy mc support for all migrate_page - 45dbef4c04f6 mm: page_eject: Add mc support during offline page - b7f5b9ef641a mm: Update PF_COREDUMP_MCS to PF_MCS - 17d08829674a mm/hwpoison: add migrate_page_mc_extra() - c10a520c4102 mm/hwpoison: introduce copy_mc_highpages - d689cd1c1479 mm/hwpoison: arm64: introduce copy_mc_highpage Add mcs support for migrate page & support disabling soft offline for HugeTLB pages. Since UCE kernel recovery is needed by this. This should be enable with the following step: - echo 1 > /proc/sys/kernel/uce_kernel_recovery Disable soft offline support for hugetlb with the following step: - echo 3 > /proc/sys/vm/enable_soft_offline Since ARCH_ENABLE_HUGEPAGE_MIGRATION is not enabled in openeuler_defconfig in arm64. There is no need to merge soft-offining related patches. Change log since v1: - add patch list from OLK-5.10 - add ARCH_ENABLE_HUGEPAGE_MIGRATION related comment - bugfix for __copy_huge_page Jiaqi Yan (1): mm/memory-failure: userspace controls soft-offlining pages Kyle Meyer (1): mm/memory-failure: support disabling soft offline for HugeTLB pages Wupeng Ma (2): uce: add copy_mc_highpage{s} arm64: mm: Add copy mc support for migrate_page .../ABI/testing/sysfs-memory-page-offline | 3 + include/linux/highmem.h | 55 +++++++++++++ include/linux/mm.h | 1 + kernel/sysctl.c | 9 +++ mm/memory-failure.c | 25 +++++- mm/migrate.c | 79 ++++++++++++++++--- 6 files changed, 162 insertions(+), 10 deletions(-) -- 2.43.0

2 5