- Kernel - mailweb.openeuler.org

[PATCH OLK-5.10] RDMA/rxe: Fix the failure of ibv_query_device() and ibv_query_device_ex() tests
by Li Xiasong 28 Feb '26

28 Feb '26

From: Zhu Yanjun <yanjun.zhu(a)linux.dev> mainline inclusion from mainline-v6.14 commit 8ce2eb9dfac8743d1c423b86339336a5b6a6069e category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/13658 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- In rdma-core, the following failures appear. " $ ./build/bin/run_tests.py -k device ssssssss....FF........s ====================================================================== FAIL: test_query_device (tests.test_device.DeviceTest.test_query_device) Test ibv_query_device() ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/ubuntu/rdma-core/tests/test_device.py", line 63, in test_query_device self.verify_device_attr(attr, dev) File "/home/ubuntu/rdma-core/tests/test_device.py", line 200, in verify_device_attr assert attr.sys_image_guid != 0 ^^^^^^^^^^^^^^^^^^^^^^^^ AssertionError ====================================================================== FAIL: test_query_device_ex (tests.test_device.DeviceTest.test_query_device_ex) Test ibv_query_device_ex() ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/ubuntu/rdma-core/tests/test_device.py", line 222, in test_query_device_ex self.verify_device_attr(attr_ex.orig_attr, dev) File "/home/ubuntu/rdma-core/tests/test_device.py", line 200, in verify_device_attr assert attr.sys_image_guid != 0 ^^^^^^^^^^^^^^^^^^^^^^^^ AssertionError " The root cause is: before a net device is set with rxe, this net device is used to generate a sys_image_guid. Fixes: 2ac5415022d1 ("RDMA/rxe: Remove the direct link to net_device") Signed-off-by: Zhu Yanjun <yanjun.zhu(a)linux.dev> Link: https://patch.msgid.link/20250302215444.3742072-1-yanjun.zhu@linux.dev Reviewed-by: Daisuke Matsuda <matsuda-daisuke(a)fujitsu.com> Tested-by: Daisuke Matsuda <matsuda-daisuke(a)fujitsu.com> Signed-off-by: Leon Romanovsky <leon(a)kernel.org> Conflicts: drivers/infiniband/sw/rxe/rxe.c [conflicts due to not merge 3225717f6dfa ("RDMA/rxe: Replace red-black trees by xarrays").] Signed-off-by: Li Xiasong <lixiasong1(a)huawei.com> --- drivers/infiniband/sw/rxe/rxe.c | 25 ++++++------------------- 1 file changed, 6 insertions(+), 19 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe.c b/drivers/infiniband/sw/rxe/rxe.c index 79740ec28399..3e8ab4563253 100644 --- a/drivers/infiniband/sw/rxe/rxe.c +++ b/drivers/infiniband/sw/rxe/rxe.c @@ -40,10 +40,8 @@ void rxe_dealloc(struct ib_device *ib_dev) } /* initialize rxe device parameters */ -static void rxe_init_device_param(struct rxe_dev *rxe) +static void rxe_init_device_param(struct rxe_dev *rxe, struct net_device *ndev) { - struct net_device *ndev; - rxe->max_inline_data = RXE_MAX_INLINE_DATA; rxe->attr.vendor_id = RXE_VENDOR_ID; @@ -74,15 +72,9 @@ static void rxe_init_device_param(struct rxe_dev *rxe) rxe->attr.max_pkeys = RXE_MAX_PKEYS; rxe->attr.local_ca_ack_delay = RXE_LOCAL_CA_ACK_DELAY; - ndev = rxe_ib_device_get_netdev(&rxe->ib_dev); - if (!ndev) - return; - addrconf_addr_eui48((unsigned char *)&rxe->attr.sys_image_guid, ndev->dev_addr); - dev_put(ndev); - rxe->max_ucontext = RXE_MAX_UCONTEXT; } @@ -115,18 +107,13 @@ static void rxe_init_port_param(struct rxe_port *port) /* initialize port state, note IB convention that HCA ports are always * numbered from 1 */ -static void rxe_init_ports(struct rxe_dev *rxe) +static void rxe_init_ports(struct rxe_dev *rxe, struct net_device *ndev) { struct rxe_port *port = &rxe->port; - struct net_device *ndev; rxe_init_port_param(port); - ndev = rxe_ib_device_get_netdev(&rxe->ib_dev); - if (!ndev) - return; addrconf_addr_eui48((unsigned char *)&port->port_guid, ndev->dev_addr); - dev_put(ndev); spin_lock_init(&port->port_lock); } @@ -210,14 +197,14 @@ static int rxe_init_pools(struct rxe_dev *rxe) } /* initialize rxe device state */ -static int rxe_init(struct rxe_dev *rxe) +static int rxe_init(struct rxe_dev *rxe, struct net_device *ndev) { int err; /* init default device parameters */ - rxe_init_device_param(rxe); + rxe_init_device_param(rxe, ndev); - rxe_init_ports(rxe); + rxe_init_ports(rxe, ndev); err = rxe_init_pools(rxe); if (err) @@ -255,7 +242,7 @@ int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name, { int err; - err = rxe_init(rxe); + err = rxe_init(rxe, ndev); if (err) return err; -- 2.34.1

2 1

[PATCH OLK-6.6] ipvlan: Make the addrs_lock be per port
by Li Xiasong 28 Feb '26

28 Feb '26

From: Dmitry Skorodumov <dskr99(a)gmail.com> stable inclusion from stable-v6.6.122 commit 04ba6de6eff61238e5397c14ac26a6578c7735a5 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/13636 CVE: CVE-2026-23103 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit d3ba32162488283c0a4c5bedd8817aec91748802 ] Make the addrs_lock be per port, not per ipvlan dev. Initial code seems to be written in the assumption, that any address change must occur under RTNL. But it is not so for the case of IPv6. So 1) Introduce per-port addrs_lock. 2) It was needed to fix places where it was forgotten to take lock (ipvlan_open/ipvlan_close) This appears to be a very minor problem though. Since it's highly unlikely that ipvlan_add_addr() will be called on 2 CPU simultaneously. But nevertheless, this could cause: 1) False-negative of ipvlan_addr_busy(): one interface iterated through all port->ipvlans + ipvlan->addrs under some ipvlan spinlock, and another added IP under its own lock. Though this is only possible for IPv6, since looks like only ipvlan_addr6_event() can be called without rtnl_lock. 2) Race since ipvlan_ht_addr_add(port) is called under different ipvlan->addrs_lock locks This should not affect performance, since add/remove IP is a rare situation and spinlock is not taken on fast paths. Fixes: 8230819494b3 ("ipvlan: use per device spinlock to protect addrs list updates") Signed-off-by: Dmitry Skorodumov <skorodumov.dmitry(a)huawei.com> Reviewed-by: Paolo Abeni <pabeni(a)redhat.com> Link: https://patch.msgid.link/20260112142417.4039566-2-skorodumov.dmitry@huawei.… Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Conflicts: drivers/net/ipvlan/ipvlan.h drivers/net/ipvlan/ipvlan_main.c [conflicts due to merged 513b49cadd13 ("ipvlan: Introduce local xmit queue for l2e mode")] Signed-off-by: Li Xiasong <lixiasong1(a)huawei.com> --- drivers/net/ipvlan/ipvlan.h | 2 +- drivers/net/ipvlan/ipvlan_core.c | 16 +++++------ drivers/net/ipvlan/ipvlan_main.c | 49 +++++++++++++++++++------------- 3 files changed, 37 insertions(+), 30 deletions(-) diff --git a/drivers/net/ipvlan/ipvlan.h b/drivers/net/ipvlan/ipvlan.h index 4cf12fbe8f62..85a1fc8e1597 100644 --- a/drivers/net/ipvlan/ipvlan.h +++ b/drivers/net/ipvlan/ipvlan.h @@ -74,7 +74,6 @@ struct ipvl_dev { DECLARE_BITMAP(mac_filters, IPVLAN_MAC_FILTER_SIZE); netdev_features_t sfeatures; u32 msg_enable; - spinlock_t addrs_lock; #if IS_ENABLED(CONFIG_IPVLAN_L2E) int local_packets_cached; unsigned long local_timeout; @@ -101,6 +100,7 @@ struct ipvl_port { struct net_device *dev; possible_net_t pnet; struct hlist_head hlhead[IPVLAN_HASH_SIZE]; + spinlock_t addrs_lock; /* guards hash-table and addrs */ struct list_head ipvlans; u16 mode; u16 flags; diff --git a/drivers/net/ipvlan/ipvlan_core.c b/drivers/net/ipvlan/ipvlan_core.c index 65da4c23f528..802b90919304 100644 --- a/drivers/net/ipvlan/ipvlan_core.c +++ b/drivers/net/ipvlan/ipvlan_core.c @@ -104,17 +104,15 @@ void ipvlan_ht_addr_del(struct ipvl_addr *addr) struct ipvl_addr *ipvlan_find_addr(const struct ipvl_dev *ipvlan, const void *iaddr, bool is_v6) { - struct ipvl_addr *addr, *ret = NULL; + struct ipvl_addr *addr; - rcu_read_lock(); - list_for_each_entry_rcu(addr, &ipvlan->addrs, anode) { - if (addr_equal(is_v6, addr, iaddr)) { - ret = addr; - break; - } + assert_spin_locked(&ipvlan->port->addrs_lock); + + list_for_each_entry(addr, &ipvlan->addrs, anode) { + if (addr_equal(is_v6, addr, iaddr)) + return addr; } - rcu_read_unlock(); - return ret; + return NULL; } bool ipvlan_addr_busy(struct ipvl_port *port, void *iaddr, bool is_v6) diff --git a/drivers/net/ipvlan/ipvlan_main.c b/drivers/net/ipvlan/ipvlan_main.c index 3870d7b2b652..ab56bebe2de7 100644 --- a/drivers/net/ipvlan/ipvlan_main.c +++ b/drivers/net/ipvlan/ipvlan_main.c @@ -128,6 +128,7 @@ static int ipvlan_port_create(struct net_device *dev) for (idx = 0; idx < IPVLAN_HASH_SIZE; idx++) INIT_HLIST_HEAD(&port->hlhead[idx]); + spin_lock_init(&port->addrs_lock); skb_queue_head_init(&port->backlog); INIT_WORK(&port->wq, ipvlan_process_multicast); ida_init(&port->ida); @@ -261,6 +262,7 @@ static void ipvlan_uninit(struct net_device *dev) static int ipvlan_open(struct net_device *dev) { struct ipvl_dev *ipvlan = netdev_priv(dev); + struct ipvl_port *port = ipvlan->port; struct ipvl_addr *addr; if (ipvlan->port->mode == IPVLAN_MODE_L3 || @@ -272,10 +274,10 @@ static int ipvlan_open(struct net_device *dev) #if IS_ENABLED(CONFIG_IPVLAN_L2E) ipvlan_local_init(dev); #endif - rcu_read_lock(); - list_for_each_entry_rcu(addr, &ipvlan->addrs, anode) + spin_lock_bh(&port->addrs_lock); + list_for_each_entry(addr, &ipvlan->addrs, anode) ipvlan_ht_addr_add(ipvlan, addr); - rcu_read_unlock(); + spin_unlock_bh(&port->addrs_lock); return 0; } @@ -291,10 +293,10 @@ static int ipvlan_stop(struct net_device *dev) #if IS_ENABLED(CONFIG_IPVLAN_L2E) ipvlan_local_uninit(dev); #endif - rcu_read_lock(); - list_for_each_entry_rcu(addr, &ipvlan->addrs, anode) + spin_lock_bh(&ipvlan->port->addrs_lock); + list_for_each_entry(addr, &ipvlan->addrs, anode) ipvlan_ht_addr_del(addr); - rcu_read_unlock(); + spin_unlock_bh(&ipvlan->port->addrs_lock); return 0; } @@ -665,7 +667,6 @@ int ipvlan_link_new(struct net *src_net, struct net_device *dev, if (!tb[IFLA_MTU]) ipvlan_adjust_mtu(ipvlan, phy_dev); INIT_LIST_HEAD(&ipvlan->addrs); - spin_lock_init(&ipvlan->addrs_lock); /* TODO Probably put random address here to be presented to the * world but keep using the physical-dev address for the outgoing @@ -743,13 +744,13 @@ void ipvlan_link_delete(struct net_device *dev, struct list_head *head) struct ipvl_dev *ipvlan = netdev_priv(dev); struct ipvl_addr *addr, *next; - spin_lock_bh(&ipvlan->addrs_lock); + spin_lock_bh(&ipvlan->port->addrs_lock); list_for_each_entry_safe(addr, next, &ipvlan->addrs, anode) { ipvlan_ht_addr_del(addr); list_del_rcu(&addr->anode); kfree_rcu(addr, rcu); } - spin_unlock_bh(&ipvlan->addrs_lock); + spin_unlock_bh(&ipvlan->port->addrs_lock); ida_simple_remove(&ipvlan->port->ida, dev->dev_id); list_del_rcu(&ipvlan->pnode); @@ -896,6 +897,8 @@ static int ipvlan_add_addr(struct ipvl_dev *ipvlan, void *iaddr, bool is_v6) { struct ipvl_addr *addr; + assert_spin_locked(&ipvlan->port->addrs_lock); + addr = kzalloc(sizeof(struct ipvl_addr), GFP_ATOMIC); if (!addr) return -ENOMEM; @@ -926,16 +929,16 @@ static void ipvlan_del_addr(struct ipvl_dev *ipvlan, void *iaddr, bool is_v6) { struct ipvl_addr *addr; - spin_lock_bh(&ipvlan->addrs_lock); + spin_lock_bh(&ipvlan->port->addrs_lock); addr = ipvlan_find_addr(ipvlan, iaddr, is_v6); if (!addr) { - spin_unlock_bh(&ipvlan->addrs_lock); + spin_unlock_bh(&ipvlan->port->addrs_lock); return; } ipvlan_ht_addr_del(addr); list_del_rcu(&addr->anode); - spin_unlock_bh(&ipvlan->addrs_lock); + spin_unlock_bh(&ipvlan->port->addrs_lock); kfree_rcu(addr, rcu); } @@ -957,14 +960,14 @@ static int ipvlan_add_addr6(struct ipvl_dev *ipvlan, struct in6_addr *ip6_addr) { int ret = -EINVAL; - spin_lock_bh(&ipvlan->addrs_lock); + spin_lock_bh(&ipvlan->port->addrs_lock); if (ipvlan_addr_busy(ipvlan->port, ip6_addr, true)) netif_err(ipvlan, ifup, ipvlan->dev, "Failed to add IPv6=%pI6c addr for %s intf\n", ip6_addr, ipvlan->dev->name); else ret = ipvlan_add_addr(ipvlan, ip6_addr, true); - spin_unlock_bh(&ipvlan->addrs_lock); + spin_unlock_bh(&ipvlan->port->addrs_lock); return ret; } @@ -1003,21 +1006,24 @@ static int ipvlan_addr6_validator_event(struct notifier_block *unused, struct in6_validator_info *i6vi = (struct in6_validator_info *)ptr; struct net_device *dev = (struct net_device *)i6vi->i6vi_dev->dev; struct ipvl_dev *ipvlan = netdev_priv(dev); + int ret = NOTIFY_OK; if (!ipvlan_is_valid_dev(dev)) return NOTIFY_DONE; switch (event) { case NETDEV_UP: + spin_lock_bh(&ipvlan->port->addrs_lock); if (ipvlan_addr_busy(ipvlan->port, &i6vi->i6vi_addr, true)) { NL_SET_ERR_MSG(i6vi->extack, "Address already assigned to an ipvlan device"); - return notifier_from_errno(-EADDRINUSE); + ret = notifier_from_errno(-EADDRINUSE); } + spin_unlock_bh(&ipvlan->port->addrs_lock); break; } - return NOTIFY_OK; + return ret; } #endif @@ -1025,14 +1031,14 @@ static int ipvlan_add_addr4(struct ipvl_dev *ipvlan, struct in_addr *ip4_addr) { int ret = -EINVAL; - spin_lock_bh(&ipvlan->addrs_lock); + spin_lock_bh(&ipvlan->port->addrs_lock); if (ipvlan_addr_busy(ipvlan->port, ip4_addr, false)) netif_err(ipvlan, ifup, ipvlan->dev, "Failed to add IPv4=%pI4 on %s intf.\n", ip4_addr, ipvlan->dev->name); else ret = ipvlan_add_addr(ipvlan, ip4_addr, false); - spin_unlock_bh(&ipvlan->addrs_lock); + spin_unlock_bh(&ipvlan->port->addrs_lock); return ret; } @@ -1074,21 +1080,24 @@ static int ipvlan_addr4_validator_event(struct notifier_block *unused, struct in_validator_info *ivi = (struct in_validator_info *)ptr; struct net_device *dev = (struct net_device *)ivi->ivi_dev->dev; struct ipvl_dev *ipvlan = netdev_priv(dev); + int ret = NOTIFY_OK; if (!ipvlan_is_valid_dev(dev)) return NOTIFY_DONE; switch (event) { case NETDEV_UP: + spin_lock_bh(&ipvlan->port->addrs_lock); if (ipvlan_addr_busy(ipvlan->port, &ivi->ivi_addr, false)) { NL_SET_ERR_MSG(ivi->extack, "Address already assigned to an ipvlan device"); - return notifier_from_errno(-EADDRINUSE); + ret = notifier_from_errno(-EADDRINUSE); } + spin_unlock_bh(&ipvlan->port->addrs_lock); break; } - return NOTIFY_OK; + return ret; } static struct notifier_block ipvlan_addr4_notifier_block __read_mostly = { -- 2.34.1

2 1

[PATCH openEuler-1.0-LTS] net/sched: Reserve TC_H_INGRESS (TC_H_CLSACT) for ingress (clsact) Qdiscs
by Dong Chenchen 28 Feb '26

28 Feb '26

From: Peilin Ye <peilin.ye(a)bytedance.com> mainline inclusion from mainline-v6.4-rc5 commit f85fa45d4a9408d98c46c8fa45ba2e3b2f4bf219 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICMJHK Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- Currently it is possible to add e.g. an HTB Qdisc under ffff:fff1 (TC_H_INGRESS, TC_H_CLSACT): $ ip link add name ifb0 type ifb $ tc qdisc add dev ifb0 parent ffff:fff1 htb $ tc qdisc add dev ifb0 clsact Error: Exclusivity flag on, cannot modify. $ drgn ... >>> ifb0 = netdev_get_by_name(prog, "ifb0") >>> qdisc = ifb0.ingress_queue.qdisc_sleeping >>> print(qdisc.ops.id.string_().decode()) htb >>> qdisc.flags.value_() # TCQ_F_INGRESS 2 Only allow ingress and clsact Qdiscs under ffff:fff1. Return -EINVAL for everything else. Make TCQ_F_INGRESS a static flag of ingress and clsact Qdiscs. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Fixes: 1f211a1b929c ("net, sched: add clsact qdisc") Tested-by: Pedro Tammela <pctammela(a)mojatatu.com> Acked-by: Jamal Hadi Salim <jhs(a)mojatatu.com> Reviewed-by: Jamal Hadi Salim <jhs(a)mojatatu.com> Reviewed-by: Vlad Buslov <vladbu(a)nvidia.com> Signed-off-by: Peilin Ye <peilin.ye(a)bytedance.com> Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Conflicts: net/sched/sch_api.c [commit 8ecc7c8a1c16 is not backport] Signed-off-by: Dong Chenchen <dongchenchen2(a)huawei.com> --- net/sched/sch_api.c | 7 ++++++- net/sched/sch_ingress.c | 4 ++-- 2 files changed, 8 insertions(+), 3 deletions(-) diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 2ea9e7753cb0..51f987df0ac7 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -1145,7 +1145,12 @@ static struct Qdisc *qdisc_create(struct net_device *dev, sch->parent = parent; if (handle == TC_H_INGRESS) { - sch->flags |= TCQ_F_INGRESS; + if (!(sch->flags & TCQ_F_INGRESS)) { + NL_SET_ERR_MSG(extack, + "Specified parent ID is reserved for ingress and clsact Qdiscs"); + err = -EINVAL; + goto err_out3; + } handle = TC_H_MAKE(TC_H_INGRESS, 0); lockdep_set_class(qdisc_lock(sch), &qdisc_rx_lock); } else { diff --git a/net/sched/sch_ingress.c b/net/sched/sch_ingress.c index ce3f55259d0d..238dacf41c22 100644 --- a/net/sched/sch_ingress.c +++ b/net/sched/sch_ingress.c @@ -130,7 +130,7 @@ static struct Qdisc_ops ingress_qdisc_ops __read_mostly = { .cl_ops = &ingress_class_ops, .id = "ingress", .priv_size = sizeof(struct ingress_sched_data), - .static_flags = TCQ_F_CPUSTATS, + .static_flags = TCQ_F_INGRESS | TCQ_F_CPUSTATS, .init = ingress_init, .destroy = ingress_destroy, .dump = ingress_dump, @@ -262,7 +262,7 @@ static struct Qdisc_ops clsact_qdisc_ops __read_mostly = { .cl_ops = &clsact_class_ops, .id = "clsact", .priv_size = sizeof(struct clsact_sched_data), - .static_flags = TCQ_F_CPUSTATS, + .static_flags = TCQ_F_INGRESS | TCQ_F_CPUSTATS, .init = clsact_init, .destroy = clsact_destroy, .dump = ingress_dump, -- 2.25.1

2 1

[PATCH OLK-6.6 0/2] CVE-2026-23096
by Wang Tao 28 Feb '26

28 Feb '26

CVE-2026-23096 Wang Tao (1): Revert "uacce: fix for cdev memory leak" Wenkai Lin (1): uacce: fix cdev handling in the cleanup path drivers/misc/uacce/uacce.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) -- 2.34.1

2 3

[PATCH OLK-6.6] sched/fair: Prefer steal within LLC and fallback to upper LLC newidle_balance on failure
by Chen Jinghuang 28 Feb '26

28 Feb '26

hulk inclusion category: perf bugzilla: https://atomgit.com/openeuler/kernel/issues/8555 -------------------------------- Implement fast load balancing with steal-task mechanism within the LLC domain; if task pulling fails, fall back to the native newidle_balance mechanism on LLC and upper scheduling domains. This optimization prioritizes task scheduling within the LLC domain to improve cache hit rate, while mitigating system load imbalance issues. Signed-off-by: Chen Jinghuang <chenjinghuang2(a)huawei.com> --- kernel/sched/fair.c | 21 +++++++++++++++++++-- kernel/sched/features.h | 1 + 2 files changed, 20 insertions(+), 2 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index e60f19cb0fee..a4592d35da8b 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -5236,6 +5236,11 @@ static inline bool steal_enabled(void) return sched_feat(STEAL) && allow; } +static inline bool steal_fail_ni_enabled(void) +{ + return sched_feat(STEAL_FAIL_NI); +} + static void overload_clear(struct rq *rq) { struct sparsemask *overload_cpus; @@ -5275,6 +5280,7 @@ static int try_steal(struct rq *this_rq, struct rq_flags *rf); static inline int try_steal(struct rq *this_rq, struct rq_flags *rf) { return 0; } static inline void overload_clear(struct rq *rq) {} static inline void overload_set(struct rq *rq) {} +static inline bool steal_fail_ni_enabled(void) { return false; } #endif #else /* CONFIG_SMP */ @@ -10343,9 +10349,15 @@ done: __maybe_unused; */ rq_idle_stamp_update(rq); - new_tasks = newidle_balance(rq, rf); - if (new_tasks == 0) + if (steal_fail_ni_enabled()) { new_tasks = try_steal(rq, rf); + if (new_tasks == 0) + new_tasks = newidle_balance(rq, rf); + } else { + new_tasks = newidle_balance(rq, rf); + if (new_tasks == 0) + new_tasks = try_steal(rq, rf); + } schedstat_end_time(rq, time); if (new_tasks) @@ -14313,6 +14325,11 @@ static int newidle_balance(struct rq *this_rq, struct rq_flags *rf) int continue_balancing = 1; u64 domain_cost; + if (steal_fail_ni_enabled()) { + if (sd->flags & SD_SHARE_PKG_RESOURCES) + continue; + } + update_next_balance(sd, &next_balance); if (this_rq->avg_idle < curr_cost + sd->max_newidle_lb_cost) diff --git a/kernel/sched/features.h b/kernel/sched/features.h index 1f665fbf0137..52ea0097c513 100644 --- a/kernel/sched/features.h +++ b/kernel/sched/features.h @@ -63,6 +63,7 @@ SCHED_FEAT(SIS_UTIL, true) * Improves CPU utilization. */ SCHED_FEAT(STEAL, false) +SCHED_FEAT(STEAL_FAIL_NI, false) #endif #ifdef CONFIG_SCHED_PARAL -- 2.33.0

2 1

[PATCH v8 0/5] arm64/riscv: Add support for crashkernel CMA reservation
by Jinjie Ruan 28 Feb '26

28 Feb '26

The crash memory allocation, and the exclude of crashk_res, crashk_low_res and crashk_cma memory are almost identical across different architectures, This patch set handle them in crash core in a general way, which eliminate a lot of duplication code. And add support for crashkernel CMA reservation for arm64 and riscv. Rebased on v7.0-rc1. Basic second kernel boot test were performed on QEMU platforms for x86, ARM64, and RISC-V architectures with the following parameters: "cma=256M crashkernel=256M crashkernel=64M,cma" Changes in v8: - Fix the build issues reported by kernel test robot and Sourabh. - Link to v7: https://lore.kernel.org/all/20260226130437.1867658-1-ruanjinjie@huawei.com/ Changes in v7: - Correct the inclusion of CMA-reserved ranges for kdump kernel in of/kexec for arm64 and riscv. - Add Acked-by. - Link to v6: https://lore.kernel.org/all/20260224085342.387996-1-ruanjinjie@huawei.com/ Changes in v6: - Update the crash core exclude code as Mike suggested. - Rebased on v7.0-rc1. - Add acked-by. - Link to v5: https://lore.kernel.org/all/20260212101001.343158-1-ruanjinjie@huawei.com/ Changes in v5: - Fix the kernel test robot build warnings. - Sort crash memory ranges before preparing elfcorehdr for powerpc - Link to v4: https://lore.kernel.org/all/20260209095931.2813152-1-ruanjinjie@huawei.com/ Changes in v4: - Move the size calculation (and the realloc if needed) into the generic crash. - Link to v3: https://lore.kernel.org/all/20260204093728.1447527-1-ruanjinjie@huawei.com/ Jinjie Ruan (4): crash: Exclude crash kernel memory in crash core crash: Use crash_exclude_core_ranges() on powerpc arm64: kexec: Add support for crashkernel CMA reservation riscv: kexec: Add support for crashkernel CMA reservation Sourabh Jain (1): powerpc/crash: sort crash memory ranges before preparing elfcorehdr .../admin-guide/kernel-parameters.txt | 16 +-- arch/arm64/kernel/machine_kexec_file.c | 39 +++---- arch/arm64/mm/init.c | 5 +- arch/loongarch/kernel/machine_kexec_file.c | 39 +++---- arch/powerpc/include/asm/kexec_ranges.h | 1 - arch/powerpc/kexec/crash.c | 5 +- arch/powerpc/kexec/ranges.c | 101 +----------------- arch/riscv/kernel/machine_kexec_file.c | 38 +++---- arch/riscv/mm/init.c | 5 +- arch/x86/kernel/crash.c | 89 +++------------ drivers/of/fdt.c | 9 +- drivers/of/kexec.c | 9 ++ include/linux/crash_core.h | 9 ++ kernel/crash_core.c | 89 ++++++++++++++- 14 files changed, 178 insertions(+), 276 deletions(-) -- 2.34.1

1 5

[PATCH OLK-6.6] sched/fair: Prefer steal within LLC and fallback to newidle_balance
by Chen Jinghuang 28 Feb '26

28 Feb '26

hulk inclusion category: perf bugzilla: https://atomgit.com/openeuler/kernel/issues/8555 -------------------------------- Implement fast load balancing with steal-task mechanism within the LLC domain; if task pulling fails, fall back to the native newidle_balance mechanism on LLC and upper scheduling domains. This optimization prioritizes task scheduling within the LLC domain to improve cache hit rate, while mitigating system load imbalance issues. Signed-off-by: Chen Jinghuang <chenjinghuang2(a)huawei.com> --- kernel/sched/fair.c | 15 +++++++++++++-- kernel/sched/features.h | 1 + 2 files changed, 14 insertions(+), 2 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index e60f19cb0fee..fd37f7648c86 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -10343,9 +10343,15 @@ done: __maybe_unused; */ rq_idle_stamp_update(rq); - new_tasks = newidle_balance(rq, rf); - if (new_tasks == 0) + if (sched_feat(STEAL_FAIL_NI)) { new_tasks = try_steal(rq, rf); + if (new_tasks == 0) + new_tasks = newidle_balance(rq, rf); + } else { + new_tasks = newidle_balance(rq, rf); + if (new_tasks == 0) + new_tasks = try_steal(rq, rf); + } schedstat_end_time(rq, time); if (new_tasks) @@ -14313,6 +14319,11 @@ static int newidle_balance(struct rq *this_rq, struct rq_flags *rf) int continue_balancing = 1; u64 domain_cost; + if (sched_feat(STEAL_FAIL_NI)) { + if (sd->flags & SD_SHARE_PKG_RESOURCES) + continue; + } + update_next_balance(sd, &next_balance); if (this_rq->avg_idle < curr_cost + sd->max_newidle_lb_cost) diff --git a/kernel/sched/features.h b/kernel/sched/features.h index 1f665fbf0137..52ea0097c513 100644 --- a/kernel/sched/features.h +++ b/kernel/sched/features.h @@ -63,6 +63,7 @@ SCHED_FEAT(SIS_UTIL, true) * Improves CPU utilization. */ SCHED_FEAT(STEAL, false) +SCHED_FEAT(STEAL_FAIL_NI, false) #endif #ifdef CONFIG_SCHED_PARAL -- 2.34.1

2 1

[PATCH OLK-6.6] sched/fair: Prefer steal within LLC and fallback to newidle_balance
by Chen Jinghuang 28 Feb '26

28 Feb '26

Offering: HULK hulk inclusion category: perf bugzilla: https://atomgit.com/openeuler/kernel/issues/8555 -------------------------------- Implement fast load balancing with steal-task mechanism within the LLC domain; if task pulling fails, fall back to the native newidle_balance mechanism on LLC and upper scheduling domains. This optimization prioritizes task scheduling within the LLC domain to improve cache hit rate, while mitigating system load imbalance issues. Signed-off-by: Chen Jinghuang <chenjinghuang2(a)huawei.com> --- kernel/sched/fair.c | 15 +++++++++++++-- kernel/sched/features.h | 1 + 2 files changed, 14 insertions(+), 2 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index e60f19cb0fee..fd37f7648c86 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -10343,9 +10343,15 @@ done: __maybe_unused; */ rq_idle_stamp_update(rq); - new_tasks = newidle_balance(rq, rf); - if (new_tasks == 0) + if (sched_feat(STEAL_FAIL_NI)) { new_tasks = try_steal(rq, rf); + if (new_tasks == 0) + new_tasks = newidle_balance(rq, rf); + } else { + new_tasks = newidle_balance(rq, rf); + if (new_tasks == 0) + new_tasks = try_steal(rq, rf); + } schedstat_end_time(rq, time); if (new_tasks) @@ -14313,6 +14319,11 @@ static int newidle_balance(struct rq *this_rq, struct rq_flags *rf) int continue_balancing = 1; u64 domain_cost; + if (sched_feat(STEAL_FAIL_NI)) { + if (sd->flags & SD_SHARE_PKG_RESOURCES) + continue; + } + update_next_balance(sd, &next_balance); if (this_rq->avg_idle < curr_cost + sd->max_newidle_lb_cost) diff --git a/kernel/sched/features.h b/kernel/sched/features.h index 1f665fbf0137..52ea0097c513 100644 --- a/kernel/sched/features.h +++ b/kernel/sched/features.h @@ -63,6 +63,7 @@ SCHED_FEAT(SIS_UTIL, true) * Improves CPU utilization. */ SCHED_FEAT(STEAL, false) +SCHED_FEAT(STEAL_FAIL_NI, false) #endif #ifdef CONFIG_SCHED_PARAL -- 2.34.1

2 1

[PATCH OLK-6.6] [Huawei] sched/fair: Prefer steal within LLC and fallback to newidle_balance
by Chen Jinghuang 28 Feb '26

28 Feb '26

Offering: HULK hulk inclusion category: perf bugzilla: https://atomgit.com/openeuler/kernel/issues/8555 -------------------------------- Implement fast load balancing with steal-task mechanism within the LLC domain; if task pulling fails, fall back to the native newidle_balance mechanism on LLC and upper scheduling domains. This optimization prioritizes task scheduling within the LLC domain to improve cache hit rate, while mitigating system load imbalance issues. Signed-off-by: Chen Jinghuang <chenjinghuang2(a)huawei.com> --- kernel/sched/fair.c | 15 +++++++++++++-- kernel/sched/features.h | 1 + 2 files changed, 14 insertions(+), 2 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index e60f19cb0fee..fd37f7648c86 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -10343,9 +10343,15 @@ done: __maybe_unused; */ rq_idle_stamp_update(rq); - new_tasks = newidle_balance(rq, rf); - if (new_tasks == 0) + if (sched_feat(STEAL_FAIL_NI)) { new_tasks = try_steal(rq, rf); + if (new_tasks == 0) + new_tasks = newidle_balance(rq, rf); + } else { + new_tasks = newidle_balance(rq, rf); + if (new_tasks == 0) + new_tasks = try_steal(rq, rf); + } schedstat_end_time(rq, time); if (new_tasks) @@ -14313,6 +14319,11 @@ static int newidle_balance(struct rq *this_rq, struct rq_flags *rf) int continue_balancing = 1; u64 domain_cost; + if (sched_feat(STEAL_FAIL_NI)) { + if (sd->flags & SD_SHARE_PKG_RESOURCES) + continue; + } + update_next_balance(sd, &next_balance); if (this_rq->avg_idle < curr_cost + sd->max_newidle_lb_cost) diff --git a/kernel/sched/features.h b/kernel/sched/features.h index 1f665fbf0137..52ea0097c513 100644 --- a/kernel/sched/features.h +++ b/kernel/sched/features.h @@ -63,6 +63,7 @@ SCHED_FEAT(SIS_UTIL, true) * Improves CPU utilization. */ SCHED_FEAT(STEAL, false) +SCHED_FEAT(STEAL_FAIL_NI, false) #endif #ifdef CONFIG_SCHED_PARAL -- 2.34.1

2 1

[PATCH OLK-6.6 0/2] CVE-2026-23163
by Hongtao Zhang 28 Feb '26

28 Feb '26

*** BLURB HERE *** Jon Doron (1): [Backport] drm/amdgpu: fix NULL pointer dereference in amdgpu_gmc_filter_faults_remove Philip Yang (1): [Backport] drm/amdkfd: Don't use sw fault filter if retry cam enabled drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) -- 2.43.0

2 3