- Kernel - mailweb.openeuler.org

[PATCH OLK-6.6] [Backport] remoteproc: Prevent crash handling to race with rproc_del()
by Chen Jinghuang 15 Sep '26

15 Sep '26

From: Bjorn Andersson <bjorn.andersson(a)oss.qualcomm.com> mainline inclusion from mainline-v7.3-rc1 commit 74ee3b2f5767447c57959994341e5b95f1079977 category: bugfix bugzilla: https://atomgit.com/openeuler/kernel/issues/9905 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- There's no synchronization between rproc_crash_handler_work() and rproc_del(), as such it's possible for a driver to be removed while crash-handler work is scheduled, or even executing - resulting in use-after-free issues. To avoid this the scheduled work need to be cancelled and synchronized against before the removal proceeds. In order to ensure that this doesn't race with the reporting, and thereby scheduling new work, a "deleting" flag is introduced. This is similar to the RPROC_DELETE state that was introduced to ensure that "start" didn't race with rproc_del(), but the existing mechanism can not be used as it's valid to call rproc_report_crash() in atomic context - and the "state" is protected by a mutex. In the event that work is cancelled the pm_stay_awake() is left unbalanced and need to be unrolled. The blocking and cancelling of crash-handler work prior to the actual rproc_shutdown() call does have the explicit side-effect that crashes resulting from the shutdown process will not enter the crash-handling path, and as such will not generate devcoredumps etc. Due to the existing mutual exclusion between these code paths there's no concrete reduction in functionality, but further work would be needed to handle this case. Assisted-by: OpenCode:GPT-5.5 Fixes: 8afd519c3470 ("remoteproc: add rproc_report_crash function to notify rproc crashes") Signed-off-by: Bjorn Andersson <bjorn.andersson(a)oss.qualcomm.com> Reviewed-by: Pradnya Dahiwale <pradnya.dahiwale(a)oss.qualcomm.com> Link: https://lore.kernel.org/r/20260723-rproc-rmmod-not-crashing-v1-2-546dfd5de0… Signed-off-by: Bjorn Andersson <andersson(a)kernel.org> Conflicts: drivers/remoteproc/remoteproc_core.c include/linux/remoteproc.h [The downstream 6.6 tree has no struct rproc::attach_work field nor rproc_attach_work(); the conflict was resolved by keeping only INIT_WORK() for crash_handler and adding spin_lock_init() for the new crash_handler_lock, dropping the upstream attach_work init line.] Signed-off-by: Chen Jinghuang <chenjinghuang2(a)huawei.com> --- drivers/remoteproc/remoteproc_core.c | 40 +++++++++++++++++++++------ drivers/remoteproc/remoteproc_sysfs.c | 1 - include/linux/remoteproc.h | 13 +++++---- 3 files changed, 40 insertions(+), 14 deletions(-) diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c index 2d4ae3b5af86..1c489b757419 100644 --- a/drivers/remoteproc/remoteproc_core.c +++ b/drivers/remoteproc/remoteproc_core.c @@ -1837,6 +1837,11 @@ int rproc_trigger_recovery(struct rproc *rproc) if (ret) return ret; + if (READ_ONCE(rproc->deleting)) { + ret = -ENODEV; + goto unlock_mutex; + } + /* State could have changed before we got the mutex */ if (rproc->state != RPROC_CRASHED) goto unlock_mutex; @@ -1869,6 +1874,11 @@ static void rproc_crash_handler_work(struct work_struct *work) mutex_lock(&rproc->lock); + if (READ_ONCE(rproc->deleting)) { + mutex_unlock(&rproc->lock); + goto out; + } + if (rproc->state == RPROC_CRASHED) { /* handle only the first crash detected */ mutex_unlock(&rproc->lock); @@ -1924,9 +1934,9 @@ int rproc_boot(struct rproc *rproc) return ret; } - if (rproc->state == RPROC_DELETED) { + if (READ_ONCE(rproc->deleting)) { ret = -ENODEV; - dev_err(dev, "can't boot deleted rproc %s\n", rproc->name); + dev_err(dev, "can't boot deleting rproc %s\n", rproc->name); goto unlock_mutex; } @@ -2497,6 +2507,7 @@ struct rproc *rproc_alloc(struct device *dev, const char *name, INIT_LIST_HEAD(&rproc->dump_segments); INIT_WORK(&rproc->crash_handler, rproc_crash_handler_work); + spin_lock_init(&rproc->crash_handler_lock); rproc->state = RPROC_OFFLINE; @@ -2556,16 +2567,21 @@ EXPORT_SYMBOL(rproc_put); */ int rproc_del(struct rproc *rproc) { + unsigned long flags; + if (!rproc) return -EINVAL; + spin_lock_irqsave(&rproc->crash_handler_lock, flags); + WRITE_ONCE(rproc->deleting, true); + spin_unlock_irqrestore(&rproc->crash_handler_lock, flags); + + if (cancel_work_sync(&rproc->crash_handler)) + pm_relax(rproc->dev.parent); + /* TODO: make sure this works with rproc->power > 1 */ rproc_shutdown(rproc); - mutex_lock(&rproc->lock); - rproc->state = RPROC_DELETED; - mutex_unlock(&rproc->lock); - rproc_delete_debug_dir(rproc); /* the rproc is downref'ed as soon as it's removed from the klist */ @@ -2677,18 +2693,26 @@ EXPORT_SYMBOL(rproc_get_by_child); */ void rproc_report_crash(struct rproc *rproc, enum rproc_crash_type type) { + unsigned long flags; + if (!rproc) { pr_err("NULL rproc pointer\n"); return; } + spin_lock_irqsave(&rproc->crash_handler_lock, flags); + if (READ_ONCE(rproc->deleting)) { + spin_unlock_irqrestore(&rproc->crash_handler_lock, flags); + return; + } + /* Prevent suspend while the remoteproc is being recovered */ pm_stay_awake(rproc->dev.parent); + queue_work(rproc_recovery_wq, &rproc->crash_handler); + spin_unlock_irqrestore(&rproc->crash_handler_lock, flags); dev_err(&rproc->dev, "crash detected in %s: type %s\n", rproc->name, rproc_crash_to_string(type)); - - queue_work(rproc_recovery_wq, &rproc->crash_handler); } EXPORT_SYMBOL(rproc_report_crash); diff --git a/drivers/remoteproc/remoteproc_sysfs.c b/drivers/remoteproc/remoteproc_sysfs.c index 8c7ea8922638..23de0490ba16 100644 --- a/drivers/remoteproc/remoteproc_sysfs.c +++ b/drivers/remoteproc/remoteproc_sysfs.c @@ -168,7 +168,6 @@ static const char * const rproc_state_string[] = { [RPROC_SUSPENDED] = "suspended", [RPROC_RUNNING] = "running", [RPROC_CRASHED] = "crashed", - [RPROC_DELETED] = "deleted", [RPROC_ATTACHED] = "attached", [RPROC_DETACHED] = "detached", [RPROC_LAST] = "invalid", diff --git a/include/linux/remoteproc.h b/include/linux/remoteproc.h index b4795698d8c2..e1634b9c7992 100644 --- a/include/linux/remoteproc.h +++ b/include/linux/remoteproc.h @@ -37,6 +37,7 @@ #include <linux/types.h> #include <linux/mutex.h> +#include <linux/spinlock.h> #include <linux/virtio.h> #include <linux/cdev.h> #include <linux/completion.h> @@ -412,7 +413,6 @@ struct rproc_ops { * a message. * @RPROC_RUNNING: device is up and running * @RPROC_CRASHED: device has crashed; need to start recovery - * @RPROC_DELETED: device is deleted * @RPROC_ATTACHED: device has been booted by another entity and the core * has attached to it * @RPROC_DETACHED: device has been booted by another entity and waiting @@ -430,10 +430,9 @@ enum rproc_state { RPROC_SUSPENDED = 1, RPROC_RUNNING = 2, RPROC_CRASHED = 3, - RPROC_DELETED = 4, - RPROC_ATTACHED = 5, - RPROC_DETACHED = 6, - RPROC_LAST = 7, + RPROC_ATTACHED = 4, + RPROC_DETACHED = 5, + RPROC_LAST = 6, }; /** @@ -527,6 +526,8 @@ enum rproc_features { * @notifyids: idr for dynamically assigning rproc-wide unique notify ids * @index: index of this rproc device * @crash_handler: workqueue for handling a crash + * @crash_handler_lock: serializes crash handler queueing and deletion + * @deleting: remoteproc deletion has begun * @crash_cnt: crash counter * @recovery_disabled: flag that state if recovery was disabled * @max_notifyid: largest allocated notify id. @@ -569,6 +570,8 @@ struct rproc { struct idr notifyids; int index; struct work_struct crash_handler; + spinlock_t crash_handler_lock; + bool deleting; unsigned int crash_cnt; bool recovery_disabled; int max_notifyid; -- 2.34.1

1 0

[PATCH OLK-6.6] kallsyms: prevent module removal when printing module name and buildid
by Tengda Wu 15 Sep '26

15 Sep '26

From: Petr Mladek <pmladek(a)suse.com> mainline inclusion from mainline-v7.0-rc1 commit 3b07086444f80c844351255fd94c2cb0a7224df2 category: bugfix bugzilla: https://atomgit.com/openeuler/kernel/issues/9956 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- kallsyms_lookup_buildid() copies the symbol name into the given buffer so that it can be safely read anytime later. But it just copies pointers to mod->name and mod->build_id which might get reused after the related struct module gets removed. The lifetime of struct module is synchronized using RCU. Take the rcu read lock for the entire __sprint_symbol(). Link: https://lkml.kernel.org/r/20251128135920.217303-8-pmladek@suse.com Signed-off-by: Petr Mladek <pmladek(a)suse.com> Reviewed-by: Aaron Tomlin <atomlin(a)atomlin.com> Cc: Alexei Starovoitov <ast(a)kernel.org> Cc: Daniel Borkman <daniel(a)iogearbox.net> Cc: Daniel Gomez <da.gomez(a)samsung.com> Cc: John Fastabend <john.fastabend(a)gmail.com> Cc: Kees Cook <kees(a)kernel.org> Cc: Luis Chamberalin <mcgrof(a)kernel.org> Cc: Marc Rutland <mark.rutland(a)arm.com> Cc: "Masami Hiramatsu (Google)" <mhiramat(a)kernel.org> Cc: Petr Pavlu <petr.pavlu(a)suse.com> Cc: Sami Tolvanen <samitolvanen(a)google.com> Cc: Steven Rostedt (Google) <rostedt(a)goodmis.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> Conflicts: kernel/kallsyms.c [Context conflict] Signed-off-by: Tengda Wu <wutengda2(a)huawei.com> --- kernel/kallsyms.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/kernel/kallsyms.c b/kernel/kallsyms.c index 78b5bda6f531..0919c4dd9d8b 100644 --- a/kernel/kallsyms.c +++ b/kernel/kallsyms.c @@ -485,6 +485,9 @@ static int __sprint_symbol(char *buffer, unsigned long address, unsigned long offset, size; int len; + /* Prevent module removal until modname and modbuildid are printed */ + guard(rcu)(); + address += symbol_offset; name = kallsyms_lookup_buildid(address, &size, &offset, &modname, &buildid, buffer); -- 2.34.1

2 1

[PATCH OLK-6.6] KVM: arm64: vgic-v3: Restrict ICH_AP1Rn_EL2 accesses to 64-bit only with vNMI
by Jinqian Yang 15 Sep '26

15 Sep '26

virt inclusion category: bugfix bugzilla: https://atomgit.com/openeuler/kernel/issues/9836 -------------------------------------------------------------------- ICH_AP1Rn_EL2 is 64-bit only when vNMI is in use, with the NMI priority in ICH_AP1R0_EL2[63:32]; otherwise the upper bits are RES0. Gate the save/restore access width on vgic_v3_vcpu_has_vnmi() which checks both vgic.has_nmi and ID_AA64PFR1_EL1.NMI, so that the NMI priority is preserved across context switch only when vNMI is actually enabled for the guest. Also drop the early assignment of kvm->arch.pfr1_nmi in kvm_arch_init_vm(). Without it vNMI is no longer enabled by default at VM creation on hosts with FEAT_NMI; instead the VMM must explicitly set ID_AA64PFR1_EL1.NMI via SET_ONE_REG to enable vNMI for the guest. Fixes: d46e52452afc ("KVM: arm64: vgic-v3: Upgrade AP1Rn to 64bit.") Signed-off-by: Jinqian Yang <yangjinqian1(a)huawei.com> --- arch/arm64/include/asm/kvm_hyp.h | 1 + arch/arm64/kvm/arm.c | 3 -- arch/arm64/kvm/hyp/vgic-v3-sr.c | 55 +++++++++++++++++++++++++++----- 3 files changed, 48 insertions(+), 11 deletions(-) diff --git a/arch/arm64/include/asm/kvm_hyp.h b/arch/arm64/include/asm/kvm_hyp.h index 127ad280e68b..c4121bf0aadc 100644 --- a/arch/arm64/include/asm/kvm_hyp.h +++ b/arch/arm64/include/asm/kvm_hyp.h @@ -82,6 +82,7 @@ void __vgic_v3_activate_traps(struct vgic_v3_cpu_if *cpu_if); void __vgic_v3_deactivate_traps(struct vgic_v3_cpu_if *cpu_if); void __vgic_v3_save_aprs(struct vgic_v3_cpu_if *cpu_if); void __vgic_v3_restore_aprs(struct vgic_v3_cpu_if *cpu_if); +bool vgic_v3_vcpu_has_vnmi(struct kvm_vcpu *vcpu); int __vgic_v3_perform_cpuif_access(struct kvm_vcpu *vcpu); #ifdef __KVM_NVHE_HYPERVISOR__ diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index 2b560eba3e0b..9a4be08730da 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -381,9 +381,6 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) /* The maximum number of VCPUs is limited by the host's GIC model */ kvm->max_vcpus = kvm_arm_default_max_vcpus(); - if (cpus_have_const_cap(ARM64_HAS_NMI) && !static_branch_unlikely(&vgic_v3_cpuif_trap)) - kvm->arch.pfr1_nmi = ID_AA64PFR1_EL1_NMI_IMP; - kvm_arm_init_hypercalls(kvm); if (!kvm_is_realm(kvm)) diff --git a/arch/arm64/kvm/hyp/vgic-v3-sr.c b/arch/arm64/kvm/hyp/vgic-v3-sr.c index 7a9e132f68db..ab406332762e 100644 --- a/arch/arm64/kvm/hyp/vgic-v3-sr.c +++ b/arch/arm64/kvm/hyp/vgic-v3-sr.c @@ -335,10 +335,40 @@ void __vgic_v3_deactivate_traps(struct vgic_v3_cpu_if *cpu_if) write_gicreg(0, ICH_HCR_EL2); } +static struct kvm_vcpu *vgic_v3_cpu_if_to_vcpu(struct vgic_v3_cpu_if *cpu_if) +{ + struct vgic_cpu *vgic_cpu = container_of(cpu_if, struct vgic_cpu, vgic_v3); + + return container_of(vgic_cpu, struct kvm_vcpu, arch.vgic_cpu); +} + +/* + * vNMI is only usable if the guest has selected NMI support + * (vgic.has_nmi) and the vcpu has FEAT_NMI (pfr1_nmi). + * + * This drives the width of the ICH_AP1Rn_EL2 accesses: with vNMI the + * registers are 64bit (the NMI priority lives in ICH_AP1R0_EL2[63]), + * without it only the low 32 bits are valid. + */ +bool vgic_v3_vcpu_has_vnmi(struct kvm_vcpu *vcpu) +{ + struct kvm *kvm = kern_hyp_va(vcpu->kvm); + + return kvm->arch.vgic.has_nmi && + kvm->arch.pfr1_nmi == ID_AA64PFR1_EL1_NMI_IMP; +} + void __vgic_v3_save_aprs(struct vgic_v3_cpu_if *cpu_if) { u64 val; u32 nr_pre_bits; + /* + * ICH_AP1Rn_EL2 are only 64bit when the vcpu has vNMI: the NMI + * priority is encoded in ICH_AP1R0_EL2[63]. Without vNMI, the + * upper bits are RES0 and only the low 32 bits must be + * preserved. + */ + bool vnmi = vgic_v3_vcpu_has_vnmi(vgic_v3_cpu_if_to_vcpu(cpu_if)); val = read_gicreg(ICH_VTR_EL2); nr_pre_bits = vtr_to_nr_pre_bits(val); @@ -357,14 +387,18 @@ void __vgic_v3_save_aprs(struct vgic_v3_cpu_if *cpu_if) switch (nr_pre_bits) { case 7: - cpu_if->vgic_ap1r[3] = __vgic_v3_read_ap1rn(3); - cpu_if->vgic_ap1r[2] = __vgic_v3_read_ap1rn(2); + cpu_if->vgic_ap1r[3] = vnmi ? __vgic_v3_read_ap1rn(3) + : (u32)__vgic_v3_read_ap1rn(3); + cpu_if->vgic_ap1r[2] = vnmi ? __vgic_v3_read_ap1rn(2) + : (u32)__vgic_v3_read_ap1rn(2); fallthrough; case 6: - cpu_if->vgic_ap1r[1] = __vgic_v3_read_ap1rn(1); + cpu_if->vgic_ap1r[1] = vnmi ? __vgic_v3_read_ap1rn(1) + : (u32)__vgic_v3_read_ap1rn(1); fallthrough; default: - cpu_if->vgic_ap1r[0] = __vgic_v3_read_ap1rn(0); + cpu_if->vgic_ap1r[0] = vnmi ? __vgic_v3_read_ap1rn(0) + : (u32)__vgic_v3_read_ap1rn(0); } } @@ -372,6 +406,7 @@ void __vgic_v3_restore_aprs(struct vgic_v3_cpu_if *cpu_if) { u64 val; u32 nr_pre_bits; + bool vnmi = vgic_v3_vcpu_has_vnmi(vgic_v3_cpu_if_to_vcpu(cpu_if)); val = read_gicreg(ICH_VTR_EL2); nr_pre_bits = vtr_to_nr_pre_bits(val); @@ -390,14 +425,18 @@ void __vgic_v3_restore_aprs(struct vgic_v3_cpu_if *cpu_if) switch (nr_pre_bits) { case 7: - __vgic_v3_write_ap1rn(cpu_if->vgic_ap1r[3], 3); - __vgic_v3_write_ap1rn(cpu_if->vgic_ap1r[2], 2); + __vgic_v3_write_ap1rn(vnmi ? cpu_if->vgic_ap1r[3] + : (u32)cpu_if->vgic_ap1r[3], 3); + __vgic_v3_write_ap1rn(vnmi ? cpu_if->vgic_ap1r[2] + : (u32)cpu_if->vgic_ap1r[2], 2); fallthrough; case 6: - __vgic_v3_write_ap1rn(cpu_if->vgic_ap1r[1], 1); + __vgic_v3_write_ap1rn(vnmi ? cpu_if->vgic_ap1r[1] + : (u32)cpu_if->vgic_ap1r[1], 1); fallthrough; default: - __vgic_v3_write_ap1rn(cpu_if->vgic_ap1r[0], 0); + __vgic_v3_write_ap1rn(vnmi ? cpu_if->vgic_ap1r[0] + : (u32)cpu_if->vgic_ap1r[0], 0); } } -- 2.33.0

2 1

[PATCH OLK-6.6] KVM: arm64: vgic-v3: Restrict ICH_AP1Rn_EL2 accesses to 64-bit only with vNMI
by Jinqian Yang 15 Sep '26

15 Sep '26

virt inclusion category: bugfix bugzilla: https://atomgit.com/openeuler/kernel/issues/9836 -------------------------------------------------------------------- ICH_AP1Rn_EL2 is 64-bit only when vNMI is in use, with the NMI priority in ICH_AP1R0_EL2[63:32]; otherwise the upper bits are RES0. Gate the save/restore access width on vgic_v3_vcpu_has_vnmi() which checks both vgic.has_nmi and ID_AA64PFR1_EL1.NMI, so that the NMI priority is preserved across context switch only when vNMI is actually enabled for the guest. Also drop the early assignment of kvm->arch.pfr1_nmi in kvm_arch_init_vm(). Without it vNMI is no longer enabled by default at VM creation on hosts with FEAT_NMI; instead the VMM must explicitly set ID_AA64PFR1_EL1.NMI via SET_ONE_REG to enable vNMI for the guest. Signed-off-by: Jinqian Yang <yangjinqian1(a)huawei.com> --- arch/arm64/include/asm/kvm_hyp.h | 1 + arch/arm64/kvm/arm.c | 3 -- arch/arm64/kvm/hyp/vgic-v3-sr.c | 55 +++++++++++++++++++++++++++----- 3 files changed, 48 insertions(+), 11 deletions(-) diff --git a/arch/arm64/include/asm/kvm_hyp.h b/arch/arm64/include/asm/kvm_hyp.h index 127ad280e68b..c4121bf0aadc 100644 --- a/arch/arm64/include/asm/kvm_hyp.h +++ b/arch/arm64/include/asm/kvm_hyp.h @@ -82,6 +82,7 @@ void __vgic_v3_activate_traps(struct vgic_v3_cpu_if *cpu_if); void __vgic_v3_deactivate_traps(struct vgic_v3_cpu_if *cpu_if); void __vgic_v3_save_aprs(struct vgic_v3_cpu_if *cpu_if); void __vgic_v3_restore_aprs(struct vgic_v3_cpu_if *cpu_if); +bool vgic_v3_vcpu_has_vnmi(struct kvm_vcpu *vcpu); int __vgic_v3_perform_cpuif_access(struct kvm_vcpu *vcpu); #ifdef __KVM_NVHE_HYPERVISOR__ diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index 2b560eba3e0b..9a4be08730da 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -381,9 +381,6 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) /* The maximum number of VCPUs is limited by the host's GIC model */ kvm->max_vcpus = kvm_arm_default_max_vcpus(); - if (cpus_have_const_cap(ARM64_HAS_NMI) && !static_branch_unlikely(&vgic_v3_cpuif_trap)) - kvm->arch.pfr1_nmi = ID_AA64PFR1_EL1_NMI_IMP; - kvm_arm_init_hypercalls(kvm); if (!kvm_is_realm(kvm)) diff --git a/arch/arm64/kvm/hyp/vgic-v3-sr.c b/arch/arm64/kvm/hyp/vgic-v3-sr.c index 7a9e132f68db..ab406332762e 100644 --- a/arch/arm64/kvm/hyp/vgic-v3-sr.c +++ b/arch/arm64/kvm/hyp/vgic-v3-sr.c @@ -335,10 +335,40 @@ void __vgic_v3_deactivate_traps(struct vgic_v3_cpu_if *cpu_if) write_gicreg(0, ICH_HCR_EL2); } +static struct kvm_vcpu *vgic_v3_cpu_if_to_vcpu(struct vgic_v3_cpu_if *cpu_if) +{ + struct vgic_cpu *vgic_cpu = container_of(cpu_if, struct vgic_cpu, vgic_v3); + + return container_of(vgic_cpu, struct kvm_vcpu, arch.vgic_cpu); +} + +/* + * vNMI is only usable if the guest has selected NMI support + * (vgic.has_nmi) and the vcpu has FEAT_NMI (pfr1_nmi). + * + * This drives the width of the ICH_AP1Rn_EL2 accesses: with vNMI the + * registers are 64bit (the NMI priority lives in ICH_AP1R0_EL2[63]), + * without it only the low 32 bits are valid. + */ +bool vgic_v3_vcpu_has_vnmi(struct kvm_vcpu *vcpu) +{ + struct kvm *kvm = kern_hyp_va(vcpu->kvm); + + return kvm->arch.vgic.has_nmi && + kvm->arch.pfr1_nmi == ID_AA64PFR1_EL1_NMI_IMP; +} + void __vgic_v3_save_aprs(struct vgic_v3_cpu_if *cpu_if) { u64 val; u32 nr_pre_bits; + /* + * ICH_AP1Rn_EL2 are only 64bit when the vcpu has vNMI: the NMI + * priority is encoded in ICH_AP1R0_EL2[63]. Without vNMI, the + * upper bits are RES0 and only the low 32 bits must be + * preserved. + */ + bool vnmi = vgic_v3_vcpu_has_vnmi(vgic_v3_cpu_if_to_vcpu(cpu_if)); val = read_gicreg(ICH_VTR_EL2); nr_pre_bits = vtr_to_nr_pre_bits(val); @@ -357,14 +387,18 @@ void __vgic_v3_save_aprs(struct vgic_v3_cpu_if *cpu_if) switch (nr_pre_bits) { case 7: - cpu_if->vgic_ap1r[3] = __vgic_v3_read_ap1rn(3); - cpu_if->vgic_ap1r[2] = __vgic_v3_read_ap1rn(2); + cpu_if->vgic_ap1r[3] = vnmi ? __vgic_v3_read_ap1rn(3) + : (u32)__vgic_v3_read_ap1rn(3); + cpu_if->vgic_ap1r[2] = vnmi ? __vgic_v3_read_ap1rn(2) + : (u32)__vgic_v3_read_ap1rn(2); fallthrough; case 6: - cpu_if->vgic_ap1r[1] = __vgic_v3_read_ap1rn(1); + cpu_if->vgic_ap1r[1] = vnmi ? __vgic_v3_read_ap1rn(1) + : (u32)__vgic_v3_read_ap1rn(1); fallthrough; default: - cpu_if->vgic_ap1r[0] = __vgic_v3_read_ap1rn(0); + cpu_if->vgic_ap1r[0] = vnmi ? __vgic_v3_read_ap1rn(0) + : (u32)__vgic_v3_read_ap1rn(0); } } @@ -372,6 +406,7 @@ void __vgic_v3_restore_aprs(struct vgic_v3_cpu_if *cpu_if) { u64 val; u32 nr_pre_bits; + bool vnmi = vgic_v3_vcpu_has_vnmi(vgic_v3_cpu_if_to_vcpu(cpu_if)); val = read_gicreg(ICH_VTR_EL2); nr_pre_bits = vtr_to_nr_pre_bits(val); @@ -390,14 +425,18 @@ void __vgic_v3_restore_aprs(struct vgic_v3_cpu_if *cpu_if) switch (nr_pre_bits) { case 7: - __vgic_v3_write_ap1rn(cpu_if->vgic_ap1r[3], 3); - __vgic_v3_write_ap1rn(cpu_if->vgic_ap1r[2], 2); + __vgic_v3_write_ap1rn(vnmi ? cpu_if->vgic_ap1r[3] + : (u32)cpu_if->vgic_ap1r[3], 3); + __vgic_v3_write_ap1rn(vnmi ? cpu_if->vgic_ap1r[2] + : (u32)cpu_if->vgic_ap1r[2], 2); fallthrough; case 6: - __vgic_v3_write_ap1rn(cpu_if->vgic_ap1r[1], 1); + __vgic_v3_write_ap1rn(vnmi ? cpu_if->vgic_ap1r[1] + : (u32)cpu_if->vgic_ap1r[1], 1); fallthrough; default: - __vgic_v3_write_ap1rn(cpu_if->vgic_ap1r[0], 0); + __vgic_v3_write_ap1rn(vnmi ? cpu_if->vgic_ap1r[0] + : (u32)cpu_if->vgic_ap1r[0], 0); } } -- 2.33.0

2 1

[PATCH OLK-5.10 0/2] CVE-2026-68396
by Zhou Minqiang 14 Sep '26

14 Sep '26

CVE-2026-68396 Zhou Minqiang (2): scsi: core: wake eh reliably when using scsi_schedule_eh scsi: core: Fix kabi breakage of struct Scsi_Host drivers/scsi/hosts.c | 13 +++++++++++++ drivers/scsi/scsi_error.c | 24 +++++++++++++++++++++++- drivers/scsi/scsi_priv.h | 1 + include/scsi/scsi_host.h | 7 ++++++- 4 files changed, 43 insertions(+), 2 deletions(-) -- 2.52.0

2 3

[PATCH OLK-5.10] scsi: core: pair EH runtime PM get and put
by Zhou Minqiang 14 Sep '26

14 Sep '26

mainline inclusion from mainline-v7.2 commit 872f486259ae0bc6b73ca4735a15d013241f73e9 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/18399 CVE: CVE-2026-74754 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- shost->eh_noresume is currently consulted twice in one error handling iteration: once before scsi_autopm_get_host() and once again before scsi_autopm_put_host(). That is racy when a PM-triggered error path flips shost->eh_noresume while the SCSI EH thread is still running. The problem flow looks like this: PM path ufshcd_set_dev_pwr_mode() shost->eh_noresume = 1 ufshcd_execute_start_stop <-- trigger EH ... shost->eh_noresume = 0 EH path scsi_error_handler() if (!shost->eh_noresume) scsi_autopm_get_host() <-- skipped ... if (!shost->eh_noresume) scsi_autopm_put_host() <-- executed later In that case one EH iteration can skip autoresume on entry and still drop a runtime PM reference on exit. That leaves an unmatched runtime PM put and can trigger a runtime PM usage count underflow. Fix this by making eh_noresume a regular bool so it can be accessed with READ_ONCE() and WRITE_ONCE(). Snapshot it once per EH iteration and use that snapshot for both runtime PM get and put decisions. Fixes: ae0751ffc77e ("[SCSI] add flag to skip the runtime PM calls on the host") Signed-off-by: Hongjie Fang <hongjiefang(a)asrmicro.com> Reviewed-by: Bart Van Assche <bvanassche(a)acm.org> Reviewed-by: Peter Wang <peter.wang(a)mediatek.com> Link: https://patch.msgid.link/20260729111614.2407559-1-hongjiefang@asrmicro.com Signed-off-by: Martin K. Petersen (Oracle) <mkp(a)kernel.org> Conflicts: drivers/scsi/scsi_error.c drivers/scsi/ufs/ufshcd.c [ctx conflicts] Signed-off-by: Zhou Minqiang <zhouminqiang2(a)huawei.com> --- drivers/scsi/scsi_error.c | 9 +++++++-- drivers/scsi/ufs/ufshcd.c | 4 ++++ 2 files changed, 11 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c index 18512b33a190..90f275ac43c6 100644 --- a/drivers/scsi/scsi_error.c +++ b/drivers/scsi/scsi_error.c @@ -2204,6 +2204,8 @@ static void scsi_unjam_host(struct Scsi_Host *shost) int scsi_error_handler(void *data) { struct Scsi_Host *shost = data; + bool eh_noresume; + unsigned long flags; /* * We use TASK_INTERRUPTIBLE so that the thread is not @@ -2245,7 +2247,10 @@ int scsi_error_handler(void *data) * what we need to do to get it up and online again (if we can). * If we fail, we end up taking the thing offline. */ - if (!shost->eh_noresume && scsi_autopm_get_host(shost) != 0) { + spin_lock_irqsave(shost->host_lock, flags); + eh_noresume = shost->eh_noresume; + spin_unlock_irqrestore(shost->host_lock, flags); + if (!eh_noresume && scsi_autopm_get_host(shost) != 0) { SCSI_LOG_ERROR_RECOVERY(1, shost_printk(KERN_ERR, shost, "scsi_eh_%d: unable to autoresume\n", @@ -2269,7 +2274,7 @@ int scsi_error_handler(void *data) * which are still online. */ scsi_restart_operations(shost); - if (!shost->eh_noresume) + if (!eh_noresume) scsi_autopm_put_host(shost); } __set_current_state(TASK_RUNNING); diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c index ae00892ba44a..b3c6f41a5284 100644 --- a/drivers/scsi/ufs/ufshcd.c +++ b/drivers/scsi/ufs/ufshcd.c @@ -8291,7 +8291,9 @@ static int ufshcd_set_dev_pwr_mode(struct ufs_hba *hba, * we are functional while we are here, skip host resume in error * handling context. */ + spin_lock_irqsave(hba->host->host_lock, flags); hba->host->eh_noresume = 1; + spin_unlock_irqrestore(hba->host->host_lock, flags); if (hba->wlun_dev_clr_ua) { ret = ufshcd_send_request_sense(hba, sdp); if (ret) @@ -8321,7 +8323,9 @@ static int ufshcd_set_dev_pwr_mode(struct ufs_hba *hba, hba->curr_dev_pwr_mode = pwr_mode; out: scsi_device_put(sdp); + spin_lock_irqsave(hba->host->host_lock, flags); hba->host->eh_noresume = 0; + spin_unlock_irqrestore(hba->host->host_lock, flags); return ret; } -- 2.52.0

2 1

[PATCH OLK-6.6 0/8] arm64: Add FEAT_TLBID support
by Zeng Heng 14 Sep '26

14 Sep '26

TLBI broadcasting all PEs introduces performance noise. By combining hardware and software, TLBID (TLBI Domain) limit TLBI to an appropriate scope, avoiding the performance overhead caused by broadcasting. Arm Architecture defines the FEAT_TLBID extension since Armv9.3, when present, allows the OSPM to target TLB invalidations to a specific domain. A TLBI domain can be contained in multiple parent domains. The TLBI table lists all the PE and SMMU components and the TLBI Domains that each component belongs to. The feature is detected at runtime, and will remain disabled if the system does not implement the feature. Jinjiang Tu (1): arm64: tlbflush: Optimize flush_tlb_mm() by using TLBID Liao Chang (1): acpi: arm64: Add basic skeleton for TLBI Domains Marc Zyngier (4): arm64: cpufeature: Add ID_AA64MMFR4_EL1 handling arm64: sysreg: Add layout for ID_AA64MMFR4_EL1 arm64: cpufeatures: Add missing ID_AA64MMFR4_EL1 to __read_sysreg_by_encoding() arm64: sysreg: Update ID_AA64MMFR4_EL1 description Zeng Heng (2): arm64: cpufeature: Add TLBID (Domain-based TLB Invalidation) detection arm64: mm: Track CPUs that a task has run on for TLBID optimization arch/arm64/Kconfig | 20 ++ arch/arm64/include/asm/cpu.h | 2 + arch/arm64/include/asm/cpufeature.h | 13 ++ arch/arm64/include/asm/tlbflush.h | 66 +++++- arch/arm64/include/asm/tlbidomain.h | 34 +++ arch/arm64/kernel/cpufeature.c | 20 ++ arch/arm64/kernel/cpuinfo.c | 4 + arch/arm64/mm/Makefile | 1 + arch/arm64/mm/context.c | 9 +- arch/arm64/mm/tlbidomain.c | 339 ++++++++++++++++++++++++++++ arch/arm64/tools/cpucaps | 2 +- arch/arm64/tools/sysreg | 64 ++++++ drivers/acpi/arm64/Makefile | 1 + drivers/acpi/arm64/init.c | 2 + drivers/acpi/arm64/init.h | 1 + drivers/acpi/arm64/tlbid.c | 107 +++++++++ drivers/acpi/tables.c | 9 + include/acpi/actbl2.h | 29 +++ include/linux/acpi.h | 1 + 19 files changed, 718 insertions(+), 6 deletions(-) create mode 100644 arch/arm64/include/asm/tlbidomain.h create mode 100644 arch/arm64/mm/tlbidomain.c create mode 100644 drivers/acpi/arm64/tlbid.c -- 2.43.0

2 9

[PATCH OLK-6.6 0/2] CVE-2026-68396
by Zhou Minqiang 14 Sep '26

14 Sep '26

CVE-2026-68396 Zhou Minqiang (2): scsi: core: wake eh reliably when using scsi_schedule_eh scsi: core: Fix kabi breakage of struct Scsi_Host drivers/scsi/hosts.c | 13 +++++++++++++ drivers/scsi/scsi_error.c | 24 +++++++++++++++++++++++- drivers/scsi/scsi_priv.h | 1 + include/scsi/scsi_host.h | 7 ++++++- 4 files changed, 43 insertions(+), 2 deletions(-) -- 2.52.0

2 3

[PATCH OLK-6.6] scsi: core: pair EH runtime PM get and put
by Zhou Minqiang 14 Sep '26

14 Sep '26

mainline inclusion from mainline-v7.2 commit 872f486259ae0bc6b73ca4735a15d013241f73e9 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/18399 CVE: CVE-2026-74754 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- [ Upstream commit 872f486259ae0bc6b73ca4735a15d013241f73e9 ] shost->eh_noresume is currently consulted twice in one error handling iteration: once before scsi_autopm_get_host() and once again before scsi_autopm_put_host(). That is racy when a PM-triggered error path flips shost->eh_noresume while the SCSI EH thread is still running. The problem flow looks like this: PM path ufshcd_set_dev_pwr_mode() shost->eh_noresume = 1 ufshcd_execute_start_stop <-- trigger EH ... shost->eh_noresume = 0 EH path scsi_error_handler() if (!shost->eh_noresume) scsi_autopm_get_host() <-- skipped ... if (!shost->eh_noresume) scsi_autopm_put_host() <-- executed later In that case one EH iteration can skip autoresume on entry and still drop a runtime PM reference on exit. That leaves an unmatched runtime PM put and can trigger a runtime PM usage count underflow. Fix this by making eh_noresume a regular bool so it can be accessed with READ_ONCE() and WRITE_ONCE(). Snapshot it once per EH iteration and use that snapshot for both runtime PM get and put decisions. Fixes: ae0751ffc77e ("[SCSI] add flag to skip the runtime PM calls on the host") Signed-off-by: Hongjie Fang <hongjiefang(a)asrmicro.com> Reviewed-by: Bart Van Assche <bvanassche(a)acm.org> Reviewed-by: Peter Wang <peter.wang(a)mediatek.com> Link: https://patch.msgid.link/20260729111614.2407559-1-hongjiefang@asrmicro.com Signed-off-by: Martin K. Petersen (Oracle) <mkp(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Conflicts: drivers/scsi/scsi_error.c drivers/ufs/core/ufshcd.c [ctx conflicts] Signed-off-by: Zhou Minqiang <zhouminqiang2(a)huawei.com> --- drivers/scsi/scsi_error.c | 9 +++++++-- drivers/ufs/core/ufshcd.c | 4 ++++ 2 files changed, 11 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c index 9b0ff9fad104..18474e441b87 100644 --- a/drivers/scsi/scsi_error.c +++ b/drivers/scsi/scsi_error.c @@ -2293,6 +2293,8 @@ static void scsi_unjam_host(struct Scsi_Host *shost) int scsi_error_handler(void *data) { struct Scsi_Host *shost = data; + bool eh_noresume; + unsigned long flags; /* * We use TASK_INTERRUPTIBLE so that the thread is not @@ -2334,7 +2336,10 @@ int scsi_error_handler(void *data) * what we need to do to get it up and online again (if we can). * If we fail, we end up taking the thing offline. */ - if (!shost->eh_noresume && scsi_autopm_get_host(shost) != 0) { + spin_lock_irqsave(shost->host_lock, flags); + eh_noresume = shost->eh_noresume; + spin_unlock_irqrestore(shost->host_lock, flags); + if (!eh_noresume && scsi_autopm_get_host(shost) != 0) { SCSI_LOG_ERROR_RECOVERY(1, shost_printk(KERN_ERR, shost, "scsi_eh_%d: unable to autoresume\n", @@ -2358,7 +2363,7 @@ int scsi_error_handler(void *data) * which are still online. */ scsi_restart_operations(shost); - if (!shost->eh_noresume) + if (!eh_noresume) scsi_autopm_put_host(shost); } __set_current_state(TASK_RUNNING); diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index 06f2626063a1..724e0225c932 100644 --- a/drivers/ufs/core/ufshcd.c +++ b/drivers/ufs/core/ufshcd.c @@ -9613,7 +9613,9 @@ static int ufshcd_set_dev_pwr_mode(struct ufs_hba *hba, * we are functional while we are here, skip host resume in error * handling context. */ + spin_lock_irqsave(hba->host->host_lock, flags); hba->host->eh_noresume = 1; + spin_unlock_irqrestore(hba->host->host_lock, flags); /* * Current function would be generally called from the power management @@ -9643,7 +9645,9 @@ static int ufshcd_set_dev_pwr_mode(struct ufs_hba *hba, } scsi_device_put(sdp); + spin_lock_irqsave(hba->host->host_lock, flags); hba->host->eh_noresume = 0; + spin_unlock_irqrestore(hba->host->host_lock, flags); return ret; } -- 2.52.0

2 1

[PATCH OLK-6.6 v2] md/raid1,raid10: fix bio accounting for split md cloned bios
by Zhou Minqiang 14 Sep '26

14 Sep '26

mainline inclusion from mainline-v7.2-rc1 commit ba976e3501111d11c550848b3b7341a73035f582 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/17965 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- Use md_cloned_bio() to control bio accounting instead of relying on r1bio_existed in raid1 or the io_accounting flag in raid10. The previous logic does not reliably reflect whether a bio is an md cloned bio. When a failed bio is split and resubmitted via bio_submit_split_bioset() on the error path, this can lead to either double accounting for md cloned bios, or missing accounting for bios returned from bio_submit_split_bioset() Fix this by using md_cloned_bio() to detect md cloned bios and skip accounting accordingly. Fixes: bb2a9acefaf9 ("md/raid1: switch to use md_account_bio() for io accounting") Fixes: 820455238366 ("md/raid10: switch to use md_account_bio() for io accounting") Signed-off-by: Abd-Alrhman Masalkhi <abd.masalkhi(a)gmail.com> Reviewed-by: Xiao Ni <xiao(a)kernel.org> Link: https://patch.msgid.link/20260501114652.590037-4-abd.masalkhi@gmail.com Signed-off-by: Yu Kuai <yukuai(a)fygo.io> Signed-off-by: Zhou Minqiang <zhouminqiang2(a)huawei.com> --- drivers/md/raid1.c | 2 +- drivers/md/raid10.c | 8 ++++---- 2 files changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index a4662d5077ba..80c1802c26e1 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -1412,7 +1412,7 @@ static void raid1_read_request(struct mddev *mddev, struct bio *bio, } r1_bio->read_disk = rdisk; - if (!r1bio_existed) { + if (likely(!md_cloned_bio(mddev, bio))) { md_account_bio(mddev, &bio); r1_bio->master_bio = bio; } diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index 316b6c96d413..e8ce8cdd7fce 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -1140,7 +1140,7 @@ static bool regular_request_wait(struct mddev *mddev, struct r10conf *conf, } static void raid10_read_request(struct mddev *mddev, struct bio *bio, - struct r10bio *r10_bio, bool io_accounting) + struct r10bio *r10_bio) { struct r10conf *conf = mddev->private; struct bio *read_bio; @@ -1218,7 +1218,7 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio, } slot = r10_bio->read_slot; - if (io_accounting) { + if (likely(!md_cloned_bio(mddev, bio))) { md_account_bio(mddev, &bio); r10_bio->master_bio = bio; } @@ -1514,7 +1514,7 @@ static bool __make_request(struct mddev *mddev, struct bio *bio, int sectors) ret = true; if (bio_data_dir(bio) == READ) - raid10_read_request(mddev, bio, r10_bio, true); + raid10_read_request(mddev, bio, r10_bio); else ret = raid10_write_request(mddev, bio, r10_bio); @@ -2839,7 +2839,7 @@ static void handle_read_error(struct mddev *mddev, struct r10bio *r10_bio) rdev_dec_pending(rdev, mddev); r10_bio->state = 0; - raid10_read_request(mddev, r10_bio->master_bio, r10_bio, false); + raid10_read_request(mddev, r10_bio->master_bio, r10_bio); /* * allow_barrier after re-submit to ensure no sync io * can be issued while regular io pending. -- 2.52.0

2 1