Will Deacon
2022-May-05 10:49 UTC
[PATCH v2] arm64: paravirt: Use RCU read locks to guard stolen_time
Hi Elliot, On Thu, Apr 28, 2022 at 11:35:36AM -0700, Elliot Berman wrote:> From: Prakruthi Deepak Heragu <quic_pheragu at quicinc.com> > > During hotplug, the stolen time data structure is unmapped and memset. > There is a possibility of the timer IRQ being triggered before memset > and stolen time is getting updated as part of this timer IRQ handler. This > causes the below crash in timer handler - > > [ 3457.473139][ C5] Unable to handle kernel paging request at virtual address ffffffc03df05148 > ... > [ 3458.154398][ C5] Call trace: > [ 3458.157648][ C5] para_steal_clock+0x30/0x50 > [ 3458.162319][ C5] irqtime_account_process_tick+0x30/0x194 > [ 3458.168148][ C5] account_process_tick+0x3c/0x280 > [ 3458.173274][ C5] update_process_times+0x5c/0xf4 > [ 3458.178311][ C5] tick_sched_timer+0x180/0x384 > [ 3458.183164][ C5] __run_hrtimer+0x160/0x57c > [ 3458.187744][ C5] hrtimer_interrupt+0x258/0x684 > [ 3458.192698][ C5] arch_timer_handler_virt+0x5c/0xa0 > [ 3458.198002][ C5] handle_percpu_devid_irq+0xdc/0x414 > [ 3458.203385][ C5] handle_domain_irq+0xa8/0x168 > [ 3458.208241][ C5] gic_handle_irq.34493+0x54/0x244 > [ 3458.213359][ C5] call_on_irq_stack+0x40/0x70 > [ 3458.218125][ C5] do_interrupt_handler+0x60/0x9c > [ 3458.223156][ C5] el1_interrupt+0x34/0x64 > [ 3458.227560][ C5] el1h_64_irq_handler+0x1c/0x2c > [ 3458.232503][ C5] el1h_64_irq+0x7c/0x80 > [ 3458.236736][ C5] free_vmap_area_noflush+0x108/0x39c > [ 3458.242126][ C5] remove_vm_area+0xbc/0x118 > [ 3458.246714][ C5] vm_remove_mappings+0x48/0x2a4 > [ 3458.251656][ C5] __vunmap+0x154/0x278 > [ 3458.255796][ C5] stolen_time_cpu_down_prepare+0xc0/0xd8 > [ 3458.261542][ C5] cpuhp_invoke_callback+0x248/0xc34 > [ 3458.266842][ C5] cpuhp_thread_fun+0x1c4/0x248 > [ 3458.271696][ C5] smpboot_thread_fn+0x1b0/0x400 > [ 3458.276638][ C5] kthread+0x17c/0x1e0 > [ 3458.280691][ C5] ret_from_fork+0x10/0x20 > > As a fix, introduce rcu lock to update stolen time structure. > > Fixes: 75df529bec91 ("arm64: paravirt: Initialize steal time when cpu is online") > Cc: stable at vger.kernel.org > Signed-off-by: Prakruthi Deepak Heragu <quic_pheragu at quicinc.com> > Signed-off-by: Elliot Berman <quic_eberman at quicinc.com> > --- > Changes since v1: https://lore.kernel.org/all/20220420204417.155194-1-quic_eberman at quicinc.com/ > - Use RCU instead of disabling interrupts > > arch/arm64/kernel/paravirt.c | 24 +++++++++++++++++++----- > 1 file changed, 19 insertions(+), 5 deletions(-)I applied this locally, but sparse is complaining because the 'kaddr' field of 'struct pv_time_stolen_time_region' is missing an '__rcu' annotation: | arch/arm64/kernel/paravirt.c:112:9: warning: cast adds address space '__rcu' to expression [sparse] | arch/arm64/kernel/paravirt.c:112:9: error: incompatible types in comparison expression (different address spaces): [sparse] | arch/arm64/kernel/paravirt.c:112:9: struct pvclock_vcpu_stolen_time [noderef] __rcu * [sparse] | arch/arm64/kernel/paravirt.c:112:9: struct pvclock_vcpu_stolen_time * [sparse] | arch/arm64/kernel/paravirt.c:67:17: warning: cast adds address space '__rcu' to expression [sparse] | arch/arm64/kernel/paravirt.c:67:17: error: incompatible types in comparison expression (different address spaces): [sparse] | arch/arm64/kernel/paravirt.c:67:17: struct pvclock_vcpu_stolen_time [noderef] __rcu * [sparse] | arch/arm64/kernel/paravirt.c:67:17: struct pvclock_vcpu_stolen_time * [sparse] | arch/arm64/kernel/paravirt.c:88:9: warning: cast adds address space '__rcu' to expression [sparse] | arch/arm64/kernel/paravirt.c:88:9: error: incompatible types in comparison expression (different address spaces): [sparse] | arch/arm64/kernel/paravirt.c:88:9: struct pvclock_vcpu_stolen_time [noderef] __rcu * [sparse] | arch/arm64/kernel/paravirt.c:88:9: struct pvclock_vcpu_stolen_time * [sparse] The diff below seems to make it happy again, but please can you take a look? Cheers, Will --->8 diff --git a/arch/arm64/kernel/paravirt.c b/arch/arm64/kernel/paravirt.c index e724ea3d86f0..57c7c211f8c7 100644 --- a/arch/arm64/kernel/paravirt.c +++ b/arch/arm64/kernel/paravirt.c @@ -35,7 +35,7 @@ static u64 native_steal_clock(int cpu) DEFINE_STATIC_CALL(pv_steal_clock, native_steal_clock); struct pv_time_stolen_time_region { - struct pvclock_vcpu_stolen_time *kaddr; + struct pvclock_vcpu_stolen_time __rcu *kaddr; }; static DEFINE_PER_CPU(struct pv_time_stolen_time_region, stolen_time_region); @@ -84,8 +84,7 @@ static int stolen_time_cpu_down_prepare(unsigned int cpu) if (!reg->kaddr) return 0; - kaddr = reg->kaddr; - rcu_assign_pointer(reg->kaddr, NULL); + kaddr = rcu_replace_pointer(reg->kaddr, NULL, true); synchronize_rcu(); memunmap(kaddr); @@ -116,8 +115,8 @@ static int stolen_time_cpu_online(unsigned int cpu) return -ENOMEM; } - if (le32_to_cpu(reg->kaddr->revision) != 0 || - le32_to_cpu(reg->kaddr->attributes) != 0) { + if (le32_to_cpu(kaddr->revision) != 0 || + le32_to_cpu(kaddr->attributes) != 0) { pr_warn_once("Unexpected revision or attributes in stolen time data\n"); return -ENXIO; }