From 28585a832602747cbfa88ad8934013177a3aae38 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Fri, 22 Sep 2017 14:10:22 -0700 Subject: [PATCH 1/4] rcu: Allow for page faults in NMI handlers A number of architecture invoke rcu_irq_enter() on exception entry in order to allow RCU read-side critical sections in the exception handler when the exception is from an idle or nohz_full CPU. This works, at least unless the exception happens in an NMI handler. In that case, rcu_nmi_enter() would already have exited the extended quiescent state, which would mean that rcu_irq_enter() would (incorrectly) cause RCU to think that it is again in an extended quiescent state. This will in turn result in lockdep splats in response to later RCU read-side critical sections. This commit therefore causes rcu_irq_enter() and rcu_irq_exit() to take no action if there is an rcu_nmi_enter() in effect, thus avoiding the unscheduled return to RCU quiescent state. This in turn should make the kernel safe for on-demand RCU voyeurism. Link: http://lkml.kernel.org/r/20170922211022.GA18084@linux.vnet.ibm.com Cc: stable@vger.kernel.org Fixes: 0be964be0 ("module: Sanitize RCU usage and locking") Reported-by: Steven Rostedt Signed-off-by: Paul E. McKenney Signed-off-by: Steven Rostedt (VMware) --- kernel/rcu/tree.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c index 51d4c3acf32d..63bee8e1b193 100644 --- a/kernel/rcu/tree.c +++ b/kernel/rcu/tree.c @@ -888,6 +888,11 @@ void rcu_irq_exit(void) RCU_LOCKDEP_WARN(!irqs_disabled(), "rcu_irq_exit() invoked with irqs enabled!!!"); rdtp = this_cpu_ptr(&rcu_dynticks); + + /* Page faults can happen in NMI handlers, so check... */ + if (READ_ONCE(rdtp->dynticks_nmi_nesting)) + return; + WARN_ON_ONCE(IS_ENABLED(CONFIG_RCU_EQS_DEBUG) && rdtp->dynticks_nesting < 1); if (rdtp->dynticks_nesting <= 1) { @@ -1020,6 +1025,11 @@ void rcu_irq_enter(void) RCU_LOCKDEP_WARN(!irqs_disabled(), "rcu_irq_enter() invoked with irqs enabled!!!"); rdtp = this_cpu_ptr(&rcu_dynticks); + + /* Page faults can happen in NMI handlers, so check... */ + if (READ_ONCE(rdtp->dynticks_nmi_nesting)) + return; + oldval = rdtp->dynticks_nesting; rdtp->dynticks_nesting++; WARN_ON_ONCE(IS_ENABLED(CONFIG_RCU_EQS_DEBUG) && From 9aadde91b3c035413c806619beb3e3ef6e697953 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (VMware)" Date: Fri, 22 Sep 2017 17:22:19 -0400 Subject: [PATCH 2/4] extable: Consolidate *kernel_text_address() functions The functionality between kernel_text_address() and _kernel_text_address() is the same except that _kernel_text_address() does a little more (that function needs a rename, but that can be done another time). Instead of having duplicate code in both, simply have _kernel_text_address() calls kernel_text_address() instead. This is marked for stable because there's an RCU bug that can happen if one of these functions gets called while RCU is not watching. That fix depends on this fix to keep from having to write the fix twice. Cc: stable@vger.kernel.org Fixes: 0be964be0 ("module: Sanitize RCU usage and locking") Acked-by: Paul E. McKenney Signed-off-by: Steven Rostedt (VMware) --- kernel/extable.c | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/kernel/extable.c b/kernel/extable.c index 38c2412401a1..a7024a494faf 100644 --- a/kernel/extable.c +++ b/kernel/extable.c @@ -102,15 +102,7 @@ int core_kernel_data(unsigned long addr) int __kernel_text_address(unsigned long addr) { - if (core_kernel_text(addr)) - return 1; - if (is_module_text_address(addr)) - return 1; - if (is_ftrace_trampoline(addr)) - return 1; - if (is_kprobe_optinsn_slot(addr) || is_kprobe_insn_slot(addr)) - return 1; - if (is_bpf_text_address(addr)) + if (kernel_text_address(addr)) return 1; /* * There might be init symbols in saved stacktraces. From e8cac8b1d10589be45671a5ade0926a639b543b7 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (VMware)" Date: Fri, 22 Sep 2017 17:36:32 -0400 Subject: [PATCH 3/4] extable: Enable RCU if it is not watching in kernel_text_address() If kernel_text_address() is called when RCU is not watching, it can cause an RCU bug because is_module_text_address(), the is_kprobe_*insn_slot() and is_bpf_text_address() functions require the use of RCU. Only enable RCU if it is not currently watching before it calls is_module_text_address(). The use of rcu_nmi_enter() is used to enable RCU because kernel_text_address() can happen pretty much anywhere (like an NMI), and even from within an NMI. It is called via save_stack_trace() that can be called by any WARN() or tracing function, which can happen while RCU is not watching (for example, going to or coming from idle, or during CPU take down or bring up). Cc: stable@vger.kernel.org Fixes: 0be964be0 ("module: Sanitize RCU usage and locking") Acked-by: Paul E. McKenney Signed-off-by: Steven Rostedt (VMware) --- kernel/extable.c | 35 ++++++++++++++++++++++++++++++----- 1 file changed, 30 insertions(+), 5 deletions(-) diff --git a/kernel/extable.c b/kernel/extable.c index a7024a494faf..9aa1cc41ecf7 100644 --- a/kernel/extable.c +++ b/kernel/extable.c @@ -119,17 +119,42 @@ int __kernel_text_address(unsigned long addr) int kernel_text_address(unsigned long addr) { + bool no_rcu; + int ret = 1; + if (core_kernel_text(addr)) return 1; + + /* + * If a stack dump happens while RCU is not watching, then + * RCU needs to be notified that it requires to start + * watching again. This can happen either by tracing that + * triggers a stack trace, or a WARN() that happens during + * coming back from idle, or cpu on or offlining. + * + * is_module_text_address() as well as the kprobe slots + * and is_bpf_text_address() require RCU to be watching. + */ + no_rcu = !rcu_is_watching(); + + /* Treat this like an NMI as it can happen anywhere */ + if (no_rcu) + rcu_nmi_enter(); + if (is_module_text_address(addr)) - return 1; + goto out; if (is_ftrace_trampoline(addr)) - return 1; + goto out; if (is_kprobe_optinsn_slot(addr) || is_kprobe_insn_slot(addr)) - return 1; + goto out; if (is_bpf_text_address(addr)) - return 1; - return 0; + goto out; + ret = 0; +out: + if (no_rcu) + rcu_nmi_exit(); + + return ret; } /* From 15516c89acce948debc4c598e03c3fee53045797 Mon Sep 17 00:00:00 2001 From: "Steven Rostedt (VMware)" Date: Thu, 21 Sep 2017 13:00:21 -0400 Subject: [PATCH 4/4] tracing: Remove RCU work arounds from stack tracer Currently the stack tracer calls rcu_irq_enter() to make sure RCU is watching when it records a stack trace. But if the stack tracer is triggered while tracing inside of a rcu_irq_enter(), calling rcu_irq_enter() unconditionally can be problematic. The reason for having rcu_irq_enter() in the first place has been fixed from within the saving of the stack trace code, and there's no reason for doing it in the stack tracer itself. Just remove it. Cc: stable@vger.kernel.org Fixes: 0be964be0 ("module: Sanitize RCU usage and locking") Acked-by: Paul E. McKenney Suggested-by: "Paul E. McKenney" Signed-off-by: Steven Rostedt (VMware) --- kernel/trace/trace_stack.c | 15 --------------- 1 file changed, 15 deletions(-) diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c index a4df67cbc711..49cb41412eec 100644 --- a/kernel/trace/trace_stack.c +++ b/kernel/trace/trace_stack.c @@ -96,23 +96,9 @@ check_stack(unsigned long ip, unsigned long *stack) if (in_nmi()) return; - /* - * There's a slight chance that we are tracing inside the - * RCU infrastructure, and rcu_irq_enter() will not work - * as expected. - */ - if (unlikely(rcu_irq_enter_disabled())) - return; - local_irq_save(flags); arch_spin_lock(&stack_trace_max_lock); - /* - * RCU may not be watching, make it see us. - * The stack trace code uses rcu_sched. - */ - rcu_irq_enter(); - /* In case another CPU set the tracer_frame on us */ if (unlikely(!frame_size)) this_size -= tracer_frame; @@ -205,7 +191,6 @@ check_stack(unsigned long ip, unsigned long *stack) } out: - rcu_irq_exit(); arch_spin_unlock(&stack_trace_max_lock); local_irq_restore(flags); }