From 6a25212dc35897e36f83860df3cefcdf5fb5c189 Mon Sep 17 00:00:00 2001 From: Prathu Baronia Date: Tue, 2 May 2023 14:32:40 +0530 Subject: kthread: fix spelling typo and grammar in comments - `If present` -> `If present,' - `reuturn` -> `return` - `function exit safely` -> `function to exit safely` Link: https://lkml.kernel.org/r/20230502090242.3037194-1-quic_pbaronia@quicinc.com Signed-off-by: Prathu Baronia Cc: Eric W. Biederman Cc: Jason A. Donenfeld Cc: Kees Cook Cc: Petr Mladek Cc: Sami Tolvanen Cc: Tejun Heo Cc: Zqiang Signed-off-by: Andrew Morton --- kernel/kthread.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/kthread.c b/kernel/kthread.c index 490792b1066e..08455fb48ba6 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -312,10 +312,10 @@ void __noreturn kthread_exit(long result) * @comp: Completion to complete * @code: The integer value to return to kthread_stop(). * - * If present complete @comp and the reuturn code to kthread_stop(). + * If present, complete @comp and then return code to kthread_stop(). * * A kernel thread whose module may be removed after the completion of - * @comp can use this function exit safely. + * @comp can use this function to exit safely. * * Does not return. */ -- cgit v1.2.3-58-ga151 From 4e2f6342ccaf9d8beb33b398cc2862769c7d2a3b Mon Sep 17 00:00:00 2001 From: Haifeng Xu Date: Mon, 8 May 2023 06:44:58 +0000 Subject: fork: optimize memcg_charge_kernel_stack() a bit Since commit f1c1a9ee00e4 ("fork: Move memcg_charge_kernel_stack() into CONFIG_VMAP_STACK"), memcg_charge_kernel_stack() has been moved into CONFIG_VMAP_STACK block, so the CONFIG_VMAP_STACK check can be removed. Furthermore, memcg_charge_kernel_stack() is only invoked by alloc_thread_stack_node() instead of dup_task_struct(). If memcg_kmem_charge_page() fails, the uncharge process is handled in memcg_charge_kernel_stack() itself instead of free_thread_stack(), so remove the incorrect comments. If memcg_charge_kernel_stack() fails to charge pages used by kernel stack, only charged pages need to be uncharged. It's unnecessary to uncharge those pages which memory cgroup pointer is NULL. [akpm@linux-foundation.org: remove assertion that PAGE_SIZE is a multiple of 1k] Link: https://lkml.kernel.org/r/20230508064458.32855-1-haifeng.xu@shopee.com Signed-off-by: Haifeng Xu Cc: Andy Lutomirski Cc: Daniel Bristot de Oliveira Cc: Sebastian Andrzej Siewior Cc: Thomas Gleixner Signed-off-by: Andrew Morton --- kernel/fork.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index ed4e01daccaa..d17995934eb4 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -252,23 +252,19 @@ static int memcg_charge_kernel_stack(struct vm_struct *vm) { int i; int ret; + int nr_charged = 0; - BUILD_BUG_ON(IS_ENABLED(CONFIG_VMAP_STACK) && PAGE_SIZE % 1024 != 0); BUG_ON(vm->nr_pages != THREAD_SIZE / PAGE_SIZE); for (i = 0; i < THREAD_SIZE / PAGE_SIZE; i++) { ret = memcg_kmem_charge_page(vm->pages[i], GFP_KERNEL, 0); if (ret) goto err; + nr_charged++; } return 0; err: - /* - * If memcg_kmem_charge_page() fails, page's memory cgroup pointer is - * NULL, and memcg_kmem_uncharge_page() in free_thread_stack() will - * ignore this page. - */ - for (i = 0; i < THREAD_SIZE / PAGE_SIZE; i++) + for (i = 0; i < nr_charged; i++) memcg_kmem_uncharge_page(vm->pages[i], 0); return ret; } -- cgit v1.2.3-58-ga151 From ff7138813ac4a20628ce4b40952c69faba835fbb Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 17 May 2023 15:10:54 +0200 Subject: locking: add lockevent_read() prototype lockevent_read() has a __weak definition and the only caller in kernel/locking/lock_events.c, plus a strong definition in qspinlock_stat.h that overrides it, but no other declaration. This causes a W=1 warning: kernel/locking/lock_events.c:61:16: error: no previous prototype for 'lockevent_read' [-Werror=missing-prototypes] Add shared prototype to avoid the warnings. Link: https://lkml.kernel.org/r/20230517131102.934196-7-arnd@kernel.org Signed-off-by: Arnd Bergmann Cc: Boqun Feng Cc: Catalin Marinas Cc: Christoph Lameter Cc: Dennis Zhou Cc: Eric Paris Cc: Heiko Carstens Cc: Helge Deller Cc: Ingo Molnar Cc: Michael Ellerman Cc: Michal Simek Cc: Palmer Dabbelt Cc: Paul Moore Cc: Pavel Machek Cc: Peter Zijlstra Cc: Rafael J. Wysocki Cc: Russell King Cc: Tejun Heo Cc: Thomas Bogendoerfer Cc: Thomas Gleixner Cc: Waiman Long Cc: Will Deacon Signed-off-by: Andrew Morton --- kernel/locking/lock_events.h | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/locking/lock_events.h b/kernel/locking/lock_events.h index 8c7e7d25f09c..a6016b91803d 100644 --- a/kernel/locking/lock_events.h +++ b/kernel/locking/lock_events.h @@ -57,4 +57,8 @@ static inline void __lockevent_add(enum lock_events event, int inc) #define lockevent_cond_inc(ev, c) #endif /* CONFIG_LOCK_EVENT_COUNTS */ + +ssize_t lockevent_read(struct file *file, char __user *user_buf, + size_t count, loff_t *ppos); + #endif /* __LOCKING_LOCK_EVENTS_H */ -- cgit v1.2.3-58-ga151 From 525bb813a995283bef759659cfc00f3fc2d51e81 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 17 May 2023 15:10:55 +0200 Subject: panic: hide unused global functions Building with W=1 shows warnings about two functions that have no declaration or caller in certain configurations: kernel/panic.c:688:6: error: no previous prototype for 'warn_slowpath_fmt' [-Werror=missing-prototypes] kernel/panic.c:710:6: error: no previous prototype for '__warn_printk' [-Werror=missing-prototypes] Enclose the definition in the same #ifdef check as the declaration. Link: https://lkml.kernel.org/r/20230517131102.934196-8-arnd@kernel.org Signed-off-by: Arnd Bergmann Cc: Boqun Feng Cc: Catalin Marinas Cc: Christoph Lameter Cc: Dennis Zhou Cc: Eric Paris Cc: Heiko Carstens Cc: Helge Deller Cc: Ingo Molnar Cc: Michael Ellerman Cc: Michal Simek Cc: Palmer Dabbelt Cc: Paul Moore Cc: Pavel Machek Cc: Peter Zijlstra Cc: Rafael J. Wysocki Cc: Russell King Cc: Tejun Heo Cc: Thomas Bogendoerfer Cc: Thomas Gleixner Cc: Waiman Long Cc: Will Deacon Signed-off-by: Andrew Morton --- kernel/panic.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/panic.c b/kernel/panic.c index 886d2ebd0a0d..10effe40a3fa 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -684,6 +684,7 @@ void __warn(const char *file, int line, void *caller, unsigned taint, add_taint(taint, LOCKDEP_STILL_OK); } +#ifdef CONFIG_BUG #ifndef __WARN_FLAGS void warn_slowpath_fmt(const char *file, int line, unsigned taint, const char *fmt, ...) @@ -722,8 +723,6 @@ void __warn_printk(const char *fmt, ...) EXPORT_SYMBOL(__warn_printk); #endif -#ifdef CONFIG_BUG - /* Support resetting WARN*_ONCE state */ static int clear_warn_once_set(void *data, u64 val) -- cgit v1.2.3-58-ga151 From e0ddec73fd4822d2ffe914d5ce3e2718f985276a Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 17 May 2023 14:49:25 +0200 Subject: kcov: add prototypes for helper functions A number of internal functions in kcov are only called from generated code and don't technically need a declaration, but 'make W=1' warns about global symbols without a prototype: kernel/kcov.c:199:14: error: no previous prototype for '__sanitizer_cov_trace_pc' [-Werror=missing-prototypes] kernel/kcov.c:264:14: error: no previous prototype for '__sanitizer_cov_trace_cmp1' [-Werror=missing-prototypes] kernel/kcov.c:270:14: error: no previous prototype for '__sanitizer_cov_trace_cmp2' [-Werror=missing-prototypes] kernel/kcov.c:276:14: error: no previous prototype for '__sanitizer_cov_trace_cmp4' [-Werror=missing-prototypes] kernel/kcov.c:282:14: error: no previous prototype for '__sanitizer_cov_trace_cmp8' [-Werror=missing-prototypes] kernel/kcov.c:288:14: error: no previous prototype for '__sanitizer_cov_trace_const_cmp1' [-Werror=missing-prototypes] kernel/kcov.c:295:14: error: no previous prototype for '__sanitizer_cov_trace_const_cmp2' [-Werror=missing-prototypes] kernel/kcov.c:302:14: error: no previous prototype for '__sanitizer_cov_trace_const_cmp4' [-Werror=missing-prototypes] kernel/kcov.c:309:14: error: no previous prototype for '__sanitizer_cov_trace_const_cmp8' [-Werror=missing-prototypes] kernel/kcov.c:316:14: error: no previous prototype for '__sanitizer_cov_trace_switch' [-Werror=missing-prototypes] Adding prototypes for these in a header solves that problem, but now there is a mismatch between the built-in type and the prototype on 64-bit architectures because they expect some functions to take a 64-bit 'unsigned long' argument rather than an 'unsigned long long' u64 type: include/linux/kcov.h:84:6: error: conflicting types for built-in function '__sanitizer_cov_trace_switch'; expected 'void(long long unsigned int, void *)' [-Werror=builtin-declaration-mismatch] 84 | void __sanitizer_cov_trace_switch(u64 val, u64 *cases); | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~ Avoid this as well with a custom type definition. Link: https://lkml.kernel.org/r/20230517124944.929997-1-arnd@kernel.org Signed-off-by: Arnd Bergmann Cc: Andrey Konovalov Cc: Dmitry Vyukov Cc: Rong Tao Signed-off-by: Andrew Morton --- include/linux/kcov.h | 17 +++++++++++++++++ kernel/kcov.c | 7 ++++--- 2 files changed, 21 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/include/linux/kcov.h b/include/linux/kcov.h index ee04256f28af..b851ba415e03 100644 --- a/include/linux/kcov.h +++ b/include/linux/kcov.h @@ -72,6 +72,23 @@ static inline void kcov_remote_stop_softirq(void) kcov_remote_stop(); } +#ifdef CONFIG_64BIT +typedef unsigned long kcov_u64; +#else +typedef unsigned long long kcov_u64; +#endif + +void __sanitizer_cov_trace_pc(void); +void __sanitizer_cov_trace_cmp1(u8 arg1, u8 arg2); +void __sanitizer_cov_trace_cmp2(u16 arg1, u16 arg2); +void __sanitizer_cov_trace_cmp4(u32 arg1, u32 arg2); +void __sanitizer_cov_trace_cmp8(kcov_u64 arg1, kcov_u64 arg2); +void __sanitizer_cov_trace_const_cmp1(u8 arg1, u8 arg2); +void __sanitizer_cov_trace_const_cmp2(u16 arg1, u16 arg2); +void __sanitizer_cov_trace_const_cmp4(u32 arg1, u32 arg2); +void __sanitizer_cov_trace_const_cmp8(kcov_u64 arg1, kcov_u64 arg2); +void __sanitizer_cov_trace_switch(kcov_u64 val, void *cases); + #else static inline void kcov_task_init(struct task_struct *t) {} diff --git a/kernel/kcov.c b/kernel/kcov.c index 84c717337df0..f9ac2e9e460f 100644 --- a/kernel/kcov.c +++ b/kernel/kcov.c @@ -279,7 +279,7 @@ void notrace __sanitizer_cov_trace_cmp4(u32 arg1, u32 arg2) } EXPORT_SYMBOL(__sanitizer_cov_trace_cmp4); -void notrace __sanitizer_cov_trace_cmp8(u64 arg1, u64 arg2) +void notrace __sanitizer_cov_trace_cmp8(kcov_u64 arg1, kcov_u64 arg2) { write_comp_data(KCOV_CMP_SIZE(3), arg1, arg2, _RET_IP_); } @@ -306,16 +306,17 @@ void notrace __sanitizer_cov_trace_const_cmp4(u32 arg1, u32 arg2) } EXPORT_SYMBOL(__sanitizer_cov_trace_const_cmp4); -void notrace __sanitizer_cov_trace_const_cmp8(u64 arg1, u64 arg2) +void notrace __sanitizer_cov_trace_const_cmp8(kcov_u64 arg1, kcov_u64 arg2) { write_comp_data(KCOV_CMP_SIZE(3) | KCOV_CMP_CONST, arg1, arg2, _RET_IP_); } EXPORT_SYMBOL(__sanitizer_cov_trace_const_cmp8); -void notrace __sanitizer_cov_trace_switch(u64 val, u64 *cases) +void notrace __sanitizer_cov_trace_switch(kcov_u64 val, void *arg) { u64 i; + u64 *cases = arg; u64 count = cases[0]; u64 size = cases[1]; u64 type = KCOV_CMP_CONST; -- cgit v1.2.3-58-ga151 From 4379e59fe5665cfda737e45b8bf2f05321ef049c Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:26 -0700 Subject: watchdog/perf: more properly prevent false positives with turbo modes Currently, in the watchdog_overflow_callback() we first check to see if the watchdog had been touched and _then_ we handle the workaround for turbo mode. This order should be reversed. Specifically, "touching" the hardlockup detector's watchdog should avoid lockups being detected for one period that should be roughly the same regardless of whether we're running turbo or not. That means that we should do the extra accounting for turbo _before_ we look at (and clear) the global indicating that we've been touched. NOTE: this fix is made based on code inspection. I am not aware of any reports where the old code would have generated false positives. That being said, this order seems more correct and also makes it easier down the line to share code with the "buddy" hardlockup detector. Link: https://lkml.kernel.org/r/20230519101840.v5.2.I843b0d1de3e096ba111a179f3adb16d576bef5c7@changeid Fixes: 7edaeb6841df ("kernel/watchdog: Prevent false positives with turbo modes") Signed-off-by: Douglas Anderson Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Lecopzer Chen Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Petr Mladek Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- kernel/watchdog_hld.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c index 247bf0b1582c..1e8a49dc956e 100644 --- a/kernel/watchdog_hld.c +++ b/kernel/watchdog_hld.c @@ -114,14 +114,14 @@ static void watchdog_overflow_callback(struct perf_event *event, /* Ensure the watchdog never gets throttled */ event->hw.interrupts = 0; + if (!watchdog_check_timestamp()) + return; + if (__this_cpu_read(watchdog_nmi_touch) == true) { __this_cpu_write(watchdog_nmi_touch, false); return; } - if (!watchdog_check_timestamp()) - return; - /* check for a hardlockup * This is done by making sure our timer interrupt * is incrementing. The timer interrupt should have -- cgit v1.2.3-58-ga151 From 810b560e8985725dbd57bbb3f188c231365eb5ae Mon Sep 17 00:00:00 2001 From: Lecopzer Chen Date: Fri, 19 May 2023 10:18:27 -0700 Subject: watchdog: remove WATCHDOG_DEFAULT No reference to WATCHDOG_DEFAULT, remove it. Link: https://lkml.kernel.org/r/20230519101840.v5.3.I6a729209a1320e0ad212176e250ff945b8f91b2a@changeid Signed-off-by: Pingfan Liu Signed-off-by: Lecopzer Chen Signed-off-by: Douglas Anderson Reviewed-by: Petr Mladek Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- kernel/watchdog.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 8e61f21e7e33..582d572e1379 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -30,10 +30,8 @@ static DEFINE_MUTEX(watchdog_mutex); #if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG) -# define WATCHDOG_DEFAULT (SOFT_WATCHDOG_ENABLED | NMI_WATCHDOG_ENABLED) # define NMI_WATCHDOG_DEFAULT 1 #else -# define WATCHDOG_DEFAULT (SOFT_WATCHDOG_ENABLED) # define NMI_WATCHDOG_DEFAULT 0 #endif -- cgit v1.2.3-58-ga151 From 730211182ed083898fa5feb4b28459ffac4c9615 Mon Sep 17 00:00:00 2001 From: Lecopzer Chen Date: Fri, 19 May 2023 10:18:28 -0700 Subject: watchdog/hardlockup: change watchdog_nmi_enable() to void Nobody cares about the return value of watchdog_nmi_enable(), changing its prototype to void. Link: https://lkml.kernel.org/r/20230519101840.v5.4.Ic3a19b592eb1ac4c6f6eade44ffd943e8637b6e5@changeid Signed-off-by: Pingfan Liu Signed-off-by: Lecopzer Chen Signed-off-by: Douglas Anderson Reviewed-by: Petr Mladek Acked-by: Nicholas Piggin Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- arch/sparc/kernel/nmi.c | 8 +++----- include/linux/nmi.h | 2 +- kernel/watchdog.c | 3 +-- 3 files changed, 5 insertions(+), 8 deletions(-) (limited to 'kernel') diff --git a/arch/sparc/kernel/nmi.c b/arch/sparc/kernel/nmi.c index 060fff95a305..5dcf31f7e81f 100644 --- a/arch/sparc/kernel/nmi.c +++ b/arch/sparc/kernel/nmi.c @@ -282,11 +282,11 @@ __setup("nmi_watchdog=", setup_nmi_watchdog); * sparc specific NMI watchdog enable function. * Enables watchdog if it is not enabled already. */ -int watchdog_nmi_enable(unsigned int cpu) +void watchdog_nmi_enable(unsigned int cpu) { if (atomic_read(&nmi_active) == -1) { pr_warn("NMI watchdog cannot be enabled or disabled\n"); - return -1; + return; } /* @@ -295,11 +295,9 @@ int watchdog_nmi_enable(unsigned int cpu) * process first. */ if (!nmi_init_done) - return 0; + return; smp_call_function_single(cpu, start_nmi_watchdog, NULL, 1); - - return 0; } /* * sparc specific NMI watchdog disable function. diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 771d77b62bc1..454fe99c4874 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -119,7 +119,7 @@ static inline int hardlockup_detector_perf_init(void) { return 0; } void watchdog_nmi_stop(void); void watchdog_nmi_start(void); int watchdog_nmi_probe(void); -int watchdog_nmi_enable(unsigned int cpu); +void watchdog_nmi_enable(unsigned int cpu); void watchdog_nmi_disable(unsigned int cpu); void lockup_detector_reconfigure(void); diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 582d572e1379..c705a18b26bf 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -93,10 +93,9 @@ __setup("nmi_watchdog=", hardlockup_panic_setup); * softlockup watchdog start and stop. The arch must select the * SOFTLOCKUP_DETECTOR Kconfig. */ -int __weak watchdog_nmi_enable(unsigned int cpu) +void __weak watchdog_nmi_enable(unsigned int cpu) { hardlockup_detector_perf_enable(); - return 0; } void __weak watchdog_nmi_disable(unsigned int cpu) -- cgit v1.2.3-58-ga151 From 1fafaa7745eeeeffef7155ab5eeb8cc83d04874f Mon Sep 17 00:00:00 2001 From: Pingfan Liu Date: Fri, 19 May 2023 10:18:29 -0700 Subject: watchdog/perf: ensure CPU-bound context when creating hardlockup detector event hardlockup_detector_event_create() should create perf_event on the current CPU. Preemption could not get disabled because perf_event_create_kernel_counter() allocates memory. Instead, the CPU locality is achieved by processing the code in a per-CPU bound kthread. Add a check to prevent mistakes when calling the code in another code path. Link: https://lkml.kernel.org/r/20230519101840.v5.5.I654063e53782b11d53e736a8ad4897ffd207406a@changeid Signed-off-by: Pingfan Liu Co-developed-by: Lecopzer Chen Signed-off-by: Lecopzer Chen Signed-off-by: Douglas Anderson Reviewed-by: Petr Mladek Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- kernel/watchdog_hld.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c index 1e8a49dc956e..2125b09e09d7 100644 --- a/kernel/watchdog_hld.c +++ b/kernel/watchdog_hld.c @@ -165,10 +165,16 @@ static void watchdog_overflow_callback(struct perf_event *event, static int hardlockup_detector_event_create(void) { - unsigned int cpu = smp_processor_id(); + unsigned int cpu; struct perf_event_attr *wd_attr; struct perf_event *evt; + /* + * Preemption is not disabled because memory will be allocated. + * Ensure CPU-locality by calling this in per-CPU kthread. + */ + WARN_ON(!is_percpu_thread()); + cpu = raw_smp_processor_id(); wd_attr = &wd_hw_attr; wd_attr->sample_period = hw_nmi_get_sample_period(watchdog_thresh); -- cgit v1.2.3-58-ga151 From 6ea0d04211a7715a926f5dca4aa1065614fa64da Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:31 -0700 Subject: watchdog/perf: rename watchdog_hld.c to watchdog_perf.c The code currently in "watchdog_hld.c" is for detecting hardlockups using perf, as evidenced by the line in the Makefile that only compiles this file if CONFIG_HARDLOCKUP_DETECTOR_PERF is defined. Rename the file to prepare for the buddy hardlockup detector, which doesn't use perf. It could be argued that the new name makes it less obvious that this is a hardlockup detector. While true, it's not hard to remember that the "perf" detector is always a hardlockup detector and it's nice not to have names that are too convoluted. Link: https://lkml.kernel.org/r/20230519101840.v5.7.Ice803cb078d0e15fb2cbf49132f096ee2bd4199d@changeid Signed-off-by: Douglas Anderson Acked-by: Nicholas Piggin Reviewed-by: Petr Mladek Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Lecopzer Chen Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- kernel/Makefile | 2 +- kernel/watchdog_hld.c | 302 ------------------------------------------------- kernel/watchdog_perf.c | 302 +++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 303 insertions(+), 303 deletions(-) delete mode 100644 kernel/watchdog_hld.c create mode 100644 kernel/watchdog_perf.c (limited to 'kernel') diff --git a/kernel/Makefile b/kernel/Makefile index b69c95315480..7eb72033143c 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -91,7 +91,7 @@ obj-$(CONFIG_FAIL_FUNCTION) += fail_function.o obj-$(CONFIG_KGDB) += debug/ obj-$(CONFIG_DETECT_HUNG_TASK) += hung_task.o obj-$(CONFIG_LOCKUP_DETECTOR) += watchdog.o -obj-$(CONFIG_HARDLOCKUP_DETECTOR_PERF) += watchdog_hld.o +obj-$(CONFIG_HARDLOCKUP_DETECTOR_PERF) += watchdog_perf.o obj-$(CONFIG_SECCOMP) += seccomp.o obj-$(CONFIG_RELAY) += relay.o obj-$(CONFIG_SYSCTL) += utsname_sysctl.o diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c deleted file mode 100644 index 2125b09e09d7..000000000000 --- a/kernel/watchdog_hld.c +++ /dev/null @@ -1,302 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 -/* - * Detect hard lockups on a system - * - * started by Don Zickus, Copyright (C) 2010 Red Hat, Inc. - * - * Note: Most of this code is borrowed heavily from the original softlockup - * detector, so thanks to Ingo for the initial implementation. - * Some chunks also taken from the old x86-specific nmi watchdog code, thanks - * to those contributors as well. - */ - -#define pr_fmt(fmt) "NMI watchdog: " fmt - -#include -#include -#include -#include - -#include -#include - -static DEFINE_PER_CPU(bool, hard_watchdog_warn); -static DEFINE_PER_CPU(bool, watchdog_nmi_touch); -static DEFINE_PER_CPU(struct perf_event *, watchdog_ev); -static DEFINE_PER_CPU(struct perf_event *, dead_event); -static struct cpumask dead_events_mask; - -static unsigned long hardlockup_allcpu_dumped; -static atomic_t watchdog_cpus = ATOMIC_INIT(0); - -notrace void arch_touch_nmi_watchdog(void) -{ - /* - * Using __raw here because some code paths have - * preemption enabled. If preemption is enabled - * then interrupts should be enabled too, in which - * case we shouldn't have to worry about the watchdog - * going off. - */ - raw_cpu_write(watchdog_nmi_touch, true); -} -EXPORT_SYMBOL(arch_touch_nmi_watchdog); - -#ifdef CONFIG_HARDLOCKUP_CHECK_TIMESTAMP -static DEFINE_PER_CPU(ktime_t, last_timestamp); -static DEFINE_PER_CPU(unsigned int, nmi_rearmed); -static ktime_t watchdog_hrtimer_sample_threshold __read_mostly; - -void watchdog_update_hrtimer_threshold(u64 period) -{ - /* - * The hrtimer runs with a period of (watchdog_threshold * 2) / 5 - * - * So it runs effectively with 2.5 times the rate of the NMI - * watchdog. That means the hrtimer should fire 2-3 times before - * the NMI watchdog expires. The NMI watchdog on x86 is based on - * unhalted CPU cycles, so if Turbo-Mode is enabled the CPU cycles - * might run way faster than expected and the NMI fires in a - * smaller period than the one deduced from the nominal CPU - * frequency. Depending on the Turbo-Mode factor this might be fast - * enough to get the NMI period smaller than the hrtimer watchdog - * period and trigger false positives. - * - * The sample threshold is used to check in the NMI handler whether - * the minimum time between two NMI samples has elapsed. That - * prevents false positives. - * - * Set this to 4/5 of the actual watchdog threshold period so the - * hrtimer is guaranteed to fire at least once within the real - * watchdog threshold. - */ - watchdog_hrtimer_sample_threshold = period * 2; -} - -static bool watchdog_check_timestamp(void) -{ - ktime_t delta, now = ktime_get_mono_fast_ns(); - - delta = now - __this_cpu_read(last_timestamp); - if (delta < watchdog_hrtimer_sample_threshold) { - /* - * If ktime is jiffies based, a stalled timer would prevent - * jiffies from being incremented and the filter would look - * at a stale timestamp and never trigger. - */ - if (__this_cpu_inc_return(nmi_rearmed) < 10) - return false; - } - __this_cpu_write(nmi_rearmed, 0); - __this_cpu_write(last_timestamp, now); - return true; -} -#else -static inline bool watchdog_check_timestamp(void) -{ - return true; -} -#endif - -static struct perf_event_attr wd_hw_attr = { - .type = PERF_TYPE_HARDWARE, - .config = PERF_COUNT_HW_CPU_CYCLES, - .size = sizeof(struct perf_event_attr), - .pinned = 1, - .disabled = 1, -}; - -/* Callback function for perf event subsystem */ -static void watchdog_overflow_callback(struct perf_event *event, - struct perf_sample_data *data, - struct pt_regs *regs) -{ - /* Ensure the watchdog never gets throttled */ - event->hw.interrupts = 0; - - if (!watchdog_check_timestamp()) - return; - - if (__this_cpu_read(watchdog_nmi_touch) == true) { - __this_cpu_write(watchdog_nmi_touch, false); - return; - } - - /* check for a hardlockup - * This is done by making sure our timer interrupt - * is incrementing. The timer interrupt should have - * fired multiple times before we overflow'd. If it hasn't - * then this is a good indication the cpu is stuck - */ - if (is_hardlockup()) { - int this_cpu = smp_processor_id(); - - /* only print hardlockups once */ - if (__this_cpu_read(hard_watchdog_warn) == true) - return; - - pr_emerg("Watchdog detected hard LOCKUP on cpu %d\n", - this_cpu); - print_modules(); - print_irqtrace_events(current); - if (regs) - show_regs(regs); - else - dump_stack(); - - /* - * Perform all-CPU dump only once to avoid multiple hardlockups - * generating interleaving traces - */ - if (sysctl_hardlockup_all_cpu_backtrace && - !test_and_set_bit(0, &hardlockup_allcpu_dumped)) - trigger_allbutself_cpu_backtrace(); - - if (hardlockup_panic) - nmi_panic(regs, "Hard LOCKUP"); - - __this_cpu_write(hard_watchdog_warn, true); - return; - } - - __this_cpu_write(hard_watchdog_warn, false); - return; -} - -static int hardlockup_detector_event_create(void) -{ - unsigned int cpu; - struct perf_event_attr *wd_attr; - struct perf_event *evt; - - /* - * Preemption is not disabled because memory will be allocated. - * Ensure CPU-locality by calling this in per-CPU kthread. - */ - WARN_ON(!is_percpu_thread()); - cpu = raw_smp_processor_id(); - wd_attr = &wd_hw_attr; - wd_attr->sample_period = hw_nmi_get_sample_period(watchdog_thresh); - - /* Try to register using hardware perf events */ - evt = perf_event_create_kernel_counter(wd_attr, cpu, NULL, - watchdog_overflow_callback, NULL); - if (IS_ERR(evt)) { - pr_debug("Perf event create on CPU %d failed with %ld\n", cpu, - PTR_ERR(evt)); - return PTR_ERR(evt); - } - this_cpu_write(watchdog_ev, evt); - return 0; -} - -/** - * hardlockup_detector_perf_enable - Enable the local event - */ -void hardlockup_detector_perf_enable(void) -{ - if (hardlockup_detector_event_create()) - return; - - /* use original value for check */ - if (!atomic_fetch_inc(&watchdog_cpus)) - pr_info("Enabled. Permanently consumes one hw-PMU counter.\n"); - - perf_event_enable(this_cpu_read(watchdog_ev)); -} - -/** - * hardlockup_detector_perf_disable - Disable the local event - */ -void hardlockup_detector_perf_disable(void) -{ - struct perf_event *event = this_cpu_read(watchdog_ev); - - if (event) { - perf_event_disable(event); - this_cpu_write(watchdog_ev, NULL); - this_cpu_write(dead_event, event); - cpumask_set_cpu(smp_processor_id(), &dead_events_mask); - atomic_dec(&watchdog_cpus); - } -} - -/** - * hardlockup_detector_perf_cleanup - Cleanup disabled events and destroy them - * - * Called from lockup_detector_cleanup(). Serialized by the caller. - */ -void hardlockup_detector_perf_cleanup(void) -{ - int cpu; - - for_each_cpu(cpu, &dead_events_mask) { - struct perf_event *event = per_cpu(dead_event, cpu); - - /* - * Required because for_each_cpu() reports unconditionally - * CPU0 as set on UP kernels. Sigh. - */ - if (event) - perf_event_release_kernel(event); - per_cpu(dead_event, cpu) = NULL; - } - cpumask_clear(&dead_events_mask); -} - -/** - * hardlockup_detector_perf_stop - Globally stop watchdog events - * - * Special interface for x86 to handle the perf HT bug. - */ -void __init hardlockup_detector_perf_stop(void) -{ - int cpu; - - lockdep_assert_cpus_held(); - - for_each_online_cpu(cpu) { - struct perf_event *event = per_cpu(watchdog_ev, cpu); - - if (event) - perf_event_disable(event); - } -} - -/** - * hardlockup_detector_perf_restart - Globally restart watchdog events - * - * Special interface for x86 to handle the perf HT bug. - */ -void __init hardlockup_detector_perf_restart(void) -{ - int cpu; - - lockdep_assert_cpus_held(); - - if (!(watchdog_enabled & NMI_WATCHDOG_ENABLED)) - return; - - for_each_online_cpu(cpu) { - struct perf_event *event = per_cpu(watchdog_ev, cpu); - - if (event) - perf_event_enable(event); - } -} - -/** - * hardlockup_detector_perf_init - Probe whether NMI event is available at all - */ -int __init hardlockup_detector_perf_init(void) -{ - int ret = hardlockup_detector_event_create(); - - if (ret) { - pr_info("Perf NMI watchdog permanently disabled\n"); - } else { - perf_event_release_kernel(this_cpu_read(watchdog_ev)); - this_cpu_write(watchdog_ev, NULL); - } - return ret; -} diff --git a/kernel/watchdog_perf.c b/kernel/watchdog_perf.c new file mode 100644 index 000000000000..8b8015758ea5 --- /dev/null +++ b/kernel/watchdog_perf.c @@ -0,0 +1,302 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Detect hard lockups on a system using perf + * + * started by Don Zickus, Copyright (C) 2010 Red Hat, Inc. + * + * Note: Most of this code is borrowed heavily from the original softlockup + * detector, so thanks to Ingo for the initial implementation. + * Some chunks also taken from the old x86-specific nmi watchdog code, thanks + * to those contributors as well. + */ + +#define pr_fmt(fmt) "NMI watchdog: " fmt + +#include +#include +#include +#include + +#include +#include + +static DEFINE_PER_CPU(bool, hard_watchdog_warn); +static DEFINE_PER_CPU(bool, watchdog_nmi_touch); +static DEFINE_PER_CPU(struct perf_event *, watchdog_ev); +static DEFINE_PER_CPU(struct perf_event *, dead_event); +static struct cpumask dead_events_mask; + +static unsigned long hardlockup_allcpu_dumped; +static atomic_t watchdog_cpus = ATOMIC_INIT(0); + +notrace void arch_touch_nmi_watchdog(void) +{ + /* + * Using __raw here because some code paths have + * preemption enabled. If preemption is enabled + * then interrupts should be enabled too, in which + * case we shouldn't have to worry about the watchdog + * going off. + */ + raw_cpu_write(watchdog_nmi_touch, true); +} +EXPORT_SYMBOL(arch_touch_nmi_watchdog); + +#ifdef CONFIG_HARDLOCKUP_CHECK_TIMESTAMP +static DEFINE_PER_CPU(ktime_t, last_timestamp); +static DEFINE_PER_CPU(unsigned int, nmi_rearmed); +static ktime_t watchdog_hrtimer_sample_threshold __read_mostly; + +void watchdog_update_hrtimer_threshold(u64 period) +{ + /* + * The hrtimer runs with a period of (watchdog_threshold * 2) / 5 + * + * So it runs effectively with 2.5 times the rate of the NMI + * watchdog. That means the hrtimer should fire 2-3 times before + * the NMI watchdog expires. The NMI watchdog on x86 is based on + * unhalted CPU cycles, so if Turbo-Mode is enabled the CPU cycles + * might run way faster than expected and the NMI fires in a + * smaller period than the one deduced from the nominal CPU + * frequency. Depending on the Turbo-Mode factor this might be fast + * enough to get the NMI period smaller than the hrtimer watchdog + * period and trigger false positives. + * + * The sample threshold is used to check in the NMI handler whether + * the minimum time between two NMI samples has elapsed. That + * prevents false positives. + * + * Set this to 4/5 of the actual watchdog threshold period so the + * hrtimer is guaranteed to fire at least once within the real + * watchdog threshold. + */ + watchdog_hrtimer_sample_threshold = period * 2; +} + +static bool watchdog_check_timestamp(void) +{ + ktime_t delta, now = ktime_get_mono_fast_ns(); + + delta = now - __this_cpu_read(last_timestamp); + if (delta < watchdog_hrtimer_sample_threshold) { + /* + * If ktime is jiffies based, a stalled timer would prevent + * jiffies from being incremented and the filter would look + * at a stale timestamp and never trigger. + */ + if (__this_cpu_inc_return(nmi_rearmed) < 10) + return false; + } + __this_cpu_write(nmi_rearmed, 0); + __this_cpu_write(last_timestamp, now); + return true; +} +#else +static inline bool watchdog_check_timestamp(void) +{ + return true; +} +#endif + +static struct perf_event_attr wd_hw_attr = { + .type = PERF_TYPE_HARDWARE, + .config = PERF_COUNT_HW_CPU_CYCLES, + .size = sizeof(struct perf_event_attr), + .pinned = 1, + .disabled = 1, +}; + +/* Callback function for perf event subsystem */ +static void watchdog_overflow_callback(struct perf_event *event, + struct perf_sample_data *data, + struct pt_regs *regs) +{ + /* Ensure the watchdog never gets throttled */ + event->hw.interrupts = 0; + + if (!watchdog_check_timestamp()) + return; + + if (__this_cpu_read(watchdog_nmi_touch) == true) { + __this_cpu_write(watchdog_nmi_touch, false); + return; + } + + /* check for a hardlockup + * This is done by making sure our timer interrupt + * is incrementing. The timer interrupt should have + * fired multiple times before we overflow'd. If it hasn't + * then this is a good indication the cpu is stuck + */ + if (is_hardlockup()) { + int this_cpu = smp_processor_id(); + + /* only print hardlockups once */ + if (__this_cpu_read(hard_watchdog_warn) == true) + return; + + pr_emerg("Watchdog detected hard LOCKUP on cpu %d\n", + this_cpu); + print_modules(); + print_irqtrace_events(current); + if (regs) + show_regs(regs); + else + dump_stack(); + + /* + * Perform all-CPU dump only once to avoid multiple hardlockups + * generating interleaving traces + */ + if (sysctl_hardlockup_all_cpu_backtrace && + !test_and_set_bit(0, &hardlockup_allcpu_dumped)) + trigger_allbutself_cpu_backtrace(); + + if (hardlockup_panic) + nmi_panic(regs, "Hard LOCKUP"); + + __this_cpu_write(hard_watchdog_warn, true); + return; + } + + __this_cpu_write(hard_watchdog_warn, false); + return; +} + +static int hardlockup_detector_event_create(void) +{ + unsigned int cpu; + struct perf_event_attr *wd_attr; + struct perf_event *evt; + + /* + * Preemption is not disabled because memory will be allocated. + * Ensure CPU-locality by calling this in per-CPU kthread. + */ + WARN_ON(!is_percpu_thread()); + cpu = raw_smp_processor_id(); + wd_attr = &wd_hw_attr; + wd_attr->sample_period = hw_nmi_get_sample_period(watchdog_thresh); + + /* Try to register using hardware perf events */ + evt = perf_event_create_kernel_counter(wd_attr, cpu, NULL, + watchdog_overflow_callback, NULL); + if (IS_ERR(evt)) { + pr_debug("Perf event create on CPU %d failed with %ld\n", cpu, + PTR_ERR(evt)); + return PTR_ERR(evt); + } + this_cpu_write(watchdog_ev, evt); + return 0; +} + +/** + * hardlockup_detector_perf_enable - Enable the local event + */ +void hardlockup_detector_perf_enable(void) +{ + if (hardlockup_detector_event_create()) + return; + + /* use original value for check */ + if (!atomic_fetch_inc(&watchdog_cpus)) + pr_info("Enabled. Permanently consumes one hw-PMU counter.\n"); + + perf_event_enable(this_cpu_read(watchdog_ev)); +} + +/** + * hardlockup_detector_perf_disable - Disable the local event + */ +void hardlockup_detector_perf_disable(void) +{ + struct perf_event *event = this_cpu_read(watchdog_ev); + + if (event) { + perf_event_disable(event); + this_cpu_write(watchdog_ev, NULL); + this_cpu_write(dead_event, event); + cpumask_set_cpu(smp_processor_id(), &dead_events_mask); + atomic_dec(&watchdog_cpus); + } +} + +/** + * hardlockup_detector_perf_cleanup - Cleanup disabled events and destroy them + * + * Called from lockup_detector_cleanup(). Serialized by the caller. + */ +void hardlockup_detector_perf_cleanup(void) +{ + int cpu; + + for_each_cpu(cpu, &dead_events_mask) { + struct perf_event *event = per_cpu(dead_event, cpu); + + /* + * Required because for_each_cpu() reports unconditionally + * CPU0 as set on UP kernels. Sigh. + */ + if (event) + perf_event_release_kernel(event); + per_cpu(dead_event, cpu) = NULL; + } + cpumask_clear(&dead_events_mask); +} + +/** + * hardlockup_detector_perf_stop - Globally stop watchdog events + * + * Special interface for x86 to handle the perf HT bug. + */ +void __init hardlockup_detector_perf_stop(void) +{ + int cpu; + + lockdep_assert_cpus_held(); + + for_each_online_cpu(cpu) { + struct perf_event *event = per_cpu(watchdog_ev, cpu); + + if (event) + perf_event_disable(event); + } +} + +/** + * hardlockup_detector_perf_restart - Globally restart watchdog events + * + * Special interface for x86 to handle the perf HT bug. + */ +void __init hardlockup_detector_perf_restart(void) +{ + int cpu; + + lockdep_assert_cpus_held(); + + if (!(watchdog_enabled & NMI_WATCHDOG_ENABLED)) + return; + + for_each_online_cpu(cpu) { + struct perf_event *event = per_cpu(watchdog_ev, cpu); + + if (event) + perf_event_enable(event); + } +} + +/** + * hardlockup_detector_perf_init - Probe whether NMI event is available at all + */ +int __init hardlockup_detector_perf_init(void) +{ + int ret = hardlockup_detector_event_create(); + + if (ret) { + pr_info("Perf NMI watchdog permanently disabled\n"); + } else { + perf_event_release_kernel(this_cpu_read(watchdog_ev)); + this_cpu_write(watchdog_ev, NULL); + } + return ret; +} -- cgit v1.2.3-58-ga151 From 81972551df9d168a8183b786ff4de06008469c2e Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:32 -0700 Subject: watchdog/hardlockup: move perf hardlockup checking/panic to common watchdog.c The perf hardlockup detector works by looking at interrupt counts and seeing if they change from run to run. The interrupt counts are managed by the common watchdog code via its watchdog_timer_fn(). Currently the API between the perf detector and the common code is a function: is_hardlockup(). When the hard lockup detector sees that function return true then it handles printing out debug info and inducing a panic if necessary. Let's change the API a little bit in preparation for the buddy hardlockup detector. The buddy hardlockup detector wants to print nearly the same debug info and have nearly the same panic behavior. That means we want to move all that code to the common file. For now, the code in the common file will only be there if the perf hardlockup detector is enabled, but eventually it will be selected by a common config. Right now, this _just_ moves the code from the perf detector file to the common file and changes the names. It doesn't make the changes that the buddy hardlockup detector will need and doesn't do any style cleanups. A future patch will do cleanup to make it more obvious what changed. With the above, we no longer have any callers of is_hardlockup() outside of the "watchdog.c" file, so we can remove it from the header, make it static, and move it to the same "#ifdef" block as our new watchdog_hardlockup_check(). While doing this, it can be noted that even if no hardlockup detectors were configured the existing code used to still have the code for counting/checking "hrtimer_interrupts" even if the perf hardlockup detector wasn't configured. We didn't need to do that, so move all the "hrtimer_interrupts" counting to only be there if the perf hardlockup detector is configured as well. This change is expected to be a no-op. Link: https://lkml.kernel.org/r/20230519101840.v5.8.Id4133d3183e798122dc3b6205e7852601f289071@changeid Signed-off-by: Douglas Anderson Reviewed-by: Petr Mladek Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Lecopzer Chen Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- include/linux/nmi.h | 5 ++- kernel/watchdog.c | 93 +++++++++++++++++++++++++++++++++++++++----------- kernel/watchdog_perf.c | 42 +---------------------- 3 files changed, 78 insertions(+), 62 deletions(-) (limited to 'kernel') diff --git a/include/linux/nmi.h b/include/linux/nmi.h index dd75e361a11f..67921436974d 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -15,7 +15,6 @@ void lockup_detector_init(void); void lockup_detector_soft_poweroff(void); void lockup_detector_cleanup(void); -bool is_hardlockup(void); extern int watchdog_user_enabled; extern int nmi_watchdog_user_enabled; @@ -88,6 +87,10 @@ extern unsigned int hardlockup_panic; static inline void hardlockup_detector_disable(void) {} #endif +#if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) +void watchdog_hardlockup_check(struct pt_regs *regs); +#endif + #if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR) # define NMI_WATCHDOG_SYSCTL_PERM 0644 #else diff --git a/kernel/watchdog.c b/kernel/watchdog.c index c705a18b26bf..12ce37d76e7d 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -85,6 +85,78 @@ __setup("nmi_watchdog=", hardlockup_panic_setup); #endif /* CONFIG_HARDLOCKUP_DETECTOR */ +#if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) + +static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts); +static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved); +static DEFINE_PER_CPU(bool, hard_watchdog_warn); +static unsigned long hardlockup_allcpu_dumped; + +static bool is_hardlockup(void) +{ + unsigned long hrint = __this_cpu_read(hrtimer_interrupts); + + if (__this_cpu_read(hrtimer_interrupts_saved) == hrint) + return true; + + __this_cpu_write(hrtimer_interrupts_saved, hrint); + return false; +} + +static void watchdog_hardlockup_kick(void) +{ + __this_cpu_inc(hrtimer_interrupts); +} + +void watchdog_hardlockup_check(struct pt_regs *regs) +{ + /* check for a hardlockup + * This is done by making sure our timer interrupt + * is incrementing. The timer interrupt should have + * fired multiple times before we overflow'd. If it hasn't + * then this is a good indication the cpu is stuck + */ + if (is_hardlockup()) { + int this_cpu = smp_processor_id(); + + /* only print hardlockups once */ + if (__this_cpu_read(hard_watchdog_warn) == true) + return; + + pr_emerg("Watchdog detected hard LOCKUP on cpu %d\n", + this_cpu); + print_modules(); + print_irqtrace_events(current); + if (regs) + show_regs(regs); + else + dump_stack(); + + /* + * Perform all-CPU dump only once to avoid multiple hardlockups + * generating interleaving traces + */ + if (sysctl_hardlockup_all_cpu_backtrace && + !test_and_set_bit(0, &hardlockup_allcpu_dumped)) + trigger_allbutself_cpu_backtrace(); + + if (hardlockup_panic) + nmi_panic(regs, "Hard LOCKUP"); + + __this_cpu_write(hard_watchdog_warn, true); + return; + } + + __this_cpu_write(hard_watchdog_warn, false); + return; +} + +#else /* CONFIG_HARDLOCKUP_DETECTOR_PERF */ + +static inline void watchdog_hardlockup_kick(void) { } + +#endif /* !CONFIG_HARDLOCKUP_DETECTOR_PERF */ + /* * These functions can be overridden if an architecture implements its * own hardlockup detector. @@ -176,8 +248,6 @@ static DEFINE_PER_CPU(unsigned long, watchdog_touch_ts); static DEFINE_PER_CPU(unsigned long, watchdog_report_ts); static DEFINE_PER_CPU(struct hrtimer, watchdog_hrtimer); static DEFINE_PER_CPU(bool, softlockup_touch_sync); -static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts); -static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved); static unsigned long soft_lockup_nmi_warn; static int __init nowatchdog_setup(char *str) @@ -312,22 +382,6 @@ static int is_softlockup(unsigned long touch_ts, } /* watchdog detector functions */ -bool is_hardlockup(void) -{ - unsigned long hrint = __this_cpu_read(hrtimer_interrupts); - - if (__this_cpu_read(hrtimer_interrupts_saved) == hrint) - return true; - - __this_cpu_write(hrtimer_interrupts_saved, hrint); - return false; -} - -static void watchdog_interrupt_count(void) -{ - __this_cpu_inc(hrtimer_interrupts); -} - static DEFINE_PER_CPU(struct completion, softlockup_completion); static DEFINE_PER_CPU(struct cpu_stop_work, softlockup_stop_work); @@ -358,8 +412,7 @@ static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer) if (!watchdog_enabled) return HRTIMER_NORESTART; - /* kick the hardlockup detector */ - watchdog_interrupt_count(); + watchdog_hardlockup_kick(); /* kick the softlockup detector */ if (completion_done(this_cpu_ptr(&softlockup_completion))) { diff --git a/kernel/watchdog_perf.c b/kernel/watchdog_perf.c index 8b8015758ea5..04415812d079 100644 --- a/kernel/watchdog_perf.c +++ b/kernel/watchdog_perf.c @@ -20,13 +20,11 @@ #include #include -static DEFINE_PER_CPU(bool, hard_watchdog_warn); static DEFINE_PER_CPU(bool, watchdog_nmi_touch); static DEFINE_PER_CPU(struct perf_event *, watchdog_ev); static DEFINE_PER_CPU(struct perf_event *, dead_event); static struct cpumask dead_events_mask; -static unsigned long hardlockup_allcpu_dumped; static atomic_t watchdog_cpus = ATOMIC_INIT(0); notrace void arch_touch_nmi_watchdog(void) @@ -122,45 +120,7 @@ static void watchdog_overflow_callback(struct perf_event *event, return; } - /* check for a hardlockup - * This is done by making sure our timer interrupt - * is incrementing. The timer interrupt should have - * fired multiple times before we overflow'd. If it hasn't - * then this is a good indication the cpu is stuck - */ - if (is_hardlockup()) { - int this_cpu = smp_processor_id(); - - /* only print hardlockups once */ - if (__this_cpu_read(hard_watchdog_warn) == true) - return; - - pr_emerg("Watchdog detected hard LOCKUP on cpu %d\n", - this_cpu); - print_modules(); - print_irqtrace_events(current); - if (regs) - show_regs(regs); - else - dump_stack(); - - /* - * Perform all-CPU dump only once to avoid multiple hardlockups - * generating interleaving traces - */ - if (sysctl_hardlockup_all_cpu_backtrace && - !test_and_set_bit(0, &hardlockup_allcpu_dumped)) - trigger_allbutself_cpu_backtrace(); - - if (hardlockup_panic) - nmi_panic(regs, "Hard LOCKUP"); - - __this_cpu_write(hard_watchdog_warn, true); - return; - } - - __this_cpu_write(hard_watchdog_warn, false); - return; + watchdog_hardlockup_check(regs); } static int hardlockup_detector_event_create(void) -- cgit v1.2.3-58-ga151 From 1610611aadc2241179e7090ba2f3fd0d763d9932 Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:33 -0700 Subject: watchdog/hardlockup: style changes to watchdog_hardlockup_check() / is_hardlockup() These are tiny style changes: - Add a blank line before a "return". - Renames two globals to use the "watchdog_hardlockup" prefix. - Store processor id in "unsigned int" rather than "int". - Minor comment rewording. - Use "else" rather than extra returns since it seemed more symmetric. Link: https://lkml.kernel.org/r/20230519101840.v5.9.I818492c326b632560b09f20d2608455ecf9d3650@changeid Reviewed-by: Petr Mladek Signed-off-by: Douglas Anderson Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Lecopzer Chen Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- kernel/watchdog.c | 32 +++++++++++++++----------------- 1 file changed, 15 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 12ce37d76e7d..169e5dffbc00 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -89,8 +89,8 @@ __setup("nmi_watchdog=", hardlockup_panic_setup); static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts); static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved); -static DEFINE_PER_CPU(bool, hard_watchdog_warn); -static unsigned long hardlockup_allcpu_dumped; +static DEFINE_PER_CPU(bool, watchdog_hardlockup_warned); +static unsigned long watchdog_hardlockup_all_cpu_dumped; static bool is_hardlockup(void) { @@ -100,6 +100,7 @@ static bool is_hardlockup(void) return true; __this_cpu_write(hrtimer_interrupts_saved, hrint); + return false; } @@ -110,21 +111,20 @@ static void watchdog_hardlockup_kick(void) void watchdog_hardlockup_check(struct pt_regs *regs) { - /* check for a hardlockup - * This is done by making sure our timer interrupt - * is incrementing. The timer interrupt should have - * fired multiple times before we overflow'd. If it hasn't + /* + * Check for a hardlockup by making sure the CPU's timer + * interrupt is incrementing. The timer interrupt should have + * fired multiple times before we overflow'd. If it hasn't * then this is a good indication the cpu is stuck */ if (is_hardlockup()) { - int this_cpu = smp_processor_id(); + unsigned int this_cpu = smp_processor_id(); - /* only print hardlockups once */ - if (__this_cpu_read(hard_watchdog_warn) == true) + /* Only print hardlockups once. */ + if (__this_cpu_read(watchdog_hardlockup_warned)) return; - pr_emerg("Watchdog detected hard LOCKUP on cpu %d\n", - this_cpu); + pr_emerg("Watchdog detected hard LOCKUP on cpu %d\n", this_cpu); print_modules(); print_irqtrace_events(current); if (regs) @@ -137,18 +137,16 @@ void watchdog_hardlockup_check(struct pt_regs *regs) * generating interleaving traces */ if (sysctl_hardlockup_all_cpu_backtrace && - !test_and_set_bit(0, &hardlockup_allcpu_dumped)) + !test_and_set_bit(0, &watchdog_hardlockup_all_cpu_dumped)) trigger_allbutself_cpu_backtrace(); if (hardlockup_panic) nmi_panic(regs, "Hard LOCKUP"); - __this_cpu_write(hard_watchdog_warn, true); - return; + __this_cpu_write(watchdog_hardlockup_warned, true); + } else { + __this_cpu_write(watchdog_hardlockup_warned, false); } - - __this_cpu_write(hard_watchdog_warn, false); - return; } #else /* CONFIG_HARDLOCKUP_DETECTOR_PERF */ -- cgit v1.2.3-58-ga151 From 77c12fc95980d100fdc49e88a5727c242d0dfedc Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:34 -0700 Subject: watchdog/hardlockup: add a "cpu" param to watchdog_hardlockup_check() In preparation for the buddy hardlockup detector where the CPU checking for lockup might not be the currently running CPU, add a "cpu" parameter to watchdog_hardlockup_check(). As part of this change, make hrtimer_interrupts an atomic_t since now the CPU incrementing the value and the CPU reading the value might be different. Technially this could also be done with just READ_ONCE and WRITE_ONCE, but atomic_t feels a little cleaner in this case. While hrtimer_interrupts is made atomic_t, we change hrtimer_interrupts_saved from "unsigned long" to "int". The "int" is needed to match the data type backing atomic_t for hrtimer_interrupts. Even if this changes us from 64-bits to 32-bits (which I don't think is true for most compilers), it doesn't really matter. All we ever do is increment it every few seconds and compare it to an old value so 32-bits is fine (even 16-bits would be). The "signed" vs "unsigned" also doesn't matter for simple equality comparisons. hrtimer_interrupts_saved is _not_ switched to atomic_t nor even accessed with READ_ONCE / WRITE_ONCE. The hrtimer_interrupts_saved is always consistently accessed with the same CPU. NOTE: with the upcoming "buddy" detector there is one special case. When a CPU goes offline/online then we can change which CPU is the one to consistently access a given instance of hrtimer_interrupts_saved. We still can't end up with a partially updated hrtimer_interrupts_saved, however, because we end up petting all affected CPUs to make sure the new and old CPU can't end up somehow read/write hrtimer_interrupts_saved at the same time. Link: https://lkml.kernel.org/r/20230519101840.v5.10.I3a7d4dd8c23ac30ee0b607d77feb6646b64825c0@changeid Signed-off-by: Douglas Anderson Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Lecopzer Chen Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Petr Mladek Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- include/linux/nmi.h | 2 +- kernel/watchdog.c | 52 +++++++++++++++++++++++++++++++------------------- kernel/watchdog_perf.c | 2 +- 3 files changed, 34 insertions(+), 22 deletions(-) (limited to 'kernel') diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 67921436974d..521bf2ee6eff 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -88,7 +88,7 @@ static inline void hardlockup_detector_disable(void) {} #endif #if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) -void watchdog_hardlockup_check(struct pt_regs *regs); +void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs); #endif #if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR) diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 169e5dffbc00..2552e224f76a 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -87,29 +87,34 @@ __setup("nmi_watchdog=", hardlockup_panic_setup); #if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) -static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts); -static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved); +static DEFINE_PER_CPU(atomic_t, hrtimer_interrupts); +static DEFINE_PER_CPU(int, hrtimer_interrupts_saved); static DEFINE_PER_CPU(bool, watchdog_hardlockup_warned); static unsigned long watchdog_hardlockup_all_cpu_dumped; -static bool is_hardlockup(void) +static bool is_hardlockup(unsigned int cpu) { - unsigned long hrint = __this_cpu_read(hrtimer_interrupts); + int hrint = atomic_read(&per_cpu(hrtimer_interrupts, cpu)); - if (__this_cpu_read(hrtimer_interrupts_saved) == hrint) + if (per_cpu(hrtimer_interrupts_saved, cpu) == hrint) return true; - __this_cpu_write(hrtimer_interrupts_saved, hrint); + /* + * NOTE: we don't need any fancy atomic_t or READ_ONCE/WRITE_ONCE + * for hrtimer_interrupts_saved. hrtimer_interrupts_saved is + * written/read by a single CPU. + */ + per_cpu(hrtimer_interrupts_saved, cpu) = hrint; return false; } static void watchdog_hardlockup_kick(void) { - __this_cpu_inc(hrtimer_interrupts); + atomic_inc(raw_cpu_ptr(&hrtimer_interrupts)); } -void watchdog_hardlockup_check(struct pt_regs *regs) +void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs) { /* * Check for a hardlockup by making sure the CPU's timer @@ -117,35 +122,42 @@ void watchdog_hardlockup_check(struct pt_regs *regs) * fired multiple times before we overflow'd. If it hasn't * then this is a good indication the cpu is stuck */ - if (is_hardlockup()) { + if (is_hardlockup(cpu)) { unsigned int this_cpu = smp_processor_id(); + struct cpumask backtrace_mask = *cpu_online_mask; /* Only print hardlockups once. */ - if (__this_cpu_read(watchdog_hardlockup_warned)) + if (per_cpu(watchdog_hardlockup_warned, cpu)) return; - pr_emerg("Watchdog detected hard LOCKUP on cpu %d\n", this_cpu); + pr_emerg("Watchdog detected hard LOCKUP on cpu %d\n", cpu); print_modules(); print_irqtrace_events(current); - if (regs) - show_regs(regs); - else - dump_stack(); + if (cpu == this_cpu) { + if (regs) + show_regs(regs); + else + dump_stack(); + cpumask_clear_cpu(cpu, &backtrace_mask); + } else { + if (trigger_single_cpu_backtrace(cpu)) + cpumask_clear_cpu(cpu, &backtrace_mask); + } /* - * Perform all-CPU dump only once to avoid multiple hardlockups - * generating interleaving traces + * Perform multi-CPU dump only once to avoid multiple + * hardlockups generating interleaving traces */ if (sysctl_hardlockup_all_cpu_backtrace && !test_and_set_bit(0, &watchdog_hardlockup_all_cpu_dumped)) - trigger_allbutself_cpu_backtrace(); + trigger_cpumask_backtrace(&backtrace_mask); if (hardlockup_panic) nmi_panic(regs, "Hard LOCKUP"); - __this_cpu_write(watchdog_hardlockup_warned, true); + per_cpu(watchdog_hardlockup_warned, cpu) = true; } else { - __this_cpu_write(watchdog_hardlockup_warned, false); + per_cpu(watchdog_hardlockup_warned, cpu) = false; } } diff --git a/kernel/watchdog_perf.c b/kernel/watchdog_perf.c index 04415812d079..4e60e8023515 100644 --- a/kernel/watchdog_perf.c +++ b/kernel/watchdog_perf.c @@ -120,7 +120,7 @@ static void watchdog_overflow_callback(struct perf_event *event, return; } - watchdog_hardlockup_check(regs); + watchdog_hardlockup_check(smp_processor_id(), regs); } static int hardlockup_detector_event_create(void) -- cgit v1.2.3-58-ga151 From ed92e1ef52224c7c9c15fba559448396b059c2ee Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:35 -0700 Subject: watchdog/hardlockup: move perf hardlockup watchdog petting to watchdog.c In preparation for the buddy hardlockup detector, which wants the same petting logic as the current perf hardlockup detector, move the code to watchdog.c. While doing this, rename the global variable to match others nearby. As part of this change we have to change the code to account for the fact that the CPU we're running on might be different than the one we're checking. Currently the code in watchdog.c is guarded by CONFIG_HARDLOCKUP_DETECTOR_PERF, which makes this change seem silly. However, a future patch will change this. Link: https://lkml.kernel.org/r/20230519101840.v5.11.I00dfd6386ee00da25bf26d140559a41339b53e57@changeid Signed-off-by: Douglas Anderson Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Lecopzer Chen Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Petr Mladek Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- include/linux/nmi.h | 5 +++-- kernel/watchdog.c | 19 +++++++++++++++++++ kernel/watchdog_perf.c | 19 ------------------- 3 files changed, 22 insertions(+), 21 deletions(-) (limited to 'kernel') diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 521bf2ee6eff..56fdc3de6894 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -88,7 +88,10 @@ static inline void hardlockup_detector_disable(void) {} #endif #if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) +void arch_touch_nmi_watchdog(void); void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs); +#elif !defined(CONFIG_HAVE_NMI_WATCHDOG) +static inline void arch_touch_nmi_watchdog(void) { } #endif #if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR) @@ -98,7 +101,6 @@ void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs); #endif #if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) -extern void arch_touch_nmi_watchdog(void); extern void hardlockup_detector_perf_stop(void); extern void hardlockup_detector_perf_restart(void); extern void hardlockup_detector_perf_disable(void); @@ -113,7 +115,6 @@ static inline void hardlockup_detector_perf_enable(void) { } static inline void hardlockup_detector_perf_cleanup(void) { } # if !defined(CONFIG_HAVE_NMI_WATCHDOG) static inline int hardlockup_detector_perf_init(void) { return -ENODEV; } -static inline void arch_touch_nmi_watchdog(void) {} # else static inline int hardlockup_detector_perf_init(void) { return 0; } # endif diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 2552e224f76a..64d7d2a0a7df 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -90,8 +90,22 @@ __setup("nmi_watchdog=", hardlockup_panic_setup); static DEFINE_PER_CPU(atomic_t, hrtimer_interrupts); static DEFINE_PER_CPU(int, hrtimer_interrupts_saved); static DEFINE_PER_CPU(bool, watchdog_hardlockup_warned); +static DEFINE_PER_CPU(bool, watchdog_hardlockup_touched); static unsigned long watchdog_hardlockup_all_cpu_dumped; +notrace void arch_touch_nmi_watchdog(void) +{ + /* + * Using __raw here because some code paths have + * preemption enabled. If preemption is enabled + * then interrupts should be enabled too, in which + * case we shouldn't have to worry about the watchdog + * going off. + */ + raw_cpu_write(watchdog_hardlockup_touched, true); +} +EXPORT_SYMBOL(arch_touch_nmi_watchdog); + static bool is_hardlockup(unsigned int cpu) { int hrint = atomic_read(&per_cpu(hrtimer_interrupts, cpu)); @@ -116,6 +130,11 @@ static void watchdog_hardlockup_kick(void) void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs) { + if (per_cpu(watchdog_hardlockup_touched, cpu)) { + per_cpu(watchdog_hardlockup_touched, cpu) = false; + return; + } + /* * Check for a hardlockup by making sure the CPU's timer * interrupt is incrementing. The timer interrupt should have diff --git a/kernel/watchdog_perf.c b/kernel/watchdog_perf.c index 4e60e8023515..547917ebd5d3 100644 --- a/kernel/watchdog_perf.c +++ b/kernel/watchdog_perf.c @@ -20,26 +20,12 @@ #include #include -static DEFINE_PER_CPU(bool, watchdog_nmi_touch); static DEFINE_PER_CPU(struct perf_event *, watchdog_ev); static DEFINE_PER_CPU(struct perf_event *, dead_event); static struct cpumask dead_events_mask; static atomic_t watchdog_cpus = ATOMIC_INIT(0); -notrace void arch_touch_nmi_watchdog(void) -{ - /* - * Using __raw here because some code paths have - * preemption enabled. If preemption is enabled - * then interrupts should be enabled too, in which - * case we shouldn't have to worry about the watchdog - * going off. - */ - raw_cpu_write(watchdog_nmi_touch, true); -} -EXPORT_SYMBOL(arch_touch_nmi_watchdog); - #ifdef CONFIG_HARDLOCKUP_CHECK_TIMESTAMP static DEFINE_PER_CPU(ktime_t, last_timestamp); static DEFINE_PER_CPU(unsigned int, nmi_rearmed); @@ -115,11 +101,6 @@ static void watchdog_overflow_callback(struct perf_event *event, if (!watchdog_check_timestamp()) return; - if (__this_cpu_read(watchdog_nmi_touch) == true) { - __this_cpu_write(watchdog_nmi_touch, false); - return; - } - watchdog_hardlockup_check(smp_processor_id(), regs); } -- cgit v1.2.3-58-ga151 From df95d3085caa5b99a60eb033d7ad6c2ff2b43dbf Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:36 -0700 Subject: watchdog/hardlockup: rename some "NMI watchdog" constants/function Do a search and replace of: - NMI_WATCHDOG_ENABLED => WATCHDOG_HARDLOCKUP_ENABLED - SOFT_WATCHDOG_ENABLED => WATCHDOG_SOFTOCKUP_ENABLED - watchdog_nmi_ => watchdog_hardlockup_ - nmi_watchdog_available => watchdog_hardlockup_available - nmi_watchdog_user_enabled => watchdog_hardlockup_user_enabled - soft_watchdog_user_enabled => watchdog_softlockup_user_enabled - NMI_WATCHDOG_DEFAULT => WATCHDOG_HARDLOCKUP_DEFAULT Then update a few comments near where names were changed. This is specifically to make it less confusing when we want to introduce the buddy hardlockup detector, which isn't using NMIs. As part of this, we sanitized a few names for consistency. [trix@redhat.com: make variables static] Link: https://lkml.kernel.org/r/20230525162822.1.I0fb41d138d158c9230573eaa37dc56afa2fb14ee@changeid Link: https://lkml.kernel.org/r/20230519101840.v5.12.I91f7277bab4bf8c0cb238732ed92e7ce7bbd71a6@changeid Signed-off-by: Douglas Anderson Signed-off-by: Tom Rix Reviewed-by: Tom Rix Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Lecopzer Chen Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Petr Mladek Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- arch/powerpc/include/asm/nmi.h | 4 +- arch/powerpc/kernel/watchdog.c | 12 ++-- arch/powerpc/platforms/pseries/mobility.c | 4 +- arch/sparc/kernel/nmi.c | 4 +- include/linux/nmi.h | 24 +++---- kernel/watchdog.c | 113 +++++++++++++++--------------- kernel/watchdog_perf.c | 2 +- 7 files changed, 81 insertions(+), 82 deletions(-) (limited to 'kernel') diff --git a/arch/powerpc/include/asm/nmi.h b/arch/powerpc/include/asm/nmi.h index c3c7adef74de..43bfd4de868f 100644 --- a/arch/powerpc/include/asm/nmi.h +++ b/arch/powerpc/include/asm/nmi.h @@ -5,10 +5,10 @@ #ifdef CONFIG_PPC_WATCHDOG extern void arch_touch_nmi_watchdog(void); long soft_nmi_interrupt(struct pt_regs *regs); -void watchdog_nmi_set_timeout_pct(u64 pct); +void watchdog_hardlockup_set_timeout_pct(u64 pct); #else static inline void arch_touch_nmi_watchdog(void) {} -static inline void watchdog_nmi_set_timeout_pct(u64 pct) {} +static inline void watchdog_hardlockup_set_timeout_pct(u64 pct) {} #endif #ifdef CONFIG_NMI_IPI diff --git a/arch/powerpc/kernel/watchdog.c b/arch/powerpc/kernel/watchdog.c index dbcc4a793f0b..edb2dd1f53eb 100644 --- a/arch/powerpc/kernel/watchdog.c +++ b/arch/powerpc/kernel/watchdog.c @@ -438,7 +438,7 @@ static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer) { int cpu = smp_processor_id(); - if (!(watchdog_enabled & NMI_WATCHDOG_ENABLED)) + if (!(watchdog_enabled & WATCHDOG_HARDLOCKUP_ENABLED)) return HRTIMER_NORESTART; if (!cpumask_test_cpu(cpu, &watchdog_cpumask)) @@ -479,7 +479,7 @@ static void start_watchdog(void *arg) return; } - if (!(watchdog_enabled & NMI_WATCHDOG_ENABLED)) + if (!(watchdog_enabled & WATCHDOG_HARDLOCKUP_ENABLED)) return; if (!cpumask_test_cpu(cpu, &watchdog_cpumask)) @@ -546,7 +546,7 @@ static void watchdog_calc_timeouts(void) wd_timer_period_ms = watchdog_thresh * 1000 * 2 / 5; } -void watchdog_nmi_stop(void) +void watchdog_hardlockup_stop(void) { int cpu; @@ -554,7 +554,7 @@ void watchdog_nmi_stop(void) stop_watchdog_on_cpu(cpu); } -void watchdog_nmi_start(void) +void watchdog_hardlockup_start(void) { int cpu; @@ -566,7 +566,7 @@ void watchdog_nmi_start(void) /* * Invoked from core watchdog init. */ -int __init watchdog_nmi_probe(void) +int __init watchdog_hardlockup_probe(void) { int err; @@ -582,7 +582,7 @@ int __init watchdog_nmi_probe(void) } #ifdef CONFIG_PPC_PSERIES -void watchdog_nmi_set_timeout_pct(u64 pct) +void watchdog_hardlockup_set_timeout_pct(u64 pct) { pr_info("Set the NMI watchdog timeout factor to %llu%%\n", pct); WRITE_ONCE(wd_timeout_pct, pct); diff --git a/arch/powerpc/platforms/pseries/mobility.c b/arch/powerpc/platforms/pseries/mobility.c index 6f30113b5468..cd632ba9ebff 100644 --- a/arch/powerpc/platforms/pseries/mobility.c +++ b/arch/powerpc/platforms/pseries/mobility.c @@ -750,7 +750,7 @@ static int pseries_migrate_partition(u64 handle) goto out; if (factor) - watchdog_nmi_set_timeout_pct(factor); + watchdog_hardlockup_set_timeout_pct(factor); ret = pseries_suspend(handle); if (ret == 0) { @@ -766,7 +766,7 @@ static int pseries_migrate_partition(u64 handle) pseries_cancel_migration(handle, ret); if (factor) - watchdog_nmi_set_timeout_pct(0); + watchdog_hardlockup_set_timeout_pct(0); out: vas_migration_handler(VAS_RESUME); diff --git a/arch/sparc/kernel/nmi.c b/arch/sparc/kernel/nmi.c index 5dcf31f7e81f..9d9e29b75c43 100644 --- a/arch/sparc/kernel/nmi.c +++ b/arch/sparc/kernel/nmi.c @@ -282,7 +282,7 @@ __setup("nmi_watchdog=", setup_nmi_watchdog); * sparc specific NMI watchdog enable function. * Enables watchdog if it is not enabled already. */ -void watchdog_nmi_enable(unsigned int cpu) +void watchdog_hardlockup_enable(unsigned int cpu) { if (atomic_read(&nmi_active) == -1) { pr_warn("NMI watchdog cannot be enabled or disabled\n"); @@ -303,7 +303,7 @@ void watchdog_nmi_enable(unsigned int cpu) * sparc specific NMI watchdog disable function. * Disables watchdog if it is not disabled already. */ -void watchdog_nmi_disable(unsigned int cpu) +void watchdog_hardlockup_disable(unsigned int cpu) { if (atomic_read(&nmi_active) == -1) pr_warn_once("NMI watchdog cannot be enabled or disabled\n"); diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 56fdc3de6894..17b2ae7103d9 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -17,8 +17,6 @@ void lockup_detector_soft_poweroff(void); void lockup_detector_cleanup(void); extern int watchdog_user_enabled; -extern int nmi_watchdog_user_enabled; -extern int soft_watchdog_user_enabled; extern int watchdog_thresh; extern unsigned long watchdog_enabled; @@ -68,17 +66,17 @@ static inline void reset_hung_task_detector(void) { } * 'watchdog_enabled' variable. Each lockup detector has its dedicated bit - * bit 0 for the hard lockup detector and bit 1 for the soft lockup detector. * - * 'watchdog_user_enabled', 'nmi_watchdog_user_enabled' and - * 'soft_watchdog_user_enabled' are variables that are only used as an + * 'watchdog_user_enabled', 'watchdog_hardlockup_user_enabled' and + * 'watchdog_softlockup_user_enabled' are variables that are only used as an * 'interface' between the parameters in /proc/sys/kernel and the internal * state bits in 'watchdog_enabled'. The 'watchdog_thresh' variable is * handled differently because its value is not boolean, and the lockup * detectors are 'suspended' while 'watchdog_thresh' is equal zero. */ -#define NMI_WATCHDOG_ENABLED_BIT 0 -#define SOFT_WATCHDOG_ENABLED_BIT 1 -#define NMI_WATCHDOG_ENABLED (1 << NMI_WATCHDOG_ENABLED_BIT) -#define SOFT_WATCHDOG_ENABLED (1 << SOFT_WATCHDOG_ENABLED_BIT) +#define WATCHDOG_HARDLOCKUP_ENABLED_BIT 0 +#define WATCHDOG_SOFTOCKUP_ENABLED_BIT 1 +#define WATCHDOG_HARDLOCKUP_ENABLED (1 << WATCHDOG_HARDLOCKUP_ENABLED_BIT) +#define WATCHDOG_SOFTOCKUP_ENABLED (1 << WATCHDOG_SOFTOCKUP_ENABLED_BIT) #if defined(CONFIG_HARDLOCKUP_DETECTOR) extern void hardlockup_detector_disable(void); @@ -120,11 +118,11 @@ static inline int hardlockup_detector_perf_init(void) { return 0; } # endif #endif -void watchdog_nmi_stop(void); -void watchdog_nmi_start(void); -int watchdog_nmi_probe(void); -void watchdog_nmi_enable(unsigned int cpu); -void watchdog_nmi_disable(unsigned int cpu); +void watchdog_hardlockup_stop(void); +void watchdog_hardlockup_start(void); +int watchdog_hardlockup_probe(void); +void watchdog_hardlockup_enable(unsigned int cpu); +void watchdog_hardlockup_disable(unsigned int cpu); void lockup_detector_reconfigure(void); diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 64d7d2a0a7df..c6790c7dc08d 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -30,17 +30,17 @@ static DEFINE_MUTEX(watchdog_mutex); #if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG) -# define NMI_WATCHDOG_DEFAULT 1 +# define WATCHDOG_HARDLOCKUP_DEFAULT 1 #else -# define NMI_WATCHDOG_DEFAULT 0 +# define WATCHDOG_HARDLOCKUP_DEFAULT 0 #endif unsigned long __read_mostly watchdog_enabled; int __read_mostly watchdog_user_enabled = 1; -int __read_mostly nmi_watchdog_user_enabled = NMI_WATCHDOG_DEFAULT; -int __read_mostly soft_watchdog_user_enabled = 1; +static int __read_mostly watchdog_hardlockup_user_enabled = WATCHDOG_HARDLOCKUP_DEFAULT; +static int __read_mostly watchdog_softlockup_user_enabled = 1; int __read_mostly watchdog_thresh = 10; -static int __read_mostly nmi_watchdog_available; +static int __read_mostly watchdog_hardlockup_available; struct cpumask watchdog_cpumask __read_mostly; unsigned long *watchdog_cpumask_bits = cpumask_bits(&watchdog_cpumask); @@ -66,7 +66,7 @@ unsigned int __read_mostly hardlockup_panic = */ void __init hardlockup_detector_disable(void) { - nmi_watchdog_user_enabled = 0; + watchdog_hardlockup_user_enabled = 0; } static int __init hardlockup_panic_setup(char *str) @@ -76,9 +76,9 @@ static int __init hardlockup_panic_setup(char *str) else if (!strncmp(str, "nopanic", 7)) hardlockup_panic = 0; else if (!strncmp(str, "0", 1)) - nmi_watchdog_user_enabled = 0; + watchdog_hardlockup_user_enabled = 0; else if (!strncmp(str, "1", 1)) - nmi_watchdog_user_enabled = 1; + watchdog_hardlockup_user_enabled = 1; return 1; } __setup("nmi_watchdog=", hardlockup_panic_setup); @@ -190,40 +190,40 @@ static inline void watchdog_hardlockup_kick(void) { } * These functions can be overridden if an architecture implements its * own hardlockup detector. * - * watchdog_nmi_enable/disable can be implemented to start and stop when + * watchdog_hardlockup_enable/disable can be implemented to start and stop when * softlockup watchdog start and stop. The arch must select the * SOFTLOCKUP_DETECTOR Kconfig. */ -void __weak watchdog_nmi_enable(unsigned int cpu) +void __weak watchdog_hardlockup_enable(unsigned int cpu) { hardlockup_detector_perf_enable(); } -void __weak watchdog_nmi_disable(unsigned int cpu) +void __weak watchdog_hardlockup_disable(unsigned int cpu) { hardlockup_detector_perf_disable(); } -/* Return 0, if a NMI watchdog is available. Error code otherwise */ -int __weak __init watchdog_nmi_probe(void) +/* Return 0, if a hardlockup watchdog is available. Error code otherwise */ +int __weak __init watchdog_hardlockup_probe(void) { return hardlockup_detector_perf_init(); } /** - * watchdog_nmi_stop - Stop the watchdog for reconfiguration + * watchdog_hardlockup_stop - Stop the watchdog for reconfiguration * * The reconfiguration steps are: - * watchdog_nmi_stop(); + * watchdog_hardlockup_stop(); * update_variables(); - * watchdog_nmi_start(); + * watchdog_hardlockup_start(); */ -void __weak watchdog_nmi_stop(void) { } +void __weak watchdog_hardlockup_stop(void) { } /** - * watchdog_nmi_start - Start the watchdog after reconfiguration + * watchdog_hardlockup_start - Start the watchdog after reconfiguration * - * Counterpart to watchdog_nmi_stop(). + * Counterpart to watchdog_hardlockup_stop(). * * The following variables have been updated in update_variables() and * contain the currently valid configuration: @@ -231,23 +231,23 @@ void __weak watchdog_nmi_stop(void) { } * - watchdog_thresh * - watchdog_cpumask */ -void __weak watchdog_nmi_start(void) { } +void __weak watchdog_hardlockup_start(void) { } /** * lockup_detector_update_enable - Update the sysctl enable bit * - * Caller needs to make sure that the NMI/perf watchdogs are off, so this - * can't race with watchdog_nmi_disable(). + * Caller needs to make sure that the hard watchdogs are off, so this + * can't race with watchdog_hardlockup_disable(). */ static void lockup_detector_update_enable(void) { watchdog_enabled = 0; if (!watchdog_user_enabled) return; - if (nmi_watchdog_available && nmi_watchdog_user_enabled) - watchdog_enabled |= NMI_WATCHDOG_ENABLED; - if (soft_watchdog_user_enabled) - watchdog_enabled |= SOFT_WATCHDOG_ENABLED; + if (watchdog_hardlockup_available && watchdog_hardlockup_user_enabled) + watchdog_enabled |= WATCHDOG_HARDLOCKUP_ENABLED; + if (watchdog_softlockup_user_enabled) + watchdog_enabled |= WATCHDOG_SOFTOCKUP_ENABLED; } #ifdef CONFIG_SOFTLOCKUP_DETECTOR @@ -288,7 +288,7 @@ __setup("nowatchdog", nowatchdog_setup); static int __init nosoftlockup_setup(char *str) { - soft_watchdog_user_enabled = 0; + watchdog_softlockup_user_enabled = 0; return 1; } __setup("nosoftlockup", nosoftlockup_setup); @@ -402,7 +402,7 @@ static int is_softlockup(unsigned long touch_ts, unsigned long period_ts, unsigned long now) { - if ((watchdog_enabled & SOFT_WATCHDOG_ENABLED) && watchdog_thresh){ + if ((watchdog_enabled & WATCHDOG_SOFTOCKUP_ENABLED) && watchdog_thresh) { /* Warn about unreasonable delays. */ if (time_after(now, period_ts + get_softlockup_thresh())) return now - touch_ts; @@ -537,7 +537,7 @@ static void watchdog_enable(unsigned int cpu) complete(done); /* - * Start the timer first to prevent the NMI watchdog triggering + * Start the timer first to prevent the hardlockup watchdog triggering * before the timer has a chance to fire. */ hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD); @@ -547,9 +547,9 @@ static void watchdog_enable(unsigned int cpu) /* Initialize timestamp */ update_touch_ts(); - /* Enable the perf event */ - if (watchdog_enabled & NMI_WATCHDOG_ENABLED) - watchdog_nmi_enable(cpu); + /* Enable the hardlockup detector */ + if (watchdog_enabled & WATCHDOG_HARDLOCKUP_ENABLED) + watchdog_hardlockup_enable(cpu); } static void watchdog_disable(unsigned int cpu) @@ -559,11 +559,11 @@ static void watchdog_disable(unsigned int cpu) WARN_ON_ONCE(cpu != smp_processor_id()); /* - * Disable the perf event first. That prevents that a large delay - * between disabling the timer and disabling the perf event causes - * the perf NMI to detect a false positive. + * Disable the hardlockup detector first. That prevents that a large + * delay between disabling the timer and disabling the hardlockup + * detector causes a false positive. */ - watchdog_nmi_disable(cpu); + watchdog_hardlockup_disable(cpu); hrtimer_cancel(hrtimer); wait_for_completion(this_cpu_ptr(&softlockup_completion)); } @@ -619,7 +619,7 @@ int lockup_detector_offline_cpu(unsigned int cpu) static void __lockup_detector_reconfigure(void) { cpus_read_lock(); - watchdog_nmi_stop(); + watchdog_hardlockup_stop(); softlockup_stop_all(); set_sample_period(); @@ -627,7 +627,7 @@ static void __lockup_detector_reconfigure(void) if (watchdog_enabled && watchdog_thresh) softlockup_start_all(); - watchdog_nmi_start(); + watchdog_hardlockup_start(); cpus_read_unlock(); /* * Must be called outside the cpus locked section to prevent @@ -668,9 +668,9 @@ static __init void lockup_detector_setup(void) static void __lockup_detector_reconfigure(void) { cpus_read_lock(); - watchdog_nmi_stop(); + watchdog_hardlockup_stop(); lockup_detector_update_enable(); - watchdog_nmi_start(); + watchdog_hardlockup_start(); cpus_read_unlock(); } void lockup_detector_reconfigure(void) @@ -725,14 +725,14 @@ static void proc_watchdog_update(void) /* * common function for watchdog, nmi_watchdog and soft_watchdog parameter * - * caller | table->data points to | 'which' - * -------------------|----------------------------|-------------------------- - * proc_watchdog | watchdog_user_enabled | NMI_WATCHDOG_ENABLED | - * | | SOFT_WATCHDOG_ENABLED - * -------------------|----------------------------|-------------------------- - * proc_nmi_watchdog | nmi_watchdog_user_enabled | NMI_WATCHDOG_ENABLED - * -------------------|----------------------------|-------------------------- - * proc_soft_watchdog | soft_watchdog_user_enabled | SOFT_WATCHDOG_ENABLED + * caller | table->data points to | 'which' + * -------------------|----------------------------------|------------------------------- + * proc_watchdog | watchdog_user_enabled | WATCHDOG_HARDLOCKUP_ENABLED | + * | | WATCHDOG_SOFTOCKUP_ENABLED + * -------------------|----------------------------------|------------------------------- + * proc_nmi_watchdog | watchdog_hardlockup_user_enabled | WATCHDOG_HARDLOCKUP_ENABLED + * -------------------|----------------------------------|------------------------------- + * proc_soft_watchdog | watchdog_softlockup_user_enabled | WATCHDOG_SOFTOCKUP_ENABLED */ static int proc_watchdog_common(int which, struct ctl_table *table, int write, void *buffer, size_t *lenp, loff_t *ppos) @@ -764,7 +764,8 @@ static int proc_watchdog_common(int which, struct ctl_table *table, int write, int proc_watchdog(struct ctl_table *table, int write, void *buffer, size_t *lenp, loff_t *ppos) { - return proc_watchdog_common(NMI_WATCHDOG_ENABLED|SOFT_WATCHDOG_ENABLED, + return proc_watchdog_common(WATCHDOG_HARDLOCKUP_ENABLED | + WATCHDOG_SOFTOCKUP_ENABLED, table, write, buffer, lenp, ppos); } @@ -774,9 +775,9 @@ int proc_watchdog(struct ctl_table *table, int write, int proc_nmi_watchdog(struct ctl_table *table, int write, void *buffer, size_t *lenp, loff_t *ppos) { - if (!nmi_watchdog_available && write) + if (!watchdog_hardlockup_available && write) return -ENOTSUPP; - return proc_watchdog_common(NMI_WATCHDOG_ENABLED, + return proc_watchdog_common(WATCHDOG_HARDLOCKUP_ENABLED, table, write, buffer, lenp, ppos); } @@ -786,7 +787,7 @@ int proc_nmi_watchdog(struct ctl_table *table, int write, int proc_soft_watchdog(struct ctl_table *table, int write, void *buffer, size_t *lenp, loff_t *ppos) { - return proc_watchdog_common(SOFT_WATCHDOG_ENABLED, + return proc_watchdog_common(WATCHDOG_SOFTOCKUP_ENABLED, table, write, buffer, lenp, ppos); } @@ -854,7 +855,7 @@ static struct ctl_table watchdog_sysctls[] = { }, { .procname = "nmi_watchdog", - .data = &nmi_watchdog_user_enabled, + .data = &watchdog_hardlockup_user_enabled, .maxlen = sizeof(int), .mode = NMI_WATCHDOG_SYSCTL_PERM, .proc_handler = proc_nmi_watchdog, @@ -871,7 +872,7 @@ static struct ctl_table watchdog_sysctls[] = { #ifdef CONFIG_SOFTLOCKUP_DETECTOR { .procname = "soft_watchdog", - .data = &soft_watchdog_user_enabled, + .data = &watchdog_softlockup_user_enabled, .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_soft_watchdog, @@ -940,8 +941,8 @@ void __init lockup_detector_init(void) cpumask_copy(&watchdog_cpumask, housekeeping_cpumask(HK_TYPE_TIMER)); - if (!watchdog_nmi_probe()) - nmi_watchdog_available = true; + if (!watchdog_hardlockup_probe()) + watchdog_hardlockup_available = true; lockup_detector_setup(); watchdog_sysctl_init(); } diff --git a/kernel/watchdog_perf.c b/kernel/watchdog_perf.c index 547917ebd5d3..9e6042a892b3 100644 --- a/kernel/watchdog_perf.c +++ b/kernel/watchdog_perf.c @@ -215,7 +215,7 @@ void __init hardlockup_detector_perf_restart(void) lockdep_assert_cpus_held(); - if (!(watchdog_enabled & NMI_WATCHDOG_ENABLED)) + if (!(watchdog_enabled & WATCHDOG_HARDLOCKUP_ENABLED)) return; for_each_online_cpu(cpu) { -- cgit v1.2.3-58-ga151 From d9b3629ade8ebffb0075e311409796a56bac8282 Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:37 -0700 Subject: watchdog/hardlockup: have the perf hardlockup use __weak functions more cleanly The fact that there watchdog_hardlockup_enable(), watchdog_hardlockup_disable(), and watchdog_hardlockup_probe() are declared __weak means that the configured hardlockup detector can define non-weak versions of those functions if it needs to. Instead of doing this, the perf hardlockup detector hooked itself into the default __weak implementation, which was a bit awkward. Clean this up. From comments, it looks as if the original design was done because the __weak function were expected to implemented by the architecture and not by the configured hardlockup detector. This got awkward when we tried to add the buddy lockup detector which was not arch-specific but wanted to hook into those same functions. This is not expected to have any functional impact. Link: https://lkml.kernel.org/r/20230519101840.v5.13.I847d9ec852449350997ba00401d2462a9cb4302b@changeid Signed-off-by: Douglas Anderson Reviewed-by: Petr Mladek Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Lecopzer Chen Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- include/linux/nmi.h | 10 ---------- kernel/watchdog.c | 30 ++++++++++++++++++------------ kernel/watchdog_perf.c | 20 ++++++++++++++------ 3 files changed, 32 insertions(+), 28 deletions(-) (limited to 'kernel') diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 17b2ae7103d9..e23b4dc01b4a 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -101,21 +101,11 @@ static inline void arch_touch_nmi_watchdog(void) { } #if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) extern void hardlockup_detector_perf_stop(void); extern void hardlockup_detector_perf_restart(void); -extern void hardlockup_detector_perf_disable(void); -extern void hardlockup_detector_perf_enable(void); extern void hardlockup_detector_perf_cleanup(void); -extern int hardlockup_detector_perf_init(void); #else static inline void hardlockup_detector_perf_stop(void) { } static inline void hardlockup_detector_perf_restart(void) { } -static inline void hardlockup_detector_perf_disable(void) { } -static inline void hardlockup_detector_perf_enable(void) { } static inline void hardlockup_detector_perf_cleanup(void) { } -# if !defined(CONFIG_HAVE_NMI_WATCHDOG) -static inline int hardlockup_detector_perf_init(void) { return -ENODEV; } -# else -static inline int hardlockup_detector_perf_init(void) { return 0; } -# endif #endif void watchdog_hardlockup_stop(void); diff --git a/kernel/watchdog.c b/kernel/watchdog.c index c6790c7dc08d..e67125f64719 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -187,27 +187,33 @@ static inline void watchdog_hardlockup_kick(void) { } #endif /* !CONFIG_HARDLOCKUP_DETECTOR_PERF */ /* - * These functions can be overridden if an architecture implements its - * own hardlockup detector. + * These functions can be overridden based on the configured hardlockdup detector. * * watchdog_hardlockup_enable/disable can be implemented to start and stop when - * softlockup watchdog start and stop. The arch must select the + * softlockup watchdog start and stop. The detector must select the * SOFTLOCKUP_DETECTOR Kconfig. */ -void __weak watchdog_hardlockup_enable(unsigned int cpu) -{ - hardlockup_detector_perf_enable(); -} +void __weak watchdog_hardlockup_enable(unsigned int cpu) { } -void __weak watchdog_hardlockup_disable(unsigned int cpu) -{ - hardlockup_detector_perf_disable(); -} +void __weak watchdog_hardlockup_disable(unsigned int cpu) { } /* Return 0, if a hardlockup watchdog is available. Error code otherwise */ int __weak __init watchdog_hardlockup_probe(void) { - return hardlockup_detector_perf_init(); + /* + * If CONFIG_HAVE_NMI_WATCHDOG is defined then an architecture + * is assumed to have the hard watchdog available and we return 0. + */ + if (IS_ENABLED(CONFIG_HAVE_NMI_WATCHDOG)) + return 0; + + /* + * Hardlockup detectors other than those using CONFIG_HAVE_NMI_WATCHDOG + * are required to implement a non-weak version of this probe function + * to tell whether they are available. If they don't override then + * we'll return -ENODEV. + */ + return -ENODEV; } /** diff --git a/kernel/watchdog_perf.c b/kernel/watchdog_perf.c index 9e6042a892b3..349fcd4d2abc 100644 --- a/kernel/watchdog_perf.c +++ b/kernel/watchdog_perf.c @@ -132,10 +132,14 @@ static int hardlockup_detector_event_create(void) } /** - * hardlockup_detector_perf_enable - Enable the local event + * watchdog_hardlockup_enable - Enable the local event + * + * @cpu: The CPU to enable hard lockup on. */ -void hardlockup_detector_perf_enable(void) +void watchdog_hardlockup_enable(unsigned int cpu) { + WARN_ON_ONCE(cpu != smp_processor_id()); + if (hardlockup_detector_event_create()) return; @@ -147,12 +151,16 @@ void hardlockup_detector_perf_enable(void) } /** - * hardlockup_detector_perf_disable - Disable the local event + * watchdog_hardlockup_disable - Disable the local event + * + * @cpu: The CPU to enable hard lockup on. */ -void hardlockup_detector_perf_disable(void) +void watchdog_hardlockup_disable(unsigned int cpu) { struct perf_event *event = this_cpu_read(watchdog_ev); + WARN_ON_ONCE(cpu != smp_processor_id()); + if (event) { perf_event_disable(event); this_cpu_write(watchdog_ev, NULL); @@ -227,9 +235,9 @@ void __init hardlockup_detector_perf_restart(void) } /** - * hardlockup_detector_perf_init - Probe whether NMI event is available at all + * watchdog_hardlockup_probe - Probe whether NMI event is available at all */ -int __init hardlockup_detector_perf_init(void) +int __init watchdog_hardlockup_probe(void) { int ret = hardlockup_detector_event_create(); -- cgit v1.2.3-58-ga151 From 1f423c905a6b43b493df1b259e6e6267e5624e62 Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:38 -0700 Subject: watchdog/hardlockup: detect hard lockups using secondary (buddy) CPUs Implement a hardlockup detector that doesn't doesn't need any extra arch-specific support code to detect lockups. Instead of using something arch-specific we will use the buddy system, where each CPU watches out for another one. Specifically, each CPU will use its softlockup hrtimer to check that the next CPU is processing hrtimer interrupts by verifying that a counter is increasing. NOTE: unlike the other hard lockup detectors, the buddy one can't easily show what's happening on the CPU that locked up just by doing a simple backtrace. It relies on some other mechanism in the system to get information about the locked up CPUs. This could be support for NMI backtraces like [1], it could be a mechanism for printing the PC of locked CPUs at panic time like [2] / [3], or it could be something else. Even though that means we still rely on arch-specific code, this arch-specific code seems to often be implemented even on architectures that don't have a hardlockup detector. This style of hardlockup detector originated in some downstream Android trees and has been rebased on / carried in ChromeOS trees for quite a long time for use on arm and arm64 boards. Historically on these boards we've leveraged mechanism [2] / [3] to get information about hung CPUs, but we could move to [1]. Although the original motivation for the buddy system was for use on systems without an arch-specific hardlockup detector, it can still be useful to use even on systems that _do_ have an arch-specific hardlockup detector. On x86, for instance, there is a 24-part patch series [4] in progress switching the arch-specific hard lockup detector from a scarce perf counter to a less-scarce hardware resource. Potentially the buddy system could be a simpler alternative to free up the perf counter but still get hard lockup detection. Overall, pros (+) and cons (-) of the buddy system compared to an arch-specific hardlockup detector (which might be implemented using perf): + The buddy system is usable on systems that don't have an arch-specific hardlockup detector, like arm32 and arm64 (though it's being worked on for arm64 [5]). + The buddy system may free up scarce hardware resources. + If a CPU totally goes out to lunch (can't process NMIs) the buddy system could still detect the problem (though it would be unlikely to be able to get a stack trace). + The buddy system uses the same timer function to pet the hardlockup detector on the running CPU as it uses to detect hardlockups on other CPUs. Compared to other hardlockup detectors, this means it generates fewer interrupts and thus is likely better able to let CPUs stay idle longer. - If all CPUs are hard locked up at the same time the buddy system can't detect it. - If we don't have SMP we can't use the buddy system. - The buddy system needs an arch-specific mechanism (possibly NMI backtrace) to get info about the locked up CPU. [1] https://lore.kernel.org/r/20230419225604.21204-1-dianders@chromium.org [2] https://issuetracker.google.com/172213129 [3] https://docs.kernel.org/trace/coresight/coresight-cpu-debug.html [4] https://lore.kernel.org/lkml/20230301234753.28582-1-ricardo.neri-calderon@linux.intel.com/ [5] https://lore.kernel.org/linux-arm-kernel/20220903093415.15850-1-lecopzer.chen@mediatek.com/ Link: https://lkml.kernel.org/r/20230519101840.v5.14.I6bf789d21d0c3d75d382e7e51a804a7a51315f2c@changeid Signed-off-by: Colin Cross Signed-off-by: Matthias Kaehlcke Signed-off-by: Guenter Roeck Signed-off-by: Tzung-Bi Shih Signed-off-by: Douglas Anderson Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Daniel Thompson Cc: "David S. Miller" Cc: Ian Rogers Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Petr Mladek Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Will Deacon Signed-off-by: Andrew Morton --- include/linux/nmi.h | 9 ++++- kernel/Makefile | 1 + kernel/watchdog.c | 29 +++++++++++---- kernel/watchdog_buddy.c | 93 +++++++++++++++++++++++++++++++++++++++++++++++++ lib/Kconfig.debug | 52 +++++++++++++++++++++++++-- 5 files changed, 173 insertions(+), 11 deletions(-) create mode 100644 kernel/watchdog_buddy.c (limited to 'kernel') diff --git a/include/linux/nmi.h b/include/linux/nmi.h index e23b4dc01b4a..1cdadc6a6cfd 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -85,8 +85,9 @@ extern unsigned int hardlockup_panic; static inline void hardlockup_detector_disable(void) {} #endif -#if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) +#if defined(CONFIG_HARDLOCKUP_DETECTOR_COUNTS_HRTIMER) void arch_touch_nmi_watchdog(void); +void watchdog_hardlockup_touch_cpu(unsigned int cpu); void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs); #elif !defined(CONFIG_HAVE_NMI_WATCHDOG) static inline void arch_touch_nmi_watchdog(void) { } @@ -116,6 +117,12 @@ void watchdog_hardlockup_disable(unsigned int cpu); void lockup_detector_reconfigure(void); +#ifdef CONFIG_HARDLOCKUP_DETECTOR_BUDDY +void watchdog_buddy_check_hardlockup(unsigned long hrtimer_interrupts); +#else +static inline void watchdog_buddy_check_hardlockup(unsigned long hrtimer_interrupts) {} +#endif + /** * touch_nmi_watchdog - manually reset the hardlockup watchdog timeout. * diff --git a/kernel/Makefile b/kernel/Makefile index 7eb72033143c..f9e3fd9195d9 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -91,6 +91,7 @@ obj-$(CONFIG_FAIL_FUNCTION) += fail_function.o obj-$(CONFIG_KGDB) += debug/ obj-$(CONFIG_DETECT_HUNG_TASK) += hung_task.o obj-$(CONFIG_LOCKUP_DETECTOR) += watchdog.o +obj-$(CONFIG_HARDLOCKUP_DETECTOR_BUDDY) += watchdog_buddy.o obj-$(CONFIG_HARDLOCKUP_DETECTOR_PERF) += watchdog_perf.o obj-$(CONFIG_SECCOMP) += seccomp.o obj-$(CONFIG_RELAY) += relay.o diff --git a/kernel/watchdog.c b/kernel/watchdog.c index e67125f64719..10947c835079 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -85,7 +85,7 @@ __setup("nmi_watchdog=", hardlockup_panic_setup); #endif /* CONFIG_HARDLOCKUP_DETECTOR */ -#if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) +#if defined(CONFIG_HARDLOCKUP_DETECTOR_COUNTS_HRTIMER) static DEFINE_PER_CPU(atomic_t, hrtimer_interrupts); static DEFINE_PER_CPU(int, hrtimer_interrupts_saved); @@ -106,6 +106,14 @@ notrace void arch_touch_nmi_watchdog(void) } EXPORT_SYMBOL(arch_touch_nmi_watchdog); +void watchdog_hardlockup_touch_cpu(unsigned int cpu) +{ + per_cpu(watchdog_hardlockup_touched, cpu) = true; + + /* Match with smp_rmb() in watchdog_hardlockup_check() */ + smp_wmb(); +} + static bool is_hardlockup(unsigned int cpu) { int hrint = atomic_read(&per_cpu(hrtimer_interrupts, cpu)); @@ -123,13 +131,16 @@ static bool is_hardlockup(unsigned int cpu) return false; } -static void watchdog_hardlockup_kick(void) +static unsigned long watchdog_hardlockup_kick(void) { - atomic_inc(raw_cpu_ptr(&hrtimer_interrupts)); + return atomic_inc_return(raw_cpu_ptr(&hrtimer_interrupts)); } void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs) { + /* Match with smp_wmb() in watchdog_hardlockup_touch_cpu() */ + smp_rmb(); + if (per_cpu(watchdog_hardlockup_touched, cpu)) { per_cpu(watchdog_hardlockup_touched, cpu) = false; return; @@ -180,11 +191,11 @@ void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs) } } -#else /* CONFIG_HARDLOCKUP_DETECTOR_PERF */ +#else /* CONFIG_HARDLOCKUP_DETECTOR_COUNTS_HRTIMER */ -static inline void watchdog_hardlockup_kick(void) { } +static inline unsigned long watchdog_hardlockup_kick(void) { return 0; } -#endif /* !CONFIG_HARDLOCKUP_DETECTOR_PERF */ +#endif /* !CONFIG_HARDLOCKUP_DETECTOR_COUNTS_HRTIMER */ /* * These functions can be overridden based on the configured hardlockdup detector. @@ -443,11 +454,15 @@ static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer) struct pt_regs *regs = get_irq_regs(); int duration; int softlockup_all_cpu_backtrace = sysctl_softlockup_all_cpu_backtrace; + unsigned long hrtimer_interrupts; if (!watchdog_enabled) return HRTIMER_NORESTART; - watchdog_hardlockup_kick(); + hrtimer_interrupts = watchdog_hardlockup_kick(); + + /* test for hardlockups */ + watchdog_buddy_check_hardlockup(hrtimer_interrupts); /* kick the softlockup detector */ if (completion_done(this_cpu_ptr(&softlockup_completion))) { diff --git a/kernel/watchdog_buddy.c b/kernel/watchdog_buddy.c new file mode 100644 index 000000000000..fee45af2e5bd --- /dev/null +++ b/kernel/watchdog_buddy.c @@ -0,0 +1,93 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include +#include +#include +#include +#include + +static cpumask_t __read_mostly watchdog_cpus; + +static unsigned int watchdog_next_cpu(unsigned int cpu) +{ + cpumask_t cpus = watchdog_cpus; + unsigned int next_cpu; + + next_cpu = cpumask_next(cpu, &cpus); + if (next_cpu >= nr_cpu_ids) + next_cpu = cpumask_first(&cpus); + + if (next_cpu == cpu) + return nr_cpu_ids; + + return next_cpu; +} + +int __init watchdog_hardlockup_probe(void) +{ + return 0; +} + +void watchdog_hardlockup_enable(unsigned int cpu) +{ + unsigned int next_cpu; + + /* + * The new CPU will be marked online before the hrtimer interrupt + * gets a chance to run on it. If another CPU tests for a + * hardlockup on the new CPU before it has run its the hrtimer + * interrupt, it will get a false positive. Touch the watchdog on + * the new CPU to delay the check for at least 3 sampling periods + * to guarantee one hrtimer has run on the new CPU. + */ + watchdog_hardlockup_touch_cpu(cpu); + + /* + * We are going to check the next CPU. Our watchdog_hrtimer + * need not be zero if the CPU has already been online earlier. + * Touch the watchdog on the next CPU to avoid false positive + * if we try to check it in less then 3 interrupts. + */ + next_cpu = watchdog_next_cpu(cpu); + if (next_cpu < nr_cpu_ids) + watchdog_hardlockup_touch_cpu(next_cpu); + + cpumask_set_cpu(cpu, &watchdog_cpus); +} + +void watchdog_hardlockup_disable(unsigned int cpu) +{ + unsigned int next_cpu = watchdog_next_cpu(cpu); + + /* + * Offlining this CPU will cause the CPU before this one to start + * checking the one after this one. If this CPU just finished checking + * the next CPU and updating hrtimer_interrupts_saved, and then the + * previous CPU checks it within one sample period, it will trigger a + * false positive. Touch the watchdog on the next CPU to prevent it. + */ + if (next_cpu < nr_cpu_ids) + watchdog_hardlockup_touch_cpu(next_cpu); + + cpumask_clear_cpu(cpu, &watchdog_cpus); +} + +void watchdog_buddy_check_hardlockup(unsigned long hrtimer_interrupts) +{ + unsigned int next_cpu; + + /* + * Test for hardlockups every 3 samples. The sample period is + * watchdog_thresh * 2 / 5, so 3 samples gets us back to slightly over + * watchdog_thresh (over by 20%). + */ + if (hrtimer_interrupts % 3 != 0) + return; + + /* check for a hardlockup on the next CPU */ + next_cpu = watchdog_next_cpu(smp_processor_id()); + if (next_cpu >= nr_cpu_ids) + return; + + watchdog_hardlockup_check(next_cpu, NULL); +} diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index ce51d4dc6803..abcad0513a32 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -1035,10 +1035,55 @@ config BOOTPARAM_SOFTLOCKUP_PANIC Say N if unsure. -config HARDLOCKUP_DETECTOR_PERF +# Both the "perf" and "buddy" hardlockup detectors count hrtimer +# interrupts. This config enables functions managing this common code. +config HARDLOCKUP_DETECTOR_COUNTS_HRTIMER bool select SOFTLOCKUP_DETECTOR +config HARDLOCKUP_DETECTOR_PERF + bool + depends on HAVE_HARDLOCKUP_DETECTOR_PERF + select HARDLOCKUP_DETECTOR_COUNTS_HRTIMER + +config HARDLOCKUP_DETECTOR_BUDDY + bool + depends on SMP + select HARDLOCKUP_DETECTOR_COUNTS_HRTIMER + +# For hardlockup detectors you can have one directly provided by the arch +# or use a "non-arch" one. If you're using a "non-arch" one that is +# further divided the perf hardlockup detector (which, confusingly, needs +# arch-provided perf support) and the buddy hardlockup detector (which just +# needs SMP). In either case, using the "non-arch" code conflicts with +# the NMI watchdog code (which is sometimes used directly and sometimes used +# by the arch-provided hardlockup detector). +config HAVE_HARDLOCKUP_DETECTOR_NON_ARCH + bool + depends on (HAVE_HARDLOCKUP_DETECTOR_PERF || SMP) && !HAVE_NMI_WATCHDOG + default y + +config HARDLOCKUP_DETECTOR_PREFER_BUDDY + bool "Prefer the buddy CPU hardlockup detector" + depends on HAVE_HARDLOCKUP_DETECTOR_NON_ARCH && HAVE_HARDLOCKUP_DETECTOR_PERF && SMP + help + Say Y here to prefer the buddy hardlockup detector over the perf one. + + With the buddy detector, each CPU uses its softlockup hrtimer + to check that the next CPU is processing hrtimer interrupts by + verifying that a counter is increasing. + + This hardlockup detector is useful on systems that don't have + an arch-specific hardlockup detector or if resources needed + for the hardlockup detector are better used for other things. + +# This will select the appropriate non-arch hardlockdup detector +config HARDLOCKUP_DETECTOR_NON_ARCH + bool + depends on HAVE_HARDLOCKUP_DETECTOR_NON_ARCH + select HARDLOCKUP_DETECTOR_BUDDY if !HAVE_HARDLOCKUP_DETECTOR_PERF || HARDLOCKUP_DETECTOR_PREFER_BUDDY + select HARDLOCKUP_DETECTOR_PERF if HAVE_HARDLOCKUP_DETECTOR_PERF && !HARDLOCKUP_DETECTOR_PREFER_BUDDY + # # Enables a timestamp based low pass filter to compensate for perf based # hard lockup detection which runs too fast due to turbo modes. @@ -1053,9 +1098,10 @@ config HARDLOCKUP_CHECK_TIMESTAMP config HARDLOCKUP_DETECTOR bool "Detect Hard Lockups" depends on DEBUG_KERNEL && !S390 - depends on HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_ARCH + depends on HAVE_HARDLOCKUP_DETECTOR_NON_ARCH || HAVE_HARDLOCKUP_DETECTOR_ARCH select LOCKUP_DETECTOR - select HARDLOCKUP_DETECTOR_PERF if HAVE_HARDLOCKUP_DETECTOR_PERF + select HARDLOCKUP_DETECTOR_NON_ARCH if HAVE_HARDLOCKUP_DETECTOR_NON_ARCH + help Say Y here to enable the kernel to act as a watchdog to detect hard lockups. -- cgit v1.2.3-58-ga151 From b17aa959330e8058452297049a0056ba4b9c72e8 Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 19 May 2023 10:18:39 -0700 Subject: watchdog/perf: add a weak function for an arch to detect if perf can use NMIs On arm64, NMI support needs to be detected at runtime. Add a weak function to the perf hardlockup detector so that an architecture can implement it to detect whether NMIs are available. Link: https://lkml.kernel.org/r/20230519101840.v5.15.Ic55cb6f90ef5967d8aaa2b503a4e67c753f64d3a@changeid Signed-off-by: Douglas Anderson Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Lecopzer Chen Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Petr Mladek Cc: Pingfan Liu Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- include/linux/nmi.h | 1 + kernel/watchdog_perf.c | 12 +++++++++++- 2 files changed, 12 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 1cdadc6a6cfd..28e65fd1de13 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -208,6 +208,7 @@ static inline bool trigger_single_cpu_backtrace(int cpu) #ifdef CONFIG_HARDLOCKUP_DETECTOR_PERF u64 hw_nmi_get_sample_period(int watchdog_thresh); +bool arch_perf_nmi_is_available(void); #endif #if defined(CONFIG_HARDLOCKUP_CHECK_TIMESTAMP) && \ diff --git a/kernel/watchdog_perf.c b/kernel/watchdog_perf.c index 349fcd4d2abc..8ea00c4a24b2 100644 --- a/kernel/watchdog_perf.c +++ b/kernel/watchdog_perf.c @@ -234,12 +234,22 @@ void __init hardlockup_detector_perf_restart(void) } } +bool __weak __init arch_perf_nmi_is_available(void) +{ + return true; +} + /** * watchdog_hardlockup_probe - Probe whether NMI event is available at all */ int __init watchdog_hardlockup_probe(void) { - int ret = hardlockup_detector_event_create(); + int ret; + + if (!arch_perf_nmi_is_available()) + return -ENODEV; + + ret = hardlockup_detector_event_create(); if (ret) { pr_info("Perf NMI watchdog permanently disabled\n"); -- cgit v1.2.3-58-ga151 From 930d8f8dbab97cb05dba30e67a2dfa0c6dbf4bc7 Mon Sep 17 00:00:00 2001 From: Lecopzer Chen Date: Fri, 19 May 2023 10:18:40 -0700 Subject: watchdog/perf: adapt the watchdog_perf interface for async model When lockup_detector_init()->watchdog_hardlockup_probe(), PMU may be not ready yet. E.g. on arm64, PMU is not ready until device_initcall(armv8_pmu_driver_init). And it is deeply integrated with the driver model and cpuhp. Hence it is hard to push this initialization before smp_init(). But it is easy to take an opposite approach and try to initialize the watchdog once again later. The delayed probe is called using workqueues. It need to allocate memory and must be proceed in a normal context. The delayed probe is able to use if watchdog_hardlockup_probe() returns non-zero which means the return code returned when PMU is not ready yet. Provide an API - lockup_detector_retry_init() for anyone who needs to delayed init lockup detector if they had ever failed at lockup_detector_init(). The original assumption is: nobody should use delayed probe after lockup_detector_check() which has __init attribute. That is, anyone uses this API must call between lockup_detector_init() and lockup_detector_check(), and the caller must have __init attribute Link: https://lkml.kernel.org/r/20230519101840.v5.16.If4ad5dd5d09fb1309cebf8bcead4b6a5a7758ca7@changeid Reviewed-by: Petr Mladek Co-developed-by: Pingfan Liu Signed-off-by: Pingfan Liu Signed-off-by: Lecopzer Chen Signed-off-by: Douglas Anderson Suggested-by: Petr Mladek Cc: Andi Kleen Cc: Catalin Marinas Cc: Chen-Yu Tsai Cc: Christophe Leroy Cc: Colin Cross Cc: Daniel Thompson Cc: "David S. Miller" Cc: Guenter Roeck Cc: Ian Rogers Cc: Marc Zyngier Cc: Mark Rutland Cc: Masayoshi Mizuma Cc: Matthias Kaehlcke Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Randy Dunlap Cc: "Ravi V. Shankar" Cc: Ricardo Neri Cc: Stephane Eranian Cc: Stephen Boyd Cc: Sumit Garg Cc: Tzung-Bi Shih Cc: Will Deacon Signed-off-by: Andrew Morton --- include/linux/nmi.h | 2 ++ kernel/watchdog.c | 67 ++++++++++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 68 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 28e65fd1de13..83577ae736cc 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -13,6 +13,7 @@ #ifdef CONFIG_LOCKUP_DETECTOR void lockup_detector_init(void); +void lockup_detector_retry_init(void); void lockup_detector_soft_poweroff(void); void lockup_detector_cleanup(void); @@ -32,6 +33,7 @@ extern int sysctl_hardlockup_all_cpu_backtrace; #else /* CONFIG_LOCKUP_DETECTOR */ static inline void lockup_detector_init(void) { } +static inline void lockup_detector_retry_init(void) { } static inline void lockup_detector_soft_poweroff(void) { } static inline void lockup_detector_cleanup(void) { } #endif /* !CONFIG_LOCKUP_DETECTOR */ diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 10947c835079..237990e8d345 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -208,7 +208,13 @@ void __weak watchdog_hardlockup_enable(unsigned int cpu) { } void __weak watchdog_hardlockup_disable(unsigned int cpu) { } -/* Return 0, if a hardlockup watchdog is available. Error code otherwise */ +/* + * Watchdog-detector specific API. + * + * Return 0 when hardlockup watchdog is available, negative value otherwise. + * Note that the negative value means that a delayed probe might + * succeed later. + */ int __weak __init watchdog_hardlockup_probe(void) { /* @@ -954,6 +960,62 @@ static void __init watchdog_sysctl_init(void) #define watchdog_sysctl_init() do { } while (0) #endif /* CONFIG_SYSCTL */ +static void __init lockup_detector_delay_init(struct work_struct *work); +static bool allow_lockup_detector_init_retry __initdata; + +static struct work_struct detector_work __initdata = + __WORK_INITIALIZER(detector_work, lockup_detector_delay_init); + +static void __init lockup_detector_delay_init(struct work_struct *work) +{ + int ret; + + ret = watchdog_hardlockup_probe(); + if (ret) { + pr_info("Delayed init of the lockup detector failed: %d\n", ret); + pr_info("Hard watchdog permanently disabled\n"); + return; + } + + allow_lockup_detector_init_retry = false; + + watchdog_hardlockup_available = true; + lockup_detector_setup(); +} + +/* + * lockup_detector_retry_init - retry init lockup detector if possible. + * + * Retry hardlockup detector init. It is useful when it requires some + * functionality that has to be initialized later on a particular + * platform. + */ +void __init lockup_detector_retry_init(void) +{ + /* Must be called before late init calls */ + if (!allow_lockup_detector_init_retry) + return; + + schedule_work(&detector_work); +} + +/* + * Ensure that optional delayed hardlockup init is proceed before + * the init code and memory is freed. + */ +static int __init lockup_detector_check(void) +{ + /* Prevent any later retry. */ + allow_lockup_detector_init_retry = false; + + /* Make sure no work is pending. */ + flush_work(&detector_work); + + return 0; + +} +late_initcall_sync(lockup_detector_check); + void __init lockup_detector_init(void) { if (tick_nohz_full_enabled()) @@ -964,6 +1026,9 @@ void __init lockup_detector_init(void) if (!watchdog_hardlockup_probe()) watchdog_hardlockup_available = true; + else + allow_lockup_detector_init_retry = true; + lockup_detector_setup(); watchdog_sysctl_init(); } -- cgit v1.2.3-58-ga151 From 4df3504e2f17cb35d2c12b07e716ae37971d0d52 Mon Sep 17 00:00:00 2001 From: Simon Horman Date: Thu, 25 May 2023 16:26:25 +0200 Subject: kexec: avoid calculating array size twice Avoid calculating array size twice in kexec_purgatory_setup_sechdrs(). Once using array_size(), and once open-coded. Flagged by Coccinelle: .../kexec_file.c:881:8-25: WARNING: array_size is already used (line 877) to compute the same size No functional change intended. Compile tested only. Link: https://lkml.kernel.org/r/20230525-kexec-array_size-v1-1-8b4bf4f7500a@kernel.org Signed-off-by: Simon Horman Acked-by: Baoquan He Cc: Eric W. Biederman Cc: Zhen Lei Signed-off-by: Andrew Morton --- kernel/kexec_file.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec_file.c b/kernel/kexec_file.c index f989f5f1933b..3f5677679744 100644 --- a/kernel/kexec_file.c +++ b/kernel/kexec_file.c @@ -867,6 +867,7 @@ static int kexec_purgatory_setup_sechdrs(struct purgatory_info *pi, { unsigned long bss_addr; unsigned long offset; + size_t sechdrs_size; Elf_Shdr *sechdrs; int i; @@ -874,11 +875,11 @@ static int kexec_purgatory_setup_sechdrs(struct purgatory_info *pi, * The section headers in kexec_purgatory are read-only. In order to * have them modifiable make a temporary copy. */ - sechdrs = vzalloc(array_size(sizeof(Elf_Shdr), pi->ehdr->e_shnum)); + sechdrs_size = array_size(sizeof(Elf_Shdr), pi->ehdr->e_shnum); + sechdrs = vzalloc(sechdrs_size); if (!sechdrs) return -ENOMEM; - memcpy(sechdrs, (void *)pi->ehdr + pi->ehdr->e_shoff, - pi->ehdr->e_shnum * sizeof(Elf_Shdr)); + memcpy(sechdrs, (void *)pi->ehdr + pi->ehdr->e_shoff, sechdrs_size); pi->sechdrs = sechdrs; offset = 0; -- cgit v1.2.3-58-ga151 From 1cba6c4309f03de570202c46f03df3f73a0d4c82 Mon Sep 17 00:00:00 2001 From: Zhen Lei Date: Sat, 27 May 2023 20:34:34 +0800 Subject: kexec: fix a memory leak in crash_shrink_memory() Patch series "kexec: enable kexec_crash_size to support two crash kernel regions". When crashkernel=X fails to reserve region under 4G, it will fall back to reserve region above 4G and a region of the default size will also be reserved under 4G. Unfortunately, /sys/kernel/kexec_crash_size only supports one crash kernel region now, the user cannot sense the low memory reserved by reading /sys/kernel/kexec_crash_size. Also, low memory cannot be freed by writing this file. For example: resource_size(crashk_res) = 512M resource_size(crashk_low_res) = 256M The result of 'cat /sys/kernel/kexec_crash_size' is 512M, but it should be 768M. When we execute 'echo 0 > /sys/kernel/kexec_crash_size', the size of crashk_res becomes 0 and resource_size(crashk_low_res) is still 256 MB, which is incorrect. Since crashk_res manages the memory with high address and crashk_low_res manages the memory with low address, crashk_low_res is shrunken only when all crashk_res is shrunken. And because when there is only one crash kernel region, crashk_res is always used. Therefore, if all crashk_res is shrunken and crashk_low_res still exists, swap them. This patch (of 6): If the value of parameter 'new_size' is in the semi-open and semi-closed interval (crashk_res.end - KEXEC_CRASH_MEM_ALIGN + 1, crashk_res.end], the calculation result of ram_res is: ram_res->start = crashk_res.end + 1 ram_res->end = crashk_res.end The operation of insert_resource() fails, and ram_res is not added to iomem_resource. As a result, the memory of the control block ram_res is leaked. In fact, on all architectures, the start address and size of crashk_res are already aligned by KEXEC_CRASH_MEM_ALIGN. Therefore, we do not need to round up crashk_res.start again. Instead, we should round up 'new_size' in advance. Link: https://lkml.kernel.org/r/20230527123439.772-1-thunder.leizhen@huawei.com Link: https://lkml.kernel.org/r/20230527123439.772-2-thunder.leizhen@huawei.com Fixes: 6480e5a09237 ("kdump: add missing RAM resource in crash_shrink_memory()") Fixes: 06a7f711246b ("kexec: premit reduction of the reserved memory size") Signed-off-by: Zhen Lei Acked-by: Baoquan He Cc: Cong Wang Cc: Eric W. Biederman Cc: Michael Holzheu Signed-off-by: Andrew Morton --- kernel/kexec_core.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index 3d578c6fefee..22acee18195a 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -1122,6 +1122,7 @@ int crash_shrink_memory(unsigned long new_size) start = crashk_res.start; end = crashk_res.end; old_size = (end == 0) ? 0 : end - start + 1; + new_size = roundup(new_size, KEXEC_CRASH_MEM_ALIGN); if (new_size >= old_size) { ret = (new_size == old_size) ? 0 : -EINVAL; goto unlock; @@ -1133,9 +1134,7 @@ int crash_shrink_memory(unsigned long new_size) goto unlock; } - start = roundup(start, KEXEC_CRASH_MEM_ALIGN); - end = roundup(start + new_size, KEXEC_CRASH_MEM_ALIGN); - + end = start + new_size; crash_free_reserved_phys_range(end, crashk_res.end); if ((start == end) && (crashk_res.parent != NULL)) -- cgit v1.2.3-58-ga151 From 6f22a744f4ee7a22be4704cf93bbe22decc7e79e Mon Sep 17 00:00:00 2001 From: Zhen Lei Date: Sat, 27 May 2023 20:34:35 +0800 Subject: kexec: delete a useless check in crash_shrink_memory() The check '(crashk_res.parent != NULL)' is added by commit e05bd3367bd3 ("kexec: fix Oops in crash_shrink_memory()"), but it's stale now. Because if 'crashk_res' is not reserved, it will be zero in size and will be intercepted by the above 'if (new_size >= old_size)'. Ago: if (new_size >= end - start + 1) Now: old_size = (end == 0) ? 0 : end - start + 1; if (new_size >= old_size) Link: https://lkml.kernel.org/r/20230527123439.772-3-thunder.leizhen@huawei.com Signed-off-by: Zhen Lei Cc: Baoquan He Cc: Cong Wang Cc: Eric W. Biederman Cc: Michael Holzheu Signed-off-by: Andrew Morton --- kernel/kexec_core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index 22acee18195a..d1ab139dd490 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -1137,7 +1137,7 @@ int crash_shrink_memory(unsigned long new_size) end = start + new_size; crash_free_reserved_phys_range(end, crashk_res.end); - if ((start == end) && (crashk_res.parent != NULL)) + if (start == end) release_resource(&crashk_res); ram_res->start = end; -- cgit v1.2.3-58-ga151 From f7f567b95b12eda7a6a273b8cb82d152491bc0da Mon Sep 17 00:00:00 2001 From: Zhen Lei Date: Sat, 27 May 2023 20:34:36 +0800 Subject: kexec: clear crashk_res if all its memory has been released If the resource of crashk_res has been released, it is better to clear crashk_res.start and crashk_res.end. Because 'end = start - 1' is not reasonable, and in some places the test is based on crashk_res.end, not resource_size(&crashk_res). Link: https://lkml.kernel.org/r/20230527123439.772-4-thunder.leizhen@huawei.com Signed-off-by: Zhen Lei Acked-by: Baoquan He Cc: Cong Wang Cc: Eric W. Biederman Cc: Michael Holzheu Signed-off-by: Andrew Morton --- kernel/kexec_core.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index d1ab139dd490..bcc86a250ab3 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -1137,15 +1137,18 @@ int crash_shrink_memory(unsigned long new_size) end = start + new_size; crash_free_reserved_phys_range(end, crashk_res.end); - if (start == end) - release_resource(&crashk_res); - ram_res->start = end; ram_res->end = crashk_res.end; ram_res->flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM; ram_res->name = "System RAM"; - crashk_res.end = end - 1; + if (start == end) { + release_resource(&crashk_res); + crashk_res.start = 0; + crashk_res.end = 0; + } else { + crashk_res.end = end - 1; + } insert_resource(&iomem_resource, ram_res); -- cgit v1.2.3-58-ga151 From 8a7db7790a3ff8413bedef886cf135ba301e88d7 Mon Sep 17 00:00:00 2001 From: Zhen Lei Date: Sat, 27 May 2023 20:34:37 +0800 Subject: kexec: improve the readability of crash_shrink_memory() The major adjustments are: 1. end = start + new_size. The 'end' here is not an accurate representation, because it is not the new end of crashk_res, but the start of ram_res, difference 1. So eliminate it and replace it with ram_res->start. 2. Use 'ram_res->start' and 'ram_res->end' as arguments to crash_free_reserved_phys_range() to indicate that the memory covered by 'ram_res' is released from the crashk. And keep it close to insert_resource(). 3. Replace 'if (start == end)' with 'if (!new_size)', clear indication that all crashk memory will be shrunken. No functional change. Link: https://lkml.kernel.org/r/20230527123439.772-5-thunder.leizhen@huawei.com Signed-off-by: Zhen Lei Acked-by: Baoquan He Cc: Cong Wang Cc: Eric W. Biederman Cc: Michael Holzheu Signed-off-by: Andrew Morton --- kernel/kexec_core.c | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index bcc86a250ab3..69fe92141b0b 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -1108,7 +1108,6 @@ ssize_t crash_get_memory_size(void) int crash_shrink_memory(unsigned long new_size) { int ret = 0; - unsigned long start, end; unsigned long old_size; struct resource *ram_res; @@ -1119,9 +1118,7 @@ int crash_shrink_memory(unsigned long new_size) ret = -ENOENT; goto unlock; } - start = crashk_res.start; - end = crashk_res.end; - old_size = (end == 0) ? 0 : end - start + 1; + old_size = !crashk_res.end ? 0 : resource_size(&crashk_res); new_size = roundup(new_size, KEXEC_CRASH_MEM_ALIGN); if (new_size >= old_size) { ret = (new_size == old_size) ? 0 : -EINVAL; @@ -1134,22 +1131,20 @@ int crash_shrink_memory(unsigned long new_size) goto unlock; } - end = start + new_size; - crash_free_reserved_phys_range(end, crashk_res.end); - - ram_res->start = end; + ram_res->start = crashk_res.start + new_size; ram_res->end = crashk_res.end; ram_res->flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM; ram_res->name = "System RAM"; - if (start == end) { + if (!new_size) { release_resource(&crashk_res); crashk_res.start = 0; crashk_res.end = 0; } else { - crashk_res.end = end - 1; + crashk_res.end = ram_res->start - 1; } + crash_free_reserved_phys_range(ram_res->start, ram_res->end); insert_resource(&iomem_resource, ram_res); unlock: -- cgit v1.2.3-58-ga151 From 5b7bfb32cbaad6ba997ab4295ee50bc9921252f2 Mon Sep 17 00:00:00 2001 From: Zhen Lei Date: Sat, 27 May 2023 20:34:38 +0800 Subject: kexec: add helper __crash_shrink_memory() No functional change, in preparation for the next patch so that it is easier to review. [akpm@linux-foundation.org: make __crash_shrink_memory() static] Link: https://lore.kernel.org/oe-kbuild-all/202305280717.Pw06aLkz-lkp@intel.com/ Link: https://lkml.kernel.org/r/20230527123439.772-6-thunder.leizhen@huawei.com Signed-off-by: Zhen Lei Acked-by: Baoquan He Cc: Cong Wang Cc: Eric W. Biederman Cc: Michael Holzheu Signed-off-by: Andrew Morton --- kernel/kexec_core.c | 51 +++++++++++++++++++++++++++++---------------------- 1 file changed, 29 insertions(+), 22 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index 69fe92141b0b..f7ff7de94a61 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -1105,11 +1105,38 @@ ssize_t crash_get_memory_size(void) return size; } +static int __crash_shrink_memory(struct resource *old_res, + unsigned long new_size) +{ + struct resource *ram_res; + + ram_res = kzalloc(sizeof(*ram_res), GFP_KERNEL); + if (!ram_res) + return -ENOMEM; + + ram_res->start = old_res->start + new_size; + ram_res->end = old_res->end; + ram_res->flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM; + ram_res->name = "System RAM"; + + if (!new_size) { + release_resource(old_res); + old_res->start = 0; + old_res->end = 0; + } else { + crashk_res.end = ram_res->start - 1; + } + + crash_free_reserved_phys_range(ram_res->start, ram_res->end); + insert_resource(&iomem_resource, ram_res); + + return 0; +} + int crash_shrink_memory(unsigned long new_size) { int ret = 0; unsigned long old_size; - struct resource *ram_res; if (!kexec_trylock()) return -EBUSY; @@ -1125,27 +1152,7 @@ int crash_shrink_memory(unsigned long new_size) goto unlock; } - ram_res = kzalloc(sizeof(*ram_res), GFP_KERNEL); - if (!ram_res) { - ret = -ENOMEM; - goto unlock; - } - - ram_res->start = crashk_res.start + new_size; - ram_res->end = crashk_res.end; - ram_res->flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM; - ram_res->name = "System RAM"; - - if (!new_size) { - release_resource(&crashk_res); - crashk_res.start = 0; - crashk_res.end = 0; - } else { - crashk_res.end = ram_res->start - 1; - } - - crash_free_reserved_phys_range(ram_res->start, ram_res->end); - insert_resource(&iomem_resource, ram_res); + ret = __crash_shrink_memory(&crashk_res, new_size); unlock: kexec_unlock(); -- cgit v1.2.3-58-ga151 From 16c6006af4d4e70ecef93977a5314409d931020b Mon Sep 17 00:00:00 2001 From: Zhen Lei Date: Sat, 27 May 2023 20:34:39 +0800 Subject: kexec: enable kexec_crash_size to support two crash kernel regions The crashk_low_res should be considered by /sys/kernel/kexec_crash_size to support two crash kernel regions shrinking if existing. While doing it, crashk_low_res will only be shrunk when the entire crashk_res is empty; and if the crashk_res is empty and crahk_low_res is not, change crashk_low_res to be crashk_res. [bhe@redhat.com: redo changelog] Link: https://lkml.kernel.org/r/20230527123439.772-7-thunder.leizhen@huawei.com Signed-off-by: Zhen Lei Acked-by: Baoquan He Cc: Cong Wang Cc: Eric W. Biederman Cc: Michael Holzheu Signed-off-by: Andrew Morton --- kernel/kexec_core.c | 43 ++++++++++++++++++++++++++++++++++++++----- 1 file changed, 38 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index f7ff7de94a61..e2f2574d8b74 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -1091,6 +1091,11 @@ __bpf_kfunc void crash_kexec(struct pt_regs *regs) } } +static inline resource_size_t crash_resource_size(const struct resource *res) +{ + return !res->end ? 0 : resource_size(res); +} + ssize_t crash_get_memory_size(void) { ssize_t size = 0; @@ -1098,8 +1103,8 @@ ssize_t crash_get_memory_size(void) if (!kexec_trylock()) return -EBUSY; - if (crashk_res.end != crashk_res.start) - size = resource_size(&crashk_res); + size += crash_resource_size(&crashk_res); + size += crash_resource_size(&crashk_low_res); kexec_unlock(); return size; @@ -1136,7 +1141,7 @@ static int __crash_shrink_memory(struct resource *old_res, int crash_shrink_memory(unsigned long new_size) { int ret = 0; - unsigned long old_size; + unsigned long old_size, low_size; if (!kexec_trylock()) return -EBUSY; @@ -1145,14 +1150,42 @@ int crash_shrink_memory(unsigned long new_size) ret = -ENOENT; goto unlock; } - old_size = !crashk_res.end ? 0 : resource_size(&crashk_res); + + low_size = crash_resource_size(&crashk_low_res); + old_size = crash_resource_size(&crashk_res) + low_size; new_size = roundup(new_size, KEXEC_CRASH_MEM_ALIGN); if (new_size >= old_size) { ret = (new_size == old_size) ? 0 : -EINVAL; goto unlock; } - ret = __crash_shrink_memory(&crashk_res, new_size); + /* + * (low_size > new_size) implies that low_size is greater than zero. + * This also means that if low_size is zero, the else branch is taken. + * + * If low_size is greater than 0, (low_size > new_size) indicates that + * crashk_low_res also needs to be shrunken. Otherwise, only crashk_res + * needs to be shrunken. + */ + if (low_size > new_size) { + ret = __crash_shrink_memory(&crashk_res, 0); + if (ret) + goto unlock; + + ret = __crash_shrink_memory(&crashk_low_res, new_size); + } else { + ret = __crash_shrink_memory(&crashk_res, new_size - low_size); + } + + /* Swap crashk_res and crashk_low_res if needed */ + if (!crashk_res.end && crashk_low_res.end) { + crashk_res.start = crashk_low_res.start; + crashk_res.end = crashk_low_res.end; + release_resource(&crashk_low_res); + crashk_low_res.start = 0; + crashk_low_res.end = 0; + insert_resource(&iomem_resource, &crashk_res); + } unlock: kexec_unlock(); -- cgit v1.2.3-58-ga151 From a94181ec064b3ad1b6f573f5953e2011f8c90292 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Wed, 7 Jun 2023 16:28:45 +0200 Subject: syscalls: add sys_ni_posix_timers prototype The sys_ni_posix_timers() definition causes a warning when the declaration is missing, so this needs to be added along with the normal syscalls, outside of the #ifdef. kernel/time/posix-stubs.c:26:17: error: no previous prototype for 'sys_ni_posix_timers' [-Werror=missing-prototypes] Link: https://lkml.kernel.org/r/20230607142925.3126422-1-arnd@kernel.org Signed-off-by: Arnd Bergmann Reviewed-by: Kees Cook Signed-off-by: Andrew Morton --- arch/alpha/kernel/osf_sys.c | 2 -- include/linux/syscalls.h | 1 + kernel/time/posix-stubs.c | 1 + 3 files changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/arch/alpha/kernel/osf_sys.c b/arch/alpha/kernel/osf_sys.c index 2a9a877a0508..d98701ee36c6 100644 --- a/arch/alpha/kernel/osf_sys.c +++ b/arch/alpha/kernel/osf_sys.c @@ -1014,8 +1014,6 @@ SYSCALL_DEFINE2(osf_settimeofday, struct timeval32 __user *, tv, return do_sys_settimeofday64(tv ? &kts : NULL, tz ? &ktz : NULL); } -asmlinkage long sys_ni_posix_timers(void); - SYSCALL_DEFINE2(osf_utimes, const char __user *, filename, struct timeval32 __user *, tvs) { diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 33a0ee3bcb2e..24871f8ec8bb 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -1280,6 +1280,7 @@ asmlinkage long sys_ni_syscall(void); #endif /* CONFIG_ARCH_HAS_SYSCALL_WRAPPER */ +asmlinkage long sys_ni_posix_timers(void); /* * Kernel code should not call syscalls (i.e., sys_xyzyyz()) directly. diff --git a/kernel/time/posix-stubs.c b/kernel/time/posix-stubs.c index 828aeecbd1e8..39769b2d1005 100644 --- a/kernel/time/posix-stubs.c +++ b/kernel/time/posix-stubs.c @@ -16,6 +16,7 @@ #include #include #include +#include #ifdef CONFIG_ARCH_HAS_SYSCALL_WRAPPER /* Architectures may override SYS_NI and COMPAT_SYS_NI */ -- cgit v1.2.3-58-ga151 From 9ec272c586b07d1abf73438524bd12b1df9c5f9b Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 26 May 2023 18:41:31 -0700 Subject: watchdog/hardlockup: keep kernel.nmi_watchdog sysctl as 0444 if probe fails Patch series "watchdog: Cleanup / fixes after buddy series v5 reviews". This patch series attempts to finish resolving the feedback received from Petr Mladek on the v5 series I posted. Probably the only thing that wasn't fully as clean as Petr requested was the Kconfig stuff. I couldn't find a better way to express it without a more major overhaul. In the very least, I renamed "NON_ARCH" to "PERF_OR_BUDDY" in the hopes that will make it marginally better. Nothing in this series is terribly critical and even the bugfixes are small. However, it does cleanup a few things that were pointed out in review. This patch (of 10): The permissions for the kernel.nmi_watchdog sysctl have always been set at compile time despite the fact that a watchdog can fail to probe. Let's fix this and set the permissions based on whether the hardlockup detector actually probed. Link: https://lkml.kernel.org/r/20230527014153.2793931-1-dianders@chromium.org Link: https://lkml.kernel.org/r/20230526184139.1.I0d75971cc52a7283f495aac0bd5c3041aadc734e@changeid Fixes: a994a3147e4c ("watchdog/hardlockup/perf: Implement init time detection of perf") Signed-off-by: Douglas Anderson Reported-by: Petr Mladek Closes: https://lore.kernel.org/r/ZHCn4hNxFpY5-9Ki@alley Reviewed-by: Petr Mladek Cc: Christophe Leroy Cc: "David S. Miller" Cc: Michael Ellerman Cc: Nicholas Piggin Signed-off-by: Andrew Morton --- include/linux/nmi.h | 6 ------ kernel/watchdog.c | 30 ++++++++++++++++++++---------- 2 files changed, 20 insertions(+), 16 deletions(-) (limited to 'kernel') diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 83577ae736cc..99b7d748ca21 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -95,12 +95,6 @@ void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs); static inline void arch_touch_nmi_watchdog(void) { } #endif -#if defined(CONFIG_HAVE_NMI_WATCHDOG) || defined(CONFIG_HARDLOCKUP_DETECTOR) -# define NMI_WATCHDOG_SYSCTL_PERM 0644 -#else -# define NMI_WATCHDOG_SYSCTL_PERM 0444 -#endif - #if defined(CONFIG_HARDLOCKUP_DETECTOR_PERF) extern void hardlockup_detector_perf_stop(void); extern void hardlockup_detector_perf_restart(void); diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 237990e8d345..4b9e31edb47f 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -880,15 +880,6 @@ static struct ctl_table watchdog_sysctls[] = { .extra1 = SYSCTL_ZERO, .extra2 = (void *)&sixty, }, - { - .procname = "nmi_watchdog", - .data = &watchdog_hardlockup_user_enabled, - .maxlen = sizeof(int), - .mode = NMI_WATCHDOG_SYSCTL_PERM, - .proc_handler = proc_nmi_watchdog, - .extra1 = SYSCTL_ZERO, - .extra2 = SYSCTL_ONE, - }, { .procname = "watchdog_cpumask", .data = &watchdog_cpumask_bits, @@ -952,10 +943,28 @@ static struct ctl_table watchdog_sysctls[] = { {} }; +static struct ctl_table watchdog_hardlockup_sysctl[] = { + { + .procname = "nmi_watchdog", + .data = &watchdog_hardlockup_user_enabled, + .maxlen = sizeof(int), + .mode = 0444, + .proc_handler = proc_nmi_watchdog, + .extra1 = SYSCTL_ZERO, + .extra2 = SYSCTL_ONE, + }, + {} +}; + static void __init watchdog_sysctl_init(void) { register_sysctl_init("kernel", watchdog_sysctls); + + if (watchdog_hardlockup_available) + watchdog_hardlockup_sysctl[0].mode = 0644; + register_sysctl_init("kernel", watchdog_hardlockup_sysctl); } + #else #define watchdog_sysctl_init() do { } while (0) #endif /* CONFIG_SYSCTL */ @@ -1011,6 +1020,8 @@ static int __init lockup_detector_check(void) /* Make sure no work is pending. */ flush_work(&detector_work); + watchdog_sysctl_init(); + return 0; } @@ -1030,5 +1041,4 @@ void __init lockup_detector_init(void) allow_lockup_detector_init_retry = true; lockup_detector_setup(); - watchdog_sysctl_init(); } -- cgit v1.2.3-58-ga151 From 6426e8d1f27417834ea37e75a9ead832d1cf7713 Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 26 May 2023 18:41:32 -0700 Subject: watchdog/hardlockup: HAVE_NMI_WATCHDOG must implement watchdog_hardlockup_probe() Right now there is one arch (sparc64) that selects HAVE_NMI_WATCHDOG without selecting HAVE_HARDLOCKUP_DETECTOR_ARCH. Because of that one architecture, we have some special case code in the watchdog core to handle the fact that watchdog_hardlockup_probe() isn't implemented. Let's implement watchdog_hardlockup_probe() for sparc64 and get rid of the special case. As a side effect of doing this, code inspection tells us that we could fix a minor bug where the system won't properly realize that NMI watchdogs are disabled. Specifically, on powerpc if CONFIG_PPC_WATCHDOG is turned off the arch might still select CONFIG_HAVE_HARDLOCKUP_DETECTOR_ARCH which selects CONFIG_HAVE_NMI_WATCHDOG. Since CONFIG_PPC_WATCHDOG was off then nothing will override the "weak" watchdog_hardlockup_probe() and we'll fallback to looking at CONFIG_HAVE_NMI_WATCHDOG. Link: https://lkml.kernel.org/r/20230526184139.2.Ic6ebbf307ca0efe91f08ce2c1eb4a037ba6b0700@changeid Signed-off-by: Douglas Anderson Suggested-by: Petr Mladek Reviewed-by: Petr Mladek Cc: Christophe Leroy Cc: "David S. Miller" Cc: Michael Ellerman Cc: Nicholas Piggin Signed-off-by: Andrew Morton --- arch/Kconfig | 3 ++- arch/sparc/kernel/nmi.c | 5 +++++ kernel/watchdog.c | 13 ------------- 3 files changed, 7 insertions(+), 14 deletions(-) (limited to 'kernel') diff --git a/arch/Kconfig b/arch/Kconfig index 205fd23e0cad..422f0ffa269e 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -405,7 +405,8 @@ config HAVE_NMI_WATCHDOG bool help The arch provides a low level NMI watchdog. It provides - asm/nmi.h, and defines its own arch_touch_nmi_watchdog(). + asm/nmi.h, and defines its own watchdog_hardlockup_probe() and + arch_touch_nmi_watchdog(). config HAVE_HARDLOCKUP_DETECTOR_ARCH bool diff --git a/arch/sparc/kernel/nmi.c b/arch/sparc/kernel/nmi.c index 9d9e29b75c43..17cdfdbf1f3b 100644 --- a/arch/sparc/kernel/nmi.c +++ b/arch/sparc/kernel/nmi.c @@ -65,6 +65,11 @@ void arch_touch_nmi_watchdog(void) } EXPORT_SYMBOL(arch_touch_nmi_watchdog); +int __init watchdog_hardlockup_probe(void) +{ + return 0; +} + static void die_nmi(const char *str, struct pt_regs *regs, int do_panic) { int this_cpu = smp_processor_id(); diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 4b9e31edb47f..62230f5b8878 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -217,19 +217,6 @@ void __weak watchdog_hardlockup_disable(unsigned int cpu) { } */ int __weak __init watchdog_hardlockup_probe(void) { - /* - * If CONFIG_HAVE_NMI_WATCHDOG is defined then an architecture - * is assumed to have the hard watchdog available and we return 0. - */ - if (IS_ENABLED(CONFIG_HAVE_NMI_WATCHDOG)) - return 0; - - /* - * Hardlockup detectors other than those using CONFIG_HAVE_NMI_WATCHDOG - * are required to implement a non-weak version of this probe function - * to tell whether they are available. If they don't override then - * we'll return -ENODEV. - */ return -ENODEV; } -- cgit v1.2.3-58-ga151 From 2711e4adef4fac2eeaee66e3c22a2f75ee86e7b3 Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 26 May 2023 18:41:33 -0700 Subject: watchdog/hardlockup: don't use raw_cpu_ptr() in watchdog_hardlockup_kick() In the patch ("watchdog/hardlockup: add a "cpu" param to watchdog_hardlockup_check()") there was no reason to use raw_cpu_ptr(). Using this_cpu_ptr() works fine. Link: https://lkml.kernel.org/r/20230526184139.3.I660e103077dcc23bb29aaf2be09cb234e0495b2d@changeid Signed-off-by: Douglas Anderson Suggested-by: Petr Mladek Reviewed-by: Petr Mladek Cc: Christophe Leroy Cc: "David S. Miller" Cc: Michael Ellerman Cc: Nicholas Piggin Signed-off-by: Andrew Morton --- kernel/watchdog.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 62230f5b8878..32dac8028753 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -133,7 +133,7 @@ static bool is_hardlockup(unsigned int cpu) static unsigned long watchdog_hardlockup_kick(void) { - return atomic_inc_return(raw_cpu_ptr(&hrtimer_interrupts)); + return atomic_inc_return(this_cpu_ptr(&hrtimer_interrupts)); } void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs) -- cgit v1.2.3-58-ga151 From 7a71d8e650b06833095e7a0d4206585e8585c00f Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 26 May 2023 18:41:34 -0700 Subject: watchdog/hardlockup: in watchdog_hardlockup_check() use cpumask_copy() In the patch ("watchdog/hardlockup: add a "cpu" param to watchdog_hardlockup_check()") we started using a cpumask to keep track of which CPUs to backtrace. When setting up this cpumask, it's better to use cpumask_copy() than to just copy the structure directly. Fix this. Link: https://lkml.kernel.org/r/20230526184139.4.Iccee2d1ea19114dafb6553a854ea4d8ab2a3f25b@changeid Signed-off-by: Douglas Anderson Suggested-by: Petr Mladek Reviewed-by: Petr Mladek Cc: Christophe Leroy Cc: "David S. Miller" Cc: Michael Ellerman Cc: Nicholas Piggin Signed-off-by: Andrew Morton --- kernel/watchdog.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 32dac8028753..85f4839b6faf 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -154,7 +154,9 @@ void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs) */ if (is_hardlockup(cpu)) { unsigned int this_cpu = smp_processor_id(); - struct cpumask backtrace_mask = *cpu_online_mask; + struct cpumask backtrace_mask; + + cpumask_copy(&backtrace_mask, cpu_online_mask); /* Only print hardlockups once. */ if (per_cpu(watchdog_hardlockup_warned, cpu)) -- cgit v1.2.3-58-ga151 From d3b62ace0f097f1d863fb6c41df3c61503e4ec9e Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 26 May 2023 18:41:36 -0700 Subject: watchdog/buddy: cleanup how watchdog_buddy_check_hardlockup() is called In the patch ("watchdog/hardlockup: detect hard lockups using secondary (buddy) CPUs"), we added a call from the common watchdog.c file into the buddy. That call could be done more cleanly. Specifically: 1. If we move the call into watchdog_hardlockup_kick() then it keeps watchdog_timer_fn() simpler. 2. We don't need to pass an "unsigned long" to the buddy for the timer count. In the patch ("watchdog/hardlockup: add a "cpu" param to watchdog_hardlockup_check()") the count was changed to "atomic_t" which is backed by an int, so we should match types. Link: https://lkml.kernel.org/r/20230526184139.6.I006c7d958a1ea5c4e1e4dc44a25596d9bb5fd3ba@changeid Signed-off-by: Douglas Anderson Suggested-by: Petr Mladek Reviewed-by: Petr Mladek Cc: Christophe Leroy Cc: "David S. Miller" Cc: Michael Ellerman Cc: Nicholas Piggin Signed-off-by: Andrew Morton --- include/linux/nmi.h | 4 ++-- kernel/watchdog.c | 15 +++++++-------- kernel/watchdog_buddy.c | 2 +- 3 files changed, 10 insertions(+), 11 deletions(-) (limited to 'kernel') diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 3625d64da6db..d35393405b24 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -114,9 +114,9 @@ void watchdog_hardlockup_disable(unsigned int cpu); void lockup_detector_reconfigure(void); #ifdef CONFIG_HARDLOCKUP_DETECTOR_BUDDY -void watchdog_buddy_check_hardlockup(unsigned long hrtimer_interrupts); +void watchdog_buddy_check_hardlockup(int hrtimer_interrupts); #else -static inline void watchdog_buddy_check_hardlockup(unsigned long hrtimer_interrupts) {} +static inline void watchdog_buddy_check_hardlockup(int hrtimer_interrupts) {} #endif /** diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 85f4839b6faf..6cc46b8e3d07 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -131,9 +131,12 @@ static bool is_hardlockup(unsigned int cpu) return false; } -static unsigned long watchdog_hardlockup_kick(void) +static void watchdog_hardlockup_kick(void) { - return atomic_inc_return(this_cpu_ptr(&hrtimer_interrupts)); + int new_interrupts; + + new_interrupts = atomic_inc_return(this_cpu_ptr(&hrtimer_interrupts)); + watchdog_buddy_check_hardlockup(new_interrupts); } void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs) @@ -195,7 +198,7 @@ void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs) #else /* CONFIG_HARDLOCKUP_DETECTOR_COUNTS_HRTIMER */ -static inline unsigned long watchdog_hardlockup_kick(void) { return 0; } +static inline void watchdog_hardlockup_kick(void) { } #endif /* !CONFIG_HARDLOCKUP_DETECTOR_COUNTS_HRTIMER */ @@ -449,15 +452,11 @@ static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer) struct pt_regs *regs = get_irq_regs(); int duration; int softlockup_all_cpu_backtrace = sysctl_softlockup_all_cpu_backtrace; - unsigned long hrtimer_interrupts; if (!watchdog_enabled) return HRTIMER_NORESTART; - hrtimer_interrupts = watchdog_hardlockup_kick(); - - /* test for hardlockups */ - watchdog_buddy_check_hardlockup(hrtimer_interrupts); + watchdog_hardlockup_kick(); /* kick the softlockup detector */ if (completion_done(this_cpu_ptr(&softlockup_completion))) { diff --git a/kernel/watchdog_buddy.c b/kernel/watchdog_buddy.c index fee45af2e5bd..3ffc5f2ede5a 100644 --- a/kernel/watchdog_buddy.c +++ b/kernel/watchdog_buddy.c @@ -72,7 +72,7 @@ void watchdog_hardlockup_disable(unsigned int cpu) cpumask_clear_cpu(cpu, &watchdog_cpus); } -void watchdog_buddy_check_hardlockup(unsigned long hrtimer_interrupts) +void watchdog_buddy_check_hardlockup(int hrtimer_interrupts) { unsigned int next_cpu; -- cgit v1.2.3-58-ga151 From 813efda23934edcad96343fc96727017378c3fe9 Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 26 May 2023 18:41:37 -0700 Subject: watchdog/buddy: don't copy the cpumask in watchdog_next_cpu() There's no reason to make a copy of the "watchdog_cpus" locally in watchdog_next_cpu(). Making a copy wouldn't make things any more race free and we're just reading the value so there's no need for a copy. Link: https://lkml.kernel.org/r/20230526184139.7.If466f9a2b50884cbf6a1d8ad05525a2c17069407@changeid Signed-off-by: Douglas Anderson Suggested-by: Petr Mladek Reviewed-by: Petr Mladek Cc: Christophe Leroy Cc: "David S. Miller" Cc: Michael Ellerman Cc: Nicholas Piggin Signed-off-by: Andrew Morton --- kernel/watchdog_buddy.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog_buddy.c b/kernel/watchdog_buddy.c index 3ffc5f2ede5a..2ef88722c5e7 100644 --- a/kernel/watchdog_buddy.c +++ b/kernel/watchdog_buddy.c @@ -10,12 +10,11 @@ static cpumask_t __read_mostly watchdog_cpus; static unsigned int watchdog_next_cpu(unsigned int cpu) { - cpumask_t cpus = watchdog_cpus; unsigned int next_cpu; - next_cpu = cpumask_next(cpu, &cpus); + next_cpu = cpumask_next(cpu, &watchdog_cpus); if (next_cpu >= nr_cpu_ids) - next_cpu = cpumask_first(&cpus); + next_cpu = cpumask_first(&watchdog_cpus); if (next_cpu == cpu) return nr_cpu_ids; -- cgit v1.2.3-58-ga151 From 28168eca3297d68faa8a9433ec93cb6acf06d2f4 Mon Sep 17 00:00:00 2001 From: Douglas Anderson Date: Fri, 26 May 2023 18:41:39 -0700 Subject: watchdog/hardlockup: move SMP barriers from common code to buddy code It's been suggested that since the SMP barriers are only potentially useful for the buddy hardlockup detector, not the perf hardlockup detector, that the barriers belong in the buddy code. Let's move them and add clearer comments about why they're needed. Link: https://lkml.kernel.org/r/20230526184139.9.I5ab0a0eeb0bd52fb23f901d298c72fa5c396e22b@changeid Signed-off-by: Douglas Anderson Suggested-by: Petr Mladek Reviewed-by: Petr Mladek Cc: Christophe Leroy Cc: "David S. Miller" Cc: Michael Ellerman Cc: Nicholas Piggin Signed-off-by: Andrew Morton --- kernel/watchdog.c | 6 ------ kernel/watchdog_buddy.c | 21 +++++++++++++++++++++ 2 files changed, 21 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 6cc46b8e3d07..a351ab0c35eb 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -109,9 +109,6 @@ EXPORT_SYMBOL(arch_touch_nmi_watchdog); void watchdog_hardlockup_touch_cpu(unsigned int cpu) { per_cpu(watchdog_hardlockup_touched, cpu) = true; - - /* Match with smp_rmb() in watchdog_hardlockup_check() */ - smp_wmb(); } static bool is_hardlockup(unsigned int cpu) @@ -141,9 +138,6 @@ static void watchdog_hardlockup_kick(void) void watchdog_hardlockup_check(unsigned int cpu, struct pt_regs *regs) { - /* Match with smp_wmb() in watchdog_hardlockup_touch_cpu() */ - smp_rmb(); - if (per_cpu(watchdog_hardlockup_touched, cpu)) { per_cpu(watchdog_hardlockup_touched, cpu) = false; return; diff --git a/kernel/watchdog_buddy.c b/kernel/watchdog_buddy.c index 2ef88722c5e7..34dbfe091f4b 100644 --- a/kernel/watchdog_buddy.c +++ b/kernel/watchdog_buddy.c @@ -51,6 +51,13 @@ void watchdog_hardlockup_enable(unsigned int cpu) if (next_cpu < nr_cpu_ids) watchdog_hardlockup_touch_cpu(next_cpu); + /* + * Makes sure that watchdog is touched on this CPU before + * other CPUs could see it in watchdog_cpus. The counter + * part is in watchdog_buddy_check_hardlockup(). + */ + smp_wmb(); + cpumask_set_cpu(cpu, &watchdog_cpus); } @@ -68,6 +75,13 @@ void watchdog_hardlockup_disable(unsigned int cpu) if (next_cpu < nr_cpu_ids) watchdog_hardlockup_touch_cpu(next_cpu); + /* + * Makes sure that watchdog is touched on the next CPU before + * this CPU disappear in watchdog_cpus. The counter part is in + * watchdog_buddy_check_hardlockup(). + */ + smp_wmb(); + cpumask_clear_cpu(cpu, &watchdog_cpus); } @@ -88,5 +102,12 @@ void watchdog_buddy_check_hardlockup(int hrtimer_interrupts) if (next_cpu >= nr_cpu_ids) return; + /* + * Make sure that the watchdog was touched on next CPU when + * watchdog_next_cpu() returned another one because of + * a change in watchdog_hardlockup_enable()/disable(). + */ + smp_rmb(); + watchdog_hardlockup_check(next_cpu, NULL); } -- cgit v1.2.3-58-ga151 From a5fcc2367e223c45c78a882438c2b8e13fe0f580 Mon Sep 17 00:00:00 2001 From: Petr Mladek Date: Fri, 16 Jun 2023 17:06:16 +0200 Subject: watchdog/hardlockup: make HAVE_NMI_WATCHDOG sparc64-specific There are several hardlockup detector implementations and several Kconfig values which allow selection and build of the preferred one. CONFIG_HARDLOCKUP_DETECTOR was introduced by the commit 23637d477c1f53acb ("lockup_detector: Introduce CONFIG_HARDLOCKUP_DETECTOR") in v2.6.36. It was a preparation step for introducing the new generic perf hardlockup detector. The existing arch-specific variants did not support the to-be-created generic build configurations, sysctl interface, etc. This distinction was made explicit by the commit 4a7863cc2eb5f98 ("x86, nmi_watchdog: Remove ARCH_HAS_NMI_WATCHDOG and rely on CONFIG_HARDLOCKUP_DETECTOR") in v2.6.38. CONFIG_HAVE_NMI_WATCHDOG was introduced by the commit d314d74c695f967e105 ("nmi watchdog: do not use cpp symbol in Kconfig") in v3.4-rc1. It replaced the above mentioned ARCH_HAS_NMI_WATCHDOG. At that time, it was still used by three architectures, namely blackfin, mn10300, and sparc. The support for blackfin and mn10300 architectures has been completely dropped some time ago. And sparc is the only architecture with the historic NMI watchdog at the moment. And the old sparc implementation is really special. It is always built on sparc64. It used to be always enabled until the commit 7a5c8b57cec93196b ("sparc: implement watchdog_nmi_enable and watchdog_nmi_disable") added in v4.10-rc1. There are only few locations where the sparc64 NMI watchdog interacts with the generic hardlockup detectors code: + implements arch_touch_nmi_watchdog() which is called from the generic touch_nmi_watchdog() + implements watchdog_hardlockup_enable()/disable() to support /proc/sys/kernel/nmi_watchdog + is always preferred over other generic watchdogs, see CONFIG_HARDLOCKUP_DETECTOR + includes asm/nmi.h into linux/nmi.h because some sparc-specific functions are needed in sparc-specific code which includes only linux/nmi.h. The situation became more complicated after the commit 05a4a95279311c3 ("kernel/watchdog: split up config options") and commit 2104180a53698df5 ("powerpc/64s: implement arch-specific hardlockup watchdog") in v4.13-rc1. They introduced HAVE_HARDLOCKUP_DETECTOR_ARCH. It was used for powerpc specific hardlockup detector. It was compatible with the perf one regarding the general boot, sysctl, and programming interfaces. HAVE_HARDLOCKUP_DETECTOR_ARCH was defined as a superset of HAVE_NMI_WATCHDOG. It made some sense because all arch-specific detectors had some common requirements, namely: + implemented arch_touch_nmi_watchdog() + included asm/nmi.h into linux/nmi.h + defined the default value for /proc/sys/kernel/nmi_watchdog But it actually has made things pretty complicated when the generic buddy hardlockup detector was added. Before the generic perf detector was newer supported together with an arch-specific one. But the buddy detector could work on any SMP system. It means that an architecture could support both the arch-specific and buddy detector. As a result, there are few tricky dependencies. For example, CONFIG_HARDLOCKUP_DETECTOR depends on: ((HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_BUDDY) && !HAVE_NMI_WATCHDOG) || HAVE_HARDLOCKUP_DETECTOR_ARCH The problem is that the very special sparc implementation is defined as: HAVE_NMI_WATCHDOG && !HAVE_HARDLOCKUP_DETECTOR_ARCH Another problem is that the meaning of HAVE_NMI_WATCHDOG is far from clear without reading understanding the history. Make the logic less tricky and more self-explanatory by making HAVE_NMI_WATCHDOG specific for the sparc64 implementation. And rename it to HAVE_HARDLOCKUP_DETECTOR_SPARC64. Note that HARDLOCKUP_DETECTOR_PREFER_BUDDY, HARDLOCKUP_DETECTOR_PERF, and HARDLOCKUP_DETECTOR_BUDDY may conflict only with HAVE_HARDLOCKUP_DETECTOR_ARCH. They depend on HARDLOCKUP_DETECTOR and it is not longer enabled when HAVE_NMI_WATCHDOG is set. Link: https://lkml.kernel.org/r/20230616150618.6073-5-pmladek@suse.com Signed-off-by: Petr Mladek Reviewed-by: Douglas Anderson Cc: Christophe Leroy Cc: "David S. Miller" Cc: Michael Ellerman Cc: Nicholas Piggin Signed-off-by: Andrew Morton --- arch/Kconfig | 18 ------------------ arch/sparc/Kconfig | 2 +- arch/sparc/Kconfig.debug | 9 +++++++++ include/linux/nmi.h | 5 ++--- kernel/watchdog.c | 2 +- lib/Kconfig.debug | 15 +++++---------- 6 files changed, 18 insertions(+), 33 deletions(-) (limited to 'kernel') diff --git a/arch/Kconfig b/arch/Kconfig index 77e5af5fda3f..6517e5477459 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -400,23 +400,8 @@ config HAVE_HARDLOCKUP_DETECTOR_PERF The arch chooses to use the generic perf-NMI-based hardlockup detector. Must define HAVE_PERF_EVENTS_NMI. -config HAVE_NMI_WATCHDOG - depends on HAVE_NMI - bool - help - The arch provides its own hardlockup detector implementation instead - of the generic ones. - - Sparc64 defines this variable without HAVE_HARDLOCKUP_DETECTOR_ARCH. - It is the last arch-specific implementation which was developed before - adding the common infrastructure for handling hardlockup detectors. - It is always built. It does _not_ use the common command line - parameters and sysctl interface, except for - /proc/sys/kernel/nmi_watchdog. - config HAVE_HARDLOCKUP_DETECTOR_ARCH bool - select HAVE_NMI_WATCHDOG help The arch provides its own hardlockup detector implementation instead of the generic ones. @@ -424,9 +409,6 @@ config HAVE_HARDLOCKUP_DETECTOR_ARCH It uses the same command line parameters, and sysctl interface, as the generic hardlockup detectors. - HAVE_NMI_WATCHDOG is selected to build the code shared with - the sparc64 specific implementation. - config HAVE_PERF_REGS bool help diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig index 8535e19062f6..7297f69635cb 100644 --- a/arch/sparc/Kconfig +++ b/arch/sparc/Kconfig @@ -33,7 +33,7 @@ config SPARC select ARCH_WANT_IPC_PARSE_VERSION select GENERIC_PCI_IOMAP select HAS_IOPORT - select HAVE_NMI_WATCHDOG if SPARC64 + select HAVE_HARDLOCKUP_DETECTOR_SPARC64 if SPARC64 select HAVE_CBPF_JIT if SPARC32 select HAVE_EBPF_JIT if SPARC64 select HAVE_DEBUG_BUGVERBOSE diff --git a/arch/sparc/Kconfig.debug b/arch/sparc/Kconfig.debug index 6b2bec1888b3..4903b6847e43 100644 --- a/arch/sparc/Kconfig.debug +++ b/arch/sparc/Kconfig.debug @@ -14,3 +14,12 @@ config FRAME_POINTER bool depends on MCOUNT default y + +config HAVE_HARDLOCKUP_DETECTOR_SPARC64 + depends on HAVE_NMI + bool + help + Sparc64 hardlockup detector is the last one developed before adding + the common infrastructure for handling hardlockup detectors. It is + always built. It does _not_ use the common command line parameters + and sysctl interface, except for /proc/sys/kernel/nmi_watchdog. diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 07bf2b813463..63acc6586774 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -9,7 +9,7 @@ #include /* Arch specific watchdogs might need to share extra watchdog-related APIs. */ -#if defined(CONFIG_HAVE_NMI_WATCHDOG) +#if defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_ARCH) || defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_SPARC64) #include #endif @@ -90,8 +90,7 @@ static inline void hardlockup_detector_disable(void) {} #endif /* Sparc64 has special implemetantion that is always enabled. */ -#if defined(CONFIG_HARDLOCKUP_DETECTOR) || \ - (defined(CONFIG_HAVE_NMI_WATCHDOG) && !defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_ARCH)) +#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_SPARC64) void arch_touch_nmi_watchdog(void); #else static inline void arch_touch_nmi_watchdog(void) { } diff --git a/kernel/watchdog.c b/kernel/watchdog.c index a351ab0c35eb..010fcc3ac141 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -29,7 +29,7 @@ static DEFINE_MUTEX(watchdog_mutex); -#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_NMI_WATCHDOG) +#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_SPARC64) # define WATCHDOG_HARDLOCKUP_DEFAULT 1 #else # define WATCHDOG_HARDLOCKUP_DEFAULT 0 diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index a0b0c4decb89..e94664339e28 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -1050,15 +1050,10 @@ config HAVE_HARDLOCKUP_DETECTOR_BUDDY # sparc64: has a custom implementation which is not using the common # hardlockup command line options and sysctl interface. # -# Note that HAVE_NMI_WATCHDOG is used to distinguish the sparc64 specific -# implementaion. It is automatically enabled also for other arch-specific -# variants which set HAVE_HARDLOCKUP_DETECTOR_ARCH. It makes the check -# of avaialable and supported variants quite tricky. -# config HARDLOCKUP_DETECTOR bool "Detect Hard Lockups" - depends on DEBUG_KERNEL && !S390 - depends on ((HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_BUDDY) && !HAVE_NMI_WATCHDOG) || HAVE_HARDLOCKUP_DETECTOR_ARCH + depends on DEBUG_KERNEL && !S390 && !HAVE_HARDLOCKUP_DETECTOR_SPARC64 + depends on HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_BUDDY || HAVE_HARDLOCKUP_DETECTOR_ARCH imply HARDLOCKUP_DETECTOR_PERF imply HARDLOCKUP_DETECTOR_BUDDY select LOCKUP_DETECTOR @@ -1079,7 +1074,7 @@ config HARDLOCKUP_DETECTOR_PREFER_BUDDY bool "Prefer the buddy CPU hardlockup detector" depends on HARDLOCKUP_DETECTOR depends on HAVE_HARDLOCKUP_DETECTOR_PERF && HAVE_HARDLOCKUP_DETECTOR_BUDDY - depends on !HAVE_NMI_WATCHDOG + depends on !HAVE_HARLOCKUP_DETECTOR_ARCH help Say Y here to prefer the buddy hardlockup detector over the perf one. @@ -1095,7 +1090,7 @@ config HARDLOCKUP_DETECTOR_PERF bool depends on HARDLOCKUP_DETECTOR depends on HAVE_HARDLOCKUP_DETECTOR_PERF && !HARDLOCKUP_DETECTOR_PREFER_BUDDY - depends on !HAVE_NMI_WATCHDOG + depends on !HAVE_HARDLOCKUP_DETECTOR_ARCH select HARDLOCKUP_DETECTOR_COUNTS_HRTIMER config HARDLOCKUP_DETECTOR_BUDDY @@ -1103,7 +1098,7 @@ config HARDLOCKUP_DETECTOR_BUDDY depends on HARDLOCKUP_DETECTOR depends on HAVE_HARDLOCKUP_DETECTOR_BUDDY depends on !HAVE_HARDLOCKUP_DETECTOR_PERF || HARDLOCKUP_DETECTOR_PREFER_BUDDY - depends on !HAVE_NMI_WATCHDOG + depends on !HAVE_HARDLOCKUP_DETECTOR_ARCH select HARDLOCKUP_DETECTOR_COUNTS_HRTIMER # -- cgit v1.2.3-58-ga151 From 47f4cb433923a08d81f1e5c065cb680215109db9 Mon Sep 17 00:00:00 2001 From: Petr Mladek Date: Fri, 16 Jun 2023 17:06:17 +0200 Subject: watchdog/sparc64: define HARDLOCKUP_DETECTOR_SPARC64 The HAVE_ prefix means that the code could be enabled. Add another variable for HAVE_HARDLOCKUP_DETECTOR_SPARC64 without this prefix. It will be set when it should be built. It will make it compatible with the other hardlockup detectors. Before, it is far from obvious that the SPARC64 variant is actually used: $> make ARCH=sparc64 defconfig $> grep HARDLOCKUP_DETECTOR .config CONFIG_HAVE_HARDLOCKUP_DETECTOR_BUDDY=y CONFIG_HAVE_HARDLOCKUP_DETECTOR_SPARC64=y After, it is more clear: $> make ARCH=sparc64 defconfig $> grep HARDLOCKUP_DETECTOR .config CONFIG_HAVE_HARDLOCKUP_DETECTOR_BUDDY=y CONFIG_HAVE_HARDLOCKUP_DETECTOR_SPARC64=y CONFIG_HARDLOCKUP_DETECTOR_SPARC64=y Link: https://lkml.kernel.org/r/20230616150618.6073-6-pmladek@suse.com Signed-off-by: Petr Mladek Reviewed-by: Douglas Anderson Cc: Christophe Leroy Cc: "David S. Miller" Cc: Michael Ellerman Cc: Nicholas Piggin Signed-off-by: Andrew Morton --- arch/sparc/Kconfig.debug | 7 ++++++- include/linux/nmi.h | 4 ++-- kernel/watchdog.c | 2 +- lib/Kconfig.debug | 2 +- 4 files changed, 10 insertions(+), 5 deletions(-) (limited to 'kernel') diff --git a/arch/sparc/Kconfig.debug b/arch/sparc/Kconfig.debug index 4903b6847e43..37e003665de6 100644 --- a/arch/sparc/Kconfig.debug +++ b/arch/sparc/Kconfig.debug @@ -16,10 +16,15 @@ config FRAME_POINTER default y config HAVE_HARDLOCKUP_DETECTOR_SPARC64 - depends on HAVE_NMI bool + depends on HAVE_NMI + select HARDLOCKUP_DETECTOR_SPARC64 help Sparc64 hardlockup detector is the last one developed before adding the common infrastructure for handling hardlockup detectors. It is always built. It does _not_ use the common command line parameters and sysctl interface, except for /proc/sys/kernel/nmi_watchdog. + +config HARDLOCKUP_DETECTOR_SPARC64 + bool + depends on HAVE_HARDLOCKUP_DETECTOR_SPARC64 diff --git a/include/linux/nmi.h b/include/linux/nmi.h index 63acc6586774..e91a1f803d55 100644 --- a/include/linux/nmi.h +++ b/include/linux/nmi.h @@ -9,7 +9,7 @@ #include /* Arch specific watchdogs might need to share extra watchdog-related APIs. */ -#if defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_ARCH) || defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_SPARC64) +#if defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_ARCH) || defined(CONFIG_HARDLOCKUP_DETECTOR_SPARC64) #include #endif @@ -90,7 +90,7 @@ static inline void hardlockup_detector_disable(void) {} #endif /* Sparc64 has special implemetantion that is always enabled. */ -#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_SPARC64) +#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HARDLOCKUP_DETECTOR_SPARC64) void arch_touch_nmi_watchdog(void); #else static inline void arch_touch_nmi_watchdog(void) { } diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 010fcc3ac141..be38276a365f 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -29,7 +29,7 @@ static DEFINE_MUTEX(watchdog_mutex); -#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HAVE_HARDLOCKUP_DETECTOR_SPARC64) +#if defined(CONFIG_HARDLOCKUP_DETECTOR) || defined(CONFIG_HARDLOCKUP_DETECTOR_SPARC64) # define WATCHDOG_HARDLOCKUP_DEFAULT 1 #else # define WATCHDOG_HARDLOCKUP_DEFAULT 0 diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index e94664339e28..f285e9cf967a 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -1052,7 +1052,7 @@ config HAVE_HARDLOCKUP_DETECTOR_BUDDY # config HARDLOCKUP_DETECTOR bool "Detect Hard Lockups" - depends on DEBUG_KERNEL && !S390 && !HAVE_HARDLOCKUP_DETECTOR_SPARC64 + depends on DEBUG_KERNEL && !S390 && !HARDLOCKUP_DETECTOR_SPARC64 depends on HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_BUDDY || HAVE_HARDLOCKUP_DETECTOR_ARCH imply HARDLOCKUP_DETECTOR_PERF imply HARDLOCKUP_DETECTOR_BUDDY -- cgit v1.2.3-58-ga151 From 4afc9a402aa3890885747b396c1adcd45f127665 Mon Sep 17 00:00:00 2001 From: Yang Li Date: Thu, 8 Jun 2023 16:23:12 +0800 Subject: kernel/time/posix-stubs.c: remove duplicated include ./kernel/time/posix-stubs.c: linux/syscalls.h is included more than once. Link: https://lkml.kernel.org/r/20230608082312.123939-1-yang.lee@linux.alibaba.com Reported-by: Abaci Robot Closes: https://bugzilla.openanolis.cn/show_bug.cgi?id=5463 Fixes: c1956519cd7e ("syscalls: add sys_ni_posix_timers prototype") Signed-off-by: Yang Li Cc: Arnd Bergmann Signed-off-by: Andrew Morton --- kernel/time/posix-stubs.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel') diff --git a/kernel/time/posix-stubs.c b/kernel/time/posix-stubs.c index 39769b2d1005..828aeecbd1e8 100644 --- a/kernel/time/posix-stubs.c +++ b/kernel/time/posix-stubs.c @@ -16,7 +16,6 @@ #include #include #include -#include #ifdef CONFIG_ARCH_HAS_SYSCALL_WRAPPER /* Architectures may override SYS_NI and COMPAT_SYS_NI */ -- cgit v1.2.3-58-ga151