diff options
27 files changed, 696 insertions, 395 deletions
diff --git a/patches/0001_random_remove_unused_irq_flags_argument_from_add_interrupt_randomness.patch b/patches/0001_random_remove_unused_irq_flags_argument_from_add_interrupt_randomness.patch new file mode 100644 index 000000000000..e596a409f7ec --- /dev/null +++ b/patches/0001_random_remove_unused_irq_flags_argument_from_add_interrupt_randomness.patch @@ -0,0 +1,97 @@ +From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Subject: random: Remove unused irq_flags argument from add_interrupt_randomness(). +Date: Tue, 07 Dec 2021 13:17:33 +0100 + +Since commit + ee3e00e9e7101 ("random: use registers from interrupted code for CPU's w/o a cycle counter") + +the irq_flags argument is no longer used. + +Remove unused irq_flags irq_flags. + +Cc: Borislav Petkov <bp@alien8.de> +Cc: Dave Hansen <dave.hansen@linux.intel.com> +Cc: Dexuan Cui <decui@microsoft.com> +Cc: H. Peter Anvin <hpa@zytor.com> +Cc: Haiyang Zhang <haiyangz@microsoft.com> +Cc: Ingo Molnar <mingo@redhat.com> +Cc: K. Y. Srinivasan <kys@microsoft.com> +Cc: Stephen Hemminger <sthemmin@microsoft.com> +Cc: Thomas Gleixner <tglx@linutronix.de> +Cc: Wei Liu <wei.liu@kernel.org> +Cc: linux-hyperv@vger.kernel.org +Cc: x86@kernel.org +Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Link: https://lore.kernel.org/r/20211207121737.2347312-2-bigeasy@linutronix.de +--- + arch/x86/kernel/cpu/mshyperv.c | 2 +- + drivers/char/random.c | 4 ++-- + drivers/hv/vmbus_drv.c | 2 +- + include/linux/random.h | 2 +- + kernel/irq/handle.c | 2 +- + 5 files changed, 6 insertions(+), 6 deletions(-) + +--- a/arch/x86/kernel/cpu/mshyperv.c ++++ b/arch/x86/kernel/cpu/mshyperv.c +@@ -79,7 +79,7 @@ DEFINE_IDTENTRY_SYSVEC(sysvec_hyperv_sti + inc_irq_stat(hyperv_stimer0_count); + if (hv_stimer0_handler) + hv_stimer0_handler(); +- add_interrupt_randomness(HYPERV_STIMER0_VECTOR, 0); ++ add_interrupt_randomness(HYPERV_STIMER0_VECTOR); + ack_APIC_irq(); + + set_irq_regs(old_regs); +--- a/drivers/char/random.c ++++ b/drivers/char/random.c +@@ -200,7 +200,7 @@ + * void add_device_randomness(const void *buf, unsigned int size); + * void add_input_randomness(unsigned int type, unsigned int code, + * unsigned int value); +- * void add_interrupt_randomness(int irq, int irq_flags); ++ * void add_interrupt_randomness(int irq); + * void add_disk_randomness(struct gendisk *disk); + * + * add_device_randomness() is for adding data to the random pool that +@@ -1242,7 +1242,7 @@ static __u32 get_reg(struct fast_pool *f + return *ptr; + } + +-void add_interrupt_randomness(int irq, int irq_flags) ++void add_interrupt_randomness(int irq) + { + struct entropy_store *r; + struct fast_pool *fast_pool = this_cpu_ptr(&irq_randomness); +--- a/drivers/hv/vmbus_drv.c ++++ b/drivers/hv/vmbus_drv.c +@@ -1381,7 +1381,7 @@ static void vmbus_isr(void) + tasklet_schedule(&hv_cpu->msg_dpc); + } + +- add_interrupt_randomness(vmbus_interrupt, 0); ++ add_interrupt_randomness(vmbus_interrupt); + } + + static irqreturn_t vmbus_percpu_isr(int irq, void *dev_id) +--- a/include/linux/random.h ++++ b/include/linux/random.h +@@ -35,7 +35,7 @@ static inline void add_latent_entropy(vo + + extern void add_input_randomness(unsigned int type, unsigned int code, + unsigned int value) __latent_entropy; +-extern void add_interrupt_randomness(int irq, int irq_flags) __latent_entropy; ++extern void add_interrupt_randomness(int irq) __latent_entropy; + + extern void get_random_bytes(void *buf, int nbytes); + extern int wait_for_random_bytes(void); +--- a/kernel/irq/handle.c ++++ b/kernel/irq/handle.c +@@ -197,7 +197,7 @@ irqreturn_t handle_irq_event_percpu(stru + + retval = __handle_irq_event_percpu(desc, &flags); + +- add_interrupt_randomness(desc->irq_data.irq, flags); ++ add_interrupt_randomness(desc->irq_data.irq); + + if (!irq_settings_no_debug(desc)) + note_interrupt(desc, retval); diff --git a/patches/0002_irq_remove_unsued_flags_argument_from___handle_irq_event_percpu.patch b/patches/0002_irq_remove_unsued_flags_argument_from___handle_irq_event_percpu.patch new file mode 100644 index 000000000000..70a3f0a993a4 --- /dev/null +++ b/patches/0002_irq_remove_unsued_flags_argument_from___handle_irq_event_percpu.patch @@ -0,0 +1,82 @@ +From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Subject: irq: Remove unsued flags argument from __handle_irq_event_percpu(). +Date: Tue, 07 Dec 2021 13:17:34 +0100 + +The __IRQF_TIMER bit from the flags argument was used in +add_interrupt_randomness() to distinguish the timer interrupt from other +interrupts. This is no longer the case. + +Remove the flags argument from __handle_irq_event_percpu(). + +Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Link: https://lore.kernel.org/r/20211207121737.2347312-3-bigeasy@linutronix.de +--- + kernel/irq/chip.c | 4 +--- + kernel/irq/handle.c | 9 ++------- + kernel/irq/internals.h | 2 +- + 3 files changed, 4 insertions(+), 11 deletions(-) + +--- a/kernel/irq/chip.c ++++ b/kernel/irq/chip.c +@@ -575,8 +575,6 @@ EXPORT_SYMBOL_GPL(handle_simple_irq); + */ + void handle_untracked_irq(struct irq_desc *desc) + { +- unsigned int flags = 0; +- + raw_spin_lock(&desc->lock); + + if (!irq_may_run(desc)) +@@ -593,7 +591,7 @@ void handle_untracked_irq(struct irq_des + irqd_set(&desc->irq_data, IRQD_IRQ_INPROGRESS); + raw_spin_unlock(&desc->lock); + +- __handle_irq_event_percpu(desc, &flags); ++ __handle_irq_event_percpu(desc); + + raw_spin_lock(&desc->lock); + irqd_clear(&desc->irq_data, IRQD_IRQ_INPROGRESS); +--- a/kernel/irq/handle.c ++++ b/kernel/irq/handle.c +@@ -136,7 +136,7 @@ void __irq_wake_thread(struct irq_desc * + wake_up_process(action->thread); + } + +-irqreturn_t __handle_irq_event_percpu(struct irq_desc *desc, unsigned int *flags) ++irqreturn_t __handle_irq_event_percpu(struct irq_desc *desc) + { + irqreturn_t retval = IRQ_NONE; + unsigned int irq = desc->irq_data.irq; +@@ -174,10 +174,6 @@ irqreturn_t __handle_irq_event_percpu(st + } + + __irq_wake_thread(desc, action); +- +- fallthrough; /* to add to randomness */ +- case IRQ_HANDLED: +- *flags |= action->flags; + break; + + default: +@@ -193,9 +189,8 @@ irqreturn_t __handle_irq_event_percpu(st + irqreturn_t handle_irq_event_percpu(struct irq_desc *desc) + { + irqreturn_t retval; +- unsigned int flags = 0; + +- retval = __handle_irq_event_percpu(desc, &flags); ++ retval = __handle_irq_event_percpu(desc); + + add_interrupt_randomness(desc->irq_data.irq); + +--- a/kernel/irq/internals.h ++++ b/kernel/irq/internals.h +@@ -103,7 +103,7 @@ extern int __irq_get_irqchip_state(struc + + extern void init_kstat_irqs(struct irq_desc *desc, int node, int nr); + +-irqreturn_t __handle_irq_event_percpu(struct irq_desc *desc, unsigned int *flags); ++irqreturn_t __handle_irq_event_percpu(struct irq_desc *desc); + irqreturn_t handle_irq_event_percpu(struct irq_desc *desc); + irqreturn_t handle_irq_event(struct irq_desc *desc); + diff --git a/patches/0003_random_split_add_interrupt_randomness.patch b/patches/0003_random_split_add_interrupt_randomness.patch new file mode 100644 index 000000000000..bd0750993a37 --- /dev/null +++ b/patches/0003_random_split_add_interrupt_randomness.patch @@ -0,0 +1,87 @@ +From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Subject: random: Split add_interrupt_randomness(). +Date: Tue, 07 Dec 2021 13:17:35 +0100 + +Split add_interrupt_randomness() into two parts: +- add_interrupt_randomness() which collects the entropy on the + invocation of a hardware interrupt and it feeds into the fast_pool, + a per-CPU variable (irq_randomness). + +- process_interrupt_randomness_pool() which feeds the fast_pool/ + irq_randomness into the entropy_store if enough entropy has been + gathered. + +This is a preparations step to ease PREEMPT_RT support. + +Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Link: https://lore.kernel.org/r/20211207121737.2347312-4-bigeasy@linutronix.de +--- + drivers/char/random.c | 47 +++++++++++++++++++++++++++-------------------- + 1 file changed, 27 insertions(+), 20 deletions(-) + +--- a/drivers/char/random.c ++++ b/drivers/char/random.c +@@ -1242,29 +1242,10 @@ static __u32 get_reg(struct fast_pool *f + return *ptr; + } + +-void add_interrupt_randomness(int irq) ++static void process_interrupt_randomness_pool(struct fast_pool *fast_pool) + { + struct entropy_store *r; +- struct fast_pool *fast_pool = this_cpu_ptr(&irq_randomness); +- struct pt_regs *regs = get_irq_regs(); + unsigned long now = jiffies; +- cycles_t cycles = random_get_entropy(); +- __u32 c_high, j_high; +- __u64 ip; +- +- if (cycles == 0) +- cycles = get_reg(fast_pool, regs); +- c_high = (sizeof(cycles) > 4) ? cycles >> 32 : 0; +- j_high = (sizeof(now) > 4) ? now >> 32 : 0; +- fast_pool->pool[0] ^= cycles ^ j_high ^ irq; +- fast_pool->pool[1] ^= now ^ c_high; +- ip = regs ? instruction_pointer(regs) : _RET_IP_; +- fast_pool->pool[2] ^= ip; +- fast_pool->pool[3] ^= (sizeof(ip) > 4) ? ip >> 32 : +- get_reg(fast_pool, regs); +- +- fast_mix(fast_pool); +- add_interrupt_bench(cycles); + + if (unlikely(crng_init == 0)) { + if ((fast_pool->count >= 64) && +@@ -1293,6 +1274,32 @@ void add_interrupt_randomness(int irq) + /* award one bit for the contents of the fast pool */ + credit_entropy_bits(r, 1); + } ++ ++void add_interrupt_randomness(int irq) ++{ ++ struct fast_pool *fast_pool = this_cpu_ptr(&irq_randomness); ++ struct pt_regs *regs = get_irq_regs(); ++ unsigned long now = jiffies; ++ cycles_t cycles = random_get_entropy(); ++ __u32 c_high, j_high; ++ __u64 ip; ++ ++ if (cycles == 0) ++ cycles = get_reg(fast_pool, regs); ++ c_high = (sizeof(cycles) > 4) ? cycles >> 32 : 0; ++ j_high = (sizeof(now) > 4) ? now >> 32 : 0; ++ fast_pool->pool[0] ^= cycles ^ j_high ^ irq; ++ fast_pool->pool[1] ^= now ^ c_high; ++ ip = regs ? instruction_pointer(regs) : _RET_IP_; ++ fast_pool->pool[2] ^= ip; ++ fast_pool->pool[3] ^= (sizeof(ip) > 4) ? ip >> 32 : ++ get_reg(fast_pool, regs); ++ ++ fast_mix(fast_pool); ++ add_interrupt_bench(cycles); ++ ++ process_interrupt_randomness_pool(fast_pool); ++} + EXPORT_SYMBOL_GPL(add_interrupt_randomness); + + #ifdef CONFIG_BLOCK diff --git a/patches/0004_random_move_the_fast_pool_reset_into_the_caller.patch b/patches/0004_random_move_the_fast_pool_reset_into_the_caller.patch new file mode 100644 index 000000000000..62e8ae44d325 --- /dev/null +++ b/patches/0004_random_move_the_fast_pool_reset_into_the_caller.patch @@ -0,0 +1,79 @@ +From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Subject: random: Move the fast_pool reset into the caller. +Date: Tue, 07 Dec 2021 13:17:36 +0100 + +The state of the fast_pool (number of added entropy, timestamp of last +addition) is reset after entropy has been consumed. + +Move the reset of the fast_pool into the caller. +This is a preparations step to ease PREEMPT_RT support. + +Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Link: https://lore.kernel.org/r/20211207121737.2347312-5-bigeasy@linutronix.de +--- + drivers/char/random.c | 29 +++++++++++++++-------------- + 1 file changed, 15 insertions(+), 14 deletions(-) + +--- a/drivers/char/random.c ++++ b/drivers/char/random.c +@@ -1242,37 +1242,35 @@ static __u32 get_reg(struct fast_pool *f + return *ptr; + } + +-static void process_interrupt_randomness_pool(struct fast_pool *fast_pool) ++static bool process_interrupt_randomness_pool(struct fast_pool *fast_pool) + { + struct entropy_store *r; +- unsigned long now = jiffies; + + if (unlikely(crng_init == 0)) { ++ bool pool_reset = false; ++ + if ((fast_pool->count >= 64) && + crng_fast_load((char *) fast_pool->pool, +- sizeof(fast_pool->pool))) { +- fast_pool->count = 0; +- fast_pool->last = now; +- } +- return; ++ sizeof(fast_pool->pool))) ++ pool_reset = true; ++ ++ return pool_reset; + } + + if ((fast_pool->count < 64) && +- !time_after(now, fast_pool->last + HZ)) +- return; ++ !time_after(jiffies, fast_pool->last + HZ)) ++ return false; + + r = &input_pool; + if (!spin_trylock(&r->lock)) +- return; ++ return false; + +- fast_pool->last = now; + __mix_pool_bytes(r, &fast_pool->pool, sizeof(fast_pool->pool)); + spin_unlock(&r->lock); + +- fast_pool->count = 0; +- + /* award one bit for the contents of the fast pool */ + credit_entropy_bits(r, 1); ++ return true; + } + + void add_interrupt_randomness(int irq) +@@ -1298,7 +1296,10 @@ void add_interrupt_randomness(int irq) + fast_mix(fast_pool); + add_interrupt_bench(cycles); + +- process_interrupt_randomness_pool(fast_pool); ++ if (process_interrupt_randomness_pool(fast_pool)) { ++ fast_pool->last = now; ++ fast_pool->count = 0; ++ } + } + EXPORT_SYMBOL_GPL(add_interrupt_randomness); + diff --git a/patches/0005_random_defer_processing_of_randomness_on_preempt_rt.patch b/patches/0005_random_defer_processing_of_randomness_on_preempt_rt.patch new file mode 100644 index 000000000000..6a1164699e75 --- /dev/null +++ b/patches/0005_random_defer_processing_of_randomness_on_preempt_rt.patch @@ -0,0 +1,116 @@ +From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Subject: random: Defer processing of randomness on PREEMPT_RT. +Date: Tue, 07 Dec 2021 13:17:37 +0100 + +On interrupt invocation, add_interrupt_randomness() adds entropy to its +per-CPU state and if it gathered enough of it then it will mix it into a +entropy_store. In order to do so, it needs to lock the pool by acquiring +entropy_store::lock which is a spinlock_t. This lock can not be acquired +on PREEMPT_RT with disabled interrupts because it is a sleeping lock. + +This lock could be made a raw_spinlock_t which will then allow to +acquire it with disabled interrupts on PREEMPT_RT. The lock is usually +hold for short amount of cycles while entropy is added to the pool and +the invocation from the IRQ handler has a try-lock which avoids spinning +on the lock if contended. The extraction of entropy (extract_buf()) +needs a few cycles more because it performs additionally few +SHA1 transformations. This takes around 5-10us on a testing box (E5-2650 +32 Cores, 2way NUMA) and is negligible. + +The frequent invocation of the IOCTLs RNDADDTOENTCNT and RNDRESEEDCRNG +on multiple CPUs in parallel leads to filling and depletion of the pool +which in turn results in heavy contention on the lock. The spinning with +disabled interrupts on multiple CPUs leads to latencies of at least +100us on the same machine which is no longer acceptable. + +Collect only the IRQ randomness in IRQ-context on PREEMPT_RT. +In threaded-IRQ context, make a copy of the per-CPU state with disabled +interrupts to ensure that it is not modified while duplicated. Pass the +copy to process_interrupt_randomness_pool() and reset the per-CPU +afterwards if needed. + +Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Link: https://lore.kernel.org/r/20211207121737.2347312-6-bigeasy@linutronix.de +--- + drivers/char/random.c | 39 ++++++++++++++++++++++++++++++++++++--- + include/linux/random.h | 1 + + kernel/irq/manage.c | 3 +++ + 3 files changed, 40 insertions(+), 3 deletions(-) + +--- a/drivers/char/random.c ++++ b/drivers/char/random.c +@@ -1273,6 +1273,32 @@ static bool process_interrupt_randomness + return true; + } + ++#ifdef CONFIG_PREEMPT_RT ++void process_interrupt_randomness(void) ++{ ++ struct fast_pool *cpu_pool; ++ struct fast_pool fast_pool; ++ ++ lockdep_assert_irqs_enabled(); ++ ++ migrate_disable(); ++ cpu_pool = this_cpu_ptr(&irq_randomness); ++ ++ local_irq_disable(); ++ memcpy(&fast_pool, cpu_pool, sizeof(fast_pool)); ++ local_irq_enable(); ++ ++ if (process_interrupt_randomness_pool(&fast_pool)) { ++ local_irq_disable(); ++ cpu_pool->last = jiffies; ++ cpu_pool->count = 0; ++ local_irq_enable(); ++ } ++ memzero_explicit(&fast_pool, sizeof(fast_pool)); ++ migrate_enable(); ++} ++#endif ++ + void add_interrupt_randomness(int irq) + { + struct fast_pool *fast_pool = this_cpu_ptr(&irq_randomness); +@@ -1296,9 +1322,16 @@ void add_interrupt_randomness(int irq) + fast_mix(fast_pool); + add_interrupt_bench(cycles); + +- if (process_interrupt_randomness_pool(fast_pool)) { +- fast_pool->last = now; +- fast_pool->count = 0; ++ /* ++ * On PREEMPT_RT the entropy can not be fed into the input_pool because ++ * it needs to acquire sleeping locks with disabled interrupts. ++ * This is deferred to the threaded handler. ++ */ ++ if (!IS_ENABLED(CONFIG_PREEMPT_RT)) { ++ if (process_interrupt_randomness_pool(fast_pool)) { ++ fast_pool->last = now; ++ fast_pool->count = 0; ++ } + } + } + EXPORT_SYMBOL_GPL(add_interrupt_randomness); +--- a/include/linux/random.h ++++ b/include/linux/random.h +@@ -36,6 +36,7 @@ static inline void add_latent_entropy(vo + extern void add_input_randomness(unsigned int type, unsigned int code, + unsigned int value) __latent_entropy; + extern void add_interrupt_randomness(int irq) __latent_entropy; ++extern void process_interrupt_randomness(void); + + extern void get_random_bytes(void *buf, int nbytes); + extern int wait_for_random_bytes(void); +--- a/kernel/irq/manage.c ++++ b/kernel/irq/manage.c +@@ -1281,6 +1281,9 @@ static int irq_thread(void *data) + if (action_ret == IRQ_WAKE_THREAD) + irq_wake_secondary(desc, action); + ++ if (IS_ENABLED(CONFIG_PREEMPT_RT)) ++ process_interrupt_randomness(); ++ + wake_threads_waitq(desc); + } + diff --git a/patches/0007_kernel_fork_only_cache_the_vmap_stack_in_finish_task_switch.patch b/patches/0007_kernel_fork_only_cache_the_vmap_stack_in_finish_task_switch.patch index 2fc9e09f60ab..723cb17a34ce 100644 --- a/patches/0007_kernel_fork_only_cache_the_vmap_stack_in_finish_task_switch.patch +++ b/patches/0007_kernel_fork_only_cache_the_vmap_stack_in_finish_task_switch.patch @@ -214,7 +214,7 @@ Link: https://lore.kernel.org/r/20211118143452.136421-8-bigeasy@linutronix.de return NULL; --- a/kernel/sched/core.c +++ b/kernel/sched/core.c -@@ -4879,8 +4879,11 @@ static struct rq *finish_task_switch(str +@@ -4881,8 +4881,11 @@ static struct rq *finish_task_switch(str if (prev->sched_class->task_dead) prev->sched_class->task_dead(prev); diff --git a/patches/0010-drm-i915-Drop-the-irqs_disabled-check.patch b/patches/0010-drm-i915-Drop-the-irqs_disabled-check.patch index 941c5be424a5..b236f2ab65eb 100644 --- a/patches/0010-drm-i915-Drop-the-irqs_disabled-check.patch +++ b/patches/0010-drm-i915-Drop-the-irqs_disabled-check.patch @@ -20,7 +20,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c -@@ -559,7 +559,6 @@ bool __i915_request_submit(struct i915_r +@@ -560,7 +560,6 @@ bool __i915_request_submit(struct i915_r RQ_TRACE(request, "\n"); @@ -28,7 +28,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> lockdep_assert_held(&engine->sched_engine->lock); /* -@@ -668,7 +667,6 @@ void __i915_request_unsubmit(struct i915 +@@ -669,7 +668,6 @@ void __i915_request_unsubmit(struct i915 */ RQ_TRACE(request, "\n"); diff --git a/patches/Add_localversion_for_-RT_release.patch b/patches/Add_localversion_for_-RT_release.patch index 8b36fd73f91a..6b1364508a7c 100644 --- a/patches/Add_localversion_for_-RT_release.patch +++ b/patches/Add_localversion_for_-RT_release.patch @@ -15,4 +15,4 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> --- /dev/null +++ b/localversion-rt @@ -0,0 +1 @@ -+-rt9 ++-rt10 diff --git a/patches/0002-drm-i915-Don-t-disable-interrupts-and-pretend-a-lock.patch b/patches/drm-i915-Don-t-disable-interrupts-and-pretend-a-lock.patch index 9fcdcb0b5b52..7d70f071b1db 100644 --- a/patches/0002-drm-i915-Don-t-disable-interrupts-and-pretend-a-lock.patch +++ b/patches/drm-i915-Don-t-disable-interrupts-and-pretend-a-lock.patch @@ -1,11 +1,12 @@ From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> -Date: Tue, 7 Jul 2020 12:25:11 +0200 -Subject: [PATCH 02/10] drm/i915: Don't disable interrupts and pretend a lock - as been acquired in __timeline_mark_lock(). +Date: Fri, 10 Dec 2021 21:44:17 +0100 +Subject: [PATCH] drm/i915: Don't disable interrupts and pretend a lock as been + acquired in __timeline_mark_lock(). This is a revert of commits d67739268cf0e ("drm/i915/gt: Mark up the nested engine-pm timeline lock as irqsafe") 6c69a45445af9 ("drm/i915/gt: Mark context->active_count as protected by timeline->mutex") + 6dcb85a0ad990 ("drm/i915: Hold irq-off for the entire fake lock period") The existing code leads to a different behaviour depending on whether lockdep is enabled or not. Any following lock that is acquired without @@ -20,12 +21,16 @@ check in intel_context_mark_active(). The other problem with disabling interrupts is that on PREEMPT_RT interrupts are also disabled which leads to problems for instance later during memory allocation. -Add a CONTEXT_IS_PARKED bit to intel_engine_cs and set_bit/clear_bit it +Add a CONTEXT_IS_PARKING bit to intel_engine_cs and set_bit/clear_bit it instead of mutex_acquire/mutex_release. Use test_bit in the two identified spots which relied on the lockdep annotation. Cc: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> +Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> +Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> +Link: https://patchwork.freedesktop.org/patch/msgid/YbO8Ie1Nj7XcQPNQ@linutronix.de --- drivers/gpu/drm/i915/gt/intel_context.h | 3 +- drivers/gpu/drm/i915/gt/intel_context_types.h | 1 @@ -41,7 +46,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> { - lockdep_assert_held(&ce->timeline->mutex); + lockdep_assert(lockdep_is_held(&ce->timeline->mutex) || -+ test_bit(CONTEXT_IS_PARKED, &ce->flags)); ++ test_bit(CONTEXT_IS_PARKING, &ce->flags)); ++ce->active_count; } @@ -51,7 +56,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> #define CONTEXT_LRCA_DIRTY 9 #define CONTEXT_GUC_INIT 10 #define CONTEXT_PERMA_PIN 11 -+#define CONTEXT_IS_PARKED 12 ++#define CONTEXT_IS_PARKING 12 struct { u64 timeout_us; @@ -110,7 +115,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> * it causing an underflow of the engine->wakeref. */ - flags = __timeline_mark_lock(ce); -+ set_bit(CONTEXT_IS_PARKED, &ce->flags); ++ set_bit(CONTEXT_IS_PARKING, &ce->flags); GEM_BUG_ON(atomic_read(&ce->timeline->active_count) < 0); rq = __i915_request_create(ce, GFP_NOWAIT); @@ -119,7 +124,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> result = false; out_unlock: - __timeline_mark_unlock(ce, flags); -+ clear_bit(CONTEXT_IS_PARKED, &ce->flags); ++ clear_bit(CONTEXT_IS_PARKING, &ce->flags); return result; } @@ -131,7 +136,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> return rcu_dereference_protected(rq->timeline, - lockdep_is_held(&rcu_access_pointer(rq->timeline)->mutex)); + lockdep_is_held(&rcu_access_pointer(rq->timeline)->mutex) || -+ test_bit(CONTEXT_IS_PARKED, &rq->context->flags)); ++ test_bit(CONTEXT_IS_PARKING, &rq->context->flags)); } static inline struct i915_gem_context * diff --git a/patches/mm-memcontro--Disable-on-PREEMPT_RT.patch b/patches/mm-memcontro--Disable-on-PREEMPT_RT.patch deleted file mode 100644 index a1b89d4f6770..000000000000 --- a/patches/mm-memcontro--Disable-on-PREEMPT_RT.patch +++ /dev/null @@ -1,25 +0,0 @@ -Subject: mm/memcontrol: Disable on PREEMPT_RT -From: Thomas Gleixner <tglx@linutronix.de> -Date: Sun, 25 Jul 2021 21:35:46 +0200 - -559271146efc ("mm/memcg: optimize user context object stock access") is a -classic example of optimizing for the cpu local BKL serialization without a -clear protection scope. - -Disable MEMCG on RT for now. - -Signed-off-by: Thomas Gleixner <tglx@linutronix.de> ---- - init/Kconfig | 1 + - 1 file changed, 1 insertion(+) - ---- a/init/Kconfig -+++ b/init/Kconfig -@@ -943,6 +943,7 @@ config PAGE_COUNTER - - config MEMCG - bool "Memory controller" -+ depends on !PREEMPT_RT - select PAGE_COUNTER - select EVENTFD - help diff --git a/patches/mm-memcontrol-Disable-on-PREEMPT_RT.patch b/patches/mm-memcontrol-Disable-on-PREEMPT_RT.patch new file mode 100644 index 000000000000..149c9877cfc0 --- /dev/null +++ b/patches/mm-memcontrol-Disable-on-PREEMPT_RT.patch @@ -0,0 +1,42 @@ +From: Thomas Gleixner <tglx@linutronix.de> +Date: Tue, 7 Dec 2021 16:52:09 +0100 +Subject: [PATCH] mm/memcontrol: Disable on PREEMPT_RT + +MEMCG has a few constructs which are not compatible with PREEMPT_RT's +requirements. This includes: +- relying on disabled interrupts from spin_lock_irqsave() locking for + something not related to lock itself (like the per-CPU counter). + +- explicitly disabling interrupts and acquiring a spinlock_t based lock + like in memcg_check_events() -> eventfd_signal(). + +- explicitly disabling interrupts and freeing memory like in + drain_obj_stock() -> obj_cgroup_put() -> obj_cgroup_release() -> + percpu_ref_exit(). + +Commit 559271146efc ("mm/memcg: optimize user context object stock +access") continued to optimize for the CPU local access which +complicates the PREEMPT_RT locking requirements further. + +Disable MEMCG on PREEMPT_RT until the whole situation can be evaluated +again. + +[ bigeasy: commit description. ] + +Signed-off-by: Thomas Gleixner <tglx@linutronix.de> +Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Link: https://lkml.kernel.org/r/20211207155208.eyre5svucpg7krxe@linutronix.de +--- + init/Kconfig | 1 + + 1 file changed, 1 insertion(+) + +--- a/init/Kconfig ++++ b/init/Kconfig +@@ -943,6 +943,7 @@ config PAGE_COUNTER + + config MEMCG + bool "Memory controller" ++ depends on !PREEMPT_RT + select PAGE_COUNTER + select EVENTFD + help diff --git a/patches/net-dev-Always-serialize-on-Qdisc-busylock-in-__dev_.patch b/patches/net-dev-Always-serialize-on-Qdisc-busylock-in-__dev_.patch new file mode 100644 index 000000000000..1c1e12aa4fd8 --- /dev/null +++ b/patches/net-dev-Always-serialize-on-Qdisc-busylock-in-__dev_.patch @@ -0,0 +1,39 @@ +From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Date: Wed, 30 Mar 2016 13:36:29 +0200 +Subject: [PATCH] net: dev: Always serialize on Qdisc::busylock in + __dev_xmit_skb() on PREEMPT_RT. + +The root-lock is dropped before dev_hard_start_xmit() is invoked and after +setting the __QDISC___STATE_RUNNING bit. If the Qdisc owner is preempted +by another sender/task with a higher priority then this new sender won't +be able to submit packets to the NIC directly instead they will be +enqueued into the Qdisc. The NIC will remain idle until the Qdisc owner +is scheduled again and finishes the job. + +By serializing every task on the ->busylock then the task will be +preempted by a sender only after the Qdisc has no owner. + +Always serialize on the busylock on PREEMPT_RT. + +Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Link: https://lkml.kernel.org/r/YbcmKeLngWW/pb1V@linutronix.de +--- + net/core/dev.c | 6 +++++- + 1 file changed, 5 insertions(+), 1 deletion(-) + +--- a/net/core/dev.c ++++ b/net/core/dev.c +@@ -3836,8 +3836,12 @@ static inline int __dev_xmit_skb(struct + * separate lock before trying to get qdisc main lock. + * This permits qdisc->running owner to get the lock more + * often and dequeue packets faster. ++ * On PREEMPT_RT it is possible to preempt the qdisc owner during xmit ++ * and then other tasks will only enqueue packets. The packets will be ++ * sent after the qdisc owner is scheduled again. To prevent this ++ * scenario the task always serialize on the lock. + */ +- contended = qdisc_is_running(q); ++ contended = IS_ENABLED(CONFIG_PREEMPT_RT) || qdisc_is_running(q); + if (unlikely(contended)) + spin_lock(&q->busylock); + diff --git a/patches/net__Dequeue_in_dev_cpu_dead_without_the_lock.patch b/patches/net__Dequeue_in_dev_cpu_dead_without_the_lock.patch deleted file mode 100644 index 132320db62cd..000000000000 --- a/patches/net__Dequeue_in_dev_cpu_dead_without_the_lock.patch +++ /dev/null @@ -1,35 +0,0 @@ -Subject: net: Dequeue in dev_cpu_dead() without the lock -From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> -Date: Wed Sep 16 16:15:39 2020 +0200 - -From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> - -Upstream uses skb_dequeue() to acquire lock of `input_pkt_queue'. The reason is -to synchronize against a remote CPU which still thinks that the CPU is online -enqueues packets to this CPU. -There are no guarantees that the packet is enqueued before the callback is run, -it just hope. -RT however complains about an not initialized lock because it uses another lock -for `input_pkt_queue' due to the IRQ-off nature of the context. - -Use the unlocked dequeue version for `input_pkt_queue'. - -Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> -Signed-off-by: Thomas Gleixner <tglx@linutronix.de> - - ---- - net/core/dev.c | 2 +- - 1 file changed, 1 insertion(+), 1 deletion(-) ---- ---- a/net/core/dev.c -+++ b/net/core/dev.c -@@ -11344,7 +11344,7 @@ static int dev_cpu_dead(unsigned int old - netif_rx_ni(skb); - input_queue_head_incr(oldsd); - } -- while ((skb = skb_dequeue(&oldsd->input_pkt_queue))) { -+ while ((skb = __skb_dequeue(&oldsd->input_pkt_queue))) { - netif_rx_ni(skb); - input_queue_head_incr(oldsd); - } diff --git a/patches/net__Remove_preemption_disabling_in_netif_rx.patch b/patches/net__Remove_preemption_disabling_in_netif_rx.patch index 291cf25e4a99..51fabc0faa4b 100644 --- a/patches/net__Remove_preemption_disabling_in_netif_rx.patch +++ b/patches/net__Remove_preemption_disabling_in_netif_rx.patch @@ -38,7 +38,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> --- --- a/net/core/dev.c +++ b/net/core/dev.c -@@ -4908,7 +4908,7 @@ static int netif_rx_internal(struct sk_b +@@ -4915,7 +4915,7 @@ static int netif_rx_internal(struct sk_b struct rps_dev_flow voidflow, *rflow = &voidflow; int cpu; @@ -47,7 +47,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> rcu_read_lock(); cpu = get_rps_cpu(skb->dev, skb, &rflow); -@@ -4918,14 +4918,14 @@ static int netif_rx_internal(struct sk_b +@@ -4925,14 +4925,14 @@ static int netif_rx_internal(struct sk_b ret = enqueue_to_backlog(skb, cpu, &rflow->last_qtail); rcu_read_unlock(); diff --git a/patches/net__Use_skbufhead_with_raw_lock.patch b/patches/net__Use_skbufhead_with_raw_lock.patch index 19dd0dca4ac7..cd7de80d0103 100644 --- a/patches/net__Use_skbufhead_with_raw_lock.patch +++ b/patches/net__Use_skbufhead_with_raw_lock.patch @@ -13,8 +13,8 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> --- include/linux/skbuff.h | 7 +++++++ - net/core/dev.c | 6 +++--- - 2 files changed, 10 insertions(+), 3 deletions(-) + net/core/dev.c | 8 ++++---- + 2 files changed, 11 insertions(+), 4 deletions(-) --- --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -58,7 +58,16 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> #endif } -@@ -11660,7 +11660,7 @@ static int __init net_dev_init(void) +@@ -11351,7 +11351,7 @@ static int dev_cpu_dead(unsigned int old + netif_rx_ni(skb); + input_queue_head_incr(oldsd); + } +- while ((skb = skb_dequeue(&oldsd->input_pkt_queue))) { ++ while ((skb = __skb_dequeue(&oldsd->input_pkt_queue))) { + netif_rx_ni(skb); + input_queue_head_incr(oldsd); + } +@@ -11667,7 +11667,7 @@ static int __init net_dev_init(void) INIT_WORK(flush, flush_backlog); diff --git a/patches/net__dev__always_take_qdiscs_busylock_in___dev_xmit_skb.patch b/patches/net__dev__always_take_qdiscs_busylock_in___dev_xmit_skb.patch deleted file mode 100644 index aee6836315b0..000000000000 --- a/patches/net__dev__always_take_qdiscs_busylock_in___dev_xmit_skb.patch +++ /dev/null @@ -1,39 +0,0 @@ -Subject: net: dev: always take qdisc's busylock in __dev_xmit_skb() -From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> -Date: Wed Mar 30 13:36:29 2016 +0200 - -From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> - -The root-lock is dropped before dev_hard_start_xmit() is invoked and after -setting the __QDISC___STATE_RUNNING bit. If this task is now pushed away -by a task with a higher priority then the task with the higher priority -won't be able to submit packets to the NIC directly instead they will be -enqueued into the Qdisc. The NIC will remain idle until the task(s) with -higher priority leave the CPU and the task with lower priority gets back -and finishes the job. - -If we take always the busylock we ensure that the RT task can boost the -low-prio task and submit the packet. - -Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> -Signed-off-by: Thomas Gleixner <tglx@linutronix.de> - - ---- - net/core/dev.c | 4 ++++ - 1 file changed, 4 insertions(+) ---- ---- a/net/core/dev.c -+++ b/net/core/dev.c -@@ -3839,7 +3839,11 @@ static inline int __dev_xmit_skb(struct - * This permits qdisc->running owner to get the lock more - * often and dequeue packets faster. - */ -+#ifdef CONFIG_PREEMPT_RT -+ contended = true; -+#else - contended = qdisc_is_running(q); -+#endif - if (unlikely(contended)) - spin_lock(&q->busylock); - diff --git a/patches/net_core__use_local_bh_disable_in_netif_rx_ni.patch b/patches/net_core__use_local_bh_disable_in_netif_rx_ni.patch index 21a37af595d4..07818d6cf5a6 100644 --- a/patches/net_core__use_local_bh_disable_in_netif_rx_ni.patch +++ b/patches/net_core__use_local_bh_disable_in_netif_rx_ni.patch @@ -23,7 +23,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> --- --- a/net/core/dev.c +++ b/net/core/dev.c -@@ -4967,11 +4967,9 @@ int netif_rx_ni(struct sk_buff *skb) +@@ -4974,11 +4974,9 @@ int netif_rx_ni(struct sk_buff *skb) trace_netif_rx_ni_entry(skb); diff --git a/patches/panic__skip_get_random_bytes_for_RT_FULL_in_init_oops_id.patch b/patches/panic__skip_get_random_bytes_for_RT_FULL_in_init_oops_id.patch deleted file mode 100644 index 74f1f2508330..000000000000 --- a/patches/panic__skip_get_random_bytes_for_RT_FULL_in_init_oops_id.patch +++ /dev/null @@ -1,30 +0,0 @@ -Subject: panic: skip get_random_bytes for RT_FULL in init_oops_id -From: Thomas Gleixner <tglx@linutronix.de> -Date: Tue Jul 14 14:26:34 2015 +0200 - -From: Thomas Gleixner <tglx@linutronix.de> - -Disable on -RT. If this is invoked from irq-context we will have problems -to acquire the sleeping lock. - -Signed-off-by: Thomas Gleixner <tglx@linutronix.de> - - ---- - kernel/panic.c | 2 ++ - 1 file changed, 2 insertions(+) ---- ---- a/kernel/panic.c -+++ b/kernel/panic.c -@@ -545,9 +545,11 @@ static u64 oops_id; - - static int init_oops_id(void) - { -+#ifndef CONFIG_PREEMPT_RT - if (!oops_id) - get_random_bytes(&oops_id, sizeof(oops_id)); - else -+#endif - oops_id++; - - return 0; diff --git a/patches/panic_remove_oops_id.patch b/patches/panic_remove_oops_id.patch new file mode 100644 index 000000000000..65ff798ce0c7 --- /dev/null +++ b/patches/panic_remove_oops_id.patch @@ -0,0 +1,55 @@ +From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Subject: panic: Remove oops_id. +Date: Thu, 02 Dec 2021 15:27:13 +0100 + +The oops id has been added as part of the end of trace marker for the +kerneloops.org project. The id is used to automatically identify duplicate +submissions of the same report. Identical looking reports with different +a id can be considered as the same oops occurred again. + +The early initialisation of the oops_id can create a warning if the +random core is not yet fully initialized. On PREEMPT_RT it is +problematic if the id is initialized on demand from non preemptible +context. + +The kernel oops project is not available since 2017. +Remove the oops_id and use 0 in the output in case parser rely on it. + +Link: https://bugs.debian.org/953172 +Cc: Arjan van de Ven <arjan@linux.intel.com> +Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Link: https://lore.kernel.org/r/20211202142713.ofadr43tawengfw4@linutronix.de +--- + kernel/panic.c | 19 +------------------ + 1 file changed, 1 insertion(+), 18 deletions(-) + +--- a/kernel/panic.c ++++ b/kernel/panic.c +@@ -538,26 +538,9 @@ void oops_enter(void) + trigger_all_cpu_backtrace(); + } + +-/* +- * 64-bit random ID for oopses: +- */ +-static u64 oops_id; +- +-static int init_oops_id(void) +-{ +- if (!oops_id) +- get_random_bytes(&oops_id, sizeof(oops_id)); +- else +- oops_id++; +- +- return 0; +-} +-late_initcall(init_oops_id); +- + static void print_oops_end_marker(void) + { +- init_oops_id(); +- pr_warn("---[ end trace %016llx ]---\n", (unsigned long long)oops_id); ++ pr_warn("---[ end trace %016llx ]---\n", 0ULL); + pr_flush(1000, true); + } + diff --git a/patches/random__Make_it_work_on_rt.patch b/patches/random__Make_it_work_on_rt.patch deleted file mode 100644 index f0ddaaa61cbe..000000000000 --- a/patches/random__Make_it_work_on_rt.patch +++ /dev/null @@ -1,173 +0,0 @@ -Subject: random: Make it work on rt -From: Thomas Gleixner <tglx@linutronix.de> -Date: Tue Aug 21 20:38:50 2012 +0200 - -From: Thomas Gleixner <tglx@linutronix.de> - -Delegate the random insertion to the forced threaded interrupt -handler. Store the return IP of the hard interrupt handler in the irq -descriptor and feed it into the random generator as a source of -entropy. - -Signed-off-by: Thomas Gleixner <tglx@linutronix.de> - - - ---- - arch/x86/kernel/cpu/mshyperv.c | 3 ++- - drivers/char/random.c | 11 +++++------ - drivers/hv/hyperv_vmbus.h | 1 + - drivers/hv/vmbus_drv.c | 5 ++++- - include/linux/irqdesc.h | 1 + - include/linux/random.h | 2 +- - kernel/irq/handle.c | 10 ++++++++-- - kernel/irq/manage.c | 6 ++++++ - 8 files changed, 28 insertions(+), 11 deletions(-) ---- ---- a/arch/x86/kernel/cpu/mshyperv.c -+++ b/arch/x86/kernel/cpu/mshyperv.c -@@ -75,11 +75,12 @@ void hv_remove_vmbus_handler(void) - DEFINE_IDTENTRY_SYSVEC(sysvec_hyperv_stimer0) - { - struct pt_regs *old_regs = set_irq_regs(regs); -+ u64 ip = regs ? instruction_pointer(regs) : 0; - - inc_irq_stat(hyperv_stimer0_count); - if (hv_stimer0_handler) - hv_stimer0_handler(); -- add_interrupt_randomness(HYPERV_STIMER0_VECTOR, 0); -+ add_interrupt_randomness(HYPERV_STIMER0_VECTOR, 0, ip); - ack_APIC_irq(); - - set_irq_regs(old_regs); ---- a/drivers/char/random.c -+++ b/drivers/char/random.c -@@ -1242,26 +1242,25 @@ static __u32 get_reg(struct fast_pool *f - return *ptr; - } - --void add_interrupt_randomness(int irq, int irq_flags) -+void add_interrupt_randomness(int irq, int irq_flags, __u64 ip) - { - struct entropy_store *r; - struct fast_pool *fast_pool = this_cpu_ptr(&irq_randomness); -- struct pt_regs *regs = get_irq_regs(); - unsigned long now = jiffies; - cycles_t cycles = random_get_entropy(); - __u32 c_high, j_high; -- __u64 ip; - - if (cycles == 0) -- cycles = get_reg(fast_pool, regs); -+ cycles = get_reg(fast_pool, NULL); - c_high = (sizeof(cycles) > 4) ? cycles >> 32 : 0; - j_high = (sizeof(now) > 4) ? now >> 32 : 0; - fast_pool->pool[0] ^= cycles ^ j_high ^ irq; - fast_pool->pool[1] ^= now ^ c_high; -- ip = regs ? instruction_pointer(regs) : _RET_IP_; -+ if (!ip) -+ ip = _RET_IP_; - fast_pool->pool[2] ^= ip; - fast_pool->pool[3] ^= (sizeof(ip) > 4) ? ip >> 32 : -- get_reg(fast_pool, regs); -+ get_reg(fast_pool, NULL); - - fast_mix(fast_pool); - add_interrupt_bench(cycles); ---- a/drivers/hv/hyperv_vmbus.h -+++ b/drivers/hv/hyperv_vmbus.h -@@ -19,6 +19,7 @@ - #include <linux/atomic.h> - #include <linux/hyperv.h> - #include <linux/interrupt.h> -+#include <linux/irq.h> - - #include "hv_trace.h" - ---- a/drivers/hv/vmbus_drv.c -+++ b/drivers/hv/vmbus_drv.c -@@ -22,6 +22,7 @@ - #include <linux/clockchips.h> - #include <linux/cpu.h> - #include <linux/sched/task_stack.h> -+#include <linux/irq.h> - - #include <linux/delay.h> - #include <linux/notifier.h> -@@ -1337,6 +1338,8 @@ static void vmbus_isr(void) - void *page_addr = hv_cpu->synic_event_page; - struct hv_message *msg; - union hv_synic_event_flags *event; -+ struct pt_regs *regs = get_irq_regs(); -+ u64 ip = regs ? instruction_pointer(regs) : 0; - bool handled = false; - - if (unlikely(page_addr == NULL)) -@@ -1381,7 +1384,7 @@ static void vmbus_isr(void) - tasklet_schedule(&hv_cpu->msg_dpc); - } - -- add_interrupt_randomness(vmbus_interrupt, 0); -+ add_interrupt_randomness(vmbus_interrupt, 0, ip); - } - - static irqreturn_t vmbus_percpu_isr(int irq, void *dev_id) ---- a/include/linux/irqdesc.h -+++ b/include/linux/irqdesc.h -@@ -68,6 +68,7 @@ struct irq_desc { - unsigned int irqs_unhandled; - atomic_t threads_handled; - int threads_handled_last; -+ u64 random_ip; - raw_spinlock_t lock; - struct cpumask *percpu_enabled; - const struct cpumask *percpu_affinity; ---- a/include/linux/random.h -+++ b/include/linux/random.h -@@ -35,7 +35,7 @@ static inline void add_latent_entropy(vo - - extern void add_input_randomness(unsigned int type, unsigned int code, - unsigned int value) __latent_entropy; --extern void add_interrupt_randomness(int irq, int irq_flags) __latent_entropy; -+extern void add_interrupt_randomness(int irq, int irq_flags, __u64 ip) __latent_entropy; - - extern void get_random_bytes(void *buf, int nbytes); - extern int wait_for_random_bytes(void); ---- a/kernel/irq/handle.c -+++ b/kernel/irq/handle.c -@@ -192,12 +192,18 @@ irqreturn_t __handle_irq_event_percpu(st - - irqreturn_t handle_irq_event_percpu(struct irq_desc *desc) - { -- irqreturn_t retval; -+ struct pt_regs *regs = get_irq_regs(); -+ u64 ip = regs ? instruction_pointer(regs) : 0; - unsigned int flags = 0; -+ irqreturn_t retval; - - retval = __handle_irq_event_percpu(desc, &flags); - -- add_interrupt_randomness(desc->irq_data.irq, flags); -+#ifdef CONFIG_PREEMPT_RT -+ desc->random_ip = ip; -+#else -+ add_interrupt_randomness(desc->irq_data.irq, flags, ip); -+#endif - - if (!irq_settings_no_debug(desc)) - note_interrupt(desc, retval); ---- a/kernel/irq/manage.c -+++ b/kernel/irq/manage.c -@@ -1281,6 +1281,12 @@ static int irq_thread(void *data) - if (action_ret == IRQ_WAKE_THREAD) - irq_wake_secondary(desc, action); - -+ if (IS_ENABLED(CONFIG_PREEMPT_RT)) { -+ migrate_disable(); -+ add_interrupt_randomness(action->irq, 0, -+ desc->random_ip ^ (unsigned long) action); -+ migrate_enable(); -+ } - wake_threads_waitq(desc); - } - diff --git a/patches/serial__8250__implement_write_atomic.patch b/patches/serial__8250__implement_write_atomic.patch index c7bf2744e82b..b08a974b1ef9 100644 --- a/patches/serial__8250__implement_write_atomic.patch +++ b/patches/serial__8250__implement_write_atomic.patch @@ -276,7 +276,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> serial8250_rpm_put(up); } -@@ -2132,14 +2132,7 @@ static void serial8250_put_poll_char(str +@@ -2125,14 +2125,7 @@ static void serial8250_put_poll_char(str struct uart_8250_port *up = up_to_u8250p(port); serial8250_rpm_get(up); @@ -292,7 +292,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> wait_for_xmitr(up, BOTH_EMPTY); /* -@@ -2152,7 +2145,7 @@ static void serial8250_put_poll_char(str +@@ -2145,7 +2138,7 @@ static void serial8250_put_poll_char(str * and restore the IER */ wait_for_xmitr(up, BOTH_EMPTY); @@ -301,7 +301,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> serial8250_rpm_put(up); } -@@ -2455,7 +2448,7 @@ void serial8250_do_shutdown(struct uart_ +@@ -2448,7 +2441,7 @@ void serial8250_do_shutdown(struct uart_ */ spin_lock_irqsave(&port->lock, flags); up->ier = 0; @@ -310,7 +310,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> spin_unlock_irqrestore(&port->lock, flags); synchronize_irq(port->irq); -@@ -2837,7 +2830,7 @@ serial8250_do_set_termios(struct uart_po +@@ -2830,7 +2823,7 @@ serial8250_do_set_termios(struct uart_po if (up->capabilities & UART_CAP_RTOIE) up->ier |= UART_IER_RTOIE; @@ -319,7 +319,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> if (up->capabilities & UART_CAP_EFR) { unsigned char efr = 0; -@@ -3303,7 +3296,7 @@ EXPORT_SYMBOL_GPL(serial8250_set_default +@@ -3296,7 +3289,7 @@ EXPORT_SYMBOL_GPL(serial8250_set_default #ifdef CONFIG_SERIAL_8250_CONSOLE @@ -328,7 +328,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> { struct uart_8250_port *up = up_to_u8250p(port); -@@ -3311,6 +3304,18 @@ static void serial8250_console_putchar(s +@@ -3304,6 +3297,18 @@ static void serial8250_console_putchar(s serial_port_out(port, UART_TX, ch); } @@ -347,7 +347,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> /* * Restore serial console when h/w power-off detected */ -@@ -3332,6 +3337,32 @@ static void serial8250_console_restore(s +@@ -3325,6 +3330,32 @@ static void serial8250_console_restore(s serial8250_out_MCR(up, UART_MCR_DTR | UART_MCR_RTS); } @@ -380,7 +380,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> /* * Print a string to the serial port trying not to disturb * any possible real use of the port... -@@ -3348,24 +3379,12 @@ void serial8250_console_write(struct uar +@@ -3341,24 +3372,12 @@ void serial8250_console_write(struct uar struct uart_port *port = &up->port; unsigned long flags; unsigned int ier; @@ -407,7 +407,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> /* check scratch reg to see if port powered off during system sleep */ if (up->canary && (up->canary != serial_port_in(port, UART_SCR))) { -@@ -3379,7 +3398,9 @@ void serial8250_console_write(struct uar +@@ -3372,7 +3391,9 @@ void serial8250_console_write(struct uar mdelay(port->rs485.delay_rts_before_send); } @@ -417,7 +417,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> /* * Finally, wait for transmitter to become empty -@@ -3392,8 +3413,7 @@ void serial8250_console_write(struct uar +@@ -3385,8 +3406,7 @@ void serial8250_console_write(struct uar if (em485->tx_stopped) up->rs485_stop_tx(up); } @@ -427,7 +427,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> /* * The receive handling will happen properly because the -@@ -3405,8 +3425,7 @@ void serial8250_console_write(struct uar +@@ -3398,8 +3418,7 @@ void serial8250_console_write(struct uar if (up->msr_saved_flags) serial8250_modem_status(up); @@ -437,7 +437,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> } static unsigned int probe_baud(struct uart_port *port) -@@ -3426,6 +3445,7 @@ static unsigned int probe_baud(struct ua +@@ -3419,6 +3438,7 @@ static unsigned int probe_baud(struct ua int serial8250_console_setup(struct uart_port *port, char *options, bool probe) { @@ -445,7 +445,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> int baud = 9600; int bits = 8; int parity = 'n'; -@@ -3435,6 +3455,8 @@ int serial8250_console_setup(struct uart +@@ -3428,6 +3448,8 @@ int serial8250_console_setup(struct uart if (!port->iobase && !port->membase) return -ENODEV; diff --git a/patches/series b/patches/series index 3a51bfbf1013..0c3315c3efdc 100644 --- a/patches/series +++ b/patches/series @@ -32,6 +32,21 @@ net-Write-lock-dev_base_lock-without-disabling-botto.patch fs-namespace-Boost-the-mount_lock.lock-owner-instead.patch 0001-blk-mq-Add-blk_mq_complete_request_direct.patch 0002-mmc-core-Use-blk_mq_complete_request_direct.patch +drm-i915-Don-t-disable-interrupts-and-pretend-a-lock.patch +kernel-locking-Use-a-pointer-in-ww_mutex_trylock.patch + +# lockdep +0001-sched-Trigger-warning-if-migration_disabled-counter-.patch +0002-locking-Remove-rt_rwlock_is_contended.patch +0003-locking-rtmutex-Squash-self-deadlock-check-for-ww_rt.patch +0004-locking-rtmutex-Add-rt_mutex_lock_nest_lock-and-rt_m.patch +0005-lockdep-Remove-softirq-accounting-on-PREEMPT_RT.patch +0006-lockdep-selftests-Avoid-using-local_lock_-acquire-re.patch +0007-lockdep-selftests-Unbalanced-migrate_disable-rcu_rea.patch +0008-lockdep-selftests-Skip-the-softirq-related-tests-on-.patch +0009-lockdep-selftests-Adapt-ww-tests-for-PREEMPT_RT.patch +0010-x86-mm-Include-spinlock_t-definition-in-pgtable.patch +0011-locking-Allow-to-include-asm-spinlock_types.h-from-l.patch ########################################################################### # Posted @@ -39,8 +54,11 @@ fs-namespace-Boost-the-mount_lock.lock-owner-instead.patch irq_poll-Use-raise_softirq_irqoff-in-cpu_dead-notifi.patch smp_wake_ksoftirqd_on_preempt_rt_instead_do_softirq.patch fscache-Use-only-one-fscache_object_cong_wait.patch -kernel-locking-Use-a-pointer-in-ww_mutex_trylock.patch tcp-Don-t-acquire-inet_listen_hashbucket-lock-with-d.patch +panic_remove_oops_id.patch +mm-memcontrol-Disable-on-PREEMPT_RT.patch +net-dev-Always-serialize-on-Qdisc-busylock-in-__dev_.patch +u64_stats-Disable-preemption-on-32bit-UP-SMP-PREEMPT.patch # sched 0001_kernel_fork_redo_ifdefs_around_task_s_handling.patch @@ -52,18 +70,12 @@ tcp-Don-t-acquire-inet_listen_hashbucket-lock-with-d.patch 0007_kernel_fork_only_cache_the_vmap_stack_in_finish_task_switch.patch 0008_kernel_fork_use_is_enabled_in_account_kernel_stack.patch -# lockdep -0001-sched-Trigger-warning-if-migration_disabled-counter-.patch -0002-locking-Remove-rt_rwlock_is_contended.patch -0003-locking-rtmutex-Squash-self-deadlock-check-for-ww_rt.patch -0004-locking-rtmutex-Add-rt_mutex_lock_nest_lock-and-rt_m.patch -0005-lockdep-Remove-softirq-accounting-on-PREEMPT_RT.patch -0006-lockdep-selftests-Avoid-using-local_lock_-acquire-re.patch -0007-lockdep-selftests-Unbalanced-migrate_disable-rcu_rea.patch -0008-lockdep-selftests-Skip-the-softirq-related-tests-on-.patch -0009-lockdep-selftests-Adapt-ww-tests-for-PREEMPT_RT.patch -0010-x86-mm-Include-spinlock_t-definition-in-pgtable.patch -0011-locking-Allow-to-include-asm-spinlock_types.h-from-l.patch +# random +0001_random_remove_unused_irq_flags_argument_from_add_interrupt_randomness.patch +0002_irq_remove_unsued_flags_argument_from___handle_irq_event_percpu.patch +0003_random_split_add_interrupt_randomness.patch +0004_random_move_the_fast_pool_reset_into_the_caller.patch +0005_random_defer_processing_of_randomness_on_preempt_rt.patch ########################################################################### # Post @@ -101,20 +113,6 @@ crypto__cryptd_-_add_a_lock_instead_preempt_disable_local_bh_disable.patch softirq__Check_preemption_after_reenabling_interrupts.patch ########################################################################### -# mm: Assorted RT bits. Need care -########################################################################### -u64_stats__Disable_preemption_on_32bit-UP_SMP_with_RT_during_updates.patch - -########################################################################### -# Disable memcontrol for now. The protection scopes are FUBARed -########################################################################### -mm-memcontro--Disable-on-PREEMPT_RT.patch -#mm_memcontrol__Disable_preemption_in___mod_memcg_lruvec_state.patch -#mm__memcontrol__Replace_disable-IRQ_locking_with_a_local_lock.patch -#mm_memcontrol__Dont_call_schedule_work_on_in_preemption_disabled_context.patch -#mm_memcontrol__Replace_local_irq_disable_with_local_locks.patch - -########################################################################### # ptrace: Revisit ########################################################################### signal__Revert_ptrace_preempt_magic.patch @@ -136,19 +134,10 @@ rcu__Delay_RCU-selftests.patch ########################################################################### net_core__use_local_bh_disable_in_netif_rx_ni.patch net__Use_skbufhead_with_raw_lock.patch -net__Dequeue_in_dev_cpu_dead_without_the_lock.patch -net__dev__always_take_qdiscs_busylock_in___dev_xmit_skb.patch - -########################################################################### -# randomness: -########################################################################### -panic__skip_get_random_bytes_for_RT_FULL_in_init_oops_id.patch -random__Make_it_work_on_rt.patch ########################################################################### # DRM: ########################################################################### -0002-drm-i915-Don-t-disable-interrupts-and-pretend-a-lock.patch 0003-drm-i915-Use-preempt_disable-enable_rt-where-recomme.patch 0004-drm-i915-Don-t-disable-interrupts-on-PREEMPT_RT-duri.patch 0005-drm-i915-Don-t-check-for-atomic-context-on-PREEMPT_R.patch diff --git a/patches/softirq-Use-a-dedicated-thread-for-timer-wakeups.patch b/patches/softirq-Use-a-dedicated-thread-for-timer-wakeups.patch index 92d63a22cd6e..12590bf0c60d 100644 --- a/patches/softirq-Use-a-dedicated-thread-for-timer-wakeups.patch +++ b/patches/softirq-Use-a-dedicated-thread-for-timer-wakeups.patch @@ -69,7 +69,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> static inline struct task_struct *this_cpu_ksoftirqd(void) --- a/kernel/softirq.c +++ b/kernel/softirq.c -@@ -623,6 +623,22 @@ static inline void tick_irq_exit(void) +@@ -624,6 +624,22 @@ static inline void tick_irq_exit(void) #endif } @@ -92,7 +92,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> static inline void __irq_exit_rcu(void) { #ifndef __ARCH_IRQ_EXIT_IRQS_DISABLED -@@ -634,6 +650,8 @@ static inline void __irq_exit_rcu(void) +@@ -635,6 +651,8 @@ static inline void __irq_exit_rcu(void) preempt_count_sub(HARDIRQ_OFFSET); if (!in_interrupt() && local_softirq_pending()) invoke_softirq(); @@ -101,7 +101,7 @@ Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> tick_irq_exit(); } -@@ -962,11 +980,69 @@ static struct smp_hotplug_thread softirq +@@ -963,11 +981,69 @@ static struct smp_hotplug_thread softirq .thread_comm = "ksoftirqd/%u", }; diff --git a/patches/softirq__Check_preemption_after_reenabling_interrupts.patch b/patches/softirq__Check_preemption_after_reenabling_interrupts.patch index 79c594270826..3251dc4fd512 100644 --- a/patches/softirq__Check_preemption_after_reenabling_interrupts.patch +++ b/patches/softirq__Check_preemption_after_reenabling_interrupts.patch @@ -61,7 +61,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> } EXPORT_SYMBOL(__dev_kfree_skb_irq); -@@ -4668,6 +4670,7 @@ static int enqueue_to_backlog(struct sk_ +@@ -4675,6 +4677,7 @@ static int enqueue_to_backlog(struct sk_ rps_unlock(sd); local_irq_restore(flags); @@ -69,7 +69,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> atomic_long_inc(&skb->dev->rx_dropped); kfree_skb(skb); -@@ -6412,12 +6415,14 @@ static void net_rps_action_and_irq_enabl +@@ -6419,12 +6422,14 @@ static void net_rps_action_and_irq_enabl sd->rps_ipi_list = NULL; local_irq_enable(); @@ -84,7 +84,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> } static bool sd_has_rps_ipi_waiting(struct softnet_data *sd) -@@ -6495,6 +6500,7 @@ void __napi_schedule(struct napi_struct +@@ -6502,6 +6507,7 @@ void __napi_schedule(struct napi_struct local_irq_save(flags); ____napi_schedule(this_cpu_ptr(&softnet_data), n); local_irq_restore(flags); @@ -92,7 +92,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> } EXPORT_SYMBOL(__napi_schedule); -@@ -11326,6 +11332,7 @@ static int dev_cpu_dead(unsigned int old +@@ -11333,6 +11339,7 @@ static int dev_cpu_dead(unsigned int old raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_enable(); diff --git a/patches/tcp-Don-t-acquire-inet_listen_hashbucket-lock-with-d.patch b/patches/tcp-Don-t-acquire-inet_listen_hashbucket-lock-with-d.patch index bf4d4ed3ec69..a0065abf3e7d 100644 --- a/patches/tcp-Don-t-acquire-inet_listen_hashbucket-lock-with-d.patch +++ b/patches/tcp-Don-t-acquire-inet_listen_hashbucket-lock-with-d.patch @@ -27,11 +27,11 @@ local_bh_disable() + spin_lock(&ilb->lock): spin_lock(&ilb->lock); acquire(&ilb->lock); -Reverse order: spin_lock(&ilb->lock) + local_bh_disable(): +Reverse order: spin_lock(&ilb2->lock) + local_bh_disable(): tcp_seq_next() listening_get_next() - spin_lock(&ilb->lock); - acquire(&ilb->lock); + spin_lock(&ilb2->lock); + acquire(&ilb2->lock); tcp4_seq_show() get_tcp4_sock() diff --git a/patches/u64_stats__Disable_preemption_on_32bit-UP_SMP_with_RT_during_updates.patch b/patches/u64_stats-Disable-preemption-on-32bit-UP-SMP-PREEMPT.patch index 417c3a241021..c89279153052 100644 --- a/patches/u64_stats__Disable_preemption_on_32bit-UP_SMP_with_RT_during_updates.patch +++ b/patches/u64_stats-Disable-preemption-on-32bit-UP-SMP-PREEMPT.patch @@ -1,25 +1,28 @@ -Subject: u64_stats: Disable preemption on 32bit-UP/SMP with RT during updates From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> -Date: Mon Aug 17 12:28:10 2020 +0200 +Date: Mon, 17 Aug 2020 12:28:10 +0200 +Subject: [PATCH] u64_stats: Disable preemption on 32bit UP+SMP PREEMPT_RT + during updates. -From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> - -On RT the seqcount_t is required even on UP because the softirq can be -preempted. The IRQ handler is threaded so it is also preemptible. +On PREEMPT_RT the seqcount_t for synchronisation is required on 32bit +architectures even on UP because the softirq (and the threaded IRQ handler) can +be preempted. -Disable preemption on 32bit-RT during value updates. There is no need to -disable interrupts on RT because the handler is run threaded. Therefore -disabling preemption is enough to guarantee that the update is not -interruped. +With the seqcount_t for synchronisation, a reader with higher priority can +preempt the writer and then spin endlessly in read_seqcount_begin() while the +writer can't make progress. -Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> -Signed-off-by: Thomas Gleixner <tglx@linutronix.de> +To avoid such a lock up on PREEMPT_RT the writer must disable preemption during +the update. There is no need to disable interrupts because no writer is using +this API in hard-IRQ context on PREEMPT_RT. +Disable preemption on 32bit-RT within the u64_stats write section. +Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> +Link: https://lore.kernel.org/r/YbO4x7vRoDGUWxrv@linutronix.de --- include/linux/u64_stats_sync.h | 42 +++++++++++++++++++++++++++-------------- 1 file changed, 28 insertions(+), 14 deletions(-) ---- + --- a/include/linux/u64_stats_sync.h +++ b/include/linux/u64_stats_sync.h @@ -66,7 +66,7 @@ @@ -27,7 +30,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> struct u64_stats_sync { -#if BITS_PER_LONG==32 && defined(CONFIG_SMP) -+#if BITS_PER_LONG==32 && (defined(CONFIG_SMP) || defined(CONFIG_PREEMPT_RT)) ++#if BITS_PER_LONG == 32 && (defined(CONFIG_SMP) || defined(CONFIG_PREEMPT_RT)) seqcount_t seq; #endif }; diff --git a/patches/x86__kvm_Require_const_tsc_for_RT.patch b/patches/x86__kvm_Require_const_tsc_for_RT.patch index 4d855fa5c859..1d473b6265e4 100644 --- a/patches/x86__kvm_Require_const_tsc_for_RT.patch +++ b/patches/x86__kvm_Require_const_tsc_for_RT.patch @@ -18,7 +18,7 @@ Signed-off-by: Thomas Gleixner <tglx@linutronix.de> --- --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c -@@ -8605,6 +8605,12 @@ int kvm_arch_init(void *opaque) +@@ -8655,6 +8655,12 @@ int kvm_arch_init(void *opaque) goto out; } |