- 17 Jun, 2019 27 commits
-
-
Peter Zijlstra authored
Nadav Amit reported that commit: b59167ac ("x86/percpu: Fix this_cpu_read()") added a bunch of constraints to all sorts of code; and while some of that was correct and desired, some of that seems superfluous. The thing is, the this_cpu_*() operations are defined IRQ-safe, this means the values are subject to change from IRQs, and thus must be reloaded. Also, the generic form: local_irq_save() __this_cpu_read() local_irq_restore() would not allow the re-use of previous values; if by nothing else, then the barrier()s implied by local_irq_*(). Which raises the point that percpu_from_op() and the others also need that volatile. OTOH __this_cpu_*() operations are not IRQ-safe and assume external preempt/IRQ disabling and could thus be allowed more room for optimization. This makes the this_cpu_*() vs __this_cpu_*() behaviour more consistent with other architectures. $ ./compare.sh defconfig-build defconfig-build1 vmlinux.o x86_pmu_cancel_txn 80 71 -9,+0 __text_poke 919 964 +45,+0 do_user_addr_fault 1082 1058 -24,+0 __do_page_fault 1194 1178 -16,+0 do_exit 2995 3027 -43,+75 process_one_work 1008 989 -67,+48 finish_task_switch 524 505 -19,+0 __schedule_bug 103 98 -59,+54 __schedule_bug 103 98 -59,+54 __sched_setscheduler 2015 2030 +15,+0 freeze_processes 203 230 +31,-4 rcu_gp_kthread_wake 106 99 -7,+0 rcu_core 1841 1834 -7,+0 call_timer_fn 298 286 -12,+0 can_stop_idle_tick 146 139 -31,+24 perf_pending_event 253 239 -14,+0 shmem_alloc_page 209 213 +4,+0 __alloc_pages_slowpath 3284 3269 -15,+0 umount_tree 671 694 +23,+0 advance_transaction 803 798 -5,+0 con_put_char 71 51 -20,+0 xhci_urb_enqueue 1302 1295 -7,+0 xhci_urb_enqueue 1302 1295 -7,+0 tcp_sacktag_write_queue 2130 2075 -55,+0 tcp_try_undo_loss 229 208 -21,+0 tcp_v4_inbound_md5_hash 438 411 -31,+4 tcp_v4_inbound_md5_hash 438 411 -31,+4 tcp_v6_inbound_md5_hash 469 411 -33,-25 tcp_v6_inbound_md5_hash 469 411 -33,-25 restricted_pointer 434 420 -14,+0 irq_exit 162 154 -8,+0 get_perf_callchain 638 624 -14,+0 rt_mutex_trylock 169 156 -13,+0 avc_has_extended_perms 1092 1089 -3,+0 avc_has_perm_noaudit 309 306 -3,+0 __perf_sw_event 138 122 -16,+0 perf_swevent_get_recursion_context 116 102 -14,+0 __local_bh_enable_ip 93 72 -21,+0 xfrm_input 4175 4161 -14,+0 avc_has_perm 446 443 -3,+0 vm_events_fold_cpu 57 56 -1,+0 vfree 68 61 -7,+0 freeze_processes 203 230 +31,-4 _local_bh_enable 44 30 -14,+0 ip_do_fragment 1982 1944 -38,+0 do_exit 2995 3027 -43,+75 __do_softirq 742 724 -18,+0 cpu_init 1510 1489 -21,+0 account_system_time 80 79 -1,+0 total 12985281 12984819 -742,+280 Reported-by:
Nadav Amit <nadav.amit@gmail.com> Signed-off-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Link: https://lkml.kernel.org/r/20181206112433.GB13675@hirez.programming.kicks-ass.netSigned-off-by:
Ingo Molnar <mingo@kernel.org>
-
Waiman Long authored
The upper bits of the count field is used as reader count. When sufficient number of active readers are present, the most significant bit will be set and the count becomes negative. If the number of active readers keep on piling up, we may eventually overflow the reader counts. This is not likely to happen unless the number of bits reserved for reader count is reduced because those bits are need for other purpose. To prevent this count overflow from happening, the most significant bit is now treated as a guard bit (RWSEM_FLAG_READFAIL). Read-lock attempts will now fail for both the fast and slow paths whenever this bit is set. So all those extra readers will be put to sleep in the wait list. Wakeup will not happen until the reader count reaches 0. Signed-off-by:
Waiman Long <longman@redhat.com> Signed-off-by:
-
Waiman Long authored
Reader optimistic spinning is helpful when the reader critical section is short and there aren't that many readers around. It makes readers relatively more preferred than writers. When a writer times out spinning on a reader-owned lock and set the nospinnable bits, there are two main reasons for that. 1) The reader critical section is long, perhaps the task sleeps after acquiring the read lock. 2) There are just too many readers contending the lock causing it to take a while to service all of them. In the former case, long reader critical section will impede the progress of writers which is usually more important for system performance. In the later case, reader optimistic spinning tends to make the reader groups that contain readers that acquire the lock together smaller leading to more of them. That may hurt performance in some cases. In other words, the setting of nonspinnable bits indicates that reader optimistic spinning may not be helpful for those workloads that cause it. Therefore, any writers that have observed the setting of the writer nonspinnable bit for a given rwsem after they fail to acquire the lock via optimistic spinning will set the reader nonspinnable bit once they acquire the write lock. Similarly, readers that observe the setting of reader nonspinnable bit at slowpath entry will also set the reader nonspinnable bit when they acquire the read lock via the wakeup path. Once the reader nonspinnable bit is on, it will only be reset when a writer is able to acquire the rwsem in the fast path or somehow a reader or writer in the slowpath doesn't observe the nonspinable bit. This is to discourage reader optmistic spinning on that particular rwsem and make writers more preferred. This adaptive disabling of reader optimistic spinning will alleviate some of the negative side effect of this feature. In addition, this patch tries to make readers in the spinning queue follow the phase-fair principle after quitting optimistic spinning by checking if another reader has somehow acquired a read lock after this reader enters the optimistic spinning queue. If so and the rwsem is still reader-owned, this reader is in the right read-phase and can attempt to acquire the lock. On a 2-socket 40-core 80-thread Skylake system, the page_fault1 test of the will-it-scale benchmark was run with various number of threads. The number of operations done before reader optimistic spinning patches, this patch and after this patch were: Threads Before rspin Before patch After patch %change ------- ------------ ------------ ----------- ------- 20 5541068 5345484 5455667 -3.5%/ +2.1% 40 10185150 7292313 9219276 -28.5%/+26.4% 60 8196733 6460517 7181209 -21.2%/+11.2% 80 9508864 6739559 8107025 -29.1%/+20.3% This patch doesn't recover all the lost performance, but it is more than half. Given the fact that reader optimistic spinning does benefit some workloads, this is a good compromise. Using the rwsem locking microbenchmark with very short critical section, this patch doesn't have too much impact on locking performance as shown by the locking rates (kops/s) below with equal numbers of readers and writers before and after this patch: # of Threads Pre-patch Post-patch ------------ --------- ---------- 2 4,730 4,969 4 4,814 4,786 8 4,866 4,815 16 4,715 4,511 32 3,338 3,500 64 3,212 3,389 80 3,110 3,044 When running the locking microbenchmark with 40 dedicated reader and writer threads, however, the reader performance is curtailed to favor the writer. Before patch: 40 readers, Iterations Min/Mean/Max = 204,026/234,309/254,816 40 writers, Iterations Min/Mean/Max = 88,515/95,884/115,644 After patch: 40 readers, Iterations Min/Mean/Max = 33,813/35,260/36,791 40 writers, Iterations Min/Mean/Max = 95,368/96,565/97,798 Signed-off-by: -
Waiman Long authored
When the rwsem is owned by reader, writers stop optimistic spinning simply because there is no easy way to figure out if all the readers are actively running or not. However, there are scenarios where the readers are unlikely to sleep and optimistic spinning can help performance. This patch provides a simple mechanism for spinning on a reader-owned rwsem by a writer. It is a time threshold based spinning where the allowable spinning time can vary from 10us to 25us depending on the condition of the rwsem. When the time threshold is exceeded, the nonspinnable bits will be set in the owner field to indicate that no more optimistic spinning will be allowed on this rwsem until it becomes writer owned again. Not even readers is allowed to acquire the reader-locked rwsem by optimistic spinning for fairness. We also want a writer to acquire the lock after the readers hold the lock for a relatively long time. In order to give preference to writers under such a circumstance, the single RWSEM_NONSPINNABLE bit is now split into two - one for reader and one for writer. When optimistic spinning is disabled, both bits will be set. When the reader count drop down to 0, the writer nonspinnable bit will be cleared to allow writers to spin on the lock, but not the readers. When a writer acquires the lock, it will write its own task structure pointer into sem->owner and clear the reader nonspinnable bit in the process. The time taken for each iteration of the reader-owned rwsem spinning loop varies. Below are sample minimum elapsed times for 16 iterations of the loop. System Time for 16 Iterations ------ ---------------------- 1-socket Skylake ~800ns 4-socket Broadwell ~300ns 2-socket ThunderX2 (arm64) ~250ns When the lock cacheline is contended, we can see up to almost 10X increase in elapsed time. So 25us will be at most 500, 1300 and 1600 iterations for each of the above systems. With a locking microbenchmark running on 5.1 based kernel, the total locking rates (in kops/s) on a 8-socket IvyBridge-EX system with equal numbers of readers and writers before and after this patch were as follows: # of Threads Pre-patch Post-patch ------------ --------- ---------- 2 1,759 6,684 4 1,684 6,738 8 1,074 7,222 16 900 7,163 32 458 7,316 64 208 520 128 168 425 240 143 474 This patch gives a big boost in performance for mixed reader/writer workloads. With 32 locking threads, the rwsem lock event data were: rwsem_opt_fail=79850 rwsem_opt_nospin=5069 rwsem_opt_rlock=597484 rwsem_opt_wlock=957339 rwsem_sleep_reader=57782 rwsem_sleep_writer=55663 With 64 locking threads, the data looked like: rwsem_opt_fail=346723 rwsem_opt_nospin=6293 rwsem_opt_rlock=1127119 rwsem_opt_wlock=1400628 rwsem_sleep_reader=308201 rwsem_sleep_writer=72281 So a lot more threads acquired the lock in the slowpath and more threads went to sleep. Signed-off-by: -
Waiman Long authored
The rwsem->owner contains not just the task structure pointer, it also holds some flags for storing the current state of the rwsem. Some of the flags may have to be atomically updated. To reflect the new reality, the owner is now changed to an atomic_long_t type. New helper functions are added to properly separate out the task structure pointer and the embedded flags. Suggested-by:
-
Waiman Long authored
This patch enables readers to optimistically spin on a rwsem when it is owned by a writer instead of going to sleep directly. The rwsem_can_spin_on_owner() function is extracted out of rwsem_optimistic_spin() and is called directly by rwsem_down_read_slowpath() and rwsem_down_write_slowpath(). With a locking microbenchmark running on 5.1 based kernel, the total locking rates (in kops/s) on a 8-socket IvyBrige-EX system with equal numbers of readers and writers before and after the patch were as follows: # of Threads Pre-patch Post-patch ------------ --------- ---------- 4 1,674 1,684 8 1,062 1,074 16 924 900 32 300 458 64 195 208 128 164 168 240 149 143 The performance change wasn't significant in this case, but this change is required by a follow-on patch. Signed-off-by: -
Waiman Long authored
Bit 1 of sem->owner (RWSEM_ANONYMOUSLY_OWNED) is used to designate an anonymous owner - readers or an anonymous writer. The setting of this anonymous bit is used as an indicator that optimistic spinning cannot be done on this rwsem. With the upcoming reader optimistic spinning patches, a reader-owned rwsem can be spinned on for a limit period of time. We still need this bit to indicate a rwsem is nonspinnable, but not setting this bit loses its meaning that the owner is known. So rename the bit to RWSEM_NONSPINNABLE to clarify its meaning. This patch also fixes a DEBUG_RWSEMS_WARN_ON() bug in __up_write(). Signed-off-by:
-
Waiman Long authored
When the front of the wait queue is a reader, other readers immediately following the first reader will also be woken up at the same time. However, if there is a writer in between. Those readers behind the writer will not be woken up. Because of optimistic spinning, the lock acquisition order is not FIFO anyway. The lock handoff mechanism will ensure that lock starvation will not happen. Assuming that the lock hold times of the other readers still in the queue will be about the same as the readers that are being woken up, there is really not much additional cost other than the additional latency due to the wakeup of additional tasks by the waker. Therefore all the readers up to a maximum of 256 in the queue are woken up when the first waiter is a reader to improve reader throughput. This is somewhat similar in concept to a phase-fair R/W lock. With a locking microbenchmark running on 5.1 based kernel, the total locking rates (in kops/s) on a 8-socket IvyBridge-EX system with equal numbers of readers and writers before and after this patch were as follows: # of Threads Pre-Patch Post-patch ------------ --------- ---------- 4 1,641 1,674 8 731 1,062 16 564 924 32 78 300 64 38 195 240 50 149 There is no performance gain at low contention level. At high contention level, however, this patch gives a pretty decent performance boost. Signed-off-by: -
Waiman Long authored
An RT task can do optimistic spinning only if the lock holder is actually running. If the state of the lock holder isn't known, there is a possibility that high priority of the RT task may block forward progress of the lock holder if it happens to reside on the same CPU. This will lead to deadlock. So we have to make sure that an RT task will not spin on a reader-owned rwsem. When the owner is temporarily set to NULL, there are two cases where we may want to continue spinning: 1) The lock owner is in the process of releasing the lock, sem->owner is cleared but the lock has not been released yet. 2) The lock was free and owner cleared, but another task just comes in and acquire the lock before we try to get it. The new owner may be a spinnable writer. So an RT task is now made to retry one more time to see if it can acquire the lock or continue spinning on the new owning writer. When testing on a 8-socket IvyBridge-EX system, the one additional retry seems to improve locking performance of RT write locking threads under heavy contentions. The table below shows the locking rates (in kops/s) with various write locking threads before and after the patch. Locking threads Pre-patch Post-patch --------------- --------- ----------- 4 2,753 2,608 8 2,529 2,520 16 1,727 1,918 32 1,263 1,956 64 889 1,343 Signed-off-by: -
Waiman Long authored
With the use of wake_q, we can do task wakeups without holding the wait_lock. There is one exception in the rwsem code, though. It is when the writer in the slowpath detects that there are waiters ahead but the rwsem is not held by a writer. This can lead to a long wait_lock hold time especially when a large number of readers are to be woken up. Remediate this situation by releasing the wait_lock before waking up tasks and re-acquiring it afterward. The rwsem_try_write_lock() function is also modified to read the rwsem count directly to avoid stale count value. Suggested-by:
-
Waiman Long authored
Because of writer lock stealing, it is possible that a constant stream of incoming writers will cause a waiting writer or reader to wait indefinitely leading to lock starvation. This patch implements a lock handoff mechanism to disable lock stealing and force lock handoff to the first waiter or waiters (for readers) in the queue after at least a 4ms waiting period unless it is a RT writer task which doesn't need to wait. The waiting period is used to avoid discouraging lock stealing too much to affect performance. The setting and clearing of the handoff bit is serialized by the wait_lock. So racing is not possible. A rwsem microbenchmark was run for 5 seconds on a 2-socket 40-core 80-thread Skylake system with a v5.1 based kernel and 240 write_lock threads with 5us sleep critical section. Before the patch, the min/mean/max numbers of locking operations for the locking threads were 1/7,792/173,696. After the patch, the figures became 5,842/6,542/7,458. It can be seen that the rwsem became much more fair, though there was a drop of about 16% in the mean locking operations done which was a tradeoff of having better fairness. Making the waiter set the handoff bit right after the first wakeup can impact performance especially with a mixed reader/writer workload. With the same microbenchmark with short critical section and equal number of reader and writer threads (40/40), the reader/writer locking operation counts with the current patch were: 40 readers, Iterations Min/Mean/Max = 1,793/1,794/1,796 40 writers, Iterations Min/Mean/Max = 1,793/34,956/86,081 By making waiter set handoff bit immediately after wakeup: 40 readers, Iterations Min/Mean/Max = 43/44/46 40 writers, Iterations Min/Mean/Max = 43/1,263/3,191 Signed-off-by:
-
Waiman Long authored
This patch modifies rwsem_spin_on_owner() to return four possible values to better reflect the state of lock holder which enables us to make a better decision of what to do next. Signed-off-by:
-
Waiman Long authored
After merging all the relevant rwsem code into one single file, there are a number of optimizations and cleanups that can be done: 1) Remove all the EXPORT_SYMBOL() calls for functions that are not accessed elsewhere. 2) Remove all the __visible tags as none of the functions will be called from assembly code anymore. 3) Make all the internal functions static. 4) Remove some unneeded blank lines. 5) Remove the intermediate rwsem_down_{read|write}_failed*() functions and rename __rwsem_down_{read|write}_failed_common() to rwsem_down_{read|write}_slowpath(). 6) Remove "__" prefix of __rwsem_mark_wake(). 7) Use atomic_long_try_cmpxchg_acquire() as much as possible. 8) Remove the rwsem_rtrylock and rwsem_wtrylock lock events as they are not that useful. That enables the compiler to do better optimization and reduce code size. The text+data size of rwsem.o on an x86-64 machine with gcc8 was reduced from 10237 bytes to 5030 bytes with this change. Suggested-by: -
Waiman Long authored
Now we only have one implementation of rwsem. Even though we still use xadd to handle reader locking, we use cmpxchg for writer instead. So the filename rwsem-xadd.c is not strictly correct. Also no one outside of the rwsem code need to know the internal implementation other than function prototypes for two internal functions that are called directly from percpu-rwsem.c. So the rwsem-xadd.c and rwsem.h files are now merged into rwsem.c in the following order: <upper part of rwsem.h> <rwsem-xadd.c> <lower part of rwsem.h> <rwsem.c> The rwsem.h file now contains only 2 function declarations for __up_read() and __down_read(). This is a code relocation patch with no code change at all except making __up_read() and __down_read() non-static functions so they can be used by percpu-rwsem.c. Suggested-by:
-
Waiman Long authored
The current way of using various reader, writer and waiting biases in the rwsem code are confusing and hard to understand. I have to reread the rwsem count guide in the rwsem-xadd.c file from time to time to remind myself how this whole thing works. It also makes the rwsem code harder to be optimized. To make rwsem more sane, a new locking scheme similar to the one in qrwlock is now being used. The atomic long count has the following bit definitions: Bit 0 - writer locked bit Bit 1 - waiters present bit Bits 2-7 - reserved for future extension Bits 8-X - reader count (24/56 bits) The cmpxchg instruction is now used to acquire the write lock. The read lock is still acquired with xadd instruction, so there is no change here. This scheme will allow up to 16M/64P active readers which should be more than enough. We can always use some more reserved bits if necessary. With that change, we can deterministically know if a rwsem has been write-locked. Looking at the count alone, however, one cannot determine for certain if a rwsem is owned by readers or not as the readers that set the reader count bits may be in the process of backing out. So we still need the reader-owned bit in the owner field to be sure. With a locking microbenchmark running on 5.1 based kernel, the total locking rates (in kops/s) of the benchmark on a 8-socket 120-core IvyBridge-EX system before and after the patch were as follows: Before Patch After Patch # of Threads wlock rlock wlock rlock ------------ ----- ----- ----- ----- 1 30,659 31,341 31,055 31,283 2 8,909 16,457 9,884 17,659 4 9,028 15,823 8,933 20,233 8 8,410 14,212 7,230 17,140 16 8,217 25,240 7,479 24,607 The locking rates of the benchmark on a Power8 system were as follows: Before Patch After Patch # of Threads wlock rlock wlock rlock ------------ ----- ----- ----- ----- 1 12,963 13,647 13,275 13,601 2 7,570 11,569 7,902 10,829 4 5,232 5,516 5,466 5,435 8 5,233 3,386 5,467 3,168 The locking rates of the benchmark on a 2-socket ARM64 system were as follows: Before Patch After Patch # of Threads wlock rlock wlock rlock ------------ ----- ----- ----- ----- 1 21,495 21,046 21,524 21,074 2 5,293 10,502 5,333 10,504 4 5,325 11,463 5,358 11,631 8 5,391 11,712 5,470 11,680 The performance are roughly the same before and after the patch. There are run-to-run variations in performance. Runs with higher variances usually have higher throughput. Signed-off-by: -
Waiman Long authored
After the following commit: 59aabfc7 ("locking/rwsem: Reduce spinlock contention in wakeup after up_read()/up_write()") the rwsem_wake() forgoes doing a wakeup if the wait_lock cannot be directly acquired and an optimistic spinning locker is present. This can help performance by avoiding spinning on the wait_lock when it is contended. With the later commit: 133e89ef ("locking/rwsem: Enable lockless waiter wakeup(s)") the performance advantage of the above optimization diminishes as the average wait_lock hold time become much shorter. With a later patch that supports rwsem lock handoff, we can no longer relies on the fact that the presence of an optimistic spinning locker will ensure that the lock will be acquired by a task soon and rwsem_wake() will be called later on to wake up waiters. This can lead to missed wakeup and application hang. So the original 59aabfc7 commit has to be reverted. Signed-off-by:
-
Waiman Long authored
The owner field in the rw_semaphore structure is used primarily for optimistic spinning. However, identifying the rwsem owner can also be helpful in debugging as well as tracing locking related issues when analyzing crash dump. The owner field may also store state information that can be important to the operation of the rwsem. So the owner field is now made a permanent member of the rw_semaphore structure irrespective of CONFIG_RWSEM_SPIN_ON_OWNER. Signed-off-by:
-
Peter Zijlstra authored
Recent probing at the Linux Kernel Memory Model uncovered a 'surprise'. Strongly ordered architectures where the atomic RmW primitive implies full memory ordering and smp_mb__{before,after}_atomic() are a simple barrier() (such as x86) fail for: *x = 1; atomic_inc(u); smp_mb__after_atomic(); r0 = *y; Because, while the atomic_inc() implies memory order, it (surprisingly) does not provide a compiler barrier. This then allows the compiler to re-order like so: atomic_inc(u); *x = 1; smp_mb__after_atomic(); r0 = *y; Which the CPU is then allowed to re-order (under TSO rules) like: atomic_inc(u); r0 = *y; *x = 1; And this very much was not intended. Therefore strengthen the atomic RmW ops to include a compiler barrier. NOTE: atomic_{or,and,xor} and the bitops already had the compiler barrier. Signed-off-by: -
Kobe Wu authored
DEBUG_LOCKS_WARN_ON() will turn off debug_locks and makes print_unlock_imbalance_bug() return directly. Remove a redundant whitespace. Signed-off-by:
Kobe Wu <kobe-cp.wu@mediatek.com> Signed-off-by:
-
Nikolay Borisov authored
All callers of lockdep_assert_held_exclusive() use it to verify the correct locking state of either a semaphore (ldisc_sem in tty, mmap_sem for perf events, i_rwsem of inode for dax) or rwlock by apparmor. Thus it makes sense to rename _exclusive to _write since that's the semantics callers care. Additionally there is already lockdep_assert_held_read(), which this new naming is more consistent with. No functional changes. Signed-off-by:
Nikolay Borisov <nborisov@suse.com> Signed-off-by:
-
Daniel Bristot de Oliveira authored
Currently, the jump label of a static key is transformed via the arch specific function: void arch_jump_label_transform(struct jump_entry *entry, enum jump_label_type type) The new approach (batch mode) uses two arch functions, the first has the same arguments of the arch_jump_label_transform(), and is the function: bool arch_jump_label_transform_queue(struct jump_entry *entry, enum jump_label_type type) Rather than transforming the code, it adds the jump_entry in a queue of entries to be updated. This functions returns true in the case of a successful enqueue of an entry. If it returns false, the caller must to apply the queue and then try to queue again, for instance, because the queue is full. This function expects the caller to sort the entries by the address before enqueueuing then. This is already done by the arch independent code, though. After queuing all jump_entries, the function: void arch_jump_label_transform_apply(void) Applies the changes in the queue. Signed-off-by:Daniel Bristot de Oliveira <bristot@redhat.com> Signed-off-by:
-
Daniel Bristot de Oliveira authored
If the architecture supports the batching of jump label updates, use it! An easy way to see the benefits of this patch is switching the schedstats on and off. For instance: -------------------------- %< ---------------------------- #!/bin/sh while [ true ]; do sysctl -w kernel.sched_schedstats=1 sleep 2 sysctl -w kernel.sched_schedstats=0 sleep 2 done -------------------------- >% ---------------------------- while watching the IPI count: -------------------------- %< ---------------------------- # watch -n1 "cat /proc/interrupts | grep Function" -------------------------- >% ---------------------------- With the current mode, it is possible to see +- 168 IPIs each 2 seconds, while with this patch the number of IPIs goes to 3 each 2 seconds. Regarding the performance impact of this patch set, I made two measurements: The time to update a key (the task that is causing the change) The time to run the int3 handler (the side effect on a thread that hits the code being changed) The schedstats static key was chosen as the key to being switched on and off. The reason being is that it is used in more than 56 places, in a hot path. The change in the schedstats static key will be done with the following command: while [ true ]; do sysctl -w kernel.sched_schedstats=1 usleep 500000 sysctl -w kernel.sched_schedstats=0 usleep 500000 done In this way, they key will be updated twice per second. To force the hit of the int3 handler, the system will also run a kernel compilation with two jobs per CPU. The test machine is a two nodes/24 CPUs box with an Intel Xeon processor @2.27GHz. Regarding the update part, on average, the regular kernel takes 57 ms to update the schedstats key, while the kernel with the batch updates takes just 1.4 ms on average. Although it seems to be too good to be true, it makes sense: the schedstats key is used in 56 places, so it was expected that it would take around 56 times to update the keys with the current implementation, as the IPIs are the most expensive part of the update. Regarding the int3 handler, the non-batch handler takes 45 ns on average, while the batch version takes around 180 ns. At first glance, it seems to be a high value. But it is not, considering that it is doing 56 updates, rather than one! It is taking four times more, only. This gain is possible because the patch uses a binary search in the vector: log2(56)=5.8. So, it was expected to have an overhead within four times. (voice of tv propaganda) But, that is not all! As the int3 handler keeps on for a shorter period (because the update part is on for a shorter time), the number of hits in the int3 handler decreased by 10%. The question then is: Is it worth paying the price of "135 ns" more in the int3 handler? Considering that, in this test case, we are saving the handling of 53 IPIs, that takes more than these 135 ns, it seems to be a meager price to be paid. Moreover, the test case was forcing the hit of the int3, in practice, it does not take that often. While the IPI takes place on all CPUs, hitting the int3 handler or not! For instance, in an isolated CPU with a process running in user-space (nohz_full use-case), the chances of hitting the int3 handler is barely zero, while there is no way to avoid the IPIs. By bounding the IPIs, we are improving a lot this scenario. Signed-off-by: -
Daniel Bristot de Oliveira authored
Currently, the patch of an address is done in three steps: -- Pseudo-code #1 - Current implementation --- 1) add an int3 trap to the address that will be patched sync cores (send IPI to all other CPUs) 2) update all but the first byte of the patched range sync cores (send IPI to all other CPUs) 3) replace the first byte (int3) by the first byte of replacing opcode sync cores (send IPI to all other CPUs) -- Pseudo-code #1 --- When a static key has more than one entry, these steps are called once for each entry. The number of IPIs then is linear with regard to the number 'n' of entries of a key: O(n*3), which is O(n). This algorithm works fine for the update of a single key. But we think it is possible to optimize the case in which a static key has more than one entry. For instance, the sched_schedstats jump label has 56 entries in my (updated) fedora kernel, resulting in 168 IPIs for each CPU in which the thread that is enabling the key is _not_ running. With this patch, rather than receiving a single patch to be processed, a vector of patches is passed, enabling the rewrite of the pseudo-code #1 in this way: -- Pseudo-code #2 - This patch --- 1) for each patch in the vector: add an int3 trap to the address that will be patched sync cores (send IPI to all other CPUs) 2) for each patch in the vector: update all but the first byte of the patched range sync cores (send IPI to all other CPUs) 3) for each patch in the vector: replace the first byte (int3) by the first byte of replacing opcode sync cores (send IPI to all other CPUs) -- Pseudo-code #2 - This patch --- Doing the update in this way, the number of IPI becomes O(3) with regard to the number of keys, which is O(1). The batch mode is done with the function text_poke_bp_batch(), that receives two arguments: a vector of "struct text_to_poke", and the number of entries in the vector. The vector must be sorted by the addr field of the text_to_poke structure, enabling the binary search of a handler in the poke_int3_handler function (a fast path). Signed-off-by:Masami Hiramatsu <mhiramat@kernel.org> Cc: Borislav Petkov <bp@alien8.de> Cc: Chris von Recklinghausen <crecklin@redhat.com> Cc: Clark Williams <williams@redhat.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: H. Peter Anvin <hpa@zytor.com> Cc: Jason Baron <jbaron@akamai.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Josh Poimboeuf <jpoimboe@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Scott Wood <swood@redhat.com> Cc: Steven Rostedt (VMware) <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Link: https://lkml.kernel.org/r/ca506ed52584c80f64de23f6f55ca288e5d079de.1560325897.git.bristot@redhat.comSigned-off-by:
-
Daniel Bristot de Oliveira authored
In the batching mode, all the entries of a given key are updated at once. During the update of a key, a hit in the int3 handler will check if the hitting code address belongs to one of these keys. To optimize the search of a given code in the vector of entries being updated, a binary search is used. The binary search relies on the order of the entries of a key by its code. Hence the keys need to be sorted by the code too, so sort the entries of a given key by the code. Signed-off-by:
-
Daniel Bristot de Oliveira authored
Move the definition of the code to be written from __jump_label_transform() to a specialized function. No functional change. Signed-off-by:
-
Daniel Bristot de Oliveira authored
Move the check if a jump_entry is valid to a function. No functional change. Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
- 16 Jun, 2019 4 commits
-
-
Linus Torvalds authored
-
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tipLinus Torvalds authored
Pull x86 fixes from Thomas Gleixner: "The accumulated fixes from this and last week: - Fix vmalloc TLB flush and map range calculations which lead to stale TLBs, spurious faults and other hard to diagnose issues. - Use fault_in_pages_writable() for prefaulting the user stack in the FPU code as it's less fragile than the current solution - Use the PF_KTHREAD flag when checking for a kernel thread instead of current->mm as the latter can give the wrong answer due to use_mm() - Compute the vmemmap size correctly for KASLR and 5-Level paging. Otherwise this can end up with a way too small vmemmap area. - Make KASAN and 5-level paging work again by making sure that all invalid bits are masked out when computing the P4D offset. This worked before but got broken recently when the LDT remap area was moved. - Prevent a NULL pointer dereference in the resource control code which can be triggered with certain mount options when the requested resource is not available. - Enforce ordering of microcode loading vs. perf initialization on secondary CPUs. Otherwise perf tries to access a non-existing MSR as the boot CPU marked it as available. - Don't stop the resource control group walk early otherwise the control bitmaps are not updated correctly and become inconsistent. - Unbreak kgdb by returning 0 on success from kgdb_arch_set_breakpoint() instead of an error code. - Add more Icelake CPU model defines so depending changes can be queued in other trees" * 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: x86/microcode, cpuhotplug: Add a microcode loader CPU hotplug callback x86/kasan: Fix boot with 5-level paging and KASAN x86/fpu: Don't use current->mm to check for a kthread x86/kgdb: Return 0 from kgdb_arch_set_breakpoint() x86/resctrl: Prevent NULL pointer dereference when local MBM is disabled x86/resctrl: Don't stop walking closids when a locksetup group is found x86/fpu: Update kernel's FPU state before using for the fsave header x86/mm/KASLR: Compute the size of the vmemmap section properly x86/fpu: Use fault_in_pages_writeable() for pre-faulting x86/CPU: Add more Icelake model numbers mm/vmalloc: Avoid rare case of flushing TLB with weird arguments mm/vmalloc: Fix calculation of direct map addr range -
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tipLinus Torvalds authored
Pull timer fixes from Thomas Gleixner: "A set of small fixes: - Repair the ktime_get_coarse() functions so they actually deliver what they are supposed to: tick granular time stamps. The current code missed to add the accumulated nanoseconds part of the timekeeper so the resulting granularity was 1 second. - Prevent the tracer from infinitely recursing into time getter functions in the arm architectured timer by marking these functions notrace - Fix a trivial compiler warning caused by wrong qualifier ordering" * 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: timekeeping: Repair ktime_get_coarse*() granularity clocksource/drivers/arm_arch_timer: Don't trace count reader functions clocksource/drivers/timer-ti-dm: Change to new style declaration -
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tipLinus Torvalds authored
Pull RAS fixes from Thomas Gleixner: "Two small fixes for RAS: - Use a proper search algorithm to find the correct element in the CEC array. The replacement was a better choice than fixing the crash causes by the original search function with horrible duct tape. - Move the timer based decay function into thread context so it can actually acquire the mutex which protects the CEC array to prevent corruption" * 'ras-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: RAS/CEC: Convert the timer callback to a workqueue RAS/CEC: Fix binary search function
-
- 15 Jun, 2019 9 commits
-
-
git://git.infradead.org/linux-platform-drivers-x86Linus Torvalds authored
Pull x86 platform driver fixes from Andy Shevchenko: - fix a couple of Mellanox driver enumeration issues - fix ASUS laptop regression with backlight - fix Dell computers that got a wrong mode (tablet versus laptop) after resume * tag 'platform-drivers-x86-v5.2-3' of git://git.infradead.org/linux-platform-drivers-x86: platform/mellanox: mlxreg-hotplug: Add devm_free_irq call to remove flow platform/x86: mlx-platform: Fix parent device in i2c-mux-reg device registration platform/x86: intel-vbtn: Report switch events when event wakes device platform/x86: asus-wmi: Only Tell EC the OS will handle display hotkeys from asus_nb_wmi
-
git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usbLinus Torvalds authored
Pull USB fixes from Greg KH: "Here are some small USB driver fixes for 5.2-rc5 Nothing major, just some small gadget fixes, usb-serial new device ids, a few new quirks, and some small fixes for some regressions that have been found after the big 5.2-rc1 merge. All of these have been in linux-next for a while with no reported issues" * tag 'usb-5.2-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: usb: typec: Make sure an alt mode exist before getting its partner usb: gadget: udc: lpc32xx: fix return value check in lpc32xx_udc_probe() usb: gadget: dwc2: fix zlp handling usb: dwc2: Set actual frame number for completed ISOC transfer for none DDMA usb: gadget: udc: lpc32xx: allocate descriptor with GFP_ATOMIC usb: gadget: fusb300_udc: Fix memory leak of fusb300->ep[i] usb: phy: mxs: Disable external charger detect in mxs_phy_hw_init() usb: dwc2: Fix DMA cache alignment issues usb: dwc2: host: Fix wMaxPacketSize handling (fix webcam regression) USB: Fix chipmunk-like voice when using Logitech C270 for recording audio. USB: usb-storage: Add new ID to ums-realtek usb: typec: ucsi: ccg: fix memory leak in do_flash USB: serial: option: add Telit 0x1260 and 0x1261 compositions USB: serial: pl2303: add Allied Telesis VT-Kit3 USB: serial: option: add support for Simcom SIM7500/SIM7600 RNDIS mode
-
git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linuxLinus Torvalds authored
Pull powerpc fixes from Michael Ellerman: "One fix for a regression introduced by our 32-bit KASAN support, which broke booting on machines with "bootx" early debugging enabled. A fix for a bug which broke kexec on 32-bit, introduced by changes to the 32-bit STRICT_KERNEL_RWX support in v5.1. Finally two fixes going to stable for our THP split/collapse handling, discovered by Nick. The first fixes random crashes and/or corruption in guests under sufficient load. Thanks to: Nicholas Piggin, Christophe Leroy, Aaro Koskinen, Mathieu Malaterre" * tag 'powerpc-5.2-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: powerpc/32s: fix booting with CONFIG_PPC_EARLY_DEBUG_BOOTX powerpc/64s: __find_linux_pte() synchronization vs pmdp_invalidate() powerpc/64s: Fix THP PMD collapse serialisation powerpc: Fix kexec failure on book3s/32
-
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-traceLinus Torvalds authored
Pull tracing fixes from Steven Rostedt: - Out of range read of stack trace output - Fix for NULL pointer dereference in trace_uprobe_create() - Fix to a livepatching / ftrace permission race in the module code - Fix for NULL pointer dereference in free_ftrace_func_mapper() - A couple of build warning clean ups * tag 'trace-v5.2-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace: ftrace: Fix NULL pointer dereference in free_ftrace_func_mapper() module: Fix livepatch/ftrace module text permissions race tracing/uprobe: Fix obsolete comment on trace_uprobe_create() tracing/uprobe: Fix NULL pointer dereference in trace_uprobe_create() tracing: Make two symbols static tracing: avoid build warning with HAVE_NOP_MCOUNT tracing: Fix out-of-range read in trace_stack_print()
-
Borislav Petkov authored
Adric Blake reported the following warning during suspend-resume: Enabling non-boot CPUs ... x86: Booting SMP configuration: smpboot: Booting Node 0 Processor 1 APIC 0x2 unchecked MSR access error: WRMSR to 0x10f (tried to write 0x0000000000000000) \ at rIP: 0xffffffff8d267924 (native_write_msr+0x4/0x20) Call Trace: intel_set_tfa intel_pmu_cpu_starting ? x86_pmu_dead_cpu x86_pmu_starting_cpu cpuhp_invoke_callback ? _raw_spin_lock_irqsave notify_cpu_starting start_secondary secondary_startup_64 microcode: sig=0x806ea, pf=0x80, revision=0x96 microcode: updated to revision 0xb4, date = 2019-04-01 CPU1 is up The MSR in question is MSR_TFA_RTM_FORCE_ABORT and that MSR is emulated by microcode. The log above shows that the microcode loader callback happens after the PMU restoration, leading to the conjecture that because the microcode hasn't been updated yet, that MSR is not present yet, leading to the #GP. Add a microcode loader-specific hotplug vector which comes before the PERF vectors and thus executes earlier and makes sure the MSR is present. Fixes: 400816f6 ("perf/x86/intel: Implement support for TSX Force Abort") Reported-by:
Adric Blake <promarbler14@gmail.com> Signed-off-by:
Borislav Petkov <bp@suse.de> Reviewed-by:
Thomas Gleixner <tglx@linutronix.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: <stable@vger.kernel.org> Cc: x86@kernel.org Link: https://bugzilla.kernel.org/show_bug.cgi?id=203637
-
git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroupLinus Torvalds authored
Pull cgroup fixes from Tejun Heo: "This has an unusually high density of tricky fixes: - task_get_css() could deadlock when it races against a dying cgroup. - cgroup.procs didn't list thread group leaders with live threads. This could mislead readers to think that a cgroup is empty when it's not. Fixed by making PROCS iterator include dead tasks. I made a couple mistakes making this change and this pull request contains a couple follow-up patches. - When cpusets run out of online cpus, it updates cpusmasks of member tasks in bizarre ways. Joel improved the behavior significantly" * 'for-5.2-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup: cpuset: restore sanity to cpuset_cpus_allowed_fallback() cgroup: Fix css_task_iter_advance_css_set() cset skip condition cgroup: css_task_iter_skip()'d iterators must be advanced before accessed cgroup: Include dying leaders with live threads in PROCS iterations cgroup: Implement css_task_iter_skip() cgroup: Call cgroup_release() before __exit_signal() docs cgroups: add another example size for hugetlb cgroup: Use css_tryget() instead of css_tryget_online() in task_get_css() -
git://anongit.freedesktop.org/drm/drmLinus Torvalds authored
Pull drm fixes from Daniel Vetter: "Nothing unsettling here, also not aware of anything serious still pending. The edid override regression fix took a bit longer since this seems to be an area with an overabundance of bad options. But the fix we have now seems like a good path forward. Next week it should be back to Dave. Summary: - fix regression on amdgpu on SI - fix edid override regression - driver fixes: amdgpu, i915, mediatek, meson, panfrost - fix writecombine for vmap in gem-shmem helper (used by panfrost) - add more panel quirks" * tag 'drm-fixes-2019-06-14' of git://anongit.freedesktop.org/drm/drm: (25 commits) drm/amdgpu: return 0 by default in amdgpu_pm_load_smu_firmware drm/amdgpu: Fix bounds checking in amdgpu_ras_is_supported() drm: add fallback override/firmware EDID modes workaround drm/edid: abstract override/firmware EDID retrieval drm/i915/perf: fix whitelist on Gen10+ drm/i915/sdvo: Implement proper HDMI audio support for SDVO drm/i915: Fix per-pixel alpha with CCS drm/i915/dmc: protect against reading random memory drm/i915/dsi: Use a fuzzy check for burst mode clock check drm/amdgpu/{uvd,vcn}: fetch ring's read_ptr after alloc drm/panfrost: Require the simple_ondemand governor drm/panfrost: make devfreq optional again drm/gem_shmem: Use a writecombine mapping for ->vaddr drm: panel-orientation-quirks: Add quirk for GPD MicroPC drm: panel-orientation-quirks: Add quirk for GPD pocket2 drm/meson: fix G12A primary plane disabling drm/meson: fix primary plane disabling drm/meson: fix G12A HDMI PLL settings for 4K60 1000/1001 variations drm/mediatek: call mtk_dsi_stop() after mtk_drm_crtc_atomic_disable() drm/mediatek: clear num_pipes when unbind driver ...
-
git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2Linus Torvalds authored
Pull gfs2 fix from Andreas Gruenbacher: "Fix rounding error in gfs2_iomap_page_prepare" * tag 'gfs2-v5.2.fixes2' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2: gfs2: Fix rounding error in gfs2_iomap_page_prepare
-
git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsiLinus Torvalds authored
Pull SCSI fix from James Bottomley: "A single bug fix for hpsa. The user visible consequences aren't clear, but the ioaccel2 raid acceleration may misfire on the malformed request assuming the payload is big enough to require chaining (more than 31 sg entries)" * tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: scsi: hpsa: correct ioaccel2 chaining
-
