sergiotarxz/nuttx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
28,687 Commits 1 Branch 0 Tags 291 MiB
49e4e62aab
Commit Graph

1092 Commits

Author SHA1 Message Date
Gregory Nutt
99bb2dda5d i.MX6: Corrects behavior of last SMP patch with i.MX6 2017-01-13 12:29:57 -06:00
Gregory Nutt
a51b5b7e17 Add REVISIT to comments. 2017-01-13 09:40:17 -06:00
Gregory Nutt
37e6e6a52e Eliminate a warning 2017-01-13 06:56:13 -06:00
Gregory Nutt
9ce4022096 SMP: Fix an error in critical section logic when performing a context switch from an interrupt handler. The g_cpu_irqset bit was not being set for the CPU so other CPUs did not know about the critical section. 2017-01-13 06:48:10 -06:00
Masayuki Ishikawa
acbfa47f70 sched_note: Fix spinlock instrumentation 2017-01-12 08:04:28 -06:00
Gregory Nutt
d5b1ca14e2 Update README and some comments. 2017-01-02 13:54:07 -06:00
Gregory Nutt
04557a1198 Eliminate a warning 2016-12-31 14:47:32 -06:00
Gregory Nutt
8f716a386b SMP: Use irq_cpu_locked() in sched_removereadytorun() and sched_setpriority() 2016-12-29 10:40:58 -06:00
Gregory Nutt
af92a67fc5 SMP: Use irq_cpu_locked() in sched_mergepending() 2016-12-29 08:53:31 -06:00
Gregory Nutt
e7d2b9f0e8 SMP: Move sharable function to common file as irq_cpu_locked(). Use irq_cpu_locked() in sched_unlock() 2016-12-29 08:17:10 -06:00
Gregory Nutt
96394f339e SMP: Fix a gap where we may try to make modifications to the task lists without being in a critical sections. That permits concurrent access to the tasks lists and many subtle problems. This fix just remains in the critical section throughout the operation (and possible until the task is restore in the event of a context switch). Makes a big difference in stability 2016-12-28 13:58:24 -06:00
Gregory Nutt
88da65ad8a sched_cpulocked: Avoid use of spinlock. That has been reported to cause a deadlock. 2016-12-28 12:10:17 -06:00
Gregory Nutt
72490f84e6 SMP: Make checks for CPU lock set more robust. There are certains conditions early in initialization on during interrupt handling where things need to be done a little differently. 2016-12-28 10:21:05 -06:00
Gregory Nutt
cfb876263a SMP: There were certain conditions that we must avoid by preventing releasing the pending tasks while withn a critical section. But this logic was incomplete; there was no logic to prevent other CPUs from adding new, running tasks while on CPU is in a critical section. 
This commit corrects this.  This is matching logic in sched_addreadytorun to avoid starting new tasks within the critical section (unless the CPU is the holder of the lock).  The holder of the IRQ lock must be permitted to do whatever it needs to do.
 2016-12-27 08:49:07 -06:00
Gregory Nutt
e0f3f4ae19 Update some coments 2016-12-26 16:12:27 -06:00
Gregory Nutt
849a5dc2a9 SMP: Enforce this rule: Tasks which are normally restored when sched_unlock() is called must remain pending (1) if we are in a critical section, i.e., g_cpu_irqlock is locked , or (2) other CPUs still have pre-emption disabled, i.e., g_cpu_schedlock is locked. In those cases, the release of the pending tasks must be deferred until those conditions are met. 2016-12-26 08:15:02 -06:00
Gregory Nutt
d45a81d643 Revert "SMP: Fix a error introduced in 36b676eea6 and fully decouple disabling of pre-emption from critical sections." 
This reverts commit 9aedf1bdec.
 2016-12-26 07:48:22 -06:00
Gregory Nutt
9aedf1bdec SMP: Fix a error introduced in 36b676eea6 and fully decouple disabling of pre-emption from critical sections. 2016-12-25 06:54:43 -06:00
Gregory Nutt
efb86382c3 SMP: Back out deferred IRQ locking. This was accidentally merged into master and it looks like it is going to be more work than I thought to get it working again. Changes will go to the irqlock branch. 2016-12-24 19:53:37 -06:00
Gregory Nutt
3af6b2a9a6 Update a comment 2016-12-24 19:01:07 -06:00
Gregory Nutt
f55bad863b SMP: Fix logic error in last change: Deferred restoration of IRQ lock only applies if the context switch was on this CPU. 2016-12-24 18:52:58 -06:00
Gregory Nutt
36b676eea6 SMP sched_unlock: Should merging depend on g_cpu_irqlock or not? Let's at least be consistent. And don't use irqcount; use g_cpu_irqlock. 2016-12-24 13:45:11 -06:00
Gregory Nutt
d75a7643d0 Fix a trivial typo 2016-12-23 13:01:20 -06:00
Gregory Nutt
e6fff09ef8 Implement deferred IRQ locking. So far only form ARMv7-M. 2016-12-23 07:55:41 -06:00
Gregory Nutt
d648f9c8b4 Add task_testcancel() 2016-12-10 16:34:14 -06:00
Gregory Nutt
f132960789 Add task_setcanceltype() 2016-12-10 16:06:14 -06:00
Gregory Nutt
5fb207eb36 Add task_setcancelstate() 2016-12-10 15:16:46 -06:00
Gregory Nutt
698597a838 task_delete() now obeys all cancellation point semantics. 2016-12-10 14:39:19 -06:00
Gregory Nutt
6997cda1b5 Grrr... cloned typos! 2016-12-10 09:45:55 -06:00
Gregory Nutt
842ec7e612 ifdef out some non-reachable code 2016-12-10 09:43:04 -06:00
Gregory Nutt
b52e4e5ecd Move cancellation point definitions to their own header file. 2016-12-10 09:08:26 -06:00
Gregory Nutt
bc3ca25cc7 Cancellation points: Close up some logic to eliminte some race conditions. 2016-12-10 08:36:58 -06:00
Gregory Nutt
e62b3bccd3 pthread_setcanceltype() and pthread_testcancel() do not have to be system calls if cancellation points are not enabled. 2016-12-10 07:40:48 -06:00
Gregory Nutt
a0f567f4a3 Update TODO, cosmetic changes, spelling -- US English spells it canceled and canceling vs cancelled and cancelling. No idea why. 2016-12-09 18:39:40 -06:00
Gregory Nutt
018eb7c1d2 Cancellation points no longer depend on EXPERIMENTAL 2016-12-09 17:44:37 -06:00
Gregory Nutt
a76e729b0d Trivial, cosmetic 2016-12-09 17:27:57 -06:00
Gregory Nutt
113d8bdcca Fix some SMP-related compilation errors 2016-12-09 17:10:59 -06:00
Gregory Nutt
78cdc9f113 Fix a typo in a debug assertion. 2016-12-09 16:59:10 -06:00
Gregory Nutt
7ee0862e67 Merge remote-tracking branch 'origin/master' into cancelpt 2016-12-09 16:54:15 -06:00
Gregory Nutt
b07964461e pthread_mutex_destroy(): Fix an error in destorynig a mutex which can occur after a pthread has been canceled while holding the mutex. 2016-12-09 16:53:29 -06:00
Gregory Nutt
7fce8022c6 Finishes all cancellation point logic 2016-12-09 16:50:34 -06:00
Gregory Nutt
d8783a7345 Update Documentation 2016-12-09 14:41:54 -06:00
Gregory Nutt
3eba0acb1c More cancellation points. 2016-12-09 13:49:36 -06:00
Gregory Nutt
c9ca97b4b5 cancellation points are basically function. More tested is needed and additional cancellation points must be implemented before this can be merged back to master. 2016-12-09 12:01:18 -06:00
Gregory Nutt
018db84567 Flesh out more cancellation point logic. 2016-12-09 10:31:40 -06:00
Gregory Nutt
d35e589d56 Flesh basic cancellation point support 2016-12-09 09:44:23 -06:00
Gregory Nutt
82a79b9c1b Add framework for cancellation point support. 2016-12-09 08:13:28 -06:00
Gregory Nutt
acaae12e8b Add pthread_testcancel(), pthread_testcancel(), and definitiions for cancellation types. 2016-12-09 07:23:00 -06:00
Gregory Nutt
6224e47533 pthread cleanup stack: Replace critical section with sched_lock/unlock(). The cleanup stack modification only needs to have the TCB stationary. The stack is never modified from interrupt level logic 2016-12-08 14:33:02 -06:00
Gregory Nutt
ab43681f15 Update TODO and some comments. 2016-12-08 10:24:40 -06:00
Gregory Nutt
a1fbc2ad0d pthreads: Add pthread_cleanup_push() and pthread_cleanup_pop() 2016-12-08 09:27:13 -06:00
Gregory Nutt
7632dfd6c7 Update some comments. 2016-12-08 07:52:02 -06:00
Gregory Nutt
a7b688e87b sched notes: Add additional note to see if/when CPU is started in SMP mode. 2016-12-07 09:08:20 -06:00
Gregory Nutt
9ed0387379 Olimex-LPC1766-STK: Enable procfs in NSH configuration. Automount /proc on startup. 2016-12-05 08:52:40 -06:00
Masayuki Ishikawa
13c9031a8d Fix DEBUGASSERT() in group_signal.c 2016-12-04 06:52:08 -06:00
Gregory Nutt
b78c3fcc30 Finish backing out previous change 2016-11-30 08:13:43 -06:00
Gregory Nutt
e6eb2e8bfa Back out the last change. I just noticed some complexities that need to be verified before this is released onto the world. 2016-11-30 07:41:48 -06:00
Gregory Nutt
8453343426 scheduler instrumentation: Add a little more protection for the SMP case 2016-11-30 07:24:15 -06:00
Gregory Nutt
9ee3f3b933 sched_note: Permit spinlock and critical section notes in in-memory buffer iff sched_not_get() interfaces is disabled. 2016-11-28 18:36:26 -06:00
Gregory Nutt
00215fbc98 sched_note: Add spinlock instrumentation; In SMP configurations, select to log only notes from certain CPUs 2016-11-28 10:33:46 -06:00
Gregory Nutt
d65be718c2 sched_note: Extend OS instrumentation to include some SMP events. 2016-11-27 17:14:57 -06:00
Gregory Nutt
e3fe320e08 SMP: Add support for linking spinlocks into a special, non-cached memory region. 2016-11-26 08:47:03 -06:00
Gregory Nutt
1d06e786e1 SMP: Clean-up and simplication of logic that I implemented late last night. 2016-11-26 07:05:27 -06:00
Gregory Nutt
a0e1af2614 SMP: Fix yet another potential deadlock 2016-11-25 23:04:27 -06:00
Gregory Nutt
eb9f8074c0 Update comments 2016-11-24 09:56:43 -06:00
Gregory Nutt
7bec4ffeec Update some comments 2016-11-23 17:40:01 -06:00
Gregory Nutt
d4037a30aa Update some comments 2016-11-23 13:20:18 -06:00
Gregory Nutt
f90525a5d1 SMP: Update some comments; trivial improvement by inlining static function. 2016-11-22 16:48:57 -06:00
Gregory Nutt
bac7153609 SMP: Add logic to avoid a deadlock condition when CPU1 is hung waiting for g_cpu_irqlock and CPU0 is waitin for g_cpu_paused 2016-11-22 11:34:16 -06:00
Gregory Nutt
130bfa3f6b Remove a assertion condition that appears to rarely cause false-alarm assertions. Teported by Petteri Aimonen 2016-11-21 14:43:56 -06:00
Gregory Nutt
558784d06f Spinlocks: Added capability to provide architecture-specific memory barriers. This was for i.MX6 but does not help with the SMP problems. It is still a good feature. 2016-11-21 11:55:59 -06:00
Gregory Nutt
2d1765a014 Remove more of those annoying empty comment blocks. 2016-11-21 07:34:27 -06:00
Gregory Nutt
0029c04dad task_restart: Make sure new task starts with pre-emption disabled and not in a critical section. 2016-11-21 07:33:23 -06:00
Gregory Nutt
cee8d59b58 Update TODO list 2016-11-20 12:26:08 -06:00
Gregory Nutt
e24f281401 This commit adds a new internal interfaces and fixes a problem with three APIs in the SMP configuration. The new internal interface is sched_cpu_pause(tcb). This function will pause a CPU if the task associated with 'tcb' is running on that CPU. This allows a different CPU to modify that OS data stuctures associated with the CPU. When the other CPU is resumed, those modifications can safely take place. 
The three fixes are to handle cases in the SMP configuration where one CPU does need to make modifications to TCB and data structures on a task that could be running running on another CPU.  Those three cases are task_delete(), task_restart(), and execution of signal handles.  In all three cases the solutions is basically the same:  (1) Call sched_cpu_pause(tcb) to pause the CPU on which the task is running, (2) perform the necessary operations, then (3) call up_cpu_resume() to restart the paused CPU.
 2016-11-20 07:57:18 -06:00
Gregory Nutt
f40fbaa4b9 sched/task: task_restart() test not supported on SMP systems. This is not fully implemented. 2016-11-19 11:41:05 -06:00
Rajan Gill
0d8e191d69 sched/clock: Correct calculation for the case of Tickless mode with a 32-bit timer. In that case, the calculation was returning millisecond accuracy. That is not good when the timer accuracy is < 1 msec. 2016-11-19 09:55:21 -06:00
Gregory Nutt
5ca9128c38 Update comments 2016-11-19 08:33:55 -06:00
Gregory Nutt
ceacacbc63 vfork(): Fix a race condition in the SMP case. Existing logic depended on the fact that the child would not run until waitpid was called because the child had the same priority as the parent. BUT in the SMP case that is not true... the child may run immediately on a different CPU. 2016-11-19 07:30:01 -06:00
Gregory Nutt
69e9f8638d Most interrupt handling logic interacts with tasks via standard mechanism such as sem_post, sigqueue, mq_send, etc. This all call enter_critical_section and are assumed to be safe in the SMP case. 
But certain logic interacts with tasks in different ways.  The only one that comes to mind are wdogs.  There is a tasking interface that to manipulate wdogs, and a different interface in the timer interrupt handling logic to manage wdog expirations.

In the normal case, this is fine.  Since the tasking level code calls enter_critical_section, interrupts are disabled an no conflicts can occur.  But that may not be the case in the SMP case.  Most architectures do not permit disabling interrupts on other CPUs so enter_critical_section must work differently:  Locks are required to protect code.

So this change adds locking (via enter_critical section) to wdog expiration logic for the the case if the SMP configuration.
 2016-11-18 13:57:30 -06:00
Gregory Nutt
78fd358023 Update some comments. 2016-11-18 08:20:52 -06:00
Gregory Nutt
8602e8a8a9 SMP: irq_csection() has a bad assumption. It assumed that the state of certain variables. That was true on entry into the interrupt handler, but might change to the execution of logic within the interrupt handler. 2016-11-18 07:38:16 -06:00
Gregory Nutt
8e029f019b Update comments 2016-11-17 17:30:12 -06:00
Gregory Nutt
47b52a2633 Fix an error in manually bringing in Sebastien's changes; Update some comments. 2016-11-17 16:15:06 -06:00
Gregory Nutt
9acd57c819 Fix up some naming and update some comments. 2016-11-17 15:50:39 -06:00
Gregory Nutt
558f4049d4 sched/irq/irq_csection: Fix a bad DEBUG assertion when SMP is enabled. 2016-11-17 15:36:27 -06:00
Gregory Nutt
2c314464a8 Fix a typo in a comment 2016-11-17 08:03:45 -06:00
Gregory Nutt
ccdc11d8ff Fix typo in variable naming. g_cpu_irqset should be g_cpu_irqlock. 2016-11-17 06:37:24 -06:00
Gregory Nutt
1c1564f0c1 Add logic to handled nested calls to enter_critical_section() from interrupts handlers (with SMP). 2016-11-16 22:45:08 -06:00
Gregory Nutt
0c3207e86e Fix backward logic in an assertion. 2016-11-16 21:48:27 -06:00
Gregory Nutt
16f0b8fa96 Fix typos in DEBUASSERT statements 2016-11-16 19:58:51 -06:00
Gregory Nutt
f1e4951a8d Fix some comments 2016-11-16 19:11:31 -06:00
Gregory Nutt
4ca653a1cc Add some experimental changes to enter/leave_critical_section to deal with the case where interrupts are disabled only on the local CPU 2016-11-16 18:28:35 -06:00
Gregory Nutt
5ce3b399d5 Correct some typos in comments 2016-11-15 16:48:40 -06:00
Gregory Nutt
40d7216a22 Back out 65ab12 and parts of 21f92b 2016-11-15 15:44:43 -06:00
Gregory Nutt
6683f01444 Trivial change to comment 2016-11-15 13:38:43 -06:00
Gregory Nutt
21f92ba601 Review some SMP logic; update comments; refresh configuration. 2016-11-15 09:24:00 -06:00
Gregory Nutt
65ab12fbb9 If SMP is enabled, if any interrupt handler calls enter_critical_section(), it should take the spinlock. 2016-11-15 08:37:58 -06:00
Gregory Nutt
b53866c872 sched/irq: Change order for SMP case in enter_critical_section: Disble local interrupts BEFORE taking spinlock. 2016-11-15 07:54:18 -06:00
Freddie Chopin
484a1b6104 sem_wait() and sem_trywait() no longer modify the errno value UNLESS an error occurs. This allows these functions to be used internallly without clobbering the errno value. 2016-11-09 07:01:49 -06:00