Commit Graph

1210 Commits

Author SHA1 Message Date
Masayuki Ishikawa
d295f11a3a SMP: Introduce a new global IRQ clearing logic and tasklist protection.
The previous implementation of clearing global IRQ in sched_addreadytorun()
and sched_removereadytorun() was done too early. As a result, nxsem_post()
would have a chance to enter the critical section even nxsem_wait() is
still not in blocked state. This patch moves clearing global IRQ controls
from sched_addreadytorun() and sched_removereadytorun() to sched_resumescheduler()
to ensure that nxsem_post() can enter the critical section correctly.

For this change, sched_resumescheduler.c is always necessary for SMP configuration.
In addition, by this change, task_exit() had to be modified so that it calls
sched_resumescheduler() because it calls sched_removescheduler() inside the
function, otherwise it will cause a deadlock.

However, I encountered another DEBUGASSERT() in sched_cpu_select() during
HTTP streaming aging test on lc823450-xgevk. Actually sched_cpu_select()
accesses the g_assignedtasks which might be changed by another CPU. Similarly,
other tasklists might be modified simultaneously if both CPUs are executing
scheduling logic. To avoid this, I introduced tasklist protetion APIs.

With these changes, SMP kernel stability has been much improved.

Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
2018-01-31 15:35:36 +09:00
Gregory Nutt
3521aaf944 Squashed commit of the following:
binfmt/, configs/, grahics/, libc/, mm/, net/, sched/:  OS references to the errno variable should always use the set_errno(), get_errno() macros
    arch/arm/src/stm32 and stm32f7:  Architecture-specific code is not permitted to modify the errno variable.  drivers/ and libc/:  OS references to the errno variable should always use the set_errno(), get_errno() macros
2018-01-30 17:57:36 -06:00
Gregory Nutt
82982f7972 sched/: Fix several inappropriate accesses to get_errno() that were missed in previous changes (some going back to nuttx-.23). Add new nxsched_setscheduler() and nxsched_getaffinity() which are equivalent to their counterparts without the nx on front. These versions do not modify the errno value. Changed all calls within the OS to use these newer versions of the functions. 2018-01-30 16:16:41 -06:00
Gregory Nutt
170a50c690 Squashed commit of the following:
sched/sched:  Correct some build issues introduced by last set of changes.
    sched/sched:  Add new internal OS function nxsched_setaffinity() that is identical to sched_isetaffinity() except that it does not modify the errno value.  All usage of sched_setaffinity() within the OS is replaced with nxsched_setaffinity().
    sched/sched:  Internal functions sched_reprioritize() and sched_setpriority() no longer movidify the errno value.  Also renamed to nxsched_reprioritize() and sched_setpriority().
    sched/sched:  Add new internal OS function nxsched_getscheduler() that is identical to sched_getscheduler() except that it does not modify the errno value.  All usage of sched_getscheduler() within the OS is replaced with nxsched_getscheduler().
    sched/sched:  Add new internal OS function nxsched_setparam() that is identical to sched_setparam() except that it does not modify the errno value.  All usage of sched_setparam() within the OS is replaced with nxsched_setparam().
    sched/sched:  Add new internal OS function nxsched_getparam() that is identical to sched_getparam() except that it does not modify the errno value (actually, the previous value erroneously neglected to set the errno value to begin with, but this fixes both issues).  All usage of sched_getparam() within the OS is replaced with nxsched_getparam().
2018-01-30 11:08:18 -06:00
Gregory Nutt
12e3e47d3c Update comments, mostly spelling errors. 2018-01-27 09:37:46 -06:00
Gregory Nutt
bcffeae4fd Update README and some comments 2018-01-20 12:59:56 -06:00
Gregory Nutt
82cc0ead67 sched/sched: Fix some priority inheritance related issues noted during review of logic. Also add some REVISIT comments for some issues noted in the design. 2018-01-20 07:44:35 -06:00
Gregory Nutt
be3c2d1e36 Change ASSERT(false) to DEBUGPANIC(). 2018-01-19 09:07:19 -06:00
Gregory Nutt
f012fe96cb sched/sched/sched_sporadic.c: fix compiler error when priority inheritance is enabled. Noted by eunb.song@samsung.com 2018-01-19 08:54:29 -06:00
Masayuki Ishikawa
99b944fb6e sched/semaphore: Add ASSERT(false) in nxsem_post() if no waiting task is found.
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
2018-01-19 22:17:58 +09:00
Masayuki Ishikawa
ab3fa89023 SMP: Introduce spin_lock_wo_note() and spin_unlock_wo_note()
These APIs are used in sched_note.c to protect instumentation data.
The deffrence between these APIs to exsiting spin_lock() and spin_unlock()
is that they do not perform insturumentation to avoid recursive call
when SCHED_INSTRUMENTATION_SPINLOCKS=y.

Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
2018-01-17 13:29:42 +09:00
Gregory Nutt
e238c8b090 sched/clock/clock_systimer.c: Change the way that the 64-bit time is sampled. Previously, we disabled interrupts before sampling the 64-bit timer since the uint64_t access is not atomic on most CPUs. However, disabling (local) interrupts does not work in the SMP case. In that case, the timer interrupt will be running on only one of the CPUs; disabling interrupts on a different CPU will provide no protection from timer rollover. To work around this, logic was added that samples 64-bit timer is sampled twice and if 32-bit rollover was detected between samples, then loops until there is no rollover. 2018-01-15 11:05:45 -06:00
Masayuki Ishikawa
0ecef4dbcb clock_systimer: Replace critical section APIs with spinlock APIs. (64bit only)
If SMP=n or SMP=y && SPINLOCK_IRQ=n, this works in the same way as before.
If SMP=y && SPINLOCK_IRQ=y, performance will be improved.

Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
2018-01-15 15:08:11 +09:00
Gregory Nutt
134bac3810 arch/arm/src/lpc54xx: Add configuration logic to the empty SPI driver. Still missing all data tranfer logic. 2018-01-13 15:38:55 -06:00
Gregory Nutt
fdf4e40b6a configs/lpcxpresso-lpc54628: Correct SRAM base address in all configurations. 2018-01-13 08:34:58 -06:00
Gregory Nutt
0080225473 sched/irq: Add a configuration option to show interrupt information via a procfs file. 2018-01-12 18:26:46 -06:00
Gregory Nutt
9f80e4ccf1 sched/: Various fixes for typos, improved parameter verification. 2018-01-12 18:26:45 -06:00
Gregory Nutt
7d131fe711 Update TODO list 2018-01-03 08:32:57 -06:00
Gregory Nutt
b2adb4917f arch/arm/src/lpc54xx: Add register level debug output for Ethernet testing. Fixed a few start up problems. Still hangs on start-up, however. 2017-12-31 11:11:57 -06:00
Masayuki Ishikawa
e1f71f988b sched/semaphore/spinlock.c: Add memory barrier operations in spin_unlock()
In ARM document regarding memory barrires, SP_DMB() must be issued
before changing a spinlock state to SP_UNLOCKED. However, we found
that SP_DSB() is also needed to ensure that spin_unlock() works
correctly for network streaming aging test.

Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
2017-12-21 17:40:08 +09:00
Masayuki Ishikawa
27d4bb43e6 Merged in masayuki2009/nuttx.nuttx/lc823450_smp_test (pull request #557)
sched/sched: Remove DEBUGASSERT() in sched_mergepending()

Because this DEBUGASSERT() assumes that while loop executes only once.

Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>

Approved-by: Gregory Nutt <gnutt@nuttx.org>
2017-12-18 12:09:06 +00:00
Masayuki Ishikawa
6150299f54 Merged in masayuki2009/nuttx.nuttx/irq_spinlock (pull request #550)
SMP: Introduce spin_lock_irqsave() and spin_unlock_irqrestore()

These APIs are simplified version of enter_critical_section() and
leave_critical_section() to protect data (e.g. registers) in SMP mode.
By using these APIs inside drivers, performace will be improved.

Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>

Approved-by: Gregory Nutt <gnutt@nuttx.org>
2017-12-07 19:39:04 +00:00
Juha Niskanen
e8ee632378 sched/task/task_exithook.c: Clear atexit() function pointer before calling it. On most archs, up_assert() calls exit() so without this change, if atexit() function triggers an assertion we are in endless loop. 2017-11-30 06:41:57 -06:00
Masayuki Ishikawa
5acd26c88f sched/semaphore/spinlock.c: Disable local interrupts in spin_setbit() and spin_clrbit() in order to avoid a deadlock condition. 2017-11-23 06:55:53 -06:00
Gregory Nutt
b9fa3e167e Changes from review of clock_nanosleep(): Misplaced right bracket, but return value in one failure case. 2017-11-11 18:01:28 -06:00
Gregory Nutt
35ee844590 sched/signal/sig_nanosleep.c and libc/time/lib_nanosleep.c: Implement clock_nanosleep(). nanosleep() is now reduced to a libc wrapper around clock_nanosleep(). 2017-11-11 17:44:59 -06:00
Juha Niskanen
cc7af2b0d0 procfs: Fix uptime being close to maximum 32-bit value in certain config 2017-11-10 09:01:39 -06:00
Gregory Nutt
93f7cd2d53 sched/sched/sched_cpuselect.c: For SMP, In order to find the cpu with the lowest priority thread, we have to remember the already found lowest priority. Noted by Anonymous in Issue #75 2017-11-06 13:33:46 -06:00
Masayuki Ishikawa
09d4874509 Merged in masayuki2009/nuttx.nuttx/lc823450 (pull request #524)
Fix DEBUGASSERT() issues with nxhello on lc823450-xgevk

* sched/task: Remove DEBUGASSERT in task_exitstatus() and task_groupexit()

* graphics: Change DEBUGASSERT condition in nx_runinstance()

Approved-by: Gregory Nutt <gnutt@nuttx.org>
2017-10-31 16:53:00 +00:00
Jussi Kivilinna
8eaa587c98 sched: move POSIX thread specific data from pthread TCB to common TCB structure. This change allows using pthread_getspecific/pthread_setspecific from main thread. Patch also enables using pthread data with config option CONFIG_DISABLE_PTHREAD=y. 2017-10-25 07:06:42 -06:00
Gregory Nutt
5b385f4d4d kthread_create(): Rename kernel_thread() to kthread_create() for better naming consistency with task_create() and kthread_delete(). 2017-10-16 11:38:00 -06:00
Gregory Nutt
b1a41b7620 task_delete(): Do not permit user applications to delete kernel threads. 2017-10-16 09:07:27 -06:00
Gregory Nutt
3c3c712887 nxsig_timedwait: Add logic to suppress the wait if there is a pending cancellation. 2017-10-12 09:53:17 -06:00
Gregory Nutt
fd58bfd9c0 sched/signal: Add logic to wake up a thread that is waiting on a signal if it is canceled. 2017-10-12 08:55:19 -06:00
Gregory Nutt
3cd41b0408 sched/task: task_exithook.c fails to link if signals are disabled because was unconditionally trying to send the SIGCHLD signal to the parent in certain configurations. Noted by Jeongchan Kim. 2017-10-11 10:43:30 -06:00
Gregory Nutt
eb79a575f6 Fix a few places where there was a semicolon following the 'if' condition, makeing the following logic unconditional. 2017-10-10 14:24:13 -06:00
Gregory Nutt
e11e3b2607 Squashed commit of the following:
Change all calls to mq_receive() and mq_timedreceive() in the OS to calls to nxmq_receive() and nxmq_timedreceive(), making appropriate changes for differences in return values.

    sched/mqueue:  Add nxmq_receive() and mxmq_timedreceive() which are functionally equivalent to the standard mq_receive and mq_timedreceive() except that (1) they do not create cancellation points, and (2) the do not modify the application's errno variable.
2017-10-10 09:57:40 -06:00
Gregory Nutt
fca07be1df Squashed commit of the following:
Change all calls to mq_send() and mq_timedsend() in the OS to calls to nxmq_send() and nxmq_timedsend(), making appropriate changes for differences in return values.

    sched/mqueue:  Add internal function nxmq_send() and nxmq_timedsend() that are equivalent to mq_send() and mq_timedsend() except that they do not create cancellation points and do to not modify the errno variable.
2017-10-10 08:44:12 -06:00
Gregory Nutt
9b61cc6b25 task: Add new cancellation point interface, check_cancellation_point(). 2017-10-09 09:43:04 -06:00
Gregory Nutt
e7c52bac60 Squashed commit of the following:
sched/mqueue:  Rename all private static functions for use the nxmq_ vs. mq_ naming.

    sched/mqueue:  Rename all OS internal functions declared in sched/mqueue/mqueue.h to begin with nxmq_ vs. mq_.  The mq_ prefix is reserved for standard application interfaces.
2017-10-09 09:06:46 -06:00
Gregory Nutt
a857cc04e4 Fix some build problems after recent separation of internal OS from application interfaces. The build problem only occurs in the PROTECTED and KERNEL builds where separate libraries are built for the applications and for use within the OS. In these cases, the correct interfaces must be used. This commit fixes a few of these, so I can get through build testing, but there are many more that need fixin'. 2017-10-08 08:13:47 -06:00
Gregory Nutt
d633c9bf8c Squashed commit of the following:
Replace all calls to sigprocmask() in the OS proper with calls to nxsig_procmask().

    sched/signal:  Add internal OS interface nxsig_procmask().  This internal interface is equivalent to the standard sigprocmask() used by applications except that it does not modify the errno value.  Also fixes a problem in that the original sigprocmask() was not setting the errno.
2017-10-07 12:16:10 -06:00
Gregory Nutt
4810499d3a Squashed commit of the following:
Replace all calls to sigqueue() in the OS proper with calls to nxsig_queue() to avoid accessing the errno variable.

    sched/signal:  Add nxsig_queue() which is functionally equivalent to sigqueue() except that it does not modify the errno variable.
2017-10-07 10:57:09 -06:00
Gregory Nutt
a20e4e7eef Eliminate a warning introduced with last commit. 2017-10-07 08:51:31 -06:00
Gregory Nutt
9e25d89223 Squashed commit of the following:
Replace all usage kill() in the OS proper with nxsig_kill().

    sched/signal:  Add nxsig_kill() which is functionally equivalent to kill() except that it does not modify the errno variable.
2017-10-07 08:22:18 -06:00
Gregory Nutt
936df1bcb5 Adds new OS internal functions nxsig_sleep() and nxsig_usleep. These differ from the standard sleep() and usleep() in that (1) they don't cause cancellation points, and (2) don't set the errno variable (if applicable). All calls to sleep() and usleep() changed to calls to nxsig_sleep() and nxsig_usleep().
Squashed commit of the following:

    Change all calls to usleep() in the OS proper to calls to nxsig_usleep()

    sched/signal:  Add a new OS internal function nxsig_usleep() that is functionally equivalent to usleep() but does not cause a cancellaption point and does not modify the errno variable.

    sched/signal:  Add a new OS internal function nxsig_sleep() that is functionally equivalent to sleep() but does not cause a cancellaption point.
2017-10-06 10:15:01 -06:00
Gregory Nutt
fdd0dcc0b6 This commit adds internal versions of the signal interfaces:
sigtimedwait() -> nxsig_timedwait()
  sigwaitinfo()  -> nxsig_waitinfo()
  nanosleep()    -> nxsig_nanosleep()

The internal OS versions differ from the standard application interfaces in that:

  - They do not create cancellation points, and
  - they do not modify the application's errno variable

Squashed commit of the following:

    sched/signal:  Replace all usage of sigwaitinfo(), sigtimedwait(), and nanosleep() with the OS internal counterparts nxsig_waitinfo(), nxsig_timedwait(), and nxsig_nanosleep().

    sched/signal:  Add nxsig_nanosleep().  This is an internal OS version of nanosleep().  It differs in that it does not set the errno varaiable and does not create a cancellation point.

    sched/signal:  Add nxsig_timedwait() and nxsig_waitinfo().  These are internal OS versions of sigtimedwait() and sigwaitinfo().  They differ in that they do not set the errno varaiable and they do not create cancellation points.
2017-10-06 08:28:20 -06:00
Gregory Nutt
21c97b16cf Revert "sem_wait.c edited online with Bitbucket. Fix some correct but useless code."
What was I thinking?  I missed that litle minus sign and the possibility that the errno might be some positive non-zero value.

This reverts commit 43880878e4.
2017-10-05 16:01:01 -06:00
Gregory Nutt
43880878e4 sem_wait.c edited online with Bitbucket. Fix some correct but useless code. 2017-10-05 21:55:24 +00:00
Gregory Nutt
8198ba6a6d This change renames all internal, private NuttX signal-related functions to use the prefix nxsig_ so that they cannot be confused with application interfaces that begin, primarily, with sig_
This is analogous to similar renaming that was done previously for semaphores.

Squashed commit of the following:

    sched/signal:  Fix a few compile warnings introduced by naming changes.

    sched/signal:  Rename all private, internal signl functions to use the nxsig_ prefix.

    sched/signal:  Rename sig_removependingsignal, sig_unmaskpendingsignal, and sig_mqnotempty to nxsig_remove_pendingsignal, nxsig_unmask_pendingsignal, and nxsig_mqnotempty to make it clear that these are OS internal interfaces.

    sched/signal:  Rename sig_findaction and sig_lowest to nxsig_find_action and nxsig_lowest to make it clear that these are OS internal interfaces.

    sched/signal:  Rename sig_allocatepingsigaction and sig_deliver to nxsig_alloc_pendingsigaction and nxsig_deliver to make it clear that these are OS internal interfaces.

    sched/signal:  Rename sig_cleanup, sig_release, sig_releasependingaction, and sig_releasependingsignal to nxsig_cleanup, nxsig_release, nxsig_release_pendingaction, and nxsig_release_pendingsignal to make it clear that these are OS internal interfaces.

    sched/signal:  Rename sig_tcbdispatch and sig_dispatch to nxsig_tcbdispatch and nxsig_dispatch to make it clear that these are OS internal interfaces.

    sched/signal:  Rename sig_releaseaction and sig_pendingset to nxsig_release_action and nxsig_pendingset to make it clear that these are OS internal interfaces.

    sched/signal:  Rename sig_initialize and sig_allocateactionblock to nxsig_initialize and nxsig_alloc_actionblock to make it clear that these are OS internal interfaces.
2017-10-05 13:25:25 -06:00