Commit Graph

266 Commits

Author SHA1 Message Date
Masayuki Ishikawa 1d958980bd Merged in masayuki2009/nuttx.nuttx/fix_smp_bugs (pull request #615)
Fix SMP related bugs

* sched/sched: Fix a deadlock in SMP mode

    Two months ago, I introduced sched_tasklist_lock() and
    sched_tasklist_unlock() to protect tasklists in SMP mode.
    Actually, this change works pretty well for HTTP audio
    streaming aging test with lc823450-xgevk.

    However, I found a deadlock in the scheduler when I tried
    similar aging tests with DVFS autonomous mode where CPU
    clock speed changed based on cpu load. In this case, call
    sequences were as follows;

    cpu1: sched_unlock()->sched_mergepending()->sched_addreadytorun()->up_cpu_pause()
    cpu0: sched_lock()->sched_mergepending()

    To avoid this deadlock, I added sched_tasklist_unlock() when calling
    up_cpu_pause() and sched_addreadytorun(). Also, added
    sched_tasklist_lock() after the call.

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

* libc: Add critical section in lib_filesem.c for SMP

    To set my_pid into fs_folder atomically in SMP mode,
    critical section API must be used.

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

* mm: Add critical section in mm_sem.c for SMP

    To set my_pid into mm_folder atomically in SMP mode,
    critical section API must be used.

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

* net: Add critical section in net_lock.c for SMP

    To set my pid (me) into fs_folder atomically in SMP mode,
    critical section API must be used.

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

Approved-by: Gregory Nutt <gnutt@nuttx.org>
2018-03-20 12:34:38 +00:00
Gregory Nutt b54ffe858a Standardization of some function headers. 2018-03-13 09:52:27 -06:00
Dmitriy Linikov a8c58607e9 Merged in hardlulz/modem-3.0-nuttx/fix-sem-EINTR (pull request #603)
Added ECANCELED condition to DEBUGASSERT-s checking sem_wait result

Approved-by: Gregory Nutt <gnutt@nuttx.org>
2018-02-20 18:24:53 +00:00
Gregory Nutt 7cf88d7dbd Make sure that labeling is used consistently in all function headers. 2018-02-01 10:00:02 -06:00
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 fef255e5be This commit adds an as-of-yet untested implemented of UDP write buffering.
Squashed commit of the following:

    net/udp:  Address most of the issues with UDP write buffering.  There is a remaining issue with handling one network going down in a multi-network environment.  None of this has been test but it is certainly ready for test.  Hence, the feature is marked EXPERIMENTAL.
    net/udp:  Some baby steps toward a corrected write buffering design.
    net/udp:  Remove pesky write buffer macros.
    Eliminate trailing space at the end of lines.
    net/udp:  A little more UDP write buffering logic.  Still at least on big gaping hole in the design.
    net/udp:  Undefined CONFIG_NET_SENDTO_TIMEOUT.
    net/udp:  Crude, naive port of the TCP write buffering logic into UDP.  This commit is certainly non-functional and is simply a starting point for the implementatin of UDP write buffering.
    net/udp:  Rename udp/udp_psock_sendto.c udp/udp_psock_sendto_unbuffered.c.
2018-01-22 18:32:02 -06:00
Gregory Nutt fb6208fbbc mm: Fix a typo in a debug assertion. 2017-11-22 07:30:48 -06:00
Gregory Nutt 514ac3fe98 mm: Add a debug assertin to check for integer overflow in malloc. 2017-11-21 07:24:10 -06:00
Gregory Nutt ef89207c8d Build system: Fix CONFIG_BUILD_KERNEL logic directories that have ubin and kbin subdirectories. Conditional logic was fine for CONFIG_BUILD_FLAT and CONFIG_BUILD_PROTECTED but generated useless dependencies if CONFIG_BUILD_KERNEL. 2017-11-15 07:54:09 -06:00
Gregory Nutt fa4722a5ac mm/mm_gran: Combine some common logic into a function (also fixes a subtle bug). 2017-11-15 07:54:08 -06:00
Gregory Nutt 1191238b17 mm/mm_gran: Fix some issues found during test of the new gran_info() interface. 2017-11-14 16:13:20 -06:00
Gregory Nutt d720711807 fs/procfs: Add logic to show the state of the page allocator in /proc/meminfo. 2017-11-14 14:59:51 -06:00
Gregory Nutt 54fad8d04f mm/mm_gran: Add a function to get information about the state of the granuale allocator. 2017-11-14 14:41:03 -06:00
Gregory Nutt 62b8026976 Remove CONFIG_GRAN_SINGLE. It adds no technical benefit (other than some minor reduction in the number of interface arguments) but adds a lot of code complexity. Better without it. 2017-11-14 11:47:12 -06:00
Jussi Kivilinna 75b53d563b mm/mm-heap: memalign: fix heap corruption caused by using unaligned chuck size. Unaligned nodes generated by memalign later cause heap corruptions when nodes are shrink further (for example, 24 bytes -> 8 bytes, when alignment is 16 bytes). 2017-10-24 15:35:52 -06:00
Gregory Nutt 34a572b226 Update last commit... Check should really use the definition MMSIZE_MAX which is really the same thing, but guaranteed to be the correct maximum size in any present and future configuration. 2017-10-17 07:34:06 -06:00
EunBong Song 196911d4fa If size is greater than (UINT32_MAX - SIZEOF_MM_ALLOCNODE), malloc size can be overflow by MM_ALIGN_UP macro. For example, if task_create() called with stack_size == -1, up_create_stack() functions allocates SIZEOF_MM_ALLOCNODE bytes for stack.
This can cause data abort in up_stack_color() function.
2017-10-17 06:37:09 -06:00
Gregory Nutt 700f1a8e8c Eliminate some warnings found in build testing. 2017-10-08 16:27:17 -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 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 91537a7722 mm/: Heap semaphore logic needs to use nxsem_* interfaces when available, but the standard semaphores only when implementing a user-space heap. Not this does introduce and issue: the memory management functions them become cancellation points because of the use of sem_wait. 2017-10-07 07:16:33 -06:00
Gregory Nutt 7cc63f90d9 sched/semaphore: sem_trywait() modifies the errno value and, hence, should not be used within the OS. Use nxsem_trywait() instead. 2017-10-05 07:59:06 -06:00
Gregory Nutt 9568600ab1 Squashed commit of the following:
This commit backs out most of commit b4747286b1.  That change was added because sem_wait() would sometimes cause cancellation points inappropriated.  But with these recent changes, nxsem_wait() is used instead and it is not a cancellation point.

    In the OS, all calls to sem_wait() changed to nxsem_wait().  nxsem_wait() does not return errors via errno so each place where nxsem_wait() is now called must not examine the errno variable.

    In all OS functions (not libraries), change sem_wait() to nxsem_wait().  This will prevent the OS from creating bogus cancellation points and from modifying the per-task errno variable.

    sched/semaphore:  Add the function nxsem_wait().  This is a new internal OS interface.  It is functionally equivalent to sem_wait() except that (1) it is not a cancellation point, and (2) it does not set the per-thread errno value on return.
2017-10-04 15:22:27 -06:00
Gregory Nutt 42a0796615 Squashed commit of the following:
sched/semaphore:  Add nxsem_post() which is identical to sem_post() except that it never modifies the errno variable.  Changed all references to sem_post in the OS to nxsem_post().

    sched/semaphore:  Add nxsem_destroy() which is identical to sem_destroy() except that it never modifies the errno variable.  Changed all references to sem_destroy() in the OS to nxsem_destroy().

    libc/semaphore and sched/semaphore:  Add nxsem_getprotocol() and nxsem_setprotocola which are identical to sem_getprotocol() and set_setprotocol() except that they never modifies the errno variable.  Changed all references to sem_setprotocol in the OS to nxsem_setprotocol().  sem_getprotocol() was not used in the OS
2017-10-03 15:35:24 -06:00
Gregory Nutt 83cdb0c552 Squashed commit of the following:
libc/semaphore:  Add nxsem_getvalue() which is identical to sem_getvalue() except that it never modifies the errno variable.  Changed all references to sem_getvalue in the OS to nxsem_getvalue().

    sched/semaphore:  Rename all internal private functions from sem_xyz to nxsem_xyz.  The sem_ prefix is (will be) reserved only for the application semaphore interfaces.

    libc/semaphore:  Add nxsem_init() which is identical to sem_init() except that it never modifies the errno variable.  Changed all references to sem_init in the OS to nxsem_init().

    sched/semaphore:  Rename sem_tickwait() to nxsem_tickwait() so that it is clear this is an internal OS function.

    sched/semaphoate:  Rename sem_reset() to nxsem_reset() so that it is clear this is an internal OS function.
2017-10-03 12:52:31 -06:00
Jussi Kivilinna 81b5118727 mm_mallinfo: do heap end debug assert check with heap semaphore held 2017-08-03 10:01:26 -06:00
Gregory Nutt 98d937104e mm/: Remove dangling space at the end of lines. 2017-06-28 13:31:21 -06:00
Gregory Nutt d9bd5ca05f Update README and some C comments 2017-05-30 09:19:04 -06:00
Gregory Nutt e9c55d8f7d IOBs: Fix a typing error mm/iob/iob.h mm/iob/iob_initialize.c 2017-05-27 08:03:00 -06:00
Gregory Nutt 2c00825dcf Porting Guide: Add description of IOBs. 2017-05-20 08:50:05 -06:00
Masayuki Ishikawa f10e10e465 IOBs: Fix build break
Signed-off-by: Masayuki Ishikawa <masayuki.ishikawa@gmail.com>
2017-05-18 13:59:03 +09:00
Gregory Nutt 5ce2ece134 syslog: Add header file inclusion to eliminate a warning; mm/iob: private function needs static storage class. 2017-05-16 12:26:23 -06:00
Gregory Nutt 6a3800f611 There can be a failure in IOB allocation to some asynchronous behavior caused by the use of sem_post(). Consider this scenario:
Task A holds an IOB.  There are no further IOBs.  The value of semcount is zero.
Task B calls iob_alloc().  Since there are not IOBs, it calls sem_wait().  The v
alue of semcount is now -1.

Task A frees the IOB.  iob_free() adds the IOB to the free list and calls sem_post() this makes Task B ready to run and sets semcount to zero NOT 1.  There is one IOB in the free list and semcount is zero.  When Task B wakes up it would increment the sem_count back to the correct value.

But an interrupt or another task runs occurs before Task B executes.  The interrupt or other tak takes the IOB off of the free list and decrements the semcount.  But since semcount is then < 0, this causes the assertion because that is an invalid state in the interrupt handler.

So I think that the root cause is that there the asynchrony between incrementing the semcount.  This change separates the list of IOBs:  Currently there is only a free list of IOBs.  The problem, I believe, is because of asynchronies due sem_post() post cause the semcount and the list content to become out of sync.  This change adds a new 'committed' list:  When there is a task waiting for an IOB, it will go into the committed list rather than the free list before the semaphore is posted.  On the waiting side, when awakened from the semaphore wait, it will expect to find its IOB in the committed list, rather than free list.

In this way, the content of the free list and the value of the semaphore count always remain in sync.
2017-05-16 11:03:35 -06:00
Gregory Nutt 0de294a586 Fix lots of occurrences of 'the the', 'the there', 'the these', 'the then', 'the they. 2017-05-11 13:35:56 -06:00
Gregory Nutt d3b9f5b37f Syslog buffering: Use IOBs to buffer data, not an on-stack buffer 2017-05-10 17:36:08 -06:00
Gregory Nutt de6bffe713 Update some README files 2017-05-09 11:32:44 -06:00
Gregory Nutt 2043e1a114 IOBs: Move from driver/iob to a better location in mm/iob 2017-05-09 07:35:30 -06:00
Gregory Nutt 0064dc52e5 realloc(): When realloc() has to fall back to calling malloc(), size including overhead was being provided to malloc(), causing a slightly larger allocation than needed. Noted by initialkjc@yahoo.com 2017-03-28 07:23:46 -06:00
Gregory Nutt 6f1c5e7b43 Add some comments. 2016-11-05 09:44:29 -06:00
Gregory Nutt 43eb04bb8f Without lowsyslog() *llinfo() is not useful. Eliminate and replace with *info(). 2016-06-20 11:59:15 -06:00
Gregory Nutt 0c8c7fecf0 Add _ to the beginning of all debug macros to avoid name collisions 2016-06-16 12:33:32 -06:00
Gregory Nutt cf4075c741 Missed macro definition in one case 2016-06-12 11:13:34 -06:00
Gregory Nutt 823b4b0cff Forget to define an info() macro 2016-06-12 11:11:08 -06:00
Gregory Nutt b29a4dd49c audio/, crypto/, libnx/, and mm/: Change some err() ERRORS to warn() WARNINGS or info() 2016-06-12 09:46:23 -06:00
Gregory Nutt a1469a3e95 Add CONFIG_DEBUG_ERROR. Change names of *dbg() * *err() 2016-06-11 15:50:49 -06:00
Gregory Nutt e99301d7c2 Rename *lldbg to *llerr 2016-06-11 14:55:27 -06:00
Gregory Nutt 86b79b33cf Reserver the name 'err' for other purposes 2016-06-11 14:40:07 -06:00
Gregory Nutt 1cdc746726 Rename CONFIG_DEBUG to CONFIG_DEBUG_FEATURES 2016-06-11 14:14:08 -06:00
Gregory Nutt fc3540cffe Replace all occurrences of vdbg with vinfo 2016-06-11 11:59:51 -06:00
Paul A. Patience f8f7b7582c Silence some warnings 2016-05-26 14:06:13 -04:00
Gregory Nutt c4493528a1 Remove some empty file section section header comments 2016-02-17 18:57:35 -06:00
Gregory Nutt 241ab98c5b Remove some empty file section section header comments 2016-02-17 18:31:56 -06:00
Gregory Nutt fccef49305 Cosmetic changes to comments 2016-02-15 13:48:08 -06:00
Gregory Nutt 046e39e2c6 nuttx/mm and libc: Replace irqsave() with enter_critical_section(); replace irqrestore() with leave_critical_section() 2016-02-14 08:57:01 -06:00
Gregory Nutt f45166af32 SMP: Fix a IDLE task semaphore operation 2016-02-12 18:03:08 -06:00
Gregory Nutt 40659d12df Rename libxx_internal.h to libxx.h 2015-12-30 07:56:56 -06:00
Gregory Nutt b90da3f27b waitpid: CRITICAL BUGFIX. Add a reference counting mechansim to prevent wait from using stale memory that was freed by the exiting task 2015-12-22 11:48:17 -06:00
Paul A. Patience bfc95c641f stdlib: Fix forgotten FARs 2015-11-18 14:22:43 -05:00
Gregory Nutt 06cd5b2fcc mm, audio, crypto, graphics: Fix various spacing/alignment issues 2015-10-08 09:10:22 -06:00
Gregory Nutt af086c40ff Remove dangling whitespace 2015-10-04 15:28:54 -06:00
Gregory Nutt 0b12dbf95d Fix some spacing problems 2015-10-04 15:04:00 -06:00
Gregory Nutt 5b51a9fcdd Standardize the width of all comment boxes in C files 2015-10-02 17:43:18 -06:00
Gregory Nutt cb9e27c3b0 Standardize naming used for public data and function groupings 2015-10-02 16:30:35 -06:00
Paul A. Patience 9f108b7b63 Fix issues detected by clang 2015-09-10 20:59:43 -04:00
Gregory Nutt 07e9a02af9 Forgot to add two files in previous commit 2015-07-10 08:46:13 -06:00
Gregory Nutt 59cc4a7a7b Protected mode: Redesign how the user space heap is accessed from the kernel code. It used to call memory management functions in user space via function pointers in the userspace interface. That is inefficient because the first thing that those memory management functions do is to trap back into the kernel to get the current PID. Worse, that operation can be fatal is certain fragile situations such as when a task is exitting.
The solution is to remove all of the memory management function calls from the interface.  Instead, the interface exports the userspace heap structure and then kernel size implementations of those memory management functions will operate on the userspace heap structure.  This avoids the unnecessary system calls and, more importantly, failures do to freeing memory when a test exits.
2015-07-10 08:37:02 -06:00
Gregory Nutt 342f5fe33d Fix references to the no-longer-existent misc/ directory in comments, README files, and documentation 2015-06-28 08:08:57 -06:00
Gregory Nutt 0476d8168a More missing semicolons after DEBUGASSERT 2015-05-27 13:32:39 -06:00
Gregory Nutt 2aa80e06f5 Remove executable flag from more .c and .h files 2015-04-09 08:20:57 -06:00
Gregory Nutt a4a6917267 Fixes some issues found by the PX4 team using Coverity. From Pavel Krienko 2015-01-19 13:01:20 -06:00
Gregory Nutt 9ac09db800 More fixes to problems noted by cppcheck. Some are kind of risky; some are real bugs. 2014-11-25 13:46:14 -06:00
Gregory Nutt f06289e4d5 If the INTR granule allocator mode is enabled, there is no
semaphore to destroy.  From Lorenz Meier.
2014-11-12 06:56:32 -06:00
Gregory Nutt 8634a5e41c Fix a typo in a Kconfig file 2014-09-27 16:11:48 -06:00
Gregory Nutt b33c2d9cef Move include/nuttx/mm.h to include/nuttx/mm/mm.h 2014-09-24 07:29:09 -06:00
Gregory Nutt 1863370672 Move include/nuttx/shm.h to include/nuttx/mm/shm.h 2014-09-24 07:05:02 -06:00
Gregory Nutt 2fa7431ee7 Move include/nuttx/gran.h to include/nuttx/mm/gran.h 2014-09-24 06:55:26 -06:00
Gregory Nutt 0f97d81f8d Completes the implemenation of the core shared memory logic: shmget(), shmctl(), shmat(), and shmdt(). This is still some unfinish platform-specific code that needs to be done before we can begin testing. 2014-09-23 16:07:12 -06:00
Gregory Nutt d546ebabe4 Remove executable mode 2014-09-23 16:06:22 -06:00
Gregory Nutt ee0d20d44c Add interfaces to support un-initializing a granule allocator. 2014-09-23 16:05:32 -06:00
Gregory Nutt 835c91b03a Add support for a per-process virtual page allocator. This is a new member of the task_group_s structure. The allocaor must be initialized when a new user process is started and uninitialize when the process group is finally destroyed. It is used by shmat() and shmdt() to pick the virtual address onto which to map the shared physical memory. 2014-09-23 16:04:39 -06:00
Gregory Nutt b542d1733f Cosmetic 2014-09-23 16:03:52 -06:00
Gregory Nutt 566dc8dd38 Add interfaces to support un-initializing a granule allocator. 2014-09-23 15:50:45 -06:00
Gregory Nutt 115634ff4d Flesh out shmctl() logic 2014-09-23 11:41:05 -06:00
Gregory Nutt 5efd5738e9 Flesh out basic logic for shmget() 2014-09-23 10:42:18 -06:00
Gregory Nutt a73a3ef99f Add shared memory initializatin logic 2014-09-23 08:46:31 -06:00
Gregory Nutt 242d5f2068 Add README files and configuration support for the shared memory logic 2014-09-23 07:11:47 -06:00
Gregory Nutt 18ce64d61e Add the build framework and skeleton files for the shared memory feature (no logic yet provided) 2014-09-22 14:53:56 -06:00
Gregory Nutt 6bed778738 Again I forget to add a file needed in a previous commit 2014-09-22 10:59:00 -06:00
Gregory Nutt c5a206666d Add a home for forthcoming shared memory logic 2014-09-22 10:58:27 -06:00
Gregory Nutt 3464089542 Move core heap allocator logic from mm/. to mm/mm_heap/. 2014-09-22 10:53:50 -06:00
Gregory Nutt 8b8c134efa Move the user heap allocator front-end from mm/. to mm/umm_heap/. 2014-09-22 10:48:58 -06:00
Gregory Nutt 77bf9d6ae7 Forgot to add a file in the last commit 2014-09-22 10:43:46 -06:00
Gregory Nutt c3a2527e56 Move the kernel heap allocator front-end from mm/. to mm/kmm_heap/. 2014-09-22 10:42:26 -06:00
Gregory Nutt 907e45752b Move granuale allocator and page allocator from mm/. to mm/mm_gran/. 2014-09-22 10:33:23 -06:00
Gregory Nutt 5d9276ed71 More changes so that the KNSH build survives the changes needed for the ELF build 2014-09-16 15:55:21 -06:00
Gregory Nutt 925fc9ad81 First round of changes to get the ELF configuration building again 2014-09-16 15:37:05 -06:00
Gregory Nutt 8993910427 Fix a typo in use of a configuration setting 2014-09-13 12:24:08 -06:00
Gregory Nutt 205c23b9d6 Add logic to initialize the per-process user heap when each user process is started 2014-09-10 15:55:36 -06:00
Gregory Nutt 0c8830f9f3 Granule allocator initialization uses wrong allocator to setting aside kernel memory 2014-09-03 13:46:05 -06:00
Gregory Nutt b05f29dbfe I love/hate conditional compilation 2014-09-03 11:43:23 -06:00