Gaurav-Aggarwal-AWS [Thu, 6 Jan 2022 05:14:01 +0000 (21:14 -0800)]
Add a cap to the queue locks (#435)
Add a cap to the queue locks
cRxLock and cTxLock members of the queue data structure count the
number items received and sent to the queue while the queue was locked.
These are later used to unblock tasks waiting on the queue when the
queue is unlocked. The data type of these members is int8_t and this can
trigger overflow check assert if an ISR continuously reads/writes to the
queue in a loop as reported in this issue: https://github.com/FreeRTOS/FreeRTOS-Kernel/issues/419.
Note due to the length of the operation is it not recommended to write to
the queue that many times from an ISR - stream buffers are a better option,
or alternatively, defer the operation to a task by just having the ISR send a
direct to task notification to unblock the task.
This PR caps the values of the cRxLock and cTxLock to the number of tasks in
the system because we cannot unblocks more tasks than there are in the system.
Note that the same assert could still be triggered is the application creates more
than 127 tasks.
Amit Tomar [Wed, 29 Dec 2021 20:53:10 +0000 (02:23 +0530)]
fix alignment exception for ullPortInterruptNesting. (#317)
* fix alignment exception for ullPortInterruptNesting.
While loading (LDR X5, ullPortInterruptNestingConst) the ullPortInterruptNesting
variable, the program control seems to be stuck and there is no abort or stack
trace observed (as there is no exception handler is installed to catch unaligned
access exception).
Program control moves forward, if one just declares this varible to be 2 bytes
aligned but then varible is not updated properly.
One of my colleague, pointed out that issue is due to the fact that
ullPortInterruptNesting must be at 8 bytes aligned address but since
"vApplicationIRQHandler" (that has 4 bytes of address) is sitting between
two 8 bytes aligned addresses that forces ullPortInterruptNesting to be at
4 byte aligned address, causing all sort of mess.
It works on QEMU (on ARM64) as it is, since there is no such check for
unaligned access but on real hardware it is prohibited.
Workaround to this problem is, either we skip 4 byets (using .align 4) after
vApplicationIRQHandler declaration or declare it the end of all declarations.
This commit does the latter one.
Signed-off-by: Amit Singh Tomar <atomar25opensource@gmail.com>
* Update portASM.S
Remove 1 tab = 4 spaces
Co-authored-by: Amit Singh Tomar <atomar25opensource@gmail.com> Co-authored-by: alfred gedeon <28123637+alfred2g@users.noreply.github.com> Co-authored-by: Joseph Julicher <jjulicher@mac.com>
Gaurav Aggarwal [Thu, 7 Oct 2021 02:00:39 +0000 (19:00 -0700)]
Add option to disable unprivileged critical sections
This commit introduces a new config
configALLOW_UNPRIVILEGED_CRITICAL_SECTIONS which enables developers to
prevent critical sections from unprivileged tasks. It defaults to 1 for
backward compatibility. Application should set it to 0 to disable
critical sections from unprivileged tasks.
Tobias Nießen [Thu, 11 Nov 2021 01:39:01 +0000 (02:39 +0100)]
Mention portMAX_DELAY in xEventGroupWaitBits docs (#411)
The public function xEventGroupWaitBits passes xTicksToWait to the
function vTaskPlaceOnUnorderedEventList, which passes the number of
ticks to prvAddCurrentTaskToDelayedList and sets xCanBlockIndefinitely
to pdTRUE, causing the latter to block indefinitely if
xTicksToWait == portMAX_DELAY and INCLUDE_vTaskSuspend == 1.
Kevin Thibedeau [Mon, 8 Nov 2021 21:51:25 +0000 (16:51 -0500)]
Build with -Wmissing-prototypes flags vPortYieldFromISR() in the Posix port. (#409)
There's already a portYIELD_FROM_ISR() macro that calls vPortYield() which wraps the FromISR code.
It doesn't appear that vPortYieldFromISR() is intended to be publicly accessible in this port so
I've marked it as private to silence the warning.
event_create() also got flagged due to missing void in prototype.
Archit Gupta [Sat, 16 Oct 2021 23:36:44 +0000 (16:36 -0700)]
Fix prvWriteMessageToBuffer on big endian (#391)
prvWriteMessageToBuffer wrote the first sbBYTES_TO_STORE_MESSAGE_LENGTH
bytes of the size_t-typed length to the buffer as the data length. While
this functions on little endian, it copies the wrong bytes on big
endian. This fix converts the length to configMESSAGE_BUFFER_LENGTH_TYPE
first, and then copies the exact amount, thus fixing the issue.
Additionally it adds an assert to verify the size is not greater than
the max value of configMESSAGE_BUFFER_LENGTH_TYPE; previously this would
truncate silently.
Gaurav-Aggarwal-AWS [Sat, 16 Oct 2021 20:45:03 +0000 (13:45 -0700)]
Update comments for the ARM_CA53_64_BIT_SRE port (#403)
Mention that FreeRTOS_IRQ_Handler should not be used for FIQs and the
reason for assuming Group 1 for Interrupt Acknowledge and End Of
Interrupt registers.
Gaurav Aggarwal [Wed, 4 Aug 2021 21:57:45 +0000 (14:57 -0700)]
Associate secure context with task handle
The secure side context management code now checks that the secure
context being saved or restored belongs to the task being switched-out
or switched-in respectively.
Gaurav Aggarwal [Wed, 4 Aug 2021 21:52:22 +0000 (14:52 -0700)]
Pre-allocate secure-side context structures
This commit improves ARMv8-M security by pre-allocating secure-side task
context structures and changing how tasks reference a secure-side
context structure when calling a secure function. The new configuration
constant secureconfigMAX_SECURE_CONTEXTS sets the number of secure
context structures to pre-allocate. secureconfigMAX_SECURE_CONTEXTS
defaults to 8 if left undefined.
Paul Adelsbach [Fri, 13 Aug 2021 00:06:41 +0000 (17:06 -0700)]
fix example usage of xMessageBufferCreateStatic and xStreamBufferCrea… (#380)
Example usage is actually correct, so remove the -1. but update
the incorrect parameter description for pucStreamBufferStorageArea
and pucMessageBufferStorageArea.
RichardBarry [Wed, 4 Aug 2021 17:05:23 +0000 (10:05 -0700)]
Indent contents of a taskENTER_CRITICAL/taskEXIT_CRITICAL block. (#348)
* Indent contents of a taskENTER_CRITICAL/taskEXIT_CRITICAL block.
Move a few configASSERT() statements out of a path where they would always be triggered to prevent "condition is always true" compiler warnings.
* Replace configASSERT() positions due to unintended semantic change from the version where asserts were at the top of the file.
* Remove all empty definitions of portCLEAN_UP_TCB( pxTCB ) and
portALLOCATE_SECURE_CONTEXT( ulSecureStackSize ) from ports.
When these are undefined, the default empty definition is defined
in FreeRTOS.h.
Gaurav-Aggarwal-AWS [Wed, 23 Jun 2021 01:51:58 +0000 (18:51 -0700)]
Add Cortex-A53 port with system register interface for CPU interface access (#357)
The difference between this port and portable/GCC/ARM_CA53_64_BIT is
that this port uses System Register interface to access CPU interface
while the other one uses Memory-mapped interface.
Signed-off-by: Gaurav Aggarwal Co-authored-by: Cobus van Eeden <35851496+cobusve@users.noreply.github.com>
RichardBarry [Mon, 14 Jun 2021 19:17:41 +0000 (12:17 -0700)]
Make the type used to hold run-time counter values configurable (#350)
* Introduce configRUN_TIME_COUNTER_TYPE which enables developers to define the type used to hold run time statistic counters. Defaults to uint32_t for backward compatibility. #define configRUN_TIME_COUNTER_TYPE to a type (for example, uint64_t) in FreeRTOSConfig.h to override the default.
Introduce ulTaskGetIdleRunTimePercent() to complement the pre-existing ulTaskGetIdleRunTimeCounter(). Whereas the pre-existing function returns the raw run time counter value, the new function returns the percentage of the entire run time consumed by the idle task. Note the amount of idle time is only a good measure of the slack time in a system if there are no other tasks executing at the idle priority, tickless
idle is not used, and configIDLE_SHOULD_YIELD is set to 0.
* Add ultaskgetidleruntimepercent to lexicon.txt.
* Update History file.
Add the MPU version of ulTaskGetIdleRunTimePercent().
* Update include/FreeRTOS.h to correct comment as per aggarg@ suggestion.
* Fix alignment in mpu_wrappers.h.
Commit changes to mpu_prototypes.h which were missed from the original commit.
Paul Bartell [Fri, 28 May 2021 14:23:29 +0000 (07:23 -0700)]
Update github Auto-Release workflow to add an option to set the main branch version in task.h (#337)
Adding text into this new field updates the following macros in task.h in the main branch:
tskKERNEL_VERSION_NUMBER (alphanumeric string)
tskKERNEL_VERSION_MAJOR (numeric only)
tskKERNEL_VERSION_MINOR (numeric only)
tskKERNEL_VERSION_BUILD (numeric only)
Create macro versions of uxListRemove() and vListInsertEnd() for use in xTaskIncrementTick(). This provides a minor optimisation to remove the need for a few function calls. (#241)
Co-authored-by: alfred gedeon <28123637+alfred2g@users.noreply.github.com> Co-authored-by: Aniruddha Kanhere <60444055+AniruddhaKanhere@users.noreply.github.com>
Jeff Tenney [Tue, 13 Apr 2021 18:05:46 +0000 (11:05 -0700)]
Remove support for tmrCOMMAND_START_DONT_TRACE (#305)
* Remove support for tmrCOMMAND_START_DONT_TRACE
Prior to this commit, the timer task used tmrCOMMAND_START_DONT_TRACE
to reload a "backlogged" auto-reload timer -- one for which a reload
operation would have started a period that had already elapsed. If the
command queue contained a stop or delete command when
tmrCOMMAND_START_DONT_TRACE was put into the queue for a reload, the
timer unexpectedly restarted when the timer task processed
tmrCOMMAND_START_DONT_TRACE. This commit implements a new method of
reloading auto-reload timers and eliminates support for
tmrCOMMAND_START_DONT_TRACE. No other code sends this private command.
However, the symbol tmrCOMMAND_START_DONT_TRACE remains defined, with
its original command value, so as not to impact trace applications.
Also fix one-shot timers that were not reliably being marked as not
active:
- when they expired before the start command could be processed
- when the expiration was processed during the timer list switch
Also improve consistency:
- Always reload auto-reload timers *before* calling the callback.
- Always call traceTIMER_EXPIRED() just prior to the callback.
* fix indent
* Revert unnecessary change to prvTimerTask()
Change was intended to faithfully work through a backlog that spanned a
list switch, and before processing a stop or delete command. But, (1)
it isn't important to do that, and (2) the code didn't accomplish the
intention when *two* auto-reload timers were backlogged across a list
switch. Best to simply leave this part of the code as it was before.
* fix style
Co-authored-by: Joseph Julicher <jjulicher@mac.com>
Paul Bartell [Sat, 10 Apr 2021 00:06:58 +0000 (17:06 -0700)]
Overwrite existing QueueRegistry entries when a handle is added multiple times. (#306)
Overwrite an existing entry for a given xQueue handle when vQueueAddToRegistry is called with an xQueue handle of a queue that is already in the QueueRegistry.
Define default values of macros before first use (#298)
configSTACK_ALLOCATION_FROM_SEPARATE_HEAP was added recently in
https://github.com/FreeRTOS/FreeRTOS-Kernel/pull/267. This macro was
used in portable.h before its default value was defined, resulting in a
warning when built with -Wundef. This changes moves the default value
definition for configSTACK_ALLOCATION_FROM_SEPARATE_HEAP to portable.h
to ensure that it is defined before first use.
portUSING_MPU_WRAPPERS check in mpu_wrappers.h was updated in
https://github.com/FreeRTOS/FreeRTOS-Kernel/pull/285. The new check
results in a warning when built with -Wundef because
portUSING_MPU_WRAPPERS is not defined yet. This changes adds the default
value definition for portUSING_MPU_WRAPPERS to portable.h to ensure that
it is defined before first use.
begriffs.com / projects / freertos / log