Integration Candidate 2020-05-20 #482
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The statvfs() call depends on the underlying filesystem supporting this. On RTEMS, the IMFS filesystem type does not, and it returns the ENOSYS errno. This is better translated to OS_ERR_NOT_IMPLEMENTED rather than OS_ERROR.
Rather than relying on the BSP to preallocate, the ram disk block devices can be created based on request. This correlates with the way RAM disks are implemented on VxWorks and is cleaner and more flexible by making it more independent of the BSP.
This string should be of OS_FS_DEV_NAME_LEN
This provides a better method of handling test abort, in a BSP-specific manner.
Remove the OS_VolumeTable definition from all BSPs, but provide a default (empty) one to support linking when OMIT_DEPRECATED is not set.
Do not assume BSP volume table provides a "/cf" directory map. Instead, create a UT-specific map of a consistent name, and use that.
Add a distinct type for FS_BASED entries, and an UNKNOWN value for when a type is actually not yet identified. This can be used to identify the FS type first at the shared layer, then in the implementation layer as a fallback option if not identifiable in the shared layer. Use the volume name prefix "RAM" as a hint that the volume is supposed to be a VOLATILE disk as opposed to a normal disk. This is done in shared layer so it applies to all OS types.
Confirm that the "selectable" flag is set before calling the underlying select() API. Also update unit tests to match.
If the "OSAL_EXT_SOURCE_DIR" cache variable is set, this location will be checked first for a BSP/OS implementation layer. This can point to an out-of-tree implementation layer if necessary. However it is discouraged from actually doing this as there is no attempt at API stability at the low level implementation layer.
Implement OS_GetResourceName, OS_ForEachObjectByType
Corrects issue when a task waiting on a binary semaphore is deleted, it left the mutex in a locked state preventing other tasks from using the mutex.
The mutex to protect the timer callback (timecb) resource table was missing.
Add a wrapper at the shared layer to provide a common location to check the status of global lock/unlock ops. All calls to OS_Lock_Global_Impl and OS_Unlock_Global_Impl from the shared modules are replaced with calls to this wrapper.
… between 0 and <MAX and output array indices vary
Fix #455, correct length of device_name in filesys
Fix #474, Add timecb global mutex
Fix #293, Expand API for object queries
Fix #450, add external source directory for OS/BSP
Fix #470, Binary sem task delete issues
Fix #447, Resolve doxygen warnings
Fix #459, dynamically create RAM disk devices on RTEMS
Fix #457, provide BSP shutdown handler
Fix #454, Better error translation on statvfs()
…sing-Tests Fix #374, Add Object Utility API functional tests
|
@yammajamma one of the unit tests failed after adding your commit for #374. Please fix and commit to this branch with the prefix "HOTFIX" in the commit message.
|
…ssing-Functional-API-Tests Hotfix #374 object utilitys missing functional api tests
Fix #476, add global lock/unlock wrapper
Use "object_id" rather than "id" for consistency with other API calls. The difference was flagged as a warning by doxygen.
- Correct merge issue with regards to the OS_Global_Unlock function - Correct use of OS_TaskGetId() - need to use Impl variant here.
Fix #367, 453, 456, 460, Deprecate OS_VolumeTable and other fixes
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Describe the contribution
Fix #293
Fix #374
Fix #447
Fix #450
Fix #454
Fix #455
Fix #457
Fix #459
Fix #470
Fix #474
Fix #476
Testing Performed
Bundle CI - https://travis-ci.com/github/nasa/cFS/builds/168822512
Expected behavior changes
PR #439 - Add functional tests for
OS_IdentifyObject,OS_ConvertToArrayIndexandOS_ForEachObjectfunctions.PR #451 - Fix doxygen warnings
PR #463 - Unit test cases which use
OS_statfsand run on anRTEMS IMFSvolume will be skipped and categorized as "NA" due toOS_ERR_NOT_IMPLEMENTEDresponse, rather than a failure.PR #464 - The device_name field was using the wrong length, it should be of
OS_FS_DEV_NAME_LENAlso correct another length check on the local path name.PR #465 - For RTEMS, will not shutdown the kernel if test abort occurs.
PR #466 - Unit tests work on RTEMS without BSP preallocating ramdisks
PR #468 - If
OSAL_EXT_SOURCE_DIRcache variable is set, this location will be checked first for a BSP/OS implementation layer.PR #469 - Implement
OS_GetResourceName()andOS_ForEachObjectOfType(), which are new functions that allow for additional query capabilities. No impact to current behavior as the FSW does not currently use any of these new APIs.PR #472 -
bin-sem-testwhich replicates the specific conditions for the observed bug to occur. Deletes the task callingOS_BinSemTake()and then attempts to use the semaphore after this.pthread"cleanup handler" to handle the situation where a task is canceled during thepthread_cond_wait()call. This ensures that themutexis unlocked as part of the cleanup, so other tasks may continue using the semaphore.mutexlocking to be a finite "timed" wait rather than an infinite wait. In all cases, the condition variable is only held for brief periods of time and should be readily available. If a task blocks for a long time, this considers the mutex "broken" and aborts, thereby avoiding deadlock. This is a "contingency" fix in that if an exception or signal or other unknown/unhandled async event occurs that leaves the mutex permanently locked.PR #475 - Adds the mutex to protect the timer callback
timecbresource table.System(s) tested on
See PRs
Additional context
Part of nasa/cFS#91
Third party code
If included, identify any third party code and provide text file of license
Contributor Info - All information REQUIRED for consideration of pull request
Joseph Hickey, Vantage Systems, Inc.
Leor Bleier, NASA-GSFC
Yasir Khan, NASA-GSFC