Spinlock validation augmentation #13800
Closed
Conversation
This file contains hidden or 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
Codecov Report
@@ Coverage Diff @@
## master #13800 +/- ##
==========================================
+ Coverage 51.97% 51.98% +<.01%
==========================================
Files 308 308
Lines 45508 45517 +9
Branches 10546 10547 +1
==========================================
+ Hits 23653 23661 +8
Misses 17055 17055
- Partials 4800 4801 +1
Continue to review full report at Codecov.
|
andrewboie
reviewed
Feb 27, 2019
| @@ -715,6 +715,7 @@ int z_spin_lock_valid(struct k_spinlock *l) | |||
| } | |||
| } | |||
| l->thread_cpu = _current_cpu->id | (u32_t)_current; | |||
| _current_cpu->spin_depth++; | |||
There was a problem hiding this comment.
needs an #ifdef since spin_depth isn't always there
| @@ -107,6 +108,10 @@ struct _cpu { | |||
| /* True when _current is allowed to context switch */ | |||
| u8_t swap_ok; | |||
| #endif | |||
|
|
|||
| #ifdef SPIN_VALIDATE | |||
There was a problem hiding this comment.
Can we make SPIN_VALIDATE a Kconfig?
| @@ -724,6 +725,7 @@ int z_spin_unlock_valid(struct k_spinlock *l) | |||
| return 0; | |||
| } | |||
| l->thread_cpu = 0; | |||
| _current_cpu->spin_depth--; | |||
There was a problem hiding this comment.
needs an #ifdef since spin_depth isn't always there
added 2 commits
March 11, 2019 21:42
Catching the error in a function is nice, but one really wants to know where it happened, and where the recursive lock was taken (or where the unowned release was actually grabbed). Add a layer of macro indirection to catch this info and log it with the assertion. Signed-off-by: Andy Ross <[email protected]>
Right now the validation layer catches mismatched locking, but another common gotcha is failing to release an outer nested lock before entering a blocking primitive. In these cases, the OS would swap away, and the next thread (if any) to try to take the outer lock would hit a "not my spinlock" error, which while correct isn't very informative. By keeping a single per-cpu nesting count we can check this at the point of the swap, which is more helpful. Signed-off-by: Andy Ross <[email protected]>
andyross
force-pushed
the
spinvalidation2
branch
from
34d266b to
61511e2
Compare
This was referenced Mar 13, 2019
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.
No need to merge this for 1.14. It's a check I added at one point when we noticed a spot where the kernel was trying to swap away with a spinlock held (recursive irqlocks worked like that, but with spinlocks you must be 100% sure there is only one lock being released in _Swap() -- and that's a good thing!).
It's cheap and easy, but as it happens everything else passes with it, so there's no rush to get it into the tree.