kern: stack watermark support #1956
Open
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
Each task has grown a new field, `low_stack_pointer`, that tracks the lowest stack pointer value seen so far (because stacks grow down). We update this value on any kernel entry (interrupt, timer, fault, syscall), so while we will miss very brief excursions of stack _between_ syscalls or interrupts, we should still get a useful value. In addition, the field `past_low_stack_pointer` tracks the lowest stack pointer value across _all previous incarnations_ of the task. This way if the task is periodically restarting, we can still get useful stack stats, including at stack overflow.
aapoalas
reviewed
Dec 16, 2024
sys/kern/src/arch/arm_m.rs
Outdated
| let current = CURRENT_TASK_PTR.load(Ordering::Relaxed); | ||
| uassert!(!current.is_null()); // irq before kernel started? | ||
|
|
||
| let current_prio = { |
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
hawkw
approved these changes
Dec 16, 2024
|
|
||
| // We don't need the current task pointer in this routine, but we'll access | ||
| // it briefly to update the stack watermark: | ||
| { |
There was a problem hiding this comment.
i wonder if this whole block should be conditional on the feature flag? I see that update_stack_watermark is a nop if the feature isn't enabled, and the compiler is probably smart enough to eliminate the CURRENT_TASK_PTR access when the function that's called on it doesn't do anything, but... I just kinda don't trust it...
There was a problem hiding this comment.
I was trying to make as little code conditional as possible, so that it always gets compiled; this is why it's only the body of update_watermark that's cfg'd. The compiler does appear to optimize this load out, fwiw.
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.
Each task has grown a new field,
low_stack_pointer, that tracks the lowest stack pointer value seen so far (because stacks grow down). We update this value on any kernel entry (interrupt, timer, fault, syscall), so while we will miss very brief excursions of stack between syscalls or interrupts, we should still get a useful value.In addition, the field
past_low_stack_pointertracks the lowest stack pointer value across all previous incarnations of the task. This way if the task is periodically restarting, we can still get useful stack stats, including at stack overflow.All of this stuff is conditional on the
stack-watermarkfeature, since it increases the kernel RAM requirement, which is explicit in Oxide's Hubris build system.