Add more assertions #5245
Open
Add more assertions #5245
+70
−163
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 ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #5245 +/- ##
==========================================
+ Coverage 82.86% 82.90% +0.04%
==========================================
Files 250 250
Lines 26897 26868 -29
==========================================
- Hits 22288 22276 -12
+ Misses 4609 4592 -17
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
ShadowCurse
force-pushed
the
misc_fixes
branch
20 times, most recently
from
70239eb to
ed878a1
Compare
ShadowCurse
force-pushed
the
misc_fixes
branch
4 times, most recently
from
169752b to
b7d4bb8
Compare
ShadowCurse
added
the
Status: Awaiting review
roypat
reviewed
Jun 25, 2025
pb8o
reviewed
Jun 25, 2025
roypat
reviewed
Jun 25, 2025
roypat
previously approved these changes
Jun 25, 2025
roypat
reviewed
Jun 25, 2025
roypat
previously approved these changes
Jun 26, 2025
Manciukic
reviewed
Jul 2, 2025
src/vmm/src/vstate/vcpu.rs
Outdated
Comment on lines
609
to
610
| #[cfg(not(test))] | ||
| assert!(std::ptr::eq(_vcpu_ptr, self)); |
There was a problem hiding this comment.
I think this is one of those cases where having a separate function return a result and then the caller decide what to do with it is better than having a assert that depends on whether we're testing or not, as it makes the code actually testable without hacks like this.
All this change would be is unwrap the result of self.reset_thread_local_data()
There was a problem hiding this comment.
On the other hand, why do we have a test that fails this assertion? Been a while since I looked at this PR, but maybe we should fix that test to actually set things up the same way we do in prod?
There was a problem hiding this comment.
added another commit which drops the VcpuDrop
There was a problem hiding this comment.
personally, a huge fan. Thanks!
| assert!(cell.get().is_none()); | ||
| cell.set(Some(self as *mut Vcpu)); | ||
| }) | ||
| } | ||
|
|
||
| /// Runs `func` for the `Vcpu` associated with the current thread. |
There was a problem hiding this comment.
I think this change is really subjective. I see no issues with the current state and I actually prefer that. What is the benefit of doing like so?
There was a problem hiding this comment.
It was used only once, so inlining it made sense. It also makes code a bit shorter.
ShadowCurse
force-pushed
the
misc_fixes
branch
2 times, most recently
from
e2ba5a2 to
fc8f40b
Compare
It is reasonable to assume that we will not have more than u32::MAX memory slots since kernel only returns i32 from a query syscall. Enforce this with `.expect` calls and change the type of `max_memslots` to u32. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
Replaces never used error result type wit optional. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
Vcpu TLS must only be initialized once. Enforce this with an assertion. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
Move setting of signal handler into vcpu init to prevent race condition between setting TLS and signal handler using TLS. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
- The reset must be called only once on vcpu drop, so move it directly into the Drop impl. - Replace errors with asserts since there is an assumption that TLS will always hold correct vcpu pointer for a given thread (except in tests). Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
Merge run_on_thread_local into signal handler since it is only used there. The error returned from run_on_thread_local was ignored, so instead replace it with logic to only use vcpu ptr if TLS is initialized, without returning any errors. The reason for not asserting on TLS being initialized here is that during Firecracker shutdown, vcpus will be destroyed and TLS will be reset. If signal will be send to Firecracker during that time, the TLS accessed from a signal handler will be empty. But this is expected, so no assertions/panics are needed. Because Rust is a good language, it does not allow to reference TLS_VCPU_PTR definded inside impl block inside the signal_handler function. So move the TLS_VCPU_PTR definition outside the Vcpu impl block. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
ShadowCurse
force-pushed
the
misc_fixes
branch
2 times, most recently
from
84032e5 to
b32041a
Compare
Instead of storing the pointer to the Vcpu in the TLS, store the mmapped `kvm_run` struct instead. This way the Drop implementation for Vcpu is no longer required, since the mmapped `kvm_run` will remain valid until TLS is destroyed by the thread Drop impl. This also removes `unsafe` code from the implementation. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
roypat
approved these changes
Jul 3, 2025
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.
Changes
Add asserts in places where we have strong assumptions about inner workings of the code.
License Acceptance
By submitting this pull request, I confirm that my contribution is made under
the terms of the Apache 2.0 license. For more information on following Developer
Certificate of Origin and signing off your commits, please check
CONTRIBUTING.md.PR Checklist
tools/devtool checkstyleto verify that the PR passes theautomated style checks.
how they are solving the problem in a clear and encompassing way.
in the PR.
CHANGELOG.md.Runbook for Firecracker API changes.
integration tests.
TODO.rust-vmm.