[Darwin] Unstored attributes should be flushed to storage on shutdown #36791
+206
−34
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
|
Review changes with |
jtung-apple
force-pushed
the
jtung/darwin-mtrdevice-should-flush-attributes-to-storage-on-invalidate
branch
from
82e5795 to
07cfafb
Compare
|
PR #36791: Size comparison from 918320e to 07cfafb Full report (69 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, cyw30739, efr32, esp32, linux, nrfconnect, nxp, psoc6, qpg, stm32, telink, tizen)
|
kiel-apple
reviewed
Dec 11, 2024
src/darwin/Framework/CHIPTests/TestHelpers/MTRTestPerControllerStorage.h
Show resolved
Hide resolved
kiel-apple
approved these changes
Dec 11, 2024
|
PR #36791: Size comparison from 918320e to d7a85ca Full report (69 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, cyw30739, efr32, esp32, linux, nrfconnect, nxp, psoc6, qpg, stm32, telink, tizen)
|
woody-apple
approved these changes
Dec 11, 2024
|
PR #36791: Size comparison from 918320e to cf88dab Full report (69 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, cyw30739, efr32, esp32, linux, nrfconnect, nxp, psoc6, qpg, stm32, telink, tizen)
|
Co-authored-by: Boris Zbarsky <[email protected]>
|
PR #36791: Size comparison from 918320e to 2306949 Full report (69 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, cyw30739, efr32, esp32, linux, nrfconnect, nxp, psoc6, qpg, stm32, telink, tizen)
|
|
PR #36791: Size comparison from 918320e to e1e64c1 Full report (25 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, nrfconnect, qpg, stm32, telink, tizen)
|
| @@ -449,6 +449,13 @@ - (void)cleanupAfterStartup | |||
| for (MTRDevice * device in devices) { | |||
| [device invalidate]; | |||
| } | |||
|
|
|||
| // Since MTRDevice invalidate may issue asynchronous writes to storage, perform a | |||
| // block synchronously on the storage delegate queue to flush those operations. | |||
There was a problem hiding this comment.
This comment still doesn't explain why we need to block here on that. To repeat my review comment, which was marked resolved without resolving:
What happens if we don't do that? Won't the writes happen anyway? Or are we worried about our caller tearing down the storage backend if we return from here, before the writes have a chance to happen? Would be good to document why this needs to be a sync point.
There was a problem hiding this comment.
Ah I didn't see it before the previous comment was resolved. And yes you were right in that I was worried about storage backend going away after shutdown returns. Let me clarify a bit more.
|
PR #36791: Size comparison from 918320e to 1fbd14b Full report (54 builds for bl602, bl702, bl702l, cc13x4_26x4, cc32xx, cyw30739, efr32, esp32, linux, nrfconnect, nxp, psoc6, qpg, stm32, telink, tizen)
|
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.
When MTRDeviceController shuts down, if an MTRDevice object still has attributes that have yet to be persisted, the persistent attribute cache may become out of date. And on next startup, subscriptions can cause more priming reports than is necessary.
This PR makes MTRDevice persist unstored attributes on invalidate, and have MTRDeviceController wait for the storage delegate has executed the write blocks before continuing with the shutdown.