Worker fixes #175
Open
Worker fixes #175
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
Nexthop events should go next to nexthop definitions in gr_nexthop.h. Move interface events before API message definitions. Signed-off-by: Robin Jarry <[email protected]>
If that thread cannot be started, grout will not function. Signed-off-by: Robin Jarry <[email protected]>
Moving one RXQ from one worker to another is only safe if the originating/source worker has stopped polling the RXQ before allowing the target/destination worker to starting polling it. Currently, after updating the queue maps of the source and target workers, we reload *all* workers without enforcing any order or operations. The target worker may very well be reloaded before the source one leading to concurrent polling of the moved RXQ by two worker threads. Fix this by always reloading the source worker first to ensure that it has given up on the RXQ, before reloading the target worker. It also has the side effect of only requiring to reload these two workers and not all of them. Signed-off-by: Robin Jarry <[email protected]>
(1 << n) evaluates to a 32 bit signed integer by default. Use GR_BIT64 which properly casts to uint64_t. Signed-off-by: Robin Jarry <[email protected]>
Instead of trying to be a smart ass about memory ordering types, always use the safest memory_order_seq_cst (Sequentially consistent ordering) which is the one used by simple atomic_load() and atomic_store() macros. If we ever encounter some performance issues, we can reconsider each individual atomic operation to see if a more precise/relaxed memory ordering would help. For the record, on x86, using atomic_store() and atomic_load() on _Atomic() variables is redundant and will not necessarily induce explicit memory barrier/fence instructions. Link: https://en.cppreference.com/w/c/atomic/memory_order Link: https://en.cppreference.com/w/c/language/atomic Signed-off-by: Robin Jarry <[email protected]>
The port_queue_assign function code is rather obscure. Add comments to explain what is done and why. Signed-off-by: Robin Jarry <[email protected]>
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 description provided.