Add MVP code for shared mempool and its example program #261
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
added 7 commits
January 10, 2023 20:38
This commit will add basic infrastructure code for share mempools Signed-off-by: Matthew Leon <[email protected]>
This commit will add shared mempool teardown code along with cleaning the shared mempool initialization code Signed-off-by: Matthew Leon <[email protected]>
This commit changes the shareable mempool code so that all of the
structures are contained within the shared memory region along with
the mempool ring and this commit adds the testing code used to test
the shared mempool
Signed-off-by: Matthew Leon <[email protected]>
This commit adds a hardcoded address for the shared mempool within the process address space and adds a script that will disable ASLR so we can increase the likelihood that mmap succeeds. Signed-off-by: Matthew Leon <[email protected]>
This commit will clean up the shared mempool example code and conform it to the standards of the project Signed-off-by: Matthew Leon <[email protected]>
This commit will clean up the shared mempool code for the linter Signed-off-by: Matthew Leon <[email protected]>
This commit adds the semaphore unlinking that is needed in the shared mempool teardown code Signed-off-by: Matthew Leon <[email protected]>
|
@leonmatt this PR is getting out of date. I was thinking we can close this one and open it later if we need too. What do you think? |
|
@KeithWiles I agree. Perhaps if I find spare cycles, we can continue working on this in the future. In case someone picks this up in the future, I think we decided that it might be worthwhile to try placing the shared memory pool in the kernel address space, marking the memory pages r/w and accessible in User-Mode, and then writing a module that creates a memory policy so that the memory pool is only mapped into processes that CNDP is configured to have access. |
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.
This PR will add basic shared mempool code to the mempool library and targets issue #221 .
The way that this works is we mmap a region of shared memory in multiple processes at the same virtual address in each process and then we store all of the required mempool data structures in the region.
To test the program, build the repo, copy the run_proc.sh script from examples/shared_mempool into the build directory's shared_mempool artifacts, and then run "sudo ./run_proc.sh". The example should print out the addresses of the data structs of the mempool.
Next steps are to decide if we want to convert the pointers to offsets for AF_XDP and if we want to build out lockless rings within the mempool data.