blog: Add GSoC'24 blog on debugging mimalloc port #448
Conversation
|
@huyang531 I see this is marked as draft. Any reason for that? |
Not really. I just marked this as ready for review. |
There was a problem hiding this comment.
Thanks @huyang531, see the inlines.
Also @razvanvirtan @RaduNichita please do a content check.
|
|
||
| ## Current Progress | ||
|
|
||
| Following porting the multithreaded `cache-scratch` stress test in [the last blog post](https://unikraft.org/blog/2024-07-12-unikraft-gsoc-test-mimalloc-on-unikraft), this past three weeks I have been debugging the it. |
There was a problem hiding this comment.
| Following porting the multithreaded `cache-scratch` stress test in [the last blog post](https://unikraft.org/blog/2024-07-12-unikraft-gsoc-test-mimalloc-on-unikraft), this past three weeks I have been debugging the it. | |
| Following porting the multithreaded `cache-scratch` stress test in [the last blog post](https://unikraft.org/blog/2024-07-12-unikraft-gsoc-test-mimalloc-on-unikraft), this past three weeks I have been debugging the it. |
This is unclear to me, can you rephrase?
| ### Debugging `pthread` | ||
|
|
||
| The [first error we ran into](https://unikraft.org/blog/2024-07-12-unikraft-gsoc-test-mimalloc-on-unikraft#testing) was a memory corruption error with the `pthread` interface. | ||
| After investigating the `pthread_join()` calls, I found that the threads were never even created correctly at all because `pthread_create()` uses the POSIX `mmap` system call, which never existed as we did not turn on virtual memory support (i.e., `ukvmem` alongside `posix-mmap`). |
There was a problem hiding this comment.
I don't think you should go that deep into technical details
There was a problem hiding this comment.
I wrote the blog in sort of chronological order - it was because of thispthread bug that I turned on virtual memory support, hence the rest. I have shortened the paragraph but still kept it here.
| In order to resolve the abovementioned issues, I still have much to study about Unikraft's virtual memory management subsystem, its boot process, and how *mimalloc* works under the hood - and I will focus on just that in the following weeks. | ||
|
|
||
| Hopefully, with those issues fixed, we could finally get the `cache-scratch` test running on multiple cores. | ||
| Once that is done, I will get back on track with my "global" TODO list: |
There was a problem hiding this comment.
Can you mention what ideas do you have for benchmarking?
There was a problem hiding this comment.
I have written much about benchmarking (i.e., which benchmarks I would use, what they are for, etc.) in my last blog post so I didn't elaborate here. I will attach a link to my last blog post
Signed-off-by: Yang Hu <[email protected]>
There was a problem hiding this comment.
Reviewed-by: Stefan Jumarea [email protected]
Approved-by: Stefan Jumarea [email protected]
Signed-off-by: Yang Hu [email protected]