-
Notifications
You must be signed in to change notification settings - Fork 11
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Minor issue in hash interface. #29
Comments
No it is not. Thanks for pointing it out. |
Hi Team, Firstly, Thanks for responding back, Really appreciate it. Why do we need DEDUP_HASH_DESC_COUNT (128)??? Wondering what could be the use case?? Many Thanks, |
Hey, right now dm-dedup worker runs in a single-threaded mode. But if it would be running in a multi-threaded mode, if we have only one slot in the table, then it would be a bottleneck. So, instead we have 128. HTH. |
Since we are now forced to use synchronous or asynchronous hash post kernel version 4.5, I believe this implementation is obsolete. We have to use asynchronous hash which exists primarily for multi-threaded mode. |
Yes, ahash and shash. I changed the interface to shash. But facing some issues. |
@venktesh-bolla yes, we are also facing some issues. Are you facing the same issue? |
Yes, exactly. it is corrupting in shash->final( ). I patched it to work with shash interface. |
Yes I too cannot think of a scenario where and why another hash_desc's tfm is getting corrupted. |
Can you please let me know how you patched to work with shash interface? |
Successfully Added Graceful fix for usage of shash interface. Tested functionality. Working fine. Will send patch/pull request. |
I have sent a pull request for this. Please have a look. and let me know your comments. |
Great work. This indeed works for ext4 but it's a workaround and we cannot use this for long term. |
Thankyou, |
Apologies, I meant that our current implementation using synchronous hash is a workaround. Your fix is correct. |
Ohh Okay, But still we are using single_thread( ) in code. Any development is in progress in that multi_thread() design now? |
Soon, once the port to latest kernel is complete and all bugs/issues are addressed. |
Found an issue.,
https://github.com/dmdedup/dmdedup/blob/master/dm-dedup-hash.c#L111
In above url link 111 line we are acquiring a slot(get_next_slot(desctable)), for calculating digest size and we are not putting it down after using it(put_slot(desctable)). Is it known issue?
Thanks,
Venkatesh.
The text was updated successfully, but these errors were encountered: