Nick Terrell
Nick Terrell
Thanks for the bug report! We'll aim to fix this one, but we're okay with not reporting all types of corruption. Our policy is that if you want error detection,...
We are planning the 1.5.4 zstd release later this year. I will plan on updating the kernel with that release. However, we don't expect too much more development before the...
> Did 1.5.3 happen, or do you have the convention that uneven means unstable? We have not released 1.5.3 publicly yet. We were experimenting with a different release process where...
> @terrelln Thank you for your confirmation! Is there any way I can assist with pre-release testing when the time comes? > Please let me know, I would be happy...
FYI I've started working on this. I expect to have a patch up next week. And I am going to be aiming for the v6.2 merge window.
Quick update: I'm preparing a patch to update the kernel to v1.5.2 now, aiming for the v6.2 merge window. If the next zstd release is out before then, I will...
> In the commit there are more indirections done due to separating functions and adding pointer calls, that's always going to hurt in kernel due to all the spectre mitigations....
I've been able to reduce the regression with PR #3295. There is still a bit of work to do, but we should be able to at least get pretty close...
The Huffman ASM has been tuned for Intel x86, where it is significantly faster than the C code. I don't remember if I benchmarked on a Zen 2 machine. I...
@felixhandte do you intend to do this before the release?