Rich Ercolani

...you do see that the backtrace of the overflowing stack has no BLAKE3 code anywhere in it? And it reproduces without the BLAKE3 PR?

@GregorKopka One big problem with functionality like this would be that when a checksum fails on a data block, you know that either the checksum or the data block is...

@richardelling that won't overwrite the old mangled block in historical copies,though, it'll just allocate a new one. Or are you suggesting reading until you get EIEIO and then digging through...

@aletus Yes, because if the block was compressed or the location of the read errors was some metadata, or many other things, that wouldn't fly. I don't think you understand...

This would be nice to see - I just added the desire to get something like this to the Hackathon project ideas not having seen this PR (or if I...

I'll give these a spin on my various sundry machines later today or tomorrow.

Of course, it's important to include the silliest part, the example benchmarks. Ryzen 5900X: ``` $ uname -r 5.18.0-0.bpo.1-amd64 $ cat /proc/spl/kstat/zfs/chksum_bench 15 0 0x01 -1 0 242790676536984 242792205359807 implementation...

I haven't dug into it, but based on the ASAN failure assertion, I was wondering if the native encryption stuff was calling into the SHA2 API somewhere in the ASM...

5.4.17 and "whatever Oracle patched in" are not at all the same version, so please mention the full version number when discussing this. I think the last person who tried...

Re: hole_birth, my understanding is that all known issues with hole_birth were fixed at the same time as the tunable was introduced, and some vendors might have on their own...