Re: O_DIRECT vs BLK_FEAT_STABLE_WRITES, was Re: [PATCH] btrfs: never trust the bio from direct IO

[Qu Wenruo <quwenruo.btrfs@gmx.com>]

在 2025/10/20 20:30, Christoph Hellwig 写道:>
>> But still, such performance drop can be very obvious, and performance
>> oriented users (who are very happy running various benchmark tools) are
>> going to notice or even complain.
> 
> I've unfortunately seen much bigger performance drops with direct I/O and
> PI on fast SSDs, but we still should be safe by default.

Off-topic a little, mind to share the performance drop with PI enabled 
on XFS?

With this patch I'm able to enable direct IO for inodes with checksums.
I thought it would easily improve the performance, but the truth is, 
it's not that different from buffered IO fall back.

It's still the old 200MiB/s (vs ~2GiB/s nodatasum), no matter falling 
back to buffered IO or not (and extra traces indeed shows it's really 
going direct IO path).

So I start wondering if it's the checksum itself causing the miserable 
performance numbers.

Thanks,
Qu