On 11/26/25 11:19, Miklos Szeredi wrote:
On Tue, 25 Nov 2025 at 02:10, Joanne Koong joannelkoong@gmail.com wrote:
Prior to the changes added in commit 0c58a97f919c ("fuse: remove tmp folio for writebacks and internal rb tree"), fuse didn't ensure that data was written back for sync. The folio was marked as not under writeback anymore, even if it was still under writeback.
This is the main point. Fuse has existed for 20 years without data integrity guarantees. Reverting to that behavior is *not* a regression, it's simply a decades old bug. And solving that bug is darn hard, which is why it's not an option at this point.
Yes, and it should be clearly spelled out in the patch descriptions that this is a fuse-only thing, and why it is a fuse-only thing. It's not a "we don't have data integrity because writeback may hang", it's a "we don't have data integrity because fuse never supported it, so waiting here is simply not required".