Re: strace log before the fix, with fsync fix and with fclose fix.

Andrii Nakryiko wrote on Mon, Oct 20, 2025 at 01:12:16PM -0700:

So unclear, which is why it would be nice for FS folks to double check. It's certainly a change in behavior, it used to work reliably before. [0] is the source code of the test (and note that we now added fsync(), without it the test is now broken).

It's a 9p bug, sorry. tentative fix: https://lkml.kernel.org/r/20251022-mmap-regression-v1-1-980365ee524e@codewre...

other thread with repro: https://lkml.kernel.org/r/CAHzjS_u_SYdt5=2gYO_dxzMKXzGMt-TfdE_ueowg-Hq5tRCAi...

I'll send the fix to Linus once someone can confirm this works for this usecase as well (and try to improve our testing a bit... maybe just run the bpf test suite for starters)