On 23/03/2025 06:38, Christoph Hellwig wrote:
On Sun, Mar 23, 2025 at 01:17:08AM +0530, Ritesh Harjani wrote:
@ Ritesh, thanks for the notice - are you ok to send a fix for this? At
a glance, it seems that those two conflicting flags won't cross paths in
practice (but obvs still need to fix this).
[full quote deleted, can you please properly trim your replies?]
So, I guess we can shift IOMAP_F_SIZE_CHANGED and IOMAP_F_STALE by
1 bit. So it will all look like..
Let's create some more space to avoid this for the next round, e.g.
count the core set flags from 31 down, and limit IOMAP_F_PRIVATE to a
single flag, which is how it is used.
