> > 
quoted text
> > But then background writeout, sync(2), etc, wouldn't update the times.
> 
> Sure it would, but only when doing the final unmap.
> 
> Did you miss the "on unmap and msync" part?

No :)

What I'm saying is that the times could be left un-updated for a long
time if program doesn't do munmap() or msync(MS_SYNC) for a long time.

If program has this pattern:

mmap()
write to map
msync(MS_ASYNC)
sleep(long)
write to map
msync(MS_ASYNC)
sleep(long)
...

Then we'd never see time updates (until the program exits, but that
could be years).

Maybe this doesn't matter, I'm just saying this is a disadvantage
compared to the "update on first dirtying" approach, which would
ensure, that times are updated at least once per 30s.

Miklos
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/