> > That would need a new page flag (PG_mmap_dirty?).  Do we have one
quoted text
> > available?
> 
> Yeah, that would be bad. We probably have flags free, but those page flags 
> are always a pain. Scratch that.
> 
> How about just setting a per-vma dirty flag, and then instead of updating 
> the mtime when taking the dirty-page fault, we just set that flag?
> 
> Then, on unmap and msync, we just do
> 
> 	if (vma->dirty-flag) {
> 		vma->dirty_flag = 0;
> 		update_file_times(vma->vm_file);
> 	}
> 
> and be done with it? 

But then background writeout, sync(2), etc, wouldn't update the times.
Dunno.  I don't think actual _physical_ writeout matters much, so it's
not worse to be 30s early with the timestamp, than to be 30s or more
late.

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/