Re: [PATCH] update ctime and mtime for mmaped write

[Miklos Szeredi]

> > Take this example:
> >
> >     fd = open()
> >     addr = mmap(.., fd)
> >     write(fd, ...)
> >     close(fd)
> >     sleep(100)
> >     msync(addr,...)
> >     munmap(addr)
> >
> > The file times will be updated in write(), but with your patch, the
> > bit in the mapping will also be set.
> >
> > Then in msync() the file times will be updated again, which is wrong,
> > since the memory was _not_ modified through the mapping.
> 
> This is correct.  I have updated my proposed patch to include the clearing
> of AS_MCTIME in the routine which updates the mtime field.

That doesn't really help.  Look at __generic_file_aio_write_nolock().
file_update_time() is called before the data is written, so after the
last write, there will be nothing to clear the flag.

And even if fixed this case by moving the file_update_time() call to
the end of the function, there's no guarantee, that some filesystem
won't do something exotic and call set_page_dirty() indenpendently of
write().  Good luck auditing all the set_page_dirty() calls ;)

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