From: Jamie Lokier <jamie@xxxxxxxxxxxxx>
Date: Fri, 23 Dec 2005 05:57:06 +0000
> I'm not sure if it's cache-safe, on architectures which have
> non-coherent, virtually-indexed caches (like MIPS etc.). There is no
> explicit step after modifying the data in user-space which which
> updates the cached page in kernel-space/dma-space. Do you know about
> that? That's unlike sendmsg(), which does it at the time, and
> sendfile(), where it's done when the file is written before
> sendfile().
pci_{map,unmap}_*() does a cache flush on such systems.
> That wouldn't work because the app also needs to know when it can
> write more data to the socket, independently of when it can recycle
> buffers.
If the buffers are free'd up by ACKs, socket space exists as well.
You can add that condition to the test to make it explicit.
> So it would need to be a separate POLLX flag. POLLWRBAND could
> probably be abused for it, if there was a desire to wedge it
> uncomfortably into the standard API.
But you just said you don't want "just POLLX event" but
"POLLX plus socket write space available", which we can
make POLL_OUT indicate when the special socket option is
set.
It's a shame that all of these conditions are very similar,
yet different enough to require explicit settings. :-/
> No other OS which has sendfile() (there are a lot) has a way to do
> check when buffers are freed, unfortunately - they don't even have
> SIOCOUTQ (not even FreeBSD).
Not really our problem :-)
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
