>
>
>On Wed, 11 Mar 1998, David S. Miller wrote:
>>
>> I'll make a pre-90-2 with this thing included, and we can remove it
>> again if it turns out that David is correct (which is likely, but
>> I'd like to hear more about this and I already applied the patch..)
>>
>> Ok, I hope my other patches I sent you last night are in there too?
>> Let's try to keep our trees in sync as best as possible.
>
>Yes, all your patches should be there too. The pre-90-2 is out there
>already, so feel free to verify that (and pre-90-2 also contains the first
>cut of my spinlock changes wrt the task lists - not all of the things I
>was talking about, but maybe enough that Ingo will be interested).
pre-2.1.90-2 works fine also with my patch rejected but then seems to be
_very_ less responsive using eth0 (at least in a recursive telnet session).
My loopback test, that stalls in pre-2.1.90-1, instead run faster
with my patch rejected (it runs 2sec/100Mbyte faster).
If my simple patch is really useful (as it seems to me) I' d like to see
this two line of log :-)
--- /usr/src/linux/net/ipv4/tcp.c Thu Mar 12 19:16:37 1998
+++ linux/net/ipv4/tcp.c Thu Mar 12 21:02:20 1998
@@ -201,6 +201,8 @@
* tcp_do_sendmsg to avoid burstiness.
* Eric Schenk : Fix fast close down bug with
* shutdown() followed by close().
+ * Andrea Arcangeli: Improved tcp performance sending an
+ * extra ack if needed in cleanup_rbuf().
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
If you think that the drop of performance in recursive telnet sessions is due
to some bug somewhere I can continue to search for a right fix (or maybe
David has just fixed this for us).
Andrea[s] Arcangeli
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu