[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 6 Jun 2022 07:58:48 +0000
From: David Laight <David.Laight@...LAB.COM>
To: 'Willem de Bruijn' <willemdebruijn.kernel@...il.com>
CC: Frederik Deweerdt <frederik.deweerdt@...il.com>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: RE: [PATCH] [doc] msg_zerocopy.rst: clarify the TCP shutdown scenario
From: Willem de Bruijn
> Sent: 03 June 2022 14:30
>
> On Fri, Jun 3, 2022 at 9:18 AM David Laight <David.Laight@...lab.com> wrote:
> >
> > From: Frederik Deweerdt <frederik.deweerdt@...il.com>
> > > Sent: 02 June 2022 00:01
> > > >
> > > > A socket must not be closed until all completion notifications have
> > > > been received.
> > > >
> > > > Calling shutdown is an optional step. It may be sufficient to simply
> > > > delay close.
> >
> > What happens if the process gets killed - eg by SIGSEGV?
>
> The skb frags hold independent references on the pages. Once skbs are
> freed on transmit completion or socket purge the pages are released if
> there are no other page references.
>
> Otherwise there is nothing zerocopy specific about closing TCP
> connections when a process crashes.
So the proposed text for the documentation is just wrong.
Something that says that retransmissions can re-read the
pages after the socket is closed would be more relevant.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
Powered by blists - more mailing lists