Re: the incredible shrinking socket

2002-07-11 Thread Alfred Perlstein
* Andrew Gallatin <[EMAIL PROTECTED]> [020711 11:32] wrote: > > Alfred Perlstein writes: > > Some time ago I noticed that there appeared to be several members > > of struct socket that were either only used by listen sockets or > > only used by data sockets. > > > > I've taken a stab at uni

Re: the incredible shrinking socket

2002-07-11 Thread Andrew Gallatin
Mike Silbersack writes: > > Speaking of competition, someone should go look at this: > > http://mail-index.netbsd.org/current-users/2002/07/03/0011.html > Its very worthwhile. Tru64 has had this for years. I think there may be a Jeff Mogul paper on it somewhere (but I don't have time t

Re: the incredible shrinking socket

2002-07-11 Thread Andrew Gallatin
Alfred Perlstein writes: > Some time ago I noticed that there appeared to be several members > of struct socket that were either only used by listen sockets or > only used by data sockets. > > I've taken a stab at unionizing the members and we wind up saving > 28 bytes per socket on i386,

Re: the incredible shrinking socket

2002-07-08 Thread Bosko Milekic
On Mon, Jul 08, 2002 at 12:37:57PM -0500, Jonathan Lemon wrote: > On Mon, Jul 08, 2002 at 10:29:37AM -0700, Luigi Rizzo wrote: > > On Mon, Jul 08, 2002 at 01:22:14PM -0400, Bosko Milekic wrote: > > ... > > > > > Speaking of competition, someone should go look at this: > > > > > > > > > > http://

Re: the incredible shrinking socket

2002-07-08 Thread Jonathan Lemon
On Mon, Jul 08, 2002 at 10:29:37AM -0700, Luigi Rizzo wrote: > On Mon, Jul 08, 2002 at 01:22:14PM -0400, Bosko Milekic wrote: > ... > > > > Speaking of competition, someone should go look at this: > > > > > > > > http://mail-index.netbsd.org/current-users/2002/07/03/0011.html > > > > > > UDP soc

Re: the incredible shrinking socket

2002-07-08 Thread Luigi Rizzo
On Mon, Jul 08, 2002 at 01:22:14PM -0400, Bosko Milekic wrote: ... > > > Speaking of competition, someone should go look at this: > > > > > > http://mail-index.netbsd.org/current-users/2002/07/03/0011.html > > > > UDP sockets have the same problem... i posted patches for that > > case around dec

Re: the incredible shrinking socket

2002-07-08 Thread Bosko Milekic
On Mon, Jul 08, 2002 at 10:16:36AM -0700, Luigi Rizzo wrote: > On Mon, Jul 08, 2002 at 11:56:22AM -0500, Mike Silbersack wrote: > ... > > Speaking of competition, someone should go look at this: > > > > http://mail-index.netbsd.org/current-users/2002/07/03/0011.html > > UDP sockets have the sam

Re: the incredible shrinking socket

2002-07-08 Thread Luigi Rizzo
On Mon, Jul 08, 2002 at 11:56:22AM -0500, Mike Silbersack wrote: ... > Speaking of competition, someone should go look at this: > > http://mail-index.netbsd.org/current-users/2002/07/03/0011.html UDP sockets have the same problem... i posted patches for that case around dec.2000 which i never en

Re: the incredible shrinking socket

2002-07-08 Thread Mike Silbersack
On Mon, 8 Jul 2002, Mike Silbersack wrote: > On Sun, 7 Jul 2002, Alfred Perlstein wrote: > > > * Jonathan Lemon <[EMAIL PROTECTED]> [020707 21:48] wrote: > > > > > > I do have a smaller TIME_WAIT structure done; it even throws the socket > > > away since it isn't needed. The savings are current

Re: the incredible shrinking socket

2002-07-07 Thread Mike Silbersack
On Sun, 7 Jul 2002, Alfred Perlstein wrote: > * Jonathan Lemon <[EMAIL PROTECTED]> [020707 21:48] wrote: > > > > I do have a smaller TIME_WAIT structure done; it even throws the socket > > away since it isn't needed. The savings are currently about 500 bytes, > > and I can and also perform some

Re: the incredible shrinking socket

2002-07-07 Thread Alfred Perlstein
* Jonathan Lemon <[EMAIL PROTECTED]> [020707 21:48] wrote: > > I do have a smaller TIME_WAIT structure done; it even throws the socket > away since it isn't needed. The savings are currently about 500 bytes, > and I can and also perform some other savings in the general case. > > I think Alfred

Re: the incredible shrinking socket

2002-07-07 Thread Jonathan Lemon
In article [EMAIL PROTECTED]> you write: > >On Sun, 7 Jul 2002, Alfred Perlstein wrote: > >> Possibly, but the additional pointer dereference would be expensive >> and a lot of code would have to change without the compatibility >> macros. >> >> I sort of did it as a proof of concept, but of cour

Re: the incredible shrinking socket

2002-07-07 Thread Mike Silbersack
On Sun, 7 Jul 2002, Alfred Perlstein wrote: > Possibly, but the additional pointer dereference would be expensive > and a lot of code would have to change without the compatibility > macros. > > I sort of did it as a proof of concept, but of course since it doesn't > completely work I haven't pr

Re: the incredible shrinking socket

2002-07-07 Thread Jon Mini
On Sun, Jul 07, 2002 at 01:14:21PM -0700, Juli Mallett wrote: > * De: Mike Silbersack <[EMAIL PROTECTED]> [ Data: 2002-07-07 ] > [ Subjecte: Re: the incredible shrinking socket ] > > > > On Sun, 7 Jul 2002, Alfred Perlstein wrote: > > > > > Some ti

Re: the incredible shrinking socket

2002-07-07 Thread Alfred Perlstein
* Mike Silbersack <[EMAIL PROTECTED]> [020707 12:36] wrote: > > On Sun, 7 Jul 2002, Alfred Perlstein wrote: > > > Some time ago I noticed that there appeared to be several members > > of struct socket that were either only used by listen sockets or > > only used by data sockets. > > > > I've tak

Re: the incredible shrinking socket

2002-07-07 Thread Juli Mallett
* De: Mike Silbersack <[EMAIL PROTECTED]> [ Data: 2002-07-07 ] [ Subjecte: Re: the incredible shrinking socket ] > > On Sun, 7 Jul 2002, Alfred Perlstein wrote: > > > Some time ago I noticed that there appeared to be several members > > of struct socket th

Re: the incredible shrinking socket

2002-07-07 Thread Mike Silbersack
On Sun, 7 Jul 2002, Alfred Perlstein wrote: > Some time ago I noticed that there appeared to be several members > of struct socket that were either only used by listen sockets or > only used by data sockets. > > I've taken a stab at unionizing the members and we wind up saving > 28 bytes per soc