From owner-freebsd-net@FreeBSD.ORG Wed Mar 21 02:59:08 2007 Return-Path: X-Original-To: freebsd-net@freebsd.org Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 3D0CD16A405 for ; Wed, 21 Mar 2007 02:59:08 +0000 (UTC) (envelope-from linuxinfoplus@gmail.com) Received: from ug-out-1314.google.com (ug-out-1314.google.com [66.249.92.173]) by mx1.freebsd.org (Postfix) with ESMTP id B540013C4C7 for ; Wed, 21 Mar 2007 02:59:07 +0000 (UTC) (envelope-from linuxinfoplus@gmail.com) Received: by ug-out-1314.google.com with SMTP id 71so112386ugh for ; Tue, 20 Mar 2007 19:59:06 -0700 (PDT) DKIM-Signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:mime-version:in-reply-to:references:content-type:message-id:content-transfer-encoding:from:subject:date:to:x-mailer; b=AOqsBWRVqgqXGlSNBUMrdlcIxzbBRIAsVpStOXje97wYDZBuGKySQrAG9b0kTIXEC/Wo+Axh3NAJxKFxPe22DxxKT2euImJ3tRfu1HRBLKycBnAfKrV1uYjKGMI/PoaI5Go/RoK46zkRkho0/l44a8DR2MtLSU9qSaGEJFt1glg= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:mime-version:in-reply-to:references:content-type:message-id:content-transfer-encoding:from:subject:date:to:x-mailer; b=Xja+nJl3P51CjJ4bSttEFBa1nmoZKPLknu2yzVKz3uNuXmWmv7NXtWEEW1rvGhfCNwU391tAc1von9xiBpC7TbYJRLPPNoPSpPYUQONqsWggojSB52zeqtYNDjFd+b5LyhzLMnKLyp86Z+g+Xlcj2K1zdL97JNnKqS8VHF5LbP0= Received: by 10.65.233.16 with SMTP id k16mr615241qbr.1174444446427; Tue, 20 Mar 2007 19:34:06 -0700 (PDT) Received: from ?192.168.3.211? ( [210.13.108.117]) by mx.google.com with ESMTP id 12sm7455432nzn.2007.03.20.19.34.03; Tue, 20 Mar 2007 19:34:05 -0700 (PDT) Mime-Version: 1.0 (Apple Message framework v752.2) In-Reply-To: <20070320120018.A6B7D16A5B3@hub.freebsd.org> References: <20070320120018.A6B7D16A5B3@hub.freebsd.org> Content-Type: text/plain; charset=GB2312; delsp=yes; format=flowed Message-Id: Content-Transfer-Encoding: quoted-printable From: rhinux Date: Wed, 21 Mar 2007 10:33:59 +0800 To: freebsd-net@freebsd.org X-Mailer: Apple Mail (2.752.2) Subject: Re: freebsd-net Digest, Vol 207, Issue 2 X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 21 Mar 2007 02:59:08 -0000 =D4=DA 2007-3-20=A3=AC=CF=C2=CE=E78:00=A3=ACfreebsd-net-request@freebsd.or= g =D0=B4=B5=C0=A3=BA > Send freebsd-net mailing list submissions to > freebsd-net@freebsd.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://lists.freebsd.org/mailman/listinfo/freebsd-net > or, via email, send a message with subject or body 'help' to > freebsd-net-request@freebsd.org > > You can reach the person managing the list at > freebsd-net-owner@freebsd.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of freebsd-net digest..." > > > Today's Topics: > > 1. Re: Wireshark (Shteryana Shopova) > 2. Interface index hack in IP_ADD_MEMBERSHIP (Bruce M Simpson) > 3. Re: Interface index hack in IP_ADD_MEMBERSHIP (Eugene Grosbein) > 4. Re: [PATCH] bge(4) patch for -STABLE (Pete French) > 5. Re: Interface index hack in IP_ADD_MEMBERSHIP (Bruce M Simpson) > 6. Re: rc.order wrong (ipfw) (Doug Barton) > 7. Re: Wireshark (Randall Stewart) > 8. Re: [PATCH] bge(4) patch for -STABLE (Mike Tancsa) > 9. Re: [PATCH] bge(4) patch for -STABLE (Pete French) > 10. [PATCH] Multicast refcounting in network stack (Bruce M Simpson) > 11. Re: [PATCH] Multicast refcounting in network stack > (Andre Oppermann) > 12. Re: netisr_direct (Keith Arner) > 13. networking code and splx() (Ignacio Rey) > 14. Re: rc.order wrong (ipfw) (David Gilbert) > 15. Re: networking code and splx() (Julian Elischer) > 16. Re: networking code and splx() (Bruce M. Simpson) > 17. Re: PMTU Discovery support (Kevin Lahey) > 18. Re: PMTU Discovery support (Kevin Lahey) > 19. Re: PMTU Discovery support (Bruce M. Simpson) > 20. Re: [PATCH] Multicast refcounting in network stack > (Bruce M Simpson) > 21. ifstated behavior (Alexandre Biancalana) > 22. Re: networking code and splx() (John Hay) > 23. Re: networking code and splx() (Eygene Ryabinkin) > 24. "em" driver shuts down interface when "ifconfig media > 100baseTX" is invoked (Infraservice hostmaster) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Mon, 19 Mar 2007 14:25:49 +0200 > From: "Shteryana Shopova" > Subject: Re: Wireshark > To: "manuel.ochoa@yahoo.com" > Cc: Max Laier , freebsd-net@freebsd.org > Message-ID: > <61b573980703190525s30f22648od0ecdecd01879d0c@mail.gmail.com> > Content-Type: text/plain; charset=3DUTF-8; format=3Dflowed > > On 3/19/07, manuel.ochoa@yahoo.com wrote: >> Max, correct me if I'm wrong but tcpdump will only give you the =20 >> headers, is that correct? This is fine most of the time but =20 >> sometimes I need to capture full frames. > > Nope - that's not correct - > > #tcpdump -s 0 > > will capture full frames. > > Shteryana > >> >> Thanks >> Manuel Ochoa CCNP MCSA MCSE MCDBA >> >> >> >> >> ----- Original Message ---- >> From: Max Laier >> To: freebsd-net@freebsd.org >> Cc: manuel.ochoa@yahoo.com >> Sent: Saturday, March 17, 2007 2:05:06 PM >> Subject: Re: Wireshark >> >> >> On Saturday 17 March 2007 19:16, manuel.ochoa@yahoo.com wrote: >>> Can someone please explain the difference between Wireshark and >>> Wireshark-lite. I would like to install a packet sniffer on my =20 >>> FreeBSD >>> box for CLI only. Thanks, >> >> What's wrong with tcpdump(8)? Other than that building either the =20= >> real or >> the -lite version with "WITHOUT_X11" defined will get you the >> cli-version. "-lite" seems to just disable a couple of dissectors =20= >> that >> have a lot of external dependencies. >> >> -- >> /"\ Best regards, | mlaier@freebsd.org >> \ / Max Laier | ICQ #67774661 >> X http://pf4freebsd.love2party.net/ | mlaier@EFnet >> / \ ASCII Ribbon Campaign | Against HTML Mail and News >> >> >> >> _____________________________________________________________________=20= >> _______________ >> Expecting? Get great news right away with email Auto-Check. >> Try the Yahoo! Mail Beta. >> http://advision.webevents.yahoo.com/mailbeta/newmail_tools.html >> _______________________________________________ >> freebsd-net@freebsd.org mailing list >> http://lists.freebsd.org/mailman/listinfo/freebsd-net >> To unsubscribe, send any mail to "freebsd-net-=20 >> unsubscribe@freebsd.org" >> > > > ------------------------------ > > Message: 2 > Date: Mon, 19 Mar 2007 14:28:52 +0000 > From: Bruce M Simpson > Subject: Interface index hack in IP_ADD_MEMBERSHIP > To: net@FreeBSD.org > Message-ID: <45FE9E24.8010201@incunabulum.net> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Hi, > > I plan to get rid of the ugly little ip_multicast_if() hack in the IP > stack.=3D > Before I do, is anyone actually using this? > > RFC 3678 specifies a protocol independent API for socket group > memberships which allow joins on interfaces referenced by index. =20 > This is > intended to support IGMPv3 and MLDv2. > > Regards, > BMS > > > > > ------------------------------ > > Message: 3 > Date: Mon, 19 Mar 2007 22:28:37 +0700 > From: Eugene Grosbein > Subject: Re: Interface index hack in IP_ADD_MEMBERSHIP > To: Bruce M Simpson > Cc: net@freebsd.org > Message-ID: <20070319152837.GA3984@svzserv.kemerovo.su> > Content-Type: text/plain; charset=3Dus-ascii > > On Mon, Mar 19, 2007 at 02:28:52PM +0000, Bruce M Simpson wrote: > >> I plan to get rid of the ugly little ip_multicast_if() hack in the IP >> stack.=3D >> Before I do, is anyone actually using this? >> >> RFC 3678 specifies a protocol independent API for socket group >> memberships which allow joins on interfaces referenced by index. =20 >> This is >> intended to support IGMPv3 and MLDv2. > > I recall that routed and ripd used to utilize something similar > long time ago. I'm not sure if they have switched to another API. > > Eugene > > > ------------------------------ > > Message: 4 > Date: Mon, 19 Mar 2007 16:28:58 +0000 > From: Pete French > Subject: Re: [PATCH] bge(4) patch for -STABLE > To: freebsd-net@FreeBSD.org, jkim@FreeBSD.org > Cc: freebsd-stable@FreeBSD.org > Message-ID: > >> I have made bge(4) patch for -STABLE (sorry, not suitable for >> RELENG_6_2): > > What dates stable is this relative to ? I am trying to apply your > patch to a cvsup of stable pulled on the day/time you sent your email, > but parts of it are failing for me unfortunately. I would like to =20 > test this > as I have a nmuber of bge interfaces running on some systems. > > -pete. > > > ------------------------------ > > Message: 5 > Date: Mon, 19 Mar 2007 16:51:29 +0000 > From: Bruce M Simpson > Subject: Re: Interface index hack in IP_ADD_MEMBERSHIP > To: Eugene Grosbein > Cc: net@freebsd.org > Message-ID: <45FEBF91.2000709@incunabulum.net> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Eugene Grosbein wrote: >> >> I recall that routed and ripd used to utilize something similar >> long time ago. I'm not sure if they have switched to another API. >> > You're right -- this would break routed on point-to-point interfaces. > > They didn't, unless it was updated at the upstream, i.e. rhyolite.com. > > This means that the RFC1724 hack can't be safely deprecated without > breaking this use case, until routed is updated to use the RFC 3678 > protocol-independent ASM API. > > Linux uses a slightly different technique to work-around this; ip_mreq > is expanded to ip_mreqn internally, and the interface index is > explicitly passed around in the kernel. > > The blocker in the FreeBSD case which prevents us simply adopting this > is the source interface selection logic in ip_output(). > > Regards, > BMS > > > ------------------------------ > > Message: 6 > Date: Mon, 19 Mar 2007 10:03:42 -0700 > From: Doug Barton > Subject: Re: rc.order wrong (ipfw) > To: Kian Mohageri > Cc: freebsd-net@freebsd.org, Mark Andrews , > freebsd-rc@freebsd.org > Message-ID: <45FEC26E.40504@FreeBSD.org> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Kian Mohageri wrote: > >> After re-reading your original idea, I think I understand a little >> better what you mean to do. For clarification, are you proposing =20 >> that >> the [early] firewall scripts do nothing if =20 >> firewall_late_enable=3DYES, and >> then have all firewalling taken care of later in the boot process =20 >> (i.e. >> post-networking) by firewall_late? >> >> I think I might have misunderstood your original proposal:) > > I think so too. :) To be clear, what I'm suggesting is that we move > ipfw and pf to a spot in the rcorder that is ahead of netif, along > with ipfilter which is already there. I am not suggesting that we > change their functionality, just the ordering. As a completely > separate thing (although they could be done at the same time) I am > suggesting _adding_ a new script for "late" firewall rules (where > "late" is defined as after netif) so that people who want to do > firewall-related things that require netif (like cloned interfaces, > FQDN rules, etc.) will have a standard way to accomplish that. > > Thanks for the opportunity to clarify, > > Doug > > --=20 > > This .signature sanitized for your protection > > > > ------------------------------ > > Message: 7 > Date: Mon, 19 Mar 2007 13:41:21 -0400 > From: Randall Stewart > Subject: Re: Wireshark > To: Shteryana Shopova > Cc: Max Laier , "manuel.ochoa@yahoo.com" > , freebsd-net@freebsd.org > Message-ID: <45FECB41.3070601@cisco.com> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Shteryana Shopova wrote: >> On 3/19/07, manuel.ochoa@yahoo.com wrote: >>> Max, correct me if I'm wrong but tcpdump will only give you the >>> headers, is that correct? This is fine most of the time but =20 >>> sometimes >>> I need to capture full frames. >> >> Nope - that's not correct - >> >> #tcpdump -s 0 >> >> will capture full frames. > > But nothing IMO beats wireshark for being able > to go in and analyze a dump .. searching on various > condition's fields etc.. > > It does not matter to me generally how its collected > wireshark/tcpdump -s 0.. > > But to analyze it.. give me wireshark any day :-D > > R > > > --=20 > Randall Stewart > NSSTG - Cisco Systems Inc. > 803-345-0369 803-317-4952 (cell) > > > ------------------------------ > > Message: 8 > Date: Mon, 19 Mar 2007 13:51:48 -0400 > From: Mike Tancsa > Subject: Re: [PATCH] bge(4) patch for -STABLE > To: Pete French , > freebsd-net@freebsd.org, jkim@freebsd.org > Cc: freebsd-stable@freebsd.org > Message-ID: <200703191753.l2JHreQ3061174@lava.sentex.ca> > Content-Type: text/plain; charset=3D"us-ascii"; format=3Dflowed > > At 12:28 PM 3/19/2007, Pete French wrote: >>> I have made bge(4) patch for -STABLE (sorry, not suitable for >>> RELENG_6_2): >> >> What dates stable is this relative to ? I am trying to apply your >> patch to a cvsup of stable pulled on the day/time you sent your =20 >> email, >> but parts of it are failing for me unfortunately. I would like to =20 >> test this >> as I have a nmuber of bge interfaces running on some systems. > > Mine applied cleanly to sources from last Friday. So far so good for > me in that I have not yet seen the watchdog timeout (previously once > every 4 days or so) but its too early to tell. Still, I have not > seen any regressions with it yet since installing it last > Saturday. This is a fairly busy recursive DNS server > > ---Mike > > > > ------------------------------ > > Message: 9 > Date: Mon, 19 Mar 2007 18:08:23 +0000 > From: Pete French > Subject: Re: [PATCH] bge(4) patch for -STABLE > To: freebsd-net@freebsd.org, jkim@freebsd.org, mike@sentex.net > Cc: freebsd-stable@freebsd.org > Message-ID: > >> Mine applied cleanly to sources from last Friday. > > O.K., that works (now I have the correct date in my supfile). Will > give it a shot... > > -pete. > > > ------------------------------ > > Message: 10 > Date: Mon, 19 Mar 2007 18:52:32 +0000 > From: Bruce M Simpson > Subject: [PATCH] Multicast refcounting in network stack > To: freebsd-net@freebsd.org > Message-ID: <45FEDBF0.4050105@incunabulum.net> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Hi, > > A patch against -CURRENT is now available: > http://people.freebsd.org/~bms/dump/multi_refcounting.diff > > This is a fairly sweeping architectural change which should resolve > memory leaks and potential panics with the network stack as a =20 > whole, to > better support interface detach at runtime. > > I'd like to check it in as soon as possible as it fixes the root cause > of the problems we have had with carp and pfsync in our stack. NetBSD > has implemented refcounting like this for some time now, so it does =20= > not > suffer from the same problems. > > Regards, > BMS > > > ------------------------------ > > Message: 11 > Date: Mon, 19 Mar 2007 20:11:56 +0100 > From: Andre Oppermann > Subject: Re: [PATCH] Multicast refcounting in network stack > To: Bruce M Simpson > Cc: freebsd-net@freebsd.org > Message-ID: <45FEE07C.4060501@freebsd.org> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Bruce M Simpson wrote: >> Hi, >> >> A patch against -CURRENT is now available: >> http://people.freebsd.org/~bms/dump/multi_refcounting.diff >> >> This is a fairly sweeping architectural change which should resolve >> memory leaks and potential panics with the network stack as a =20 >> whole, to >> better support interface detach at runtime. >> >> I'd like to check it in as soon as possible as it fixes the root =20 >> cause >> of the problems we have had with carp and pfsync in our stack. NetBSD >> has implemented refcounting like this for some time now, so it =20 >> does not >> suffer from the same problems. > > Patch looks good. :-) > > --=20 > Andre > > > > ------------------------------ > > Message: 12 > Date: Mon, 19 Mar 2007 15:54:55 -0400 > From: "Keith Arner" > Subject: Re: netisr_direct > To: "Robert Watson" > Cc: net@freebsd.org > Message-ID: > <8e552a500703191254qba4e194y810eb99a9f07bed8@mail.gmail.com> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > On 3/11/07, Robert Watson wrote: >> >> >> There are several ways we could start to reduce contention on that =20= >> lock: >> >> (3) Move towards greater granularity of locking for the tcbinfo: =20 >> instead >> of >> a single mutex, move to more than one locks, so that different >> connections >> processed simultaneously are likely to involve different =20 >> locks. For >> listen sockets, we would have to have a special case, such as a >> single >> lock to serialize simultaneous lock acquisitions of multiple =20 >> chain >> locks >> (for example). >> > > I've been thinking about this approach, and it does sound like the =20 > simplest > to > implement of the three proposed. However, the special case of the =20 > listen > socket > seems like it would complicate matters. > > It seems to me, however, that the complication of the listen socket =20= > could be > simplified > if the listen sockets were maintained in a separate pcb list from =20 > the rest > of the TCP > sockets (the fully connected sockets). If the two types of sockets =20= > were > thus separated, > the code would acquire the lock on the bucket in the connect hash, and > search the connect > hash; if there was a miss, acquire the lock on the listen list and =20 > then > search the listen list. > > The lock on the listen list should follow the locks on the connect =20 > buckets > in the locking > order. The connection bucket should be locked first, as delivery =20 > of data > would be > the common case. To create a new connection in the tcp_input path, =20= > both the > connect > bucket and the listen list would need to be locked (connect bucket =20 > as a new > entry would > be added to the list, and listen list as the accept socket would be > protected from going away). > > Keith > > > ------------------------------ > > Message: 13 > Date: Mon, 19 Mar 2007 20:37:09 +0100 > From: Ignacio Rey > Subject: networking code and splx() > To: freebsd-net@freebsd.org > Message-ID: <20070319203709.1272a470@debian> > Content-Type: text/plain; charset=3DUS-ASCII > > Hello everyone, > > I'm studying a bit the FreeBSD networking code. > > I've read "TCP/IP illustrated vol 2" by G. R. Wright and W. R. =20 > Stevens, > which describes code in 4.4BSD-lite. > > Now I'm taking a look at FreeBSD 6.2 release. Some things are =20 > different, > many others kept the same. What I'm confused about is that in 4.4BSD > there are many calls to the splx() family of functions, and I =20 > didn't see > any of them in the networking code in FreeBSD. However the 'grep' =20 > program > showed me that they are used in other parts of the kernel. > > > The question is: Have calls to these functions been wrapped? or are =20= > they > simply not used in this context? > > > Thanks in advance, > > Ignacio > > > ------------------------------ > > Message: 14 > Date: Mon, 19 Mar 2007 15:12:52 -0500 > From: David Gilbert > Subject: Re: rc.order wrong (ipfw) > To: Doug Barton > Cc: freebsd-net@freebsd.org, Mark Andrews , = Kian > Mohageri , freebsd-rc@freebsd.org > Message-ID: <17918.61124.353668.804988@canoe.dclg.ca> > Content-Type: text/plain; charset=3Dus-ascii > >>>>>> "Doug" =3D=3D Doug Barton writes: > > Doug> Kian Mohageri wrote: >>> I agree VERY MUCH with this sort of approach. It would be a much >>> cleaner solution than completely separate handling of all of these >>> different problems. I'm trying to get an idea of what all of the >>> major problems with the current order are, and these are the ones >>> I'm aware of: >>> >>> - ipfw blocks by default (names unresolvable, rtsol breaks) - >>> ipf/pf pass by default (services are unprotected) >>> >>> I think a firewall_boot script (similar to what you've proposed) >>> could potentially solve all of these problems. > > Doug> exception, not the rule. Furthermore (and I'm betraying a > Doug> prejudice here) I think that firewall rules that rely on name > Doug> resolution are absolutely nuts, and I say that with many years > Doug> of experience as a professional DNS and system administrator. > > I think you're misreading the above. The poster is saying that > because ipfw's default behaviour is block, loading it at the wrong > time can break other startup items because they require name > resolution or the sending of packets (rtsol). > > Dave. > > --=20 > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=20 > =3D=3D=3D=3D=3D=3D > |David Gilbert, Independent Contractor. | Two things can =20 > be | > |Mail: dave@daveg.ca | equal if and only =20 > if they | > |http://daveg.ca | are precisely =20 > opposite. | > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3DGLO=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=20 > =3D=3D=3D=3D=3D=3D > > > ------------------------------ > > Message: 15 > Date: Mon, 19 Mar 2007 13:23:17 -0700 > From: Julian Elischer > Subject: Re: networking code and splx() > To: Ignacio Rey > Cc: freebsd-net@freebsd.org > Message-ID: <45FEF135.2050203@elischer.org> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Ignacio Rey wrote: >> Hello everyone, >> >> I'm studying a bit the FreeBSD networking code. >> >> I've read "TCP/IP illustrated vol 2" by G. R. Wright and W. R. =20 >> Stevens, >> which describes code in 4.4BSD-lite. >> >> Now I'm taking a look at FreeBSD 6.2 release. Some things are =20 >> different, >> many others kept the same. What I'm confused about is that in 4.4BSD >> there are many calls to the splx() family of functions, and I =20 >> didn't see >> any of them in the networking code in FreeBSD. However the 'grep' =20 >> program >> showed me that they are used in other parts of the kernel. > > There was a major change in the way that synchronization was done =20 > bewteen FreeBSD 4.x > and FreeBSD 5.X. > The changes were to support real multiprocessor operation and were =20 > extensive. > The 'spl" method of operation was only able to protect operation on =20= > a single CPU. > > You should probably get a copy of "The design of the FreeBSD (um =20 > 5.2 I think) > Operating System by Kirk McKusick and George Neville-Neil. > It goes into some of the changes. (many changes have happenned =20 > since then too). > the New locking largely depends on Mutexes. > >> >> >> The question is: Have calls to these functions been wrapped? or =20 >> are they >> simply not used in this context? >> >> >> Thanks in advance, >> >> Ignacio >> _______________________________________________ >> freebsd-net@freebsd.org mailing list >> http://lists.freebsd.org/mailman/listinfo/freebsd-net >> To unsubscribe, send any mail to "freebsd-net-=20 >> unsubscribe@freebsd.org" > > > > ------------------------------ > > Message: 16 > Date: Mon, 19 Mar 2007 22:14:33 +0000 > From: "Bruce M. Simpson" > Subject: Re: networking code and splx() > To: Ignacio Rey > Cc: freebsd-net@freebsd.org > Message-ID: <45FF0B49.9060008@FreeBSD.org> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Ignacio Rey wrote: >> ... >> The question is: Have calls to these functions been wrapped? or =20 >> are they >> simply not used in this context? >> > splx() and friends have been no-ops since FreeBSD 5.x was branched. > Synchronization is now done using other mechanisms such as mutexes and > spin locks. See the new man page locking(9) in -CURRENT. > > Regards, > BMS > > > ------------------------------ > > Message: 17 > Date: Mon, 19 Mar 2007 14:54:22 -0700 > From: Kevin Lahey > Subject: Re: PMTU Discovery support > To: freebsd-net@freebsd.org > Message-ID: <20070319145422.39bfddcd@yakko.patheticgeek.net> > Content-Type: text/plain; charset=3DUS-ASCII > > On Tue, 6 Mar 2007 10:35:42 +0530 > "aditya kiran" wrote: > >> RFC 1191 says to increase the PMTU at some itnerval (15 minutes =20 >> default) > > 10 minutes. > >> next time a packet is sent, this will be used... and if PMTU is =20 >> really >> increased, >> no ICMP error will be recieved. that shows an increase in the =20 >> PMTU. I'm >> trying to >> understand if this mechanism is there in freebsd. any on this is =20 >> appreicated > > It looks to me as though FreeBSD stores per-host MTU data in the > hostcache, which gets purged after five minutes of inactivity. If > that's actually how it works, then, yes, FreeBSD should indeed > periodically probe for larger PMTUs. > > Of course, the real test is to set up a few hosts and see what > happens, rather than speculating based on a quick perusal of the > code. :-) > > Kevin > kml@patheticgeek.net > > > ------------------------------ > > Message: 18 > Date: Mon, 19 Mar 2007 17:21:56 -0700 > From: Kevin Lahey > Subject: Re: PMTU Discovery support > To: freebsd-net@freebsd.org > Message-ID: <20070319172156.68cba0a9@yakko.patheticgeek.net> > Content-Type: text/plain; charset=3DUS-ASCII > > On Mon, 19 Mar 2007 14:54:22 -0700 > Kevin Lahey wrote: > >> Of course, the real test is to set up a few hosts and see what >> happens, rather than speculating based on a quick perusal of the >> code. :-) > > After my slap-dash read of the current FreeBSD code, I was a little > concerned that I'd missed something. As penance, I set up a quick > experiment with four hosts connected in a line, A <-> B <-> C <-> D, > set the MTU on the links from B to C to 512, and ran ttcp from A to D. > PMTUD worked correctly. Then I suspended the ttcp process, went away > for an hour, and resumed it. Watching tcpdump, it appears that 512 > octet packets continued to be sent, with no attempt at probing. > > That would seem to be a bug. > > The boxes were running FreeBSD-6.1, but I can't really vouch for the > particular kernel configuration. It could well be that the problem is > with the loose nut behind the wheel, rather than with FreeBSD. :-) > > Kevin > kml@patheticgeek.net > > > ------------------------------ > > Message: 19 > Date: Tue, 20 Mar 2007 01:47:27 +0000 > From: "Bruce M. Simpson" > Subject: Re: PMTU Discovery support > To: Kevin Lahey > Cc: freebsd-net@freebsd.org > Message-ID: <45FF3D2F.3040000@FreeBSD.org> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Kevin Lahey wrote: >> >> The boxes were running FreeBSD-6.1, but I can't really vouch for the >> particular kernel configuration. It could well be that the =20 >> problem is >> with the loose nut behind the wheel, rather than with FreeBSD. :-) >> > > I believe PMTU measurements may only be relied upon for active TCP > connections, but it's been a while since I read this code. > > It would be useful if non-TCP drivers such as gre(4) could be extended > to perform PMTU discovery and auto-tune their MTU based on this, as > manually setting the MTU is a bit random and can result in horrible > fragmentation when going across the big-I Internet. > > I imagine doing this would require changes to the icmp input path =20 > and a > bit of abstraction. > > Regards, > BMS > > > ------------------------------ > > Message: 20 > Date: Tue, 20 Mar 2007 01:48:00 +0000 > From: Bruce M Simpson > Subject: Re: [PATCH] Multicast refcounting in network stack > To: Andre Oppermann > Cc: freebsd-net@freebsd.org > Message-ID: <45FF3D50.3050604@incunabulum.net> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Andre Oppermann wrote: >>> http://people.freebsd.org/~bms/dump/multi_refcounting.diff >> Patch looks good. :-) > Committed, with some changes. > > Regards, > BMS > > > ------------------------------ > > Message: 21 > Date: Tue, 20 Mar 2007 00:22:55 -0300 > From: Alexandre Biancalana > Subject: ifstated behavior > To: freebsd-net@freebsd.org > Message-ID: <45FF538F.1050405@seudns.net> > Content-Type: text/plain; charset=3DISO-8859-1; format=3Dflowed > > Hi list, > > First, excuse-me by the off-topic message, I asked this on -=20 > questions > but I don't have any answer. > > I'm trying to setup ifstated to check two links and if some go down, > do some actions like change pf rules and machine's route. > > My doubt is about the execution order/repetition of the states =20 > body of > ifstated.conf, in all configs that I tried just the last check is > executed always, follow and example: > > ifstated.conf: > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D > loglevel debug > > ping1 =3D '( "ping -q -c 1 -t 3 www.site1.com > > /dev/null" every 10 ) ' > ping2 =3D '( "ping -q -c 1 -t 3 www.site2.com > > /dev/null" every 10 ) ' > > state one { > if ! ( $ping1 && $ping2 ) { > set-state two > } > } > > state two { > > init { > run "logger -p console.notice -t ifstated 'Restarting > network !'" > } > > if ( $ping && $ping2 ) { > set-state one > } > } > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D > > # ifstated -dv > ping1 =3D "( "ping -q -c 1 -t 3 www.site1.com > > /dev/null" every 10 ) " > ping2 =3D "( "ping -q -c 1 -t 3 www.site2.com > > /dev/null" every 10 ) " > ifstated: initial state: one > ifstated: changing state to one > ifstated: running ping -q -c 1 -t 3 www.site1.com www.site1.com> >> /dev/null > ifstated: running ping -q -c 1 -t 3 www.site2.com www.site2.com> >> /dev/null > ifstated: started > ifstated: changing state to two > ifstated: running ping -q -c 1 -t 3 www.site1.com www.site1.com> >> /dev/null > ifstated: running ping -q -c 1 -t 3 www.site2.com www.site2.com> >> /dev/null > ifstated: running ping -q -c 1 -t 3 www.site2.com www.site2.com> >> /dev/null > ifstated: running ping -q -c 1 -t 3 www.site2.com www.site2.com> >> /dev/null > > > As you can see, after change state ifstated execute only the *last* > check command of the statement (ping2) forever.... > > This is the expected behavior ? > > I'm running 6-STABLE + ifstated-20050505 (instaled via > /usr/ports/net/ifstated) > > Thanks for any help. > > Alexandre > > > ------------------------------ > > Message: 22 > Date: Tue, 20 Mar 2007 09:31:50 +0200 > From: John Hay > Subject: Re: networking code and splx() > To: "Bruce M. Simpson" > Cc: freebsd-net@freebsd.org > Message-ID: <20070320073150.GA19859@zibbi.meraka.csir.co.za> > Content-Type: text/plain; charset=3Dus-ascii > > On Mon, Mar 19, 2007 at 10:14:33PM +0000, Bruce M. Simpson wrote: >> Ignacio Rey wrote: >>> ... >>> The question is: Have calls to these functions been wrapped? or =20 >>> are they >>> simply not used in this context? >>> >> splx() and friends have been no-ops since FreeBSD 5.x was branched. >> Synchronization is now done using other mechanisms such as mutexes =20= >> and >> spin locks. See the new man page locking(9) in -CURRENT. > > It does not seem to get installed: > > Doing a grep for locking in /usr/src/share/man/man9/Makefile produce > nothing. > > John > --=20 > John Hay -- John.Hay@meraka.csir.co.za / jhay@FreeBSD.org > > > ------------------------------ > > Message: 23 > Date: Tue, 20 Mar 2007 10:40:51 +0300 > From: Eygene Ryabinkin > Subject: Re: networking code and splx() > To: John Hay > Cc: freebsd-net@freebsd.org, "Bruce M. Simpson" > Message-ID: <20070320074051.GH96806@codelabs.ru> > Content-Type: text/plain; charset=3Dkoi8-r > > John, good day. > > Tue, Mar 20, 2007 at 09:31:50AM +0200, John Hay wrote: >>>> >>> splx() and friends have been no-ops since FreeBSD 5.x was branched. >>> Synchronization is now done using other mechanisms such as =20 >>> mutexes and >>> spin locks. See the new man page locking(9) in -CURRENT. >> >> It does not seem to get installed: > > The locking.9 is not the part of the current build as the comment > of the initial commit of that file says. > >> Doing a grep for locking in /usr/src/share/man/man9/Makefile produce >> nothing. > > But if you're running -CURRENT, then you can view the page from > the sources (assuming that you have the system sources in /usr/src/): > $ groff -Tascii -mandoc /usr/src/share/man/man9/locking.9 | less > > Our you can download the locking.9 from > http://www.freebsd.org/cgi/cvsweb.cgi/src/share/man/man9/locking.9 > --=20 > Eygene > > > ------------------------------ > > Message: 24 > Date: Tue, 20 Mar 2007 04:19:55 -0400 (EDT) > From: Infraservice hostmaster > Subject: "em" driver shuts down interface when "ifconfig media > 100baseTX" is invoked > To: freebsd-net@freebsd.org > Message-ID: > Content-Type: text/plain; charset=3Dus-ascii > > > FreeBSD cg-gw 6.2-PRERELEASE FreeBSD 6.2-PRERELEASE #10: Tue Jan 16 =20= > 01:40:27 EST 2007 root@taint:/usr/obj/usr/src/sys/FW-YQ i386 > > > I did: > "ifconfig em2 media 100baseTX mediaopt full-duplex" > > ifconfig em2 now reports: > ... > media: Ethernet 100baseTX (autoselect) > status: no carrier > > "ifconfig em2 media 100baseTX" also does this > > > We tried it on 2 different 100baseTX interfaces with the same result > > "ifconfig em2 media autonegotiate" brings the interface back up > > > It doesn't seem to happen on another interface which autonegotiates > 10baseT half-duplex, however > > > This seems to indicate there might be a driver problem since > the 2 interfaces connect to very different kinds of equipment > > > Any suggestions? > > > > ------------------------------ > > _______________________________________________ > freebsd-net@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-net > To unsubscribe, send any mail to "freebsd-net-unsubscribe@freebsd.org" > > End of freebsd-net Digest, Vol 207, Issue 2 > *******************************************