From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Lord Subject: Re: 2.6.25-rc8: FTP transfer errors Date: Thu, 10 Apr 2008 17:05:11 -0400 Message-ID: <47FE8107.4050400@rtr.ca> References: <47FCF9DD.6080007@rtr.ca> <20080410.023045.16227424.yoshfuji@linux-ipv6.org> <47FD138B.2060801@rtr.ca> <20080409.152933.132174258.davem@davemloft.net> <47FD590C.5020003@rtr.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , yoshfuji@linux-ipv6.org, Jeff Garzik , rjw@sisk.pl, LKML , linux-net@vger.kernel.org, Netdev To: =?UTF-8?B?SWxwbyBKw6RydmluZW4=?= Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Ilpo J=C3=A4rvinen wrote: > On Wed, 9 Apr 2008, Mark Lord wrote: >=20 >> David Miller wrote: >>> From: Mark Lord >>> Date: Wed, 09 Apr 2008 15:05:47 -0400 >>> >>>> But it would be far more useful for whoever has been working on th= e >>>> stack to suggest some possible/likely commits to look at instead. >>> Personally all I see is that one side closes the socket before all >>> data packets received have been read into the application, resultin= g >>> in a (correct) reset going out. >>> >>> I can't think of any change we've made over the course of this >>> release that would change behvaior in that area. >>> >>> So you will likely need to bisect. >> .. >> >> Or I can ignore it, like the net developers, since I have a workarou= nd. >> And then we'll see what other apps are broken upon 2.6.25 final rele= ase. >> >> Really, folks. Bug reports are intended to *help* the developers, >> not something to be thrown back in their faces. >> >> There do seem to have been a *lot* of changes around the tcp closing= /close >> code (as I see from diff'ing 2.6.24 against latest -git). =2E. > I might help if would add netdev on cc list in case you really want t= o=20 > reac net developers, otherwise they might just end up "ignoring it"..= =2E ;-) =2E. Oh.. I didn't know about that list. How does that differ from linux-ne= t ? (Thanks) >=20 >> reducing the mountain of commits to a big handful or two. >=20 > Those touching fin/close are mostly whitespace/move things, so I doub= t=20 > that you find these useful but in case you insist, here's the list: > > 056834d9f6f6eaf4cc7268569e53acab957aac27 [TCP]: cleanup tcp_{in,out}p= ut.c style > 058dc3342b71ffb3531c4f9df7c35f943f392b8d [TCP]: reduce tcp_output's i= ndentation levels a bit > 490d5046930276aae50dd16942649bfc626056f7 [TCP]: Uninline tcp_set_stat= e >=20 > In addition, there's this one (...though I have read it number of tim= es=20 > through and still cannot catch something that would cause the wrongne= ss=20 > you're seeing): >=20 > e870a8efcddaaa3da7e180b6ae21239fb96aa2bb [TCP]: Perform setting of co= mmon=20 > control fields in one place >=20 > There's very little really on interesting side I can think of, mostly= =20 > thinks are congestion control related changes... ...maybe either one = of=20 > these could cause something unpleasant in some corner case: >=20 > bd515c3e48ececd774eb3128e81b669dbbd32637 [TCP]: Fix TSO deferring > 0e3a4803aa06cd7bc2cfc1d04289df4f6027640a [TCP]: Force TSO splits to M= SS boundaries >=20 > ...e.g., if the latter causes a return with zero limit under some=20 > conditions, tso_fragment might generate, well, interesting packets an= d=20 > never finish if the condition persists but. =2E. That matches my own assessment there, too: lot's of whitespace changes, and not much real code difference on most paths. Bummer. :) -ml