From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ralf Baechle DO1GRB Subject: Re: ax25 kernel problem sometimes accur Date: Wed, 11 Dec 2002 13:30:17 +0100 Sender: linux-hams-owner@vger.kernel.org Message-ID: <20021211133016.A14287@linux-mips.org> References: <20021211024019.A28178@linux-mips.org> Mime-Version: 1.0 Return-path: Content-Disposition: inline In-Reply-To: ; from tpm@prkele.tky.hut.fi on Wed, Dec 11, 2002 at 12:29:12PM +0200 List-Id: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Tomi Manninen OH2BNS Cc: linux-hams@vger.kernel.org On Wed, Dec 11, 2002 at 12:29:12PM +0200, Tomi Manninen OH2BNS wrote: > Ok, I understand. Anyway some kind of progress report would be nice. Even > something like: "don't hold your breath", if that is appropriate... :-) Holding breath is definately something that could negativly impact your health ;-) > I have gotten surprisingly many private queries about this lately, so to > recap what I have understood so far: the 2.4 code is bad, full of races, > works only if treated nicely. Your SOCKOPS_WRAP removal obviously makes > things even worse until you redo the locking stuff. When that is done we > have a perfect ax.25 stack again. Right? ;-) In some sense the improments that are necessary to make the AX.25 stack "future proof" are ridiculous. AX.25 is about the worst protocol design I've ran over in my life, so making a multi-threaded implementation of it is an amusing thought. But the other legacy protocols like Appletalk are going through the same pain ... 73 de DO1GRB op Ralf -- Loc. JN47BS / CQ 14 / ITU 28 / DOK A21