Date: Mon, 10 Feb 2003 11:16:55 +0100 (CET) From: Harti Brandt <brandt@fokus.fraunhofer.de> To: Vincent Jardin <vjardin@wanadoo.fr> Cc: "Andrew R. Reiter" <arr@watson.org>, "Matthew N. Dodd" <winter@jurai.net>, atm@FreeBSD.ORG Subject: Re: New version of ngATM Message-ID: <20030210110915.E33486@beagle.fokus.gmd.de> In-Reply-To: <3E26DAA600DC2A2B@mel-rta10.wanadoo.fr> (added by postmaster@wanadoo.fr) References: <Pine.NEB.3.96L.1030207134646.80673B-100000@fledge.watson.org> <20030208163348.I27486@beagle.fokus.gmd.de> <3E26DAA600DC2A2B@mel-rta10.wanadoo.fr> (added by postmaster@wanadoo.fr)
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 9 Feb 2003, Vincent Jardin wrote: VJ>I already use the HARP stack for PPPoE and PPPoEoA whithout any modification VJ>of the HARP source code ;-) PPPoA? VJ>> With regard to drivers: I'm going to busdmaify the en driver. I think I VJ>> have an LE155 card somewher here, so I will tackle the IDT driver after VJ>> that. When this is done, we have all that is supported with HARP also in VJ>> ngATM. VJ> VJ>The LE155 board uses the IDT77211 SAR chipset, doesn't it ? Then it is VJ>already supported by the HARP stack, moreover this driver uses DMA. It VJ>already supports UBR and CBR. [/usr/src/sys/dev/idt] # grep bus_dma [/usr/src/sys/dev/idt] # idt.c: buf = vtophys(idt->fixbuf); /* Transmit Status Queue Base */ *(volatile u_long *)(idt->virt_baseaddr + REGTSQB) = buf; This will fail on sparc64: vtophys is not the correct way to get a DMA-able address (the sparc has an IOMMU). u_long is probably wrong and means uint32_t. On sparc u_long is 64 bit. VJ>According to me, the HARP stack is a very stable code, it has lot of VJ>features and their drivers are working very well for a long time. If we VJ>need to improve the ATM support of FreeBSD, one should be more focused VJ>on the current code in order to improve the HARP stack and to add some VJ>new drivers. My feeling is, that nobody is going to do serious work on HARP. We can keep it running as long as possible. VJ>FreeBSD has already 2 ATM stacks : a basic one (NATM) and a full VJ>featured one (HARP), why is a third ATM stack required ? I received requests and questions from people outside the FreeBSD developers community for this ATM stuff. There is certainly interest in this. This will attract new users to the system. Forget about NATM (although I still find it to be a clever hack). harti -- harti brandt, http://www.fokus.gmd.de/research/cc/cats/employees/hartmut.brandt/private brandt@fokus.fraunhofer.de, harti@freebsd.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-atm" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030210110915.E33486>