From owner-freebsd-stable Mon Oct 29 5:17:44 2001 Delivered-To: freebsd-stable@freebsd.org Received: from freebsd.dk (fw-rl0.freebsd.dk [212.242.86.114]) by hub.freebsd.org (Postfix) with ESMTP id 36BE437B405; Mon, 29 Oct 2001 05:17:37 -0800 (PST) Received: (from sos@localhost) by freebsd.dk (8.11.6/8.11.3) id f9TDHQN45559; Mon, 29 Oct 2001 14:17:26 +0100 (CET) (envelope-from sos) From: Søren Schmidt Message-Id: <200110291317.f9TDHQN45559@freebsd.dk> Subject: Re: "VIA '686b southbridge fix applied" & -stable In-Reply-To: <20011029140547.D398-100000@howie.ncptiddische.net> To: Nils Holland Date: Mon, 29 Oct 2001 14:17:26 +0100 (CET) Cc: Randall Hopper , multimedia@FreeBSD.ORG, stable@FreeBSD.ORG Reply-To: sos@freebsd.dk X-Mailer: ELM [version 2.4ME+ PL94b (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=ISO-8859-1 Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG It seems Nils Holland wrote: > On Mon, 29 Oct 2001, Søren Schmidt wrote: > > > Depends, it seems this bug is a combination of mb vendors not > > knowing how to set the config regs of the VIA chips, and the > > SBLive not conforming to the specs. I've had reports that it > > works, but also afew that it doesn't... > > I guess the SBLive is not absolutely neccessary to make this bug occur. I > haven't gotten into this issue too deep, but from what I have read, it > seems that this bug might very well also occur if there is no SBLive in > the system. One article claimed that the SBLive puts the PCI bus under > heavier-than-normal stress, which in turn triggers the bug more often. The bug is a "feature" of misprogramming the VIA parts, or rather make use of some advance features that partly doesn't work as advertised and partly dopesn't work randomly together. The problem is related PCI bus traffic, what generates it is of less importance... The fix in -current just makes sure these features are set in a way that is known to work. > Furthermore, I guess that all "patches", like VIAs 4-in-1 driver for > Windows (never used that since I don't have Windows) and the FreeBSD patch > are only workarounds. The core of the problem has to be found (and fixed) > in the BIOS, AFAIK. From my own experience, a mobo I own was initially > subject to the bug, but once I updated the BIOS provided by the > manufacturer, I could do whatever I wanted to without ever having any > problems again... Well, the problem can be fixed anywhere, its just a matter of writing the right values in a couble of PCI config regs. It should be fixed in the BIOS, but some vendors have chosen ONLY to apply this fix if a SBLive is detected, and that is NOT enough to make sure things work, and the reason I did this "hack". Its a workaround no matter where its done :) -Søren To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message