Date: Wed, 10 May 1995 00:27:34 -0700 From: "Justin T. Gibbs" <gibbs@estienne.CS.Berkeley.EDU> To: "Daniel M. Eischen" <deischen@iworks.InterWorks.org> Cc: peppe@unipg.it, questions@FreeBSD.org Subject: Re: 950412-SNAP on a Plato P90 with Adaptec 2940 Message-ID: <199505100727.AAA17435@estienne.cs.berkeley.edu> In-Reply-To: Your message of "Tue, 09 May 1995 22:22:39 CDT." <9505100322.AA09494@iworks.InterWorks.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> >>The problem is not "completely" solved. >> >>What I did try at the end is to use all the "new" >>aic7* files that drive the Adaptec 2940 with the >>"old" 950412-SNAP kernel. Now the things work, >>but I discovered, what, probabilly, is the problem >>with the Intel Plato Board. >> >>The new version of the aic7* files in -current >>showed me that the Intel Plato is using "edge-triggered" >>interrupts by default on PCI, instead of "level-sensitive". >>The Plato's BIOS initialize the card that way. > >Not correct. The -current aic7xxx drivers show that message for the benefit >of the aic7770-based cards (aha274x and aha284x). For these cards, the interr >upt >trigger type is meaningful. The interrupt trigger type is taken from the >HCNTRL register which is on the same location for both chips (7770 and 7870), >but the interrupt trigger type bit in HCNTRL is defined as a spare R/W bit >with no current hardware assignment for the 7870. For the 7770 it does have >meaning. So the message your seeing regarding the interrupt type is not >correct. Well, you could still be using edge triggered interrupts, but I >doubt anything would work. > >I have a 2940W and run both Linux and FreeBSD on it. The FLinux driver >always assigns a 1 to this bit which is interpreted as Level-sensitieve >interrupts. The FreeBSD driver leaves the bit unchanged (0) and interprets >it as Edge-triggered. What both drivers should do is ignore the bit >for the case of the 7870 (2940 cards) chip and not display any message >regarding interrupt trigger settings. > >This is just a nit and does not affect the performance of the driver in >any way. > >Dan Eischen >deischen@iworks.InterWorks.org I mentioned the exact same thing to you about two months ago (that the bit was undefined for the aic7870). When I told you, you responded that it did make a difference because you tried to set it to 0 on your system and it hung. If the Data Book is really correct in this regard, I'll kill the message for the aic7870 (I don't have a card right now to test it on). -- Justin T. Gibbs ============================================== TCS Instructional Group - Programmer/Analyst 1 Cory | Po | Danube | Volga | Parker | Torus ==============================================
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199505100727.AAA17435>