Date: Tue, 25 Aug 1998 11:48:01 -0600 (MDT) From: "Kenneth D. Merry" <ken@plutotech.com> To: garbanzo@hooked.net (Alex) Cc: freebsd-current@FreeBSD.ORG Subject: Re: pciconf Message-ID: <199808251748.LAA15408@panzer.plutotech.com> In-Reply-To: <Pine.BSF.4.00.9808242314230.18055-100000@zippy.dyn.ml.org> from Alex at "Aug 24, 98 11:15:51 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Alex wrote... > On Sun, 23 Aug 1998, Kenneth D. Merry wrote: > [...] > > The problem is that the PCIOCGETCONF ioctl isn't supported in > > -current. There's nothing wrong with your machine. I've modified the PCI > > code in the CAM tree to support that ioctl again. So 'pciconf -l' now > > produces output like: > [..] > > The new version of that ioctl also supports matching against bus, > > device, function, device name and unit number. > > Will this perhaps be merged into the -current tree before the rest of CAM? > I know I'd like to try the CAM stuff out for other reasons, but I don't > quite feel like dealing with huge patches (or a whole 'nother source > tree). If someone wants to do it, that's fine with me. It'll take a couple of things first, though: - clear it with Stefan, since he's the PCI dude - realize that there are a couple of limitations to it Last week I finished work on some CAM transport layer matching code that has a number of similarities to the pciconf code I wrote. (it's similar, but it's far more complex than the pci stuff) Anyway, I realized that the way I did the match/result buffer in the PCI code isn't exactly the best way to do things. In the CAM matching code, I've got two separate buffers, one for patterns to match against, the other for match results. In the PCI code, there's only one buffer that's passed in. The user passes in a matching pattern, and gets back the result in the same buffer. It would be better to have two separate buffers so if the buffer the user passed in isn't big enough for the results, the user won't have to copy the pattern back in to the buffer before sending it again. One other thing that might be nice to have in the code is the ability to match against vendor and device id's. (Justin asked for this a few weeks ago, but I haven't had time/inclination to do it yet.) Ken -- Kenneth Merry ken@plutotech.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199808251748.LAA15408>