Date: Fri, 26 Dec 2003 16:34:54 -0700 From: Scott Long <scottl@freebsd.org> To: David Wolfskill <david@catwhisker.org> Cc: current@freebsd.org Subject: Re: "no driver attached" for Realtek 8129-based NIC Message-ID: <3FECC59E.2030605@freebsd.org> In-Reply-To: <200312262058.hBQKwC5Y019071@bunrab.catwhisker.org> References: <200312262058.hBQKwC5Y019071@bunrab.catwhisker.org>
next in thread | previous in thread | raw e-mail | index | archive | help
David Wolfskill wrote: > Ugh. Looks as if my cut/paste of the console output ran across a lone > '.' and somewhat truncated the output; worse, the next stuff was > interpreted as Cc: addresses. Sorry about that.... :-( > > The console output that was dropped was not relevant to the problem, > so I'll not bother with it further. > > I will point out that the kernel does still have the "rl" device; it > also has the "re" device defined. So rev 1.118 of sys/pci/if_rl.c is what breaks your configuration? You can't back that change out, but if you '#if 0' the test that changed in that rev (checking hwrev for certain values), does it make a difference? If it does, you and Bill Paul will have to get together and figure out if the test is correct. Possibilities include either the test being wrong, or the re(4) driver needing to be modified to recognise your card. Also, the PCI database contained in pciconf is completely separate from the drivers that are in the kernel. There is absolutely no connection between the two. pciconf works by parsing the PCI config space itself and matching the data that it finds to entries in it's database. There is nothing odd about what you are seeing; there are probably hundreds of devices that pciconf can identify but have no correspinding driver in FreeBSD at all. The same applies to the PCI configuration code in the kernel. Parsing the PCI config space is easy and has no bearing on what drivers may or may not attach to the hardware. Scott
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3FECC59E.2030605>