From mboxrd@z Thu Jan 1 00:00:00 1970 From: "David S. Miller" Date: Mon, 05 Jan 2004 09:17:45 +0000 Subject: Re: [SPARC32] Early kernel panic with 2.6.1rc1 Message-Id: <20040105011745.3a6603e2.davem@redhat.com> List-Id: References: <1073231988.22653.23.camel@zorak> In-Reply-To: <1073231988.22653.23.camel@zorak> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: sparclinux@vger.kernel.org On Mon, 5 Jan 2004 03:05:58 -0500 (EST) Ricky Beam wrote: > On Sun, 4 Jan 2004, David S. Miller wrote: > >We cannot just depend upon PCI however for the 8250 driver options > >as it may appear also in ISA and onboard forms. > > No sparc32 or sparc64 system (that I know of) contain an actual 8250. > Furthermore, the only systems to have an ISA bus are PCI systems (PCI-ISA > bridged.) Granted, I've not used an SBUS based system in several years. The 8250 drives all variants, including the kind used for the 'su' devices in PCI sparc64 systems, this is what Peter was talking about when he alluded to merging the sunsu.c driver into the 8250.c code. Also, any system that has a PCI bus can therefore host one of those serial/parallel PCI cards which use 8250 chips often. So I stand where I started, that adding conditions to the 8250 Kconfig option is totally pointless as no selection of checks are totally suitable.