[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20180724140203.GV17271@n2100.armlinux.org.uk>
Date: Tue, 24 Jul 2018 15:02:04 +0100
From: Russell King - ARM Linux <linux@...linux.org.uk>
To: Alexei Colin <acolin@....edu>
Cc: linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
Arnd Bergmann <arnd@...db.de>,
Alex Bounine <alex.bou9@...il.com>,
Barry Wood <barry.wood@....com>,
John Paul Walters <jwalters@....edu>
Subject: Re: [PATCH] arm: enable RapidIO config options in Kconfig
On Tue, Jul 24, 2018 at 09:41:26AM -0400, Alexei Colin wrote:
> ARM SoCs with a PCI bus offer the RapiodIO config menu; SoCs with
> RapidIO IP blocks but without a PCI bus, need to add "select
> HAS_RAPIDIO" to the Kconfig entry for that SoC (e.g. ARCH_*).
>
> HAS_RAPIDIO was chosen over HAVE_RAPIDIO to be consistent with
> other architectures which already define this flag (powerpc).
Is there any reason we can't have the RAPIDIO and HAS_RAPIDIO config
blocks in drivers/rapidio/Kconfig (maybe without the || PCI - see
below), and then have architecture Kconfig files do:
config (ARM or PPC)
select HAS_RAPIDIO if PCI
or in the case of MIPS:
config MIPS
...
select HAS_RAPIDIO
That would avoid duplicating almost identical RAPIDIO and HAS_RAPIDIO
config blocks in architecture Kconfig files.
Why should rapidio be available just because we have PCI selected? If
it behaves as a PCI add-in card, then why isn't this available for any
architecture with PCI?
What about build coverage - shouldn't RAPIDIO be selectable for (eg)
other architectures via COMPILE_TEST if it's supposed to be generic?
--
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 13.8Mbps down 630kbps up
According to speedtest.net: 13Mbps down 490kbps up
Powered by blists - more mailing lists