On Tue, 7 Aug 2007, linux-mips@xxxxxxxxxxxxxx wrote: > Author: Ralf Baechle <ralf@xxxxxxxxxxxxxx> Tue Aug 7 17:18:28 2007 +0100 > Commit: 00cc123703425aa362b0af75616134cbad4e0689 > Gitweb: http://www.linux-mips.org/g/linux/00cc1237 > Branch: master > > Signed-off-by: Ralf Baechle <ralf@xxxxxxxxxxxxxx> > > --- > > include/asm-mips/mach-generic/ide.h | 76 ++++++++++++----------------------- > 1 files changed, 25 insertions(+), 51 deletions(-) This change breaks the SWARM -- depending on the setting of CONFIG_IDE_GENERIC, either a bus error happens because of blind probing or the onboard IDE interface gets designated as "ide2". I cannot really see a dependency between "idebus=" (which merely sets a variable somewhere in drivers/ide/ide.c) and code affected by this change -- what's the reason behind it? Also the comment is misleading -- there is almost nothing about IDE in the PCI spec; certainly nothing that would make I/O ports at 0x1f0 and 0x170 special. They may be special for a IDE controller PCI devices, but one cannot simply assume such a device is present somewhere on the bus (or that something will subtractively decode unclaimed addresses). Maciej