On Sat, May 22, 2004 at 02:23:32PM +0200, Thiemo Seufer wrote: > Christoph Hellwig wrote: > [snip] > > > Because they have a single upstream, while the kernel has several > > > for all the architectures. > > > > That's bullshit. You have a single upstream and patches vs it from > > different sources. > > I regard the linux-mips.org CVS as MIPS upstream, and I don't want to > handle patch conflicts with architectures I know nothing about.
Apparently your regard does not extend to looking what's really in there. Care to check the diff between linux-mips repository and upstream? Ralf et.al. are doing a very good job keeping it in sync with 2.6; right now it's pretty much vanilla 2.6.6 + patches in arch/mips, include/asm-mips + changes in/additions of mips-only stuff in drivers/*, include/linux + some noise they'd forgotten to drop (e.g. in fs/stat.c - && !defined(CONFIG_ARCH_S390) && !defined(__hppa__) \ + && !defined(CONFIG_ARCH_S390) && !defined(__hppa__) && !defined(__mips__) \ && !defined(__arm__) && !defined(CONFIG_V850) && !defined(__powerpc64__) \ && !defined(__mips__) adding !defined(__mips__) even though it's already in there). There are several contention points (fbmem.c, as usual) but they are trivial to handle. Please, have at least some respect to mips guys - they know what they are doing and talk about "patch conflicts with architectures you know nothing about" is a BS. Their stuff is well-behaving and localized enough to avoid that.
