On Thursday 24 May 2012 10:55:32 pm Greg Ungerer wrote: > > This looks pretty good overall. Only a couple of comments below. > Otherwise I'd like to commit it to the m68knommu git tree (after the > current merge window closes).
Sure. Actually what I'd like to do is after the merge window closes submit a revised patch set (I have to re-submit the fec patch anyway). Same for the 5253. Some other minor patches as well. > > > > +config M5441x > > + bool "MCF5441x" > > + select MMU_COLDFIRE if MMU > > Maybe we should make this "depends on !MMU" until we really do > have the MMU support in place. There is pretty much no way it will > compile and work with the MMU enabled yet. Yeah, I tried compiling it with the MMU turned on. The lack of an FPU is going to make it challenging. Although I dont understand why the m68k/mmu code is so insistent that one needs an fpu... > > > > diff --git a/arch/m68k/Kconfig.machine b/arch/m68k/Kconfig.machine > > index 7cdf6b0..7031173 100644 > > --- a/arch/m68k/Kconfig.machine > > +++ b/arch/m68k/Kconfig.machine > > @@ -428,7 +428,6 @@ config SAVANTrosie1 > > help > > Support for the Savant Rosie1 board. > > > > - > > I like clean ups... But they should go in separate patches of their own. ;-P I had defined a config selection for the twr-mcf5441x board, but decided it should be in its own branch and didnt notice that sneaking in until it came to commit. > > > if !MMU || COLDFIRE > > > > comment "Machine Options" > > diff --git a/arch/m68k/Makefile b/arch/m68k/Makefile > > index 804f139..6940baf 100644 > > --- a/arch/m68k/Makefile > > +++ b/arch/m68k/Makefile > > @@ -34,6 +34,7 @@ cpuflags-$(CONFIG_M68030) := > > cpuflags-$(CONFIG_M68020) := > > cpuflags-$(CONFIG_M68360) := -m68332 > > cpuflags-$(CONFIG_M68000) := -m68000 > > +cpuflags-$(CONFIG_M5441x) := $(call cc-option,-mcpu=54455,-mcfv4e) > > What if the compiler supports neither -54455 or -mcfv4e? > Admittedly it would have to be a old compiler. But the thinking here > was that the second arg is the last resort fallback if the compiler > doesn't support the ideal compiler switch. It would have to be really old compiler. The -mcfv4e option appeared in gcc 3.4.0, released in april of '04. Now, I'm not saying everyone needs to use the latest gcc - I went back to 4.3.3 for my Coldfire projects after getting some strange crashes with 4.6.1, but I think anyone who tries to build the latest linux kernel for a relatively new processor variant with a an 8+ year old version of gcc deserves what they get. ;) On the other hand, I now think this is incorrect for this processor. Freescale, in some of their documentation refer to the 5441x as a v4m and CodeWarrior lists v4e devices as 547x and 548x and puts the 54411x and 54455 together with the 540x. _______________________________________________ uClinux-dev mailing list uClinux-dev@uclinux.org http://mailman.uclinux.org/mailman/listinfo/uclinux-dev This message was resent by uclinux-dev@uclinux.org To unsubscribe see: http://mailman.uclinux.org/mailman/options/uclinux-dev
