On Thu, 7 Feb 2008, Brad Boyer wrote:
> On Fri, Feb 08, 2008 at 12:41:42AM +0100, Michael Schmitz wrote: > > at should not do direct hardware clock calls. If it really does, we'd > > need a fake-hwclk kernel module to keep it happy (fixing the hwclock > > kernel module for mac would be prefered, but that requires real > > hardware to test). > > It's not really that much code. We had it working for most Mac models in > the 2.2 tree. I think the Q900/Q950 were the only ones where it was > broken. The switch was based on ADB type, but that's not really the > right way to tell. In particular, the IIfx has the same clock > arrangement as the MacII style while the Q900/Q950 are a one-off thing > that nobody ever quite figured out completely. That explains why the patch below never worked on my Q950. I never sent the patch because I figured there was something, well, fixable. It works on CUDA. I think I tested Mac II ADB too, but I don't remember the results. I'm pretty sure that Roman's "move to adb/via init" comment* related to a build failure that Al Viro since fixed**. -f * http://linux-m68k-cvs.ubb.ca/c/cvsweb/linux/arch/m68k/mac/config.c#rev1.4 ** http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=3272244c2b1a8f13cec83c04b8245fa7fcb47a27 > > Brad Boyer > [EMAIL PROTECTED] > Index: linux-2.6.22/arch/m68k/mac/config.c =================================================================== --- linux-2.6.22.orig/arch/m68k/mac/config.c 2007-11-13 10:37:41.000000000 +1100 +++ linux-2.6.22/arch/m68k/mac/config.c 2007-11-13 10:38:09.000000000 +1100 @@ -166,10 +166,7 @@ void __init config_mac(void) mach_init_IRQ = mac_init_IRQ; mach_get_model = mac_get_model; mach_gettimeoffset = mac_gettimeoffset; -#warning move to adb/via init -#if 0 mach_hwclk = mac_hwclk; -#endif mach_set_clock_mmss = mac_set_clock_mmss; mach_reset = mac_reset; mach_halt = mac_poweroff; -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
