Roman Zippel - Re: [RTL, ColdFire 24/63] Add support for a MODE_INDEX_REG_CLASS macro (original) (raw)

This is the mail archive of the gcc-patches@gcc.gnu.orgmailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Hi,

On Wed, 14 Feb 2007, Jeffrey Law wrote:

So, from an optimisation perspective, I think it is reasonable to accept indexed addresses for SImode and not for SFmode. I'm still not sure whether you disagree with that. I disagree with that if the port allows integer values in FP regsters and the hardware can't do FP indexed loads/stores because G_I_L_A could be passed SImode but generate an FP load/store.

m68k doesn't allow integer values in fp register (rejected in m68k_regno_mode_ok), thus...

Now G_I_L_A has a mode argument precisely so that the legitimacy of an address can depend on the mode. However, G_I_L_A is not the only backend interface for describing addressing modes; we also have the BASE_REG_CLASS and INDEX_REG_CLASS macros. These latter macros are all that reload looks at when legitimizing addresses, so they must of course agree with G_I_L_A. I disagree on this point and as I've stated before I think you're using this to avoid fixing the backend to deal with a broken G_I_L_A.

...our G_I_L_A isn't broken in this regard.

bye, Roman

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]