This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Global Approval (Re: Patch to fix several compat failures onirix6.5)
> OK to install? If so, I'll update the 3.4 changes.html entry.
>
> Richard
>
>
> * calls.c (initialize_argument_information): If an argument has no
> stack space associated with it, and BLOCK_REG_PADDING is defined,
> use it to decide at which end the argument should be padded.
> * function.c (assign_parms): Allocate BLKmode stack slots.
> * config/mips/mips-protos.h (mips_pad_arg_upward): Declare.
> (mips_pad_reg_upward): Declare.
> * config/mips/mips.h (PAD_VARARGS_DOWN): Use FUNCTION_ARG_PADDING.
> (CUMULATIVE_ARGS): Remove num_adjusts and adjusts.
> (FUNCTION_ARG_PADDING): Use mips_pad_arg_upward.
> (BLOCK_REG_PADDING): Use mips_pad_reg_upward.
> * config/mips/mips.c (struct mips_arg_info): Remove struct_p.
> (mips_expand_call): Remove code for generating structure shifts.
> (mips_arg_info): Don't set struct_p. Don't set fpr_p for non-float
> types unless using the EABI.
> (function_arg_advance): Don't generate shift instructions.
> (function_arg): Don't return them. Don't short-circuit the
> check for double structure chunks for DFmode arguments.
> (mips_pad_arg_upward, mips_pad_reg_upward): New functions.
> (mips_expand_prologue): Remove code to emit structure shifts.
> * config/mips/irix6-libc-compat.c: Remove workarounds for buggy
> structure passing (inet_ntoa, inet_lnaof, inet_netof). Update
> comments to say that only structure returns are a problem.
Looks good to me, even the target independent bits. Need someone to ack
it with global write permission though...
-eric
--
Eric Christopher <echristo@redhat.com>