This is the mail archive of the gcc-patches@gcc.gnu.org mailing 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]

Re: [patch] Fix PR/13430 (compile time hog in bb-reorder.c)


On Monday 22 December 2003 07:52, Brad Lucier wrote:
> Shouldn't this magic number 8 be changed to use the --param mechanism?

Apart from the magic number, I just don't like the approach.  Is it possible 
to just not run the first bb reordering if STACK_REGS is defined?

AFAICT the only targets that defines STACK_REGS are i[3456]86.  These targets 
do not define INSN_SCHEDULING or LEAF_REGISTERS, ifcvt2 probably shouldn't be 
there (see Bug 12617), and flag_branch_target_load_optimize2 is always false. 
So there is nothing between rest_of_handle_reorder_blocks() and 
rest_of_handle_stack_regs() that is called for these targets.

Gr.
Steven


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