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]

Re: A loop.c patch


On Wed, Jul 01, 1998 at 01:45:59PM -0700, H.J. Lu wrote:
> I don't think so. It compares
> 
> 	rtx_equal_p (XEXP (loop_store_mems[i], 0), SET_SRC (pat))
> 
> not
> 	rtx_equal_p (loop_store_mems[i], SET_SRC (pat))
> 
> Basically, it checks if SET_SRC (pat) has the address of
> loop_store_mems[i].

Oh, I see.  No, this is something that find_base_value is 
supposed to determine.

Anyway, I can't make the test case fail, because true_dependence
_does_ recognize that
	(mem/s:SI (plus:SI (reg:SI 32) (const_int 4)))
and
	(mem:DI (plus:SI (reg:SI 6 %ebp) (const_int -8)))
conflict.


r~


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