This is the mail archive of the gcc@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: inline asm in an inline function

To: Richard Henderson <rth at redhat dot com>
Subject: Re: inline asm in an inline function
From: Aldy Hernandez <aldyh at redhat dot com>
Date: 10 Jun 2001 14:19:37 +0100
Cc: gcc at gcc dot gnu dot org
References: <991761446.2753.26.camel.cygnus.gcc@culebra.cygnus.com><jmn17m6df9.fsf@geoffk.org> <m3pucgpmrm.fsf_-_@culebra.cygnus.com> <20010608110427.A2926@redhat.com>


> 4.  Supress this change for GET_CODE (SET_SRC (set)) == ASM_OPERANDS.

alrighty.  scan_flow no longer tries to combine inline asms.  tested on
x86.  no regressions.

ok for branch? trunk?

cheers
aldy
 
--
Aldy Hernandez			E-mail: aldyh@redhat.com
Professional Gypsy
Red Hat, Inc.

2001-06-08  Aldy Hernandez  <aldyh@redhat.com>

	* loop.c (scan_loop): Do not combine asm statements.

Index: loop.c
===================================================================
RCS file: /cvs/gcc/egcs/gcc/loop.c,v
retrieving revision 1.322.4.9
diff -c -r1.322.4.9 loop.c
*** loop.c	2001/05/12 20:32:37	1.322.4.9
--- loop.c	2001/06/10 13:14:44
***************
*** 772,777 ****
--- 772,778 ----
  		  && (REGNO_LAST_UID (regno)
  		      == INSN_UID (regs->array[regno].single_usage))
  		  && regs->array[regno].set_in_loop == 1
+ 		  && GET_CODE (SET_SRC (set)) != ASM_OPERANDS
  		  && ! side_effects_p (SET_SRC (set))
  		  && ! find_reg_note (p, REG_RETVAL, NULL_RTX)
  		  && (! SMALL_REGISTER_CLASSES

References:
- Re: inline asm for powerpc question
  - From: Geoff Keating
- Re: inline asm in an inline function
  - From: Aldy Hernandez
- Re: inline asm in an inline function
  - From: Richard Henderson

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