[lra] rework deactivation of fp2sp elimination [PR120424]

Deactivating the fp2sp elimination in lra_update_fp2sp_elimination
prevents update_reg_eliminate from propagating the fp2sp elimination
offset to the next chosen elimination, so it may retain -1 as the
prev_offset, and prev_offset will be taken as an already-applied
offset that needs to be compensated in the next round of spilling and
reloading.  This affects, for example, crtbegin.o's
__do_global_dtors_aux on arm-linux-gnueabihf in a {BOOT_C,T}FLAGS='-O2
-g -fnon-call-exceptions -fstack-clash-protection' bootstrap.

Alas, just retaining that elimination causes spills to use the fp2sp
elimination, including applying sp offsets, which breaks e.g. an
x86_64-linux-gnu native bootstrap with ix86_frame_pointer_required
modified to return true on nonzero frame size.

The middle-ground solution is to keep the elimination active, so that
its offsets are applied and propagated on to the subsequent fp
elimination, but without introducing sp offsets, so that
e.g. pr103973-18.c on the modified x86_64-linux-gnu doesn't get
adjacent argument pushes of two adjacent on-stack temporaries ending
up pushing the same temporary because of undesired adjustments.

for  gcc/ChangeLog

	PR rtl-optimization/120424
	* lra-eliminations.cc (lra_update_fp2sp_elimination):
	Avoid sp offsets in further fp2sp eliminations...
	(update_reg_eliminate): ... and restore to_rtx before assert
	checking.

diff --git a/gcc/lra-eliminations.cc b/gcc/lra-eliminations.cc
index 6663d1c37e8ba7085d816b678200332cbd51dc75..0a702a43a5a17be8e5522a5683a82f097e35bdf2 100644 (file)
--- a/gcc/lra-eliminations.cc
+++ b/gcc/lra-eliminations.cc
@@ -1172,7 +1172,16 @@ update_reg_eliminate (bitmap insns_with_changed_offsets)
       /* If it is a currently used elimination: update the previous
         offset.  */
       if (elimination_map[ep->from] == ep)
-       ep->previous_offset = ep->offset;
+       {
+         ep->previous_offset = ep->offset;
+         /* Restore the stack_pointer_rtx into to_rtx, that
+            lra_update_fp2sp_elimination set to from_rtx, so that the assert
+            below still checks what it was supposed to check.  */
+         if (ep->from_rtx == ep->to_rtx
+             && ep->from != ep->to
+             && ep->from == FRAME_POINTER_REGNUM)
+           ep->to_rtx = stack_pointer_rtx;
+       }
 
       prev = ep->prev_can_eliminate;
       setup_can_eliminate (ep, targetm.can_eliminate (ep->from, ep->to));
@@ -1418,7 +1427,12 @@ lra_update_fp2sp_elimination (int *spilled_pseudos)
   ep = elimination_map[FRAME_POINTER_REGNUM];
   if (ep->to == STACK_POINTER_REGNUM)
     {
-      elimination_map[FRAME_POINTER_REGNUM] = NULL;
+      /* Prevent any further uses of fp, say in spill addresses, from being
+        eliminated to sp and affected by sp offsets.  Alas, deactivating the
+        elimination altogether causes the next chosen fp elimination to miss
+        the offset propagation, so it may keep -1 as its prev_offset, and that
+        will make subsequent offsets incorrect.  */
+      ep->to_rtx = ep->from_rtx;
       setup_can_eliminate (ep, false);
     }
   else