* [PATCH] for PR16461
@ 2004-07-11 23:34 Zdenek Dvorak
2004-07-12 19:27 ` Jeffrey A Law
0 siblings, 1 reply; 4+ messages in thread
From: Zdenek Dvorak @ 2004-07-11 23:34 UTC (permalink / raw)
To: gcc-patches
Hello,
as Diego noted, tree-tailcall.c:independent_of_stmt_p must handle
non-SSA name arguments of phi nodes.
Bootstrapped & regtested on ia64.
Zdenek
PR tree-optimization/16461
* tree-tailcall.c (independent_of_stmt_p): Handle non-SSA name
arguments of phi nodes.
Index: tree-tailcall.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/tree-tailcall.c,v
retrieving revision 2.14
diff -c -3 -p -r2.14 tree-tailcall.c
*** tree-tailcall.c 8 Jul 2004 16:16:41 -0000 2.14
--- tree-tailcall.c 10 Jul 2004 16:55:20 -0000
*************** independent_of_stmt_p (tree expr, tree a
*** 238,243 ****
--- 238,245 ----
abort ();
expr = PHI_ARG_DEF_FROM_EDGE (at, e);
+ if (is_gimple_min_invariant (expr))
+ break;
}
/* Unmark the blocks. */
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] for PR16461
2004-07-11 23:34 [PATCH] for PR16461 Zdenek Dvorak
@ 2004-07-12 19:27 ` Jeffrey A Law
2004-07-12 19:38 ` Zdenek Dvorak
0 siblings, 1 reply; 4+ messages in thread
From: Jeffrey A Law @ 2004-07-12 19:27 UTC (permalink / raw)
To: Zdenek Dvorak; +Cc: gcc-patches
On Sun, 2004-07-11 at 14:01, Zdenek Dvorak wrote:
> Hello,
>
> as Diego noted, tree-tailcall.c:independent_of_stmt_p must handle
> non-SSA name arguments of phi nodes.
>
> Bootstrapped & regtested on ia64.
>
> Zdenek
>
> PR tree-optimization/16461
> * tree-tailcall.c (independent_of_stmt_p): Handle non-SSA name
> arguments of phi nodes.
I'd suggest using
if (TREE_CODE (expr) != SSA_NAME)
break;
Instead since only an SSA_NAME is valid for EXPR in that loop. Why
check is_gimple_min_invariant when it's clear that the only valid
node is an SSA_NAME?
jeff
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] for PR16461
2004-07-12 19:27 ` Jeffrey A Law
@ 2004-07-12 19:38 ` Zdenek Dvorak
2004-07-12 19:50 ` Jeffrey A Law
0 siblings, 1 reply; 4+ messages in thread
From: Zdenek Dvorak @ 2004-07-12 19:38 UTC (permalink / raw)
To: Jeffrey A Law; +Cc: gcc-patches
Hello,
> > as Diego noted, tree-tailcall.c:independent_of_stmt_p must handle
> > non-SSA name arguments of phi nodes.
> >
> > Bootstrapped & regtested on ia64.
> >
> > Zdenek
> >
> > PR tree-optimization/16461
> > * tree-tailcall.c (independent_of_stmt_p): Handle non-SSA name
> > arguments of phi nodes.
> I'd suggest using
>
> if (TREE_CODE (expr) != SSA_NAME)
> break;
>
> Instead since only an SSA_NAME is valid for EXPR in that loop. Why
> check is_gimple_min_invariant when it's clear that the only valid
> node is an SSA_NAME?
sure. Ok with the change?
Zdenek
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] for PR16461
2004-07-12 19:38 ` Zdenek Dvorak
@ 2004-07-12 19:50 ` Jeffrey A Law
0 siblings, 0 replies; 4+ messages in thread
From: Jeffrey A Law @ 2004-07-12 19:50 UTC (permalink / raw)
To: Zdenek Dvorak; +Cc: gcc-patches
On Mon, 2004-07-12 at 07:07, Zdenek Dvorak wrote:
> Hello,
>
> > > as Diego noted, tree-tailcall.c:independent_of_stmt_p must handle
> > > non-SSA name arguments of phi nodes.
> > >
> > > Bootstrapped & regtested on ia64.
> > >
> > > Zdenek
> > >
> > > PR tree-optimization/16461
> > > * tree-tailcall.c (independent_of_stmt_p): Handle non-SSA name
> > > arguments of phi nodes.
> > I'd suggest using
> >
> > if (TREE_CODE (expr) != SSA_NAME)
> > break;
> >
> > Instead since only an SSA_NAME is valid for EXPR in that loop. Why
> > check is_gimple_min_invariant when it's clear that the only valid
> > node is an SSA_NAME?
>
> sure. Ok with the change?
Yes, OK with that change.
jeff
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2004-07-12 13:13 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-07-11 23:34 [PATCH] for PR16461 Zdenek Dvorak
2004-07-12 19:27 ` Jeffrey A Law
2004-07-12 19:38 ` Zdenek Dvorak
2004-07-12 19:50 ` Jeffrey A Law
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).