This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [SFN] Bootstrap broken
- From: Alexandre Oliva <oliva at gnu dot org>
- To: Andreas Schwab <schwab at linux-m68k dot org>
- Cc: David Edelsohn <dje dot gcc at gmail dot com>, Rainer Orth <ro at cebitec dot uni-bielefeld dot de>, Jeffrey Law <law at redhat dot com>, Richard Biener <richard dot guenther at gmail dot com>, Jason Merrill <jason at redhat dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Fri, 15 Dec 2017 00:38:56 -0200
- Subject: Re: [SFN] Bootstrap broken
- Authentication-results: sourceware.org; auth=none
- References: <CAGWvny=4dmX11q5dBvy2SpAtRZYBygpb87JOwpzBZN6WS_qhDQ@mail.gmail.com> <yddo9n3n5ni.fsf@CeBiTec.Uni-Bielefeld.DE> <CAGWvnykHQmrxBU6rF=N+4MQskLdKW3vFqYwOQgqPo4jFrmLQvA@mail.gmail.com> <ormv2n6q7b.fsf@lxoliva.fsfla.org> <or1sjx2nde.fsf@lxoliva.fsfla.org> <ord13gn4gu.fsf@lxoliva.fsfla.org>
On Dec 14, 2017, Alexandre Oliva <oliva@gnu.org> wrote:
> On Dec 14, 2017, Alexandre Oliva <aoliva@redhat.com> wrote:
>> I'll arrange for markers to be moved past labels, even in gimple
> Here's a patch that implements this, and reverts all the changes I could
> find that had been introduced to support debug markers before labels and
> between BBs.
> I have *not* fully tested it yet
I reverted too much :-( see
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83396#c77 for details, or
just apply this patchlet on top of the earlier one (it reinstates some
of the SFN changes that turned out to be needed for more than accepting
markers before labels)
diff --git a/gcc/tree-cfg.c b/gcc/tree-cfg.c
index 4ec356d204d1..7fe0a1eec4c8 100644
--- a/gcc/tree-cfg.c
+++ b/gcc/tree-cfg.c
@@ -561,14 +561,22 @@ make_blocks_1 (gimple_seq seq, basic_block bb)
{
gimple_stmt_iterator i = gsi_start (seq);
gimple *stmt = NULL;
+ gimple *prev_stmt = NULL;
bool start_new_block = true;
bool first_stmt_of_seq = true;
while (!gsi_end_p (i))
{
- gimple *prev_stmt;
-
- prev_stmt = stmt;
+ /* PREV_STMT should only be set to a debug stmt if the debug
+ stmt is before nondebug stmts. Once stmt reaches a nondebug
+ nonlabel, prev_stmt will be set to it, so that
+ stmt_starts_bb_p will know to start a new block if a label is
+ found. However, if stmt was a label after debug stmts only,
+ keep the label in prev_stmt even if we find further debug
+ stmts, for there may be other labels after them, and they
+ should land in the same block. */
+ if (!prev_stmt || !stmt || !is_gimple_debug (stmt))
+ prev_stmt = stmt;
stmt = gsi_stmt (i);
if (stmt && is_gimple_call (stmt))
@@ -583,6 +591,7 @@ make_blocks_1 (gimple_seq seq, basic_block bb)
gsi_split_seq_before (&i, &seq);
bb = create_basic_block (seq, bb);
start_new_block = false;
+ prev_stmt = NULL;
}
/* Now add STMT to BB and create the subgraphs for special statement
@@ -2703,6 +2712,13 @@ stmt_starts_bb_p (gimple *stmt, gimple *prev_stmt)
if (stmt == NULL)
return false;
+ /* PREV_STMT is only set to a debug stmt if the debug stmt is before
+ any nondebug stmts in the block. We don't want to start another
+ block in this case: the debug stmt will already have started the
+ one STMT would start if we weren't outputting debug stmts. */
+ if (prev_stmt && is_gimple_debug (prev_stmt))
+ return false;
+
/* Labels start a new basic block only if the preceding statement
wasn't a label of the same type. This prevents the creation of
consecutive blocks that have nothing but a single label. */
--
Alexandre Oliva, freedom fighter http://FSFLA.org/~lxoliva/
You must be the change you wish to see in the world. -- Gandhi
Be Free! -- http://FSFLA.org/ FSF Latin America board member
Free Software Evangelist|Red Hat Brasil GNU Toolchain Engineer