From patchwork Wed Mar 9 10:07:21 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Richard Biener X-Patchwork-Id: 51812 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 9367A3858423 for ; Wed, 9 Mar 2022 10:07:56 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 9367A3858423 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1646820476; bh=HJRuxAw/UceLuhQcvEzd99GbF4TZt0V8Qz8WZ4cnRLs=; h=Date:To:Subject:List-Id:List-Unsubscribe:List-Archive:List-Post: List-Help:List-Subscribe:From:Reply-To:Cc:From; b=EMp5MlPoFhZTtXtFwPpPrie+fKRPfPgOf1yVSM4GNCILSE245Olzo3Rn6yPhwaanK mfhIRVFLTB82of4tQIXC7VBOP20g9sHOVaGyIW+CAA+NJjGRZAYDcLQNfNLBWZ+66i +ht2Wjkszy3eTkPM/8mNWAY1+QAWNeucesKXBLyM= X-Original-To: gcc-patches@gcc.gnu.org Delivered-To: gcc-patches@gcc.gnu.org Received: from smtp-out2.suse.de (smtp-out2.suse.de [195.135.220.29]) by sourceware.org (Postfix) with ESMTPS id 9C88E385803F for ; Wed, 9 Mar 2022 10:07:27 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 9C88E385803F Received: from imap2.suse-dmz.suse.de (imap2.suse-dmz.suse.de [192.168.254.74]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (P-521) server-digest SHA512) (No client certificate requested) by smtp-out2.suse.de (Postfix) with ESMTPS id 60E431F380; Wed, 9 Mar 2022 10:07:22 +0000 (UTC) Received: from imap2.suse-dmz.suse.de (imap2.suse-dmz.suse.de [192.168.254.74]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (P-521) server-digest SHA512) (No client certificate requested) by imap2.suse-dmz.suse.de (Postfix) with ESMTPS id 473B113D79; Wed, 9 Mar 2022 10:07:22 +0000 (UTC) Received: from dovecot-director2.suse.de ([192.168.254.65]) by imap2.suse-dmz.suse.de with ESMTPSA id WXc6EFp8KGLGPAAAMHmgww (envelope-from ); Wed, 09 Mar 2022 10:07:22 +0000 Date: Wed, 9 Mar 2022 11:07:21 +0100 (CET) To: gcc-patches@gcc.gnu.org Subject: [PATCH] middle-end/104786 - ICE with asm and VLA MIME-Version: 1.0 Message-Id: <20220309100722.473B113D79@imap2.suse-dmz.suse.de> X-Spam-Status: No, score=-11.8 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, SPF_HELO_NONE, SPF_PASS, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: gcc-patches@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Richard Biener via Gcc-patches From: Richard Biener Reply-To: Richard Biener Cc: Jakub Jelinek Errors-To: gcc-patches-bounces+patchwork=sourceware.org@gcc.gnu.org Sender: "Gcc-patches" The following fixes an ICE observed with a MEM_REF allows_mem asm operand. There's code expecting INDIRECT_REFs that are now never going to appear. The following simply treats all tcc_reference operands the same. A more conservative approach would be to use INDIRECT_REF || MEM_REF which would fix this particular testcase, but all non-register typed kinds of reference trees could appear here and I'm not sure why former INDIRECT_REFs should be OK to go this path but not say a.b as we also handle plain DECL_P here albeit with some extra constraints. For the VLA case in the PR another spot to adjust would be || TREE_ADDRESSABLE (type) noting that we also cannot create a temporary for not constant-sized types. That said - INDIRECT_REF catched my eye here, this might have caused quite some extra copies eventually so I did want to fix that. Btw, it seems to me the DECL_P case disallowing promoted variables should only apply to register types which should be SSA names here unless TREE_ADDRESSABLE in which case we let them through anyway. With the same reasoning a generic REFERENCE_CLASS_P should be OK, not needing such special-casing. Bootstrap and regtest running on x86_64-unknown-linux-gnu. OK for trunk? Thanks, Richard. 2022-03-09 Richard Biener PR middle-end/104786 * cfgexpand.cc (expand_asm_stmt): Specialize REFERENCE_CLASS_P operands rather than INDIRECT_REFs. * gcc.dg/pr104786.c: New testcase. --- gcc/cfgexpand.cc | 2 +- gcc/testsuite/gcc.dg/pr104786.c | 8 ++++++++ 2 files changed, 9 insertions(+), 1 deletion(-) create mode 100644 gcc/testsuite/gcc.dg/pr104786.c diff --git a/gcc/cfgexpand.cc b/gcc/cfgexpand.cc index 87536ec7ccd..aaccf1a0906 100644 --- a/gcc/cfgexpand.cc +++ b/gcc/cfgexpand.cc @@ -3290,7 +3290,7 @@ expand_asm_stmt (gasm *stmt) generating_concat_p = 0; - if ((TREE_CODE (val) == INDIRECT_REF && allows_mem) + if ((REFERENCE_CLASS_P (val) && allows_mem) || (DECL_P (val) && (allows_mem || REG_P (DECL_RTL (val))) && ! (REG_P (DECL_RTL (val)) diff --git a/gcc/testsuite/gcc.dg/pr104786.c b/gcc/testsuite/gcc.dg/pr104786.c new file mode 100644 index 00000000000..3076d236d21 --- /dev/null +++ b/gcc/testsuite/gcc.dg/pr104786.c @@ -0,0 +1,8 @@ +/* { dg-do compile } */ +/* { dg-options "-std=gnu90" } */ + +void h(void *di, int num) +{ + char (*t)[num] = di; + __asm__ ("" : "=X"( *t)); +}