On Tue, 11 May 2021, Maxim Kuvyrkov wrote:
Hi Richard,
Your patch a076632e274abe344ca7648b7c7f299273d4cbe0 appears to have broken bootstrap-O3 for 32-bit armhf. Do you have an AArch32-capable machine to reproduce/investigate this on? Let me know if not, and I'll make a proper bug report with a testcase.
There is certainly the possibility of more fallout - please make a proper bugreport with a testcase. The cp/call.c hunk in provides a hint for a fix - the issue is that go..C479 (wherever that's created) is not TREE_ADDRESSABLE but we take its address, probably in some (global) initializer CTOR.
Richard.
ICE: 00:33:32 In function ‘syscall.forkExec’: 00:33:32 go1: error: address taken, but ADDRESSABLE bit not set 00:33:32 PHI argument 00:33:32 &go..C479; 00:33:32 for PHI node 00:33:32 err$__object_78 = PHI <err$__object_76(58), &go..C479(59)> 00:33:32 during GIMPLE pass: fre 00:33:32 go1: internal compiler error: verify_ssa failed 00:33:32 0x9c18d7 verify_ssa(bool, bool) 00:33:32 /home/tcwg-buildslave/workspace/tcwg_gnu_0/abe/snapshots/gcc.git~master/gcc/tree-ssa.c:1214 00:33:32 0x6f8d5b execute_function_todo 00:33:32 /home/tcwg-buildslave/workspace/tcwg_gnu_0/abe/snapshots/gcc.git~master/gcc/passes.c:2049 00:33:32 0x6f9abf do_per_function 00:33:32 /home/tcwg-buildslave/workspace/tcwg_gnu_0/abe/snapshots/gcc.git~master/gcc/passes.c:1687 00:33:32 0x6f9abf execute_todo 00:33:32 /home/tcwg-buildslave/workspace/tcwg_gnu_0/abe/snapshots/gcc.git~master/gcc/passes.c:2096 00:33:32 Please submit a full bug report, 00:33:32 with preprocessed source if appropriate. 00:33:32 Please include the complete backtrace with any bug report. 00:33:32 See https://gcc.gnu.org/bugs/ for instructions. 00:33:32 Makefile:3001: recipe for target 'syscall.lo' failed
Full build log: https://ci.linaro.org/job/tcwg_gcc-bisect-gnu-master-arm-bootstrap_O3/16/art...
Regards,
-- Maxim Kuvyrkov https://www.linaro.org
On May 11, 2021, at 8:15 AM, tcwg-jira (Jira) projects@linaro.org wrote:
There is 1 comment. GNU Toolchain / <Mail Attachment.png> GNU-692 IN PROGRESS Regressions from tcwg_binutils/tcwg_cross/tcwg_gnu CI View issue · Add comment 1 comment tcwg-jira on 11/May/21 5:04 AM Successfully identified regression in gcc in CI configuration tcwg_gnu/gnu-master-arm-bootstrap_O3. So far, this commit has regressed CI configurations: • tcwg_gnu/gnu-master-arm-bootstrap_O3 Culprit:
<cut> commit a076632e274abe344ca7648b7c7f299273d4cbe0 Author: Richard Biener <rguenther@suse.de> Date: Fri May 7 09:51:18 2021 +0200 middle-end/100464 - avoid spurious TREE_ADDRESSABLE in folding debug stmts canonicalize_constructor_val was setting TREE_ADDRESSABLE on bases of ADDR_EXPRs but that's futile when we're dealing with CTOR values in debug stmts. This rips out the code which was added for Java and should have been an assertion when we didn't have debug stmts. To not regress g++.dg/tree-ssa/array-temp1.C we have to adjust the testcase to not look for a no longer applied invalid optimization. 2021-05-10 Richard Biener <rguenther@suse.de> PR middle-end/100464 PR c++/100468 gcc/ • gimple-fold.c (canonicalize_constructor_val): Do not set TREE_ADDRESSABLE. gcc/cp/ • call.c (set_up_extended_ref_temp): Mark the temporary addressable if the TARGET_EXPR was. gcc/testsuite/ • gcc.dg/pr100464.c: New testcase. • g++.dg/tree-ssa/array-temp1.C: Adjust. </cut> Details: https://ci.linaro.org/job/tcwg_gcc-bisect-gnu-master-arm-bootstrap_O3/16/artifact/artifacts/jenkins/mail-body.txt/*view*/ Even more details: https://ci.linaro.org/job/tcwg_gcc-bisect-gnu-master-arm-bootstrap_O3/16/artifact/artifacts/ This message was sent by Atlassian Jira (v8.11.1#811002-sha1:94cd716) Jira is improving email notifications, share your feedback! Get Jira notifications on your phone! Download the Jira Server app for Android or iOS.