This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
ARM testsuite failures
- From: Martin Michlmayr <tbm at cyrius dot com>
- To: gcc at gcc dot gnu dot org
- Cc: Paul Brook <paul at codesourcery dot com>, Richard Earnshaw <rearnsha at arm dot com>, fortran at gcc dot gnu dot org
- Date: Tue, 30 Oct 2007 09:28:34 +0100
- Subject: ARM testsuite failures
I compile gcc trunk from two days ago on ARM (OABI, v3, Debian) for
all languages and looked at some of the testsuite failures. I wanted
to ask whether people are aware of these issues and working on them
before I file PRs.
Testsuite results:
http://gcc.gnu.org/ml/gcc-testresults/2007-10/msg01361.html
gcc
---
A number of ICEs like this one:
| gcc.c-torture/execute/20050121-1.c:33 (set (reg:HI 0 r0)
| (subreg:HI (mem/c:CQI (plus:SI (reg/f:SI 11 fp)
| (const_int -24 [0xffffffe8])) [0 S2 A8]) 0)) -1 (nil))
| gcc.c-torture/execute/20050121-1.c:33: internal compiler error: in extract_insn, at recog.c:1990
Many neon testcases fail with:
| arm_neon.h: In function 'test_vdupQ_np16':
| arm_neon.h:5385: internal compiler error: in copy_to_mode_reg, at explow.c:644
None of them are reported. Paul/Richard, are you aware of those?
libffi
------
A number of "execution test" test failures. Then some output failures:
| FAIL: libffi.call/cls_3byte1.c -O0 -W -Wall output pattern test, is 1 15 58900 157: 58901 172
| FAIL: libffi.call/cls_3byte1.c -O2 output pattern test, is 1 15 34316 224: 34317 239
| FAIL: libffi.call/cls_2byte.c -O3 output pattern test, is 1 13 4 166: 5 179
| FAIL: libffi.call/cls_3byte1.c -O3 output pattern test, is 1 15 46604 137: 46605 152
| FAIL: libffi.call/cls_3byte1.c -Os output pattern test, is 1 15 58896 184: 58897 199
Who might be interested in taking a look?
gfortran
--------
gfortran has 7442 unexpected failures. Most of them are due to "test for
excess errors". Many are simply because of this:
| warning: 'const' attribute directive ignored
| warning: 'nothrow' attribute directive ignored
which seems to be mentioned in PR21185 (comment #20). Is that problem
still on the radar of the gfortran developers?
But then we also have e.g.:
| gfortran.dg/PR19754_1.f90:7.7-12:
| x = x + y ! { dg-error "Shapes for operands at" }
| 1 2
| Error: Shapes for operands at (1) and (2) are not conformable
for which I cannot find a PR. Also:
| access_spec_2.f90:9.13:
| public :: x ! { dg-error "was already specified" }
| 1
| Error: ACCESS specification at (1) was already specified
and:
| access_spec_2.f90:18.19:
| integer, public :: y ! { dg-error "Fortran 2003: Attribute PUBLIC" }
| 1
| Error: Fortran 2003: Attribute PUBLIC at (1) in a TYPE definition
Are these known (but unreported) issues or should I file PRs for them?
obj-c++
-------
There are a number of longstanding but known issue that show up on all or
most architectures:
ICE: internal compiler error: tree check: expected tree that contains 'decl
with RTL' structure, have 'field_decl' in assemble_external_real, at
varasm.c:2220
PR31032
ICE: tree check: expected class 'type', have 'exceptional' (error_mark)
in setup_one_parameter, at tree-inline.c:1483
PR23716
--
Martin Michlmayr
http://www.cyrius.com/