LAST_UPDATED: Fri Jan 9 01:21:29 UTC 2009 (revision 143198) $ ../trunk/configure --with-arch=native --disable-java-awt --without-x --enable-__cxa_atexit --disable-libgomp --disable-static --enable-languages=c,c++,java --disable-fixed-point --enable-checking=release --with-gmp=/home/daney/mp --with-mpfr=/home/daney/mp --disable-libmudflap --disable-multilib As noted in these test results: http://gcc.gnu.org/ml/gcc-testresults/2009-01/msg01307.html We have FAIL: gcc.target/mips/dpaq_sa_l_w.c (test for excess errors) Here is the log of the failure: Executing on host: /home/daney/gccsvn/trunk-build/gcc/xgcc -B/home/daney/gccsvn/trunk-build/gcc/ /home/daney/gccsvn/trunk/gcc/testsuite/\ gcc.target/mips/dpaq_sa_l_w.c -DNOMIPS16=__attribute__((nomips16)) -mabi=32 -mips32r2 -mgp32 -O2 -mdsp -S -o dpaq_sa_l_w.s (timeou\ t = 300) /home/daney/gccsvn/trunk/gcc/testsuite/gcc.target/mips/dpaq_sa_l_w.c:5: error: fixed-point types not supported for this target /home/daney/gccsvn/trunk/gcc/testsuite/gcc.target/mips/dpaq_sa_l_w.c:6: error: fixed-point types not supported for this target . . . The native CPU is an R5000 My untested hypothesis is that configuring with --with-arch=native has incorrect interaction with -m* from the gcc command line.
This is at most a documentation problem.
The fixed-point tests fail to compile when configure with --disable-fixed-point. I think the easiest would be to extend Richard's new infrastructure with a new keyword and then skip these tests completely. I take a look if this sounds good.
It was actually Richard's mips.exp rewrite that removed { target { fixed_point } } from these tests. Richard, was this intentional? It seems to me that since fixed-point is not dependent on a command-line flag (-mdsp is orthogonal) the effective-target check is correct here. Should I revert it like this: Index: fixed-vector-type.c =================================================================== --- fixed-vector-type.c (revision 143450) +++ fixed-vector-type.c (working copy) @@ -1,5 +1,5 @@ /* Test vector fixed-point instructions */ -/* { dg-do compile } */ +/* { dg-do compile { target { fixed_point } } } */ /* { dg-options "-mdspr2 -O2" } */ /* { dg-final { scan-assembler-times "\taddq_s.ph\t" 2 } } */ /* { dg-final { scan-assembler-times "\tsubq_s.ph\t" 2 } } */ Index: fixed-scalar-type.c =================================================================== --- fixed-scalar-type.c (revision 143450) +++ fixed-scalar-type.c (working copy) @@ -1,5 +1,5 @@ /* Test scalar fixed-point instructions */ -/* { dg-do compile } */ +/* { dg-do compile { target { fixed_point } } } */ /* { dg-options "-mdspr2 -O2" } */ /* { dg-final { scan-assembler-times "\taddu\t" 10 } } */ /* { dg-final { scan-assembler-times "\tsubu\t" 10 } } */ Index: dpaq_sa_l_w.c =================================================================== --- dpaq_sa_l_w.c (revision 143450) +++ dpaq_sa_l_w.c (working copy) @@ -1,4 +1,4 @@ -/* { dg-do compile } */ +/* { dg-do compile { target { fixed_point } } } */ /* { dg-options "-O2 -mgp32 -mdsp" } */ /* { dg-final { scan-assembler-times "\tdpaq_sa.l.w\t\\\$ac" 3 } } */ Index: dpsq_sa_l_w.c =================================================================== --- dpsq_sa_l_w.c (revision 143450) +++ dpsq_sa_l_w.c (working copy) @@ -1,4 +1,4 @@ -/* { dg-do compile } */ +/* { dg-do compile { target { fixed_point } } } */ /* { dg-options "-O2 -mgp32 -mdsp" } */ /* { dg-final { scan-assembler-times "\tdpsq_sa.l.w\t\\\$ac" 2 } } */
Subject: Re: Incorrect interaction between --with-arch=native and -mARCH "nemet at gcc dot gnu dot org" <gcc-bugzilla@gcc.gnu.org> writes: > ------- Comment #3 from nemet at gcc dot gnu dot org 2009-01-24 02:47 ------- > It was actually Richard's mips.exp rewrite that removed { target { fixed_point > } } from these tests. > > Richard, was this intentional? It seems to me that since fixed-point is not > dependent on a command-line flag (-mdsp is orthogonal) the effective-target > check is correct here. > > Should I revert it like this: Yeah. Sorry, I hadn't realised you could actually disable _cc1_ support for fixed point. (I knew you could disable libgcc support, but that's not a problem for these compile-only tests.) Richard
Adjust the summary.
Subject: Bug 38864 Author: nemet Date: Tue Jan 27 07:41:46 2009 New Revision: 143694 URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=143694 Log: PR testsuite/38864 * gcc.target/mips/fixed-vector-type.c: Add { target { fixed_point} } to dg-do compile. * gcc.target/mips/fixed-scalar-type.c: Likewise. * gcc.target/mips/dpaq_sa_l_w.c: Likewise. * gcc.target/mips/dpsq_sa_l_w.c: Likewise. Modified: trunk/gcc/testsuite/ChangeLog trunk/gcc/testsuite/gcc.target/mips/dpaq_sa_l_w.c trunk/gcc/testsuite/gcc.target/mips/dpsq_sa_l_w.c trunk/gcc/testsuite/gcc.target/mips/fixed-scalar-type.c trunk/gcc/testsuite/gcc.target/mips/fixed-vector-type.c
Fixed.