| Message ID | 5614CC24.7070706@st.com |
|---|---|
| State | New |
| Headers | show |
On 10/07/2015 01:39 AM, Christian Bruel wrote: > A regression test to check -fno-align-functions -O2 -mthumb > > thanks > > align4.patch > > > 2015-09-29 Christian Bruel<christian.bruel@st.com> > > PR target/67880 > * gcc.target/arm/no-align.c: Likewise. If this currently passes, then it's fine for the trunk. If it needs your patch for 67745, then include it as part of the work to fix 67745. jeff
On Wed, Oct 7, 2015 at 5:57 PM, Jeff Law <law@redhat.com> wrote: > On 10/07/2015 01:39 AM, Christian Bruel wrote: >> >> A regression test to check -fno-align-functions -O2 -mthumb >> >> thanks >> >> align4.patch >> >> >> 2015-09-29 Christian Bruel<christian.bruel@st.com> >> >> PR target/67880 >> * gcc.target/arm/no-align.c: Likewise. > > If this currently passes, then it's fine for the trunk. > > If it needs your patch for 67745, then include it as part of the work to fix > 67745. To my mind this really reads like a dup of PR67745. Ramana > > jeff >
On 10/07/2015 11:05 PM, Ramana Radhakrishnan wrote: > On Wed, Oct 7, 2015 at 5:57 PM, Jeff Law <law@redhat.com> wrote: >> On 10/07/2015 01:39 AM, Christian Bruel wrote: >>> >>> A regression test to check -fno-align-functions -O2 -mthumb >>> >>> thanks >>> >>> align4.patch >>> >>> >>> 2015-09-29 Christian Bruel<christian.bruel@st.com> >>> >>> PR target/67880 >>> * gcc.target/arm/no-align.c: Likewise. >> >> If this currently passes, then it's fine for the trunk. >> >> If it needs your patch for 67745, then include it as part of the work to fix >> 67745. > > To my mind this really reads like a dup of PR67745. Partially yes. I opened a different tracker because this test failed for a long time even without the use of attributes. It is fixed by the common subpatch (pr67745 1/2) and does not need (pr67745 2/2). Never mind, the important thing is that it's been in the testsuite for tracking. I'll can go together after PR67745 patches if it's OK. > > Ramana > >> >> jeff >>
On 10/08/2015 01:35 AM, Christian Bruel wrote: > > > On 10/07/2015 11:05 PM, Ramana Radhakrishnan wrote: >> On Wed, Oct 7, 2015 at 5:57 PM, Jeff Law <law@redhat.com> wrote: >>> On 10/07/2015 01:39 AM, Christian Bruel wrote: >>>> >>>> A regression test to check -fno-align-functions -O2 -mthumb >>>> >>>> thanks >>>> >>>> align4.patch >>>> >>>> >>>> 2015-09-29 Christian Bruel<christian.bruel@st.com> >>>> >>>> PR target/67880 >>>> * gcc.target/arm/no-align.c: Likewise. >>> >>> If this currently passes, then it's fine for the trunk. >>> >>> If it needs your patch for 67745, then include it as part of the work >>> to fix >>> 67745. >> >> To my mind this really reads like a dup of PR67745. > > Partially yes. I opened a different tracker because this test failed for > a long time even without the use of attributes. It is fixed by the > common subpatch (pr67745 1/2) and does not need (pr67745 2/2). > > Never mind, the important thing is that it's been in the testsuite for > tracking. I'll can go together after PR67745 patches if it's OK. Yea, let's bundle it in the 67745 patches. jeff
2015-09-29 Christian Bruel <christian.bruel@st.com> PR target/67880 * gcc.target/arm/no-align.c: Likewise. Index: gcc.target/arm/no-align.c =================================================================== --- gcc.target/arm/no-align.c (revision 0) +++ gcc.target/arm/no-align.c (working copy) @@ -0,0 +1,12 @@ +/* PR target/67745 + Verify that -mthumb code is not aligned. */ +/* { dg-do compile } */ +/* { dg-options "-O2 -mthumb -fno-align-functions" } */ +/* { dg-skip-if "" { ! { arm_thumb1_ok || arm_thumb2_ok } } } */ + +void +foo() +{ +} + +/* { dg-final { scan-assembler-not ".align\[ \t]2" } } */