Correct gcc m33 floating point handling #6146
Conversation
|
@cmonr why is this needs: work? because no description provided? I did not spot Travis failures, so that would explain it |
|
@0xc0170 Marked as needs: work because the PR is still not stable in Travis. |
theotherjimmy
force-pushed
the
fix-m33-fp
branch
2 times, most recently
from
bc691c6 to
5dd00f3
Compare
|
|
| elif target.core.startswith("Cortex-M33F"): | ||
| self.cpu = ["-mcpu=cortex-m33"] | ||
| elif target.core.startswith("Cortex-M33"): | ||
| self.cpu = ["-march=armv8-m.main"] |
There was a problem hiding this comment.
Interesting. Why is this not the same as line 60?
There was a problem hiding this comment.
That's actually the bug this fixes. -mcpu=cortex-m33 turns on DSP and FP instructions, which are not required.
|
Still not stable in Travis. Assuming that rebuild is stable, will launch CI. |
theotherjimmy
force-pushed
the
fix-m33-fp
branch
from
5dd00f3 to
c29207a
Compare
|
@cmonr Travis build stabilized. Did I answer your question? |
|
/morph build |
|
Dang it. Just remembered I should not have done this. |
|
Stopped the build for now. Will restart once a window opens. |
Build : FAILUREBuild number : 1315 |
|
/morph build |
Build : SUCCESSBuild number : 1339 Triggering tests/morph test |
Exporter Build : SUCCESSBuild number : 995 |
Test : SUCCESSBuild number : 1122 |
No description provided.