diff options
| author | Jakob Stoklund Olesen <stoklund@2pi.dk> | 2012-01-10 01:34:59 +0000 |
|---|---|---|
| committer | Jakob Stoklund Olesen <stoklund@2pi.dk> | 2012-01-10 01:34:59 +0000 |
| commit | dae412bd320f1522bd1d850be0276212fca0331f (patch) | |
| tree | addc865b66b7088f8f71d7839c70155a7dd1f55e /test/Transforms/LoopStrengthReduce | |
| parent | 9cce24a257108ec50f65652d233b9ffadc691682 (diff) | |
| download | external_llvm-dae412bd320f1522bd1d850be0276212fca0331f.zip external_llvm-dae412bd320f1522bd1d850be0276212fca0331f.tar.gz external_llvm-dae412bd320f1522bd1d850be0276212fca0331f.tar.bz2 | |
Accurately model hardware alignment rounding.
On Thumb, the displacement computation hardware uses the address of the
current instruction rouned down to a multiple of 4. Include this
rounding in the UserOffset we compute for each instruction.
When inline asm is present, the instruction alignment may not be known.
Constrain the maximum displacement instead in that case.
This makes it possible for CreateNewWater() and OffsetIsInRange() to
agree about the valid displacements. When they disagree, infinite
looping happens.
As always, test cases for this stuff are insane.
<rdar://problem/10660175>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147825 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/Transforms/LoopStrengthReduce')
0 files changed, 0 insertions, 0 deletions