|
|
| Bugzilla Link |
1576 |
| Resolution |
FIXED |
| Resolved on |
Mar 25, 2008 16:27 |
| Version |
trunk |
| OS |
All |
| Reporter |
LLVM Bugzilla Contributor |
Extended Description
Probably just need a new addPassesToEmitFile() parameter that allows llvm-gcc to informs codegen to optimize for size. There are potentially some minor codegen changes that can save a bit of code size. e.g. X86 backend. Instead of aligning functions on 16-byte boundaries, it can just align on 4-byte boundaries.