#4211: LLVM: Stack alignment on OSX
--------------------------------+-------------------------------------------
Reporter: dterei | Owner: dterei
Type: task | Status: new
Priority: normal | Milestone: 7.4.1
Component: Compiler (LLVM) | Version: 6.13
Keywords: | Testcase:
Blockedby: | Difficulty:
Os: MacOS X | Blocking:
Architecture: x86 | Failure: None/Unknown
--------------------------------+-------------------------------------------
Changes (by dmp):
* cc: dmp@… (added)
Comment:
I have made the changes to keep the stack 16+8 byte aligned throughout
STG code for x86_64 targets. The changes are available in my github
repo on the 4211 branch
https://github.com/dmpots/ghc/tree/4211
I can also attach the patches to this ticket if that would work better.
The results look promising. I get no new regressions on a validate run
and a testsuite run with `WAY=optllvm` goes from 36 failures to 6
failures, plus two unexpected passes (one of which is the 5250 test
simonmar mentioned above).
I would love to see these changes merged to the mainline GHC. I
suppose that the main sticking point is probably that I made the
change for 64-bit targets, but not the 32-bit targets. Is there any
chance it will get merged for the 64-bit targets? My motivation (and
therefore availble time) is much less for 32-bit targets and I'm
worried that dealing with the `stdcall` convetion could be a bit more
compilicated.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4211#comment:11>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
