[LLVMbugs] [Bug 20292] New: x86_stdcallcc: Assertion hit when passing more than 2^16 bytes on the stack

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=20292

            Bug ID: 20292
           Summary: x86_stdcallcc: Assertion hit when passing more than
                    2^16 bytes on the stack
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: X86
          Assignee: unassignedbugs at nondot.org
          Reporter: code at klickverbot.at
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified

Generating code for the following piece of IR currently asserts as
X86TargetLowering::LowerReturn tries to fit the number of bytes to pop on
function exit into a 16 bit constant:

---
target datalayout =
"e-p:32:32:32-S128-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f16:16:16-f32:32:32-f64:32:64-f128:128:128-v64:64:64-v128:128:128-a0:0:64-f80:32:32-n8:16:32"
target triple = "i386-unknown-linux-gnu"

%bar = type { [16385 x i32] }

define x86_stdcallcc void @foo(%bar* byval %s_arg) {
  ret void
}
---

Of course, passing such a large object on the stack is rather nonsensical, but
I think this is somewhat problematic as IR that passes the validator/… silently
leads to wrong code being generated ("ret $4") when LLVM is built without
assertions.

Is it worth supporting this, or should there simply be a fatal error in
X86TargetLowering::LowerReturn when encountering such large sizes?

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20140713/ff05408a/attachment.html>