[llvm-bugs] [Bug 47824] Clang generates wrong alignment on packed structs (looks like #5598 again)

via llvm-bugs llvm-bugs at lists.llvm.org
Tue Oct 13 13:06:41 PDT 2020


https://bugs.llvm.org/show_bug.cgi?id=47824

Richard Smith <richard-llvm at metafoo.co.uk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |INVALID
             Status|NEW                         |RESOLVED

--- Comment #1 from Richard Smith <richard-llvm at metafoo.co.uk> ---
The frontend behavior is correct. __attribute__((packed)) reduces the alignment
of the 'x' field of 'a' to 1. Extending your example a little:

struct __attribute__((packed)) a { int x; };
struct b { int x; };

void test(struct a *a, struct b *b)
{
    a->x = b->x;
}

struct c { int n; char x; struct a a; };
int main() {
  struct c c;
  struct b b;
  test(&c.a, &b);
}

... results in a->x not being 4-byte aligned in practice: struct c has 4 byte
alignment, and the offset of a.x in c is 5. So the "align 1" is necessary for
correctness.

Also, bug#5598 was that we did not properly mark accesses as "align 1" in this
case, not that we did incorrectly mark them.

-- 
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/20201013/6df7ea19/attachment.html>


More information about the llvm-bugs mailing list