[Bug inline-asm/34833] New: rejects "i"(&var) with -fpic -m32

[Bug inline-asm/34833] New: rejects "i"(&var) with -fpic -m32

[stsp at users dot sourceforge dot net]

Hi.

Opening per suggestion in Bug 23200.
Not entirely sure if this is really a bug,
but it looks like the one.

The following program:
---
static char var;
int main()
{
    asm volatile ("" :: "i"(&var));
    return 0;
}
---

gives this when compiled with -fpic -m32:
---
asvar1.c: In function 'main':
asvar1.c:4: warning: asm operand 0 probably doesn't match constraints
asvar1.c:4: error: impossible constraint in 'asm'
---

Without -m32 it compiles.
On what targets the relocations are supported?


-- 
           Summary: rejects "i"(&var) with -fpic -m32
           Product: gcc
           Version: 4.1.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: inline-asm
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: stsp at users dot sourceforge dot net
  GCC host triplet: x86_64


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34833

[Bug inline-asm/34833] rejects "i"(&var) with -fpic -m32

[stsp at users dot sourceforge dot net]

------- Comment #1 from stsp at users dot sourceforge dot net  2008-01-17 18:39 -------
Created an attachment (id=14962)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14962&action=view)
test case


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34833

[Bug inline-asm/34833] rejects "i"(&var) with -fpic -m32

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2008-01-18 22:02 -------
I don't think this is a bug as &var is not a constant for PIC mode.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34833

[Bug inline-asm/34833] rejects "i"(&var) with -fpic -m32

[stsp at users dot sourceforge dot net]

------- Comment #3 from stsp at users dot sourceforge dot net  2008-01-19 08:25 -------
Yes, I know, but... without -m32 it compiles.
So either way it might be a bug - maybe it
should not compile without -m32 too? Why does
-m32 make the difference here?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34833

[Bug inline-asm/34833] rejects "i"(&var) with -fpic -m32

[pinskia at gcc dot gnu dot org]

------- Comment #4 from pinskia at gcc dot gnu dot org  2008-01-19 12:08 -------
Well with x86_64, PIC addressing global variables is simpler as you have access
to the rip register.  Try this on any other target, you will get a rejection.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34833

[Bug inline-asm/34833] rejects "i"(&var) with -fpic -m32

[stsp at users dot sourceforge dot net]

------- Comment #5 from stsp at users dot sourceforge dot net  2008-01-19 12:47 -------
OK, I understand, thank you.
Closing it then?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34833

[Bug inline-asm/34833] rejects "i"(&var) with -fpic -m32

[rth at gcc dot gnu dot org]

------- Comment #6 from rth at gcc dot gnu dot org  2008-01-31 19:06 -------
Yes, closing as not-a-bug for 32-bit i386.


-- 

rth at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34833