This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Another movdf_hardfloat32 patch
- To: David Edelsohn <dje at watson dot ibm dot com>
- Subject: Re: Another movdf_hardfloat32 patch
- From: Richard Henderson <rth at cygnus dot com>
- Date: Wed, 23 Jun 1999 15:31:14 -0700
- Cc: law at cygnus dot com, egcs-patches at egcs dot cygnus dot com
- References: <law@cygnus.com> <9906230415.AA47648@marc.watson.ibm.com> <19990623152744.A15074@cygnus.com>
On Wed, Jun 23, 1999 at 03:27:44PM -0700, Richard Henderson wrote:
> One of the values for `const' that the linker can choose is 32764.
> If you've emitted `@l(sym)+4' the instruction field will overflow
> and you'll get -32768 instead of what you expected.
Hmm. Actually, there is one mitigating factor to this -- DFmode must
be 8-byte aligned. So that means that 32764 is _not_ a value that the
linker can chose.
So while lo_sum is not generally offsettable, it is in this case
becuase of the known alignment of the data.
r~