#13 Assembly error in floating-point library

[Anonymous]

When compiling brickos-0.2.6.10.6, I am receiving the
following error in the floating-point library:

/usr/local/bin/h8300-hitachi-hms-as fixsfsi.s -o fixsfsi.o
fixsfsi.s: Assembler messages:
fixsfsi.s:195: Error: invalid operands
make[2]: *** [fixsfsi.o] Error 1

There are also some warnings earlier in the same
section. I have attached a complete log.

I am using GCC 3.3.2 and GNU binutils 2.14.90.0.8.

[Nobody/Anonymous]

Build log

[Andrew Keyser]

Logged In: YES
user_id=830119

I have the same problem; after I got the cross-compiler to
compile right (gcc 3.2.3), same error as reported above.

[Sřren Holm]

Logged In: YES
user_id=735254

I allso share the problem. Anyone know any good
assembly-tutorials out there, then I'll fix the problem myself.

I'm running gcc 3.3.4 and binutils 2.15

[Michael Edenfield]

Logged In: YES
user_id=67958

The following patch fixes this specific problem, but there's
more trouble later on. Still working on that one.

Index: brickos/lib/float/fixsfsi.s

RCS file: /cvsroot/brickos/brickos/lib/float/fixsfsi.s,v
retrieving revision 1.2
diff -u -r1.2 fixsfsi.s
--- brickos/lib/float/fixsfsi.s 16 Feb 2004 02:46:17 -0000
1.2
+++ brickos/lib/float/fixsfsi.s 18 Oct 2004 15:21:46 -0000
@@ -192,7 +192,7 @@

; Subtract bias + 31 from exponent

- add.b #-158,r4l ; subtract bias + 31 from
exponent
+ add.b #256-158,r4l ; subtract bias + 31
from exponent
addx.b #-1,r4h

; Is exponent < 0 ?

[Björn Lindqvist]

Logged In: YES
user_id=51702

I get the exact same error using gcc 3.4.3 and binutils 2.15

[Nobody/Anonymous]

Logged In: NO

any solution to this bug??

[Martin v. Löwis]

Logged In: YES
user_id=21627

The work-around is to use binutils 2.9.5.