aboutsummaryrefslogtreecommitdiff
path: root/gdb/config/sparc/tm-sparc.h
diff options
context:
space:
mode:
Diffstat (limited to 'gdb/config/sparc/tm-sparc.h')
-rw-r--r--gdb/config/sparc/tm-sparc.h12
1 files changed, 4 insertions, 8 deletions
diff --git a/gdb/config/sparc/tm-sparc.h b/gdb/config/sparc/tm-sparc.h
index 9dcc7a4..d4acb9a 100644
--- a/gdb/config/sparc/tm-sparc.h
+++ b/gdb/config/sparc/tm-sparc.h
@@ -25,18 +25,14 @@ Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
#define IEEE_FLOAT
/* When passing a structure to a function, Sun cc passes the address
- in a register, not the structure itself. It (under SunOS4) creates
- two symbols, so we get a LOC_ARG saying the address is on the stack
- (a lie, and a serious one since we don't know which register to
- use), and a LOC_REGISTER saying that the struct is in a register
- (sort of a lie, but fixable with REG_STRUCT_HAS_ADDR). Gcc version
- two (as of 1.92) behaves like sun cc. REG_STRUCT_HAS_ADDR is smart
- enough to distinguish between Sun cc, gcc version 1 and gcc version 2.
+ not the structure itself. It (under SunOS4) creates two symbols,
+ which we need to combine to a LOC_REGPARM. Gcc version two (as of
+ 1.92) behaves like sun cc. REG_STRUCT_HAS_ADDR is smart enough to
+ distinguish between Sun cc, gcc version 1 and gcc version 2.
This still doesn't work if the argument is not one passed in a
register (i.e. it's the 7th or later argument). */
#define REG_STRUCT_HAS_ADDR(gcc_p) (gcc_p != 1)
-#define STRUCT_ARG_SYM_GARBAGE(gcc_p) (gcc_p != 1)
/* If Pcc says that a parameter is a short, it's a short. This is
because the parameter does get passed in in a register as an int,