From: uweigand@de.ibm.com
To: gdb-patches@sourceware.org
Subject: [rfc][28/37] Eliminate builtin_type_ macros: Update infcall.c routines
Date: Sun, 31 Aug 2008 18:15:00 -0000 [thread overview]
Message-ID: <20080831175135.489007000@de.ibm.com> (raw)
In-Reply-To: <20080831175045.128504000@de.ibm.com>
[-- Attachment #1: diff-type-infcall --]
[-- Type: text/plain, Size: 4737 bytes --]
Hello,
this patch removes builtin_type_ macros from infcall.c. The instances
in value_arg_coerce are simply made per-arch by passing in the architecture
from the (sole) call site.
find_function_addr used to return a default "builtin_type_int" if the
actual function return type could not be detected from debug data.
However: most callers to not even look at the return type computed
by find_function_addr, and the call site in eval.c implements its own
unknown return type handling. Thus the patch moves the return type
default setting to the sole remaining caller, call_function_by_hand.
Bye,
Ulrich
ChangeLog:
* infcall.c (value_arg_coerce): Add GDBARCH parameter. Use its
associates types instead of builtin_type_ macros.
(find_function_addr): Leave output VALUE_TYPE NULL if unknown.
(call_function_by_hand): Use per-architecture "int" type as
fall-back if find_function_addr returns NULL VALUE_TYPE.
Update call to value_arg_coerce.
Index: gdb-head/gdb/infcall.c
===================================================================
--- gdb-head.orig/gdb/infcall.c
+++ gdb-head/gdb/infcall.c
@@ -100,9 +100,10 @@ Unwinding of stack if a signal is receiv
its value as needed). */
static struct value *
-value_arg_coerce (struct value *arg, struct type *param_type,
- int is_prototyped, CORE_ADDR *sp)
+value_arg_coerce (struct gdbarch *gdbarch, struct value *arg,
+ struct type *param_type, int is_prototyped, CORE_ADDR *sp)
{
+ const struct builtin_type *builtin = builtin_type (gdbarch);
struct type *arg_type = check_typedef (value_type (arg));
struct type *type
= param_type ? check_typedef (param_type) : arg_type;
@@ -141,22 +142,22 @@ value_arg_coerce (struct value *arg, str
/* If we don't have a prototype, coerce to integer type if necessary. */
if (!is_prototyped)
{
- if (TYPE_LENGTH (type) < TYPE_LENGTH (builtin_type_int))
- type = builtin_type_int;
+ if (TYPE_LENGTH (type) < TYPE_LENGTH (builtin->builtin_int))
+ type = builtin->builtin_int;
}
/* Currently all target ABIs require at least the width of an integer
type for an argument. We may have to conditionalize the following
type coercion for future targets. */
- if (TYPE_LENGTH (type) < TYPE_LENGTH (builtin_type_int))
- type = builtin_type_int;
+ if (TYPE_LENGTH (type) < TYPE_LENGTH (builtin->builtin_int))
+ type = builtin->builtin_int;
break;
case TYPE_CODE_FLT:
if (!is_prototyped && coerce_float_to_double_p)
{
- if (TYPE_LENGTH (type) < TYPE_LENGTH (builtin_type_double))
- type = builtin_type_double;
- else if (TYPE_LENGTH (type) > TYPE_LENGTH (builtin_type_double))
- type = builtin_type_long_double;
+ if (TYPE_LENGTH (type) < TYPE_LENGTH (builtin->builtin_double))
+ type = builtin->builtin_double;
+ else if (TYPE_LENGTH (type) > TYPE_LENGTH (builtin->builtin_double))
+ type = builtin->builtin_long_double;
}
break;
case TYPE_CODE_FUNC:
@@ -199,7 +200,7 @@ find_function_addr (struct value *functi
{
struct type *ftype = check_typedef (value_type (function));
enum type_code code = TYPE_CODE (ftype);
- struct type *value_type;
+ struct type *value_type = NULL;
CORE_ADDR funaddr;
/* If it's a member function, just look at the function
@@ -223,8 +224,6 @@ find_function_addr (struct value *functi
¤t_target);
value_type = TYPE_TARGET_TYPE (ftype);
}
- else
- value_type = builtin_type_int;
}
else if (code == TYPE_CODE_INT)
{
@@ -251,8 +250,6 @@ find_function_addr (struct value *functi
/* Handle integer used as address of a function. */
funaddr = (CORE_ADDR) value_as_long (function);
}
-
- value_type = builtin_type_int;
}
else
error (_("Invalid data type for function to be called."));
@@ -472,6 +469,9 @@ call_function_by_hand (struct value *fun
funaddr = find_function_addr (function, &values_type);
CHECK_TYPEDEF (values_type);
+ if (!values_type)
+ values_type = builtin_type (gdbarch)->builtin_int;
+
/* Are we returning a value using a structure return (passing a
hidden argument pointing to storage) or a normal value return?
There are two cases: language-mandated structure return and
@@ -590,7 +590,8 @@ call_function_by_hand (struct value *fun
else
param_type = NULL;
- args[i] = value_arg_coerce (args[i], param_type, prototyped, &sp);
+ args[i] = value_arg_coerce (gdbarch, args[i],
+ param_type, prototyped, &sp);
if (param_type != NULL && language_pass_by_reference (param_type))
args[i] = value_addr (args[i]);
--
Dr. Ulrich Weigand
GNU Toolchain for Linux on System z and Cell BE
Ulrich.Weigand@de.ibm.com
next prev parent reply other threads:[~2008-08-31 18:15 UTC|newest]
Thread overview: 98+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-08-31 17:53 [rfc][00/37] Eliminate builtin_type_ macros uweigand
2008-08-31 17:52 ` [rfc][29/37] Eliminate builtin_type_ macros: Update valarith.c routines uweigand
2008-08-31 17:52 ` [rfc][09/37] Eliminate builtin_type_ macros: Make argument promotion explicit uweigand
2008-08-31 17:52 ` [rfc][21/37] Eliminate builtin_type_ macros: Platform-neutral builtin_type_void uweigand
2008-09-06 0:38 ` Joel Brobecker
2008-09-06 4:12 ` Daniel Jacobowitz
2008-09-06 14:00 ` Joel Brobecker
2008-09-07 15:59 ` Ulrich Weigand
2008-09-13 15:23 ` Daniel Jacobowitz
2008-09-13 17:23 ` Joel Brobecker
2008-08-31 17:52 ` [rfc][23/37] Eliminate builtin_type_ macros: Platform-neutral types for internal variables uweigand
2008-09-02 12:43 ` Daniel Jacobowitz
2008-09-02 21:55 ` Ulrich Weigand
2008-09-02 22:00 ` Daniel Jacobowitz
2008-09-02 23:53 ` Ulrich Weigand
2008-08-31 17:52 ` [rfc][13/37] Eliminate builtin_type_ macros: Update EVAL_SKIP dummy return type uweigand
2008-09-05 22:56 ` Joel Brobecker
2008-09-07 15:40 ` Ulrich Weigand
2008-09-07 15:49 ` Joel Brobecker
2008-08-31 17:52 ` [rfc][15/37] Eliminate builtin_type_ macros: Dereferencing of integer types uweigand
2008-09-01 7:19 ` Tom Tromey
2008-09-02 23:34 ` Ulrich Weigand
2008-09-05 23:02 ` Joel Brobecker
2008-08-31 17:52 ` [rfc][07/37] Eliminate builtin_type_ macros: Use expression arch for size_t type uweigand
2008-09-05 18:18 ` Joel Brobecker
2008-09-05 20:16 ` Ulrich Weigand
2008-08-31 17:52 ` [rfc][19/37] Eliminate builtin_type_ macros: Ada range type handling uweigand
2008-09-06 0:24 ` Joel Brobecker
2008-09-07 15:43 ` Ulrich Weigand
2008-09-09 18:00 ` Joel Brobecker
2008-09-09 20:21 ` Ulrich Weigand
2008-09-09 22:08 ` Joel Brobecker
2008-09-09 22:32 ` Ulrich Weigand
2008-09-10 6:09 ` Joel Brobecker
2008-09-10 9:51 ` Ulrich Weigand
2008-08-31 17:52 ` [rfc][26/37] Eliminate builtin_type_ macros: Use per-frame architecture uweigand
2008-08-31 17:52 ` [rfc][05/37] Eliminate builtin_type_ macros: Replace LA_BOOL_TYPE macro uweigand
2008-09-05 18:08 ` Joel Brobecker
2008-08-31 17:52 ` [rfc][16/37] Eliminate builtin_type_ macros: Ada fixed/double conversions uweigand
2008-09-05 23:13 ` Joel Brobecker
2008-08-31 17:52 ` [rfc][32/37] Eliminate builtin_type_ macros: Update value-printing code uweigand
2008-08-31 17:52 ` [rfc][30/37] Eliminate builtin_type_ macros: Remove gdbarch_name_of_malloc uweigand
2008-08-31 17:52 ` [rfc][02/37] Eliminate builtin_type_ macros: Introduce expression architecture uweigand
2008-08-31 17:52 ` [rfc][33/37] Eliminate builtin_type_ macros: Default target word size uweigand
2008-08-31 17:52 ` [rfc][11/37] Eliminate builtin_type_ macros: Update ax-gdb expression evaluator uweigand
2008-08-31 17:52 ` [rfc][01/37] Eliminate builtin_type_ macros: Unused write_exp_msymbol parameters uweigand
2008-08-31 17:52 ` [rfc][24/37] Eliminate builtin_type_ macros: Platform-neutral generic integers uweigand
2008-09-06 3:15 ` Joel Brobecker
2008-09-07 16:44 ` Ulrich Weigand
2008-08-31 17:53 ` [rfc][36/37] Eliminate builtin_type_ macros: Use target arch in solib code uweigand
2008-08-31 17:53 ` [rfc][17/37] Eliminate builtin_type_ macros: Ada pos_atr result type uweigand
2008-08-31 17:53 ` [rfc][12/37] Eliminate builtin_type_ macros: Remove redundant coerce_enum/coerce_number uweigand
2008-08-31 17:53 ` [rfc][22/37] Eliminate builtin_type_ macros: Platform-neutral "true char" types uweigand
2008-08-31 17:53 ` [rfc][35/37] Eliminate builtin_type_ macros: Use target arch in bsd-uthread.c uweigand
2008-08-31 17:53 ` [rfc][20/37] Eliminate builtin_type_ macros: Objective-C expression evaluation uweigand
2008-08-31 17:53 ` [rfc][08/37] Eliminate builtin_type_ macros: Make pointer arithmetic explicit uweigand
2008-09-02 12:38 ` Daniel Jacobowitz
2008-09-02 21:48 ` Ulrich Weigand
2008-09-02 21:52 ` Daniel Jacobowitz
2008-09-04 22:32 ` Tom Tromey
2008-09-05 18:21 ` Joel Brobecker
2008-08-31 17:53 ` [rfc][04/37] Eliminate builtin_type_ macros: Introduce java_language_arch_info uweigand
2008-08-31 17:53 ` [rfc][27/37] Eliminate builtin_type_ macros: Update C++ ABI handling uweigand
2008-09-05 20:18 ` Ulrich Weigand
2008-08-31 17:53 ` [rfc][10/37] Eliminate builtin_type_ macros: Use expression arch for argument promotion uweigand
2008-09-05 22:39 ` Joel Brobecker
2008-08-31 17:53 ` [rfc][37/37] Eliminate builtin_type_ macros: Delete the macros uweigand
2008-08-31 17:53 ` [rfc][06/37] Eliminate builtin_type_ macros: Make OP_COMPLEX type explicit uweigand
2008-08-31 17:53 ` [rfc][18/37] Eliminate builtin_type_ macros: Ada System.Address special handling uweigand
2008-08-31 17:53 ` [rfc][03/37] Eliminate builtin_type_ macros: Extract bitstring subscript handling uweigand
2008-09-05 18:16 ` Joel Brobecker
2008-09-05 20:17 ` Ulrich Weigand
2008-08-31 17:53 ` [rfc][14/37] Eliminate builtin_type_ macros: Implicit dereferencing of references uweigand
2008-08-31 18:12 ` [rfc][34/37] Eliminate builtin_type_ macros: Use target arch in procfs.c uweigand
2008-08-31 18:13 ` [rfc][31/37] Eliminate builtin_type_ macros: Inferior call argument types uweigand
2008-09-06 1:37 ` Joel Brobecker
2008-08-31 18:15 ` uweigand [this message]
2008-09-02 12:48 ` [rfc][28/37] Eliminate builtin_type_ macros: Update infcall.c routines Daniel Jacobowitz
2008-09-02 21:56 ` Ulrich Weigand
2008-08-31 18:16 ` [rfc][25/37] Eliminate builtin_type_ macros: Update *-tdep.c files uweigand
2008-08-31 22:20 ` [rfc][00/37] Eliminate builtin_type_ macros Mark Kettenis
2008-09-01 3:46 ` David Miller
2008-09-01 18:57 ` Ulrich Weigand
2008-09-02 10:22 ` Mark Kettenis
2008-09-02 12:30 ` Daniel Jacobowitz
2008-09-02 21:37 ` Ulrich Weigand
2008-09-02 12:50 ` Daniel Jacobowitz
2008-09-02 22:02 ` Ulrich Weigand
2008-09-02 22:12 ` Daniel Jacobowitz
2008-09-06 3:16 ` Joel Brobecker
2008-09-07 16:43 ` Ulrich Weigand
2008-09-09 18:05 ` Joel Brobecker
2008-09-09 20:21 ` Ulrich Weigand
2008-09-09 21:18 ` Joel Brobecker
2008-09-09 22:12 ` Ulrich Weigand
2008-09-10 6:18 ` Joel Brobecker
2008-09-10 9:43 ` Ulrich Weigand
2008-09-10 16:25 ` Joel Brobecker
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20080831175135.489007000@de.ibm.com \
--to=uweigand@de.ibm.com \
--cc=gdb-patches@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox