diff mbox

[review] Add scalar_storage_order support for floating point

Message ID gerrit.1574880479000.Ic02eb711d80ce678ef0ecf8c506a626e441b8440@gnutoolchain-gerrit.osci.io
State New, archived
Headers

Commit Message

Simon Marchi (Code Review) Nov. 27, 2019, 6:48 p.m. UTC 
  Change URL: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/727
......................................................................

Add scalar_storage_order support for floating point

Testing the scalar_storage_order patch pointed out that it does not
handle floating point properly.  This patch fixes this problem.

gdb/ChangeLog
2019-11-27  Tom Tromey  <tromey@adacore.com>

	* dwarf2read.c (dwarf2_init_float_type)
	(dwarf2_init_complex_target_type): Add byte_order parameter.
	(read_base_type): Compute byte order earlier.
	* gdbtypes.c (init_float_type): Add byte_order parameter.
	* gdbtypes.h (init_float_type): Add byte_order parameter.

gdb/testsuite/ChangeLog
2019-11-27  Tom Tromey  <tromey@adacore.com>

	* gdb.base/endianity.c (struct otherendian) <f>: New field.
	(main): Initialize it.
	* gdb.base/endianity.exp: Update.

Change-Id: Ic02eb711d80ce678ef0ecf8c506a626e441b8440
---
M gdb/ChangeLog
M gdb/dwarf2read.c
M gdb/gdbtypes.c
M gdb/gdbtypes.h
M gdb/testsuite/ChangeLog
M gdb/testsuite/gdb.base/endianity.c
M gdb/testsuite/gdb.base/endianity.exp
7 files changed, 52 insertions(+), 24 deletions(-)

Comments

Simon Marchi (Code Review) Nov. 27, 2019, 10:02 p.m. UTC | #1 
Simon Marchi has posted comments on this change.

Change URL: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/727
......................................................................


Patch Set 1:

(4 comments)

| --- gdb/dwarf2read.c
| +++ gdb/dwarf2read.c
| @@ -17681,8 +17682,19 @@ read_base_type (struct die_info *die, struct dwarf2_cu *cu)
|    arch = get_objfile_arch (objfile);
| +  enum bfd_endian byte_order = gdbarch_byte_order (arch);
| +  switch (endianity)
| +    {
| +      case DW_END_big:
| +	byte_order = BFD_ENDIAN_BIG;
| +        break;
| +      case DW_END_little:
| +	byte_order = BFD_ENDIAN_LITTLE;
| +        break;

PS1, Line 17691:

Not really related to this patch, since this is pre-exising code, but
what do you think of complaining if `endianity` has another value than
_big or _little?

| +    }
| +
|    switch (encoding)
|      {
|        case DW_ATE_address:
|  	/* Turn DW_ATE_address into a void * pointer.  */
|  	type = init_type (objfile, TYPE_CODE_VOID, TARGET_CHAR_BIT, NULL);
|  	type = init_pointer_type (objfile, bits, name, type);
|  	break;
| --- gdb/gdbtypes.c
| +++ gdb/gdbtypes.c
| @@ -2898,18 +2898,15 @@ init_boolean_type (struct objfile *objfile,
|    if (unsigned_p)
|      TYPE_UNSIGNED (t) = 1;
|  
|    return t;
|  }
|  
|  /* Allocate a TYPE_CODE_FLT type structure associated with OBJFILE.
|     BIT is the type size in bits; if BIT equals -1, the size is
|     determined by the floatformat.  NAME is the type name.  Set the
|     TYPE_FLOATFORMAT from FLOATFORMATS.  */

PS1, Line 2907:

Could you document the new parameter (especially the behavior for
_UNKNOWN)?

|  
|  struct type *
|  init_float_type (struct objfile *objfile,
|  		 int bit, const char *name,
| -		 const struct floatformat **floatformats)
| -{
| -  struct gdbarch *gdbarch = get_objfile_arch (objfile);
| -  const struct floatformat *fmt = floatformats[gdbarch_byte_order (gdbarch)];
| +		 const struct floatformat **floatformats,
| --- gdb/testsuite/gdb.base/endianity.c
| +++ gdb/testsuite/gdb.base/endianity.c
| @@ -15,17 +15,19 @@ /* This testcase is part of GDB, the GNU debugger.
|     You should have received a copy of the GNU General Public License
|     along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
|  
|  /* This tests the handling of dwarf attributes:
|      DW_AT_endianity, DW_END_big, and DW_END_little.  */
|  struct otherendian
|  {
|    int v;
|    short w;
| +  float f;

PS1, Line 24:

I wonder if we should test double too, or if it would be superfluous?

| +  __complex__ float cplx;
|  }
|  #if defined __GNUC__ && (__GNUC__ >= 6)
|  #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
|  __attribute__( ( scalar_storage_order( "big-endian" ) ) )
|  #else
|  __attribute__( ( scalar_storage_order( "little-endian" ) ) )
|  #endif
|  #endif
| --- gdb/testsuite/gdb.base/endianity.exp
| +++ gdb/testsuite/gdb.base/endianity.exp
| @@ -25,16 +25,17 @@ if ![runto "endianity.c:$bp_location" ] then {
|    return -1
|  }
|  
| -gdb_test "print o" "= {v = 3, w = 2}" "print o before assignment"
| +gdb_test "print o" "= {v = 3, w = 2, f = 23.5, cplx = 1.25 \\+ 7.25 \\* I}" \
| +    "print o before assignment"
|  
|  gdb_test "print o.v = 4" "= 4"
|  gdb_test "print o.w = 3" "= 3"
| +gdb_test "print o.f = 1.5" "= 1.5"

PS1, Line 33:

Just wondering why we don't set cplx, does gdb not offer a way to set
complex?

|  
|  # scalar_storage_order requires gcc >= 6
|  if { ([test_compiler_info {gcc-[0-5]-*}] || ![test_compiler_info gcc*]) } {
|    setup_xfail "*-*-*"
|  }
|  gdb_test "x/x &o.v" "0x04000000"
|  gdb_test "x/xh &o.w" "0x0300"
|  
| -gdb_test "print o" "= {v = 4, w = 3}" "print o after assignment"
Simon Marchi (Code Review) Dec. 3, 2019, 9:30 p.m. UTC | #2 
Tom Tromey has posted comments on this change.

Change URL: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/727
......................................................................


Patch Set 1:

(1 comment)

| --- gdb/testsuite/gdb.base/endianity.exp
| +++ gdb/testsuite/gdb.base/endianity.exp
| @@ -25,16 +25,17 @@ if ![runto "endianity.c:$bp_location" ] then {
|    return -1
|  }
|  
| -gdb_test "print o" "= {v = 3, w = 2}" "print o before assignment"
| +gdb_test "print o" "= {v = 3, w = 2, f = 23.5, cplx = 1.25 \\+ 7.25 \\* I}" \
| +    "print o before assignment"
|  
|  gdb_test "print o.v = 4" "= 4"
|  gdb_test "print o.w = 3" "= 3"
| +gdb_test "print o.f = 1.5" "= 1.5"

PS1, Line 33:

> Just wondering why we don't set cplx, does gdb not offer a way to set complex?

I don't think there's a way.  I went through the GCC node about
complex
types, and basically gdb doesn't implement any of it, and in fact
the way gdb print complex numbers for C is not a way that GCC can
read.

|  
|  # scalar_storage_order requires gcc >= 6
|  if { ([test_compiler_info {gcc-[0-5]-*}] || ![test_compiler_info gcc*]) } {
|    setup_xfail "*-*-*"
|  }
|  gdb_test "x/x &o.v" "0x04000000"
|  gdb_test "x/xh &o.w" "0x0300"
|  
| -gdb_test "print o" "= {v = 4, w = 3}" "print o after assignment"
Simon Marchi (Code Review) Dec. 4, 2019, 3:33 p.m. UTC | #3 
Simon Marchi has posted comments on this change.

Change URL: https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/727
......................................................................


Patch Set 2: Code-Review+2
diff mbox

Patch

diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index f1ed42a..66afaea 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,5 +1,13 @@ 
 2019-11-27  Tom Tromey  <tromey@adacore.com>
 
+	* dwarf2read.c (dwarf2_init_float_type)
+	(dwarf2_init_complex_target_type): Add byte_order parameter.
+	(read_base_type): Compute byte order earlier.
+	* gdbtypes.c (init_float_type): Add byte_order parameter.
+	* gdbtypes.h (init_float_type): Add byte_order parameter.
+
+2019-11-27  Tom Tromey  <tromey@adacore.com>
+
 	* dwarf2read.h (struct dwarf2_per_objfile): Remove unnecessary
 	backslashes.
 	* cp-support.c: Remove unnecessary backslashes.
diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index 1ca801c..cc815a2 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -17552,7 +17552,7 @@ 
 
 static struct type *
 dwarf2_init_float_type (struct objfile *objfile, int bits, const char *name,
-			const char *name_hint)
+			const char *name_hint, enum bfd_endian byte_order)
 {
   struct gdbarch *gdbarch = get_objfile_arch (objfile);
   const struct floatformat **format;
@@ -17560,7 +17560,7 @@ 
 
   format = gdbarch_floatformat_for_type (gdbarch, name_hint, bits);
   if (format)
-    type = init_float_type (objfile, bits, name, format);
+    type = init_float_type (objfile, bits, name, format, byte_order);
   else
     type = init_type (objfile, TYPE_CODE_ERROR, bits, name);
 
@@ -17599,7 +17599,8 @@ 
 static struct type *
 dwarf2_init_complex_target_type (struct dwarf2_cu *cu,
 				 struct objfile *objfile,
-				 int bits, const char *name_hint)
+				 int bits, const char *name_hint,
+				 enum bfd_endian byte_order)
 {
   gdbarch *gdbarch = get_objfile_arch (objfile);
   struct type *tt = nullptr;
@@ -17648,7 +17649,7 @@ 
     tt = nullptr;
 
   const char *name = (tt == nullptr) ? nullptr : TYPE_NAME (tt);
-  return dwarf2_init_float_type (objfile, bits, name, name_hint);
+  return dwarf2_init_float_type (objfile, bits, name, name_hint, byte_order);
 }
 
 /* Find a representation of a given base type and install
@@ -17679,6 +17680,17 @@ 
     endianity = DW_UNSND (attr);
 
   arch = get_objfile_arch (objfile);
+  enum bfd_endian byte_order = gdbarch_byte_order (arch);
+  switch (endianity)
+    {
+      case DW_END_big:
+	byte_order = BFD_ENDIAN_BIG;
+        break;
+      case DW_END_little:
+	byte_order = BFD_ENDIAN_LITTLE;
+        break;
+    }
+
   switch (encoding)
     {
       case DW_ATE_address:
@@ -17690,14 +17702,15 @@ 
 	type = init_boolean_type (objfile, bits, 1, name);
 	break;
       case DW_ATE_complex_float:
-	type = dwarf2_init_complex_target_type (cu, objfile, bits / 2, name);
+	type = dwarf2_init_complex_target_type (cu, objfile, bits / 2, name,
+						byte_order);
 	type = init_complex_type (objfile, name, type);
 	break;
       case DW_ATE_decimal_float:
 	type = init_decfloat_type (objfile, bits, name);
 	break;
       case DW_ATE_float:
-	type = dwarf2_init_float_type (objfile, bits, name, name);
+	type = dwarf2_init_float_type (objfile, bits, name, name, byte_order);
 	break;
       case DW_ATE_signed:
 	type = dwarf2_init_integer_type (cu, objfile, bits, 0, name);
@@ -17755,17 +17768,7 @@ 
 
   maybe_set_alignment (cu, die, type);
 
-  switch (endianity)
-    {
-      case DW_END_big:
-        if (gdbarch_byte_order (arch) == BFD_ENDIAN_LITTLE)
-          TYPE_ENDIANITY_NOT_DEFAULT (type) = 1;
-        break;
-      case DW_END_little:
-        if (gdbarch_byte_order (arch) == BFD_ENDIAN_BIG)
-          TYPE_ENDIANITY_NOT_DEFAULT (type) = 1;
-        break;
-    }
+  TYPE_ENDIANITY_NOT_DEFAULT (type) = gdbarch_byte_order (arch) != byte_order;
 
   return set_die_type (die, type, cu);
 }
diff --git a/gdb/gdbtypes.c b/gdb/gdbtypes.c
index 31c1a7b..ff89009 100644
--- a/gdb/gdbtypes.c
+++ b/gdb/gdbtypes.c
@@ -2909,10 +2909,15 @@ 
 struct type *
 init_float_type (struct objfile *objfile,
 		 int bit, const char *name,
-		 const struct floatformat **floatformats)
+		 const struct floatformat **floatformats,
+		 enum bfd_endian byte_order)
 {
-  struct gdbarch *gdbarch = get_objfile_arch (objfile);
-  const struct floatformat *fmt = floatformats[gdbarch_byte_order (gdbarch)];
+  if (byte_order == BFD_ENDIAN_UNKNOWN)
+    {
+      struct gdbarch *gdbarch = get_objfile_arch (objfile);
+      byte_order = gdbarch_byte_order (gdbarch);
+    }
+  const struct floatformat *fmt = floatformats[byte_order];
   struct type *t;
 
   bit = verify_floatformat (bit, fmt);
diff --git a/gdb/gdbtypes.h b/gdb/gdbtypes.h
index 8fc770c..dc4cb3e 100644
--- a/gdb/gdbtypes.h
+++ b/gdb/gdbtypes.h
@@ -1804,7 +1804,8 @@ 
 extern struct type *init_boolean_type (struct objfile *, int, int,
 				       const char *);
 extern struct type *init_float_type (struct objfile *, int, const char *,
-				     const struct floatformat **);
+				     const struct floatformat **,
+				     enum bfd_endian = BFD_ENDIAN_UNKNOWN);
 extern struct type *init_decfloat_type (struct objfile *, int, const char *);
 extern struct type *init_complex_type (struct objfile *, const char *,
 				       struct type *);
diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index 93e94f8..9931534 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,3 +1,9 @@ 
+2019-11-27  Tom Tromey  <tromey@adacore.com>
+
+	* gdb.base/endianity.c (struct otherendian) <f>: New field.
+	(main): Initialize it.
+	* gdb.base/endianity.exp: Update.
+
 2019-11-27  Andrew Burgess  <andrew.burgess@embecosm.com>
 
 	* gdb.mi/mi-fortran-modules-2.f90: New file.
diff --git a/gdb/testsuite/gdb.base/endianity.c b/gdb/testsuite/gdb.base/endianity.c
index 5ab090d..57378c2 100644
--- a/gdb/testsuite/gdb.base/endianity.c
+++ b/gdb/testsuite/gdb.base/endianity.c
@@ -21,6 +21,8 @@ 
 {
   int v;
   short w;
+  float f;
+  __complex__ float cplx;
 }
 #if defined __GNUC__ && (__GNUC__ >= 6)
 #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
@@ -39,7 +41,7 @@ 
 int
 main (void)
 {
-  struct otherendian o = {3,2};
+  struct otherendian o = {3, 2, 23.5, 1.25 + 7.25i};
 
   do_nothing (&o); /* START */
 }
diff --git a/gdb/testsuite/gdb.base/endianity.exp b/gdb/testsuite/gdb.base/endianity.exp
index 80df012..a6ddea3 100644
--- a/gdb/testsuite/gdb.base/endianity.exp
+++ b/gdb/testsuite/gdb.base/endianity.exp
@@ -25,10 +25,12 @@ 
   return -1
 }
 
-gdb_test "print o" "= {v = 3, w = 2}" "print o before assignment"
+gdb_test "print o" "= {v = 3, w = 2, f = 23.5, cplx = 1.25 \\+ 7.25 \\* I}" \
+    "print o before assignment"
 
 gdb_test "print o.v = 4" "= 4"
 gdb_test "print o.w = 3" "= 3"
+gdb_test "print o.f = 1.5" "= 1.5"
 
 # scalar_storage_order requires gcc >= 6
 if { ([test_compiler_info {gcc-[0-5]-*}] || ![test_compiler_info gcc*]) } {
@@ -37,4 +39,5 @@ 
 gdb_test "x/x &o.v" "0x04000000"
 gdb_test "x/xh &o.w" "0x0300"
 
-gdb_test "print o" "= {v = 4, w = 3}" "print o after assignment"
+gdb_test "print o" "= {v = 4, w = 3, f = 1.5, cplx = 1.25 \\+ 7.25 \\* I}" \
+    "print o after assignment"