[PATCH] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[PATCH] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Andrew Burgess]

This patch addresses some regressions that Sergio identified in this mail:
   https://sourceware.org/ml/gdb-patches/2019-05/msg00060.html

It turns out these regressions are all related to bug PR gdb/18644
which I've resolved in this patch.  I'm posting this as a new thread
rather than a reply for more visibility.

Thanks,
Andrew

---

PR gdb/18644 is caused by GDB using the wrong floating point format
for gfortran's 16-byte floating point type, including when the 16-byte
float is used as the component of a 32-byte complex type.

This commit addresses the issue in two places, first in i386-tdep.c
there already exists some code to force the use of
floatformats_ia64_quad for specific named types.  This code is
extended to include the type names that gfortran uses for its 16-byte
floats.

Second, the builtin 16-byte float type (in f-lang.c) is changed to use
floatformats_ia64_quad instead of the long double type from gdbarch.
On some targets (for example i386) long double is not 16-bytes, and so
is inappropriate.

This patch was tested on X86-64/GNU-Linux using '--target_board=unix'
and '--target_board=unix/-m32', and resolves all of the known failures
associated with PR gdb/18644.  I've also added the test case from the
original bug report.

gdb/ChangeLog:

	PR gdb/18644:
	* f-lang.c (build_fortran_types): Use floatformats_ia64_quad for
	16-byte floats.
	* i386-tdep.c (i386_floatformat_for_type): Use
	floatformats_ia64_quad for the 16-byte floating point component
	within a fortran 32-byte complex number.

gdb/testsuite/ChangeLog:

	PR gdb/18644
	* gdb.fortran/complex.exp: Remove setup_kfail calls.
	* gdb.fortran/printing-types.exp: Add new test.
	* gdb.fortran/printing-types.f90: Add 16-byte real variable for
	testing.
	* gdb.fortran/type-kinds.exp (test_cast_1_to_type_kind): Remove
	setup_kfail call.
---
 gdb/ChangeLog                                |  9 +++++++++
 gdb/f-lang.c                                 |  3 +--
 gdb/i386-tdep.c                              |  4 +++-
 gdb/testsuite/ChangeLog                      | 10 ++++++++++
 gdb/testsuite/gdb.fortran/complex.exp        |  2 --
 gdb/testsuite/gdb.fortran/printing-types.exp |  1 +
 gdb/testsuite/gdb.fortran/printing-types.f90 |  2 ++
 gdb/testsuite/gdb.fortran/type-kinds.exp     |  6 ------
 8 files changed, 26 insertions(+), 11 deletions(-)

diff --git a/gdb/f-lang.c b/gdb/f-lang.c
index 9da6fdb3e1c..d7b322f759d 100644
--- a/gdb/f-lang.c
+++ b/gdb/f-lang.c
@@ -728,8 +728,7 @@ build_fortran_types (struct gdbarch *gdbarch)
     = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
 		       "real*8", gdbarch_double_format (gdbarch));
   builtin_f_type->builtin_real_s16
-    = arch_float_type (gdbarch, gdbarch_long_double_bit (gdbarch),
-		       "real*16", gdbarch_long_double_format (gdbarch));
+    = arch_float_type (gdbarch, 128, "real*16", floatformats_ia64_quad);
 
   builtin_f_type->builtin_complex_s8
     = arch_complex_type (gdbarch, "complex*8",
diff --git a/gdb/i386-tdep.c b/gdb/i386-tdep.c
index 54d9dd873b8..66fc6d679e2 100644
--- a/gdb/i386-tdep.c
+++ b/gdb/i386-tdep.c
@@ -8158,7 +8158,9 @@ i386_floatformat_for_type (struct gdbarch *gdbarch,
   if (len == 128 && name)
     if (strcmp (name, "__float128") == 0
 	|| strcmp (name, "_Float128") == 0
-	|| strcmp (name, "complex _Float128") == 0)
+	|| strcmp (name, "complex _Float128") == 0
+	|| strcmp (name, "complex(kind=16)") == 0
+	|| strcmp (name, "real(kind=16)") == 0)
       return floatformats_ia64_quad;
 
   return default_floatformat_for_type (gdbarch, name, len);
diff --git a/gdb/testsuite/gdb.fortran/complex.exp b/gdb/testsuite/gdb.fortran/complex.exp
index 136f1c4df79..94ac53afc70 100644
--- a/gdb/testsuite/gdb.fortran/complex.exp
+++ b/gdb/testsuite/gdb.fortran/complex.exp
@@ -33,7 +33,6 @@ gdb_test "print c4" " = \\(1000,-50\\)"
 gdb_test "print c8" " = \\(321,-22\\)"
 gdb_test "print dc" " = \\(321,-22\\)"
 
-setup_kfail gdb/18644 "*-*-*"
 gdb_test "print c16" " = \\(-874,19\\)"
 
 gdb_test "whatis c" "type = complex\\(kind=4\\)"
@@ -53,7 +52,6 @@ gdb_test "print \$_creal (dc)" " = 321"
 gdb_test "whatis \$" " = real\\*8"
 
 gdb_test "whatis c16" "type = complex\\(kind=16\\)"
-setup_kfail gdb/18644 "*-*-*"
 gdb_test "print \$_creal (c16)" " = -874"
 gdb_test "whatis \$" " = real\\*16"
 
diff --git a/gdb/testsuite/gdb.fortran/printing-types.exp b/gdb/testsuite/gdb.fortran/printing-types.exp
index 2f6be4ec249..6394e45f4c3 100644
--- a/gdb/testsuite/gdb.fortran/printing-types.exp
+++ b/gdb/testsuite/gdb.fortran/printing-types.exp
@@ -33,3 +33,4 @@ gdb_test "print oneByte"	" = 1"
 gdb_test "print twobytes"	" = 2"
 gdb_test "print chvalue"	" = \'a\'"
 gdb_test "print logvalue"	" = \.TRUE\."
+gdb_test "print rVal"		" = 2000"
diff --git a/gdb/testsuite/gdb.fortran/printing-types.f90 b/gdb/testsuite/gdb.fortran/printing-types.f90
index b4ff928604f..36b63532c8e 100644
--- a/gdb/testsuite/gdb.fortran/printing-types.f90
+++ b/gdb/testsuite/gdb.fortran/printing-types.f90
@@ -18,10 +18,12 @@ program prog
   integer(2) :: twoBytes
   character  :: chValue
   logical(1) :: logValue
+  real(kind=16) :: rVal
 
   oneByte  = 1
   twoBytes = 2
   chValue  = 'a'
   logValue = .true.
+  rVal = 2000
   write(*,*) s
 end
diff --git a/gdb/testsuite/gdb.fortran/type-kinds.exp b/gdb/testsuite/gdb.fortran/type-kinds.exp
index 1ae15b96f1a..9d19a9ceb39 100644
--- a/gdb/testsuite/gdb.fortran/type-kinds.exp
+++ b/gdb/testsuite/gdb.fortran/type-kinds.exp
@@ -27,12 +27,6 @@ if { [skip_fortran_tests] } { continue }
 proc test_cast_1_to_type_kind {base_type type_kind cast_result size_result} {
     set type_string "$base_type (kind=$type_kind)"
     gdb_test "p (($type_string) 1)" " = $cast_result"
-
-    if {($base_type == "real" || $base_type == "complex")
-	&& $type_kind == 16} {
-	setup_kfail gdb/18644 "*-*-*"
-    }
-
     gdb_test "p sizeof (($type_string) 1)" " = $size_result"
 }
 
-- 
2.14.5

[PATCHv2] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Andrew Burgess]

This new version moves the use of floatformats_ia64_quad out of
f-lang.c, which I think is an improvement.  With this done I plan to
push this patch shortly unless anyone complains.

Thanks,
Andrew


---

PR gdb/18644 is caused by GDB using the wrong floating point format
for gfortran's 16-byte floating point type, including when the 16-byte
float is used as the component of a 32-byte complex type.

This commit addresses the issue in two places, first in i386-tdep.c,
there is already some code to force the use of floatformats_ia64_quad
for specific named types, this is extended to include the type names
that gfortran uses for its 16-byte floats.

Second, the builtin 16-byte float type (in f-lang.c) is changed so it
no longer uses gdbarch_long_double_format.  On i386 this type is not
16-bytes, but is smaller, this is not what gfortran is expecting.
Instead we now use gdbarch_floatformat_for_type and ask for a
16-byte (128 bit) type using the common gfortran type name.  This is
then spotted in i386-tdep.c (thanks to the first change above) and we
again get floatformats_ia64_quad returned.

This patch was tested on X86-64/GNU-Linux using '--target_board=unix'
and '--target_board=unix/-m32', and resolves all of the known failures
associated with PR gdb/18644.  I've also added the test case from the
original bug report.

gdb/ChangeLog:

	PR gdb/18644:
	* f-lang.c (build_fortran_types): Use floatformats_ia64_quad for
	16-byte floats.
	* i386-tdep.c (i386_floatformat_for_type): Use
	floatformats_ia64_quad for the 16-byte floating point component
	within a fortran 32-byte complex number.

gdb/testsuite/ChangeLog:

	PR gdb/18644
	* gdb.fortran/complex.exp: Remove setup_kfail calls.
	* gdb.fortran/printing-types.exp: Add new test.
	* gdb.fortran/printing-types.f90: Add 16-byte real variable for
	testing.
	* gdb.fortran/type-kinds.exp (test_cast_1_to_type_kind): Remove
	setup_kfail call.
---
 gdb/ChangeLog                                |  9 +++++++++
 gdb/f-lang.c                                 |  4 ++--
 gdb/i386-tdep.c                              |  4 +++-
 gdb/testsuite/ChangeLog                      | 10 ++++++++++
 gdb/testsuite/gdb.fortran/complex.exp        |  2 --
 gdb/testsuite/gdb.fortran/printing-types.exp |  1 +
 gdb/testsuite/gdb.fortran/printing-types.f90 |  2 ++
 gdb/testsuite/gdb.fortran/type-kinds.exp     |  6 ------
 8 files changed, 27 insertions(+), 11 deletions(-)

diff --git a/gdb/f-lang.c b/gdb/f-lang.c
index 9da6fdb3e1c..5855c68b38c 100644
--- a/gdb/f-lang.c
+++ b/gdb/f-lang.c
@@ -727,9 +727,9 @@ build_fortran_types (struct gdbarch *gdbarch)
   builtin_f_type->builtin_real_s8
     = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
 		       "real*8", gdbarch_double_format (gdbarch));
+  auto fmt = gdbarch_floatformat_for_type (gdbarch, "real(kind=16)", 128);
   builtin_f_type->builtin_real_s16
-    = arch_float_type (gdbarch, gdbarch_long_double_bit (gdbarch),
-		       "real*16", gdbarch_long_double_format (gdbarch));
+    = arch_float_type (gdbarch, 128, "real*16", fmt);
 
   builtin_f_type->builtin_complex_s8
     = arch_complex_type (gdbarch, "complex*8",
diff --git a/gdb/i386-tdep.c b/gdb/i386-tdep.c
index 54d9dd873b8..66fc6d679e2 100644
--- a/gdb/i386-tdep.c
+++ b/gdb/i386-tdep.c
@@ -8158,7 +8158,9 @@ i386_floatformat_for_type (struct gdbarch *gdbarch,
   if (len == 128 && name)
     if (strcmp (name, "__float128") == 0
 	|| strcmp (name, "_Float128") == 0
-	|| strcmp (name, "complex _Float128") == 0)
+	|| strcmp (name, "complex _Float128") == 0
+	|| strcmp (name, "complex(kind=16)") == 0
+	|| strcmp (name, "real(kind=16)") == 0)
       return floatformats_ia64_quad;
 
   return default_floatformat_for_type (gdbarch, name, len);
diff --git a/gdb/testsuite/gdb.fortran/complex.exp b/gdb/testsuite/gdb.fortran/complex.exp
index 136f1c4df79..94ac53afc70 100644
--- a/gdb/testsuite/gdb.fortran/complex.exp
+++ b/gdb/testsuite/gdb.fortran/complex.exp
@@ -33,7 +33,6 @@ gdb_test "print c4" " = \\(1000,-50\\)"
 gdb_test "print c8" " = \\(321,-22\\)"
 gdb_test "print dc" " = \\(321,-22\\)"
 
-setup_kfail gdb/18644 "*-*-*"
 gdb_test "print c16" " = \\(-874,19\\)"
 
 gdb_test "whatis c" "type = complex\\(kind=4\\)"
@@ -53,7 +52,6 @@ gdb_test "print \$_creal (dc)" " = 321"
 gdb_test "whatis \$" " = real\\*8"
 
 gdb_test "whatis c16" "type = complex\\(kind=16\\)"
-setup_kfail gdb/18644 "*-*-*"
 gdb_test "print \$_creal (c16)" " = -874"
 gdb_test "whatis \$" " = real\\*16"
 
diff --git a/gdb/testsuite/gdb.fortran/printing-types.exp b/gdb/testsuite/gdb.fortran/printing-types.exp
index 2f6be4ec249..6394e45f4c3 100644
--- a/gdb/testsuite/gdb.fortran/printing-types.exp
+++ b/gdb/testsuite/gdb.fortran/printing-types.exp
@@ -33,3 +33,4 @@ gdb_test "print oneByte"	" = 1"
 gdb_test "print twobytes"	" = 2"
 gdb_test "print chvalue"	" = \'a\'"
 gdb_test "print logvalue"	" = \.TRUE\."
+gdb_test "print rVal"		" = 2000"
diff --git a/gdb/testsuite/gdb.fortran/printing-types.f90 b/gdb/testsuite/gdb.fortran/printing-types.f90
index b4ff928604f..36b63532c8e 100644
--- a/gdb/testsuite/gdb.fortran/printing-types.f90
+++ b/gdb/testsuite/gdb.fortran/printing-types.f90
@@ -18,10 +18,12 @@ program prog
   integer(2) :: twoBytes
   character  :: chValue
   logical(1) :: logValue
+  real(kind=16) :: rVal
 
   oneByte  = 1
   twoBytes = 2
   chValue  = 'a'
   logValue = .true.
+  rVal = 2000
   write(*,*) s
 end
diff --git a/gdb/testsuite/gdb.fortran/type-kinds.exp b/gdb/testsuite/gdb.fortran/type-kinds.exp
index 1ae15b96f1a..9d19a9ceb39 100644
--- a/gdb/testsuite/gdb.fortran/type-kinds.exp
+++ b/gdb/testsuite/gdb.fortran/type-kinds.exp
@@ -27,12 +27,6 @@ if { [skip_fortran_tests] } { continue }
 proc test_cast_1_to_type_kind {base_type type_kind cast_result size_result} {
     set type_string "$base_type (kind=$type_kind)"
     gdb_test "p (($type_string) 1)" " = $cast_result"
-
-    if {($base_type == "real" || $base_type == "complex")
-	&& $type_kind == 16} {
-	setup_kfail gdb/18644 "*-*-*"
-    }
-
     gdb_test "p sizeof (($type_string) 1)" " = $size_result"
 }
 
-- 
2.14.5

Re: [PATCHv2] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Andrew Burgess]

* Andrew Burgess <andrew.burgess@embecosm.com> [2019-05-16 17:00:58 +0100]:

> This new version moves the use of floatformats_ia64_quad out of
> f-lang.c, which I think is an improvement.  With this done I plan to
> push this patch shortly unless anyone complains.

This has now been pushed.

Thanks,
Andrew



> 
> Thanks,
> Andrew
> 
> 
> ---
> 
> PR gdb/18644 is caused by GDB using the wrong floating point format
> for gfortran's 16-byte floating point type, including when the 16-byte
> float is used as the component of a 32-byte complex type.
> 
> This commit addresses the issue in two places, first in i386-tdep.c,
> there is already some code to force the use of floatformats_ia64_quad
> for specific named types, this is extended to include the type names
> that gfortran uses for its 16-byte floats.
> 
> Second, the builtin 16-byte float type (in f-lang.c) is changed so it
> no longer uses gdbarch_long_double_format.  On i386 this type is not
> 16-bytes, but is smaller, this is not what gfortran is expecting.
> Instead we now use gdbarch_floatformat_for_type and ask for a
> 16-byte (128 bit) type using the common gfortran type name.  This is
> then spotted in i386-tdep.c (thanks to the first change above) and we
> again get floatformats_ia64_quad returned.
> 
> This patch was tested on X86-64/GNU-Linux using '--target_board=unix'
> and '--target_board=unix/-m32', and resolves all of the known failures
> associated with PR gdb/18644.  I've also added the test case from the
> original bug report.
> 
> gdb/ChangeLog:
> 
> 	PR gdb/18644:
> 	* f-lang.c (build_fortran_types): Use floatformats_ia64_quad for
> 	16-byte floats.
> 	* i386-tdep.c (i386_floatformat_for_type): Use
> 	floatformats_ia64_quad for the 16-byte floating point component
> 	within a fortran 32-byte complex number.
> 
> gdb/testsuite/ChangeLog:
> 
> 	PR gdb/18644
> 	* gdb.fortran/complex.exp: Remove setup_kfail calls.
> 	* gdb.fortran/printing-types.exp: Add new test.
> 	* gdb.fortran/printing-types.f90: Add 16-byte real variable for
> 	testing.
> 	* gdb.fortran/type-kinds.exp (test_cast_1_to_type_kind): Remove
> 	setup_kfail call.
> ---
>  gdb/ChangeLog                                |  9 +++++++++
>  gdb/f-lang.c                                 |  4 ++--
>  gdb/i386-tdep.c                              |  4 +++-
>  gdb/testsuite/ChangeLog                      | 10 ++++++++++
>  gdb/testsuite/gdb.fortran/complex.exp        |  2 --
>  gdb/testsuite/gdb.fortran/printing-types.exp |  1 +
>  gdb/testsuite/gdb.fortran/printing-types.f90 |  2 ++
>  gdb/testsuite/gdb.fortran/type-kinds.exp     |  6 ------
>  8 files changed, 27 insertions(+), 11 deletions(-)
> 
> diff --git a/gdb/f-lang.c b/gdb/f-lang.c
> index 9da6fdb3e1c..5855c68b38c 100644
> --- a/gdb/f-lang.c
> +++ b/gdb/f-lang.c
> @@ -727,9 +727,9 @@ build_fortran_types (struct gdbarch *gdbarch)
>    builtin_f_type->builtin_real_s8
>      = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
>  		       "real*8", gdbarch_double_format (gdbarch));
> +  auto fmt = gdbarch_floatformat_for_type (gdbarch, "real(kind=16)", 128);
>    builtin_f_type->builtin_real_s16
> -    = arch_float_type (gdbarch, gdbarch_long_double_bit (gdbarch),
> -		       "real*16", gdbarch_long_double_format (gdbarch));
> +    = arch_float_type (gdbarch, 128, "real*16", fmt);
>  
>    builtin_f_type->builtin_complex_s8
>      = arch_complex_type (gdbarch, "complex*8",
> diff --git a/gdb/i386-tdep.c b/gdb/i386-tdep.c
> index 54d9dd873b8..66fc6d679e2 100644
> --- a/gdb/i386-tdep.c
> +++ b/gdb/i386-tdep.c
> @@ -8158,7 +8158,9 @@ i386_floatformat_for_type (struct gdbarch *gdbarch,
>    if (len == 128 && name)
>      if (strcmp (name, "__float128") == 0
>  	|| strcmp (name, "_Float128") == 0
> -	|| strcmp (name, "complex _Float128") == 0)
> +	|| strcmp (name, "complex _Float128") == 0
> +	|| strcmp (name, "complex(kind=16)") == 0
> +	|| strcmp (name, "real(kind=16)") == 0)
>        return floatformats_ia64_quad;
>  
>    return default_floatformat_for_type (gdbarch, name, len);
> diff --git a/gdb/testsuite/gdb.fortran/complex.exp b/gdb/testsuite/gdb.fortran/complex.exp
> index 136f1c4df79..94ac53afc70 100644
> --- a/gdb/testsuite/gdb.fortran/complex.exp
> +++ b/gdb/testsuite/gdb.fortran/complex.exp
> @@ -33,7 +33,6 @@ gdb_test "print c4" " = \\(1000,-50\\)"
>  gdb_test "print c8" " = \\(321,-22\\)"
>  gdb_test "print dc" " = \\(321,-22\\)"
>  
> -setup_kfail gdb/18644 "*-*-*"
>  gdb_test "print c16" " = \\(-874,19\\)"
>  
>  gdb_test "whatis c" "type = complex\\(kind=4\\)"
> @@ -53,7 +52,6 @@ gdb_test "print \$_creal (dc)" " = 321"
>  gdb_test "whatis \$" " = real\\*8"
>  
>  gdb_test "whatis c16" "type = complex\\(kind=16\\)"
> -setup_kfail gdb/18644 "*-*-*"
>  gdb_test "print \$_creal (c16)" " = -874"
>  gdb_test "whatis \$" " = real\\*16"
>  
> diff --git a/gdb/testsuite/gdb.fortran/printing-types.exp b/gdb/testsuite/gdb.fortran/printing-types.exp
> index 2f6be4ec249..6394e45f4c3 100644
> --- a/gdb/testsuite/gdb.fortran/printing-types.exp
> +++ b/gdb/testsuite/gdb.fortran/printing-types.exp
> @@ -33,3 +33,4 @@ gdb_test "print oneByte"	" = 1"
>  gdb_test "print twobytes"	" = 2"
>  gdb_test "print chvalue"	" = \'a\'"
>  gdb_test "print logvalue"	" = \.TRUE\."
> +gdb_test "print rVal"		" = 2000"
> diff --git a/gdb/testsuite/gdb.fortran/printing-types.f90 b/gdb/testsuite/gdb.fortran/printing-types.f90
> index b4ff928604f..36b63532c8e 100644
> --- a/gdb/testsuite/gdb.fortran/printing-types.f90
> +++ b/gdb/testsuite/gdb.fortran/printing-types.f90
> @@ -18,10 +18,12 @@ program prog
>    integer(2) :: twoBytes
>    character  :: chValue
>    logical(1) :: logValue
> +  real(kind=16) :: rVal
>  
>    oneByte  = 1
>    twoBytes = 2
>    chValue  = 'a'
>    logValue = .true.
> +  rVal = 2000
>    write(*,*) s
>  end
> diff --git a/gdb/testsuite/gdb.fortran/type-kinds.exp b/gdb/testsuite/gdb.fortran/type-kinds.exp
> index 1ae15b96f1a..9d19a9ceb39 100644
> --- a/gdb/testsuite/gdb.fortran/type-kinds.exp
> +++ b/gdb/testsuite/gdb.fortran/type-kinds.exp
> @@ -27,12 +27,6 @@ if { [skip_fortran_tests] } { continue }
>  proc test_cast_1_to_type_kind {base_type type_kind cast_result size_result} {
>      set type_string "$base_type (kind=$type_kind)"
>      gdb_test "p (($type_string) 1)" " = $cast_result"
> -
> -    if {($base_type == "real" || $base_type == "complex")
> -	&& $type_kind == 16} {
> -	setup_kfail gdb/18644 "*-*-*"
> -    }
> -
>      gdb_test "p sizeof (($type_string) 1)" " = $size_result"
>  }
>  
> -- 
> 2.14.5
> 

Re: [PATCHv2] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Tom Tromey]

>>>>> "Andrew" == Andrew Burgess <andrew.burgess@embecosm.com> writes:

>> This new version moves the use of floatformats_ia64_quad out of
>> f-lang.c, which I think is an improvement.  With this done I plan to
>> push this patch shortly unless anyone complains.

Andrew> This has now been pushed.

This caused a crash while testing gdb master using the internal AdaCore
test suite.

In particular what happened is that the ARM target (in fact many
targets) uses default_floatformat_for_type, returns NULL.  Then, if
build_fortran_types is called, it will crash gdb in arch_float_type:

    (top) bt
    #0  0x0000000000b4e3b1 in arch_float_type (gdbarch=0x32ef480, bit=128, name=0x1617675 "real*16", floatformats=0x0)
        at ../../binutils-gdb/gdb/gdbtypes.c:5169
    #1  0x0000000000b14b11 in build_fortran_types (gdbarch=0x32ef480) at ../../binutils-gdb/gdb/f-lang.c:735
    #2  0x0000000000b40d64 in gdbarch_data (gdbarch=0x32ef480, data=0x2d84bc0) at ../../binutils-gdb/gdb/gdbarch.c:5269
    #3  0x0000000000b14bb5 in builtin_f_type (gdbarch=0x32ef480) at ../../binutils-gdb/gdb/f-lang.c:755
    #4  0x0000000000b1378e in f_language_arch_info (gdbarch=0x32ef480, lai=0x33155e8) at ../../binutils-gdb/gdb/f-lang.c:171
    #5  0x0000000000be010d in language_gdbarch_post_init (gdbarch=0x32ef480) at ../../binutils-gdb/gdb/language.c:978
    #6  0x0000000000b40d64 in gdbarch_data (gdbarch=0x32ef480, data=0x31a43d0) at ../../binutils-gdb/gdb/gdbarch.c:5269
    #7  0x0000000000be0140 in language_string_char_type (la=0x15a20e0 <ada_language_defn>, gdbarch=0x32ef480)
        at ../../binutils-gdb/gdb/language.c:990
    #8  0x00000000008ac86c in type_char (par_state=0x7fffffffd400) at ../../binutils-gdb/gdb/ada-exp.y:1453
    #9  0x00000000008a7d85 in ada_yylex () at ../../binutils-gdb/gdb/ada-lex.l:151

In this case it happens via the Ada parser, but it can happen in many
ways, I think.

I'm testing the appended but I'm not really sure this is the best
approach.

Tom

diff --git a/gdb/f-lang.c b/gdb/f-lang.c
index 097e8906d1f..5c8eeab6f6b 100644
--- a/gdb/f-lang.c
+++ b/gdb/f-lang.c
@@ -731,8 +731,12 @@ build_fortran_types (struct gdbarch *gdbarch)
     = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
 		       "real*8", gdbarch_double_format (gdbarch));
   auto fmt = gdbarch_floatformat_for_type (gdbarch, "real(kind=16)", 128);
-  builtin_f_type->builtin_real_s16
-    = arch_float_type (gdbarch, 128, "real*16", fmt);
+  if (fmt != nullptr)
+    builtin_f_type->builtin_real_s16
+      = arch_float_type (gdbarch, 128, "real*16", fmt);
+  else
+    builtin_f_type->builtin_real_s16
+      = arch_type (gdbarch, TYPE_CODE_ERROR, 128, "real*16");
 
   builtin_f_type->builtin_complex_s8
     = arch_complex_type (gdbarch, "complex*8",

Re: [PATCHv2] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Andrew Burgess]

* Tom Tromey <tromey@adacore.com> [2019-05-21 12:06:39 -0400]:

> >>>>> "Andrew" == Andrew Burgess <andrew.burgess@embecosm.com> writes:
> 
> >> This new version moves the use of floatformats_ia64_quad out of
> >> f-lang.c, which I think is an improvement.  With this done I plan to
> >> push this patch shortly unless anyone complains.
> 
> Andrew> This has now been pushed.
> 
> This caused a crash while testing gdb master using the internal AdaCore
> test suite.
> 
> In particular what happened is that the ARM target (in fact many
> targets) uses default_floatformat_for_type, returns NULL.  Then, if
> build_fortran_types is called, it will crash gdb in arch_float_type:
> 
>     (top) bt
>     #0  0x0000000000b4e3b1 in arch_float_type (gdbarch=0x32ef480, bit=128, name=0x1617675 "real*16", floatformats=0x0)
>         at ../../binutils-gdb/gdb/gdbtypes.c:5169
>     #1  0x0000000000b14b11 in build_fortran_types (gdbarch=0x32ef480) at ../../binutils-gdb/gdb/f-lang.c:735
>     #2  0x0000000000b40d64 in gdbarch_data (gdbarch=0x32ef480, data=0x2d84bc0) at ../../binutils-gdb/gdb/gdbarch.c:5269
>     #3  0x0000000000b14bb5 in builtin_f_type (gdbarch=0x32ef480) at ../../binutils-gdb/gdb/f-lang.c:755
>     #4  0x0000000000b1378e in f_language_arch_info (gdbarch=0x32ef480, lai=0x33155e8) at ../../binutils-gdb/gdb/f-lang.c:171
>     #5  0x0000000000be010d in language_gdbarch_post_init (gdbarch=0x32ef480) at ../../binutils-gdb/gdb/language.c:978
>     #6  0x0000000000b40d64 in gdbarch_data (gdbarch=0x32ef480, data=0x31a43d0) at ../../binutils-gdb/gdb/gdbarch.c:5269
>     #7  0x0000000000be0140 in language_string_char_type (la=0x15a20e0 <ada_language_defn>, gdbarch=0x32ef480)
>         at ../../binutils-gdb/gdb/language.c:990
>     #8  0x00000000008ac86c in type_char (par_state=0x7fffffffd400) at ../../binutils-gdb/gdb/ada-exp.y:1453
>     #9  0x00000000008a7d85 in ada_yylex () at ../../binutils-gdb/gdb/ada-lex.l:151
> 
> In this case it happens via the Ada parser, but it can happen in many
> ways, I think.

Sorry for breaking this.

> 
> I'm testing the appended but I'm not really sure this is the best
> approach.
> 
> Tom
> 
> diff --git a/gdb/f-lang.c b/gdb/f-lang.c
> index 097e8906d1f..5c8eeab6f6b 100644
> --- a/gdb/f-lang.c
> +++ b/gdb/f-lang.c
> @@ -731,8 +731,12 @@ build_fortran_types (struct gdbarch *gdbarch)
>      = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
>  		       "real*8", gdbarch_double_format (gdbarch));
>    auto fmt = gdbarch_floatformat_for_type (gdbarch, "real(kind=16)", 128);
> -  builtin_f_type->builtin_real_s16
> -    = arch_float_type (gdbarch, 128, "real*16", fmt);
> +  if (fmt != nullptr)
> +    builtin_f_type->builtin_real_s16
> +      = arch_float_type (gdbarch, 128, "real*16", fmt);
> +  else
> +    builtin_f_type->builtin_real_s16
> +      = arch_type (gdbarch, TYPE_CODE_ERROR, 128, "real*16");

Maybe we should switch back to the original code in the else block,
which would just make use of long double format?

Something like:

diff --git a/gdb/f-lang.c b/gdb/f-lang.c
index 5855c68b38c..b0c57878fec 100644
--- a/gdb/f-lang.c
+++ b/gdb/f-lang.c
@@ -728,8 +728,13 @@ build_fortran_types (struct gdbarch *gdbarch)
     = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
 		       "real*8", gdbarch_double_format (gdbarch));
   auto fmt = gdbarch_floatformat_for_type (gdbarch, "real(kind=16)", 128);
-  builtin_f_type->builtin_real_s16
-    = arch_float_type (gdbarch, 128, "real*16", fmt);
+  if (fmt != nullptr)
+    builtin_f_type->builtin_real_s16
+      = arch_float_type (gdbarch, 128, "real*16", fmt);
+  else
+    builtin_f_type->builtin_real_s16
+      = arch_float_type (gdbarch, gdbarch_long_double_bit (gdbarch),
+			 "real*16", gdbarch_long_double_format (gdbarch));
 
   builtin_f_type->builtin_complex_s8
     = arch_complex_type (gdbarch, "complex*8",

Or maybe we should be even more restrictive, and only use long double
format if its the correct length, like:

diff --git a/gdb/f-lang.c b/gdb/f-lang.c
index 5855c68b38c..e612eeda7f7 100644
--- a/gdb/f-lang.c
+++ b/gdb/f-lang.c
@@ -728,8 +728,16 @@ build_fortran_types (struct gdbarch *gdbarch)
     = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
 		       "real*8", gdbarch_double_format (gdbarch));
   auto fmt = gdbarch_floatformat_for_type (gdbarch, "real(kind=16)", 128);
-  builtin_f_type->builtin_real_s16
-    = arch_float_type (gdbarch, 128, "real*16", fmt);
+  if (fmt != nullptr)
+    builtin_f_type->builtin_real_s16
+      = arch_float_type (gdbarch, 128, "real*16", fmt);
+  else if (gdbarch_long_double_bit (gdbarch) == 128)
+    builtin_f_type->builtin_real_s16
+      = arch_float_type (gdbarch, gdbarch_long_double_bit (gdbarch),
+			 "real*16", gdbarch_long_double_format (gdbarch));
+  else
+    builtin_f_type->builtin_real_s16
+      = arch_type (gdbarch, TYPE_CODE_ERROR, 128, "real*16");
 
   builtin_f_type->builtin_complex_s8
     = arch_complex_type (gdbarch, "complex*8",

What do you think?

Again, sorry for the breakage.

Thanks,
Andrew

Re: [PATCHv2] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Tom Tromey]

Andrew> Sorry for breaking this.

Thanks!  Don't feel too bad, though, I think the occasional bug like
this is going to creep in.

Andrew> Or maybe we should be even more restrictive, and only use long double
Andrew> format if its the correct length, like:

This seems to make the most sense to me, but I don't actually know
Fortran, so I'm just guessing.

thanks,
Tom

Re: [PATCHv2] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Andrew Burgess]

* Tom Tromey <tromey@adacore.com> [2019-05-21 13:05:40 -0400]:

> Andrew> Sorry for breaking this.
> 
> Thanks!  Don't feel too bad, though, I think the occasional bug like
> this is going to creep in.
> 
> Andrew> Or maybe we should be even more restrictive, and only use long double
> Andrew> format if its the correct length, like:
> 
> This seems to make the most sense to me, but I don't actually know
> Fortran, so I'm just guessing.

Honestly, I don't know the specifics here either.  I suspect the real
answer is that different targets and maybe even different Fortran
compilers could use different formats for 16-byte floats.

The intention with my change was to allow each target to find
something suitable (clearly that didn't quite work).  I think falling
back to long double means we're no worse off than we were.

If long double turns out to be wrong for other targets they can always
implement gdbarch_floatformat_for_type like i386 does.

Would you like me to push a fix, or do you want to do it?

Thanks,
Andrew

Re: [PATCHv2] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Tom Tromey]

>>>>> "Andrew" == Andrew Burgess <andrew.burgess@embecosm.com> writes:

Andrew> Would you like me to push a fix, or do you want to do it?

Go for it.

Tom

Re: [PATCHv2] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Andrew Burgess]

* Tom Tromey <tromey@adacore.com> [2019-05-21 16:18:18 -0400]:

> >>>>> "Andrew" == Andrew Burgess <andrew.burgess@embecosm.com> writes:
> 
> Andrew> Would you like me to push a fix, or do you want to do it?
> 
> Go for it.

I pushed the below.

Thanks,
Andrew

---

commit dc42e902cc54af2b7e7b54a1171d562f867342d5
Author: Andrew Burgess <andrew.burgess@embecosm.com>
Date:   Tue May 21 22:14:05 2019 +0100

    gdb/fortran: Handle gdbarch_floatformat_for_type returning nullptr
    
    In this commit:
    
      commit 34d11c682fd96c7dbe3ebd6cd9033e65d51ec7a3
      Date:   Fri May 3 15:23:55 2019 +0100
    
          gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats
    
    GDB was changed such that the Fortran's 16-byte float format was
    obtained by calling gdbarch_floatformat_for_type instead of just using
    gdbarch_long_double_format as it was before.
    
    The problem with this default_floatformat_for_type can return NULL in
    some cases, and the code introduced in 34d11c682f didn't consider
    this.
    
    This commit introduces several alternative strategies for finding a
    suitable 16-byte floating point type.  First GDB calls
    gdbarch_floatformat_for_type (this was what 34d11c682f added), if this
    returns null GDB will use gdbarch_long_double_format if it is the
    correct size (this was the format used before 34d11c682f).  Finally,
    if neither of the above provides a suitable type then GDB will create
    a new dummy type.
    
    This final dummy type is unlikely to provide an correct debug
    experience as far as examining the 16-byte floats, but it should
    prevent GDB crashing.
    
    gdb/ChangeLog:
    
            PR gdb/18644:
            * f-lang.c (build_fortran_types): Handle the case where
            gdbarch_floatformat_for_type returns a nullptr.

diff --git a/gdb/f-lang.c b/gdb/f-lang.c
index 5855c68b38c..e612eeda7f7 100644
--- a/gdb/f-lang.c
+++ b/gdb/f-lang.c
@@ -728,8 +728,16 @@ build_fortran_types (struct gdbarch *gdbarch)
     = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
 		       "real*8", gdbarch_double_format (gdbarch));
   auto fmt = gdbarch_floatformat_for_type (gdbarch, "real(kind=16)", 128);
-  builtin_f_type->builtin_real_s16
-    = arch_float_type (gdbarch, 128, "real*16", fmt);
+  if (fmt != nullptr)
+    builtin_f_type->builtin_real_s16
+      = arch_float_type (gdbarch, 128, "real*16", fmt);
+  else if (gdbarch_long_double_bit (gdbarch) == 128)
+    builtin_f_type->builtin_real_s16
+      = arch_float_type (gdbarch, gdbarch_long_double_bit (gdbarch),
+			 "real*16", gdbarch_long_double_format (gdbarch));
+  else
+    builtin_f_type->builtin_real_s16
+      = arch_type (gdbarch, TYPE_CODE_ERROR, 128, "real*16");
 
   builtin_f_type->builtin_complex_s8
     = arch_complex_type (gdbarch, "complex*8",

Re: [PATCHv2] gdb/fortran: Use floatformats_ia64_quad for fortran 16-byte floats

[Tom Tromey]

>> Go for it.

Andrew> I pushed the below.

Thank you very much!

Tom