I've use a few intrinsics before with GNAT, but I get an error for __builtin_cpu_is when trying to pass in an Chars_Ptr:
error: parameter to builtin must be a string constant or literal
I also tried plugging the "amd" target parameter in directly, but that didn't work.
with Ada.Text_IO;
with Interfaces.C.Strings;
procedure Intrinsics is
procedure CPU_Init;
pragma Import (Intrinsic, CPU_Init, "__builtin_cpu_init");
function Is_CPU (CPU_Name : Interfaces.C.Strings.chars_ptr) return Interfaces.C.Int;
pragma Import (Intrinsic, Is_CPU, "__builtin_cpu_is");
Target : constant Interfaces.C.Strings.Chars_Ptr := Interfaces.C.Strings.New_String ("amd");
begin
CPU_Init;
-- ERROR from the below line, from Is_CPU
Ada.Text_IO.Put_Line (Interfaces.C.Int'Image (Is_CPU (Target)));
end Intrinsics;
References I've been looking at:
GCC Built-ins
Learn Ada, Interfacing w/ C
I think you hit a (current) limitation in importing GCC intrinsics in Ada programs (at least for version GCC/GNAT FSF 11.2). The best workaround is to wrap the builtin/intrinsic with string literal in a C function and then import that C wrapper function in the Ada program.
The error is thrown by the GCC back-end (see here). The built-in only accepts a string literal. This is not clear from the equivalent C signature of the built-in. The equivalent C signature suggests that any constant pointer-to-char is accepted:
int __builtin_cpu_is (const char *cpuname)
However, a simple test shows that this works:
#include <stdbool.h>
bool is_amd() {
return __builtin_cpu_is("amd") != 0;
}
But this doesn't:
#include <stdbool.h>
bool is_cpu(const char *cpuname) {
return __builtin_cpu_is(cpuname) != 0;
}
During compilation the abstract syntax tree is analyzed and the reference to the built-in is being matched along with the actual parameter that is passed in. This actual parameter must be a string literal (a specific tree node type). The string literal is then parsed/matched by GCC. Upon success, the call to the built-in in the syntax tree is (as a whole) replaced by a comparison (done here).
$ gcc -c is_amd.c --dump-tree-original && cat is_amd.c.005t.original
;; Function is_amd (null)
;; enabled by -tree-original
{
return __cpu_model.__cpu_vendor == 2 ? 1 : 0;
}
Now, it seems that the GNAT front-end is currently unable to generate the exact nodes (or node pattern) in the syntax tree that will match those expected by the built-in parser. This is likely because of the declared signature of the built-in and the fact that Ada makes a clear distinction between string values and string pointers.
The GNAT front-end compares the binding to __builtin_cpu_is with the signature declared internally by the GCC back-end and concludes that the cpuname argument must be a constant pointer-to-string. So, something like this:
function Is_CPU (CPU_Name : access constant String) return Integer;
pragma Import (Intrinsic, Is_CPU, "__builtin_cpu_is");
However, when using this signature, you cannot pass a string literal directly; you must use some indirection:
AMD : aliased constant String := "amd"
and then
Is_CPU (AMD'Access);
This indirection is (as far as I can see) preserved before the GNAT front-end hands over the syntax tree to the GCC back-end; GNAT will not "inline" the string literal (that is: will not remove the indirection; which I guess is actually a good thing as you do not want a constant string to be inlined into function calls in general: multiple functions might reference the string and if the string is very big, the effect of inlining might cause the program size to grow significantly).
On the other hand, if you want to pass a string literal directly in Ada, then you need a signature similar to
function Is_CPU (CPU_Name : String) return Integer;
pragma Import (Intrinsic, Is_CPU, "__builtin_cpu_is");
This signature, however, conflicts with the signature declared by the GCC back-end. Moreover, the GNAT front-end will complain that a string literal cannot be passed-by-copy (something that is likely required for the call to be accepted and recognized by the back-end).
So, I guess some additional logic for handling GCC built-ins with string arguments would have to be added to the GNAT front-end in order for this to work and allow something like this to compile:
function Is_AMD return Boolean is
function Is_CPU (CPU_Name : String) return Integer;
pragma Import (Intrinsic, Is_CPU, "__builtin_cpu_is");
begin
return Is_CPU ("amd") /= 0;
end Is_AMD;
Until then, wrapping the intrinsic with string literal in a separate C function (like the is_amd() example above) and then importing this C wrapper function in the Ada program will be the way to go.
Eric found a working solution:
with Ada.Unchecked_Conversion;
with Ada.Text_IO;
with Interfaces.C.Strings;
procedure Main is
procedure CPU_Init;
pragma Import (Intrinsic, CPU_Init, "__builtin_cpu_init");
function Is_CPU (CPU_Name : Interfaces.C.Strings.chars_ptr) return Integer;
pragma Import (Intrinsic, Is_CPU, "__builtin_cpu_is");
function To_Chars_Ptr is
new Ada.Unchecked_Conversion (String, Interfaces.C.Strings.chars_ptr);
begin
CPU_Init;
Ada.Text_IO.Put_Line (Integer'Image (Is_CPU (To_Chars_Ptr ("intel"))));
end;
How about trying Target as shown below
Target : constant Interfaces.C.Char_Ptr := Interfaces.C.To_C ("amd");
Related
I'm attempting to pass a structure from (x86) assembler to Ada on the stack. I've been able to successfully use this pattern in C to accept to wrap a large number of arguments passed from assembly inside a struct and I'm wondering if this will work in a similar way in Ada.
Here is a (contrived, minimal) example:
When I do this, debugging the callee shows that the passed record contains uninitialised data. It appears that Ada is interpreting the C calling convention differently despite the export directive.
The RM contains information about passing structs from Ada to C, saying that it will automatically pass a record as a pointer type, but the inverse does not appear to be true. If you accept a single access type it will simply be filled with the first value on the stack, as one would expect from cdecl.
( Please excuse any minor errors, this isn't my actual code. )
#####################################################################
# Caller
#
# This pushes the values onto the stack and calls the Ada function
#####################################################################
.global __example_function
.type __example_function, #function
__example_function:
push $1
push $2
push $3
push $4
call accepts_struct
ret
----------------------------------------------------------------------------
-- Accepts_Struct
--
-- Purpose:
-- Attempts to accept arguments pass on the stack as a struct.
----------------------------------------------------------------------------
procedure Accepts_Struct (
Struct : Struct_Passed_On_Stack
)
with Export,
Convention => C,
External_Name => "accepts_struct";
----------------------------------------------------------------------------
-- Ideally the four variables passed on the stack would be accepted as
-- the values of this struct.
----------------------------------------------------------------------------
type Struct_Passed_On_Stack is
record
A : Unsigned_32;
B : Unsigned_32;
C : Unsigned_32;
D : Unsigned_32;
end record
with Convention => C;
On the other hand, this works just fine:
procedure Accepts_Struct (
A : Unsigned_32;
B : Unsigned_32;
C : Unsigned_32;
D : Unsigned_32
)
with Export,
Convention => C,
External_Name => "accepts_struct";
That's not a big deal in this minimal case, but if I'm passing 16 or more variables it gets a bit onerous. If you're wondering why I'm doing this, it's an exception handler where the processor automatically passes variables onto the stack to show register states.
Any help here would be greatly appreciated.
The record version does not work because a record is not stored on the stack. Instead 4 Unsigned_32 elements are stored on the stack. If you really want to work with a record instead of four separate unsigned integer values you can assign the four values to the members of your record within the call to "accepts_struct".
Ada expects the first entry in the stack to be a record, not an unsigned_32.
The Ada LRM, section 6.4.1 states:
For the evaluation of a parameter_association: The actual parameter is
first evaluated. For an access parameter, the access_definition is
elaborated, which creates the anonymous access type. For a parameter
(of any mode) that is passed by reference (see 6.2), a view conversion
of the actual parameter to the nominal subtype of the formal parameter
is evaluated, and the formal parameter denotes that conversion. For an
in or in out parameter that is passed by copy (see 6.2), the formal
parameter object is created, and the value of the actual parameter is
converted to the nominal subtype of the formal parameter and assigned
to the formal.
Furthermore, the passing mode for parameters is described in section 6.2:
6.2 Formal Parameter Modes
A parameter_specification declares a formal parameter of mode in, in
out, or out. Static Semantics
A parameter is passed either by copy or by reference. When a parameter
is passed by copy, the formal parameter denotes a separate object from
the actual parameter, and any information transfer between the two
occurs only before and after executing the subprogram. When a
parameter is passed by reference, the formal parameter denotes (a view
of) the object denoted by the actual parameter; reads and updates of
the formal parameter directly reference the actual parameter object.
A type is a by-copy type if it is an elementary type, or if it is a
descendant of a private type whose full type is a by-copy type. A
parameter of a by-copy type is passed by copy, unless the formal
parameter is explicitly aliased.
A type is a by-reference type if it is a descendant of one of the
following:
a tagged type;
a task or protected type;
an explicitly limited record type;
a composite type with a subcomponent of a by-reference type;
a private type whose full type is a by-reference type.
A parameter of a by-reference type is passed by reference, as is an
explicitly aliased parameter of any type. Each value of a by-reference
type has an associated object. For a parenthesized expression,
qualified_expression, or type_conversion, this object is the one
associated with the operand. For a conditional_expression, this object
is the one associated with the evaluated dependent_expression.
For other parameters, it is unspecified whether the parameter is
passed by copy or by reference.
It appears that your compiler is trying to pass the struct by reference rather than by copy. In C all parameters are passed by copy.
Maybe you already solved the problem, but if not, then you might also want to have at look at the interrupt function attribute provided by GCC (see here). I've translated a test of the GCC testsuite which pushes values to the stack (as described in section 6.12 of the Intel SDM) and reads them back in an ISR. The translated Ada version seems to work well. See here for the original C version. See the GCC ChangeLog for some additional info.
main.adb
with PR68037_1;
procedure Main is
begin
PR68037_1.Run;
end Main;
pr68037_1.ads
package PR68037_1 is
procedure Run;
end PR68037_1;
pr68037_1.adb
with System.Machine_Code;
with Ada.Assertions;
with Interfaces.C;
with GNAT.OS_Lib;
package body PR68037_1 is
-- Ada-like re-implementation of
-- gcc/testsuite/gcc.dg/guality/pr68037-1.c
subtype uword_t is Interfaces.C.unsigned_long; -- for x86-64
ERROR : constant uword_t := 16#1234567_0#;
IP : constant uword_t := 16#1234567_1#;
CS : constant uword_t := 16#1234567_2#;
FLAGS : constant uword_t := 16#1234567_3#;
SP : constant uword_t := 16#1234567_4#;
SS : constant uword_t := 16#1234567_5#;
type interrupt_frame is
record
ip : uword_t;
cs : uword_t;
flags : uword_t;
sp : uword_t;
ss : uword_t;
end record
with Convention => C;
procedure fn (frame : interrupt_frame; error : uword_t)
with Export, Convention => C, Link_Name => "__fn";
pragma Machine_Attribute (fn, "interrupt");
--------
-- fn --
--------
procedure fn (frame : interrupt_frame; error : uword_t) is
use Ada.Assertions;
use type uword_t;
begin
-- Using the assertion function here. In general, be careful when
-- calling subprograms from an ISR. For now it's OK as we will not
-- return from the ISR and not continue the execution of an interrupted
-- program.
Assert (frame.ip = IP , "Mismatch IP");
Assert (frame.cs = CS , "Mismatch CS");
Assert (frame.flags = FLAGS, "Mismatch FLAGS");
Assert (frame.sp = SP , "Mismatch SP");
Assert (frame.ss = SS , "Mismatch SS");
-- At the end of this function IRET will be executed. This will
-- result in a segmentation fault as the value for EIP is nonsense.
-- Hence, abort the program before IRET is executed.
GNAT.OS_Lib.OS_Exit (0);
end fn;
---------
-- Run --
---------
procedure Run is
use System.Machine_Code;
use ASCII;
begin
-- Mimic the processor behavior when an ISR is invoked. See also:
--
-- Intel (R) 64 and IA-32 Architectures / Software Developer's Manual
-- Volume 3 (3A, 3B, 3C & 3D) : System Programming Guide
-- Section 6.12: Exception and Interrupt Handling
--
-- Push the data to the stack and jump unconditionally to the
-- interrupt service routine.
Asm
(Template =>
"push %0" & LF &
"push %1" & LF &
"push %2" & LF &
"push %3" & LF &
"push %4" & LF &
"push %5" & LF &
"jmp __fn",
Inputs =>
(uword_t'Asm_Input ("l", SS),
uword_t'Asm_Input ("l", SP),
uword_t'Asm_Input ("l", FLAGS),
uword_t'Asm_Input ("l", CS),
uword_t'Asm_Input ("l", IP),
uword_t'Asm_Input ("l", ERROR)),
Volatile => True);
end Run;
end PR68037_1;
I compiled the program in GNAT CE 2019 with compiler options -g -mgeneral-regs-only (copied from the GCC test). Note that the parameter interrupt_frame will be passed by reference (see RM B.3 69/2).
I have an unusual situation in which elaboration code is simply not being executed at all. This is not an elaboration order issue, but rather an elaboration at all issue.
The problem is that I don't "with" the unit in question whatsoever, yet in theory it should still be accessible, as long as its elaboration occurs.
Of course I could just add a useless "with" for the unit in question, but in my real use case there are a large number of units that I would have to do that with.
My question is if there is any way either in the code, through pragmas, in the gpr project file, or through command-line switches that I could force the compiler to include a file even though it thinks the file isn't referenced?
Here is a minimal working example:
as.ads:
package As is
type A is tagged null record;
type Nothing is null record;
function Create (Ignored : not null access Nothing) return A;
function Image (From : A) return String;
end As;
as.adb:
package body As is
function Create (Ignored : not null access Nothing) return A is
(null record);
function Image (From : A) return String is ("A");
end As;
finder.ads:
with Ada.Tags;
package Finder is
procedure Register (Name : String; Tag : Ada.Tags.Tag);
function Find (Name : String; Default : Ada.Tags.Tag) return Ada.Tags.Tag;
end Finder;
finder.adb:
with Ada.Containers.Indefinite_Vectors;
package body Finder is
type Name_Tag (Size : Natural) is
record
Name : String (1 .. Size);
To : Ada.Tags.Tag;
end record;
package Name_Tag_Vectors is new Ada.Containers.Indefinite_Vectors (Positive, Name_Tag);
Name_Tags : Name_Tag_Vectors.Vector := Name_Tag_Vectors.Empty_Vector;
procedure Register (Name : String; Tag : Ada.Tags.Tag) is begin
Name_Tags.Append ((Name'Length, Name, Tag));
end Register;
function Find (Name : String; Default : Ada.Tags.Tag) return Ada.Tags.Tag is begin
for Tag of Name_Tags loop
if Tag.Name = Name then
return Tag.To;
end if;
end loop;
return Default;
end Find;
end Finder;
bs.ads:
with As;
package Bs is
type B is new As.A with null record;
function Create (Ignored : not null access As.Nothing) return B;
function Image (From : B) return String;
end Bs;
bs.adb:
with Finder;
package body Bs is
function Create (Ignored : not null access As.Nothing) return B is
(As.Create (Ignored) with null record);
function Image (From : B) return String is ("B");
begin
Finder.Register ("B", B'Tag);
end Bs;
test.adb:
with As; use As;
-- with Bs; -- (uncommenting this line solves my problem, but what if I had the rest of the alphabet?)
with Finder;
with Ada.Tags.Generic_Dispatching_Constructor;
with Ada.Text_IO;
procedure Test is
function Constructor is new Ada.Tags.Generic_Dispatching_Constructor (
T => A,
Parameters => Nothing,
Constructor => Create);
Nada : aliased Nothing := (null record);
What : A'Class := Constructor (Finder.Find ("B", A'Tag), Nada'Access);
begin
Ada.Text_IO.Put_Line (What.Image);
end Test;
The compiler thinks your package Bs isn't referenced because it isn't. You don't have a with clause for it, so it's not part of your program.
A simple example:
a.ads
package A is
procedure Blah;
end A;
a.adb
with Ada.Text_IO;
package body A is
procedure Blah is begin null; end Blah;
begin
Ada.Text_IO.Put_Line("Elaborate A");
end A;
b.ads
package B is
procedure Blah;
end B;
b.adb
with Ada.Text_IO;
package body B is
procedure Blah is begin null; end Blah;
begin
Ada.Text_IO.Put_Line("Elaborate B");
end B;
main.adb
with Ada.Text_IO;
with A;
procedure Main is
begin
Ada.Text_IO.Put_Line("Main");
end Main;
When I run main, it prints
Elaborate A
Main
It doesn't print Elaborate B because that package isn't part of the program; it's just a couple of source files in the same directory.
The obvious solution is to add the with clauses.
I don't know whether there's a less obvious solution. If there is, it's probably compiler-specific. But I'm not sure why a compiler would have a feature that lets you incorporate an otherwise unused package into a program.
What I’ve done (e.g. here ff) is to actually reference the units in the main program (with pragma Unreferenced to prevent warnings).
Alternatively, you could have a package e.g. Required_Units with all the necessary withs included, and then with that from the main program.
Even if there was some alternative process, you’d have to tell it what units you need to have included; might as well go with the flow and do it in Ada!
Since the package Bs is invisible to your program, so is the type B.
So the next question is: why do you need to register type B if it is not used anywhere?
If an Ada compiler did elaborate all units (packages or standalone subprograms) that are irrelevant to a main program, but are visible through source path, it would become really messy!...
I'm interfacing to a USB device (on Debian Stretch) using hidraw, and I need to process some information supplied by the USB device in the form of wchar_t* which I need to convert into (Ada) Wide_String. This is giving some trouble and I'm not seeing a clean way forward using the facilities in Interfaces.C and Interfaces.C.Strings.
All files are edited down without destroying their consistency. They will build, but without one of these, they won't actually run.
The problem is that device information like Serial Number and Product Name are presented by the Linux device driver as an access stddef_h.wchar_t from which I want to return a Wide_String or even a normal String) and I'm not seeing any good way to get there.
Interfaces.C.Strings has function Value (Item : in chars_ptr) return String; but no equivalent exists for Wide_String that I can see. So I think I need an equivalent Value function for wide characters.
The approach below uses To_Ada (from Interfaces.C) to return a Wide_String given a wchar_array. It fails, of course, because an access wchar_t is not convertible to a wchar_array.
-- helper function to deal with wchar_t * to wide_string
function Value (P : access stddef_h.wchar_t) return Wide_String is
temp : Wide_String(1 .. 256);
count : natural := 0;
-- ugliness to convert pointer types
type sd_wchar_ptr is access all stddef_h.wchar_t;
type wchar_array_ptr is access wchar_array;
Function To_Wchar_Array_Ptr is new Ada.Unchecked_Conversion(sd_wchar_ptr, wchar_array_ptr);
-- this does NOT create the required wchar_array pointer
WCP : wchar_array_ptr := To_Wchar_Array_Ptr(sd_wchar_ptr(P));
begin
Put_Line("Wide string");
To_Ada(WCP.all, temp, count);
Put_Line("Wide string length " & natural'image(count));
return temp(1..count);
end Value;
and the inevitable result
./test_hid
Wide string
Execution terminated by unhandled exception raised STORAGE_ERROR :
stack overflow or erroneous memory access
A similar character by character approach would be possible ... if (and I can't believe I'm saying this!) you could increment access types...
Feels like there's something missing from Interfaces.C here... what am I missing? any ideas to get round this relatively trivial seeming stumbling block?
EDIT : I'm leaning towards some brazen theft from the Interfaces.C.Strings sources with appropriate changes, but I'd welcome alternative suggestions.
The rest of this below is the full story so far (including all code necessary to reproduce)
Step 1 : generate low level Ada bindings automatically using gcc.
gcc -c -fdump-ada-spec-slim /usr/include/hidapi/hidapi.h
producing the low level binding package hidapi_hidapi_h
pragma Ada_2005;
pragma Style_Checks (Off);
with Interfaces.C; use Interfaces.C;
with Interfaces.C.Strings;
with stddef_h;
with System;
package hidapi_hidapi_h is
-- see source file /usr/include/hidapi/hidapi.h
type hid_device_info is record
path : Interfaces.C.Strings.chars_ptr; -- /usr/include/hidapi/hidapi.h:51
vendor_id : aliased unsigned_short; -- /usr/include/hidapi/hidapi.h:53
product_id : aliased unsigned_short; -- /usr/include/hidapi/hidapi.h:55
serial_number : access stddef_h.wchar_t; -- /usr/include/hidapi/hidapi.h:57
release_number : aliased unsigned_short; -- /usr/include/hidapi/hidapi.h:60
manufacturer_string : access stddef_h.wchar_t; -- /usr/include/hidapi/hidapi.h:62
product_string : access stddef_h.wchar_t; -- /usr/include/hidapi/hidapi.h:64
usage_page : aliased unsigned_short; -- /usr/include/hidapi/hidapi.h:67
usage : aliased unsigned_short; -- /usr/include/hidapi/hidapi.h:70
interface_number : aliased int; -- /usr/include/hidapi/hidapi.h:75
next : access hid_device_info; -- /usr/include/hidapi/hidapi.h:78
end record;
pragma Convention (C_Pass_By_Copy, hid_device_info); -- /usr/include/hidapi/hidapi.h:49
function hid_enumerate (arg1 : unsigned_short; arg2 : unsigned_short) return access hid_device_info; -- /usr/include/hidapi/hidapi.h:132
pragma Import (C, hid_enumerate, "hid_enumerate");
end hidapi_hidapi_h;
This is a low level binding, exposing C types (and the binding generator has decided that the wchar_t in Interfaces.C isn't good enough, it wants one from stddef.h too, so...
pragma Ada_2005;
pragma Style_Checks (Off);
with Interfaces.C; use Interfaces.C;
package stddef_h is
-- unsupported macro: NULL ((void *)0)
subtype size_t is unsigned_long; -- /usr/lib/gcc/x86_64-linux-gnu/6/include/stddef.h:216
subtype wchar_t is int; -- /usr/lib/gcc/x86_64-linux-gnu/6/include/stddef.h:328
end stddef_h;
Because it is a low level binding; we want to hide it (and implement RAII etc) behind a simpler and more usable high level binding, so ... (below)
with Ada.Finalization; use Ada.Finalization;
private with hidapi_hidapi_h;
private with System;
package hidapi is
type id is new natural range 0 .. 2**16 - 1;
type hid_device is new Limited_Controlled with private;
-- find first matching devices by enumeration : the RA part of RAII.
function enumerate (vendor_id, product_id : id) return hid_device;
-- accessors for device characteristics on enumerated device
function Serial_No (D : hid_device) return Wide_String;
function Product_String (D : hid_device) return Wide_String;
private
type hid_device is new Limited_Controlled with record
member : access hidapi_hidapi_h.hid_device_info;
addr : System.Address;
end record;
end hidapi;
and its implementation, containing the problem function value to return a Wide_String.
with hidapi_hidapi_h;
with Interfaces.C; use Interfaces.C;
with Ada.Text_IO; use Ada.Text_IO;
with Ada.Unchecked_Conversion;
with stddef_h;
package body hidapi is
function enumerate (vendor_id, product_id : id) return hid_device is
use hidapi_hidapi_h;
first : access hid_device_info;
begin
first := hid_enumerate(unsigned_short(vendor_id), unsigned_short(product_id));
if first /= null then
return H : hid_device do
H.member := first;
H.addr := System.Null_Address;
end return;
else raise Program_Error;
end if;
end enumerate;
-- helper function to deal with wchar_t * to wide_string
function Value (P : access stddef_h.wchar_t) return Wide_String is
temp : Wide_String(1 .. 256);
count : natural := 0;
type sd_wchar_ptr is access all stddef_h.wchar_t;
type wchar_array_ptr is access wchar_array;
Function To_Wchar_Array_Ptr is new Ada.Unchecked_Conversion(sd_wchar_ptr, wchar_array_ptr);
WCP : wchar_array_ptr := To_Wchar_Array_Ptr(sd_wchar_ptr(P));
begin
Put_Line("Wide string");
To_Ada(WCP.all, temp, count);
Put_Line("Wide string length " & natural'image(count));
return temp(1..count);
end Value;
function Serial_No (D : hid_device) return Wide_String is
use hidapi_hidapi_h;
begin
return Value(D.member.serial_number);
end Serial_No;
function Product_String (D : hid_device) return Wide_String is
use hidapi_hidapi_h;
begin
return Value(D.member.product_string);
end Product_String;
end hidapi;
And of course a test case to exercise it...
with Hidapi;
with Ada.Wide_Text_IO;
procedure Test_Hid is
usbrelay_vendor_id : constant Hidapi.id := 16#16c0#;
usbrelay_product_id : constant Hidapi.id := 16#05df#;
Device : Hidapi.hid_device := Hidapi.Enumerate(usbrelay_vendor_id, usbrelay_product_id);
begin
Ada.Wide_Text_IO.Put_Line("Serial : " & Device.Serial_No);
Ada.Wide_Text_IO.Put_Line("Product : " & Device.Product_String);
end Test_Hid;
One answer, slavishly copying the approach in the package body for Tnterfaces.C.Strings with necessary changes.
The naughty stuff is in functions "+" and Peek which use Unchecked Conversions on pointers,
to permit address arithmetic. Not pointer increment, but pointer+offset. One change is that the offset has to be scaled for 4 byte characters. I haven't set that scaling in a portable manner, but note that "+" will overload for each different return type so that offsets will be scaled appropriately for different named access types.
to allow the stddef_h.wchar_t to be viewed as a Wide_Wide_Character in the absence of any type conversion function. Whether the representation is correct is another matter (here, it is) but this technique could also be used to fake the input type of a suitable conversion function like To_Ada in Interfaces.C.
The remainder is straightforward character by character handling. One other change (so far) is to return Wide_Wide_Character rather than Wide_Character (because as the stddef_h package above reveals, the stored characters are 32 bit, same size as Interfaces.C.int. I'm happy to change my interface, but Wide_String could be easily handled by Ada.Strings packages.
type sd_wchar_ptr is access all stddef_h.wchar_t;
type w_w_char_ptr is access all char32_t;
-- Two Unchecked_Conversions to allow pointer arithmetic
-- And a third to allow the resulting storage to be interpreted as Wide_Wide_Char
function To_Sd_wchar_ptr is new Ada.Unchecked_Conversion (System.Address, sd_wchar_ptr);
function To_Address is new Ada.Unchecked_Conversion (sd_wchar_ptr, System.Address);
function To_Wchar_Ptr is new Ada.Unchecked_Conversion (sd_wchar_ptr, w_w_char_ptr);
-- pointer + offset arithmetic, with offset scaled for size of stddef_h.wchar_t;
-- TODO: attempted better way of computing word size ran into type errors
function "+" (Left : sd_wchar_ptr; Right : size_t) return sd_wchar_ptr is
begin
return To_Sd_wchar_ptr (To_Address (Left) + Storage_Offset (Right) * 4);
end "+";
function Peek (From : sd_wchar_ptr) return char32_t is
begin
return To_Wchar_Ptr(From).all;
end Peek;
function Strlen (Item : sd_wchar_ptr) return size_t is
Item_Index : size_t := 0;
begin
if Item = Null then
raise Program_Error;
end if;
loop
if Peek (Item + Item_Index) = char32_nul then
return Item_Index;
end if;
Item_Index := Item_Index + 1;
end loop;
end Strlen;
function Value (Item : sd_wchar_ptr) return char32_array is
Result : char32_array (0 .. Strlen (Item));
begin
if Item = Null then
raise Program_Error;
end if;
Put_Line("String length " & size_t'image(Strlen(Item)));
-- Note that the following loop will also copy the terminating Nul
for J in Result'Range loop
Result (J) := Peek (Item + J);
end loop;
return Result;
end Value;
-- helper function to deal with wchar_t * to wide_wide_string
function Value (Item : access stddef_h.wchar_t) return Wide_Wide_String is
begin
return To_Ada (Value (sd_wchar_ptr(Item)));
end Value;
How do I fix this Seg Fault in my DLL?
I'm generating a Windows DLL (in Ada) and using the DLL from an Ada
program. I’m using AdaCore’s GNAT GPS v6.0.1 IDE for both the DLL
and an Ada program to test the DLL, running on a Windows 7 machine.
Two separate project files are used, one for the DLL, the other for
the test driver. The DLL does not have any DLLMain nor initialization
or finalization routines.
As a first step (because I've never created a DLL or used GPS prior to this, do know some Ada though), I coded two very simple functions for the DLL. One function returns a pointer to a string, the other function returns a fixed length string.
The test program successfully calls the DLL function that returns a fixed-length
string, however when calling the function that returns a string pointer, a
segmentation fault occurs. Here is the gcc debug output:
Program received signal SIGSEGV, Segmentation fault.
0x6b81dd2c in system.secondary_stack.ss_mark () from C:\GNAT\2014\bin\libgnat-2014.dll
(gdb) quit
Here is the code:
DLL Spec
with Ada.Strings.Fixed; use Ada.Strings.Fixed;
package String_Utils is
type String_Ptr_T is access String;
type Spec_Str is new String (1..7);
function Int_Trim_Left( IntToTrim : Integer) return String_Ptr_T;
pragma Export(DLL, Int_Trim_Left, "String_Utils__Int_Trim_Left");
function Spec( Input_Int : Integer) return Spec_Str;
pragma Export(DLL, Spec, "String_Utils__Spec");
end String_Utils;
DLL Body
package body String_Utils is
function Int_Trim_Left( IntToTrim : Integer) return String_Ptr_T is
String_Ptr : String_Ptr_T;
begin
Text_IO.Put_Line("About to call new String in DLL.");
String_Ptr := new String'(
Ada.Strings.Fixed.Trim(Integer'Image(IntToTrim),
Ada.Strings.Left));
return String_Ptr;
end;
--
function Spec( Input_Int : Integer) return Spec_Str
is
Result_Spec : String := "ABC-UNK";
begin
case Input_Int is
when 1 => return "ABC-STD"; -- Standard
when 2 => return "ABC-PRF"; -- Performance
when 3 => return "DEF-DTL"; -- Detailed
when Others => return "ABC-UNK";
end case;
end;
DLL Project File
project HAGUtils is
for Library_Name use "HAGUtils";
for Library_Dir use "libdir";
for Library_Version use "0.01";
for Library_Kind use "dynamic";
for Object_Dir use "obj";
for Source_Dirs use ("src");
for Source_Files use ("string_utils.adb", "string_utils.ads");
end HAGUtils;
Test Driver
-- Driver for DLL
with Text_IO; use Text_IO;
procedure test_ada_dll is
type String_Ptr_T is access String;
subtype String7 is String(1..7);
input_val : Integer := 0;
Spec_Str : String7 := (Others => ' ');
Int_String_Ptr : String_Ptr_T:= null;
-- Import
function Int_Trim_Left ( IntToTrim : Integer) return String_Ptr_T
is
function Inner_Int_Trim_Left ( IntToTrim : Integer) return String_Ptr_T;
pragma Import (DLL, Inner_Int_Trim_Left, "String_Utils__Int_Trim_Left");
begin
return Inner_Int_Trim_Left (IntToTrim);
end Int_Trim_Left;
-- Import
function Spec ( Input_Int : Integer) return String7
is
function Inner_Spec ( Input_Int : Integer) return String7;
pragma Import (DLL, Inner_Spec, "String_Utils__Spec");
begin
return Inner_Spec (Input_Int);
end Spec;
begin
input_val := 3;
Spec_Str := Spec(input_val);
Text_IO.Put_Line("The Spec is -- " & Spec_Str);
Text_IO.Put_Line("Calling Int_Trim_Left with --" & Integer'Image(input_val));
Int_String_Ptr := Int_Trim_Left(input_val);
Text_IO.Put_Line("After call --" & Int_String_Ptr.all);
end;
I think that the SEGV happened because your DLL wasn’t initialized. The Ada runtime system needs initialization, which in the absence of DLLs would be called up in the GNAT bind process (you may have seen calls to gnatbind or gprbind flashing up the screen).
However, you have a DLL that requires the RTS to be initialized (the part that deals with the secondary stack, which is where GNAT constructs temporary unconstrained objects such as strings); but the binder isn’t aware of this because of the way you’ve linked your program (you don’t say, but I suspect you’ve specified the DLL via -lHAGutils?).
The way to get GNAT to handle this for you is to write a project file for the test program and have it with your DLL’s project:
with "HAGutils";
project Test_Ada_Dll is
for Main use ("test_ada_dll.adb");
for Exec_Dir use ".";
for Source_Files use ("test_ada_dll.adb");
for Object_Dir use ".build";
end Test_Ada_Dll;
This then makes the interfaces of HAGlib visible to test_ada_dll, so you can change it to say
with Text_IO; use Text_IO;
with String_Utils;
procedure test_ada_dll is
input_val : Integer := 0;
Spec_Str : String_Utils.Spec_Str := (Others => ' ');
Int_String_Ptr : String_Utils.String_Ptr_T:= null;
begin
input_val := 3;
Spec_Str := String_Utils.Spec(input_val);
Text_IO.Put_Line("The Spec is -- " & String (Spec_Str));
Text_IO.Put_Line("Calling Int_Trim_Left with --" & Integer'Image(input_val));
Int_String_Ptr := String_Utils.Int_Trim_Left(input_val);
Text_IO.Put_Line("After call --" & Int_String_Ptr.all);
end;
(note, the conversion in Text_IO.Put_Line("The Spec is -- " & String (Spec_Str)); is because Spec_Str is a derived type; I think it’s be more normal in this case to make it a subtype).
Further, you no longer need to use the pragma Exports in String_Utils’s spec.
The result of this is that the binder is aware of the properties of your HAGutils DLL, and can arrange for the necessary initializations to happen.
There is a way in which you can make your original code work, which is to use the GPR attribute Library_Auto_Init in HAGutils.gpr:
for Library_Auto_Init use “true”;
but I think you’d have to make HAGlib a proper standalone library. This is quite complex to get right, and not necessary to get the library working to start with.
I want to print the address of an access variable (pointer) for debugging purposes.
type Node is private;
type Node_Ptr is access Node;
procedure foo(n: in out Node_Ptr) is
package Address_Node is new System.Address_To_Access_Conversions(Node);
use Address_Node;
begin
Put_Line("node at address " & System.Address_Image(To_Address(n)));
end foo;
Address_Image returns the string representation of an address.
System.Address_To_Access_Conversions is a generic package to convert between addresses and access types (see ARM 13.7.2), defined as follows:
generic
type Object(<>) is limited private;
package System.Address_To_Access_Conversions is
-- [...]
type Object_Pointer is access all Object;
-- [...]
function To_Address(Value : Object_Pointer) return Address;
-- [...]
end System.Address_To_Access_Conversions;
gnat gives me the following errors for procedure foo defined above:
expected type "System.Address_To_Access_Conversions.Object_Pointer" from instance at line...
found type "Node_Ptr" defined at ...
Object_Pointer ist definied as access all Object. From my understanding the type Object is Node, therefore Object_Ptr is access all Node. What is gnat complaining about?
I guess my understanding of Ada generics is flawed and I am not using System.Address_To_Access_Conversions correctly.
EDIT:
I compiled my code with "gnatmake -gnatG" to see the generic instantiation:
package address_node is
subtype btree__clear__address_node__object__2 is btree__node;
type btree__clear__address_node__object_pointer__2 is access
all btree__clear__address_node__object__2;
function to_address (value :
btree__clear__address_node__object_pointer__2) return
system__address;
end address_node;
btree__node is the mangled name of the type Node as defined above, so I really think the parameter type for to_address() is correct, yet gnat is complaining (see above).
I don't have a compiler in front of me at the moment, but doesn't this work?
procedure foo(n: in out Node_Ptr) is
begin
Put_Line("node at address " & System.Address_Image(n.all'address)); --'
end foo;
Ok, explicit type conversion does the trick:
procedure Foo(n: in out Node_Ptr) is
package Address_Node is new System.Address_To_Access_Conversions(Node);
use Address_Node;
p : Address_Node.Object_Pointer;
begin
p := Address_Node.Object_Pointer(n);
Put_Line("node at address " & System.Address_Image(To_Address(p)));
end Foo;
Takes some time getting used to Ada... ;-)