Make the Ada programming language available on GNU/Hurd in its GCC GNAT implementation, and enable Hurd-specific features.
There is a FOSS Factory bounty (p259) on this task.
Part I
First, make the language functional, have its test suite pass without errors.
Using, for example, the build instructions from GCC, you'll need to add
ada
to --enable-languages
.
Original GSoC Task Description
Obsolete
This is no longer valid as a Google Summer of Code project; existing work needs to be integrated first.
An initial port of the GNU Ada Translator (GNAT) is available for the Hurd.
The goal of this project is to continue improving and polishing that port and getting it integrated in the upstream GCC repository. It requires implementing some explicitly system-specific stuff in GNAT (mostly in its runtime libraries), and for that also address a number of issues in Hurd and other libraries. Knowledge of Ada is a must; some Hurd knowledge will have to be acquired while working on the project.
Designing and implementing language bindings is a follow-up project.
Possible mentors: Samuel Thibault (youpi), Thomas Schwinge (tschwinge).
Exercise: Fix one of the known issues of GNAT on the Hurd.
Open Issue page. Entry in the GCC wiki.
Debian GCC
There has a patch been added for GNU/kFreeBSD:
bfe081336914729fc0180c07ab4afa41965100f2
, `git-svn-id:
svn://svn.debian.org/gcccvs/branches/sid@5638
6ca36cf4-e1d1-0310-8c6f-e303bb2178ca'
IRC, freenode, #hurd, 2012-07-17
<gnu_srs> I've found the remaining problem with gnat backtrace for Hurd!
Related to the stack frame.
<gnu_srs> This version does not work: one relying on static assumptions
about the frame layout
<gnu_srs> Causing segfaults.
<gnu_srs> Any interest to create a test case out of that piece of code,
taken from gcc/ada/tracebak.c?
<braunr> gnu_srs: sure
IRC, freenode, #hurd, 2012-07-18
<braunr> "Digging further revealed that the GNU/Hurd stack frame does not
seem to
<braunr> be static enough to define USE_GENERIC_UNWINDER in
gcc/ada/tracebak.c.
<braunr> "
<braunr> what do you mean by a "stack frame does not seem to be static
enough" ?
<gnu_srs> I can qoute from the source file if you want. Otherwise look at
the code yourself: gcc/ada/tracebak,c
<gnu_srs> I mean that something is wrong with the stack frame for
Hurd. This is the code I wanted to use as a test case for the stack.
<gnu_srs> Remember?
<braunr> more or less
<braunr> ah, "static assumptions"
<braunr> all right, i don't think anything is "wrong" with stack frames
<braunr> but if you use a recent version of gcc, as indicated in the code,
-fomit-frame-pointer is enabled by default
<braunr> so your stack frame won't look like it used to be without the
option
<braunr> hence the need for USE_GCC_UNWINDER
<braunr> http://en.wikipedia.org/wiki/Call_stack explains this very well
<gnu_srs> However, kfreebsd does not seem to need USE_GCC_UNWINDER, how
come?
<braunr> i guess they don't omit the frame pointer
<braunr> your fix is good btw
<gnu_srs> thanks
IRC, freenode, #hurd, 2012-07-19
<gnu_srs> tschwinge: The bug in #681998 should go upstream. Applied in
Debian already. Hopefully this is the last patch needed for the port of
GNAT to Hurd.
2015-11-27
Is this related to:
gcc/ada/tracebak.c
:
#if (defined (__x86_64__) || defined (__linux__)) && !defined (__USING_SJLJ_EXCEPTIONS__)
#define USE_GCC_UNWINDER
#else
#define USE_GENERIC_UNWINDER
Svante's patch
A basic port has been done by Svante, Debian bug #668425, Debian bug #681998,
id:"1333104917.2962.439.camel@s1499.it.kth.se"
, but there's still
lots of work remaining.
The port is not yet upstream: the maintainer raised some concerns that
I have not yet found the time to follow up on, id:"1339857758-5032-1-git-send-email-thomas@codesourcery.com"
.
While the test
results of the GCC/GNAT testsuite don't look bad (but there are a few
unresolved issues, and the testsuite appears to be a rather small one), I don't
know if the port has yet seen any real-world usage, such as using it for any
bigger Ada code bases, or any Ada testsuites.
"try2"
id:"1456875776.5852.194.camel@gmail.com"
.
GCC trunk r234252.
gcc/ada/s-osinte-gnu.adb
: r235467 (on trunk)/r235468 (on gcc-6-branch).
getpagesize
patch: r235896 (on trunk)/r235897 (on gcc-6-branch).
Still incomplete? (Changes to generic files?)
2016-12-14
Based on commit 17f0f5acdcfb9181203ca9b6d8f8cbc64e5ce843, tschwinge tried
to build with --enable-languages=all,ada
, but both with --disable-bootstrap
and --enable-bootstrap
we run into:
[...]
make[3]: Entering directory '/media/erich/home/thomas/tmp/gcc/hurd/master.build/gcc/ada/tools'
../../gnatbind -I- -I../rts -I. -I/home/thomas/tmp/gcc/hurd/master/gcc/ada -o b_gnatm.adb gnatmake.ali
error: "system.ads" has restriction No_Implicit_Dynamic_Code
error: but the following files violate this restriction:
error: "make.adb"
error: "makeutl.adb"
error: "prj.adb"
error: "prj-env.adb"
error: "prj-conf.adb"
error: "prj-nmsc.adb"
../gcc-interface/Makefile:2987: recipe for target 'b_gnatm.adb' failed
make[3]: *** [b_gnatm.adb] Error 5
make[3]: Leaving directory '/media/erich/home/thomas/tmp/gcc/hurd/master.build/gcc/ada/tools'
Makefile:194: recipe for target 'gnattools-native' failed
make[2]: *** [gnattools-native] Error 2
make[2]: Leaving directory '/media/erich/home/thomas/tmp/gcc/hurd/master.build/gnattools'
Makefile:9650: recipe for target 'all-gnattools' failed
make[1]: *** [all-gnattools] Error 2
make[1]: Leaving directory '/media/erich/home/thomas/tmp/gcc/hurd/master.build'
Makefile:895: recipe for target 'all' failed
make: *** [all] Error 2
(Partial) review of differences:
gcc/ada/gcc-interface/Makefile.in
, Linux vs. Hurd
LIBGNAT_TARGET_PAIRS = \
a-intnam.ads<a-intnam-linux.ads vs. a-intnam.ads<a-intnam-freebsd.ads
a-synbar.adb<a-synbar-posix.adb vs. nil
a-synbar.ads<a-synbar-posix.ads vs. nil
s-inmaop.adb<s-inmaop-posix.adb
s-intman.adb<s-intman-posix.adb
s-tpopsp.adb<s-tpopsp-tls.adb vs. s-tpopsp.adb<s-tpopsp-posix-foreign.adb
g-sercom.adb<g-sercom-linux.adb vs. nil
a-exetim.adb<a-exetim-posix.adb vs. nil
a-exetim.ads<a-exetim-default.ads vs. nil
s-linux.ads<s-linux.ads vs. nil
s-osinte.adb<s-osinte-posix.adb vs. s-osinte.adb<s-osinte-gnu.adb
$(ATOMICS_TARGET_PAIRS)
system.ads<system-linux-x86.ads vs. system.ads<system-freebsd-x86.ads
$(X86_TARGET_PAIRS)
s-mudido.adb<s-mudido-affinity.adb vs. nil
s-osinte.ads<s-osinte-linux.ads vs. s-osinte.ads<s-osinte-gnu.ads
s-osprim.adb<s-osprim-posix.adb
s-taprop.adb<s-taprop-linux.adb vs. s-taprop.adb<s-taprop-posix.adb
s-tasinf.ads<s-tasinf-linux.ads vs. nil
s-tasinf.adb<s-tasinf-linux.adb vs. nil
s-taspri.ads<s-taspri-posix.ads
EH_MECHANISM=-gcc
THREADSLIB = -lpthread -lrt vs. THREADSLIB = -lpthread
EXTRA_GNATRTL_NONTASKING_OBJS += g-sse.o g-ssvety.o
EXTRA_GNATRTL_TASKING_OBJS=s-linux.o a-exetim.o vs. nil
TOOLS_TARGET_PAIRS = \
mlib-tgt-specific.adb<mlib-tgt-specific-linux.adb
indepsw.adb<indepsw-gnu.adb
GNATLIB_SHARED = gnatlib-shared-dual
MISCLIB = -ldl vs. nil
GMEM_LIB = gmemlib
LIBRARY_VERSION := $(LIB_VERSION)
TOOLS1_LIBS=-ldl vs. nil
Eventually tschwinge found http://gcc.gnu.org/PR57188, and similarly fixed the issue.
getcontext
/makecontext
/setcontext
/swapcontext
usage analysis
In context of tls-threadvar. Looking at GCC trunk commit f6568ea476aa52a6e23c6db43b3e240cde55783a (2013-04-26).
gcc/ada/init.c: sigaltstack (&stack, NULL);
gcc/ada/init.c: sigaltstack (&stack, NULL);
gcc/ada/init.c: sigaltstack (&stack, NULL);
gcc/ada/s-osinte-aix.ads: function sigaltstack
gcc/ada/s-osinte-aix.ads: pragma Import (C, sigaltstack, "sigaltstack");
gcc/ada/s-osinte-android.ads: function sigaltstack
gcc/ada/s-osinte-android.ads: pragma Import (C, sigaltstack, "sigaltstack");
gcc/ada/s-osinte-darwin.ads: function sigaltstack
gcc/ada/s-osinte-darwin.ads: pragma Import (C, sigaltstack, "sigaltstack");
gcc/ada/s-osinte-freebsd.ads: function sigaltstack
gcc/ada/s-osinte-freebsd.ads: pragma Import (C, sigaltstack, "sigaltstack");
gcc/ada/s-osinte-hpux.ads: function sigaltstack
gcc/ada/s-osinte-hpux.ads: pragma Import (C, sigaltstack, "sigaltstack");
gcc/ada/s-osinte-kfreebsd-gnu.ads: function sigaltstack
gcc/ada/s-osinte-kfreebsd-gnu.ads: pragma Import (C, sigaltstack, "sigaltstack");
gcc/ada/s-osinte-linux.ads: function sigaltstack
gcc/ada/s-osinte-linux.ads: pragma Import (C, sigaltstack, "sigaltstack");
gcc/ada/s-osinte-rtems.adb: -- sigaltstack --
gcc/ada/s-osinte-rtems.adb: function sigaltstack
gcc/ada/s-osinte-rtems.adb: end sigaltstack;
gcc/ada/s-osinte-rtems.ads: function sigaltstack
gcc/ada/s-osinte-solaris-posix.ads: function sigaltstack
gcc/ada/s-osinte-solaris-posix.ads: pragma Import (C, sigaltstack, "sigaltstack");
gcc/ada/s-taprop-linux.adb: Result := sigaltstack (Stack'Access, null);
gcc/ada/s-taprop-posix.adb: Result := sigaltstack (Stack'Access, null);
gcc/ada/init.c: stack.ss_sp = __gnat_alternate_stack;
gcc/ada/init.c: stack.ss_sp = __gnat_alternate_stack;
gcc/ada/init.c: stack.ss_sp = __gnat_alternate_stack;
gcc/ada/s-osinte-aix.ads: ss_sp : System.Address;
gcc/ada/s-osinte-android.ads: ss_sp : System.Address;
gcc/ada/s-osinte-darwin.ads: ss_sp : System.Address;
gcc/ada/s-osinte-darwin.ads: uc_stack : stack_t; -- Stack Used By This Context
gcc/ada/s-osinte-freebsd.ads: ss_sp : System.Address;
gcc/ada/s-osinte-hpux.ads: ss_sp : System.Address;
gcc/ada/s-osinte-kfreebsd-gnu.ads: ss_sp : System.Address;
gcc/ada/s-osinte-linux.ads: ss_sp : System.Address;
gcc/ada/s-osinte-rtems.ads: ss_sp : System.Address;
gcc/ada/s-osinte-solaris-posix.ads: ss_sp : System.Address;
gcc/ada/s-osinte-solaris.ads: ss_sp : System.Address;
gcc/ada/s-osinte-solaris.ads: uc_stack : record_type_2;
gcc/ada/s-taprop-linux.adb: Stack.ss_sp := Self_ID.Common.Task_Alternate_Stack;
gcc/ada/s-taprop-posix.adb: Stack.ss_sp := Self_ID.Common.Task_Alternate_Stack;
gcc/ada/init.c: act.sa_flags |= SA_ONSTACK;
gcc/ada/init.c: act.sa_flags |= SA_ONSTACK;
gcc/ada/init.c: act.sa_flags = SA_SIGINFO | SA_ONSTACK;
gcc/ada/init.c: act.sa_flags |= SA_ONSTACK;
gcc/ada/s-intman-posix.adb: act.sa_flags := act.sa_flags + SA_ONSTACK;
gcc/ada/s-linux-alpha.ads: SA_ONSTACK : constant := 16#01#;
gcc/ada/s-linux-hppa.ads: SA_ONSTACK : constant := 16#01#;
gcc/ada/s-linux-mipsel.ads: SA_ONSTACK : constant := 16#08000000#;
gcc/ada/s-linux-sparc.ads: SA_ONSTACK : constant := 16#001#;
gcc/ada/s-linux.ads: SA_ONSTACK : constant := 16#08000000#;
gcc/ada/s-osinte-aix.ads: SA_ONSTACK : constant := 16#0001#;
gcc/ada/s-osinte-android.ads: SA_ONSTACK : constant := System.Linux.SA_ONSTACK;
gcc/ada/s-osinte-darwin.ads: SA_ONSTACK : constant := 16#0001#;
gcc/ada/s-osinte-freebsd.ads: SA_ONSTACK : constant := 16#0001#;
gcc/ada/s-osinte-hpux-dce.ads: SA_ONSTACK : constant := 16#01#;
gcc/ada/s-osinte-hpux.ads: SA_ONSTACK : constant := 16#01#;
gcc/ada/s-osinte-kfreebsd-gnu.ads: SA_ONSTACK : constant := 16#0001#;
gcc/ada/s-osinte-linux.ads: SA_ONSTACK : constant := System.Linux.SA_ONSTACK;
gcc/ada/s-osinte-lynxos.ads: SA_ONSTACK : constant := 16#00#;
gcc/ada/s-osinte-lynxos.ads: -- SA_ONSTACK is not defined on LynxOS, but it is referred to in the POSIX
gcc/ada/s-osinte-rtems.ads: SA_ONSTACK : constant := 16#00#;
gcc/ada/s-osinte-rtems.ads: -- SA_ONSTACK is not defined on RTEMS, but it is referred to in the POSIX
gcc/ada/s-osinte-solaris-posix.ads: SA_ONSTACK : constant := 16#0001#;
gcc/ada/s-osinte-vxworks.ads: SA_ONSTACK : constant := 16#0004#;
TODO
diff --git ./gcc/ada/init.c ./gcc/ada/init.c
@@ -2002,6 +2021,15 @@ __gnat_install_handler (void)
trap_0_entry->inst_fourth = 0xa1480000;
#endif
{+#if (defined (i386) || defined (__i386__)) && !defined (VTHREADS)+}
{+ /* By experiment, found that sysModel () returns the following string+}
{+ prefix for vxsim when running on Linux and Windows. */+}
{+ model = sysModel ();+}
{+ if ((strncmp (model, "Linux", 5) == 0)+}
{+ || (strncmp (model, "Windows", 7) == 0))+}
{+ is_vxsim = 1;+}
{+#endif+}
__gnat_handler_installed = 1;
}
diff --git ./gcc/ada/sigtramp.h ./gcc/ada/sigtramp.h
@@ -62,8 +62,14 @@ typedef struct ucontext
system headers so call it something unique. */
typedef void __sigtramphandler_t (int signo, void *siginfo, void *sigcontext);
{+#if CPU == SIMNT || CPU == SIMPENTIUM || CPU == SIMLINUX+}
{+ /* Vxsim requires a specially compiled handler. */+}
{+ void __gnat_sigtramp_vxsim (int signo, void *siginfo, void *sigcontext,+}
{+ __sigtramphandler_t * handler);+}
{+#else+}
void __gnat_sigtramp (int signo, void *siginfo, void *sigcontext,
__sigtramphandler_t * handler);
{+#endif+}
diff --git ./gcc/ada/gcc-interface/Makefile.in ./gcc/ada/gcc-interface/Makefile.in
@@ -923,13 +928,49 @@ ifeq ($(strip $(filter-out %86 wrs vxworks vxworks7,$(target_cpu) $(target_vendo
{+ VXSIM_CPU =+}
{+ ifeq ($(strip $(filter-out vxworks rtp rtp-smp,$(target_os) $(THREAD_KIND))),)+}
{+ VXSIM_CPU = SIMPENTIUM+}
{+ else+}
{+ ifeq ($(strip $(filter-out kernel kernel-smp rtp rtp-smp,$(THREAD_KIND))),)+}
{+ ifeq ($(strip $(filter-out linux%,$(host_os))),)+}
{+ # Linux+}
{+ VXSIM_CPU = SIMLINUX+}
{+ else+}
{+ # Windows+}
{+ VXSIM_CPU = SIMNT+}
{+ endif+}
{+ endif+}
diff --git ./gcc/ada/gcc-interface/Makefile.in ./gcc/ada/gcc-interface/Makefile.in
@@ -2413,6 +2468,14 @@ ifeq ($(filter a-except%,$(LIBGNAT_TARGET_PAIRS)),)
{+# Configuration of host tools+}
{+# Under linux, host tools need to be linked with -ldl+}
{+ifeq ($(strip $(filter-out linux%,$(host_os))),)+}
{+ TOOLS1_LIBS=-ldl+}
{+endif+}
diff --git ./gcc/ada/s-osinte-linux.ads ./gcc/ada/s-osinte-linux.ads
@@ -224,6 +224,10 @@ package System.OS_Interface is
{+function clock_gettime+}
{+ (clock_id : clockid_t; tp : access timespec) return int;+}
{+ pragma Import (C, clock_gettime, "clock_gettime");+}
diff --git ./gcc/ada/s-taprop-linux.adb ./gcc/ada/s-taprop-linux.adb
@@ -39,7 +39,6 @@ pragma Polling (Off);
with Interfaces.C;
[-with Interfaces.C.Extensions;-]
@@ -64,7 +63,6 @@ package body System.Task_Primitives.Operations is
use Interfaces.C;
[- use Interfaces.C.Extensions;-]
@@ -629,30 +627,14 @@ package body System.Task_Primitives.Operations is
function Monotonic_Clock return Duration is
[-use Interfaces;-]
[- procedure timeval_to_duration-]
[- (T : not null access timeval;-]
[- sec : not null access C.Extensions.long_long;-]
[- usec : not null access C.long);-]
[- pragma Import (C, timeval_to_duration, "__gnat_timeval_to_duration");-]
[- Micro : constant := 10**6;-]
[- sec : aliased C.Extensions.long_long;-]
[- usec : aliased C.long;-]
[- TV-]{+TS+} : aliased [-timeval;-]{+timespec;+}
Result : int;
[- function gettimeofday-]
[- (Tv : access timeval;-]
[- Tz : System.Address := System.Null_Address) return int;-]
[- pragma Import (C, gettimeofday, "gettimeofday");-]
begin
Result := [-gettimeofday (TV'Access, System.Null_Address);-]{+clock_gettime+}
{+ (clock_id => OSC.CLOCK_RT_Ada, tp => TS'Unchecked_Access);+}
pragma Assert (Result = 0);
[-timeval_to_duration (TV'Access, sec'Access, usec'Access);-]
return [-Duration (sec) + Duration (usec) / Micro;-]{+To_Duration (TS);+}
end Monotonic_Clock;
diff --git ./gcc/ada/adaint.c ./gcc/ada/adaint.c
@@ -3220,6 +3220,107 @@ __gnat_kill (int pid, int sig, int close ATTRIBUTE_UNUSED)
#endif
}
{+void __gnat_killprocesstree (int pid, int sig_num)+}
{+{+}
{+[...]+}
{+#elif defined (__linux__)+}
{+ DIR *dir;+}
{+ struct dirent *d;+}
{+ /* read all processes' pid and ppid */+}
{+ dir = opendir ("/proc");+}
{+[...]+}
{+ /* kill process */+}
{+ __gnat_kill (pid, sig_num, 1);+}
{+#else+}
{+ __gnat_kill (pid, sig_num, 1);+}
{+#endif+}
{+[...]+}
{+}+}
diff --git ./gcc/ada/s-os_lib.ads ./gcc/ada/s-os_lib.ads
@@ -740,6 +744,19 @@ package System.OS_Lib is
{+ procedure Kill_Process_Tree (Pid : Process_Id; Hard_Kill : Boolean := True);+}
{+ -- Kill the process designated by Pid and all it's children processes.+}
{+ -- [...]
{+ -- Note that this routine is not atomic and is supported only on Linux+}
{+ -- and Windows. On other OS it will only kill the process identified by+}
{+ -- Pid.+}
GCC includes (TODO: some version of) the Ada Conformity Assessment Test Suite (ACATS), http://ada-auth.org/acats.html. Additional tests to be found at http://ada-auth.org/submit.html. TODO: results?
Part II
Next, Hurd-specific features can be added. Add an interface to the language/environment for being able to do RPC calls, in order to program translators natively in Ada.
Original GSoC Task Description
The main idea of the Hurd design is giving users the ability to easily modify/extend the system's functionality (extensible system). This is done by creating filesystem translators and other kinds of Hurd servers.
However, in practice this is not as easy as it should, because creating translators and other servers is quite involved -- the interfaces for doing that are not exactly simple, and available only for C programs. Being able to easily create simple translators in RAD languages is highly desirable, to really be able to reap the advantages of the Hurd architecture.
Originally Lisp was meant to be the second system language besides C in the GNU system; but that doesn't mean we are bound to Lisp. Bindings for any popular high-level language, that helps quickly creating simple programs, are highly welcome.
Several approaches are possible when creating such bindings. One way is simply to provide wrappers to all the available C libraries (libtrivfs, libnetfs etc.). While this is easy (it requires relatively little consideration), it may not be the optimal solution. It is preferable to hook in at a lower level, thus being able to create interfaces that are specially adapted to make good use of the features available in the respective language.
These more specialized bindings could hook in at some of the lower level library interfaces (libports, glibc, etc.); use the MIG-provided RPC stubs directly; or even create native stubs directly from the interface definitions. The lisp bindings created by Flavio Cruz as his 2008 GSoC project mostly use the latter approach, and can serve as a good example. In his 2011 GSoC project, Jérémie Koenig designed and began implementing an object-oriented interface; see his Java status page for details.
The task is to create easy to use Hurd bindings for a language of the student's choice, and some example servers to prove that it works well in practice. This project will require gaining a very good understanding of the various Hurd interfaces. Skills in designing nice programming interfaces are a must.
Anatoly A. Kazantsev has started working on Python bindings last year -- if Python is your language of choice, you probably should take his work and complete it.
There was also some previous work on Perl bindings, which might serve as a reference if you want to work on Perl.
Possible mentors: Anatoly A. Kazantsev (anatoly) for Python
Discussion
Java
IRC, freenode, #hurd, 2013-12-19
<antrik_> teythoon_: I don't think wrapping libtrivfs etc. for guile
bindings is really desirable... for the lisp bindings, we agreed that
it's better to hook in at a lower level, and build more lispish
abstractions
<antrik> trivfs is a C framework; it probably doesn't map very well to
other languages -- especially non-imperative ones...
<antrik> (it is arguable whether trivfs is really a good abstraction even
for C... but that's another discussion :-) )
<antrik> ArneBab: same for Python bindings. when I suggested ignoring
libtrivfs etc., working around the pthread problem was just a side effect
-- the real goal has always been nicer abstraction
<anatoly> antrik: agree with you
<anatoly> antrik: about nicer abstractions
<teythoon_> antrik: I agree too, but wrapping libtrivfs is much easier
<teythoon_> otherwise, one needs to reimplement lots of stuff to get some
basic functionality
<teythoon_> like a mig that emits your language
<braunr> i agree with antrik too
<braunr> yes, the best would be mig handling multiple languages
<antrik> teythoon_: not exactly. for dynamic languages, code generation is
silly. just handle the marshalling on the fly. that's what the Lisp
bindings are doing (AFAIK)
<teythoon> antrik: ok, but you'd still need to parse the rpc definitions,
no?
<antrik> teythoon: yeah, you still need to parse the .defs -- unless we add
reflection to RPC interfaces...
<antrik> err, I mean introspection