Commit Graph

1368 Commits

Author SHA1 Message Date
Michael Matz
b0f8ca5e03 Git should ignore tests2 executables. 2014-04-05 22:52:17 +02:00
Michael Matz
3d18c9aa64 tests2: Build executables as well
The individual tests in tests2 are checked only with -run.  Build
(and check) executables as well, to test also building executables.
2014-04-05 17:35:00 +02:00
Michael Matz
0688afdd34 arm: Handle R_ARM_NONE relocs
These relocations are used to express a dependency on a certain
symbol (e.g. for EABIs exception handling to the
__aeabi_unwind_cpp_pr{0,1,2} routines).  Just ignore them in
reloc processing.
2014-04-04 23:33:04 +02:00
grischka
5879c854fb tccgen: x86_64: fix garbage in the SValue upper bits
This was going wrong (case TOK_LAND in unary: computed labels)
-        vset(&s->type, VT_CONST | VT_SYM, 0);
-        vtop->sym = s;

This does the right thing and is shorter:

+        vpushsym(&s->type, s);


Test case was:

    int main(int argc, char **argv)
    {
        int x;
        static void *label_return = &&lbl_return;
        printf("label_return = %p\n", label_return);
        goto *label_return; //<<<<< here segfault on linux X86_64 without the memset on vset
        printf("unreachable\n");
    lbl_return:
        return 0;
    }


Also::
- Rename "void* CValue.ptr" to more usable "addr_t ptr_offset"
  and start to use it in obvious cases.

- use __attribute__ ((noreturn)) only with gnu compiler

- Revert CValue memsets ("After several days searching ...")
  commit 4bc83ac393

Doesn't mean that the vsetX/vpush thingy isn't brittle and
there still might be bugs as to differences in how the CValue
union  was set and is then interpreted later on.

However the big memset hammer was just too slow (-3% overall).
2014-04-04 20:20:44 +02:00
Michael Matz
2024c44541 run: Always create .got relocs
When output is memory we applied the correct GOT offset for certain
relocations (e.g. _GOT32), but we forgot to actually fill the got
entries with the final symbol values, so unconditionally create relocs
against .got as well.
2014-04-04 17:54:52 +02:00
Michael Matz
f2c8491fc0 ELF: Make first PT_LOAD cover headers
This makes it so that the first PT_LOAD segment covers
ELF and program header and .interp (contained in the same page anyway,
right before the start of the first loaded section).  binutils
strip creates invalid output otherwise (which strictly is a binutils
bug, but let's be nice anyway).
2014-04-03 18:00:44 +02:00
Michael Matz
a913ee6082 x86-64: Use correct ELF values
The x86-64 uses different segment alignment (2MB) and a different
start address.
2014-04-03 17:59:41 +02:00
Michael Matz
ea2805f097 shared libs: Build libtcc1.a with -fPIC
TCCs runtime library must be compiled as position independend code,
so it can be linked into shared libraries.
2014-04-02 21:27:22 +02:00
Thomas Preud'homme
3e56584223 Allow local redefinition of enumerator 2014-03-31 22:59:10 +08:00
Vincent Lefevre
a620b12dc1 Fixed typo from commit 0ac8aaab1b 2014-03-31 15:24:32 +02:00
Thomas Preud'homme
01cf514e59 Update Changelog from git changelog entries 2014-03-31 19:42:42 +08:00
Michael Matz
0bd1282059 x86-64: shared libs improvement
This correctly resolves local references to global functions from
shared libs to their PLT slot (instead of directly to the target
symbol), so that interposition works.

This is still not 100% conforming (executables don't export symbols
that are also defined in linked shared libs, as they must), but
normal shared lib situations work.
2014-03-31 05:36:12 +02:00
Michael Matz
080ad7e62a x86-64: Add basic shared lib support
Initial support for shared libraries on x86-64.
2014-03-31 03:45:35 +02:00
mingodad
5a5fee867a Add __attribute__ ((noreturn)) to tcc_error and expect functions.
This make use of static analysis tools like scan-build report less false positives.
2014-03-30 10:18:18 +01:00
Thomas Preud'homme
80811671d4 Add tests for previous fixes
Add tests for the fixes made in commits
76cb1144ef,
a465b7f58f,
0f522fb32a,
82969f045c and
673befd2d7.
2014-03-30 12:56:55 +08:00
Thomas Preud'homme
fdc31e152b Update Changelog from git changelog entries 2014-03-30 12:35:22 +08:00
Daniel Glöckner
3900b235e0 x86_64: pass va_list as pointer
The ABI requires that va_list is passed as a pointer although its
contents is a kept in a structure. Therefore make it a single element
array.
2014-03-30 00:13:58 +01:00
grischka
0ac8aaab1b tccpp: reorder some tokens
... and make future reordering possibly easier

related to 9a6ee577f6
2014-03-29 19:37:26 +01:00
Daniel Glöckner
1075087241 ARM: Fix passing arrays to varadic functions
TinyCC miscompiled

void g(int,...);

void f(void)
{
        char b[4000];
        g(1, 2, 3, 4, b);
}

in two ways:
 1. It didn't align the stack to 8 bytes before the call
 2. It added sizeof(b) to the stack pointer after the call
2014-03-29 17:50:40 +01:00
Thomas Preud'homme
f272407353 Fix typo in code added by b018bac9c8 2014-03-29 14:57:59 +08:00
Thomas Preud'homme
b018bac9c8 Fix again GOT32 + PLT32 reloc commit
Fix commit aa561d7011 by setting
has_plt_entry once the plt has been created, not before.
2014-03-29 14:46:26 +08:00
Thomas Preud'homme
b125743323 Create bcheck region for argv and arge argument
For program manipulating argv or arge as pointer with construct such as:

(while *argv++) {
  do_something_with_argv;
}

it is necessary to have argv and arge inside a region. This patch create
regions argv and arge) if main is declared with those parameters.
2014-03-29 14:46:26 +08:00
Thomas Preud'homme
9a6ee577f6 Make get_tok_str support NULL as second param.
As was pointed out on tinycc-devel, many uses of get_tok_str gives as
second parameter the value NULL. However, that pointer was
unconditionally dereferenced in get_tok_ptr. This commit explicitely add
support for thas case.
2014-03-29 14:46:26 +08:00
mingodad
5c233f2cf3 The hack to allow valgrind works with tcc compiled programs
have the undesired side effect of programs compiled with
debug info segfaulting after debug info been striped
more tought must be done here
2014-03-28 21:07:06 +00:00
mingodad
c025478d7c New implementation of va_list/va_start/var_copy that do not use dynamic memory, with this when compiling fossil-scm with tcc on linux X86_64 it works fine. 2014-03-28 20:28:19 +00:00
mingodad
700c2f769b Remove the fix from my last commit, it was pointed by scan-build and is a false positive, thanks to grischka for pointing it. 2014-03-28 20:25:39 +00:00
mingodad
0ba7c8670c This allow valgrind to work on linux, some how the PHDR is missing and then valgrind complain with:
Inconsistency detected by ld.so: rtld.c: 1284: dl_main: Assertion `_rtld_local._dl_rtld_map.l_libname' failed!
2014-03-28 11:44:00 +00:00
mingodad
14bb8302c4 Fix a incorrect size for malloc. 2014-03-27 22:15:45 +00:00
mingodad
4bc83ac393 After several days searching why my code refactoring to remove globals was crashing,
I found the problem it was because CValue stack variables have rubish as it inital values
and assigning to a member that is smaller than the big union item and trying to
recover it later as a different member gives bak garbage.

ST_FUNC void vset(TCCState* tcc_state, CType *type, int r, int v)
{
    CValue cval;
    memset(&cval, 0, sizeof(CValue));

    cval.i = v; //,<<<<<<<<<<< here is the main bug that mix with garbage
    vsetc(tcc_state, type, r, &cval);
}

/* store a value or an expression directly in global data or in local array */
static void init_putv(TCCState* tcc_state, CType *type, Section *sec, unsigned long c,
                      int v, int expr_type)
{
...
        case VT_PTR:
            if (tcc_state->tccgen_vtop->r & VT_SYM) {
                greloc(tcc_state, sec, tcc_state->tccgen_vtop->sym, c, R_DATA_PTR);
            }

//<<< on the next line is where we try to get the assigned value to cvalue.i as cvalue.ull

            *(addr_t *)ptr |= (tcc_state->tccgen_vtop->c.ull & bit_mask) << bit_pos;
            break;

Also this patch makes vla tests pass on linux 32 bits
2014-03-26 20:18:48 +00:00
Thomas Preud'homme
aa561d7011 Simplify and fix GOT32 + PLT32 reloc commit
Introduce a new attribute to check the existence of a PLT entry for a
given symbol has the presence of an entry for that symbol in the dynsym
section is not proof that a PLT entry exists.

This fixes commit dc8ea93b13.
2014-03-26 23:13:28 +08:00
mingodad
bed865275d Add the generated executables ending with "-cc" and "-tcc" to the makefile "clean" 2014-03-26 14:32:03 +00:00
mingodad
ad9568060e A possible fix for the memory leak reported by valgrind when running tcctest.c with tcc. 2014-03-25 18:06:14 +00:00
Thomas Preud'homme
dc8ea93b13 Support GOT32 and PLT32 reloc for same symbol
Some symbol (such as __gmon_start__ but this one does not matter to tcc)
can have both a R_386_GOT32 and R_386_PLT32 relocation. It is thus not
enough to test if a GOT reloc was already done when deciding whether to
return early from put_got_entry.
2014-03-25 22:35:11 +08:00
Thomas Preud'homme
078ba241d9 Always link libtcc1.a in (useful for va_* on x86)
On x86 tcc call to function in libtcc1.a to implement va_* functions.
2014-03-25 21:18:57 +08:00
Thomas Preud'homme
f1f45a47ef Add test for previous commit
* Adapt tests2 Makefile to support testing tcc error reporting
* Add test for previous commit
2014-03-25 20:54:19 +08:00
Thomas Preud'homme
6f6ed8acc7 Warn about soft float ABI not being supported
For ARM target, tcc uses the soft float ABI when not asked to use hard
float ABI. This means machine without a VFP co-processor generate code
that they cannot run. This commit add a warning for such cases.
2014-03-25 19:58:12 +08:00
Thomas Preud'homme
b8610f14b0 Deprecate FPA and OABI support for ARM 2014-03-25 19:58:03 +08:00
Thomas Preud'homme
b68499e971 Make parse_btype only accept one basic type
This makes int char c; and struct {} int c; generate an error. Thanks
Mobi Phil for reporting.
2014-03-24 23:40:39 +08:00
Thomas Preud'homme
ec1c83081d Fix relocation of __bound_init
When bound check is enabled, tcc tries to relocate a call to
__bound_init in _init. This means that relocation (in tcc_add_bcheck)
must be done after libtcc1.a (which countains __bound_init) is loaded
but before crtn.o is loaded as this finalize _init.
2014-03-17 23:14:38 +08:00
Thomas Preud'homme
40e3859739 Fix __clear_cache implementation
Forgot to give the parameters to syscall function, doh!
2014-03-11 22:57:22 +08:00
Thomas Preud'homme
b2192fc50b Adjust relocation offset for thumb to ARM veneer 2014-03-11 21:45:52 +08:00
Thomas Preud'homme
d3d89900f6 Don't hardcode gcc in tests Makefile 2014-03-09 22:54:48 +08:00
Thomas Preud'homme
62d1da1b3e Fix warning of clang 2014-03-09 22:54:30 +08:00
Thomas Preud'homme
98afe11c85 Use intptr_t to cast pointer 2014-03-09 22:22:43 +08:00
Thomas Preud'homme
73ac39c317 Undefine __va* in libtcc1 to avoid errors w/ clang 2014-03-09 22:22:21 +08:00
Thomas Preud'homme
e50f08faa1 Make condition in libtcc1 based on target
Prior to this commit runtime library was compiled according to the host
because of the macro used to detec what architecture to choose. This
commit fixes this by using the TARGET_* macro instead.
2014-03-09 22:15:01 +08:00
Thomas Preud'homme
33cea54dc7 Fix type_to_str test for unsigned int 2014-03-09 13:32:40 +08:00
Thomas Preud'homme
fdb3b10d06 Fix various errors uncovered by static analysis
Reported-by: Carlos Montiers <cmontiers@gmail.com>
2014-03-08 18:38:49 +08:00
Austin English
ba286136bf libtcc: ignore linker optizimization and as-needed options. This allows compiling some packages from Gentoo's portage 2014-03-06 12:29:19 -08:00
Thomas Preud'homme
361ec4f98e Call fill_got_entry unconditionally
Call fill_got_entry unconditionally from fill_got so as to avoid
warnings on !x86-64 architectures. This can be done since this code path
is only followed by x86-64 architecture anyway.
2014-02-10 21:35:00 +08:00