Commit Graph

1975 Commits

Author SHA1 Message Date
Edmund Grimley Evans
f75f89fc8f tccgen.c: In parse_btype, handle type qualifiers applied to arrays.
Also add some test cases in tests/tests2/39_typedef.c.
2016-01-11 07:51:58 +00:00
Edmund Grimley Evans
541b33591c CodingStyle: Remove reference to misaligned struct CString.
This was fixed by 1c2dfa1 on 2015-11-21.
2016-01-06 19:09:28 +00:00
Thomas Preud'homme
933c2235e5 i386: Add support for new psABI relocation
R_386_GOT32X can occur in object files assembled by new binutils, and in
particular do appear in glibc startup code (crt*.o). This patch is
modeled after the x86_64 one, handling the new relocation in the same
trivial way.
2015-12-27 12:09:45 +08:00
Michael Matz
f15c0a9333 x86-64: fix shared libs
The introduction of read32le everywhere created a subtle issue, going
from
   x = *(int*)p;
to
   x = read32le(p);
is not equivalent if x is a larger than 32bit quantity, like an
address on x86_64, because read32le returns an unsigned int.  The first
sign extends, the latter zero extends.  This broke shared library
creation for gawk.  It's enough to amend the case of the above
situation, cases like "write32le(p, read32le(p) +- something)" are okay,
no extensions happen or matter.
2015-12-17 19:41:20 +01:00
Michael Matz
e264243adc x86-64: Define symbol constant for new relocs
Whoops, we have our own <elf.h> copy, so I can just as well add
the symbol defines for the relocs instead of hard-coding numbers
in tccelf.c.
2015-12-17 07:30:35 +01:00
Michael Matz
c4d0498b3a x86-64: Add support for new psABI relocations
R_X86_64_GOTPCRELX and R_X86_64_REX_GOTPCRELX can occur in object files
comiled by new binutils.  They are not dynamic relocations, so normally
wouldn't be a problem for tcc (one doesn't normally mix object files
created by different compiler/binutils, static archives are so out :)).
If it weren't for the glibc startup code, crt*.o, of course.  They now
do contain such relocs --> boom.  Handle them in the trivial way.
2015-12-17 07:17:34 +01:00
Vincent Lefevre
d1e15514aa Fixed a dependency (error with make -j8).
Signed-off-by: Vincent Lefevre <vincent@vinc17.net>
2015-12-15 13:47:07 +01:00
Edmund Grimley Evans
1c2dfa1f4b Change the way struct CStrings are handled.
A CString used to be copied into a token string, which is an int array.
On a 64-bit architecture the pointers were misaligned, so ASan gave
lots of warnings. On a 64-bit architecture that required memory
accesses to be correctly aligned it would not work at all.

The CString is now included in CValue instead.
2015-11-26 12:40:50 +00:00
Edmund Grimley Evans
4886d2c640 TODO: Add two issues. 2015-11-26 12:31:23 +00:00
Edmund Grimley Evans
99372bb1d3 tccgen.c: Give error if statement expression found when const wanted.
Some test cases:

#define SE ({ switch (0) { } 0; })

// Should give error:
int x = SE;
void f(void) { static int x = SE; }
void f(void) { enum e { a = SE }; }
void f(void) { switch (0) { case SE: break; } }

// Correct:
int f(void) { return SE; }
int f(void) { return sizeof(SE); }
2015-11-26 12:28:42 +00:00
Edmund Grimley Evans
51c3465a49 tcc.h: Change order of built-in include paths: put TCC's own dir first.
There may be compiler-specific header files that should override
system headers. See TCC's include paths by running "tcc -vv".
2015-11-23 12:50:16 +00:00
Edmund Grimley Evans
7301b42e36 tccgen.c: Try to make sizeof(!x) work.
tests/tests2/27_sizeof.*: Add test.
2015-11-22 00:00:36 +00:00
Edmund Grimley Evans
737f984213 tccgen.c: Bug fix for 992cbda and 3ff77a1: set nocode_wanted.
tests/tests2/78_vla_label.*: Add test.
2015-11-21 23:58:58 +00:00
Edmund Grimley Evans
cfef9ac3f5 TODO: Add note on handling of floating-point values. 2015-11-21 10:35:47 +00:00
Edmund Grimley Evans
dd40d6a068 TODO: Add some issues. 2015-11-21 00:04:58 +00:00
Edmund Grimley Evans
5bd5fd488d CodingStyle: Add notes on language and testing. 2015-11-20 23:41:01 +00:00
Edmund Grimley Evans
3ff77a1d6f Improve constant propagation with "&&" and "||". 2015-11-20 23:33:49 +00:00
Edmund Grimley Evans
c7067aeb84 tccelf.c: On arm64, use read64le, and use uint64_t to check range. 2015-11-20 23:29:08 +00:00
Edmund Grimley Evans
4ae626451e Bug fix for commit 553242c18a.
In gtst, vtop->c.i is not usually zero, but it is when compiling:

int f(void) { return 1 && 1 ? 1 : 1; }
2015-11-20 23:17:24 +00:00
grischka
8dd1859176 tccpp: allow .. in token stream
for gas comments lonely on a line such as

    # .. more stuff

where tcc would try to parse .. as a preprocessor directive

See also: 0b3612631f
2015-11-20 18:25:00 +01:00
grischka
0b3612631f tccpp: cleanup #include_next
tcc_normalize_inc_dirs: normally no problem to be absolutly
gcc compatible as long as it can be done the tiny way.

This reverts to the state before recent related commits and
reimplements a (small) part of it to fix the reported problem.


Also: Revert "parsing "..." sequence"
c3975cf27c

	&& p[1] == '.'

is not a reliable way to lookahead
2015-11-20 12:05:55 +01:00
grischka
54cf57ab1a tccgen: asm_label cleanup
- avoid memory allocation by using its (int) token number
- avoid additional function parameter by using Attribute

Also: fix some strange looking error messages
2015-11-20 11:22:56 +01:00
Edmund Grimley Evans
992cbda8d0 tccgen.c: Recognise constant expressions with conditional operator.
tests/tests2/78_vla_label.c: Check that int a[1 ? 1 : 1] is not a VLA.
2015-11-20 00:24:46 +00:00
Edmund Grimley Evans
30c54c9d43 tccgen.c: In parse_btype, handle typedef types with added type qualifiers.
In a case like

    typedef int T[1];
    const T x;

we must make a copy of the typedef type so that we can add the type
qualifiers to it.

The following code used to give

error: incompatible types for redefinition of 'f'

    typedef int T[1];
    void f(const int [1]);
    void f(const T);
2015-11-19 23:45:33 +00:00
Edmund Grimley Evans
58a34d22c9 tccgen.c: Improvements to type_to_str (only used for error messages).
1. Handle array types.
2. Print the type qualifiers of pointers.
2015-11-19 23:35:36 +00:00
Edmund Grimley Evans
ba99a70cd8 Trivial changes to avoid some compiler warnings. 2015-11-19 18:26:47 +00:00
Edmund Grimley Evans
553242c18a Replace pointer casts with calls to (read|write)(16|32|64)le.
This stops UBSan from giving runtime misaligned address errors
and might eventually allow building on a non-little-endian host.
2015-11-19 18:21:14 +00:00
Edmund Grimley Evans
5d496b1695 tccgen.c: Avoid undefined behaviour in constant propagation. 2015-11-17 19:34:31 +00:00
Edmund Grimley Evans
569fba6db9 Merge the integer members of union CValue into "uint64_t i". 2015-11-17 19:09:35 +00:00
Edmund Grimley Evans
3712c958f4 tests/tests2/79_vla_continue.c: Fix off-by-one error. 2015-11-13 21:49:29 +00:00
Edmund Grimley Evans
c39bc9caa7 libtcc.c: Completely replace tcc_normalize_inc_dirs. 2015-11-11 21:18:01 +00:00
Edmund Grimley Evans
a7334f791d arm64-gen.c: Avoid some cases of undefined behaviour.
Also make some functions more portable.
2015-11-09 23:06:05 +00:00
Edmund Grimley Evans
c52128c581 tccelf.c: Avoid two trivial instances of undefined behaviour. 2015-11-09 22:57:58 +00:00
Avi Halachmi (:avih)
951c23f257 win: include dirs: add some docs, minor refactor 2015-11-08 16:28:00 +02:00
seyko
37e815eee1 lib/Makefile: filter-out -b flag from XFLAGS
to build tcc with bound checking
    ./configure --cc=tcc --extra-cflags-b
2015-11-08 12:26:17 +03:00
seyko
763dd22b35 fix tcc_mormalize_inc_dirs naming 2015-11-08 11:42:56 +03:00
Avi Halachmi (:avih)
ec67dc482f win: fix incorrect directory detection
Allow other flags too...
2015-11-08 04:33:30 +02:00
Avi Halachmi (:avih)
acbab220c8 win: fix dir comparison during include dirs processing
Use native Windows API via wrapper with stat-like API.
2015-11-08 03:56:25 +02:00
Avi Halachmi (:avih)
9d33388b29 win: libm: add implementation for round/fmin/fmax and variants
round and fmin/fmax are relatively commonly used functions but were not
implemented anywhere in the tcc Windows distribution package. Newer mingw(64)
math.h stil doesn't include these implementations.

Add C implementations for these functions and place it as inline functions at
win32/include/tcc/tcc_libm.h - which is already included from math.h .

The code is mostly taken from musl-libc rs-1.0 (MIT) [*],

musl-libc: http://git.musl-libc.org/cgit/musl/tree/src/math?h=rs-1.0
license:   http://git.musl-libc.org/cgit/musl/tree/COPYRIGHT?h=rs-1.0

Potential enhancements:
- Check how many useful libm implementations are still missing and consider
  adding them (some of them already work via the MS runtime).
- Consider putting libm implementations in an actual libm.a file, or add a dummy
  one such that build processes which try to link with libm will not fail.
2015-11-07 22:31:57 +02:00
Avi Halachmi (:avih)
9c52ba48b3 win: math.h: fix fpclassify/signbit/etc - use C instead of broken asm
The asm code cannot currently be used with tcc since tcc doesn't support 't'
constraint.

Use inline C implementation instead, place it win32/include/tcc/tcc_libm.h, and
include it from win32/include/math.h.

Since fpclassify now works, it also fixes few other macros which depend on it.
Implicitly fixed: isfinite, isinf, isnan, isnormal.

The implementations were taken from musl-libc rs-1.0 (MIT license).

musl-libc: http://git.musl-libc.org/cgit/musl/tree/src/math?h=rs-1.0
license:   http://git.musl-libc.org/cgit/musl/tree/COPYRIGHT?h=rs-1.0
2015-11-07 22:18:46 +02:00
Avi Halachmi (:avih)
6e261a107c win: math.h: isnan: use macro, similar to others (still broken)
It was broken due to tcc not able to compile asm with 't' constraint, and it's
still broken because fpclassify on which it now depends has the same issue. Next
commit will fix this.
2015-11-07 22:17:21 +02:00
Avi Halachmi (:avih)
7307a6a3cb Revert "win32/include/math.h: remoing a "t" modifier usage"
This reverts commit 45bc505968.

The new asm code did not work as expected. Coordinated with seiko.
2015-11-07 21:30:56 +02:00
seyko
8fc9c79705 TOK_INCLUDE: fix for the "normalize inc dirs"
A case for the absolute path: prevent an error after openening
2015-11-06 02:50:36 +03:00
seyko
7cb921a44b TOK_INCLUDE: streamline
goto removed
2015-11-06 02:40:14 +03:00
Edmund Grimley Evans
eb00777309 tcctok.h: Put TOK_memmove in the correct places (I hope).
This should have been part of b051549. Someone should test on ARM.
2015-11-05 19:37:04 +00:00
Edmund Grimley Evans
7f0b798418 tcctok.h: Revert 41408f2, which moved TOK_memmove. 2015-11-05 19:33:46 +00:00
seyko
97916c9d04 tcc.h: remove CONFIG_TCCBOOT part
curremtly no one will try to compile a linux kernel
    from the boot loader. With current tcc it is not
    possible w/o additional tuning.
2015-11-05 20:30:51 +03:00
seyko
41408f2104 fix for the "tccgen.c: Use memmove for struct assignment"
./configure --enable-cross
    make
    tcc -o i386-tcc tcc.c
    tcc -o x86_64-tcc tcc.c
    tcc -o i386-win-tcc tcc.c
    In file included from tcc.c:22:
    In file included from libtcc.c:39:
    tccgen.c:2580: error: 'TOK_memmove' undeclared
2015-11-05 20:24:04 +03:00
seyko
0ad87a094c fix for the previous commit
a cross-compilation from unix to win32 need a "sys/stat.h"
    include file
2015-11-05 20:14:42 +03:00
seyko
a6276b7a78 normalize inc dirs, symplify include_next
include dirs are prepared as in gcc
    - for each duplicate path keep just the first one
    - remove each include_path that exists in sysinclude_paths

    include_next streamlined by introducing inc_path_index
    in the BufferedFile
2015-11-05 19:52:49 +03:00