Commit Graph

2282 Commits

Author SHA1 Message Date
Michael Matz
38e5cf0983 tccpp: Fix macro_is_equal
When tokens in macro definitions need cstr_buf inside get_tok_str,
the second might overwrite the first (happens when tokens are
multi-character non-identifiers, see testcase) in macro_is_equal,
failing to diagnose a difference.  Use a real local buffer.
2016-12-15 17:47:05 +01:00
Michael Matz
8080401ab0 tccpp: free defines also with PP_BENCH
When benchmarking preprocessing of multiple files we need to
free the defines like when not benchmarking.
2016-12-15 17:47:05 +01:00
Michael Matz
e2f489aaff x86-asm: Get rid of OPC_JMP and OPC_SHORTJMP
Those two insn types are nicer to handle as operand types, because
the pressure for bits on instr_type is higher than for operands.
2016-12-15 17:47:05 +01:00
Michael Matz
4094f7c5fc x86-64-asm: Tidy 2016-12-15 17:47:05 +01:00
Michael Matz
58963828ab x86-asm: Correct mem64->xmm movq
Now we can express prefixes with 0x0fxx opcodes we can correct the
movq mem64->xmm opcode, and restrict the movq xmm->mem64 movq to
not invalidly accept mmx.
2016-12-15 17:47:05 +01:00
Michael Matz
5a222588a8 x86-asm: Remove OPC_D16
Now that we can store prefixes even for 0x0fXX opcodes we can remove
the OPC_D16 bit.
2016-12-15 17:47:05 +01:00
Michael Matz
8a10a442ff x86-asm: Fix register order
Inserting random registers in the middle of the 8-blocks
breaks register assignment.
2016-12-15 17:47:05 +01:00
Michael Matz
bde802df29 x86-asm: Reorganize instr_type
Disjoint instruction types don't need to be a bit field, so
introduce an enumeration (3 bits).  Also the 0x0f prefix can
be expressed by a bit, doesn't need a byte in the opcode field.
That enables to encode further prefixes still in 16 bit.
To not have to touch all insns do some macro fiddling filtering
out a 0x0f byte in the second position.
2016-12-15 17:47:05 +01:00
Michael Matz
4af6e087dd x86-asm: move stats code
The old place (tccasm.c) didn't have access to the variables anymore
and was ifdefed out.  Move it to i386-asm.c.
2016-12-15 17:47:05 +01:00
Michael Matz
ed35ac841b x86-asm: Add more SSE2 instructions
In particular those that are extensions of existing mmx (or sse1)
instructions by a simple 0x66 prefix.  There's one caveat for
x86-64: as we don't yet correctly handle the 0xf3 prefix
the movq mem64->xmm is wrong (tested in asmtest.S).  Needs
some refactoring of the instr_type member.
2016-12-15 17:47:05 +01:00
grischka
2b7ee000cd tests: add .so/.dll creation test
Also remove bitfield test from tcctest.c because gcc
versions don't agree among each other.
2016-12-15 17:04:07 +01:00
grischka
ca92bfc3c6 tccelf: some linker cleanup
- generate and use SYM@PLT for plt addresses
- get rid of patch_dynsym_undef hack (no idea what it did on FreeBSD)
- use sym_attrs instead of symtab_to_dynsym
- special case for function pointers into .so on i386
- libtcc_test: test tcc_add_symbol with data object
- move target specicic code to *-link.c files
- add R_XXX_RELATIVE (needed for PE)
2016-12-15 17:01:22 +01:00
Thomas Preud'homme
fe6453f8f0 Use functions to get relocation info
MSVC does not support array designator so cannot compile source using
relocs_info. This commit replace the relocs_info array into a set of
functions, each returning the value given by a given field of the struct
reloc_info.
2016-12-10 18:14:10 +00:00
Thomas Preud'homme
d31226c873 Remove now useless pltoff_addend reloc info
Last use for pltoff_addend field of relocs_info array was removed in
commit 25927df3b7. It is now useless so
this commit removes it and all initialization related to it.
2016-12-10 18:13:23 +00:00
Thomas Preud'homme
ee2108d07d Add missing relocation info for C67 target
Fill in relocs_info table for C67 and fix R_C60_NUM value to really be
greater than all relocation values known to TCC.
2016-12-10 17:31:20 +00:00
Thomas Preud'homme
11747fe5d0 Error out in put_got_entry if no dynamic symbol 2016-12-10 19:22:02 +08:00
Thomas Preud'homme
dfed9babfc Allow PLT/GOT entry for weak static symbol 2016-12-10 19:12:36 +08:00
Thomas Preud'homme
0bf262864c Fix PLT creation for i386
i386 target does not have PC relative loads. Its ABI therefore require
ebx register to points to the GOT when executing a PLT entry. This means
that PLT entry cannot be used transparently, the compiler needs to
expect execution of a PLT entry to be able to use one, that is a PLT
entry should only be created if the relocation explicitely asks for it
(eg. R_386_PLT32).

This patch creates a new target macro PCRELATIVE_DLLPLT to indicate
whether a target can do a PC relative load in PLT entry when building a
dynamic library. Executable do not normally pose a problem because they
are loaded at a fixed address and thus the absolute address of GOT can
be used.

Note that in such a case, if the compiler does not use a PLT aware
relocation for external access then the code relocation will fall on the
dynamic loader since there is no PLT entry to relocate too.
2016-12-10 09:44:09 +00:00
Edmund Grimley Evans
e0fe69050d arm64: Fix regression introduced by 6245db9. 2016-12-05 23:29:25 +00:00
Thomas Preud'homme
2372639e9d Fix set but not used error in arm64-link.c 2016-12-05 21:34:30 +00:00
Thomas Preud'homme
3811794048 Fix tcc_error params for R_AARCH64_(JUMP|CALL)26 2016-12-05 21:34:30 +00:00
Thomas Preud'homme
e22249b81c Error on unrecognized relocations 2016-12-05 20:58:00 +00:00
Thomas Preud'homme
557c5c1f11 Add relocs_info array to c67 backend 2016-12-05 20:52:02 +00:00
Thomas Preud'homme
59391d5520 Fix relocs_info declaration in tcc.h
C standard specifies that array should be declared with a non null size
or with * for standard array. Declaration of relocs_info in tcc.h was
not respecting this rule. This commit add a R_NUM macro that maps to the
R_<ARCH>_NUM macros and declare relocs_info using it. This commit also
moves all linker-related macros from <arch>-gen.c files to <arch>-link.c
ones.
2016-12-05 20:51:10 +00:00
Thomas Preud'homme
097cf3aa5e Control symbol table of which to relocate symbols
Pass pointer to symbol table to relocate the symbols of in relocate_syms
2016-12-03 17:26:51 +00:00
Thomas Preud'homme
c4bec037be Code simplification in relocate_syms 2016-12-03 17:26:51 +00:00
Thomas Preud'homme
25927df3b7 Consolidate all relocations in relocate_section
Static relocation of functions in dynamic libraries must use the PLT
entry as the target. Before this commit, it used to be done in 2 parts
for ARM, with the offset of the PLT entry from the beginning of the PLT
being put in the relocated place in build_got_entries () and then the
address of the PLT being added in relocate_section.

This led to code dealing with reading the offset of a bl instruction in
build_got_entries. Furthermore, the addition of the address of the start
of the PLT was done based on the relocation type which does not convey
whether a PLT entry should be used to reach the symbol.

This commit moves the decision to use the PLT as the target in
relocate_section, therefore having the instruction aware code contained
to the target-specific bit of that function (in <target>-link.c).

Note that relocate_syms is *not* the right place to do this because two
different relocations for the same symbol can make different decision.
This is the case in tcc -run mode where the static and dynamic
relocation are done by tcc.

Storing the PLT entry address in the symbol's st_value field and relying
on the specific relocation type being used for dynamic relocation would
work but the PLT entry address would then appear in the static symbol
table (symtab). This would also make the static symbol table entry
differ from the dynamic symbol table entry.
2016-12-03 17:26:51 +00:00
Thomas Preud'homme
2c38800bbe Allow to get sym attr and fail if no entry
Change alloc_sym_attr into get_sym_attr and add a parameter to control
whether to allocate a new symattr structure or return NULL if symbol is
not found;
2016-12-03 17:26:51 +00:00
Thomas Preud'homme
a11b0a67e3 Consolidate GOT creation in build_got_entries
Currently GOT/PLT creation happens in two locations depending on whether
the GOT/PLT [entry] is required by the symbol or the relocation:

- bind_exe_dynsym for relocations to undefined symbol
- build_got_entries/put_got_entry for relocations that require a GOT/PLT
  entry

This commit consolidate GOT/PLT creation in build_got_entries by
reducing bind_exe_dynsym's job to create a dynamic symbol for undefined
symbols. build_got_entries then invoke put_got_entry if the symbol being
relocated is undefined or the relocation asks for a PLT or GOT [entry].
put_got_entry is also modified to only export a symbol in the dynamic
symbol table when we are in the case of PLT/GOT [entry] required by the
relocation (since undefined symbol are already exported by
bind_exe_dynsym).
2016-12-03 17:26:51 +00:00
Thomas Preud'homme
1c811a4d1d Make build_got_entries more target independent
Factor most of common logic between targets in build_got_entries by
defining target specific info into structures in the backends.
2016-12-03 17:26:51 +00:00
Thomas Preud'homme
523b55d82d Only create GOT or GOT entry when needed
Currently we always build a GOT when we recognize a relocation in
build_got_entries even if the relocation does not require one. In the
same spirit, when the relocation does require one we always create a GOT
entry even if not entry is necessary. This patch restricts the creation
of a GOT and a GOT entry to relocations that needs it, ie:
- do not create a GOT if relocation is not related to GOT and symbol is
  not UNDEF
- do not create a GOT entry if relocation only relates to beginning of
  GOT
2016-12-03 17:26:51 +00:00
Thomas Preud'homme
6cd23d1d8c Recognize more relocations as needing GOT/PLT entry 2016-12-03 17:26:51 +00:00
Thomas Preud'homme
cb273fdad8 Do section relocation in architecture backend 2016-12-03 17:26:51 +00:00
Thomas Preud'homme
60374d01ae Add address of GOT + 8 in PLT + 16 and fix PLT0
On ARM targets, the jump to ld.so resolution routine is done in PLT0 by
loading the offset to the GOT found in PLT+16 and from there loading the
address in GOT+8 and jumping to it.

Currently tcc starts the first regular PLT entry at PLT+16 which thus
does not contain the offset to the GOT. This commit fixes that.

Note that calls via PLT still worked nonetheless because of some missing
dynamic tag which makes ld.so behaves as if RTLD_BIND_NOW was specified
in the environment for all executable created by tcc.
2016-12-03 17:26:51 +00:00
Thomas Preud'homme
f924d0ca96 Improve put_got_entry doc and structure 2016-12-03 17:26:51 +00:00
Thomas Preud'homme
64b5ee2dea Rename add_elf_sym to set_elf_sym
add_elf_sym is a confusing name because it is not clear what the
function does compared to put_elf_sym. As a matter of fact, put_elf_sym
also adds a symbol in a symbol table. Besides, "add_elf_sym" fails to
convey that the function can be used to update a symbol (for instance
its value). "set_elf_sym" seems like a more appropriate name: it will
set a symbol to a given set of properties (value, size, etc.) and create
a new one if non exist for that name as one would expect.
2016-12-03 17:26:51 +00:00
Thomas Preud'homme
9e6610b0aa Improve comments for symbol export and binding 2016-12-03 17:26:51 +00:00
Thomas Preud'homme
e9769a7249 Do not add symbol if it is already there
Do not create a new symbol in add_elf_sym if a symbol with same properties
(value, size, info, etc.) already exists. This prevents symbols from
being exported twice in the dynamic symbol table.
2016-12-03 17:26:51 +00:00
Thomas Preud'homme
bf692af31b Fix error logic for undefined reference in library
Prior to this patch, an error would only be given when a library has an
unresolved undefined symbol if there is no undefined reference for the
same symbol in the executable itself. This patch changes the logic to
check both that the executable has the symbol in its static symbol table
*and* that it is defined to decide if the error path should be followed.
2016-12-03 17:26:51 +00:00
root
ccf9ed7d54 Clear SHF_GROUP flag when linking
SHF_GROUP flag set on a section indicates that it is part of a section
group and that if the section is removed, the other sections in the same
group should be removed as well [1]. Since section group are guide for
the linking process, they do not have any meaning after linking has
occured. TCC rightfully [2] discard such sections (by not recognizing the
section type) but keeps the SHF_GROUP flag set on sections that were
part of a section group which confuses binutils (objdump and gdb at
least). Clearing that bit makes objdump and gdb accept binaries created
by TCC.

[1] https://docs.oracle.com/cd/E19683-01/816-1386/chapter7-26/index.html
[2] GNU ld does the same
2016-12-03 17:26:50 +00:00
Christian Jullien
ed99f3608d Remove warning when __builtin_frame_address is used with gcc >= 6. 2016-11-30 06:18:48 +01:00
David Mertens
d2e2f42382 Implement gcc bitfield algorithm; add -mms-bitfields 2016-11-28 09:01:12 -05:00
David Mertens
3c68a8c6c0 Minor grammar fixes to docs 2016-11-28 08:59:53 -05:00
grischka
0d14e7e698 configure: prefer lib64 on 64-bit platforms
use lib64 if
- "/usr/lib/multi-arch-triplet" does not work and
- we are on a 64-bit platform and
- lib64 exists and does contain crti.o
2016-11-28 14:48:54 +01:00
grischka
a52a39179a tccelf: introduce add32/64le() 2016-11-20 14:52:56 +01:00
grischka
4a3741bf02 x86_64-asm: =m operand fixes
The problem was with tcctest.c:

    unsigned set;
    __asm__("btsl %1,%0" : "=m"(set) : "Ir"(20) : "cc");

when with tcc compiled with the HAVE_SELINUX option, run with
tcc -run, it would use large addresses far beyond the 32bits
range when tcc did not use the pc-relative mode for accessing
'set' in global data memory.  In fact the assembler did not
know about %rip at all.

Changes:
- memory operands use (%rax) not (%eax)
- conversion from VT_LLOCAL: use type VT_PTR
- support 'k' modifier
- support %rip register
- support X(%rip) pc-relative addresses

The test in tcctest.c is from Michael Matz.
2016-11-20 14:50:56 +01:00
Thomas Stalder
47fd807f9b arm: Fix relocate_section with TCC_OUTPUT_MEMORY 2016-11-13 11:52:28 +01:00
Thomas Stalder
4af25aed92 remove warnings 2016-11-13 11:45:55 +01:00
grischka
59216d3db0 tccgen: fix inline_functions double free fix 2016-11-11 20:25:13 +01:00
grischka
7c28c9b13f tccgen: inline_functions double free fix
Fix double free of the inline function token_string which
could happen when an error/longjmp occurred while compiling
the inline function.
2016-11-11 18:29:45 +01:00