aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Support local register variablesMichael Matz2016-12-154-6/+58
| | | | | | | | | Similar to GCC a local asm register variable enforces the use of a specified register in asm operands (and doesn't otherwise matter). Works only if the variable is directly mentioned as operand. For that we now generally store a backpointer from an SValue to a Sym when the SValue was the result of unary() parsing a symbol identifier.
* Fix const folding of 64bit pointer constantsMichael Matz2016-12-152-2/+8
| | | | See testcase.
* x86-asm: Correct register size for pointer opsMichael Matz2016-12-152-1/+10
| | | | | A pointer is 64 bit as well, so it needs a full register for register operands.
* struct-init: Copy relocs for compound literalsMichael Matz2016-12-152-0/+55
| | | | | When copying the content of compound literals we must include relocations as well.
* x86_64-asm: fix copy-out registersMichael Matz2016-12-152-1/+11
| | | | | | If the destination is an indirect pointer access (which ends up as VT_LLOCAL) the intermediate pointer must be loaded as VT_PTR, not as whatever the pointed to type is.
* x86_64: Add -mno-sse optionMichael Matz2016-12-153-4/+29
| | | | | This disables generation of any SSE instructions (in particular in stdarg function prologues). Necessary for kernel compiles.
* x86-64: relocation addend is 64bitMichael Matz2016-12-153-6/+6
| | | | | | | | Some routines were using the wrong type (int) in passing addends, truncating it. This matters when bit 31 isn't set and the high 32 bits are set: the truncation would make it unsigned where in reality it's signed (happen e.g. on the x86-64 with it's load address at top-2GB).
* x86-64: Prefer 32S relocationsMichael Matz2016-12-152-2/+2
| | | | | | | | | | | | | | | | This target has _32 and _32S relocs (the latter being for signed 32 bit entities). All instruction displacements have to use the 32S variants. Normal references like .long s normally would use the _32 variant. For normal executables this doesn't matter. For shared libraries neither (which use PC-relative relocs). But it matters for things like the kernel that are linked to high addresses (signed ones). There the GNU linker would error out on overflow for the _32 variant. To keep life simple we simply switch from _32 to _32S altogether. Strictly speaking it's still wrong, but in practice using _32 is more often wrong than using _32S ;)
* opt: Don't emit inline functions from dead codeMichael Matz2016-12-152-1/+9
| | | | | Inside dead code don't regard inline functions as being referenced.
* Handle __builtin_extract_return_addrMichael Matz2016-12-152-0/+5
| | | | | Our architectures don't need anything special for this builtin, just pass through the argument.
* opt: constprop also 'cond && 0'Michael Matz2016-12-152-62/+113
| | | | | | We didn't handle constants in logical expressions when they weren't the first operand. Some reordering in the loop structure is enough to handle them.
* opt: Make break and goto not fallthroughMichael Matz2016-12-152-2/+68
| | | | | As we can optimize dead code a bit already it's fitting to disable code emission after break and goto.
* opt: Start optimizing dead code a bitMichael Matz2016-12-154-2/+182
| | | | | | | If a condition is always zero/non-zero we can omit the then or else code. This is complicated a bit by having to deal with labels that might make such code reachable without us yet knowing during parsing.
* Revert "Reject jumping inside stmtexprs"Michael Matz2016-12-151-7/+1
| | | | | | | | | | | | | Not fully thought out. You can't jump inside stmt exprs, but you can jump out of them. So there's a difference between undefined but declared labels at the end of stmt exprs and those defined inside. Additionally it should also be checked if a label defined inside a stmt expr was tentatively created as declared from outside. I'm not prepared doing that right now, so simply revert. This reverts commit 9160e4cab9147d77840cc44a285031fdb4640cf9.
* Factor out const condition detectionMichael Matz2016-12-151-10/+22
| | | | Creating condition_3way for this.
* Reject jumping inside stmtexprsMichael Matz2016-12-151-1/+7
| | | | | | | | | | | | One can't jump into statement expressions from outside them, like the following: int i = ({ label: foo(); 42; }); goto label; We reject this by making the labels simply not available outside (GCC has a nicer error message about jumping into a statement expression).
* Fix more nocode_wanted jump problemsMichael Matz2016-12-153-4/+46
| | | | | | | In statement expression we really mustn't emit backward jumps under nocode_wanted (they will form infinte loops as no expressions are evaluated). Do-while and explicit loop with gotos weren't handled.
* Fix aliases on 64 bitMichael Matz2016-12-152-2/+3
| | | | Use correct width ELF structure.
* Fix sizeof(char[a])Michael Matz2016-12-152-12/+25
| | | | | The sizes of VLAs need to be evaluated even inside sizeof, i.e. when nocode_wanted is set.
* Fix __builtin_constant_p(1000/x)Michael Matz2016-12-152-0/+27
| | | | | was incorrectly treated as constant because the vpop removed all traces of non-constness.
* tccasm: Support refs to anon symbols from asmMichael Matz2016-12-154-13/+52
| | | | | | | This happens when e.g. string constants (or other static data) are passed as operands to inline asm as immediates. The produced symbol ref wouldn't be found. So tighten the connection between C and asm-local symbol table even more.
* tccasm: Don't ignore # in preprocessor directivesMichael Matz2016-12-154-25/+49
| | | | | | | Our preprocessor throws away # line-comments in asm mode. It did so also inside preprocessor directives, thereby removing stringification. Parse defines in non-asm mode (but retain '.' as identifier character inside macro definitions).
* x86-asm: Accept all 32bit immediatesMichael Matz2016-12-152-1/+3
| | | | | In particular don't care if they're signed or unsigned, they're all acceptable as immediates.
* Fix enum bitfields passed to stdarg functionsMichael Matz2016-12-152-0/+6
| | | | | VT_ENUM types use the .ref member and can be VT_BITFIELD, so we need to copy it as well. Simply do it always.
* Addresses of non-weak symbols are non-zeroMichael Matz2016-12-152-1/+13
| | | | Use this fact in some foldings of comparisons. See testcase.
* Fix access-after-free with statement expressionsMichael Matz2016-12-154-29/+76
| | | | | | | | | | | | The return value of statement expressions might refer to local symbols, so those can't be popped. The old error message always was just a band-aid, and since disabling it for pointer types it wasn't effective anyway. It also never considered that also the vtop->sym member might have referred to such symbols (see the testcase with the local static, that used to segfault). For fixing this (can be seen better with valgrind and SYM_DEBUG) simply leave local symbols of stmt exprs on the stack.
* tccpp: Allow computed include like 42.hMichael Matz2016-12-153-29/+50
| | | | | | The include directive needs to be parsed as pp-tokens, not as token (i.e. no conversion to TOK_STR or TOK_NUM). Also fix parsing computed includes using quoted strings.
* x86-asm: Correctly infer register size for boolsMichael Matz2016-12-152-1/+12
| | | | | | Register operands of type _Bool weren't correctly getting the 8-bit sized registers (but rather used the default 32-bit ones).
* x86-64-asm: Implement cmpxchg16bMichael Matz2016-12-152-0/+6
|
* x86-64: Allow loads from some structs/unionsMichael Matz2016-12-152-0/+22
| | | | | | GCC allows register loads for asms based on type mode, and correctly sized structs/union have an allowed mode (basically 1,2,4,8 sized aggregates).
* tccasm: Lookup C symbols from ASM blocksMichael Matz2016-12-152-19/+71
| | | | | It's now possible to use symbols defined in C code to be used from later inline asm blocks. See testcase.
* tccasm: Implement .set sym, exprMichael Matz2016-12-153-16/+61
| | | | | | | | | | That, as well as "sym = expr", if expr contains symbols. Slightly tricky because a definition from .set is overridable, whereas proper definitions aren't. This doesn't yet allow using this for override tricks from C and global asm blocks because the symbol tables from C and asm are separate.
* enums and ints are compatibleMichael Matz2016-12-154-14/+35
| | | | | | | But like GCC do warn about changes in signedness. The latter leads to some changes in gen_assign_cast to not also warn about unsigned* = int* (where GCC warns, but only with extra warnings).
* enums and ints are compatibleMichael Matz2016-12-152-0/+10
|
* x86-64-asm: Fix ltr/str and push/pop operandsMichael Matz2016-12-153-2/+12
| | | | | str accepts rm16/r32/r64, and push/pop defaults to 64 when given memory operands (to 32 on i386).
* x86-64-asm: Implement high %cr registersMichael Matz2016-12-152-4/+13
|
* struct-init: Support range inits for local varsMichael Matz2016-12-153-17/+33
| | | | Implement missing support for range init for local variables.
* x86-64-asm: Support high registers %r8 - %r15Michael Matz2016-12-153-43/+215
| | | | | | | This requires correctly handling the REX prefix. As bonus we now also support the four 8bit registers spl,bpl,sil,dil, which are decoded as ah,ch,dh,bh in non-long-mode (and require a REX prefix as well).
* inline-asm: Accept "flags" clobberMichael Matz2016-12-152-3/+4
|
* struct-init: Allow member initialization from qualified lvaluesMichael Matz2016-12-153-4/+33
| | | | See testcase.
* struct-init: Correctly parse unnamed member initializersMichael Matz2016-12-153-21/+78
| | | | | | | | | | | | | For union U { struct {int a,b}; int c; }; union U u = {{ 1, 2, }}; The unnamed first member of union U needs to actually exist in the structure so initializer parsing isn't confused about the double braces. That means also the a and b members must be part of _that_, not of union U directly. Which in turn means we need to do a bit more work for field lookup. See the testcase extension for more things that need to work.
* struct-init: Cleanup some moreMichael Matz2016-12-151-98/+103
| | | | | Some parameters aren't actually necessary. Also join the two parsing loops for the initializer list of arrays and structs.
* struct-init: CleanupMichael Matz2016-12-154-103/+25
| | | | | | | Remove dead code and variables. Properly check for unions when skipping fields in initializers. Make tests2/*.expect depend on the .c files so they are automatically rebuilt when the latter change.
* struct-init: Implement initializing subaggregatesMichael Matz2016-12-154-40/+226
| | | | | | E.g. "struct { struct S s; int a;} = { others, 42 };" if 'others' is also a 'struct S'. Also when the value is a compound literal. See added testcases.
* struct-init: ReimplementMichael Matz2016-12-151-78/+106
| | | | | | Start reimplementing the whole initializer handling to be conforming to ISO C. This patch just reimplements current functionality to prepare for further changes, all tests pass.
* Support attribute between double pointer starsMichael Matz2016-12-152-4/+12
| | | | "int * __attribute__((something)) *" is supported by GCC.
* Fix function to pointer conversionMichael Matz2016-12-152-0/+20
| | | | | | | | This snippet is valid: void foo(void); ... foo + 42 ... the function designator is converted to pointer to function implicitely. gen_op didn't do that and bailed out.
* Fix parsing array typedefs of unknown sizeMichael Matz2016-12-152-0/+22
| | | | | | | | | | | This must compile: typedef int arrtype1[]; arrtype1 sinit19 = {1}; arrtype1 sinit20 = {2,3}; and generate two arrays of one resp. two elements. Before the fix the determined size of the first array was encoded in the type directly, so sinit20 couldn't be parsed anymore (because arrtype1 was thought to be only one element long).
* tccpp: Implement __BASE_FILE__ macroMichael Matz2016-12-153-0/+22
| | | | | Like __FILE__ but always refers to the command line file name also from inside headers.
* Implement __builtin_choose_exprMichael Matz2016-12-153-0/+43
| | | | Follows GCC implementation.