aboutsummaryrefslogtreecommitdiffstats
path: root/backend/Selection.v
Commit message (Collapse)AuthorAgeFilesLines
* Merge branch 'master' (Absint 3.8) into kvx-work-merge3.8David Monniaux2020-11-181-6/+17
|\
| * Add support for __builtin_fabsfXavier Leroy2020-07-271-0/+2
| |
| * Eliminate known builtins whose result is ignoredXavier Leroy2020-06-251-6/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A typical example is `(void) __builtin_sel(a, b, c)`. It is safe to generate zero code for these uses of builtins because builtins whose semantics are known to the compiler are pure. Other builtins with side effects (e.g. `__builtin_trap`) are not known and will remain in the compiled code. It is useful to generate zero code for these uses of builtins because some of them (e.g. `__builtin_sel`) must be transformed into proper CminorSel expressions during instruction selection. Otherwise, they propagate all the way to ExpandAsm, causing a "not implemented" error there.
* | __builtin_expect seems to workDavid Monniaux2020-04-071-11/+30
| |
* | start implementing expect as exprDavid Monniaux2020-04-071-1/+2
| |
* | __builtin_expect defined as its first argumentDavid Monniaux2019-09-251-1/+2
| |
* | Merge branch 'master' of https://github.com/AbsInt/CompCert into ↵David Monniaux2019-07-191-9/+47
|\| | | | | | | mppa-work-upstream-merge
| * Give formal semantics to some built-in functions and run-time functionsXavier Leroy2019-07-171-6/+47
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commit adds mechanisms to - recognize certain built-in and run-time functions by name and signature; - associate semantics to these functions, as a partial function from list of values to values; - interpret external calls to these functions according to this semantics (pure function from values to values, memory unchanged, no observable events in the trace); - external calls to unknown built-in and run-time functions remain interpreted as generating observable events and possibly changing memory, like before. The description of the built-ins is split into a target-independent part (in common/Builtins0.v) and a target-specific part (in $ARCH/Builtins1.v). Instruction selection uses the new mechanism in order to - recognize some built-in functions and turn them into operations of the target processor. Currently, this is done for __builtin_sel and __builtin_fabs; more to come. - remove the axioms about int64 helper functions from the standard library. More precisely, the behavior of these functions is still axiomatized, but now it is specified using the more general machinery introduced in this commit, rather than ad-hoc axioms in backend/SplitLongproof. The only built-ins currently described are __builtin_fsqrt (for all platforms) and __builtin_fmin / __builtin_fmax (for x86). More built-ins will be added later.
| * If-conversion optimizationXavier Leroy2019-06-061-9/+84
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Extends the instruction selection pass with an if-conversion optimization: some if/then/else statements are converted into "select" operations, which in turn can be compiled down to branchless instruction sequences if the target architecture supports them. The statements that are converted are of the form if (cond) { x = a1; } else { x = a2; } if (cond) { x = a1; } if (cond) { /*skip*/; } else { x = a2; } where a1, a2 are "safe" expressions, containing no operations that can fail at run-time, such as memory loads or integer divisions. A heuristic in backend/Selectionaux.ml controls when the optimization occurs, depending on command-line flags and the complexity of the "then" and "else" branches.
* | rm old select/selectl/selectf/selectfsDavid Monniaux2019-06-031-61/+0
| |
* | Merge branch 'if-conversion' of https://github.com/AbsInt/CompCert into ↵David Monniaux2019-06-031-10/+83
|\ \ | | | | | | | | | mppa-if-conversion
| * | If-conversion optimizationXavier Leroy2019-05-311-9/+84
| |/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Extends the instruction selection pass with an if-conversion optimization: some if/then/else statements are converted into "select" operations, which in turn can be compiled down to branchless instruction sequences if the target architecture supports them. The statements that are converted are of the form if (cond) { x = a1; } else { x = a2; } if (cond) { x = a1; } if (cond) { /*skip*/; } else { x = a2; } where a1, a2 are "safe" expressions, containing no operations that can fail at run-time, such as memory loads or integer divisions. A heuristic in backend/Selectionaux.ml controls when the optimization occurs, depending on command-line flags and the complexity of the "then" and "else" branches.
* | preparing for the builtin connectorDavid Monniaux2019-04-051-3/+69
| |
* | removed the unproved hack to get builtins, will be reinstated laterDavid Monniaux2019-04-051-66/+6
| |
* | Oselectf, Oselectfs with conditionDavid Monniaux2019-04-051-8/+2
| |
* | selectl with conditionDavid Monniaux2019-04-051-4/+1
| |
* | more on selectDavid Monniaux2019-04-041-4/+1
| |
* | select_soundDavid Monniaux2019-04-041-1/+1
| |
* | ternary ops for float/doubleDavid Monniaux2019-04-031-2/+34
| |
* | problem in ValueAOpDavid Monniaux2019-04-031-1/+17
| |
* | begin implementing ternary builtinDavid Monniaux2019-04-031-4/+13
| |
* | selection of builtin.. progress...David Monniaux2019-04-031-13/+14
| |
* | some more on builtinsDavid Monniaux2019-04-031-3/+14
| |
* | attempts at generating builtins, startDavid Monniaux2019-04-031-3/+6
| |
* | ça recompile sur x86David Monniaux2019-03-221-1/+1
| |
* | la division flottante fonctionneDavid Monniaux2019-03-201-1/+4
| |
* | added helper functions but strangeDavid Monniaux2019-03-191-1/+6
|/ | | | idiv.c: error: __compcert_i32_sdiv: missing or incorrect declaration
* Prefixed runtime functions.Bernhard Schommer2017-08-251-15/+15
| | | | | | | The runtime functions are prefixed with compcert in order to avoid potential clashes with runtime/builtin functions of other compilers. Bug 22062
* Hybrid 64bit/32bit PowerPC portBernhard Schommer2017-05-031-2/+2
| | | | | | | | | | | | | This commit adds code generation for 64bit PowerPC architectures which execute 32bit applications. The main difference to the normal 32bit PowerPC port is that it uses the available 64bit instructions instead of using the runtime library functions. However pointers are still 32bit and the 32bit calling convention is used. In order to use this port the target architecture must be either in Server execution mode or if in Embedded execution mode the high order 32 bits of GPRs must be implemented in 32-bit mode. Furthermore the operating system must preserve the high order 32 bits of GPRs.
* Turn 64-bit integer division and modulus by constants into multiply-highXavier Leroy2016-10-041-1/+4
| | | | | | This trick was already implemented for 32-bit integer division and modulus. Here we extend it to the 64-bit case. For 32-bit target processors, the runtime library must implement 64-bit multiply-high (signed and unsigned). Tentative implementations are provided for IA32 and PowerPC, but need testing.
* Improve code generation for 64-bit signed integer divisionXavier Leroy2016-10-021-3/+3
| | | | | | Implement the 'shift right extended' trick, both in the generic implementation (backend/SplitLong) and in the IA32 port. Note that now SelectDiv depends on SelectLong, and that some work was moved from SelectLong to SelectDiv.
* Support for 64-bit architectures: generic supportXavier Leroy2016-10-011-36/+28
| | | | | | | | | | | - Introduce Archi.ptr64 parameter. - Define module Ptrofs of integers as wide as a pointer (64 if Archi.ptr64, 32 otherwise). - Use Ptrofs.int as the offset type for Vptr values and anywhere pointer offsets are manipulated. - Modify Val operations that handle pointers (e.g. Val.add, Val.sub, Val.cmpu) so that in 64-bit pointer mode it is the "long" operation (e.g. Val.addl, Val.subl, Val.cmplu) that handles pointers. - Update the memory model accordingly. - Modify C operations that handle pointers (e.g. addition, subtraction, comparisons) accordingly. - Make it possible to turn off the splitting of 64-bit integers into pairs of 32-bit integers. - Update the compiler front-end and back-end accordingly.
* Update the back-end proofs to the new linking framework.Xavier Leroy2016-03-061-37/+31
|
* Updated PR by removing whitespaces. Bug 17450.Bernhard Schommer2015-10-201-6/+6
|
* Use Coq strings instead of idents to name external and builtin functions.Xavier Leroy2015-10-111-21/+89
| | | | | | | | | | The AST.ident type represents source-level identifiers as unique positive numbers. However, the mapping identifiers <-> AST.ident differs between runs of CompCert on different source files. This is problematic when we need to produce or recognize external functions and builtin functions with fixed names, for example: * in $ARCH/Machregs.v to define the register conventions for builtin functions; * in the VST program logic from Princeton to treat thread primitives specially. So far, we used AST.ident_of_string to recover the ident associated with a string. However, this function is defined in OCaml and doesn't execute within Coq. This is a problem both for VST and for future executability of CompCert within Coq. This commit replaces "ident" by "string" in the arguments of EF_external, EF_builtin, EF_inline_asm, EF_annot, and EF_annot_val. This provides stable names for externals and builtins, as needed. For inline asm and annotations, it's a matter of taste, but using strings feels more natural. EF_debug keeps using idents, since some kinds of EF_debug annotations talk about program variables.
* Refactoring of builtins and annotations in the back-end.Xavier Leroy2015-08-211-17/+25
| | | | | | | | | | | | | | | | | | | | | | | | Before, the back-end languages had distinct instructions - Iannot for annotations, taking structured expressions (annot_arg) as arguments, and producing no results' - Ibuiltin for other builtins, using simple pseudoregs/locations/registers as arguments and results. This branch enriches Ibuiltin instructions so that they take structured expressions (builtin_arg and builtin_res) as arguments and results. This way, - Annotations fit the general pattern of builtin functions, so Iannot instructions are removed. - EF_vload_global and EF_vstore_global become useless, as the same optimization can be achieved by EF_vload/vstore taking a structured argument of the "address of global" kind. - Better code can be generated for builtin_memcpy between stack locations, or volatile accesses to stack locations. Finally, this commit also introduces a new kind of external function, EF_debug, which is like EF_annot but produces no observable events. It will be used later to transport debug info through the back-end, without preventing optimizations.
* Updated the Caml part. Added some more tests in annot1.c.Xavier Leroy2015-03-271-1/+12
|
* Extend annotations so that they can keep track of global variables and local ↵Xavier Leroy2015-03-271-1/+9
| | | | | | | | | | variables whose address is taken. - CminorSel, RTL: add "annot" instructions. - CminorSel to Asm: use type "annot_arg" for arguments of "annot" instructions. - AST, Events: simplify EF_annot because constants are now part of the arguments. Implementation is not complete yet.
* Protect against redefinition of the __i64_xxx helper library functions.Xavier Leroy2015-01-201-27/+26
| | | | | | | | | | This is achieved by declaring these functions in C2C.ml, then re-checking their declarations in the global environment during the Selection pass. In passing, the "hf" parameter for SelectLong functions was removed and replaced by fixed identifiers corresponding to the actual names of the helper functions.
* - Support "switch" statements over 64-bit integersxleroy2014-08-171-33/+94
| | | | | | | | | | | | | (in CompCert C to Cminor, included) - Translation of "switch" to decision trees or jumptables made generic over the sizes of integers and moved to the Cminor->CminorSel pass instead of CminorSel->RTL as before. - CminorSel: add "exitexpr" to support the above. - ValueDomain: more precise analysis of comparisons against an integer literal. E.g. "x >=u 0" is always true. git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@2565 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Merge of "newspilling" branch:xleroy2014-07-231-0/+15
| | | | | | | | | | | | | | | - Support single-precision floats as first-class values - Introduce chunks Many32, Many64 and types Tany32, Tany64 to support saving and restoring registers without knowing the exact types (int/single/float) of their contents, just their sizes. - Memory model: generalize the opaque encoding of pointers to apply to any value, not just pointers, if chunks Many32/Many64 are selected. - More properties of FP arithmetic proved. git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@2537 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Optimize integer divisions by positive constants, turning them intoxleroy2013-07-291-0/+1
| | | | | | | multiply-high and shifts. git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@2300 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Treat casts int64 -> float32 as primitive operations instead of twoxleroy2013-07-031-0/+2
| | | | | | | | casts int64 -> float64 -> float32. The latter causes double rounding errors. git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@2290 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Expand 64-bit integer comparisons into 32-bit integer comparisons.xleroy2013-04-291-7/+9
| | | | git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@2218 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Big merge of the newregalloc-int64 branch. Lots of changes in two directions:xleroy2013-04-201-9/+44
| | | | | | | | | 1- new register allocator (+ live range splitting, spilling&reloading, etc) based on a posteriori validation using the Rideau-Leroy algorithm 2- support for 64-bit integer arithmetic (type "long long"). git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@2200 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Remove some useless "Require".xleroy2012-12-301-4/+0
| | | | | | | Update ARM port. git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@2085 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Merge of branch seq-and-or. See Changelog for details.xleroy2012-10-061-50/+6
| | | | git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@2059 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Merge of the newmem branch:xleroy2012-05-211-10/+20
| | | | | | | | | | - Revised memory model with Max and Cur permissions, but without bounds - Constant propagation of 'const' globals - Function inlining at RTL level - (Unprovable) elimination of unreferenced static definitions git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@1899 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Take advantage of Cmaskzero and Cmasknotzero.xleroy2012-02-241-1/+4
| | | | git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@1825 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
* Merge of the "volatile" branch:xleroy2012-02-041-0/+3
| | | | | | | | | | | | - native treatment of volatile accesses in CompCert C's semantics - translation of volatile accesses to built-ins in SimplExpr - native treatment of struct assignment and passing struct parameter by value - only passing struct result by value remains emulated - in cparser, remove emulations that are no longer used - added C99's type _Bool and used it to express || and && more efficiently. git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@1814 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
generated by cgit (git 2.34.1) at 2024-05-20 07:53:23 +0000