Christoph Mallon [Wed, 12 Aug 2009 17:36:19 +0000 (17:36 +0000)]
Minor corrections.
[r26331]
Christoph Mallon [Wed, 12 Aug 2009 11:59:03 +0000 (11:59 +0000)]
Remove unnecessary #include "phiclass.h".
[r26330]
Matthias Braun [Wed, 12 Aug 2009 09:14:55 +0000 (09:14 +0000)]
fix rbitset_is_empty; put opening brace of functions on an own line
[r26329]
Matthias Braun [Tue, 11 Aug 2009 11:46:32 +0000 (11:46 +0000)]
- implement output constraint enforcement for new register allocator
- rename bespilloptions.* to bespill.* and bespill.* to bespillutil.*
[r26327]
Matthias Braun [Tue, 11 Aug 2009 11:45:44 +0000 (11:45 +0000)]
SwitchJmp doesn't have 1 'none' output
[r26326]
Matthias Braun [Tue, 11 Aug 2009 11:45:18 +0000 (11:45 +0000)]
add rbitsets_have_common function
[r26325]
Matthias Braun [Thu, 6 Aug 2009 16:36:43 +0000 (16:36 +0000)]
fix
[r26321]
Matthias Braun [Thu, 6 Aug 2009 14:49:04 +0000 (14:49 +0000)]
- Rewrite internal backend logic for querying register constraints.
(We have a separate get_input, and get_output callback for now).
This should make the code faster for now and is a first step towards
changing the interface to query register constraints on the mode_T node
itself instead of the Proj nodes.
- Handle middleend node constraints and stuff in benode.c instead of in each
backend
- Remove irn_class_branch we already had is_cfop in the middleend
- Fix a bunch of bugs/problems in the process
[r26320]
Matthias Braun [Wed, 5 Aug 2009 15:16:07 +0000 (15:16 +0000)]
fix read from uninitialized memory
[r26318]
Matthias Braun [Wed, 5 Aug 2009 10:53:46 +0000 (10:53 +0000)]
reorganize backend headers (kill some _t variants in favor of a be_types.h)
[r26317]
Matthias Braun [Tue, 4 Aug 2009 11:36:09 +0000 (11:36 +0000)]
fix register allocation verifier not checking output constraints for mode_T nodes
[r26315]
Matthias Braun [Mon, 3 Aug 2009 17:26:53 +0000 (17:26 +0000)]
initial work for phi-nodes (ifs should work, loops not yet)
[r26314]
Matthias Braun [Mon, 3 Aug 2009 12:40:06 +0000 (12:40 +0000)]
more work on permutate values (not finished yet)
[r26313]
Matthias Braun [Sun, 2 Aug 2009 18:09:24 +0000 (18:09 +0000)]
fix warning and compilation error in optimized mode
[r26311]
Matthias Braun [Sun, 2 Aug 2009 18:09:07 +0000 (18:09 +0000)]
fix a few bugs in benewalloc, split codegen timer in 2
[r26310]
Matthias Braun [Fri, 31 Jul 2009 12:39:52 +0000 (12:39 +0000)]
cleanup backend: make pre_spill_prepare_constraint independent of chordal allocator structures (and rework it a bit in the process)
[r26304]
Matthias Braun [Fri, 31 Jul 2009 12:39:14 +0000 (12:39 +0000)]
unnecessary include
[r26303]
Matthias Braun [Fri, 31 Jul 2009 12:39:02 +0000 (12:39 +0000)]
fix warning
[r26302]
Matthias Braun [Fri, 31 Jul 2009 12:38:45 +0000 (12:38 +0000)]
fix warning
[r26301]
Matthias Braun [Fri, 31 Jul 2009 10:35:46 +0000 (10:35 +0000)]
fix last commit
[r26299]
Matthias Braun [Fri, 31 Jul 2009 10:27:57 +0000 (10:27 +0000)]
declare load/store alternative in iroptimize header
[r26298]
Matthias Braun [Fri, 31 Jul 2009 10:22:42 +0000 (10:22 +0000)]
less abreviations: rebitset_cpy => rebitset_copy
[r26297]
Sebastian Buchwald [Wed, 15 Jul 2009 15:01:42 +0000 (15:01 +0000)]
condeval is called jump threading now
[r26280]
Matthias Braun [Fri, 10 Jul 2009 13:56:22 +0000 (13:56 +0000)]
condeval is called Jump Threading now
[r26260]
Michael Beck [Mon, 6 Jul 2009 18:39:53 +0000 (18:39 +0000)]
- hopefully fixed the lftr now
[r26241]
Michael Beck [Sun, 5 Jul 2009 17:38:32 +0000 (17:38 +0000)]
- fixed warnings regarding unused irg's
[r26239]
Michael Beck [Sun, 5 Jul 2009 17:14:32 +0000 (17:14 +0000)]
- remove all irg parameter from node constructors having a block
- SymConst's are now ALWAYS placed in the start block
[r26236]
Matthias Braun [Mon, 29 Jun 2009 09:09:31 +0000 (09:09 +0000)]
remove unused common/firmwalk.* code
[r26224]
Matthias Braun [Mon, 29 Jun 2009 09:06:41 +0000 (09:06 +0000)]
remove pointless and unused common/statistics.* code
[r26223]
Matthias Braun [Mon, 29 Jun 2009 09:04:43 +0000 (09:04 +0000)]
Cleanup: remove firm_common_t.h (and the PRECISE_EXC_CONTEXT define)
[r26222]
Matthias Braun [Mon, 29 Jun 2009 08:58:41 +0000 (08:58 +0000)]
fix bad merge in last commit
[r26221]
Matthias Braun [Mon, 29 Jun 2009 08:19:31 +0000 (08:19 +0000)]
rework python generators: we can make use of the classes/inheritance language features to nicely define our nodes
[r26220]
Matthias Braun [Fri, 26 Jun 2009 08:41:09 +0000 (08:41 +0000)]
mark get_Block_irg inline to avoid warnings
[r26190]
Michael Beck [Thu, 25 Jun 2009 22:48:39 +0000 (22:48 +0000)]
Add an irg attribute the Bad nodes: so get_Block_irg() can operate on
Bad nodes just like on Block nodes.
This simplifies get_irn_irg() and implements the idea the the irg can
be retrieved from "the block" of a node.
[r26188]
Michael Beck [Wed, 24 Jun 2009 19:48:06 +0000 (19:48 +0000)]
- fixed warning
[r26181]
Michael Beck [Wed, 24 Jun 2009 19:41:25 +0000 (19:41 +0000)]
- add an inline version of get_Block_irg()
[r26180]
Michael Beck [Wed, 24 Jun 2009 19:37:01 +0000 (19:37 +0000)]
- add some doxygen comments laying around
[r26179]
Matthias Braun [Fri, 19 Jun 2009 15:08:39 +0000 (15:08 +0000)]
another instance of wrong block in new_Proj
[r26164]
Matthias Braun [Fri, 19 Jun 2009 13:44:10 +0000 (13:44 +0000)]
verify that Projs are in the same Block as their predecessors (and fix bugs in firm where this wasn't the case)
[r26163]
Matthias Braun [Fri, 19 Jun 2009 12:10:54 +0000 (12:10 +0000)]
- Get rid of the infamous Start->ProjX->Start backedge
[r26162]
Matthias Braun [Thu, 18 Jun 2009 14:47:25 +0000 (14:47 +0000)]
transform assert to an if, as some students tend to hit it
[r26157]
Matthias Braun [Thu, 18 Jun 2009 11:34:52 +0000 (11:34 +0000)]
fix wrong assert, expor copy_node_attr function to public API
[r26153]
Matthias Braun [Thu, 18 Jun 2009 11:34:27 +0000 (11:34 +0000)]
small fixes
[r26152]
Matthias Braun [Wed, 3 Jun 2009 09:10:04 +0000 (09:10 +0000)]
- remove modecode concept which is pointless nowadays
[r26093]
Michael Beck [Tue, 2 Jun 2009 22:38:55 +0000 (22:38 +0000)]
- svn properties added
- typos fixed
- C99 features removed
[r26089]
Michael Beck [Tue, 2 Jun 2009 22:31:52 +0000 (22:31 +0000)]
- C99 features removed
[r26088]
Michael Beck [Tue, 2 Jun 2009 22:26:36 +0000 (22:26 +0000)]
- fixed Win32 build after r26081
[r26087]
Matthias Braun [Mon, 1 Jun 2009 09:53:40 +0000 (09:53 +0000)]
add note/error message about critical edge splitting and IJmps
[r26084]
Matthias Braun [Mon, 1 Jun 2009 08:53:20 +0000 (08:53 +0000)]
fix a bunch of bugs in my last commit (no idea why I didn't see them on my mac)
[r26082]
Matthias Braun [Mon, 1 Jun 2009 08:52:55 +0000 (08:52 +0000)]
rework architecture specific timer code
[r26081]
Matthias Braun [Mon, 1 Jun 2009 08:34:19 +0000 (08:34 +0000)]
Rework Block labels: They are entities now so we don't need a special symconst type for them
[r26080]
Matthias Braun [Mon, 1 Jun 2009 08:29:57 +0000 (08:29 +0000)]
small fixes and improvements to mysql support
[r26079]
Michael Beck [Sat, 30 May 2009 15:30:16 +0000 (15:30 +0000)]
- fixed warning
[r26075]
Michael Beck [Sat, 30 May 2009 15:29:03 +0000 (15:29 +0000)]
- lftr() is still buggy
[r26074]
Michael Beck [Sat, 30 May 2009 13:28:18 +0000 (13:28 +0000)]
- removed permanent debug output
- try to fix linear-function-test-replacement
[r26073]
Michael Beck [Sat, 30 May 2009 12:14:20 +0000 (12:14 +0000)]
- BugFix: when a P<->I conversion for an induction variable took place, run a post pass that fixes probably generated AddI(Iu,Is), SubI(IuIs).
- this fixes fehler169.c
[r26072]
Johannes Franz [Fri, 29 May 2009 14:56:42 +0000 (14:56 +0000)]
removed unnecessary test
[r26070]
Matthias Braun [Fri, 29 May 2009 09:17:32 +0000 (09:17 +0000)]
remove double declarataion spotted by cparser
[r26065]
Michael Beck [Fri, 29 May 2009 01:53:01 +0000 (01:53 +0000)]
- implemented i_mapper_bswap()
[r26062]
Matthias Braun [Thu, 28 May 2009 18:31:38 +0000 (18:31 +0000)]
hacky workaround for some builtin_return_address deficiencies (make gcc testsuite work...)
[r26061]
Matthias Braun [Wed, 27 May 2009 08:12:25 +0000 (08:12 +0000)]
libtool adjust to improve win32 support
[r26057]
Matthias Braun [Wed, 27 May 2009 08:11:37 +0000 (08:11 +0000)]
remove firmnet and ycomp remote support (avoid unnecesary/complicated dependencies for something noone uses)
[r26056]
Matthias Braun [Wed, 20 May 2009 13:33:30 +0000 (13:33 +0000)]
we don't depend on libcore anymore
[r26012]
Michael Beck [Tue, 19 May 2009 11:51:32 +0000 (11:51 +0000)]
- BugFix: add missing check for frame entities. This fixes testsuite/opt/
20001212-1.c
[r26007]
Matthias Braun [Tue, 19 May 2009 09:08:45 +0000 (09:08 +0000)]
remove bogus assert
[r26004]
Matthias Braun [Sun, 17 May 2009 10:19:56 +0000 (10:19 +0000)]
fix my last commit
[r25998]
Matthias Braun [Sun, 17 May 2009 10:13:54 +0000 (10:13 +0000)]
change livense to LGPL
[r25997]
Matthias Braun [Sun, 17 May 2009 10:13:21 +0000 (10:13 +0000)]
- newer gcc warned me about errors in our strncat usage - this function is very
unintuitive as it writes up to n+1 bytes, so you have to do:
strncat(buf, X, sizeof(buf)-1);
- fixed a bunch of other warnings in release mode
[r25996]
Michael Beck [Fri, 15 May 2009 23:52:56 +0000 (23:52 +0000)]
- more docu added
[r25986]
Michael Beck [Fri, 15 May 2009 23:52:29 +0000 (23:52 +0000)]
- removed useless space
[r25985]
Michael Beck [Fri, 15 May 2009 23:51:40 +0000 (23:51 +0000)]
- make errors and warnings more understandable
[r25984]
Matthias Braun [Fri, 15 May 2009 22:09:38 +0000 (22:09 +0000)]
improve sched_info datastructure, saving space and making it more efficient
[r25982]
Matthias Braun [Fri, 15 May 2009 22:08:43 +0000 (22:08 +0000)]
update filelist script
[r25981]
Matthias Braun [Fri, 15 May 2009 22:08:34 +0000 (22:08 +0000)]
update configure, Makefile.am
[r25980]
Matthias Braun [Fri, 15 May 2009 11:59:42 +0000 (11:59 +0000)]
add helper script for updating Makefile.am, update Makefile.am
[r25972]
Matthias Braun [Fri, 15 May 2009 11:20:44 +0000 (11:20 +0000)]
fix mode querying for mode_T div nodes
[r25967]
Matthias Braun [Fri, 15 May 2009 08:53:39 +0000 (08:53 +0000)]
generate a few more things automatically
[r25965]
Matthias Braun [Fri, 15 May 2009 08:53:30 +0000 (08:53 +0000)]
NEWS udpate
[r25964]
Matthias Braun [Thu, 14 May 2009 22:23:11 +0000 (22:23 +0000)]
when inlining and return types of method-type/call-type don't match create a conv (or give up)
[r25962]
Michael Beck [Thu, 14 May 2009 18:18:11 +0000 (18:18 +0000)]
- fix r25941
[r25956]
Christoph Mallon [Thu, 14 May 2009 13:13:59 +0000 (13:13 +0000)]
Fix compile errors.
[r25950]
Matthias Braun [Thu, 14 May 2009 11:09:23 +0000 (11:09 +0000)]
eclipse C project is out of date and unmaintained
[r25945]
Matthias Braun [Thu, 14 May 2009 11:04:26 +0000 (11:04 +0000)]
add some hacks that fixes register pressure faithfullness of Asm nodes in common cases (fixes asm_test6.c)
[r25944]
Michael Beck [Thu, 14 May 2009 09:20:17 +0000 (09:20 +0000)]
- removed unused header attribute from node_entry
- improved move_loads_out_of_loops() by calculating the avail set
[r25942]
Michael Beck [Thu, 14 May 2009 09:17:29 +0000 (09:17 +0000)]
- cleaned up irp functions a bit
[r25941]
Matthias Braun [Thu, 14 May 2009 08:07:40 +0000 (08:07 +0000)]
always set graph state to KEEP_MUX in lower_mode_b
[r25939]
Matthias Braun [Wed, 13 May 2009 11:44:51 +0000 (11:44 +0000)]
- Introduce the concept of the graph state bitset: This is a set of flags that indicate the current "state" of a graph.
(Used so far to indicate wether muxes are lowered and should not be touched anymore and wether architecture dependent mul and div with constant implementations are used)
We should use this to indicate dom/loop/... consistency in the fututure
- fixed a bug where lower_mode_b was creating a "set" instruction with the help of a Mux. (Maybe we should create a new Node for this and not use a 0/1 mux?)
[r25937]
Matthias Braun [Tue, 12 May 2009 13:54:05 +0000 (13:54 +0000)]
improve comments, remove unnecessary test
[r25931]
Matthias Braun [Mon, 11 May 2009 14:36:04 +0000 (14:36 +0000)]
add a newline to avoid warning
[r25926]
Matthias Braun [Mon, 11 May 2009 14:06:27 +0000 (14:06 +0000)]
more fixes for strange op entries
[r25924]
Matthias Braun [Mon, 11 May 2009 13:57:08 +0000 (13:57 +0000)]
more fixes (sorry for the breakages...)
[r25923]
Matthias Braun [Mon, 11 May 2009 12:32:30 +0000 (12:32 +0000)]
fix a bunch of stuff in my last commit
[r25921]
Matthias Braun [Mon, 11 May 2009 11:09:49 +0000 (11:09 +0000)]
- generate more stuff from ir_spec.py
- make some functions consistent: new_r(d)_Unknown, new_r(d)_Dummy
actually take graphs and debug infos now
[r25919]
Michael Beck [Fri, 8 May 2009 13:26:29 +0000 (13:26 +0000)]
- C99 feature removed
[r25903]
Michael Beck [Fri, 8 May 2009 10:44:04 +0000 (10:44 +0000)]
- changed handling of Mux nodes: these are now optimized using the iropt instead of additional runs in the if-conv
if-conv:
- through away analysis info only when graph was changed
- fixed handling of Raise
- removed useless dominator computation
- add some docu
[r25901]
Michael Beck [Fri, 8 May 2009 10:39:58 +0000 (10:39 +0000)]
- fixed memory leaks
- add missing static
- fixed documentation
[r25900]
Michael Beck [Fri, 8 May 2009 10:38:44 +0000 (10:38 +0000)]
- let loop run backwards
[r25899]
Michael Beck [Fri, 8 May 2009 10:37:27 +0000 (10:37 +0000)]
- turn while into do-while
[r25898]
Michael Beck [Fri, 8 May 2009 10:35:30 +0000 (10:35 +0000)]
- do not run node_is_in_irg_storage() in release mode, this seems to use a remarkable amount of time
[r25897]
Michael Beck [Fri, 8 May 2009 10:30:42 +0000 (10:30 +0000)]
- removed needless cast
- fixed type of iterator variable
- iterate backwards
[r25896]