libfirm
13 years agokeeps behind a series of phis are fine
Matthias Braun [Wed, 2 Mar 2011 09:56:31 +0000 (10:56 +0100)]
keeps behind a series of phis are fine

13 years agoremove minir dump experiments
Matthias Braun [Wed, 2 Mar 2011 13:23:01 +0000 (14:23 +0100)]
remove minir dump experiments

13 years agoremove unmaintained/broken bespillbelady2
Matthias Braun [Wed, 2 Mar 2011 12:53:34 +0000 (13:53 +0100)]
remove unmaintained/broken bespillbelady2

13 years agofix code after TLS-type change
Matthias Braun [Fri, 4 Mar 2011 13:17:02 +0000 (14:17 +0100)]
fix code after TLS-type change

13 years agoFix some -Wmissing-declarations warnings
Matthias Braun [Thu, 3 Mar 2011 09:26:10 +0000 (10:26 +0100)]
Fix some -Wmissing-declarations warnings

13 years agofix more size_t warnings
Matthias Braun [Fri, 4 Mar 2011 12:04:36 +0000 (13:04 +0100)]
fix more size_t warnings

13 years agoadd compound parameter lower to lower_for_target
Matthias Braun [Fri, 4 Mar 2011 11:43:00 +0000 (12:43 +0100)]
add compound parameter lower to lower_for_target

13 years agoFixed size_t related warnings.
Michael Beck [Thu, 3 Mar 2011 21:43:58 +0000 (22:43 +0100)]
Fixed size_t related warnings.

13 years agoFurther pushed size_t: structure analysis uses size_t now.
Michael Beck [Thu, 3 Mar 2011 21:31:45 +0000 (22:31 +0100)]
Further pushed size_t: structure analysis uses size_t now.

- simplified code
- deleted macros and useless casts
- fixed some debug messages

13 years agoIterate over all segment types when detecting static allocated objects in rta.
Michael Beck [Thu, 3 Mar 2011 20:39:27 +0000 (21:39 +0100)]
Iterate over all segment types when detecting static allocated objects in rta.

13 years agoTLS type must be a struct type, as stated by the documentation.
Michael Beck [Thu, 3 Mar 2011 20:38:20 +0000 (21:38 +0100)]
TLS type must be a struct type, as stated by the documentation.

13 years agoFix rapid type analysis: TLS type is a class type.
Julian Oppermann [Thu, 3 Mar 2011 10:19:21 +0000 (11:19 +0100)]
Fix rapid type analysis: TLS type is a class type.

13 years agoFixed size_t related warnings.
Michael Beck [Wed, 2 Mar 2011 22:53:34 +0000 (23:53 +0100)]
Fixed size_t related warnings.

Fixed typo.

13 years agoFixed size_t related warnings.
Michael Beck [Wed, 2 Mar 2011 22:43:49 +0000 (23:43 +0100)]
Fixed size_t related warnings.

13 years agoFixed size_t related warnings in dumping code.
Michael Beck [Wed, 2 Mar 2011 22:38:08 +0000 (23:38 +0100)]
Fixed size_t related warnings in dumping code.

13 years agoFurther pushed size_t: callgraph functions uses size_t now.
Michael Beck [Wed, 2 Mar 2011 22:34:57 +0000 (23:34 +0100)]
Further pushed size_t: callgraph functions uses size_t now.

13 years agoFixed size_t related warnings.
Michael Beck [Wed, 2 Mar 2011 08:31:25 +0000 (09:31 +0100)]
Fixed size_t related warnings.

13 years agoFurther pushed size_t: trouts functions uses size_t now.
Michael Beck [Wed, 2 Mar 2011 08:30:52 +0000 (09:30 +0100)]
Further pushed size_t: trouts functions uses size_t now.

13 years agoFixed size_t related warnings.
Michael Beck [Tue, 1 Mar 2011 23:33:43 +0000 (00:33 +0100)]
Fixed size_t related warnings.

13 years agoFurther pushed size_t: ir_initializer_t uses size_t now.
Michael Beck [Tue, 1 Mar 2011 23:18:47 +0000 (00:18 +0100)]
Further pushed size_t: ir_initializer_t uses size_t now.

13 years agoFixed size_t related warnings.
Michael Beck [Tue, 1 Mar 2011 23:16:46 +0000 (00:16 +0100)]
Fixed size_t related warnings.

13 years agoC99 features removed.
Michael Beck [Tue, 1 Mar 2011 01:47:33 +0000 (02:47 +0100)]
C99 features removed.

13 years agofix warnings because compare_attr gets const nodes now
Matthias Braun [Mon, 28 Feb 2011 18:48:29 +0000 (19:48 +0100)]
fix warnings because compare_attr gets const nodes now

13 years agoia32: prefere != over < or > relation where possible
Matthias Braun [Mon, 28 Feb 2011 15:53:53 +0000 (16:53 +0100)]
ia32: prefere != over < or > relation where possible

13 years agofix amd64 backend since negated_relation may return unordered now
Matthias Braun [Mon, 28 Feb 2011 14:43:19 +0000 (15:43 +0100)]
fix amd64 backend since negated_relation may return unordered now

13 years agofix confirms being broken with Cmp changes
Matthias Braun [Mon, 28 Feb 2011 14:36:38 +0000 (15:36 +0100)]
fix confirms being broken with Cmp changes

13 years agoperform gigo first so no optimization sees bad inputs with wrong mode
Matthias Braun [Mon, 28 Feb 2011 12:26:25 +0000 (13:26 +0100)]
perform gigo first so no optimization sees bad inputs with wrong mode

13 years agofix mux_allowed expecting Proj(Cmp) instead of Cmp
Matthias Braun [Mon, 28 Feb 2011 12:26:10 +0000 (13:26 +0100)]
fix mux_allowed expecting Proj(Cmp) instead of Cmp

13 years agoFixed warning about strange position of be_Keeps in the schedule.
Sebastian Buchwald [Wed, 23 Feb 2011 16:04:44 +0000 (17:04 +0100)]
Fixed warning about strange position of be_Keeps in the schedule.

The warning arises from the removal of the barrier before peephole
optimization. An operand of the barrier is now kept by a new be_Keep
which is scheduled directly after the operand.

13 years agoFixed some typos.
Sebastian Buchwald [Thu, 17 Feb 2011 09:38:51 +0000 (10:38 +0100)]
Fixed some typos.

13 years agostarted adding a relation to the Cmp node
Matthias Braun [Mon, 31 Jan 2011 00:02:17 +0000 (01:02 +0100)]
started adding a relation to the Cmp node

13 years agoSimplified the code and fixed a comment.
Michael Beck [Thu, 24 Feb 2011 22:36:40 +0000 (22:36 +0000)]
Simplified the code and fixed a comment.

No functional change.

[r28460]

13 years agocombo uses now the node_cmp_attr_func for lambda.opcode.
Michael Beck [Thu, 24 Feb 2011 22:21:01 +0000 (22:21 +0000)]
combo uses now the node_cmp_attr_func for lambda.opcode.

This simplifies the code and removes some divergencies between the CSE and
combo.

[r28459]

13 years agoConstify node_cmp_attr_func.
Michael Beck [Thu, 24 Feb 2011 22:20:50 +0000 (22:20 +0000)]
Constify node_cmp_attr_func.

[r28458]

13 years agoSplit the divmod attribute into one div and one mod attribute.
Michael Beck [Thu, 24 Feb 2011 22:20:25 +0000 (22:20 +0000)]
Split the divmod attribute into one div and one mod attribute.

[r28457]

13 years agogeneralize bittest pattern
Matthias Braun [Thu, 24 Feb 2011 18:41:52 +0000 (19:41 +0100)]
generalize bittest pattern

13 years agodo gigo earlier to avoid optimisations seeing bad inputs with wrong more
Matthias Braun [Thu, 24 Feb 2011 18:17:02 +0000 (19:17 +0100)]
do gigo earlier to avoid optimisations seeing bad inputs with wrong more

13 years agofix a bunch of whitespace errors in headerfiles
Matthias Braun [Thu, 24 Feb 2011 12:00:10 +0000 (13:00 +0100)]
fix a bunch of whitespace errors in headerfiles

13 years agoUse Lea instead of Add in the ia32 transformer.
Christoph Mallon [Thu, 24 Feb 2011 14:30:17 +0000 (14:30 +0000)]
Use Lea instead of Add in the ia32 transformer.

[r28454]

13 years agoia32_Lea cannot use immediates, set the offset instead.
Christoph Mallon [Thu, 24 Feb 2011 14:30:15 +0000 (14:30 +0000)]
ia32_Lea cannot use immediates, set the offset instead.

[r28453]

13 years agofix broken bt instruction after ia32 cc rewrite
Matthias Braun [Thu, 24 Feb 2011 10:29:09 +0000 (10:29 +0000)]
fix broken bt instruction after ia32 cc rewrite

[r28447]

13 years agoFix beabi call sorting
Andreas Zwinkau [Tue, 22 Feb 2011 10:24:27 +0000 (10:24 +0000)]
Fix beabi call sorting

Difference of idx for Calls with no order relation is stable,
but may lead to circular dependencies. Fixed by looking at
the heights first.

Fixes backend/transform_loop.c (with -O0)

[r28436]

13 years agoremove confusing, unused and broken function
Matthias Braun [Mon, 21 Feb 2011 15:52:05 +0000 (15:52 +0000)]
remove confusing, unused and broken function

[r28435]

13 years agoSkip shift dependency if no other users
Andreas Zwinkau [Mon, 21 Feb 2011 15:27:12 +0000 (15:27 +0000)]
Skip shift dependency if no other users

If a shift operation has a dependency operand (64bit shifts),
then instruction selection inserts a dependency edge between them,
so the scheduler orders them correctly. However, if the operand
is not used, the scheduler is confused about the dependency edge, which
is never resolved.

Fixes opt/longlong.c testcase

[r28434]

13 years agoFix opt/fehler180
Andreas Zwinkau [Mon, 21 Feb 2011 12:58:44 +0000 (12:58 +0000)]
Fix opt/fehler180

Respect block dominance in localopt. Node a may not be live in b's block.

[r28430]

13 years agofix wrong usage of ircons functions
Matthias Braun [Mon, 21 Feb 2011 11:44:34 +0000 (11:44 +0000)]
fix wrong usage of ircons functions

[r28426]

13 years agoBugFix: ensure that two ASM nodes are never congruent in combo.
Michael Beck [Fri, 18 Feb 2011 22:44:01 +0000 (22:44 +0000)]
BugFix: ensure that two ASM nodes are never congruent in combo.

Without parsing the text of a ASM node it is not possible to decide whether two
ASM nodes are congruent. Don't even try it.

This fixes testsuite/opt/comboattrs.c

[r28422]

13 years agoRemoved C99 feature, adapt to Firm coding conventions.
Michael Beck [Fri, 18 Feb 2011 22:22:29 +0000 (22:22 +0000)]
Removed C99 feature, adapt to Firm coding conventions.

[r28421]

13 years agoFix opt/git_error.c testcase
Andreas Zwinkau [Fri, 18 Feb 2011 14:24:25 +0000 (14:24 +0000)]
Fix opt/git_error.c testcase

Extend the fix_adds_and_subs pass after strength reduction, which
inserts Convs to make the verifier happy. In case the index
variable is inverted (i=1-i) on each loop iteration, the optimization
may yield a SubP(ConstIu,PhiP), which is rejected by the verifier.

[r28409]

13 years agoFixed iterations broken with size_t refactoring.
Michael Beck [Thu, 17 Feb 2011 22:20:23 +0000 (22:20 +0000)]
Fixed iterations broken with size_t refactoring.

Some iterations must be done backwards.

[r28401]

13 years agoFinally fix the "wrong" assert of r28379.
Michael Beck [Thu, 17 Feb 2011 20:46:37 +0000 (20:46 +0000)]
Finally fix the "wrong" assert of r28379.

Introduce a get_irn_resmode() and use it at all places in combo where the
part of the opcode is meaned.

[r28400]

13 years agoFixes testcase opt/gzip_error
Andreas Zwinkau [Thu, 17 Feb 2011 12:56:12 +0000 (12:56 +0000)]
Fixes testcase opt/gzip_error

The backend sometimes creates ProjM with a temporarily broken pn,
so the verifier must not check them.

[r28392]

13 years agofixed broken builtins and added some comments
Matthias Braun [Thu, 17 Feb 2011 12:52:36 +0000 (12:52 +0000)]
fixed broken builtins and added some comments

[r28391]

13 years agoRevert reversion of r28379. Maybe this way it's clear why the assert was wrong.
Matthias Braun [Thu, 17 Feb 2011 00:13:54 +0000 (00:13 +0000)]
Revert reversion of r28379. Maybe this way it's clear why the assert was wrong.

[r28387]

13 years agoreworked ia32 condition code handling, remove backend Bound support (really someone...
Matthias Braun [Thu, 17 Feb 2011 00:13:53 +0000 (00:13 +0000)]
reworked ia32 condition code handling, remove backend Bound support (really someone should lower it before the backend

[r28386]

13 years agoRevert r28379.
Michael Beck [Wed, 16 Feb 2011 21:08:35 +0000 (21:08 +0000)]
Revert r28379.

This optimization is unsafe and breaks Load-Store-optimization.

It is not enough that both loads are connected by the memory edge,
if one Load has an exception-edge, they cannot be parallelized, because
this would destroy exception order.
load-store-opt handle this case and many others, but expect all
Loads in order yet.
opt_parallelize_mem() does this optimization in a wider range.

[r28385]

13 years agoRevert r28379.
Michael Beck [Wed, 16 Feb 2011 20:56:43 +0000 (20:56 +0000)]
Revert r28379.

All leaders of a partition must have the same opcode, which is the pair
of <irn_opcode, ir_mode> in Firm.

[r28384]

13 years agofix docu
Matthias Braun [Wed, 16 Feb 2011 17:43:15 +0000 (17:43 +0000)]
fix docu

[r28383]

13 years agoimprove bittest normalization
Matthias Braun [Wed, 16 Feb 2011 15:41:29 +0000 (15:41 +0000)]
improve bittest normalization

[r28382]

13 years agonormalize some bittest constructs
Matthias Braun [Wed, 16 Feb 2011 15:39:24 +0000 (15:39 +0000)]
normalize some bittest constructs

[r28381]

13 years agoRemove broken assert
Andreas Zwinkau [Wed, 16 Feb 2011 15:00:19 +0000 (15:00 +0000)]
Remove broken assert

[r28379]

13 years agoNew localopt: parallelize non-volatile loads
Andreas Zwinkau [Wed, 16 Feb 2011 15:00:18 +0000 (15:00 +0000)]
New localopt: parallelize non-volatile loads

[r28378]

13 years agoRemoved: list_sched_single_block
Andreas Zwinkau [Wed, 16 Feb 2011 12:17:16 +0000 (12:17 +0000)]
Removed: list_sched_single_block

Was used by the (removed) ILP scheduler. Dead code now.

[r28377]

13 years agoFurther spread size_t.
Michael Beck [Fri, 11 Feb 2011 15:21:45 +0000 (15:21 +0000)]
Further spread size_t.

[r28375]

13 years agoFixed comments.
Michael Beck [Fri, 11 Feb 2011 15:18:07 +0000 (15:18 +0000)]
Fixed comments.

[r28374]

13 years agoFixed backward iteration using size_t again.
Michael Beck [Fri, 11 Feb 2011 15:17:57 +0000 (15:17 +0000)]
Fixed backward iteration using size_t again.

[r28373]

13 years agouse long for now so at least not half the testsuite is broken
Matthias Braun [Fri, 11 Feb 2011 15:10:15 +0000 (15:10 +0000)]
use long for now so at least not half the testsuite is broken

[r28372]

13 years agotarval_quo has been removed in favor of tarval_div
Matthias Braun [Fri, 11 Feb 2011 15:04:34 +0000 (15:04 +0000)]
tarval_quo has been removed in favor of tarval_div

[r28371]

13 years agoSome size_t related fixes.
Michael Beck [Fri, 11 Feb 2011 14:03:03 +0000 (14:03 +0000)]
Some size_t related fixes.

[r28369]

13 years agoSimplified implementation by removing unused args and impossible conditions.
Michael Beck [Fri, 11 Feb 2011 14:02:53 +0000 (14:02 +0000)]
Simplified implementation by removing unused args and impossible conditions.

There seems to be no "open" methods at all (at least now, this concept seems
to be dead).

[r28368]

13 years agocomparing timesteps in different blocks is pointless and here more conservative than...
Matthias Braun [Fri, 11 Feb 2011 13:52:47 +0000 (13:52 +0000)]
comparing timesteps in different blocks is pointless and here more conservative than necessary

[r28367]

13 years agoAmendment to r28345-r28346.
Michael Beck [Fri, 11 Feb 2011 13:19:11 +0000 (13:19 +0000)]
Amendment to r28345-r28346.

Check for mode int on several places where Div nodes are handled:
- fixed arch_dep_replace_div_by_const() (works for int only)
- fixed constant folding for Div nodes (different for float and int)
- fixed transform_node_Div() (some code pathes allowed only for int)
- fixed comment and one missing check in Div mode verifycation

[r28366]

13 years agoRemoved C99 feature.
Michael Beck [Fri, 11 Feb 2011 13:18:59 +0000 (13:18 +0000)]
Removed C99 feature.

[r28365]

13 years agoWindows has no ssize_t ...
Michael Beck [Fri, 11 Feb 2011 13:18:49 +0000 (13:18 +0000)]
Windows has no ssize_t ...

[r28364]

13 years agoFixed typo.
Sebastian Buchwald [Thu, 10 Feb 2011 12:24:14 +0000 (12:24 +0000)]
Fixed typo.

[r28361]

13 years agoremove Quot node (just use Div instead)
Matthias Braun [Wed, 9 Feb 2011 13:55:18 +0000 (13:55 +0000)]
remove Quot node (just use Div instead)

[r28346]

13 years agoremove DivMod
Matthias Braun [Wed, 9 Feb 2011 13:55:16 +0000 (13:55 +0000)]
remove DivMod

[r28345]

13 years agodo not output no_dead_strip for anonymous entities on macho targets
Matthias Braun [Wed, 9 Feb 2011 13:55:14 +0000 (13:55 +0000)]
do not output no_dead_strip for anonymous entities on macho targets

[r28344]

13 years agorecursively optimize phi to get real minimal SSA-form for reducible CFGs
Matthias Braun [Wed, 9 Feb 2011 13:55:13 +0000 (13:55 +0000)]
recursively optimize phi to get real minimal SSA-form for reducible CFGs

[r28343]

13 years agoebx is no problem with pic when compiling in 64bit (but the pushl+popl are)
Matthias Braun [Wed, 9 Feb 2011 13:55:12 +0000 (13:55 +0000)]
ebx is no problem with pic when compiling in 64bit (but the pushl+popl are)

[r28342]

13 years agotranslate comment to english
Matthias Braun [Wed, 9 Feb 2011 13:55:11 +0000 (13:55 +0000)]
translate comment to english

[r28341]

13 years agofurther spread size_t (all warnings on linux/gcc fixed)
Matthias Braun [Wed, 9 Feb 2011 13:55:10 +0000 (13:55 +0000)]
further spread size_t (all warnings on linux/gcc fixed)

[r28340]

13 years agofix bugs introduced in last commit
Matthias Braun [Wed, 9 Feb 2011 13:55:08 +0000 (13:55 +0000)]
fix bugs introduced in last commit

[r28339]

13 years agoFurther spread size_t.
Michael Beck [Wed, 9 Feb 2011 00:03:31 +0000 (00:03 +0000)]
Further spread size_t.

[r28338]

13 years agofurther spread size_t
Matthias Braun [Tue, 8 Feb 2011 18:08:39 +0000 (18:08 +0000)]
further spread size_t

[r28337]

13 years agoavoid some warnings by not using macros
Matthias Braun [Tue, 8 Feb 2011 18:08:38 +0000 (18:08 +0000)]
avoid some warnings by not using macros

[r28336]

13 years agoHandle some cases of </<=/>=/> comparisons in the fixpoint vrp.
Christoph Mallon [Sun, 6 Feb 2011 18:36:30 +0000 (18:36 +0000)]
Handle some cases of </<=/>=/> comparisons in the fixpoint vrp.

[r28334]

13 years agoAdd a hint about the infamous pn_Cmp_Lg/Ne mixup in the assertion message of verify_n...
Christoph Mallon [Sun, 6 Feb 2011 18:36:27 +0000 (18:36 +0000)]
Add a hint about the infamous pn_Cmp_Lg/Ne mixup in the assertion message of verify_node_Proj_Cmp().

[r28333]

13 years agoProperly check whether sscanf() succeeded.
Christoph Mallon [Sun, 6 Feb 2011 18:36:24 +0000 (18:36 +0000)]
Properly check whether sscanf() succeeded.

[r28332]

13 years agoReplace strcpy()+strcat() song and dance by a single sprintf().
Christoph Mallon [Sun, 6 Feb 2011 18:36:21 +0000 (18:36 +0000)]
Replace strcpy()+strcat() song and dance by a single sprintf().

[r28331]

13 years agoFix several size_t related signed/unsigned warnings.
Christoph Mallon [Sun, 6 Feb 2011 18:36:17 +0000 (18:36 +0000)]
Fix several size_t related signed/unsigned warnings.

[r28330]

13 years agoRemove surplus {}.
Christoph Mallon [Sun, 6 Feb 2011 18:36:10 +0000 (18:36 +0000)]
Remove surplus {}.

[r28329]

13 years agoRemoved unused parameter from_step of be_get_next_use().
Michael Beck [Sun, 6 Feb 2011 15:05:57 +0000 (15:05 +0000)]
Removed unused parameter from_step of be_get_next_use().

Additionally
- switched the type of a visitor counter to ir_visited_t
- add some doxygen docu

[r28328]

13 years agoFixed size_t related warnings by isolating PTR_TO_INT macros.
Michael Beck [Sun, 6 Feb 2011 12:15:30 +0000 (12:15 +0000)]
Fixed size_t related warnings by isolating PTR_TO_INT macros.

Additionally:
- rename some variables to reflect its content
- add some doxygen comments

[r28327]

13 years agoSprinkle const on the in array of node constructors with variable arity.
Christoph Mallon [Sat, 5 Feb 2011 15:40:02 +0000 (15:40 +0000)]
Sprinkle const on the in array of node constructors with variable arity.

[r28310]

13 years agoFixed more size_T related warnings.
Michael Beck [Thu, 3 Feb 2011 21:51:17 +0000 (21:51 +0000)]
Fixed more size_T related warnings.

[r28292]

13 years agoBugFix: when we create a new Conv node in combo's exchange_leader(), this irn has...
Michael Beck [Thu, 3 Feb 2011 21:50:54 +0000 (21:50 +0000)]
BugFix: when we create a new Conv node in combo's exchange_leader(), this irn has no node_t.

When we create a new Conv node in combo's exchange_leader(), this irn has no node_t.
This is ok, replace the original node in a post walker.
However, as the new node has a visited count of 0 it might be revisited
again.
Fix this by copying the visited count.

This fixes testsuite/opt/fehler179.c

[r28291]

13 years agoadd #include <stddef.h>
Julius Neuffer [Thu, 3 Feb 2011 13:48:09 +0000 (13:48 +0000)]
add #include <stddef.h>

[r28289]

13 years agoCorrect typo in comment.
Christoph Mallon [Mon, 31 Jan 2011 10:18:11 +0000 (10:18 +0000)]
Correct typo in comment.

[r28287]

13 years agoRemove pointless assert: size_t nelts is always >= 0.
Christoph Mallon [Mon, 31 Jan 2011 10:18:10 +0000 (10:18 +0000)]
Remove pointless assert: size_t nelts is always >= 0.

[r28286]