Hi,
this patch finishes transition of IPA passes to be per-function (and thus
behaving sanely
at LTO time when different flags are mixed).
It also goes through common.opt and fixes attributes of quite few flags:
fauto-inc-dec, fdelete-dead-exceptions, ffunction-cse,
fgraphite, fstrict-volatile-bitfields, fira-algorithm, fira-region,
fira-share-save-slots, fira-share-spill-slots,
fmodulo-sched-allow-regmoves, fpartial-inlining,
sched-stalled-insns, fsched-stalled-insns-dep, fstrict-overflow,
ftracer, ftree-parallelize-loops, fassociative-math,
freciprocal-math, fvect-cost-model, fsimd-cost-model, flag_stack_reuse
are all function local properties and thus should be marked as Optimization.
while
fauto-profile, fcommon, fdata-sections, fipa-icf-variables,
ftoplevel-reorder, funit-at-a-time, fwhole-program
currentwly won't work in optimization attributes (and won't get properly
maintained
to LTO). This is becuase either they affect variables that are not annotated or
because
they are decided globally for whole compilation unit.
There are few cases I need to look into still:
flag-function-sections
- Here we can support optimizaiton attribute but currently don't - basically
it
needs revisiting the few flag_function_sections uses and making the
opt_for_fn.
flag_gnu_tm
flag_gnu_unique
- Those probably can be turned into Optimization but I have no clue.
flag_proflie_reorder_functions
- can be handled but isn't at the moment.
flag_split_stack
- this one needs some unit finalization
flag_stack_protector
- Probably can be supported as Optimization
flag_strict_aliasing
- This is marked as Optimization, but I blieve it does not work that way
because
alias classes are assigned to types that gets shared. It would be great to
support
this since a lot of real code is mixing the settings.
flag_toplevel_reorder
- This can be supported but isn't
Similarly to flag_proflie_reorder_functions needs a bit tweaking of lto and
cgraphunit.
fp_contract_mode
- This probalby should be optimization but doing so makes the awk machinery
to fail.
Bootstrapped/regtested x86_64-linux, will commit it later today if there are no
complains.
Honza
* ipa-reference.c (set_reference_optimization_summary,
ipa_reference_get_not_written_global): Do nothing if ipa-reference is
disabled.
(propagate_bits): If ipa-reference is disabled, do not look into local
properties.
(analyze_function): Disable analysis when ipa_reference is disabled.
(generate_summary): Do not dump when reference is disabled.
(get_read_write_all_from_node): When ipa-reference is disabled, use the
node flags.
(gate): Enable for LTO.
* optc-save-gen.awk: Handle optimize_debug correctly.
* opth-gen.awk: Likewise.
* common.opt (fauto-inc-dec, fdelete-dead-exceptions, ffunction-cse,
fgraphite, fstrict-volatile-bitfields, fira-algorithm, fira-region,
fira-share-save-slots, fira-share-spill-slots,
fmodulo-sched-allow-regmoves, fpartial-inlining,
sched-stalled-insns, fsched-stalled-insns-dep, fstrict-overflow,
ftracer, ftree-parallelize-loops, fassociative-math,
freciprocal-math, fvect-cost-model, fsimd-cost-model): Mark as
Optimization
(fauto-profile, fcommon, fdata-sections, fipa-icf-variables,
ftoplevel-reorder, funit-at-a-time, fwhole-program): Do not mark as
Optimization.
* ipa-icf.c (gate, sem_item_optimizer::filter_removed_items):
Fix for IPA.
Index: ipa-reference.c
===================================================================
--- ipa-reference.c (revision 219756)
+++ ipa-reference.c (working copy)
@@ -198,6 +198,9 @@ set_reference_optimization_summary (stru
bitmap
ipa_reference_get_not_read_global (struct cgraph_node *fn)
{
+ if (!opt_for_fn (fn->decl, flag_ipa_reference)
+ || !opt_for_fn (current_function_decl, flag_ipa_reference))
+ return NULL;
ipa_reference_optimization_summary_t info =
get_reference_optimization_summary (fn->function_symbol (NULL));
if (info)
@@ -216,6 +219,9 @@ ipa_reference_get_not_read_global (struc
bitmap
ipa_reference_get_not_written_global (struct cgraph_node *fn)
{
+ if (!opt_for_fn (fn->decl, flag_ipa_reference)
+ || !opt_for_fn (current_function_decl, flag_ipa_reference))
+ return NULL;
ipa_reference_optimization_summary_t info =
get_reference_optimization_summary (fn);
if (info)
@@ -381,8 +387,9 @@ propagate_bits (ipa_reference_global_var
/* Only look into nodes we can propagate something. */
int flags = flags_from_decl_or_type (y->decl);
- if (avail > AVAIL_INTERPOSABLE
- || (avail == AVAIL_INTERPOSABLE && (flags & ECF_LEAF)))
+ if (opt_for_fn (y->decl, flag_ipa_reference)
+ && (avail > AVAIL_INTERPOSABLE
+ || (avail == AVAIL_INTERPOSABLE && (flags & ECF_LEAF))))
{
if (get_reference_vars_info (y))
{
@@ -474,6 +481,8 @@ analyze_function (struct cgraph_node *fn
int i;
tree var;
+ if (!opt_for_fn (fn->decl, flag_ipa_reference))
+ return;
local = init_function_info (fn);
for (i = 0; fn->iterate_reference (i, ref); i++)
{
@@ -570,7 +579,8 @@ generate_summary (void)
if (dump_file)
FOR_EACH_DEFINED_FUNCTION (node)
- if (node->get_availability () >= AVAIL_INTERPOSABLE)
+ if (node->get_availability () >= AVAIL_INTERPOSABLE
+ && opt_for_fn (node->decl, flag_ipa_reference))
{
ipa_reference_local_vars_info_t l;
unsigned int index;
@@ -607,7 +617,7 @@ read_write_all_from_decl (struct cgraph_
tree decl = node->decl;
int flags = flags_from_decl_or_type (decl);
if ((flags & ECF_LEAF)
- && node->get_availability () <= AVAIL_INTERPOSABLE)
+ && node->get_availability () < AVAIL_INTERPOSABLE)
;
else if (flags & ECF_CONST)
;
@@ -640,7 +650,8 @@ get_read_write_all_from_node (struct cgr
struct cgraph_edge *e, *ie;
/* When function is overwritable, we can not assume anything. */
- if (node->get_availability () <= AVAIL_INTERPOSABLE)
+ if (node->get_availability () <= AVAIL_INTERPOSABLE
+ || (node->analyzed && !opt_for_fn (node->decl, flag_ipa_reference)))
read_write_all_from_decl (node, read_all, write_all);
for (e = node->callees;
@@ -650,7 +661,8 @@ get_read_write_all_from_node (struct cgr
enum availability avail;
struct cgraph_node *callee = e->callee->function_symbol (&avail);
gcc_checking_assert (callee);
- if (avail <= AVAIL_INTERPOSABLE)
+ if (avail <= AVAIL_INTERPOSABLE
+ || (callee->analyzed && !opt_for_fn (callee->decl,
flag_ipa_reference)))
read_write_all_from_decl (callee, read_all, write_all);
}
@@ -1178,7 +1190,7 @@ public:
/* opt_pass methods: */
virtual bool gate (function *)
{
- return (flag_ipa_reference
+ return ((in_lto_p || flag_ipa_reference)
/* Don't bother doing anything if the program has errors. */
&& !seen_error ());
}
Index: optc-save-gen.awk
===================================================================
--- optc-save-gen.awk (revision 219756)
+++ optc-save-gen.awk (working copy)
@@ -84,15 +84,17 @@ print "void";
print "cl_optimization_save (struct cl_optimization *ptr, struct gcc_options
*opts)";
print "{";
-n_opt_char = 2;
+n_opt_char = 3;
n_opt_short = 0;
n_opt_int = 0;
n_opt_enum = 1;
n_opt_other = 0;
var_opt_char[0] = "optimize";
var_opt_char[1] = "optimize_size";
+var_opt_char[2] = "optimize_debug";
var_opt_range["optimize"] = "0, 255";
-var_opt_range["optimize_size"] = "0, 255";
+var_opt_range["optimize_size"] = "0, 1";
+var_opt_range["optimize_debug"] = "0, 1";
var_opt_enum[0] = "flag_fp_contract_mode";
# Sort by size to mimic how the structure is laid out to be friendlier to the
@@ -734,11 +736,13 @@ for (i = 0; i < n_target_val; i++) {
print "}";
-n_opt_val = 2;
+n_opt_val = 3;
var_opt_val[0] = "x_optimize"
var_opt_val_type[0] = "char "
var_opt_val[1] = "x_optimize_size"
+var_opt_val[2] = "x_optimize_debug"
var_opt_val_type[1] = "char "
+var_opt_val_type[2] = "char "
for (i = 0; i < n_opts; i++) {
if (flag_set_p("Optimization", flags[i])) {
name = var_name(flags[i])
Index: opth-gen.awk
===================================================================
--- opth-gen.awk (revision 219756)
+++ opth-gen.awk (working copy)
@@ -132,13 +132,14 @@ print "/* Structure to save/restore opti
print "struct GTY(()) cl_optimization";
print "{";
-n_opt_char = 2;
+n_opt_char = 3;
n_opt_short = 0;
n_opt_int = 0;
n_opt_enum = 1;
n_opt_other = 0;
var_opt_char[0] = "unsigned char x_optimize";
var_opt_char[1] = "unsigned char x_optimize_size";
+var_opt_char[2] = "unsigned char x_optimize_debug";
var_opt_enum[0] = "enum fp_contract_mode x_flag_fp_contract_mode";
for (i = 0; i < n_opts; i++) {
Index: common.opt
===================================================================
--- common.opt (revision 219756)
+++ common.opt (working copy)
@@ -912,11 +912,11 @@ Common Report Var(flag_asynchronous_unwi
Generate unwind tables that are exact at each instruction boundary
fauto-inc-dec
-Common Report Var(flag_auto_inc_dec) Init(1)
+Common Report Var(flag_auto_inc_dec) Init(1) Optimization
Generate auto-inc/dec instructions
fauto-profile
-Common Report Var(flag_auto_profile) Optimization
+Common Report Var(flag_auto_profile)
Use sample profile information for call graph node weights. The default
profile file is fbdata.afdo in 'pwd'.
@@ -981,7 +981,7 @@ Common Report Var(flag_combine_stack_adj
Looks for opportunities to reduce stack adjustments and stack references.
fcommon
-Common Report Var(flag_no_common,0) Optimization
+Common Report Var(flag_no_common,0)
Do not put uninitialized globals in the common section
fcompare-debug
@@ -1029,7 +1029,7 @@ Common Report Var(flag_cx_fortran_rules)
Complex multiplication and division follow Fortran rules
fdata-sections
-Common Report Var(flag_data_sections) Optimization
+Common Report Var(flag_data_sections)
Place data items into their own section
fdbg-cnt-list
@@ -1059,7 +1059,7 @@ Common Report Var(flag_delayed_branch) O
Attempt to fill delay slots of branch instructions
fdelete-dead-exceptions
-Common Report Var(flag_delete_dead_exceptions) Init(0)
+Common Report Var(flag_delete_dead_exceptions) Init(0) Optimization
Delete dead instructions that may throw exceptions
fdelete-null-pointer-checks
@@ -1268,7 +1268,7 @@ Enum(fp_contract_mode) String(fast) Valu
; Used for compiling the Unix kernel, where strange substitutions are
; done on the assembly output.
ffunction-cse
-Common Report Var(flag_no_function_cse,0)
+Common Report Var(flag_no_function_cse,0) Optimization
Allow function addresses to be held in registers
ffunction-sections
@@ -1299,7 +1299,7 @@ has finished
; This option is not documented yet as its semantics will change.
fgraphite
-Common Report Var(flag_graphite)
+Common Report Var(flag_graphite) Opitmization
Enable in and out of Graphite representation
fgraphite-identity
@@ -1348,7 +1348,7 @@ Common Report Var(flag_loop_optimize_isl
Enable the ISL based loop nest optimizer
fstrict-volatile-bitfields
-Common Report Var(flag_strict_volatile_bitfields) Init(-1)
+Common Report Var(flag_strict_volatile_bitfields) Init(-1) Optimization
Force bitfield accesses to match their type width
fguess-branch-probability
@@ -1477,7 +1477,7 @@ Common Report Var(flag_ipa_icf_functions
Perform Identical Code Folding for functions
fipa-icf-variables
-Common Report Var(flag_ipa_icf_variables) Optimization
+Common Report Var(flag_ipa_icf_variables)
Perform Identical Code Folding for variables
fipa-reference
@@ -1493,7 +1493,7 @@ Common Ignore
Does nothing. Preserved for backward compatibility.
fira-algorithm=
-Common Joined RejectNegative Enum(ira_algorithm) Var(flag_ira_algorithm)
Init(IRA_ALGORITHM_CB)
+Common Joined RejectNegative Enum(ira_algorithm) Var(flag_ira_algorithm)
Init(IRA_ALGORITHM_CB) Optimization
-fira-algorithm=[CB|priority] Set the used IRA algorithm
Enum
@@ -1506,7 +1506,7 @@ EnumValue
Enum(ira_algorithm) String(priority) Value(IRA_ALGORITHM_PRIORITY)
fira-region=
-Common Joined RejectNegative Enum(ira_region) Var(flag_ira_region)
Init(IRA_REGION_AUTODETECT)
+Common Joined RejectNegative Enum(ira_region) Var(flag_ira_region)
Init(IRA_REGION_AUTODETECT) Optimization
-fira-region=[one|all|mixed] Set regions for IRA
Enum
@@ -1532,11 +1532,11 @@ Use IRA based register pressure calculat
in RTL loop optimizations.
fira-share-save-slots
-Common Report Var(flag_ira_share_save_slots) Init(1)
+Common Report Var(flag_ira_share_save_slots) Init(1) Optimization
Share slots for saving different hard registers.
fira-share-spill-slots
-Common Report Var(flag_ira_share_spill_slots) Init(1)
+Common Report Var(flag_ira_share_spill_slots) Init(1) Optimization
Share stack slots for spilled pseudo-registers.
fira-verbose=
@@ -1658,7 +1658,7 @@ Common Report Var(flag_modulo_sched) Opt
Perform SMS based modulo scheduling before the first scheduling pass
fmodulo-sched-allow-regmoves
-Common Report Var(flag_modulo_sched_allow_regmoves)
+Common Report Var(flag_modulo_sched_allow_regmoves) Optimization
Perform SMS based modulo scheduling with register moves allowed
fmove-loop-invariants
@@ -1719,7 +1719,7 @@ Common Report Var(flag_optimize_sibling_
Optimize sibling and tail recursive calls
fpartial-inlining
-Common Report Var(flag_partial_inlining)
+Common Report Var(flag_partial_inlining) Optimization
Perform partial inlining
fpre-ipa-mem-report
@@ -1966,7 +1966,7 @@ Common Report Var(flag_sched_stalled_ins
Allow premature scheduling of queued insns
fsched-stalled-insns=
-Common RejectNegative Joined UInteger
+Common RejectNegative Joined UInteger Optimization
-fsched-stalled-insns=<number> Set number of queued insns that can be
prematurely scheduled
; sched_stalled_insns_dep controls how many recently scheduled cycles will
@@ -1978,7 +1978,7 @@ Common Report Var(flag_sched_stalled_ins
Set dependence distance checking in premature scheduling of queued insns
fsched-stalled-insns-dep=
-Common RejectNegative Joined UInteger
+Common RejectNegative Joined UInteger Optimization
-fsched-stalled-insns-dep=<number> Set dependence distance checking in
premature scheduling of queued insns
fsched-group-heuristic
@@ -2114,7 +2114,7 @@ Common Report Var(flag_strict_aliasing)
Assume strict aliasing rules apply
fstrict-overflow
-Common Report Var(flag_strict_overflow)
+Common Report Var(flag_strict_overflow) Optimization
Treat signed overflow as undefined
fsync-libcalls
@@ -2157,11 +2157,11 @@ EnumValue
Enum(tls_model) String(local-exec) Value(TLS_MODEL_LOCAL_EXEC)
ftoplevel-reorder
-Common Report Var(flag_toplevel_reorder) Init(2) Optimization
+Common Report Var(flag_toplevel_reorder) Init(2)
Reorder top level functions, variables, and asms
ftracer
-Common Report Var(flag_tracer)
+Common Report Var(flag_tracer) Optimization
Perform superblock formation via tail duplication
; Zero means that floating-point math operations cannot generate a
@@ -2285,7 +2285,7 @@ Common Report Var(flag_tree_loop_optimiz
Enable loop optimizations on tree level
ftree-parallelize-loops=
-Common Report Joined RejectNegative UInteger Var(flag_tree_parallelize_loops)
Init(1)
+Common Report Joined RejectNegative UInteger Var(flag_tree_parallelize_loops)
Init(1) Optimization
Enable automatic parallelization of loops
ftree-phiprop
@@ -2337,7 +2337,7 @@ Common Report Var(flag_tree_vrp) Init(0)
Perform Value Range Propagation on trees
funit-at-a-time
-Common Report Var(flag_unit_at_a_time) Init(1) Optimization
+Common Report Var(flag_unit_at_a_time) Init(1)
Compile whole compilation unit at a time
funroll-loops
@@ -2356,12 +2356,12 @@ Common Report Var(flag_unsafe_loop_optim
Allow loop optimizations to assume that the loops behave in normal way
fassociative-math
-Common Report Var(flag_associative_math) SetByCombined
+Common Report Var(flag_associative_math) SetByCombined Optimization
Allow optimization for floating-point arithmetic which may change the
result of the operation due to rounding.
freciprocal-math
-Common Report Var(flag_reciprocal_math) SetByCombined
+Common Report Var(flag_reciprocal_math) SetByCombined Optimization
Same as -fassociative-math for expressions which include division.
; Nonzero means that unsafe floating-point math optimizations are allowed
@@ -2439,11 +2439,11 @@ Common Report Var(flag_tree_slp_vectoriz
Enable basic block vectorization (SLP) on trees
fvect-cost-model=
-Common Joined RejectNegative Enum(vect_cost_model) Var(flag_vect_cost_model)
Init(VECT_COST_MODEL_DEFAULT)
+Common Joined RejectNegative Enum(vect_cost_model) Var(flag_vect_cost_model)
Init(VECT_COST_MODEL_DEFAULT) Optimization
Specifies the cost model for vectorization
fsimd-cost-model=
-Common Joined RejectNegative Enum(vect_cost_model) Var(flag_simd_cost_model)
Init(VECT_COST_MODEL_UNLIMITED)
+Common Joined RejectNegative Enum(vect_cost_model) Var(flag_simd_cost_model)
Init(VECT_COST_MODEL_UNLIMITED) Optimization
Specifies the vectorization cost model for code marked with a simd directive
Enum
@@ -2540,7 +2540,7 @@ Common Report Var(flag_tree_builtin_call
Enable conditional dead code elimination for builtin calls
fwhole-program
-Common Report Var(flag_whole_program) Init(0) Optimization
+Common Report Var(flag_whole_program) Init(0)
Perform whole program optimizations
fwrapv
Index: ipa-icf.c
===================================================================
--- ipa-icf.c (revision 219756)
+++ ipa-icf.c (working copy)
@@ -1652,7 +1652,8 @@ sem_item_optimizer::filter_removed_items
{
sem_item *item = m_items[i];
- if (!flag_ipa_icf_functions && item->type == FUNC)
+ if (item->type == FUNC
+ && !opt_for_fn (item->node->decl, flag_ipa_icf_functions))
{
remove_item (item);
continue;
@@ -2499,7 +2500,7 @@ public:
/* opt_pass methods: */
virtual bool gate (function *)
{
- return flag_ipa_icf_variables || flag_ipa_icf_functions;
+ return in_lto_p || flag_ipa_icf_variables || flag_ipa_icf_functions;
}
virtual unsigned int execute (function *)
