[PATCH 0/4] assorted cleanups

[PATCH 0/4] assorted cleanups

[Luc Van Oostenryck]

This serie is only a small collections of independent
cleanups, mainly removal of unused struct members.

Patches 1-4 have no functional changes.
Patch 4 make show_statement() slightly more useful.


*** BLURB HERE ***

Luc Van Oostenryck (4):
  cleanup: there is no 'struct phi' to allocate
  remove unused field 'multijmp' in struct statement
  remove unused field 'goto_bb' in struct statement
  fix show-parse()'s labels

 allocate.h   | 1 -
 parse.h      | 9 ---------
 show-parse.c | 6 +++---
 3 files changed, 3 insertions(+), 13 deletions(-)

-- 
2.11.0

[PATCH 1/4] cleanup: there is no 'struct phi' to allocate

[Luc Van Oostenryck]

This wasn't removed when the 'struct phi' was replaced
(commit 6e4960a3: "Remove "struct phi", replace with instruction that generates a pseudo.")

Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
---
 allocate.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/allocate.h b/allocate.h
index 9f1dc8cc4..9e51ed7da 100644
--- a/allocate.h
+++ b/allocate.h
@@ -75,7 +75,6 @@ DECLARE_ALLOCATOR(basic_block);
 DECLARE_ALLOCATOR(entrypoint);
 DECLARE_ALLOCATOR(instruction);
 DECLARE_ALLOCATOR(multijmp);
-DECLARE_ALLOCATOR(phi);
 DECLARE_ALLOCATOR(pseudo);
 
 #endif
-- 
2.11.0

[PATCH 2/4] remove unused field 'multijmp' in struct statement

[Luc Van Oostenryck]

Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
---
 parse.h | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/parse.h b/parse.h
index 195f0fdf3..f2463f5fd 100644
--- a/parse.h
+++ b/parse.h
@@ -108,11 +108,6 @@ struct statement {
 			struct expression *bb_conditional;
 			struct symbol *bb_target;
 		};
-		struct /* multijmp */ {
-			struct expression *multi_from;
-			struct expression *multi_to;
-			struct symbol *multi_target;
-		};
 		struct /* asm */ {
 			struct expression *asm_string;
 			struct expression_list *asm_outputs;
-- 
2.11.0

[PATCH 3/4] remove unused field 'goto_bb' in struct statement

[Luc Van Oostenryck]

Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
---
 parse.h | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/parse.h b/parse.h
index f2463f5fd..a2b3e3889 100644
--- a/parse.h
+++ b/parse.h
@@ -104,10 +104,6 @@ struct statement {
 			struct expression *goto_expression;
 			struct symbol_list *target_list;
 		};
-		struct /* goto_bb */ {
-			struct expression *bb_conditional;
-			struct symbol *bb_target;
-		};
 		struct /* asm */ {
 			struct expression *asm_string;
 			struct expression_list *asm_outputs;
-- 
2.11.0

[PATCH 4/4] fix show-parse()'s labels

[Luc Van Oostenryck]

show_statement() sometimes displays BBs as "L0x0" because
it uses the symbol::bb_target field which is never set when
show_statement() is usefull (before or during linearization).

Fix this by using the address of the underlaying symbol like
done when displaying others BBs (for example for the label at
the beginning of a BB).

Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
---
 show-parse.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/show-parse.c b/show-parse.c
index fb5437555..d365d737f 100644
--- a/show-parse.c
+++ b/show-parse.c
@@ -529,14 +529,14 @@ static void show_switch_statement(struct statement *stmt)
 			} else
 				printf("    what?");
 		}
-		printf(": .L%p\n", sym->bb_target);
+		printf(": .L%p\n", sym);
 	} END_FOR_EACH_PTR(sym);
 	printf("# end case table\n");
 
 	show_statement(stmt->switch_statement);
 
 	if (stmt->switch_break->used)
-		printf(".L%p:\n", stmt->switch_break->bb_target);
+		printf(".L%p:\n", stmt->switch_break);
 }
 
 static void show_symbol_decl(struct symbol_list *syms)
@@ -683,7 +683,7 @@ int show_statement(struct statement *stmt)
 			int val = show_expression(stmt->goto_expression);
 			printf("\tgoto\t\t*v%d\n", val);
 		} else {
-			printf("\tgoto\t\t.L%p\n", stmt->goto_label->bb_target);
+			printf("\tgoto\t\t.L%p\n", stmt->goto_label);
 		}
 		break;
 	case STMT_ASM:
-- 
2.11.0

Re: [PATCH 0/4] assorted cleanups

[Christopher Li]

On Thu, Feb 16, 2017 at 6:13 AM, Luc Van Oostenryck
<luc.vanoostenryck@gmail.com> wrote:
> This serie is only a small collections of independent
> cleanups, mainly removal of unused struct members.
>
> Patches 1-4 have no functional changes.
> Patch 4 make show_statement() slightly more useful.

Thanks,

Applied and pushed to sparse-next.

Do you have more patches coming? If not I am going to cut
a sparse release RC soon.

Chris

Re: [PATCH 0/4] assorted cleanups

[Luc Van Oostenryck]

On Thu, Feb 16, 2017 at 10:29:47AM +0800, Christopher Li wrote:
> On Thu, Feb 16, 2017 at 6:13 AM, Luc Van Oostenryck
> <luc.vanoostenryck@gmail.com> wrote:
> > This serie is only a small collections of independent
> > cleanups, mainly removal of unused struct members.
> >
> > Patches 1-4 have no functional changes.
> > Patch 4 make show_statement() slightly more useful.
> 
> Thanks,
> 
> Applied and pushed to sparse-next.
Thanks.
 
> Do you have more patches coming? If not I am going to cut
> a sparse release RC soon.
I have lots of patches in diverse state of completeness but only
a dozen or so really ready.

In fact I was going to send some of them while I saw your email.
And I'll probably send a few others in the coming days.

Would it be possible to have the previous sparse-next promoted
to the master branch? I could then merge my features branches
with an up-to-date official branch that I'm sure is stable before
making some testing.

Luc

Re: [PATCH 0/4] assorted cleanups

[Luc Van Oostenryck]

On Thu, Feb 16, 2017 at 3:29 AM, Christopher Li <sparse@chrisli.org> wrote:
> On Thu, Feb 16, 2017 at 6:13 AM, Luc Van Oostenryck
> <luc.vanoostenryck@gmail.com> wrote:
>> This serie is only a small collections of independent
>> cleanups, mainly removal of unused struct members.
>>
>> Patches 1-4 have no functional changes.
>> Patch 4 make show_statement() slightly more useful.
>
> Thanks,
>
> Applied and pushed to sparse-next.
>
> Do you have more patches coming? If not I am going to cut
> a sparse release RC soon.
>
> Chris

What I sent one hour ago was my last pending fix.

Luc