[patch][tuples] fix warning

[patch][tuples] fix warning

[Rafael Espindola]

[-- Attachment #1: Type: text/plain, Size: 370 bytes --]

The attached patch fixes an uninitialized variable warning. Committed
as obvious.

2008-02-21  Rafael Espindola  <espindola@google.com>

       * tree-tailcall.c (adjust_accumulator_values): Initialize phi.

Cheers,
-- 
Rafael Avila de Espindola

Google Ireland Ltd.
Gordon House
Barrow Street
Dublin 4
Ireland

Registered in Dublin, Ireland
Registration Number: 368047

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: fix.patch --]
[-- Type: text/x-patch; name=fix.patch, Size: 491 bytes --]

diff --git a/gcc/tree-tailcall.c b/gcc/tree-tailcall.c
index 32ebed1..cc6394b 100644
--- a/gcc/tree-tailcall.c
+++ b/gcc/tree-tailcall.c
@@ -527,7 +527,7 @@ find_tail_calls (basic_block bb, struct tailcall **ret)
 static void
 adjust_accumulator_values (gimple_stmt_iterator bsi, tree m, tree a, edge back)
 {
-  gimple stmt, phi;
+  gimple stmt, phi = NULL;
   tree var, tmp;
   tree ret_type = TREE_TYPE (DECL_RESULT (current_function_decl));
   tree a_acc_arg = a_acc, m_acc_arg = m_acc;

Re: [patch][tuples] fix warning

[Diego Novillo]

On 2/21/08 6:18 AM, Rafael Espindola wrote:
> The attached patch fixes an uninitialized variable warning. Committed
> as obvious.
> 
> 2008-02-21  Rafael Espindola  <espindola@google.com>
> 
>        * tree-tailcall.c (adjust_accumulator_values): Initialize phi.

PHI should be initialized before each of those loops in the function 
body.  The call to add_phi_arg looks like it's always expecting the 
block to have PHI nodes, and that we actually found a PHI node.  Please 
put an assert before that and initialize PHI before each loop.

The file hasn't been converted yet, so I guess it's not too important atm.


Thanks.  Diego.

Re: [patch][tuples] fix warning

[Rafael Espindola]

[-- Attachment #1: Type: text/plain, Size: 612 bytes --]

> PHI should be initialized before each of those loops in the function
>  body.  The call to add_phi_arg looks like it's always expecting the
>  block to have PHI nodes, and that we actually found a PHI node.  Please
>  put an assert before that and initialize PHI before each loop.
>
>  The file hasn't been converted yet, so I guess it's not too important atm.

Please review the attached patch. It is compiling right now.

>
>  Thanks.  Diego.
>

Thanks,
-- 
Rafael Avila de Espindola

Google Ireland Ltd.
Gordon House
Barrow Street
Dublin 4
Ireland

Registered in Dublin, Ireland
Registration Number: 368047

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: fix.patch --]
[-- Type: text/x-patch; name=fix.patch, Size: 1361 bytes --]

diff --git a/gcc/tree-tailcall.c b/gcc/tree-tailcall.c
index cc6394b..ab539d0 100644
--- a/gcc/tree-tailcall.c
+++ b/gcc/tree-tailcall.c
@@ -527,7 +527,7 @@ find_tail_calls (basic_block bb, struct tailcall **ret)
 static void
 adjust_accumulator_values (gimple_stmt_iterator bsi, tree m, tree a, edge back)
 {
-  gimple stmt, phi = NULL;
+  gimple stmt, phi;
   tree var, tmp;
   tree ret_type = TREE_TYPE (DECL_RESULT (current_function_decl));
   tree a_acc_arg = a_acc, m_acc_arg = m_acc;
@@ -577,6 +577,7 @@ adjust_accumulator_values (gimple_stmt_iterator bsi, tree m, tree a, edge back)
   if (a_acc)
     { 
       phis = phi_nodes (back->dest);
+      phi = NULL;
       for (psi = gsi_start (phis); !gsi_end_p (psi); gsi_next (&psi))
 	{
 	  phi = gsi_stmt (psi);
@@ -584,12 +585,14 @@ adjust_accumulator_values (gimple_stmt_iterator bsi, tree m, tree a, edge back)
 	    break;
 	}
 
+      gcc_assert (phi != NULL);
       add_phi_arg (phi, a_acc_arg, back);
     }
 
   if (m_acc)
     {
       phis = phi_nodes (back->dest);
+      phi = NULL;
       for (psi = gsi_start (phis); !gsi_end_p (psi); gsi_next (&psi))
 	{
 	  phi = gsi_stmt (psi);
@@ -597,6 +600,7 @@ adjust_accumulator_values (gimple_stmt_iterator bsi, tree m, tree a, edge back)
 	    break;
 	}
 
+      gcc_assert (phi != NULL);
       add_phi_arg (phi, m_acc_arg, back);
     }
 }

Re: [patch][tuples] fix warning

[Diego Novillo]

On 2/21/08 9:25 AM, Rafael Espindola wrote:

> Please review the attached patch. It is compiling right now.

Looks good.  Thanks.


Diego.