[vta/trunk] don't violate strict aliasing rules in type_hash_add

[vta/trunk] don't violate strict aliasing rules in type_hash_add

[Alexandre Oliva]

[-- Attachment #1: Type: text/plain, Size: 397 bytes --]

I noticed we cast a (void**) to (struct type_hash **) and then assign
to the dereferenced pointer.  This is wrong.  What we want is to cast
the (struct type_hash *) object to (void*) and then assign to the
dereferenced (void**), which mirrors the way we read from the hash
table: dereferencing the void** and then casting the void* to (struct
type_hash *).

Ok for the trunk?  Installing in vta.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: gcc-tree-type-hash-alias-safety.patch --]
[-- Type: text/x-patch, Size: 629 bytes --]

for  gcc/ChangeLog.vta
from  Alexandre Oliva  <aoliva@redhat.com>

	* tree.c (type_hash_add): Don't violate strict aliasing rules.

Index: gcc/tree.c
===================================================================
--- gcc/tree.c.orig	2007-11-11 20:33:13.000000000 -0200
+++ gcc/tree.c	2007-11-11 20:32:52.000000000 -0200
@@ -4685,7 +4685,7 @@ type_hash_add (hashval_t hashcode, tree 
   h->hash = hashcode;
   h->type = type;
   loc = htab_find_slot_with_hash (type_hash_table, h, hashcode, INSERT);
-  *(struct type_hash **) loc = h;
+  *loc = (void*)h;
 }
 
 /* Given TYPE, and HASHCODE its hash code, return the canonical

[-- Attachment #3: Type: text/plain, Size: 249 bytes --]


-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: [vta/trunk] don't violate strict aliasing rules in type_hash_add

[Alexandre Oliva]

On Nov 12, 2007, Alexandre Oliva <aoliva@redhat.com> wrote:

> Ok for the trunk?  Installing in vta.

> 	* tree.c (type_hash_add): Don't violate strict aliasing rules.

> Index: gcc/tree.c
> -  *(struct type_hash **) loc = h;
> +  *loc = (void*)h;

I'm going ahead and checking this in as obviously correct.

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: [vta/trunk] don't violate strict aliasing rules in type_hash_add

[Diego Novillo]

On 11/26/07 1:37 AM, Alexandre Oliva wrote:
> On Nov 12, 2007, Alexandre Oliva <aoliva@redhat.com> wrote:
> 
>> Ok for the trunk?  Installing in vta.
> 
>> 	* tree.c (type_hash_add): Don't violate strict aliasing rules.
> 
>> Index: gcc/tree.c
>> -  *(struct type_hash **) loc = h;
>> +  *loc = (void*)h;

Space after 'void'.


Diego.

Re: [vta/trunk] don't violate strict aliasing rules in type_hash_add

[Alexandre Oliva]

On Dec  5, 2007, Diego Novillo <dnovillo@google.com> wrote:

> On 11/26/07 1:37 AM, Alexandre Oliva wrote:
>> On Nov 12, 2007, Alexandre Oliva <aoliva@redhat.com> wrote:
>> 
>>> Ok for the trunk?  Installing in vta.
>> 
>>> * tree.c (type_hash_add): Don't violate strict aliasing rules.
>> 
>>> Index: gcc/tree.c
>>> -  *(struct type_hash **) loc = h;
>>> +  *loc = (void*)h;

> Space after 'void'.

Thanks, I'm checking this in.  Sorry about the delay.

for gcc/ChangeLog
from  Alexandre Oliva  <aoliva@redhat.com>

	* tree.c (type_hash_add): Fix whitespace.

Index: gcc/tree.c
===================================================================
--- gcc/tree.c	(revision 130958)
+++ gcc/tree.c	(working copy)
@@ -4677,7 +4677,7 @@
   h->hash = hashcode;
   h->type = type;
   loc = htab_find_slot_with_hash (type_hash_table, h, hashcode, INSERT);
-  *loc = (void*)h;
+  *loc = (void *)h;
 }
 
 /* Given TYPE, and HASHCODE its hash code, return the canonical


-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}