c/7325: Test "if ( a && b )" where a, b are 1-bit fields, could be better

[falk at efalk dot org]

>Number:         7325
>Category:       c
>Synopsis:       Test "if ( a && b )" where a, b are 1-bit fields, could be better
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    unassigned
>State:          open
>Class:          pessimizes-code
>Submitter-Id:   net
>Arrival-Date:   Mon Jul 15 22:46:00 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator:     falk@efalk.org
>Release:        2.96
>Organization:
>Environment:
RH 7.2
>Description:
Imagine a structure like this:

   typedef struct {
     unsigned int f1:1 ;
     unsigned int f2:1 ;
   } Foo ;

This test:

   if( foo->f1 || foo->f2 ) { ... }

Compiles into something like this:

   if( (*foo & 3) != 0 ) { ... }

Generating code like this:

   testb $3, (%ebx)
   je .L34
   ...

Which is pretty good.  However, if the test is

   if( foo->f1 && foo->f2 ) { ... }

The code generated is

   if( (*foo & 3) == 3 ) { ... }

which compiles as

   movb (%ebx), %al
   andl $3, %eax
   cmpb $3, %al
   jne .L34
   ...

which is not so hot.

It would be better if the test was translated to

   if( (*foo & ~3) == 0 ) { ... }

which would compile something like this:

   testb $-4, (%ebx)
   jne .L35
   ...
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:
----gnatsweb-attachment----
Content-Type: application/octet-stream; name="tester.c"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="tester.c"

CnR5cGVkZWYgc3RydWN0IGZvbyB7CgkgIHVuc2lnbmVkIGludCBmMToxIDsKCSAgdW5zaWduZWQg
aW50IGYyOjEgOwoJfSBGb28gOwoKdm9pZApiYXIoRm9vICpmb28pCnsKCWlmKCBmb28tPmYxIHx8
IGZvby0+ZjIgKSBiYXIxKCkgOwoJaWYoIGZvby0+ZjEgJiYgZm9vLT5mMiApIGJhcjIoKSA7Cn0K