Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
* Avoid gcc warning on sparc solaris 9
@ 2007-10-23 16:46 Pedro Alves
  2007-10-23 17:56 ` Ulrich Weigand
  0 siblings, 1 reply; 7+ messages in thread
From: Pedro Alves @ 2007-10-23 16:46 UTC (permalink / raw)
  To: gdb-patches

[-- Attachment #1: Type: text/plain, Size: 2454 bytes --]

Hi,

Building HEAD on a native gdb on sparc solaris 9 with gcc 3.4.6 stops due to:

gcc -c -g -O2    -I. -I../../src/gdb -I../../src/gdb/config
-DLOCALEDIR="\"/usr/local/share/locale\"" -DHAVE_CONFIG_H
-I../../src/gdb/../include/opcode -I../../src/gdb/../readline/..
-I../bfd -I../../src/gdb/../bfd -I../../src/gdb/../include -I./../intl
 -DMI_OUT=1 -DTUI=1 -I/usr/local/include -Wall
-Wdeclaration-after-statement -Wpointer-arith -Wformat-nonliteral
-Wno-unused -Wno-switch -Wno-char-subscripts -Werror
../../src/gdb/sol-thread.c
../../src/gdb/sol-thread.c: In function `sol_thread_fetch_registers':
../../src/gdb/sol-thread.c:533: warning: dereferencing type-punned
pointer will break strict-aliasing rules
make: *** [sol-thread.o] Error 1

gcc -v
Reading specs from /usr/local/lib/gcc/sparc-sun-solaris2.9/3.4.6/specs
Configured with: ../configure --with-as=/usr/ccs/bin/as
--with-ld=/usr/ccs/bin/ld --enable-shared --enable-languages=c,c++,f77
Thread model: posix
gcc version 3.4.6

uname -a
SunOS ol154 5.9 Generic_117171-13 sun4u sparc SUNW,Sun-Fire-480R Solaris

The attached patch gets me going.

config/sparc/nm-sol2.h has these, so it looks safe on sparc:

#define GDB_GREGSET_T prgregset_t
#define GDB_FPREGSET_T prfpregset_t

Here's a simplified preprocessed output of sol-thread.c:

typedef prgregset_t gdb_gregset_t;
typedef prfpregset_t gdb_fpregset_t;

extern void supply_gregset (struct regcache *regcache,
       const gdb_gregset_t *gregs);

extern void supply_fpregset (struct regcache *regcache,
        const gdb_fpregset_t *fpregs);

static void
sol_thread_fetch_registers (struct regcache *regcache, int regnum)
{
  prgregset_t gregset;
  prfpregset_t fpregset;

  supply_gregset (regcache, (const gdb_gregset_t *) &gregset);
  supply_fpregset (regcache, (const gdb_fpregset_t *) &fpregset);
}

But I'm not sure it's correct, since the i386 sparc config files don't set
GDB_{G|FP}REGSET, so they should be defaulting to:

gregset.h:

#ifndef GDB_GREGSET_T
#define GDB_GREGSET_T gregset_t
#endif

#ifndef GDB_FPREGSET_T
#define GDB_FPREGSET_T fpregset_t
#endif

I ran the testsuite on it, and this is what I got:

                === gdb Summary ===

# of expected passes            10898
# of unexpected failures        81
# of unexpected successes       2
# of expected failures          47
# of known failures             42
# of unresolved testcases       3
# of untested testcases         8
# of unsupported tests          25

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: solaris.diff --]
[-- Type: text/x-patch; name=solaris.diff, Size: 1368 bytes --]

2007-10-23  Pedro Alves  <pedro_alves@portugalmail.pt>

	* sol-thread.c (sol_thread_fetch_registers): Avoid alias warning.
	Add assertion to verify correct size of gdb_gregset_t and
	gdb_fpregset_t.

---
 gdb/sol-thread.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

Index: src/gdb/sol-thread.c
===================================================================
--- src.orig/gdb/sol-thread.c	2007-10-23 15:14:44.529038000 +0100
+++ src/gdb/sol-thread.c	2007-10-23 15:15:49.819263000 +0100
@@ -484,6 +484,10 @@ sol_thread_fetch_registers (struct regca
   td_err_e val;
   prgregset_t gregset;
   prfpregset_t fpregset;
+
+  gdb_assert (sizeof (prgregset_t) == sizeof (gdb_gregset_t));
+  gdb_assert (sizeof (prfpregset_t) == sizeof (gdb_fpregset_t));
+
 #if 0
   int xregsize;
   caddr_t xregset;
@@ -530,8 +534,8 @@ sol_thread_fetch_registers (struct regca
      calling the td routines because the td routines call ps_lget*
      which affect the values stored in the registers array.  */
 
-  supply_gregset (regcache, (const gdb_gregset_t *) &gregset);
-  supply_fpregset (regcache, (const gdb_fpregset_t *) &fpregset);
+  supply_gregset (regcache, (const gdb_gregset_t *) (void *) &gregset);
+  supply_fpregset (regcache, (const gdb_fpregset_t *) (void *) &fpregset);
 
 #if 0
   /* FIXME: libthread_db doesn't seem to handle this right.  */

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Avoid gcc warning on sparc solaris 9
  2007-10-23 16:46 Avoid gcc warning on sparc solaris 9 Pedro Alves
@ 2007-10-23 17:56 ` Ulrich Weigand
  2007-10-24 13:28   ` Pedro Alves
  0 siblings, 1 reply; 7+ messages in thread
From: Ulrich Weigand @ 2007-10-23 17:56 UTC (permalink / raw)
  To: Pedro Alves; +Cc: gdb-patches

Pedro Alves wrote:

> Building HEAD on a native gdb on sparc solaris 9 with gcc 3.4.6 stops due to:
> 
> ../../src/gdb/sol-thread.c: In function `sol_thread_fetch_registers':
> ../../src/gdb/sol-thread.c:533: warning: dereferencing type-punned
> pointer will break strict-aliasing rules

We've seen this on other platforms with gcc 3.4 as well.  I think
the best fix would be replace the problematic construct
  supply_gregset (regcache, (const gdb_gregset_t *) &gregset);
with

 gdb_gregset_t *gregset_p = &gregset;

 supply_gregset (regcache, (const gdb_gregset_t *) gregset_p);

That seemed to fix the problems e.g. in i386-linux-nat.c (fetch_regs).


Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Avoid gcc warning on sparc solaris 9
  2007-10-23 17:56 ` Ulrich Weigand
@ 2007-10-24 13:28   ` Pedro Alves
  2007-10-24 13:45     ` Ulrich Weigand
  0 siblings, 1 reply; 7+ messages in thread
From: Pedro Alves @ 2007-10-24 13:28 UTC (permalink / raw)
  To: Ulrich Weigand; +Cc: gdb-patches

[-- Attachment #1: Type: text/plain, Size: 871 bytes --]

Ulrich Weigand wrote:
> Pedro Alves wrote:
>
> > Building HEAD on a native gdb on sparc solaris 9 with gcc 3.4.6 stops due to:
> >
> > ../../src/gdb/sol-thread.c: In function `sol_thread_fetch_registers':
> > ../../src/gdb/sol-thread.c:533: warning: dereferencing type-punned
> > pointer will break strict-aliasing rules
>
> We've seen this on other platforms with gcc 3.4 as well.  I think
> the best fix would be replace the problematic construct
>   supply_gregset (regcache, (const gdb_gregset_t *) &gregset);
> with
>
>  gdb_gregset_t *gregset_p = &gregset;
>
>  supply_gregset (regcache, (const gdb_gregset_t *) gregset_p);
>
> That seemed to fix the problems e.g. in i386-linux-nat.c (fetch_regs).
>

Indeed it does fix it here too.  Much nicer.  Thanks for the pointer.

Testsuite shows the same results as the previous version.

OK now?

-- 
Cheers,
Pedro Alves

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: solaris.diff --]
[-- Type: text/x-patch; name=solaris.diff, Size: 1244 bytes --]

2007-10-24  Pedro Alves  <pedro_alves@portugalmail.pt>

	* sol-thread.c (sol_thread_fetch_registers): Work around gcc 3.4
	alias warning bug.

---
 gdb/sol-thread.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

Index: src/gdb/sol-thread.c
===================================================================
--- src.orig/gdb/sol-thread.c	2007-10-23 17:07:07.604730000 +0100
+++ src/gdb/sol-thread.c	2007-10-24 13:35:37.380474000 +0100
@@ -484,6 +484,9 @@ sol_thread_fetch_registers (struct regca
   td_err_e val;
   prgregset_t gregset;
   prfpregset_t fpregset;
+  gdb_gregset_t *gregset_p = &gregset;
+  gdb_fpregset_t *fpregset_p = &fpregset;
+
 #if 0
   int xregsize;
   caddr_t xregset;
@@ -530,8 +533,8 @@ sol_thread_fetch_registers (struct regca
      calling the td routines because the td routines call ps_lget*
      which affect the values stored in the registers array.  */
 
-  supply_gregset (regcache, (const gdb_gregset_t *) &gregset);
-  supply_fpregset (regcache, (const gdb_fpregset_t *) &fpregset);
+  supply_gregset (regcache, (const gdb_gregset_t *) gregset_p);
+  supply_fpregset (regcache, (const gdb_fpregset_t *) fpregset_p);
 
 #if 0
   /* FIXME: libthread_db doesn't seem to handle this right.  */

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Avoid gcc warning on sparc solaris 9
  2007-10-24 13:28   ` Pedro Alves
@ 2007-10-24 13:45     ` Ulrich Weigand
  2007-10-24 13:48       ` Daniel Jacobowitz
  0 siblings, 1 reply; 7+ messages in thread
From: Ulrich Weigand @ 2007-10-24 13:45 UTC (permalink / raw)
  To: Pedro Alves; +Cc: gdb-patches

Pedro Alves wrote:

>2007-10-24  Pedro Alves  <pedro_alves@portugalmail.pt>
>
>	* sol-thread.c (sol_thread_fetch_registers): Work around gcc 3.4
>	alias warning bug.

This is OK, thanks.

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Avoid gcc warning on sparc solaris 9
  2007-10-24 13:45     ` Ulrich Weigand
@ 2007-10-24 13:48       ` Daniel Jacobowitz
  2007-10-24 13:58         ` Pedro Alves
  0 siblings, 1 reply; 7+ messages in thread
From: Daniel Jacobowitz @ 2007-10-24 13:48 UTC (permalink / raw)
  To: Ulrich Weigand; +Cc: Pedro Alves, gdb-patches

On Wed, Oct 24, 2007 at 03:39:38PM +0200, Ulrich Weigand wrote:
> Pedro Alves wrote:
> 
> >2007-10-24  Pedro Alves  <pedro_alves@portugalmail.pt>
> >
> >	* sol-thread.c (sol_thread_fetch_registers): Work around gcc 3.4
> >	alias warning bug.
> 
> This is OK, thanks.

This is also PR gdb/2341.  Pedro, could you add that to the
changelog, and also commit to the 6.7 branch?  Or ask me if you don't
have it checked out and I can handle it.

-- 
Daniel Jacobowitz
CodeSourcery


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Avoid gcc warning on sparc solaris 9
  2007-10-24 13:48       ` Daniel Jacobowitz
@ 2007-10-24 13:58         ` Pedro Alves
  2007-10-24 14:56           ` Daniel Jacobowitz
  0 siblings, 1 reply; 7+ messages in thread
From: Pedro Alves @ 2007-10-24 13:58 UTC (permalink / raw)
  To: Ulrich Weigand, Pedro Alves, gdb-patches

Daniel Jacobowitz wrote:
> On Wed, Oct 24, 2007 at 03:39:38PM +0200, Ulrich Weigand wrote:
> > Pedro Alves wrote:
> >
> > >2007-10-24  Pedro Alves  <pedro_alves@portugalmail.pt>
> > >
> > >     * sol-thread.c (sol_thread_fetch_registers): Work around gcc 3.4
> > >     alias warning bug.
> >
> > This is OK, thanks.
>
> This is also PR gdb/2341.  Pedro, could you add that to the
> changelog, and also commit to the 6.7 branch?  Or ask me if you don't
> have it checked out and I can handle it.
>

Could you please?  I don't have access to cvs from here.

Thanks!

Cheers,
Pedro Alves


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Avoid gcc warning on sparc solaris 9
  2007-10-24 13:58         ` Pedro Alves
@ 2007-10-24 14:56           ` Daniel Jacobowitz
  0 siblings, 0 replies; 7+ messages in thread
From: Daniel Jacobowitz @ 2007-10-24 14:56 UTC (permalink / raw)
  To: gdb-patches

On Wed, Oct 24, 2007 at 02:48:44PM +0100, Pedro Alves wrote:
> Could you please?  I don't have access to cvs from here.

Done.

-- 
Daniel Jacobowitz
CodeSourcery


^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2007-10-24 13:58 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-10-23 16:46 Avoid gcc warning on sparc solaris 9 Pedro Alves
2007-10-23 17:56 ` Ulrich Weigand
2007-10-24 13:28   ` Pedro Alves
2007-10-24 13:45     ` Ulrich Weigand
2007-10-24 13:48       ` Daniel Jacobowitz
2007-10-24 13:58         ` Pedro Alves
2007-10-24 14:56           ` Daniel Jacobowitz

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox