1. Patchew
  2. QEMU
  3. View series

[PATCH] plugins: fix -Werror=maybe-uninitialized false-positive

marcandre.lureau@redhat.com posted 1 patch 2 months, 3 weeks ago 
contrib/plugins/cache.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
[PATCH] plugins: fix -Werror=maybe-uninitialized false-positive
Posted by marcandre.lureau@redhat.com 2 months, 3 weeks ago 
From: Marc-André Lureau <marcandre.lureau@redhat.com>

../contrib/plugins/cache.c:638:9: error: ‘l2_cache’ may be used uninitialized [-Werror=maybe-uninitialized]
  638 |         append_stats_line(rep, l1_dmem_accesses, l1_dmisses,
      |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Is a false-positive, since cores > 1, so the variable is set in the
above loop.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 contrib/plugins/cache.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/contrib/plugins/cache.c b/contrib/plugins/cache.c
index 512ef6776b..c2c274cfcd 100644
--- a/contrib/plugins/cache.c
+++ b/contrib/plugins/cache.c
@@ -609,7 +609,7 @@ static int l2_cmp(gconstpointer a, gconstpointer b)
 static void log_stats(void)
 {
     int i;
-    Cache *icache, *dcache, *l2_cache;
+    Cache *icache, *dcache, *l2_cache = NULL;
 
     g_autoptr(GString) rep = g_string_new("core #, data accesses, data misses,"
                                           " dmiss rate, insn accesses,"
-- 
2.47.0
Re: [PATCH] plugins: fix -Werror=maybe-uninitialized false-positive
Posted by Alex Bennée 2 months ago 
marcandre.lureau@redhat.com writes:

> From: Marc-André Lureau <marcandre.lureau@redhat.com>
>
> ../contrib/plugins/cache.c:638:9: error: ‘l2_cache’ may be used uninitialized [-Werror=maybe-uninitialized]
>   638 |         append_stats_line(rep, l1_dmem_accesses, l1_dmisses,
>       |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> Is a false-positive, since cores > 1, so the variable is set in the
> above loop.
>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>

Queued to plugins/next, thanks.

-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro
Re: [PATCH] plugins: fix -Werror=maybe-uninitialized false-positive
Posted by Pierrick Bouvier 2 months, 2 weeks ago 
On 1/14/25 02:48, marcandre.lureau@redhat.com wrote:
> From: Marc-André Lureau <marcandre.lureau@redhat.com>
> 
> ../contrib/plugins/cache.c:638:9: error: ‘l2_cache’ may be used uninitialized [-Werror=maybe-uninitialized]
>    638 |         append_stats_line(rep, l1_dmem_accesses, l1_dmisses,
>        |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> Is a false-positive, since cores > 1, so the variable is set in the
> above loop.
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   contrib/plugins/cache.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/contrib/plugins/cache.c b/contrib/plugins/cache.c
> index 512ef6776b..c2c274cfcd 100644
> --- a/contrib/plugins/cache.c
> +++ b/contrib/plugins/cache.c
> @@ -609,7 +609,7 @@ static int l2_cmp(gconstpointer a, gconstpointer b)
>   static void log_stats(void)
>   {
>       int i;
> -    Cache *icache, *dcache, *l2_cache;
> +    Cache *icache, *dcache, *l2_cache = NULL;
>   
>       g_autoptr(GString) rep = g_string_new("core #, data accesses, data misses,"
>                                             " dmiss rate, insn accesses,"

By the way,

which compiler (version?) are you using?
Just curious to check if we don't have any other false positives in the 
code base with it if it's a newer version.
Re: [PATCH] plugins: fix -Werror=maybe-uninitialized false-positive
Posted by Marc-André Lureau 2 months, 2 weeks ago 
Hi

On Tue, Jan 14, 2025 at 8:00 PM Pierrick Bouvier
<pierrick.bouvier@linaro.org> wrote:
>
> On 1/14/25 02:48, marcandre.lureau@redhat.com wrote:
> > From: Marc-André Lureau <marcandre.lureau@redhat.com>
> >
> > ../contrib/plugins/cache.c:638:9: error: ‘l2_cache’ may be used uninitialized [-Werror=maybe-uninitialized]
> >    638 |         append_stats_line(rep, l1_dmem_accesses, l1_dmisses,
> >        |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >
> > Is a false-positive, since cores > 1, so the variable is set in the
> > above loop.
> >
> > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > ---
> >   contrib/plugins/cache.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/contrib/plugins/cache.c b/contrib/plugins/cache.c
> > index 512ef6776b..c2c274cfcd 100644
> > --- a/contrib/plugins/cache.c
> > +++ b/contrib/plugins/cache.c
> > @@ -609,7 +609,7 @@ static int l2_cmp(gconstpointer a, gconstpointer b)
> >   static void log_stats(void)
> >   {
> >       int i;
> > -    Cache *icache, *dcache, *l2_cache;
> > +    Cache *icache, *dcache, *l2_cache = NULL;
> >
> >       g_autoptr(GString) rep = g_string_new("core #, data accesses, data misses,"
> >                                             " dmiss rate, insn accesses,"
>
> By the way,
>
> which compiler (version?) are you using?
> Just curious to check if we don't have any other false positives in the
> code base with it if it's a newer version.


GCC v14.2.1 from fc41.
I use '--enable-debug' '--enable-asan' '-Doptimization=g'
'-Ddebug=true' in my configure flags.
All targets do not compile with those flags.
ASAN reports various errors during make test.

Help welcome!
Re: [PATCH] plugins: fix -Werror=maybe-uninitialized false-positive
Posted by Pierrick Bouvier 2 months, 2 weeks ago 
Hi Marc-André,

I could successfully build all targets with gcc 14.2.0 (using a debian 
sid container) with ubsan and asan enabled, and with -O2.

Pierrick

On 1/14/25 22:40, Marc-André Lureau wrote:
> Hi
> 
> On Tue, Jan 14, 2025 at 8:00 PM Pierrick Bouvier
> <pierrick.bouvier@linaro.org> wrote:
>>
>> On 1/14/25 02:48, marcandre.lureau@redhat.com wrote:
>>> From: Marc-André Lureau <marcandre.lureau@redhat.com>
>>>
>>> ../contrib/plugins/cache.c:638:9: error: ‘l2_cache’ may be used uninitialized [-Werror=maybe-uninitialized]
>>>     638 |         append_stats_line(rep, l1_dmem_accesses, l1_dmisses,
>>>         |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>>
>>> Is a false-positive, since cores > 1, so the variable is set in the
>>> above loop.
>>>
>>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>>> ---
>>>    contrib/plugins/cache.c | 2 +-
>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/contrib/plugins/cache.c b/contrib/plugins/cache.c
>>> index 512ef6776b..c2c274cfcd 100644
>>> --- a/contrib/plugins/cache.c
>>> +++ b/contrib/plugins/cache.c
>>> @@ -609,7 +609,7 @@ static int l2_cmp(gconstpointer a, gconstpointer b)
>>>    static void log_stats(void)
>>>    {
>>>        int i;
>>> -    Cache *icache, *dcache, *l2_cache;
>>> +    Cache *icache, *dcache, *l2_cache = NULL;
>>>
>>>        g_autoptr(GString) rep = g_string_new("core #, data accesses, data misses,"
>>>                                              " dmiss rate, insn accesses,"
>>
>> By the way,
>>
>> which compiler (version?) are you using?
>> Just curious to check if we don't have any other false positives in the
>> code base with it if it's a newer version.
> 
> 
> GCC v14.2.1 from fc41.
> I use '--enable-debug' '--enable-asan' '-Doptimization=g'
> '-Ddebug=true' in my configure flags.
> All targets do not compile with those flags.
> ASAN reports various errors during make test.
> 
> Help welcome!
>
Re: [PATCH] plugins: fix -Werror=maybe-uninitialized false-positive
Posted by Pierrick Bouvier 2 months, 2 weeks ago 
Hi Marc-André,

On 1/14/25 02:48, marcandre.lureau@redhat.com wrote:
> From: Marc-André Lureau <marcandre.lureau@redhat.com>
> 
> ../contrib/plugins/cache.c:638:9: error: ‘l2_cache’ may be used uninitialized [-Werror=maybe-uninitialized]
>    638 |         append_stats_line(rep, l1_dmem_accesses, l1_dmisses,
>        |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> Is a false-positive, since cores > 1, so the variable is set in the
> above loop.
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   contrib/plugins/cache.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/contrib/plugins/cache.c b/contrib/plugins/cache.c
> index 512ef6776b..c2c274cfcd 100644
> --- a/contrib/plugins/cache.c
> +++ b/contrib/plugins/cache.c
> @@ -609,7 +609,7 @@ static int l2_cmp(gconstpointer a, gconstpointer b)
>   static void log_stats(void)
>   {
>       int i;
> -    Cache *icache, *dcache, *l2_cache;
> +    Cache *icache, *dcache, *l2_cache = NULL;
>   
>       g_autoptr(GString) rep = g_string_new("core #, data accesses, data misses,"
>                                             " dmiss rate, insn accesses,"

It does not hurt to initialize it, even though it's a false positive.
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>

Thanks,
Pierrick

Patchew 2.1-devel-b67e4c2

© 2016 - 2025 Red Hat, Inc.