[PATCH v2] softmmu/physmem: Silence GCC 10 maybe-uninitialized error

[PATCH v2] softmmu/physmem: Silence GCC 10 maybe-uninitialized error

[Philippe Mathieu-Daudé]

When building with GCC 10.2 configured with --extra-cflags=-Os, we get:

  softmmu/physmem.c: In function ‘address_space_translate_for_iotlb’:
  softmmu/physmem.c:643:26: error: ‘notifier’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
    643 |         notifier->active = true;
        |                          ^
  softmmu/physmem.c:608:23: note: ‘notifier’ was declared here
    608 |     TCGIOMMUNotifier *notifier;
        |                       ^~~~~~~~

Initialize 'notifier' to silence the warning.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
---
v2: Remove pointless assert (Peter Maydell)

Yet another hole in our CI.

Supersedes: <20210117160754.4086411-1-f4bug@amsat.org>
---
 softmmu/physmem.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/softmmu/physmem.c b/softmmu/physmem.c
index 6301f4f0a5c..cdcd197656f 100644
--- a/softmmu/physmem.c
+++ b/softmmu/physmem.c
@@ -605,7 +605,7 @@ static void tcg_register_iommu_notifier(CPUState *cpu,
      * when the IOMMU tells us the mappings we've cached have changed.
      */
     MemoryRegion *mr = MEMORY_REGION(iommu_mr);
-    TCGIOMMUNotifier *notifier;
+    TCGIOMMUNotifier *notifier = NULL;
     int i;
 
     for (i = 0; i < cpu->iommu_notifiers->len; i++) {
-- 
2.26.2

Re: [PATCH v2] softmmu/physmem: Silence GCC 10 maybe-uninitialized error

[Paolo Bonzini]

On 17/01/21 18:04, Philippe Mathieu-Daudé wrote:
> When building with GCC 10.2 configured with --extra-cflags=-Os, we get:
> 
>    softmmu/physmem.c: In function ‘address_space_translate_for_iotlb’:
>    softmmu/physmem.c:643:26: error: ‘notifier’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
>      643 |         notifier->active = true;
>          |                          ^
>    softmmu/physmem.c:608:23: note: ‘notifier’ was declared here
>      608 |     TCGIOMMUNotifier *notifier;
>          |                       ^~~~~~~~
> 
> Initialize 'notifier' to silence the warning.
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> ---
> v2: Remove pointless assert (Peter Maydell)
> 
> Yet another hole in our CI.
> 
> Supersedes: <20210117160754.4086411-1-f4bug@amsat.org>
> ---
>   softmmu/physmem.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/softmmu/physmem.c b/softmmu/physmem.c
> index 6301f4f0a5c..cdcd197656f 100644
> --- a/softmmu/physmem.c
> +++ b/softmmu/physmem.c
> @@ -605,7 +605,7 @@ static void tcg_register_iommu_notifier(CPUState *cpu,
>        * when the IOMMU tells us the mappings we've cached have changed.
>        */
>       MemoryRegion *mr = MEMORY_REGION(iommu_mr);
> -    TCGIOMMUNotifier *notifier;
> +    TCGIOMMUNotifier *notifier = NULL;
>       int i;
>   
>       for (i = 0; i < cpu->iommu_notifiers->len; i++) {
> 

Queued, thanks.

Paolo

Re: [PATCH v2] softmmu/physmem: Silence GCC 10 maybe-uninitialized error

[Thomas Huth]

On 17/01/2021 18.04, Philippe Mathieu-Daudé wrote:
> When building with GCC 10.2 configured with --extra-cflags=-Os, we get:
> 
>    softmmu/physmem.c: In function ‘address_space_translate_for_iotlb’:
>    softmmu/physmem.c:643:26: error: ‘notifier’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
>      643 |         notifier->active = true;
>          |                          ^
>    softmmu/physmem.c:608:23: note: ‘notifier’ was declared here
>      608 |     TCGIOMMUNotifier *notifier;
>          |                       ^~~~~~~~
> 
> Initialize 'notifier' to silence the warning.
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> ---
> v2: Remove pointless assert (Peter Maydell)
> 
> Yet another hole in our CI.


I wouldn't call this a hole in the CI. AFAIU we don't support compiling with 
anything else than the default -O2 (and maybe -O0 for debugging?). -O3 is 
known to produce a lot of compiler warnings, and apparently -Os has such 
"problems", too. As far as I can see, it's a false positive warning here, 
"notifier" should always get initialized, the compiler just fails to see it 
correctly. Anyway, initializing the variable also can not hurt, so:

Reviewed-by: Thomas Huth <thuth@redhat.com>