Re: [PATCH] webservices: free the potentially allocated pointer (Coverity)

[Hans Leidekker]

On Sat, 2016-12-17 at 12:48 +0100, Marcus Meissner wrote:
> 1374356 Resource leak
>
> Signed-off-by: Marcus Meissner <marcus(a)jet.franken.de>
> ---
>  dlls/webservices/reader.c | 26 ++++++++++++++++++++++----
>  1 file changed, 22 insertions(+), 4 deletions(-)
>
> diff --git a/dlls/webservices/reader.c b/dlls/webservices/reader.c
> index d67c47a701..219683b2af 100644
> --- a/dlls/webservices/reader.c
> +++ b/dlls/webservices/reader.c
> @@ -4701,15 +4701,33 @@ static ULONG get_field_size( const WS_FIELD_DESCRIPTION *desc )
>
> static HRESULT read_param( struct reader *reader, const WS_FIELD_DESCRIPTION *desc, WS_HEAP *heap, void *ret )
>  {
> -    if (!ret && !(ret = ws_alloc_zero( heap, get_field_size(desc) ))) return WS_E_QUOTA_EXCEEDED;
> -    return read_type_struct_field( reader, desc, heap, ret, 0 );
> +    HRESULT hr;
> +    VOID    *xret = NULL;
> +
> +    if (!ret)
> +    {
> +        if (!(xret = ws_alloc_zero( heap, get_field_size(desc) ))) return WS_E_QUOTA_EXCEEDED;
> +        ret = xret;
> +    }
> +    hr = read_type_struct_field( reader, desc, heap, ret, 0 );
> +    ws_free( heap, xret);
> +    return hr;
>  }

It's not a leak. The memory is allocated on the heap passed to this function and it will
be released when the heap is destroyed. I guess Coverity doesn't see that.

[PATCH] webservices: free the potentially allocated pointer (Coverity)

[Marcus Meissner]

On Mon, Dec 19, 2016 at 09:14:16AM +0100, Hans Leidekker wrote:
> On Sat, 2016-12-17 at 12:48 +0100, Marcus Meissner wrote:
> > 1374356 Resource leak
> >
> > Signed-off-by: Marcus Meissner <marcus(a)jet.franken.de>
> > ---
> >  dlls/webservices/reader.c | 26 ++++++++++++++++++++++----
> >  1 file changed, 22 insertions(+), 4 deletions(-)
> >
> > diff --git a/dlls/webservices/reader.c b/dlls/webservices/reader.c
> > index d67c47a701..219683b2af 100644
> > --- a/dlls/webservices/reader.c
> > +++ b/dlls/webservices/reader.c
> > @@ -4701,15 +4701,33 @@ static ULONG get_field_size( const WS_FIELD_DESCRIPTION *desc )
> >
> > static HRESULT read_param( struct reader *reader, const WS_FIELD_DESCRIPTION *desc, WS_HEAP *heap, void *ret )
> >  {
> > -    if (!ret && !(ret = ws_alloc_zero( heap, get_field_size(desc) ))) return WS_E_QUOTA_EXCEEDED;
> > -    return read_type_struct_field( reader, desc, heap, ret, 0 );
> > +    HRESULT hr;
> > +    VOID    *xret = NULL;
> > +
> > +    if (!ret)
> > +    {
> > +        if (!(xret = ws_alloc_zero( heap, get_field_size(desc) ))) return WS_E_QUOTA_EXCEEDED;
> > +        ret = xret;
> > +    }
> > +    hr = read_type_struct_field( reader, desc, heap, ret, 0 );
> > +    ws_free( heap, xret);
> > +    return hr;
> >  }
>
> It's not a leak. The memory is allocated on the heap passed to this function and it will
> be released when the heap is destroyed. I guess Coverity doesn't see that.

Yes, it does not think in Heaps. Will mark the issue as invalid.

Ciao, Marcus