Re: [PATCH 1/2] d3d10/effect: Parse through value expressions.

[Nikolay Sivov]

On 11/19/21 4:41 PM, Matteo Bruni wrote:
> On Tue, Nov 16, 2021 at 6:15 PM Nikolay Sivov  wrote:
> > Signed-off-by: Nikolay Sivov
> >
> > dlls/d3d10/d3d10_private.h |  1 +
> >  dlls/d3d10/effect.c        | 35 +++++++++++++++++
> >  dlls/d3d10/tests/effect.c  | 78 ++++++++++++++++++++++++++++++++++++++
> >  3 files changed, 114 insertions(+)
>
> Sorry for the delay.
> > diff --git a/dlls/d3d10/d3d10_private.h b/dlls/d3d10/d3d10_private.h
> > index 7f83fb1182a..b35711b55e5 100644
> > --- a/dlls/d3d10/d3d10_private.h
> > +++ b/dlls/d3d10/d3d10_private.h
> > @@ -52,6 +52,7 @@ enum d3d10_effect_object_operation
> >      D3D10_EOO_CONST_INDEX = 3,
> >      D3D10_EOO_VAR_INDEX = 4,
> >      D3D10_EOO_INDEX_EXPRESSION = 5,
> > - D3D10_EOO_VALUE_EXPRESSION = 6,
> > D3D10_EOO_ANONYMOUS_SHADER = 7,
> >
> > };
> >
> > diff --git a/dlls/d3d10/effect.c b/dlls/d3d10/effect.c
> > index 3f35e9cc28a..9c2012768dd 100644
> > --- a/dlls/d3d10/effect.c
> > +++ b/dlls/d3d10/effect.c
> > @@ -289,6 +289,10 @@ struct d3d10_effect_prop_dependency
> >              struct d3d10_effect_variable *v;
> >              struct d3d10_effect_preshader index;
> >          } index_expr;
> > -
> >    struct
> > -
> >    {
> > -
> >        struct d3d10_effect_preshader value;
> > -
> >    } value_expr;
> > };
> >
> > };
> >
> > @@ -392,6 +396,9 @@ static void d3d10_effect_clear_prop_dependencies(struct d3d10_effect_prop_depend
> >              case D3D10_EOO_INDEX_EXPRESSION:
> >                  d3d10_effect_preshader_clear(&dep->index_expr.index);
> >                  break;
> > -
> >        case D3D10_EOO_VALUE_EXPRESSION:
> > -
> >            d3d10_effect_preshader_clear(&dep->value_expr.value);
> > -
> >            break;
> >    }
> > }
> >    heap_free(d->entries);
> >
> > @@ -2632,6 +2639,34 @@ static HRESULT parse_fx10_property_assignment(const char *data, size_t data_size
> >          return d3d10_effect_add_prop_dependency(d, &dep);
> >
> > -
> >    case D3D10_EOO_VALUE_EXPRESSION:
> > -
> > -
> >        if (value_offset >= data_size || !require_space(value_offset, 1, sizeof(uint16_t), data_size))
> > -
> >        {
> > -
> >            WARN("Invalid offset %#x (data size %#lx).\n", value_offset, (long)data_size);
> > -
> >            return E_FAIL;
> > -
> >        }
>
> Pretty sure you want sizeof(DWORD) here.
> > -
> > -
> >        data_ptr = data + value_offset;
> > -
> >        read_dword(&data_ptr, &blob_size);
> > -
> > -
> >        if (!require_space(value_offset, 1, sizeof(uint16_t) + blob_size, data_size))
> > -
> >        {
> > -
> >            WARN("Invalid offset %#x (data size %#lx).\n", value_offset, (long)data_size);
> > -
> >            return E_FAIL;
> > -
> >        }
>
> Not sure about this sizeof(uint16_t) either, what's the idea?
>
> BTW, we probably want this kind of require_space(..., blob_size, ...)
> in the D3D10_EOO_INDEX_EXPRESSION case too.

Yes, it should have been uint32_t in both places.

First check is to read 4 bytes at "value_offset" offset, which would
give blob_size. Additional size in second check is to account for
blob_size that's been read already. Alternatively this could do
value_offset += 4.