On 2017-12-02, at 13:20:09 +0300, Dan Carpenter wrote:
> On Fri, Dec 01, 2017 at 05:19:37PM +0000, Jeremy Sowden wrote:
> > -#define DEFAULT_PIPE_INFO \
> > -{ \
> > - {IA_CSS_BINARY_DEFAULT_FRAME_INFO}, /* output_info */ \
> > - {IA_CSS_BINARY_DEFAULT_FRAME_INFO}, /* vf_output_info */ \
> > - IA_CSS_BINARY_DEFAULT_FRAME_INFO, /* raw_output_info */ \
> > - { 0, 0}, /* output system in res */ \
> > - DEFAULT_SHADING_INFO, /* shading_info */ \
> > - DEFAULT_GRID_INFO, /* grid_info */ \
> > - 0 /* num_invalid_frames */ \
> > -}
> > +#define DEFAULT_PIPE_INFO ( \
>
> Why does this have a ( now? That can't compile can it??
It does.
> > + (struct ia_css_pipe_info) { \
> > + .output_info = {IA_CSS_BINARY_DEFAULT_FRAME_INFO}, \
> > + .vf_output_info = {IA_CSS_BINARY_DEFAULT_FRAME_INFO}, \
> > + .raw_output_info = IA_CSS_BINARY_DEFAULT_FRAME_INFO, \
> > + .output_system_in_res_info = { 0, 0 }, \
> > + .shading_info = DEFAULT_SHADING_INFO, \
> > + .grid_info = DEFAULT_GRID_INFO, \
> > + .num_invalid_frames = 0 \
> > + } \
> > +)
Checkpatch got quite shouty, e.g.:
ERROR: Macros with complex values should be enclosed in parentheses
#826: FILE: drivers/staging/media/atomisp/pci/atomisp2/css2400/isp/kernels/sdis/common/ia_css_sdis_common_types.h:215:
+#define DEFAULT_DVS_STAT_PUBLIC_DVS_GLOBAL_CFG \
+(struct dvs_stat_public_dvs_global_cfg) { \
+ .kappa = 0, \
+ .match_shift = 0, \
+ .ybin_mode = 0, \
+}
so I just wrapped all of them.
> We need to get better compile test coverage on this... :/ There are
> some others as well.
I have run a test-compilation. Some of the code doesn't get built
because it's #ifdeffed off. I did try adding -DISP2401 (which enables
most of it), that that just causes unrelated compilation failures.
J.
Attachment:
signature.asc
Description: PGP signature
