Message ID | 1520432900-9601-1-git-send-email-patrick.delaunay@st.com |
---|---|
State | Changes Requested |
Delegated to: | Masahiro Yamada |
Headers | show |
Series | [U-Boot,v2] common: add a prototype for mach_cpu_init() | expand |
On Wed, Mar 07, 2018 at 03:28:20PM +0100, Patrick Delaunay wrote: > avoid warning: no previous prototype for ‘mach_cpu_init’ > > Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Tom Rini <trini@konsulko.com>
2018-03-07 23:45 GMT+09:00 Tom Rini <trini@konsulko.com>: > On Wed, Mar 07, 2018 at 03:28:20PM +0100, Patrick Delaunay wrote: > >> avoid warning: no previous prototype for ‘mach_cpu_init’ >> >> Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> > > Reviewed-by: Tom Rini <trini@konsulko.com> > > -- People tend to put all sort of misc thingy into common.h but this is one of the ugliest parts in U-Boot. Most of files parse <common.h> that contains unrelated / unnecessary defines. Please realize this madness. I have tried to slim it down several times, but people have added more and more. So this task never ends. How about splitting out platform init hooks? (<init.h> or anything suitable is OK.) Also, I'd like to ban new additions to <common.h> like we do for config_whitelist.txt (for example, record the number of lines of <common.h> then build fails if somebody increases it.)
On Thu, Mar 08, 2018 at 09:57:58AM +0900, Masahiro Yamada wrote: > 2018-03-07 23:45 GMT+09:00 Tom Rini <trini@konsulko.com>: > > On Wed, Mar 07, 2018 at 03:28:20PM +0100, Patrick Delaunay wrote: > > > >> avoid warning: no previous prototype for ‘mach_cpu_init’ > >> > >> Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> > > > > Reviewed-by: Tom Rini <trini@konsulko.com> > > > > -- > > > People tend to put all sort of misc thingy into common.h > but this is one of the ugliest parts in U-Boot. > > Most of files parse <common.h> that contains > unrelated / unnecessary defines. > Please realize this madness. > > I have tried to slim it down several times, > but people have added more and more. So this task never ends. > > How about splitting out platform init hooks? > (<init.h> or anything suitable is OK.) OK, sure, lets move these and relevant bits over to init.h, or init_helpers.h which already exists and is in a few of the appropriate locations already. > Also, I'd like to ban new additions to <common.h> > like we do for config_whitelist.txt > (for example, record the number of lines of <common.h> > then build fails if somebody increases it.) Sure, thanks!
diff --git a/include/common.h b/include/common.h index 0fe9439..57ae790 100644 --- a/include/common.h +++ b/include/common.h @@ -471,9 +471,23 @@ int is_core_valid (unsigned int); * CPU- or SoC- specific init needed to continue the init sequence. See * board_f.c for where it is called. If this is not provided, a default * version (which does nothing) will be used. + * + * @return: 0 on success, otherwise error */ int arch_cpu_init(void); +/** + * mach_cpu_init() - SoC/machine dependent CPU setup + * + * This is called after arch_cpu_init(). It should handle any + * SoC or machine specific init needed to continue the init sequence. See + * board_f.c for where it is called. If this is not provided, a default + * version (which does nothing) will be used. + * + * @return: 0 on success, otherwise error + */ +int mach_cpu_init(void); + void s_init(void); int checkcpu (void);
avoid warning: no previous prototype for ‘mach_cpu_init’ Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> --- Changes in v2: - add return info for functions mach_cpu_init() and arch_cpu_init() include/common.h | 14 ++++++++++++++ 1 file changed, 14 insertions(+)