[PATCH libdrm 1/8] Add static qualifier to local functions

Emil Velikov emil.l.velikov at gmail.com
Sat Mar 14 06:12:17 PDT 2015


On 14/03/15 00:08, Jan Vesely wrote:
> On Fri, 2015-03-13 at 23:07 +0000, Emil Velikov wrote:
>> On 27 February 2015 at 18:07, Jan Vesely <jan.vesely at rutgers.edu> wrote:
>>> Signed-off-by: Jan Vesely <jan.vesely at rutgers.edu>
>>> ---
>>>  tests/drmstat.c             |  8 ++++----
>>>  tests/kmstest/main.c        |  2 +-
>>>  tests/modeprint/modeprint.c | 18 +++++++++---------
>>>  tests/proptest/proptest.c   |  2 +-
>>>  tests/radeon/radeon_ttm.c   |  4 ++--
>>>  xf86drm.c                   |  2 +-
>>>  xf86drmMode.c               |  2 +-
>>>  7 files changed, 19 insertions(+), 19 deletions(-)
>>>
>>> diff --git a/tests/drmstat.c b/tests/drmstat.c
>>> index 5935d07..36cc70d 100644
>>> --- a/tests/drmstat.c
>>> +++ b/tests/drmstat.c
>>> @@ -81,13 +81,13 @@ static void getversion(int fd)
>>>         printf( "No driver available\n" );
>>>      }
>>>  }
>>> -
>>> -void handler(int fd, void *oldctx, void *newctx)
>>> +
>>> +static void handler(int fd, void *oldctx, void *newctx)
>>>  {
>>>      printf("Got fd %d\n", fd);
>>>  }
>>>
>> It's only "user" was commented out as a transition to libdrm2 afaict.
>> Should be safe to nuke alongside the commented out caller.
> 
> This one got commented out in 8/8 if you prefer I can nuke it, and apply
> 8/8 before this one.
> 
>>
>>> -void process_sigio(char *device)
>>> +static void process_sigio(char *device)
>>>  {
>>>      int              fd;
>>>
>>> @@ -427,7 +427,7 @@ int main(int argc, char **argv)
>>>      return r;
>>>  }
>>>
>>> -void DRM_PRINTFLIKE(4, 0)
>>> +static void DRM_PRINTFLIKE(4, 0)
>>>  xf86VDrvMsgVerb(int scrnIndex, int type, int verb, const char *format,
>>>                  va_list args)
>> Think don't need to bother making this static and just nuke it. It
>> seems like it was added by mistake (commit c3092ead642) and never
>> used.
> 
> same here. it is removed in 8/8 I can reorder the patches tog et rid of
> this artifact
> 
That's exactly I was thinking. They seems toe be unused for 4+ years so
might as well kill them off in a single blow.

>>
>> ...
>>> diff --git a/xf86drm.c b/xf86drm.c
>>> index e117bc6..016247f 100644
>>> --- a/xf86drm.c
>>> +++ b/xf86drm.c
>>> @@ -131,7 +131,7 @@ drmMsg(const char *format, ...)
>>>      }
>>>  }
>>>
>>> -void
>>> +static void
>>>  drmSetDebugMsgFunction(debug_msg_func_t debug_msg_ptr)
>>>  {
>>>      drm_debug_print = debug_msg_ptr;
>>> diff --git a/xf86drmMode.c b/xf86drmMode.c
>>> index 9ea8fe7..1c06a19 100644
>>> --- a/xf86drmMode.c
>>> +++ b/xf86drmMode.c
>>> @@ -76,7 +76,7 @@ static inline int DRM_IOCTL(int fd, unsigned long cmd, void *arg)
>>>   * Util functions
>>>   */
>>>
>>> -void* drmAllocCpy(void *array, int count, int entry_size)
>>> +static void* drmAllocCpy(void *array, int count, int entry_size)
>> Strictly speaking these could still be used, despite never being part
>> of the API. Although my vote (fwiw) would be to that we're safe.
> 
> This one is heavily used in the same file.
> 
I might have been a bit vague there. I'm not saying a word against any
of it, just pointing out that there is a very small chance it may break
some unusual apps.


-Emil



More information about the dri-devel mailing list