Messages in this thread | | | From | Nick Desaulniers <> | Date | Thu, 7 Jun 2018 11:51:59 -0700 | Subject | Re: [PATCH v2 1/2] compiler-gcc.h: add gnu_inline to all inline declarations |
| |
On Thu, Jun 7, 2018 at 11:31 AM Joe Perches <joe@perches.com> wrote: > > On Thu, 2018-06-07 at 10:26 -0700, Nick Desaulniers wrote: > > I get the feeling that the use of __inline__ or __inline (vs inline) > > in the kernel may be wrong and their use should be eradicated in the > > follow up patch set, but it would be cool if others have additional > > insight. > > __inline is easy and useful to remove as it's used in > just a few files. > > But __inline__ is used in a lot of files and locations: > > $ git grep -w --name-only __inline__ | wc -l > 154 > $ git grep -w __inline__ | wc -l > 503 > > Some of these files are used in asm includes as > well where __inline__ may be preferred by gcc > > https://gcc.gnu.org/onlinedocs/gcc/Alternate-Keywords.html#Alternate-Keywords > > so perhaps asm exclusions would be necessary.
Oops, just saw this email after sending v3: https://lkml.org/lkml/2018/6/7/1055
Guess I should re-add __inline__ and __inline then to 1/3? https://lkml.org/lkml/2018/6/7/1060
ie:
+#define __inline__ __inline__ inline instead of +#define __inline__ inline
?
On the other hand, I was able to assemble with gcc+binutils and v3 as it stands.
Reading the link you sent, it seems more about compiling with -pedantic or non-gnu C standards being used (I'd call that unlikely for the kernel). So it still seems incorrect to me the use of __inline__ (but I do appreciate the link!) at least in the kernel.
| |