On Wed, Feb 12, 2020 at 6:14 PM Simon Glass sjg@chromium.org wrote:
Hi Masahiro,
On Wed, 12 Feb 2020 at 06:14, Masahiro Yamada masahiroy@kernel.org wrote:
On Mon, Jan 13, 2020 at 4:08 AM Simon Glass sjg@chromium.org wrote:
This function name conflicts with our desire to #define free() to something else on sandbox. Since it deals with resources, rename it to rfree().
Signed-off-by: Simon Glass sjg@chromium.org
I noticed this commit was merged recently.
Now 'free' is a reserved keyword you cannot use in U-Boot.
Commit cc92c3c thru cf23c7c are horrible.
Commit cfda60f should have used 'static inline' instead of #define.
I cannot believe it.
Are you sure you understand the problem I was trying to solve? I am using dlmalloc's existing means of adding a prefix, but I'm sure we could change it to another way.
If we define malloc() as dlmalloc() in dlmalloc.c, then we could add a declaration in dlmalloc.h that uses static inline to convert calls to malloc() to call dlmalloc(). Then anything that doesn't include malloc.h would still call the C library malloc(). Is that what you are thinking?
There is no "malloc()" in dlmalloc.c. It is called "mALLOc()" and by defining USE_DL_PREFIX, you already have converted that to be linked as "dlmalloc()".
I think there should be no difference in who calls what when converting your defines to static inline functions.
And yes, I also dislike the other patches that remove all occurrences of 'free'. I think without knowing the backgrounds of your patches, that just looks strange.
Regards, Simon
I did look at using a link script instead but it is pretty messy.
What do you mean by 'free' being a reserved keyword? Where? So is 'rfree' a good substitute or do you suggest something else?
Regards, Simon