On 2022/5/3 9:06, Tong Tiangen wrote:
在 2022/5/2 22:24, Christophe Leroy 写道:
Le 20/04/2022 à 05:04, Tong Tiangen a écrit :
x86/powerpc has it's implementation of copy_mc_to_user but not use
#define
to declare.
This may cause problems, for example, if other architectures open
CONFIG_ARCH_HAS_COPY_MC, but want to use copy_mc_to_user() outside the
architecture, the code add to include/linux/uaddess.h is as follows:
#ifndef copy_mc_to_user
static inline unsigned long __must_check
copy_mc_to_user(void *dst, const void *src, size_t cnt)
{
...
}
#endif
Then this definition will conflict with the implementation of
x86/powerpc
and cause compilation errors as follow:
Fixes: ec6347bb4339 ("x86, powerpc: Rename memcpy_mcsafe() to
copy_mc_to_{user, kernel}()")
I don't understand, what does it fix really ? What was the
(existing/real) bug introduced by that patch and that your are fixing ?
If those defined had been expected and missing, we would have had a
build failure. If you have one, can you describe it ?
It could prevent future problems when patch3 is introduced, and yes,for
now,
this patch won't fix any issue,we could drop the fix tag, and update the
changelog.
There will be build failure after patch 3 is added, there is a little
confusing for a reader of this commit in isolation.
In the next version, I will put this patch after patch 3.
This is an alternative.
Thanks,
Tong.
.