On 06/12/2023 8:46 am, Jan Beulich wrote:
> On 06.12.2023 08:10, Juergen Gross wrote:
>> Instead of defining get_unaligned() and put_unaligned() in a way that
>> is only supporting architectures allowing unaligned accesses, use the
>> same approach as the Linux kernel and let the compiler do the
>> decision how to generate the code for probably unaligned data accesses.
>>
>> Update include/xen/unaligned.h from include/asm-generic/unaligned.h of
>> the Linux kernel.
>>
>> The generated code has been checked to be the same on x86.
>>
>> Modify the Linux variant to not use underscore prefixed identifiers,
>> avoid unneeded parentheses and drop the 24-bit accessors.
>>
>> Signed-off-by: Arnd Bergmann <a...@arndb.de>
>> Origin: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
>> 803f4e1eab7a
>> Signed-off-by: Juergen Gross <jgr...@suse.com>
> Acked-by: Jan Beulich <jbeul...@suse.com>
>
> Nevertheless ...
>
>> @@ -15,67 +7,82 @@
>> #include <asm/byteorder.h>
>> #endif
>>
>> -#define get_unaligned(p) (*(p))
>> -#define put_unaligned(val, p) (*(p) = (val))
>> +/*
>> + * This is the most generic implementation of unaligned accesses
>> + * and should work almost anywhere.
>> + */
>> +
>> +#define get_unaligned_t_(type, ptr) ({
>> \
> ..., do we need the trailing underscores here in addition to the already
> sufficiently clear _t suffixes? (Leaving aside that ..._t generally is to
> denote types, not macros or functions.)
_t is fine. It's what we use for {min,max}_t() and friends too.
~Andrew
