Re: [PATCH v6 3/3] selftests/lam: Update LAM tag width and cleanup names

[Maciej Wieczor-Retman]

On 2026-04-22 at 11:11:26 -0700, Sohil Mehta wrote:
>On 4/10/2026 2:55 AM, Maciej Wieczor-Retman wrote:
>> diff --git a/tools/testing/selftests/x86/lam.c b/tools/testing/selftests/x86/lam.c
>> index 1919fa6daec0..f14600e74d8c 100644
>> --- a/tools/testing/selftests/x86/lam.c
>> +++ b/tools/testing/selftests/x86/lam.c
>> @@ -24,11 +24,15 @@
>>  # error This test is 64-bit only
>>  #endif
>>
>> +#define GENMASK(h, l) (((~0UL) << (l)) & (~0UL >> (__BITS_PER_LONG - 1 - (h))))
>> +
>
>
>Did you run into any problem when including linux/bits.h? It would be
>preferable to avoid redefining generic macros. I see the GENMASK defines
>in tools/include/linux/bits.h and include/uapi/linux/bits.h.

Yeah, including it breaks something else in some other header that's included.

>If you are running into issues with the x86 selftests header includes,
>the uapi one should be safe to use.
>
>You probably just need:
>#define GENMASK(h, l)		__GENMASK(h, l)

Okay, that is a better way, thanks :)

>
>>  /* LAM modes, these definitions were copied from kernel code */
>>  #define LAM_NONE                0
>> -#define LAM_U57_BITS            6
>> +#define LAM_TAG_BITS		4
>> +#define LAM_LS_BIT		57
>> +#define LAM_MS_BIT		(LAM_LS_BIT + LAM_TAG_BITS - 1) /* 60 */
>> +#define LAM_UNTAG_MASK		~GENMASK(LAM_MS_BIT, LAM_LS_BIT)
>>
>
>Same comment as patch 1:
>#define LAM_UNTAG_MASK		~GENMASK(60, 57)

Sure.

>
>
>> -#define LAM_U57_MASK            (0x3fULL << 57)
>>  /* arch prctl for LAM */
>>  #define ARCH_GET_UNTAG_MASK     0x4001
>>  #define ARCH_ENABLE_TAGGED_ADDR 0x4002
>> @@ -51,8 +55,8 @@
>>  #define GET_USER_KERNEL         3
>
>The rest of the patch looks fine.
>
>With the above addressed,
>
>Reviewed-by: Sohil Mehta <sohil.mehta@xxxxxxxxx>

Thanks!

-- 
Kind regards
Maciej Wieczór-Retman