This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH 6/8] tunables: Use glibc.tune.hwcap_mask tunable instead of _dl_hwcap_mask


On Thursday 11 May 2017 07:50 PM, Siddhesh Poyarekar wrote:
>>  static void
>>  get_common_indeces (struct cpu_features *cpu_features,
>> @@ -316,7 +317,11 @@ no_cpuid:
>>    /* Reuse dl_platform, dl_hwcap and dl_hwcap_mask for x86.  */
>>    GLRO(dl_platform) = NULL;
>>    GLRO(dl_hwcap) = 0;
>> +#if HAVE_TUNABLES
>> +  TUNABLE_UPDATE_VAL (glibc, tune, hwcap_mask, &(uint64_t){HWCAP_IMPORTANT});
>> +#else
>>    GLRO(dl_hwcap_mask) = HWCAP_IMPORTANT;
>> +#endif
>>  
>>  # ifdef __x86_64__
>>    if (cpu_features->kind == arch_kind_intel)
>>
>>
>> But even though build succeeds, the elf/tst-env-setuid fails on x86_64 (still
>> investigating why).

OK, I figured out what happened - I did not take into account HJ's
recent patch that introduced hwcap_mask into cpu-features.c, breaking
all of this.

The correct fix here is to remove the dl_hwcap_mask initialization for
tunables because tunables already initializes it correctly.  I'll fix up
and repost the entire series minus the first patch which is relatively
simple anyway.

Siddhesh


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]