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 1/2] Add framework for tunables


On Sun, Jul 3, 2016 at 10:13 AM, Siddhesh Poyarekar
<siddhesh@sourceware.org> wrote:
> On Sun, Jul 03, 2016 at 07:44:25AM -0700, H.J. Lu wrote:
>> > +
>> > +/* Compare environment names, bounded by the name hardcoded in glibc.  */
>> > +static bool
>> > +is_name (const char *orig, const char *envname)
>> > +{
>> > +  for (;*orig != '\0' && *envname != '\0'; envname++, orig++)
>> > +    if (*orig != *envname)
>> > +      break;
>> > +
>> > +  /* The ENVNAME is immediately followed by a value.  */
>> > +  if (*orig == '\0' && *envname == '=')
>> > +    return true;
>> > +  else
>> > +    return false;
>> > +}
>>
>> Can you add the length parameter to is_name and make is_name
>> processor specific so that misaligned short/int/long long load can
>> be used, similar to equal in
>
> Will it really make that much of an impact?  These will be really
> short comparisons (less than 16 bytes in most cases) and will really
> only happen once during startup.
>
> Maybe we could do it as an addon patch if the tunables initialization
> really turns out to be that much of a bottleneck.
>

Sure.

-- 
H.J.


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