Re: [PATCH] jhash: Deinline jhash, jhash2 and __jhash_nwords

[Daniel Borkmann <daniel@iogearbox.net>]

On 07/16/2015 02:15 PM, Denys Vlasenko wrote:
> On 07/16/2015 12:41 PM, Thomas Graf wrote:
>> On 07/16/15 at 12:02pm, Denys Vlasenko wrote:
>>> +/* jhash - hash an arbitrary key
>>> + * @k: sequence of bytes as key
>>> + * @length: the length of the key
>>> + * @initval: the previous hash, or an arbitray value
>>> + *
>>> + * The generic version, hashes an arbitrary sequence of bytes.
>>> + * No alignment or length assumptions are made about the input key.
>>> + *
>>> + * Returns the hash value of the key. The result depends on endianness.
>>> + */
>>> +u32 jhash(const void *key, u32 length, u32 initval)
>>
>> Shouldn't these live in lib/jhash.c or something? Otherwise
>> everyone needs to depend on CONFIG_RHASHTABLE
>
> There is no CONFIG_RHASHTABLE, rhashtable.c is compiled unconditionally.
>
> I will send an alternative patch, which creates jhash.c;
> apply whichever version you like most.

Please also Cc netdev as networking subsys is one of the main users
of jhash in various critical paths. Did you run e.g. some *_RR work
load benchmarks as well to make sure there's no regression?