Re: [PATCH v3] hash: add XOR32 hash function

[Bili Dong <qobilidop@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 1464 bytes --]

Got it. Will fix. Thanks!

On Mon, Feb 20, 2023 at 12:10 PM Medvedkin, Vladimir <
vladimir.medvedkin@intel.com> wrote:

> Hi Bill,
>
> On 15/02/2023 11:06, Bili Dong wrote:
> > An XOR32 hash is needed in the Software Switch (SWX) Pipeline for its
> > use case in P4. We implement it in this patch so it could be easily
> > registered in the pipeline later.
> >
> > Signed-off-by: Bili Dong <qobilidop@gmail.com>
> > ---
> > +static inline uint32_t
> > +rte_hash_xor(const void *data, uint32_t data_len, uint32_t init_val)
> > +{
> > +     uint32_t i;
> > +     uintptr_t pd = (uintptr_t) data;
> > +     init_val = rte_cpu_to_be_32(init_val);
> > +
> > +     for (i = 0; i < data_len / 4; i++) {
> > +             init_val ^= *(const uint32_t *)pd;
> > +             pd += 4;
> > +     }
> > +
> > +     if (data_len & 0x2) {
> > +             init_val ^= *(const uint32_t *)pd & LEFT16b_MASK;
>
> Here you are reading 2 bytes after the data buffer, which can sometimes
> lead to segfault. I think it would be better just to:
>
> init_val ^= *(const uint16_t *)pd << 16;
>
> The same with the section bellow
>
> > +             pd += 2;
> > +     }
> > +
> > +     if (data_len & 0x1)
> > +             init_val ^= *(const uint32_t *)pd & LEFT8b_MASK;
> > +
> > +     init_val = rte_be_to_cpu_32(init_val);
> > +     return init_val;
> > +}
> > +
> > +#ifdef __cplusplus
> > +}
> > +#endif
> > +
> > +#endif /* _RTE_HASH_XOR_H_ */
>
> --
> Regards,
> Vladimir
>
>

[-- Attachment #2: Type: text/html, Size: 2169 bytes --]