Re: [dpdk-dev] [PATCH 1/9] cryptodev: add feature flag for non-byte aligned data

["Trahe, Fiona" <fiona.trahe@intel.com>]

Hi Akhil,

> -----Original Message-----
> From: Akhil Goyal <akhil.goyal@nxp.com>
> Sent: Tuesday, May 5, 2020 11:13 AM
> To: Trahe, Fiona <fiona.trahe@intel.com>; dev@dpdk.org
> Cc: Ruifeng.Wang@arm.com; Doherty, Declan <declan.doherty@intel.com>; asomalap@amd.com;
> anoobj@marvell.com; Zhang, Roy Fan <roy.fan.zhang@intel.com>; tdu@semihalf.com;
> rnagadheeraj@marvell.com; adwivedi@marvell.com; Gagandeep Singh <G.Singh@nxp.com>; Hemant
> Agrawal <hemant.agrawal@nxp.com>; jianjay.zhou@huawei.com; De Lara Guarch, Pablo
> <pablo.de.lara.guarch@intel.com>
> Subject: RE: [PATCH 1/9] cryptodev: add feature flag for non-byte aligned data
> 
> Hi Fiona,
> >
> > Hi Akhil,
> >
> > >
> > > Some wireless algos like SNOW, ZUC may support input
> > > data in bits which are not byte aligned. However, not
> > > all PMDs can support this requirement. Hence added a
> > > new feature flag RTE_CRYPTODEV_FF_NON_BYTE_ALIGNED_DATA
> > > to identify which all PMDs can support non-byte aligned
> > > data.
> > [Fiona] I know it's a bit pedantic, but shouldn't this really be called
> > NON_BYTE_MULTIPLE?
> >
> > As a non-byte-multiple data set could have its start byte-aligned, but still not be
> > supported.
> >
> The intention is to have both start as well as end to be byte aligned.
> Is there some PMD which can support unaligned byte data but still multiples of byte?
[Fiona] no, true - anything which can support non-byte-aligned can presumably support non-byte-multiples.
So I withdraw my suggestion. In that case:
Acked-by: Fiona Trahe <fiona.trahe@intel.com>
 
> BTW I did not see any example in the current test cases which support unaligned bytes data.
> 
> Regards,
> Akhil