Re: [PATCH v2 1/5] crypto/dpaa2_sec: remove crypto device in bus device

[David Marchand <david.marchand@redhat.com>]

Hello,

On Mon, 17 Nov 2025 at 09:36, Maxime Leroy <maxime@leroys.fr> wrote:
> > @@ -4526,10 +4524,14 @@ cryptodev_dpaa2_sec_probe(struct rte_dpaa2_driver *dpaa2_drv __rte_unused,
> >  static int
> >  cryptodev_dpaa2_sec_remove(struct rte_dpaa2_device *dpaa2_dev)
> >  {
> > +       char cryptodev_name[RTE_CRYPTODEV_NAME_MAX_LEN];
> >         struct rte_cryptodev *cryptodev;
> >         int ret;
> >
> > -       cryptodev = dpaa2_dev->cryptodev;
> > +       snprintf(cryptodev_name, sizeof(cryptodev_name), "dpsec-%d",
> > +                       dpaa2_dev->object_id);
> > +
> > +       cryptodev = rte_cryptodev_pmd_get_named_dev(cryptodev_name);
> >         if (cryptodev == NULL)
> >                 return -ENODEV;
> >
> > --
> > 2.51.0
> >
> Should we add a function to generate the cryptodev name instead of
> duplicating the code, similar to how it’s done for PCI devices (i.e.,
> rte_pci_device_name)?

I am not sure I get the parallel..
rte_pci_device_name is a bus level name, I don't see how we could
define a fslmc bus level helper that would help here.
Ah, you mean adding a helper internal to this cryptodev driver?

I can respin this today, but after this, I'll be offline for the rest
of the week (ovscon 2025).

Btw, this series is a cleanup (those drivers have been like this for a
long time), and it can be deferred to next release if we are too short
on time.
On the other hand, the fixes from
https://patchwork.dpdk.org/project/dpdk/list/?series=36702 seem valid
for 25.11.0 (or at least backport material for 25.11.1).


-- 
David Marchand