[dpdk-dev] [PATCH v4 02/41] bus/dpaa: introduce NXP DPAA Bus driver skeleton
Shreyansh Jain
shreyansh.jain at nxp.com
Tue Sep 19 15:14:27 CEST 2017
Hello Ferruh,
On Monday 18 September 2017 08:17 PM, Ferruh Yigit wrote:
> On 9/9/2017 12:20 PM, Shreyansh Jain wrote:
>> Signed-off-by: Shreyansh Jain <shreyansh.jain at nxp.com>
>> Signed-off-by: Hemant Agrawal <hemant.agrawal at nxp.com>
>
> <...>
>
>> diff --git a/drivers/bus/dpaa/rte_bus_dpaa_version.map b/drivers/bus/dpaa/rte_bus_dpaa_version.map
>> new file mode 100644
>> index 0000000..d97a009
>> --- /dev/null
>> +++ b/drivers/bus/dpaa/rte_bus_dpaa_version.map
>> @@ -0,0 +1,7 @@
>> +DPDK_17.11 {
>> + global:
>> +
>> + rte_dpaa_driver_register;
>> + rte_dpaa_driver_unregister;
>
> "local *;" ?
Agree. I will change this.
Currently rte_dpaa_driver_* functions are being used locally within
bus/dpaa.
>
> <...>
>
>> +struct rte_dpaa_device {
>> + TAILQ_ENTRY(rte_dpaa_device) next;
>> + struct rte_device device;
>> + union {
>> + struct rte_eth_dev *eth_dev;
>> + struct rte_cryptodev *crypto_dev;
>> + };
>
> Bus struct should be independt from functionality, this has been done in
> PCI, can same thing be done for dpaa bus too?
Sorry, I didn't get your point. This is the rte_dpaa_bus structure:
struct rte_dpaa_bus {
struct rte_bus bus;
struct rte_dpaa_device_list device_list;
struct rte_dpaa_driver_list driver_list;
int device_count;
};
If you are referring to unlinking eth/crypto functionality from
rte_dpaa_device - that is something which needs investigation. I have
seen patches on PCI from Gaetan. Can that be an incremental change over
this?
>
> <...>
>
>> + * @return
>> + * 0 in case of success, error otherwise
>> + */
>> +int rte_dpaa_portal_init(void *arg);
>
> Definition is not in this patch.
>
>> +
>> +/**
>> + * Cleanup a DPAA Portal
>> + */
>> +void dpaa_portal_finish(void *arg);
>
> Definition is not in this patch.
>
> <...>
>
Yes, this is my mistake. I will fix this.
More information about the dev
mailing list