DPDK patches and discussions
 help / color / mirror / Atom feed
From: Jan Blunck <jblunck@infradead.org>
To: Shreyansh Jain <shreyansh.jain@nxp.com>
Cc: dev@dpdk.org, David Marchand <david.marchand@6wind.com>
Subject: Re: [dpdk-dev] [PATCH v2 8/8] ethdev: Decouple interrupt handling from PCI device
Date: Tue, 20 Dec 2016 11:51:21 +0100	[thread overview]
Message-ID: <CALe+Z01Nv12g41v4Fpe=ddSNxru39Y=VXpjF+D9WRbpDF2AJLg@mail.gmail.com> (raw)
In-Reply-To: <84460a6f-3574-8da2-3ccc-6f1e23fb50e7@nxp.com>

On Tue, Nov 22, 2016 at 1:57 PM, Shreyansh Jain <shreyansh.jain@nxp.com> wrote:
> On Monday 21 November 2016 10:25 PM, Jan Blunck wrote:
>>
>> The struct rte_intr_handle is an abstraction layer for different types of
>> interrupt mechanisms. It is embedded in the low-level device (e.g. PCI).
>> On allocation of a struct rte_eth_dev a reference to the intr_handle
>> should be stored for devices supporting interrupts.
>>
>> Signed-off-by: Jan Blunck <jblunck@infradead.org>
>> ---
>>  lib/librte_ether/rte_ethdev.c | 18 ++++++++++++++++--
>>  lib/librte_ether/rte_ethdev.h |  1 +
>>  2 files changed, 17 insertions(+), 2 deletions(-)
>>
>> diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
>> index 4288577..4ecea50 100644
>> --- a/lib/librte_ether/rte_ethdev.c
>> +++ b/lib/librte_ether/rte_ethdev.c
>> @@ -258,6 +258,7 @@ rte_eth_dev_pci_probe(struct rte_pci_driver *pci_drv,
>>                         rte_panic("Cannot allocate memzone for private
>> port data\n");
>>         }
>>         eth_dev->pci_dev = pci_dev;
>> +       eth_dev->intr_handle = &pci_dev->intr_handle;
>>         eth_dev->driver = eth_drv;
>>         eth_dev->data->rx_mbuf_alloc_failed = 0;
>>
>> @@ -2543,7 +2544,13 @@ rte_eth_dev_rx_intr_ctl(uint8_t port_id, int epfd,
>> int op, void *data)
>>         RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
>>
>>         dev = &rte_eth_devices[port_id];
>> -       intr_handle = &dev->pci_dev->intr_handle;
>> +
>> +       if (!dev->intr_handle) {
>> +               RTE_PMD_DEBUG_TRACE("RX Intr handle unset\n");
>> +               return -ENOTSUP;
>> +       }
>> +
>> +       intr_handle = dev->intr_handle;
>>         if (!intr_handle->intr_vec) {
>>                 RTE_PMD_DEBUG_TRACE("RX Intr vector unset\n");
>>                 return -EPERM;
>> @@ -2603,7 +2610,12 @@ rte_eth_dev_rx_intr_ctl_q(uint8_t port_id, uint16_t
>> queue_id,
>>                 return -EINVAL;
>>         }
>>
>> -       intr_handle = &dev->pci_dev->intr_handle;
>> +       if (!dev->intr_handle) {
>> +               RTE_PMD_DEBUG_TRACE("RX Intr handle unset\n");
>> +               return -ENOTSUP;
>> +       }
>> +
>> +       intr_handle = dev->intr_handle;
>>         if (!intr_handle->intr_vec) {
>>                 RTE_PMD_DEBUG_TRACE("RX Intr vector unset\n");
>>                 return -EPERM;
>> @@ -3205,6 +3217,8 @@ rte_eth_copy_pci_info(struct rte_eth_dev *eth_dev,
>> struct rte_pci_device *pci_de
>>                 return;
>>         }
>>
>> +       eth_dev->intr_handle = &pci_dev->intr_handle;
>> +
>>         eth_dev->data->dev_flags = 0;
>>         if (pci_dev->driver->drv_flags & RTE_PCI_DRV_INTR_LSC)
>>                 eth_dev->data->dev_flags |= RTE_ETH_DEV_INTR_LSC;
>> diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
>> index 3adbb2b..f1f656a 100644
>> --- a/lib/librte_ether/rte_ethdev.h
>> +++ b/lib/librte_ether/rte_ethdev.h
>> @@ -1629,6 +1629,7 @@ struct rte_eth_dev {
>>         const struct eth_driver *driver;/**< Driver for this device */
>>         const struct eth_dev_ops *dev_ops; /**< Functions exported by PMD
>> */
>>         struct rte_pci_device *pci_dev; /**< PCI info. supplied by probing
>> */
>> +       struct rte_intr_handle *intr_handle; /**< Device interrupt handle
>> */
>>         /** User application callbacks for NIC interrupts */
>>         struct rte_eth_dev_cb_list link_intr_cbs;
>>         /**
>>
>
> Is there another patch which replaces all uses of
> eth_dev->pci_dev->intr_handle with that of eth_dev->intr_handle?
>
> Now that eth_dev has a reference which is initialized as early as probe, we
> should use that rather than pci_dev->intr_handle for PCI PMDs.
>

I've added this indirection because it is required for the ethdev
function. The drivers shouldn't use the indirection through ethdev to
access the intr_handle because they do have direct access to the
device.


> OR maybe, ETH_DEV_INTR_HANDLE() like macro which you have introduced in
> i40e_ethdev.h.
>
> -
> Shreyansh

  reply	other threads:[~2016-12-20 10:51 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-11-21 16:55 [dpdk-dev] [PATCH v2 1/8] eal: define container_of macro Jan Blunck
2016-11-21 16:55 ` [dpdk-dev] [PATCH v2 2/8] ethdev: Helper to convert to struct rte_pci_device Jan Blunck
2016-11-22 12:46   ` Shreyansh Jain
2016-11-21 16:55 ` [dpdk-dev] [PATCH v2 3/8] drivers: Use ETH_DEV_PCI_DEV() helper Jan Blunck
2016-11-22 11:21   ` Shreyansh Jain
2016-11-21 16:55 ` [dpdk-dev] [PATCH v2 4/8] virtio: Don't fill dev_info->driver_name Jan Blunck
2016-11-21 16:55 ` [dpdk-dev] [PATCH v2 5/8] virtio: Add vtpci_intr_handle() helper to get rte_intr_handle Jan Blunck
2016-11-21 16:55 ` [dpdk-dev] [PATCH v2 6/8] virtio: Don't depend on struct rte_eth_dev's pci_dev Jan Blunck
2016-11-21 16:55 ` [dpdk-dev] [PATCH v2 7/8] ethdev: Move filling of rte_eth_dev_info->pci_dev to dev_infos_get() Jan Blunck
2016-11-21 16:55 ` [dpdk-dev] [PATCH v2 8/8] ethdev: Decouple interrupt handling from PCI device Jan Blunck
2016-11-22 12:57   ` Shreyansh Jain
2016-12-20 10:51     ` Jan Blunck [this message]
2016-12-20 12:45       ` Shreyansh Jain
2016-11-22 10:03 ` [dpdk-dev] [PATCH v2 1/8] eal: define container_of macro Shreyansh Jain
2016-11-22 10:19   ` Thomas Monjalon
2016-11-22 11:26     ` Shreyansh Jain
2016-11-22 12:33       ` Jan Viktorin
2016-11-22 12:36         ` Shreyansh Jain
2016-12-08  7:01 ` Shreyansh Jain
2016-12-09  7:36   ` Shreyansh Jain

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CALe+Z01Nv12g41v4Fpe=ddSNxru39Y=VXpjF+D9WRbpDF2AJLg@mail.gmail.com' \
    --to=jblunck@infradead.org \
    --cc=david.marchand@6wind.com \
    --cc=dev@dpdk.org \
    --cc=shreyansh.jain@nxp.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).