From: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
To: David Marchand <david.marchand@6wind.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>, Felix Marti <felix@chelsio.com>,
Nirranjan Kirubaharan <nirranjan@chelsio.com>,
Kumar Sanghvi <kumaras@chelsio.com>
Subject: Re: [dpdk-dev] [PATCH v2 1/3] nic_uio: Fix to allow any device to be bound to nic_uio
Date: Mon, 20 Jul 2015 17:37:27 +0530 [thread overview]
Message-ID: <20150720120725.GA3771@scalar.blr.asicdesigners.com> (raw)
In-Reply-To: <CALwxeUt5hFdpCjhizf1EPMV53+kXJ0VsM0p5yJ5W1udJf4oj5w@mail.gmail.com>
Hi David,
On Mon, Jul 20, 2015 at 09:43:57 +0200, David Marchand wrote:
> Hum, what bothers me is that you do not rely on the same criteria to
> re-attach the devices to nic_uio.
> See below.
>
> lib/librte_eal/bsdapp/nic_uio/nic_uio.c | 48
> +++++++++------------------------
> 1 file changed, 13 insertions(+), 35 deletions(-)
>
> diff --git a/lib/librte_eal/bsdapp/nic_uio/nic_uio.c
> b/lib/librte_eal/bsdapp/nic_uio/nic_uio.c
> index 2354e84..f868dc8 100644
> --- a/lib/librte_eal/bsdapp/nic_uio/nic_uio.c
> +++ b/lib/librte_eal/bsdapp/nic_uio/nic_uio.c
> [snip]
> @@ -195,11 +177,10 @@ nic_uio_probe (device_t dev)
> {
> int i;
>
> - for (i = 0; i < NUM_DEVICES; i++)
> - if (pci_get_vendor(dev) == devices[i].vend &&
> - pci_get_device(dev) == devices[i].dev) {
> -
> - device_set_desc(dev, "Intel(R) DPDK PCI
> Device");
> + for (i = 0; i < num_detached; i++)
> + if (pci_get_vendor(dev) ==
> pci_get_vendor(detached_devices[i]) &&
> + pci_get_device(dev) ==
> pci_get_device(detached_devices[i])) {
> + device_set_desc(dev, "DPDK PCI Device");
> return BUS_PROBE_SPECIFIC;
> }
>
> When going through the probe stuff, the device vendor and type are used as
> the matching criteria.
>
> @@ -256,7 +237,6 @@ static void
> nic_uio_load(void)
> {
> uint32_t bus, device, function;
> - int i;
> device_t dev;
> char bdf_str[256];
> char *token, *remaining;
> @@ -295,17 +275,15 @@ nic_uio_load(void)
> if (dev == NULL)
> continue;
>
> - for (i = 0; i < NUM_DEVICES; i++)
> - if (pci_get_vendor(dev) == devices[i].vend &&
> - pci_get_device(dev) ==
> devices[i].dev) {
> - if (num_detached <
> MAX_DETACHED_DEVICES) {
> -
> printf("nic_uio_load: detaching and storing dev=%p\n", dev);
> -
> detached_devices[num_detached++] = dev;
> - } else
> -
> printf("nic_uio_load: reached MAX_DETACHED_DEVICES=%d. dev=%p won't be
> reattached\n",
> -
> MAX_DETACHED_DEVICES, dev);
> - device_detach(dev);
> - }
> + if (num_detached < MAX_DETACHED_DEVICES) {
> + printf("nic_uio_load: detaching and storing
> dev=%p\n",
> + dev);
> + detached_devices[num_detached++] = dev;
> + } else {
> + printf("nic_uio_load: reached
> MAX_DETACHED_DEVICES=%d. dev=%p won't be reattached\n",
> + MAX_DETACHED_DEVICES, dev);
> + }
> + device_detach(dev);
> }
> }
>
> But here at init time, the bdfs informations are used to detach the pci
> devices.
> I would say this is safer we have the same criteria in both cases.
> I think that the pci addresses are the best criteria since this is what
> the user gives.
> Don't we have them in the dev pointer ?
It looks like we can get them via pci_get_bus(), pci_get_slot(), and
pci_get_function(). Will add check for these 3 info instead of vendor
and device in probe to make it consistent.
>
> Btw, with this change, we would then be limited to MAX_DETACHED_DEVICES
> devices even if 128 pci devices looks quite big enough to me.
> This part could be reworked (later).
> --
> David Marchand
Thanks,
Rahul
next prev parent reply other threads:[~2015-07-20 12:08 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-07-17 18:20 [dpdk-dev] [PATCH 0/3] cxgbe: Fix compilation and enable FreeBSD support for CXGBE PMD Rahul Lakkireddy
2015-07-17 18:20 ` [dpdk-dev] [PATCH 1/3] nic_uio: Fix to allow any device to be bound to nic_uio Rahul Lakkireddy
2015-07-19 23:23 ` Thomas Monjalon
2015-07-17 18:20 ` [dpdk-dev] [PATCH 2/3] cxgbe: Enable and fix FreeBSD compilation for CXGBE PMD Rahul Lakkireddy
2015-07-17 18:20 ` [dpdk-dev] [PATCH 3/3] doc: Update documentation to reflect FreeBSD support " Rahul Lakkireddy
2015-07-20 6:41 ` [dpdk-dev] [PATCH v2 0/3] cxgbe: Fix compilation and enable " Rahul Lakkireddy
2015-07-20 6:41 ` [dpdk-dev] [PATCH v2 1/3] nic_uio: Fix to allow any device to be bound to nic_uio Rahul Lakkireddy
2015-07-20 7:43 ` David Marchand
2015-07-20 12:07 ` Rahul Lakkireddy [this message]
2015-07-20 6:41 ` [dpdk-dev] [PATCH v2 2/3] cxgbe: Enable and fix FreeBSD compilation for CXGBE PMD Rahul Lakkireddy
2015-07-20 6:41 ` [dpdk-dev] [PATCH v2 3/3] doc: Update documentation to reflect FreeBSD support " Rahul Lakkireddy
2015-07-20 17:31 ` [dpdk-dev] [PATCH v3 0/3] cxgbe: Fix compilation and enable " Rahul Lakkireddy
2015-07-20 17:31 ` [dpdk-dev] [PATCH v3 1/3] nic_uio: Fix to allow any device to be bound to nic_uio Rahul Lakkireddy
2015-07-21 6:14 ` David Marchand
2015-07-20 17:31 ` [dpdk-dev] [PATCH v3 2/3] cxgbe: Enable and fix FreeBSD compilation for CXGBE PMD Rahul Lakkireddy
2015-07-20 17:31 ` [dpdk-dev] [PATCH v3 3/3] doc: Update documentation to reflect FreeBSD support " Rahul Lakkireddy
2015-07-21 9:40 ` [dpdk-dev] [PATCH v3 0/3] cxgbe: Fix compilation and enable " Bruce Richardson
2015-07-22 13:11 ` Thomas Monjalon
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=20150720120725.GA3771@scalar.blr.asicdesigners.com \
--to=rahul.lakkireddy@chelsio.com \
--cc=david.marchand@6wind.com \
--cc=dev@dpdk.org \
--cc=felix@chelsio.com \
--cc=kumaras@chelsio.com \
--cc=nirranjan@chelsio.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).