From: Igor Ryzhov <iryzhov@nfware.com>
To: Ferruh Yigit <ferruh.yigit@intel.com>
Cc: dev <dev@dpdk.org>, Dan Gora <dg@adax.com>
Subject: Re: [dpdk-dev] [PATCH v4 1/2] kni: rework rte_kni_update_link using ioctl
Date: Tue, 6 Jul 2021 12:14:31 +0300 [thread overview]
Message-ID: <CAF+s_Fz3Jr9Yy+3+mmGzq5PoZvHCR=wSSWMBS1d8DoyOaPZ6eA@mail.gmail.com> (raw)
In-Reply-To: <415de2ea-8ea0-18c3-af10-280a9f4fe8da@intel.com>
Hi Ferruh,
Thanks for the review. My comments inline. I'll send a new version later
this week.
On Mon, Jul 5, 2021 at 2:58 PM Ferruh Yigit <ferruh.yigit@intel.com> wrote:
> On 7/4/2021 6:06 PM, Igor Ryzhov wrote:
> > Current implementation doesn't allow us to update KNI carrier if the
> > interface is not yet UP in kernel. It means that we can't use it in the
> > same thread which is processing rte_kni_ops.config_network_if, which is
> > very convenient, because it allows us to have correct carrier status
> > of the interface right after we enabled it and we don't have to use any
> > additional thread to track link status.
> >
> > Propagating speed/duplex/autoneg to the kernel module also allows us to
> > implement ethtool_ops.get_link_ksettings callback.
> >
> > Suggested-by: Dan Gora <dg@adax.com>
> > Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
> > ---
> > app/test/test_kni.c | 32 ++++++++++---------------
> > examples/kni/main.c | 2 +-
> > kernel/linux/kni/compat.h | 4 ----
> > kernel/linux/kni/kni_dev.h | 5 ++++
> > kernel/linux/kni/kni_misc.c | 47 +++++++++++++++++++++++++++++++++++++
> > kernel/linux/kni/kni_net.c | 15 ------------
> > lib/kni/rte_kni.c | 38 ++++++++----------------------
> > lib/kni/rte_kni.h | 10 ++++----
> > lib/kni/rte_kni_common.h | 9 +++++++
> > 9 files changed, 89 insertions(+), 73 deletions(-)
> >
> > diff --git a/app/test/test_kni.c b/app/test/test_kni.c
> > index 96733554b6c4..f9300552b3f6 100644
> > --- a/app/test/test_kni.c
> > +++ b/app/test/test_kni.c
> > @@ -125,6 +125,7 @@ kni_change_mtu(uint16_t port_id, unsigned int
> new_mtu)
> > static int
> > test_kni_link_change(void)
> > {
> > + struct rte_eth_link link;
> > int ret;
> > int pid;
> >
> > @@ -135,42 +136,35 @@ test_kni_link_change(void)
> > }
> >
> > if (pid == 0) {
> > + link.link_speed = ETH_SPEED_NUM_10G;
> > + link.link_duplex = ETH_LINK_FULL_DUPLEX;
> > + link.link_autoneg = ETH_LINK_AUTONEG;
> > +
> > printf("Starting KNI Link status change tests.\n");
> > if (system(IFCONFIG TEST_KNI_PORT" up") == -1) {
> > ret = -1;
> > goto error;
> > }
> >
> > - ret = rte_kni_update_link(test_kni_ctx, 1);
> > + link.link_status = ETH_LINK_UP;
> > + ret = rte_kni_update_link(test_kni_ctx, &link);
> > if (ret < 0) {
> > printf("Failed to change link state to Up
> ret=%d.\n",
> > ret);
> > goto error;
> > }
> > rte_delay_ms(1000);
> > - printf("KNI: Set LINKUP, previous state=%d\n", ret);
> > -
> > - ret = rte_kni_update_link(test_kni_ctx, 0);
> > - if (ret != 1) {
> > - printf(
> > - "Failed! Previous link state should be 1, returned %d.\n",
> > - ret);
> > - goto error;
> > - }
> > - rte_delay_ms(1000);
> > - printf("KNI: Set LINKDOWN, previous state=%d\n", ret);
> > + printf("KNI: Set LINKUP\n");
> >
> > - ret = rte_kni_update_link(test_kni_ctx, 1);
> > - if (ret != 0) {
> > - printf(
> > - "Failed! Previous link state should be 0, returned %d.\n",
> > + link.link_status = ETH_LINK_DOWN;
> > + ret = rte_kni_update_link(test_kni_ctx, &link);
> > + if (ret < 0) {
> > + printf("Failed to change link state to Down
> ret=%d.\n",
> > ret);
> > goto error;
> > }
>
> Is there a way to verify the link status of the KNI interface, to double
> check
> setting link status succedded?
>
Yes, we can still read from the carrier file to check the current status.
I'll add this to the test.
>
> > - printf("KNI: Set LINKUP, previous state=%d\n", ret);
> > -
> > - ret = 0;
> > rte_delay_ms(1000);
> > + printf("KNI: Set LINKDOWN\n");
> >
> > error:
> > if (system(IFCONFIG TEST_KNI_PORT" down") == -1)
> > diff --git a/examples/kni/main.c b/examples/kni/main.c
> > index beabb3c848aa..5833037cf1c9 100644
> > --- a/examples/kni/main.c
> > +++ b/examples/kni/main.c
> > @@ -755,7 +755,7 @@ monitor_all_ports_link_status(void *arg)
> > }
> > for (i = 0; i < p[portid]->nb_kni; i++) {
> > prev =
> rte_kni_update_link(p[portid]->kni[i],
> > - link.link_status);
> > + &link);
> > log_link_state(p[portid]->kni[i], prev,
> &link);
> > }
> > }
> > diff --git a/kernel/linux/kni/compat.h b/kernel/linux/kni/compat.h
> > index 5f65640d5ed2..995d109c275a 100644
> > --- a/kernel/linux/kni/compat.h
> > +++ b/kernel/linux/kni/compat.h
> > @@ -61,10 +61,6 @@
> > #define kni_sock_map_fd(s) sock_map_fd(s, 0)
> > #endif
> >
> > -#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 9, 0)
> > -#define HAVE_CHANGE_CARRIER_CB
> > -#endif
> > -
> > #if LINUX_VERSION_CODE < KERNEL_VERSION(3, 14, 0)
> > #define ether_addr_copy(dst, src) memcpy(dst, src, ETH_ALEN)
> > #endif
> > diff --git a/kernel/linux/kni/kni_dev.h b/kernel/linux/kni/kni_dev.h
> > index c15da311ba25..969108cc30f8 100644
> > --- a/kernel/linux/kni/kni_dev.h
> > +++ b/kernel/linux/kni/kni_dev.h
> > @@ -88,6 +88,11 @@ struct kni_dev {
> > void *alloc_va[MBUF_BURST_SZ];
> >
> > struct task_struct *usr_tsk;
> > +
> > + /* correct when netif_carrier_ok */
> > + uint32_t speed;
> > + uint8_t duplex;
> > + uint8_t autoneg;
> > };
> >
> > #ifdef HAVE_IOVA_TO_KVA_MAPPING_SUPPORT
> > diff --git a/kernel/linux/kni/kni_misc.c b/kernel/linux/kni/kni_misc.c
> > index 2b464c438113..30ee69661935 100644
> > --- a/kernel/linux/kni/kni_misc.c
> > +++ b/kernel/linux/kni/kni_misc.c
> > @@ -481,6 +481,50 @@ kni_ioctl_release(struct net *net, uint32_t
> ioctl_num,
> > return ret;
> > }
> >
> > +static int
> > +kni_ioctl_link(struct net *net, uint32_t ioctl_num,
> > + unsigned long ioctl_param)
> > +{
> > + struct kni_net *knet = net_generic(net, kni_net_id);
> > + int ret = -EINVAL;
> > + struct kni_dev *dev, *n;
> > + struct rte_kni_link_info link_info;
> > + struct net_device *netdev;
> > +
> > + if (_IOC_SIZE(ioctl_num) > sizeof(link_info))
> > + return -EINVAL;
> > +
> > + if (copy_from_user(&link_info, (void *)ioctl_param,
> sizeof(link_info)))
> > + return -EFAULT;
> > +
> > + if (strlen(link_info.name) == 0)
> > + return -EINVAL;
> > +
> > + down_read(&knet->kni_list_lock);
> > + list_for_each_entry_safe(dev, n, &knet->kni_list_head, list) {
> > + if (strncmp(dev->name, link_info.name, RTE_KNI_NAMESIZE)
> != 0)
> > + continue;
> > +
> > + netdev = dev->net_dev;
> > +
> > + if (link_info.status) {
> > + netif_carrier_on(netdev);
> > +
> > + dev->speed = link_info.speed;
> > + dev->duplex = link_info.duplex;
> > + dev->autoneg = link_info.autoneg;
> > + } else {
> > + netif_carrier_off(netdev);
> > + }
> > +
> > + ret = 0;
> > + break;
> > + }
> > + up_read(&knet->kni_list_lock);
> > +
> > + return ret;
> > +}
> > +
> > static int
> > kni_ioctl(struct inode *inode, uint32_t ioctl_num, unsigned long
> ioctl_param)
> > {
> > @@ -502,6 +546,9 @@ kni_ioctl(struct inode *inode, uint32_t ioctl_num,
> unsigned long ioctl_param)
> > case _IOC_NR(RTE_KNI_IOCTL_RELEASE):
> > ret = kni_ioctl_release(net, ioctl_num, ioctl_param);
> > break;
> > + case _IOC_NR(RTE_KNI_IOCTL_LINK):
> > + ret = kni_ioctl_link(net, ioctl_num, ioctl_param);
> > + break;
> > default:
> > pr_debug("IOCTL default\n");
> > break;
> > diff --git a/kernel/linux/kni/kni_net.c b/kernel/linux/kni/kni_net.c
> > index 611719b5ee27..99da8d37dd6b 100644
> > --- a/kernel/linux/kni/kni_net.c
> > +++ b/kernel/linux/kni/kni_net.c
> > @@ -778,18 +778,6 @@ kni_net_set_mac(struct net_device *netdev, void *p)
> > return (ret == 0 ? req.result : ret);
> > }
> >
> > -#ifdef HAVE_CHANGE_CARRIER_CB
> > -static int
> > -kni_net_change_carrier(struct net_device *dev, bool new_carrier)
> > -{
> > - if (new_carrier)
> > - netif_carrier_on(dev);
> > - else
> > - netif_carrier_off(dev);
> > - return 0;
> > -}
> > -#endif
> > -
> > static const struct header_ops kni_net_header_ops = {
> > .create = kni_net_header,
> > .parse = eth_header_parse,
> > @@ -808,9 +796,6 @@ static const struct net_device_ops
> kni_net_netdev_ops = {
> > .ndo_change_mtu = kni_net_change_mtu,
> > .ndo_tx_timeout = kni_net_tx_timeout,
> > .ndo_set_mac_address = kni_net_set_mac,
> > -#ifdef HAVE_CHANGE_CARRIER_CB
> > - .ndo_change_carrier = kni_net_change_carrier,
> > -#endif
>
> What do you think keeping the 'carrier' interface? Even the API not using
> it, it
> can be useful, what do you think?
>
We can keep it, but if people continue to write to the carrier file instead
of using the API,
then speed/duplex/autoneg values will have unspecified values. We can
reflect this in
the docs.
>
> > };
> >
> > static void kni_get_drvinfo(struct net_device *dev,
> > diff --git a/lib/kni/rte_kni.c b/lib/kni/rte_kni.c
> > index eb24b0d0ae4e..0a7b562abf11 100644
> > --- a/lib/kni/rte_kni.c
> > +++ b/lib/kni/rte_kni.c
> > @@ -784,43 +784,25 @@ rte_kni_unregister_handlers(struct rte_kni *kni)
> > }
> >
> > int
> > -rte_kni_update_link(struct rte_kni *kni, unsigned int linkup)
> > +rte_kni_update_link(struct rte_kni *kni, struct rte_eth_link *link)
> > {
> > - char path[64];
> > - char old_carrier[2];
> > - const char *new_carrier;
> > - int old_linkup;
> > - int fd, ret;
> > + struct rte_kni_link_info link_info;
> >
> > if (kni == NULL)
> > return -1;
> >
> > - snprintf(path, sizeof(path), "/sys/devices/virtual/net/%s/carrier",
> > - kni->name);
> > + snprintf(link_info.name, RTE_KNI_NAMESIZE, "%s", kni->name);
> > + link_info.speed = link->link_speed;
> > + link_info.duplex = link->link_duplex;
> > + link_info.autoneg = link->link_autoneg;
> > + link_info.status = link->link_status;
> >
> > - fd = open(path, O_RDWR);
> > - if (fd == -1) {
> > - RTE_LOG(ERR, KNI, "Failed to open file: %s.\n", path);
> > + if (ioctl(kni_fd, RTE_KNI_IOCTL_LINK, &link_info) < 0) {
> > + RTE_LOG(ERR, KNI, "Fail to update KNI link\n");
> > return -1;
> > }
> >
> > - ret = read(fd, old_carrier, 2);
> > - if (ret < 1) {
> > - close(fd);
> > - return -1;
> > - }
> > - old_linkup = (old_carrier[0] == '1');
> > -
> > - new_carrier = linkup ? "1" : "0";
> > - ret = write(fd, new_carrier, 1);
> > - if (ret < 1) {
> > - RTE_LOG(ERR, KNI, "Failed to write file: %s.\n", path);
> > - close(fd);
> > - return -1;
> > - }
> > -
> > - close(fd);
> > - return old_linkup;
> > + return 0;
> > }
>
> Can you please add new API to KNI PMD too? To set the link up when port
> started.
>
I'll check KNI PMD and update.
>
> >
> > void
> > diff --git a/lib/kni/rte_kni.h b/lib/kni/rte_kni.h
> > index b0eaf4610416..ab5990fd210d 100644
> > --- a/lib/kni/rte_kni.h
> > +++ b/lib/kni/rte_kni.h
> > @@ -21,6 +21,7 @@
> > #include <rte_memory.h>
> > #include <rte_mempool.h>
> > #include <rte_ether.h>
> > +#include <rte_ethdev.h>
> >
> > #include <rte_kni_common.h>
> >
> > @@ -245,18 +246,15 @@ int rte_kni_unregister_handlers(struct rte_kni
> *kni);
> > * @param kni
> > * pointer to struct rte_kni.
> > * @param linkup
> > - * New link state:
> > - * 0 for linkdown.
> > - * > 0 for linkup.
> > + * new link state, speed, duplex, autoneg.
> > *
> > * @return
> > + * On success: 0
> > * On failure: -1
> > - * Previous link state == linkdown: 0
> > - * Previous link state == linkup: 1
> > */
>
> This was useful to detect the change in the status and log according.
> With current change KNI example will keep logging the link state.
>
> Also in the KNI sample it periodically reads the link status from the
> device and
> applies it to the kni device, with current implementation it will keep
> updating
> the link status. Can we add a logic in the KNI sample so that update the
> KNI
> interface only if physical device link status changed?
>
We can still read from the carrier file and return the old state if it's
really needed.
I'll check the KNI sample code and try to find the best solution.
>
> > __rte_experimental
> > int
> > -rte_kni_update_link(struct rte_kni *kni, unsigned int linkup);
> > +rte_kni_update_link(struct rte_kni *kni, struct rte_eth_link *link);
> >
> > /**
> > * Close KNI device.
> > diff --git a/lib/kni/rte_kni_common.h b/lib/kni/rte_kni_common.h
> > index b547ea550171..35a1620e0bd2 100644
> > --- a/lib/kni/rte_kni_common.h
> > +++ b/lib/kni/rte_kni_common.h
> > @@ -130,10 +130,19 @@ struct rte_kni_device_info {
> > uint8_t iova_mode;
> > };
> >
> > +struct rte_kni_link_info {
> > + char name[RTE_KNI_NAMESIZE];
> > + unsigned int speed;
> > + unsigned char duplex;
> > + unsigned char autoneg;
> > + unsigned char status;
> > +};
>
> Should we prefer fixed size storage types "uint32_t speed" or "unsigned
> int",
> did you select current ones explicitly?
>
Sure, I'll update. Not sure why I chose it initially, it was a long time
ago :)
>
> > +
> > #define KNI_DEVICE "kni"
> >
> > #define RTE_KNI_IOCTL_TEST _IOWR(0, 1, int)
> > #define RTE_KNI_IOCTL_CREATE _IOWR(0, 2, struct rte_kni_device_info)
> > #define RTE_KNI_IOCTL_RELEASE _IOWR(0, 3, struct rte_kni_device_info)
> > +#define RTE_KNI_IOCTL_LINK _IOWR(0, 4, struct rte_kni_link_info)
> >
> > #endif /* _RTE_KNI_COMMON_H_ */
> >
>
>
next prev parent reply other threads:[~2021-07-06 9:14 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-24 19:25 [dpdk-dev] [PATCH] " Igor Ryzhov
2019-09-24 19:33 ` [dpdk-dev] [PATCH v2] " Igor Ryzhov
2019-09-24 20:37 ` Aaron Conole
2019-09-25 9:00 ` Igor Ryzhov
2019-09-25 9:36 ` [dpdk-dev] [PATCH v3] " Igor Ryzhov
2019-10-14 16:10 ` Ferruh Yigit
2019-10-14 16:17 ` Ferruh Yigit
2019-10-14 19:01 ` Dan Gora
2019-10-14 20:55 ` Dan Gora
2019-10-27 20:16 ` Igor Ryzhov
2021-06-28 12:55 ` Ferruh Yigit
2021-06-28 13:16 ` Igor Ryzhov
2021-07-04 16:06 ` [dpdk-dev] [PATCH v4 1/2] " Igor Ryzhov
2021-07-04 16:06 ` [dpdk-dev] [PATCH v4 2/2] kni: implement basic get_link_ksettings callback Igor Ryzhov
2021-07-05 11:58 ` [dpdk-dev] [PATCH v4 1/2] kni: rework rte_kni_update_link using ioctl Ferruh Yigit
2021-07-06 9:14 ` Igor Ryzhov [this message]
2021-08-26 15:19 ` [dpdk-dev] [PATCH v5 1/3] " Igor Ryzhov
2021-08-26 15:19 ` [dpdk-dev] [PATCH v5 2/3] kni: implement basic get_link_ksettings callback Igor Ryzhov
2021-08-26 15:19 ` [dpdk-dev] [PATCH v5 3/3] app/test: fix return value of test_kni_link_change Igor Ryzhov
2021-08-26 17:15 ` [dpdk-dev] [PATCH v5 1/3] kni: rework rte_kni_update_link using ioctl Stephen Hemminger
2021-08-26 17:46 ` Igor Ryzhov
2021-08-26 18:06 ` Stephen Hemminger
2021-08-30 18:05 ` Igor Ryzhov
2021-08-30 14:27 ` [dpdk-dev] [PATCH v6 " Igor Ryzhov
2021-08-30 14:27 ` [dpdk-dev] [PATCH v6 2/3] kni: implement basic get_link_ksettings callback Igor Ryzhov
2021-08-30 14:27 ` [dpdk-dev] [PATCH v6 3/3] app/test: fix return value of test_kni_link_change Igor Ryzhov
2023-06-29 17:05 ` [dpdk-dev] [PATCH v6 1/3] kni: rework rte_kni_update_link using ioctl Stephen Hemminger
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='CAF+s_Fz3Jr9Yy+3+mmGzq5PoZvHCR=wSSWMBS1d8DoyOaPZ6eA@mail.gmail.com' \
--to=iryzhov@nfware.com \
--cc=dev@dpdk.org \
--cc=dg@adax.com \
--cc=ferruh.yigit@intel.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).