netdev - [PATCH] netlink: add real_num_[tr]x

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <20200130175749.GA31391@zeil-osx>
Date:   Thu, 30 Jan 2020 20:57:49 +0300
From:   Dmitry Yakunin <zeil@...dex-team.ru>
To:     davem@...emloft.net, kuba@...nel.org
Cc:     netdev@...r.kernel.org
Subject: [PATCH] netlink: add real_num_[tr]x_queues to ifinfo

This patch adds the number of active tx/rx queues to ifinfo message.

Now there are two ways of determining the number of active queues:
1) by counting entries in /sys/class/net/eth[0-9]+/queues/
2) by ioctl syscall if a driver implements ethtool_ops->get_channels()

Default mq qdisc sets up pfifo_fast only for active queues. So, if we want
to reproduce this behavior with custom leaf qdiscs, we should use one
of these methods which are foreign for the code where netlink was used.

Signed-off-by: Dmitry Yakunin <zeil@...dex-team.ru>
---
 include/uapi/linux/if_link.h | 2 ++
 net/core/rtnetlink.c         | 6 ++++++
 2 files changed, 8 insertions(+)

diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
index 8aec876..6566b63 100644
--- a/include/uapi/linux/if_link.h
+++ b/include/uapi/linux/if_link.h
@@ -169,6 +169,8 @@ enum {
 	IFLA_MAX_MTU,
 	IFLA_PROP_LIST,
 	IFLA_ALT_IFNAME, /* Alternative ifname */
+	IFLA_REAL_NUM_TX_QUEUES,
+	IFLA_REAL_NUM_RX_QUEUES,
 	__IFLA_MAX
 };
 
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index d9001b5..3ebae18 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -1013,7 +1013,9 @@ static noinline size_t if_nlmsg_size(const struct net_device *dev,
 	       + nla_total_size(1) /* IFLA_CARRIER */
 	       + nla_total_size(4) /* IFLA_PROMISCUITY */
 	       + nla_total_size(4) /* IFLA_NUM_TX_QUEUES */
+	       + nla_total_size(4) /* IFLA_REAL_NUM_TX_QUEUES */
 	       + nla_total_size(4) /* IFLA_NUM_RX_QUEUES */
+	       + nla_total_size(4) /* IFLA_REAL_NUM_RX_QUEUES */
 	       + nla_total_size(4) /* IFLA_GSO_MAX_SEGS */
 	       + nla_total_size(4) /* IFLA_GSO_MAX_SIZE */
 	       + nla_total_size(1) /* IFLA_OPERSTATE */
@@ -1687,10 +1689,12 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb,
 	    nla_put_u32(skb, IFLA_GROUP, dev->group) ||
 	    nla_put_u32(skb, IFLA_PROMISCUITY, dev->promiscuity) ||
 	    nla_put_u32(skb, IFLA_NUM_TX_QUEUES, dev->num_tx_queues) ||
+	    nla_put_u32(skb, IFLA_REAL_NUM_TX_QUEUES, dev->real_num_tx_queues) ||
 	    nla_put_u32(skb, IFLA_GSO_MAX_SEGS, dev->gso_max_segs) ||
 	    nla_put_u32(skb, IFLA_GSO_MAX_SIZE, dev->gso_max_size) ||
 #ifdef CONFIG_RPS
 	    nla_put_u32(skb, IFLA_NUM_RX_QUEUES, dev->num_rx_queues) ||
+	    nla_put_u32(skb, IFLA_REAL_NUM_RX_QUEUES, dev->real_num_rx_queues) ||
 #endif
 	    put_master_ifindex(skb, dev) ||
 	    nla_put_u8(skb, IFLA_CARRIER, netif_carrier_ok(dev)) ||
@@ -1803,7 +1807,9 @@ static const struct nla_policy ifla_policy[IFLA_MAX+1] = {
 	[IFLA_EXT_MASK]		= { .type = NLA_U32 },
 	[IFLA_PROMISCUITY]	= { .type = NLA_U32 },
 	[IFLA_NUM_TX_QUEUES]	= { .type = NLA_U32 },
+	[IFLA_REAL_NUM_TX_QUEUES] = { .type = NLA_U32 },
 	[IFLA_NUM_RX_QUEUES]	= { .type = NLA_U32 },
+	[IFLA_REAL_NUM_RX_QUEUES] = { .type = NLA_U32 },
 	[IFLA_GSO_MAX_SEGS]	= { .type = NLA_U32 },
 	[IFLA_GSO_MAX_SIZE]	= { .type = NLA_U32 },
 	[IFLA_PHYS_PORT_ID]	= { .type = NLA_BINARY, .len = MAX_PHYS_ITEM_ID_LEN },
-- 
2.7.4