* [PATCH net-next 08/22] net: apple: fix return type of ndo_start_xmit function
From: YueHaibing @ 2018-09-20 12:32 UTC (permalink / raw)
To: davem, dmitry.tarnyagin, wg, mkl, michal.simek, hsweeten,
madalin.bucur, pantelis.antoniou, claudiu.manoil, leoyang.li,
linux, sammy, ralf, nico, steve.glendinning, f.fainelli,
grygorii.strashko, w-kwok2, m-karicheri2, t.sailer, jreuter, kys,
haiyangz, wei.liu2, paul.durrant, arvid.brodin, pshelar
Cc: linux-kernel, netdev, linux-can, linux-arm-kernel, linuxppc-dev,
linux-mips, linux-omap, linux-hams, devel, linux-usb, xen-devel,
dev, YueHaibing
In-Reply-To: <20180920123306.14772-1-yuehaibing@huawei.com>
The method ndo_start_xmit() is defined as returning an 'netdev_tx_t',
which is a typedef for an enum type, so make sure the implementation in
this driver has returns 'netdev_tx_t' value, and change the function
return type to netdev_tx_t.
Found by coccinelle.
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
---
drivers/net/ethernet/apple/bmac.c | 4 ++--
drivers/net/ethernet/apple/mace.c | 4 ++--
drivers/net/ethernet/apple/macmace.c | 4 ++--
3 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/apple/bmac.c b/drivers/net/ethernet/apple/bmac.c
index 024998d..6a8e256 100644
--- a/drivers/net/ethernet/apple/bmac.c
+++ b/drivers/net/ethernet/apple/bmac.c
@@ -154,7 +154,7 @@ struct bmac_data {
static irqreturn_t bmac_rxdma_intr(int irq, void *dev_id);
static void bmac_set_timeout(struct net_device *dev);
static void bmac_tx_timeout(struct timer_list *t);
-static int bmac_output(struct sk_buff *skb, struct net_device *dev);
+static netdev_tx_t bmac_output(struct sk_buff *skb, struct net_device *dev);
static void bmac_start(struct net_device *dev);
#define DBDMA_SET(x) ( ((x) | (x) << 16) )
@@ -1456,7 +1456,7 @@ static int bmac_close(struct net_device *dev)
spin_unlock_irqrestore(&bp->lock, flags);
}
-static int
+static netdev_tx_t
bmac_output(struct sk_buff *skb, struct net_device *dev)
{
struct bmac_data *bp = netdev_priv(dev);
diff --git a/drivers/net/ethernet/apple/mace.c b/drivers/net/ethernet/apple/mace.c
index 0b5429d..68b9ee4 100644
--- a/drivers/net/ethernet/apple/mace.c
+++ b/drivers/net/ethernet/apple/mace.c
@@ -78,7 +78,7 @@ struct mace_data {
static int mace_open(struct net_device *dev);
static int mace_close(struct net_device *dev);
-static int mace_xmit_start(struct sk_buff *skb, struct net_device *dev);
+static netdev_tx_t mace_xmit_start(struct sk_buff *skb, struct net_device *dev);
static void mace_set_multicast(struct net_device *dev);
static void mace_reset(struct net_device *dev);
static int mace_set_address(struct net_device *dev, void *addr);
@@ -525,7 +525,7 @@ static inline void mace_set_timeout(struct net_device *dev)
mp->timeout_active = 1;
}
-static int mace_xmit_start(struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t mace_xmit_start(struct sk_buff *skb, struct net_device *dev)
{
struct mace_data *mp = netdev_priv(dev);
volatile struct dbdma_regs __iomem *td = mp->tx_dma;
diff --git a/drivers/net/ethernet/apple/macmace.c b/drivers/net/ethernet/apple/macmace.c
index 137cbb4..376f2c2 100644
--- a/drivers/net/ethernet/apple/macmace.c
+++ b/drivers/net/ethernet/apple/macmace.c
@@ -89,7 +89,7 @@ struct mace_frame {
static int mace_open(struct net_device *dev);
static int mace_close(struct net_device *dev);
-static int mace_xmit_start(struct sk_buff *skb, struct net_device *dev);
+static netdev_tx_t mace_xmit_start(struct sk_buff *skb, struct net_device *dev);
static void mace_set_multicast(struct net_device *dev);
static int mace_set_address(struct net_device *dev, void *addr);
static void mace_reset(struct net_device *dev);
@@ -444,7 +444,7 @@ static int mace_close(struct net_device *dev)
* Transmit a frame
*/
-static int mace_xmit_start(struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t mace_xmit_start(struct sk_buff *skb, struct net_device *dev)
{
struct mace_data *mp = netdev_priv(dev);
unsigned long flags;
--
1.8.3.1
^ permalink raw reply related
* [PATCH net-next 05/22] net: sgi: fix return type of ndo_start_xmit function
From: YueHaibing @ 2018-09-20 12:32 UTC (permalink / raw)
To: davem, dmitry.tarnyagin, wg, mkl, michal.simek, hsweeten,
madalin.bucur, pantelis.antoniou, claudiu.manoil, leoyang.li,
linux, sammy, ralf, nico, steve.glendinning, f.fainelli,
grygorii.strashko, w-kwok2, m-karicheri2, t.sailer, jreuter, kys,
haiyangz, wei.liu2, paul.durrant, arvid.brodin, pshelar
Cc: linux-kernel, netdev, linux-can, linux-arm-kernel, linuxppc-dev,
linux-mips, linux-omap, linux-hams, devel, linux-usb, xen-devel,
dev, YueHaibing
In-Reply-To: <20180920123306.14772-1-yuehaibing@huawei.com>
The method ndo_start_xmit() is defined as returning an 'netdev_tx_t',
which is a typedef for an enum type, so make sure the implementation in
this driver has returns 'netdev_tx_t' value, and change the function
return type to netdev_tx_t.
Found by coccinelle.
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
---
drivers/net/ethernet/sgi/ioc3-eth.c | 4 ++--
drivers/net/ethernet/sgi/meth.c | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/sgi/ioc3-eth.c b/drivers/net/ethernet/sgi/ioc3-eth.c
index 18d533f..3140999 100644
--- a/drivers/net/ethernet/sgi/ioc3-eth.c
+++ b/drivers/net/ethernet/sgi/ioc3-eth.c
@@ -99,7 +99,7 @@ struct ioc3_private {
static int ioc3_ioctl(struct net_device *dev, struct ifreq *rq, int cmd);
static void ioc3_set_multicast_list(struct net_device *dev);
-static int ioc3_start_xmit(struct sk_buff *skb, struct net_device *dev);
+static netdev_tx_t ioc3_start_xmit(struct sk_buff *skb, struct net_device *dev);
static void ioc3_timeout(struct net_device *dev);
static inline unsigned int ioc3_hash(const unsigned char *addr);
static inline void ioc3_stop(struct ioc3_private *ip);
@@ -1390,7 +1390,7 @@ static void ioc3_remove_one(struct pci_dev *pdev)
.remove = ioc3_remove_one,
};
-static int ioc3_start_xmit(struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t ioc3_start_xmit(struct sk_buff *skb, struct net_device *dev)
{
unsigned long data;
struct ioc3_private *ip = netdev_priv(dev);
diff --git a/drivers/net/ethernet/sgi/meth.c b/drivers/net/ethernet/sgi/meth.c
index ea55abd..703fbbe 100644
--- a/drivers/net/ethernet/sgi/meth.c
+++ b/drivers/net/ethernet/sgi/meth.c
@@ -697,7 +697,7 @@ static void meth_add_to_tx_ring(struct meth_private *priv, struct sk_buff *skb)
/*
* Transmit a packet (called by the kernel)
*/
-static int meth_tx(struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t meth_tx(struct sk_buff *skb, struct net_device *dev)
{
struct meth_private *priv = netdev_priv(dev);
unsigned long flags;
--
1.8.3.1
^ permalink raw reply related
* [PATCH net-next 00/22] net: fix return type of ndo_start_xmit function
From: YueHaibing @ 2018-09-20 12:32 UTC (permalink / raw)
To: davem, dmitry.tarnyagin, wg, mkl, michal.simek, hsweeten,
madalin.bucur, pantelis.antoniou, claudiu.manoil, leoyang.li,
linux, sammy, ralf, nico, steve.glendinning, f.fainelli,
grygorii.strashko, w-kwok2, m-karicheri2, t.sailer, jreuter, kys,
haiyangz, wei.liu2, paul.durrant, arvid.brodin, pshelar
Cc: linux-kernel, netdev, linux-can, linux-arm-kernel, linuxppc-dev,
linux-mips, linux-omap, linux-hams, devel, linux-usb, xen-devel,
dev, YueHaibing
The method ndo_start_xmit() is defined as returning an 'netdev_tx_t',
which is a typedef for an enum type, so make sure the implementation in
this driver has returns 'netdev_tx_t' value, and change the function
return type to netdev_tx_t.
YueHaibing (22):
net: micrel: fix return type of ndo_start_xmit function
net: freescale: fix return type of ndo_start_xmit function
net: seeq: fix return type of ndo_start_xmit function
net: cirrus: fix return type of ndo_start_xmit function
net: sgi: fix return type of ndo_start_xmit function
net: wiznet: fix return type of ndo_start_xmit function
net: i825xx: fix return type of ndo_start_xmit function
net: apple: fix return type of ndo_start_xmit function
net: smsc: fix return type of ndo_start_xmit function
net: ti: fix return type of ndo_start_xmit function
net: faraday: fix return type of ndo_start_xmit function
net: ovs: fix return type of ndo_start_xmit function
net: xen-netback: fix return type of ndo_start_xmit function
net: caif: fix return type of ndo_start_xmit function
net: hamradio: fix return type of ndo_start_xmit function
usbnet: ipheth: fix return type of ndo_start_xmit function
hv_netvsc: fix return type of ndo_start_xmit function
can: xilinx: fix return type of ndo_start_xmit function
net: plip: fix return type of ndo_start_xmit function
rionet: fix return type of ndo_start_xmit function
l2tp: fix return type of ndo_start_xmit function
net: hsr: fix return type of ndo_start_xmit function
drivers/net/caif/caif_hsi.c | 10 +++++-----
drivers/net/caif/caif_serial.c | 7 +++++--
drivers/net/caif/caif_spi.c | 6 +++---
drivers/net/caif/caif_virtio.c | 2 +-
drivers/net/can/xilinx_can.c | 2 +-
drivers/net/ethernet/apple/bmac.c | 4 ++--
drivers/net/ethernet/apple/mace.c | 4 ++--
drivers/net/ethernet/apple/macmace.c | 4 ++--
drivers/net/ethernet/cirrus/ep93xx_eth.c | 2 +-
drivers/net/ethernet/cirrus/mac89x0.c | 4 ++--
drivers/net/ethernet/faraday/ftgmac100.c | 4 ++--
drivers/net/ethernet/faraday/ftmac100.c | 7 ++++---
drivers/net/ethernet/freescale/dpaa/dpaa_eth.c | 3 ++-
drivers/net/ethernet/freescale/fec_mpc52xx.c | 3 ++-
drivers/net/ethernet/freescale/fs_enet/fs_enet-main.c | 3 ++-
drivers/net/ethernet/freescale/gianfar.c | 4 ++--
drivers/net/ethernet/freescale/ucc_geth.c | 3 ++-
drivers/net/ethernet/i825xx/ether1.c | 5 +++--
drivers/net/ethernet/i825xx/lib82596.c | 4 ++--
drivers/net/ethernet/i825xx/sun3_82586.c | 6 ++++--
drivers/net/ethernet/micrel/ks8695net.c | 2 +-
drivers/net/ethernet/micrel/ks8851_mll.c | 4 ++--
drivers/net/ethernet/seeq/ether3.c | 5 +++--
drivers/net/ethernet/seeq/sgiseeq.c | 3 ++-
drivers/net/ethernet/sgi/ioc3-eth.c | 4 ++--
drivers/net/ethernet/sgi/meth.c | 2 +-
drivers/net/ethernet/smsc/smc911x.c | 3 ++-
drivers/net/ethernet/smsc/smc91x.c | 3 ++-
drivers/net/ethernet/smsc/smsc911x.c | 3 ++-
drivers/net/ethernet/ti/cpmac.c | 2 +-
drivers/net/ethernet/ti/davinci_emac.c | 2 +-
drivers/net/ethernet/ti/netcp_core.c | 8 ++++----
drivers/net/ethernet/wiznet/w5100.c | 2 +-
drivers/net/ethernet/wiznet/w5300.c | 2 +-
drivers/net/hamradio/baycom_epp.c | 3 ++-
drivers/net/hamradio/dmascc.c | 4 ++--
drivers/net/hyperv/netvsc_drv.c | 10 +++++++---
drivers/net/plip/plip.c | 4 ++--
drivers/net/rionet.c | 3 ++-
drivers/net/usb/ipheth.c | 2 +-
drivers/net/xen-netback/interface.c | 3 ++-
net/caif/chnl_net.c | 3 ++-
net/hsr/hsr_device.c | 2 +-
net/l2tp/l2tp_eth.c | 3 ++-
net/openvswitch/vport-internal_dev.c | 5 +++--
45 files changed, 100 insertions(+), 74 deletions(-)
--
1.8.3.1
^ permalink raw reply
* Re: [PATCH] net-next: mscc: remove unused ocelot_dev_gmii.h
From: David Miller @ 2018-09-20 18:25 UTC (permalink / raw)
To: clabbe; +Cc: alexandre.belloni, linux-kernel, netdev
In-Reply-To: <1537384013-2353-1-git-send-email-clabbe@baylibre.com>
From: Corentin Labbe <clabbe@baylibre.com>
Date: Wed, 19 Sep 2018 19:06:53 +0000
> The header ocelot_dev_gmii.h is unused since the inclusion of the driver.
> It is unused, lets just remove it.
>
> Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Yeah if it's unused, let's toss it. If future code uses it we can bring it
back, if necessary.
Thanks!
^ permalink raw reply
* Re: [PATCH v7 4/4] gpiolib: Implement fast processing path in get/set array
From: Dan Carpenter @ 2018-09-20 18:05 UTC (permalink / raw)
To: Janusz Krzysztofik
Cc: Andrew Lunn, Ulf Hansson, linux-doc, linux-iio, Linus Walleij,
Dominik Brodowski, Yegor Yefremov, David S. Miller, linux-i2c,
Peter Meerwald-Stadler, Marek Szyprowski, devel, Florian Fainelli,
Jonathan Corbet, Krzysztof Kozlowski, Kishon Vijay Abraham I,
Tony Lindgren, Peter Korsgaard, Geert Uytterhoeven, linux-serial,
Jiri Slaby, Michael Hennerich, Uwe Kleine-König
In-Reply-To: <9860023.SlBYqtbjDV@z50>
On Thu, Sep 20, 2018 at 05:48:22PM +0200, Janusz Krzysztofik wrote:
> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> index a53d17745d21..5bc3447949c9 100644
> --- a/drivers/gpio/gpiolib.c
> +++ b/drivers/gpio/gpiolib.c
> @@ -2880,7 +2880,7 @@ int gpiod_get_array_value_complex(bool raw, bool can_sleep,
> __set_bit(hwgpio, mask);
>
> if (array_info)
> - find_next_zero_bit(array_info->get_mask,
> + i = find_next_zero_bit(array_info->get_mask,
> array_size, i);
We could mark find_next_zero_bit() and friends as a __must_check
functions so we avoid this bug in the future. I have a more complicated
idea how to detect these bugs in a generic way using Smatch but it will
take longer to implement.
regards,
dan carpenter
^ permalink raw reply
* Re: [PATCH] qtnfmac: remove set but not used variable 'vif'
From: Kalle Valo @ 2018-09-20 12:12 UTC (permalink / raw)
To: YueHaibing
Cc: Igor Mitsyanko, Avinash Patil, Sergey Matyukevich,
Sergei Maksimenko, Vasily Ulyanov, Andrey Shevchenko, YueHaibing,
linux-wireless, netdev, kernel-janitors
In-Reply-To: <1536668904-13110-1-git-send-email-yuehaibing@huawei.com>
YueHaibing <yuehaibing@huawei.com> wrote:
> Fixes gcc '-Wunused-but-set-variable' warning:
>
> drivers/net/wireless/quantenna/qtnfmac/cfg80211.c: In function 'qtnf_dump_survey':
> drivers/net/wireless/quantenna/qtnfmac/cfg80211.c:694:19: warning:
> variable 'vif' set but not used [-Wunused-but-set-variable]
>
> Signed-off-by: YueHaibing <yuehaibing@huawei.com>
> Reviewed-by: Sergey Matyukevich <sergey.matyukevich.os@quantenna.com>
Patch applied to wireless-drivers-next.git, thanks.
93ed990e3a6e qtnfmac: remove set but not used variable 'vif'
--
https://patchwork.kernel.org/patch/10595609/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
^ permalink raw reply
* Re: [PATCH] brcm80211: remove redundant condition check before debugfs_remove_recursive
From: Kalle Valo @ 2018-09-20 12:10 UTC (permalink / raw)
To: zhong jiang
Cc: davem, arend.vanspriel, franky.lin, hante.meuleman, chi-hsien.lin,
wright.feng, linux-wireless, brcm80211-dev-list.pdl, netdev
In-Reply-To: <1536414043-32146-1-git-send-email-zhongjiang@huawei.com>
zhong jiang <zhongjiang@huawei.com> wrote:
> debugfs_remove_recursive has taken IS_ERR_OR_NULL into account. So just
> remove the condition check before debugfs_remove_recursive.
>
> Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Patch applied to wireless-drivers-next.git, thanks.
761cb7cdef1a brcm80211: remove redundant condition check before debugfs_remove_recursive
--
https://patchwork.kernel.org/patch/10593061/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
^ permalink raw reply
* Re: [PATCH] bpf: remove redundant null pointer check before consume_skb
From: Song Liu @ 2018-09-20 17:53 UTC (permalink / raw)
To: zhongjiang; +Cc: Alexei Starovoitov, Daniel Borkmann, Networking, open list
In-Reply-To: <1537436772-42545-1-git-send-email-zhongjiang@huawei.com>
On Thu, Sep 20, 2018 at 3:00 AM zhong jiang <zhongjiang@huawei.com> wrote:
>
> consume_skb has taken the null pointer into account. hence it is safe
> to remove the redundant null pointer check before consume_skb.
>
> Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Acked-by: Song Liu <songliubraving@fb.com>
> ---
> kernel/bpf/sockmap.c | 6 ++----
> 1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c
> index 488ef96..a9359cb 100644
> --- a/kernel/bpf/sockmap.c
> +++ b/kernel/bpf/sockmap.c
> @@ -590,8 +590,7 @@ static int free_sg(struct sock *sk, int start,
> if (i == MAX_SKB_FRAGS)
> i = 0;
> }
> - if (md->skb)
> - consume_skb(md->skb);
> + consume_skb(md->skb);
>
> return free;
> }
> @@ -973,8 +972,7 @@ static int bpf_tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
>
> if (!sg->length && md->sg_start == md->sg_end) {
> list_del(&md->list);
> - if (md->skb)
> - consume_skb(md->skb);
> + consume_skb(md->skb);
> kfree(md);
> }
> }
> --
> 1.7.12.4
>
^ permalink raw reply
* Re: [PATCH] docs: fix some broken documentation references
From: Jonathan Corbet @ 2018-09-20 17:10 UTC (permalink / raw)
To: Mauro Carvalho Chehab
Cc: linux-kernel, Linux Media Mailing List, Mauro Carvalho Chehab,
Jan Kara, Stephen Hemminger, David S. Miller, Michael Ellerman,
Greg Kroah-Hartman, Andrew Morton, Arnd Bergmann, linux-doc,
linux-ext4, bridge, netdev
In-Reply-To: <6b47bf56b898c48a0dc3cd42283c9e5c7c23367a.1537210894.git.mchehab+samsung@kernel.org>
On Mon, 17 Sep 2018 15:02:34 -0400
Mauro Carvalho Chehab <mchehab+samsung@kernel.org> wrote:
> Some documentation files received recent changes and are
> pointing to wrong places.
>
> Those references can easily fixed with the help of a
> script:
>
> $ ./scripts/documentation-file-ref-check --fix
>
> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Applied, thanks.
jon
^ permalink raw reply
* [PATCH net-next,3/3] hv_netvsc: Update document for LRO/RSC support
From: Haiyang Zhang @ 2018-09-20 17:07 UTC (permalink / raw)
To: davem, netdev
Cc: haiyangz, kys, sthemmin, olaf, vkuznets, devel, linux-kernel
In-Reply-To: <20180920170701.9574-1-haiyangz@linuxonhyperv.com>
From: Haiyang Zhang <haiyangz@microsoft.com>
Update document for LRO/RSC support, and the command line info to
change the setting.
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
---
Documentation/networking/netvsc.txt | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/Documentation/networking/netvsc.txt b/Documentation/networking/netvsc.txt
index 92f5b31392fa..3bfa635bbbd5 100644
--- a/Documentation/networking/netvsc.txt
+++ b/Documentation/networking/netvsc.txt
@@ -45,6 +45,15 @@ Features
like packets and significantly reduces CPU usage under heavy Rx
load.
+ Large Receive Offload (LRO), or Receive Side Coalescing (RSC)
+ -------------------------------------------------------------
+ The driver supports LRO/RSC in the vSwitch feature. It reduces the per packet
+ processing overhead by coalescing multiple TCP segments when possible. The
+ feature is enabled by default on VMs running on Windows Server 2019 and
+ later. It may be changed by ethtool command:
+ ethtool -K eth0 lro on
+ ethtool -K eth0 lro off
+
SR-IOV support
--------------
Hyper-V supports SR-IOV as a hardware acceleration option. If SR-IOV
--
2.18.0
^ permalink raw reply related
* [PATCH net-next,2/3] hv_netvsc: Add handler for LRO setting change
From: Haiyang Zhang @ 2018-09-20 17:07 UTC (permalink / raw)
To: davem, netdev; +Cc: olaf, sthemmin, haiyangz, linux-kernel, devel, vkuznets
In-Reply-To: <20180920170701.9574-1-haiyangz@linuxonhyperv.com>
From: Haiyang Zhang <haiyangz@microsoft.com>
This patch adds the handler for LRO setting change, so that a user
can use ethtool command to enable / disable LRO feature.
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
---
drivers/net/hyperv/hyperv_net.h | 4 ++++
drivers/net/hyperv/netvsc_drv.c | 30 ++++++++++++++++++++++++++++++
drivers/net/hyperv/rndis_filter.c | 11 ++++++++---
3 files changed, 42 insertions(+), 3 deletions(-)
diff --git a/drivers/net/hyperv/hyperv_net.h b/drivers/net/hyperv/hyperv_net.h
index 7f1603dc8128..ef6f766f6389 100644
--- a/drivers/net/hyperv/hyperv_net.h
+++ b/drivers/net/hyperv/hyperv_net.h
@@ -185,6 +185,7 @@ struct rndis_device {
/* Interface */
struct rndis_message;
+struct ndis_offload_params;
struct netvsc_device;
struct netvsc_channel;
struct net_device_context;
@@ -218,6 +219,9 @@ void rndis_filter_device_remove(struct hv_device *dev,
struct netvsc_device *nvdev);
int rndis_filter_set_rss_param(struct rndis_device *rdev,
const u8 *key);
+int rndis_filter_set_offload_params(struct net_device *ndev,
+ struct netvsc_device *nvdev,
+ struct ndis_offload_params *req_offloads);
int rndis_filter_receive(struct net_device *ndev,
struct netvsc_device *net_dev,
struct netvsc_channel *nvchan,
diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c
index f8c18370d9d1..ec699741170b 100644
--- a/drivers/net/hyperv/netvsc_drv.c
+++ b/drivers/net/hyperv/netvsc_drv.c
@@ -1006,6 +1006,8 @@ static void netvsc_init_settings(struct net_device *dev)
ndc->speed = SPEED_UNKNOWN;
ndc->duplex = DUPLEX_FULL;
+
+ dev->features = NETIF_F_LRO;
}
static int netvsc_get_link_ksettings(struct net_device *dev,
@@ -1733,6 +1735,33 @@ static int netvsc_set_ringparam(struct net_device *ndev,
return ret;
}
+static int netvsc_set_features(struct net_device *ndev,
+ netdev_features_t features)
+{
+ netdev_features_t change = features ^ ndev->features;
+ struct net_device_context *ndevctx = netdev_priv(ndev);
+ struct netvsc_device *nvdev = rtnl_dereference(ndevctx->nvdev);
+ struct ndis_offload_params offloads;
+
+ if (!nvdev || nvdev->destroy)
+ return -ENODEV;
+
+ if (!(change & NETIF_F_LRO))
+ return 0;
+
+ memset(&offloads, 0, sizeof(struct ndis_offload_params));
+
+ if (features & NETIF_F_LRO) {
+ offloads.rsc_ip_v4 = NDIS_OFFLOAD_PARAMETERS_RSC_ENABLED;
+ offloads.rsc_ip_v6 = NDIS_OFFLOAD_PARAMETERS_RSC_ENABLED;
+ } else {
+ offloads.rsc_ip_v4 = NDIS_OFFLOAD_PARAMETERS_RSC_DISABLED;
+ offloads.rsc_ip_v6 = NDIS_OFFLOAD_PARAMETERS_RSC_DISABLED;
+ }
+
+ return rndis_filter_set_offload_params(ndev, nvdev, &offloads);
+}
+
static u32 netvsc_get_msglevel(struct net_device *ndev)
{
struct net_device_context *ndev_ctx = netdev_priv(ndev);
@@ -1776,6 +1805,7 @@ static const struct net_device_ops device_ops = {
.ndo_start_xmit = netvsc_start_xmit,
.ndo_change_rx_flags = netvsc_change_rx_flags,
.ndo_set_rx_mode = netvsc_set_rx_mode,
+ .ndo_set_features = netvsc_set_features,
.ndo_change_mtu = netvsc_change_mtu,
.ndo_validate_addr = eth_validate_addr,
.ndo_set_mac_address = netvsc_set_mac_addr,
diff --git a/drivers/net/hyperv/rndis_filter.c b/drivers/net/hyperv/rndis_filter.c
index 7bc8ade5c2bd..f94892edaa21 100644
--- a/drivers/net/hyperv/rndis_filter.c
+++ b/drivers/net/hyperv/rndis_filter.c
@@ -716,7 +716,7 @@ int rndis_filter_set_device_mac(struct netvsc_device *nvdev,
return ret;
}
-static int
+int
rndis_filter_set_offload_params(struct net_device *ndev,
struct netvsc_device *nvdev,
struct ndis_offload_params *req_offloads)
@@ -1246,8 +1246,13 @@ static int rndis_netdev_set_hwcaps(struct rndis_device *rndis_device,
if (hwcaps.rsc.ip4 && hwcaps.rsc.ip6) {
net->hw_features |= NETIF_F_LRO;
- offloads.rsc_ip_v4 = NDIS_OFFLOAD_PARAMETERS_RSC_ENABLED;
- offloads.rsc_ip_v6 = NDIS_OFFLOAD_PARAMETERS_RSC_ENABLED;
+ if (net->features & NETIF_F_LRO) {
+ offloads.rsc_ip_v4 = NDIS_OFFLOAD_PARAMETERS_RSC_ENABLED;
+ offloads.rsc_ip_v6 = NDIS_OFFLOAD_PARAMETERS_RSC_ENABLED;
+ } else {
+ offloads.rsc_ip_v4 = NDIS_OFFLOAD_PARAMETERS_RSC_DISABLED;
+ offloads.rsc_ip_v6 = NDIS_OFFLOAD_PARAMETERS_RSC_DISABLED;
+ }
}
/* In case some hw_features disappeared we need to remove them from
--
2.18.0
^ permalink raw reply related
* [PATCH net-next,1/3] hv_netvsc: Add support for LRO/RSC in the vSwitch
From: Haiyang Zhang @ 2018-09-20 17:06 UTC (permalink / raw)
To: davem, netdev
Cc: haiyangz, kys, sthemmin, olaf, vkuznets, devel, linux-kernel
In-Reply-To: <20180920170701.9574-1-haiyangz@linuxonhyperv.com>
From: Haiyang Zhang <haiyangz@microsoft.com>
LRO/RSC in the vSwitch is a feature available in Windows Server 2019
hosts and later. It reduces the per packet processing overhead by
coalescing multiple TCP segments when possible. This patch adds netvsc
driver support for this feature.
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
---
drivers/net/hyperv/hyperv_net.h | 47 +++++++++++++---
drivers/net/hyperv/netvsc.c | 18 +++++--
drivers/net/hyperv/netvsc_drv.c | 28 +++++-----
drivers/net/hyperv/rndis_filter.c | 90 ++++++++++++++++++++++++++-----
4 files changed, 145 insertions(+), 38 deletions(-)
diff --git a/drivers/net/hyperv/hyperv_net.h b/drivers/net/hyperv/hyperv_net.h
index a32ded5b4f41..7f1603dc8128 100644
--- a/drivers/net/hyperv/hyperv_net.h
+++ b/drivers/net/hyperv/hyperv_net.h
@@ -186,6 +186,7 @@ struct rndis_device {
/* Interface */
struct rndis_message;
struct netvsc_device;
+struct netvsc_channel;
struct net_device_context;
extern u32 netvsc_ring_bytes;
@@ -203,10 +204,7 @@ void netvsc_linkstatus_callback(struct net_device *net,
struct rndis_message *resp);
int netvsc_recv_callback(struct net_device *net,
struct netvsc_device *nvdev,
- struct vmbus_channel *channel,
- void *data, u32 len,
- const struct ndis_tcp_ip_checksum_info *csum_info,
- const struct ndis_pkt_8021q_info *vlan);
+ struct netvsc_channel *nvchan);
void netvsc_channel_cb(void *context);
int netvsc_poll(struct napi_struct *napi, int budget);
@@ -222,7 +220,7 @@ int rndis_filter_set_rss_param(struct rndis_device *rdev,
const u8 *key);
int rndis_filter_receive(struct net_device *ndev,
struct netvsc_device *net_dev,
- struct vmbus_channel *channel,
+ struct netvsc_channel *nvchan,
void *data, u32 buflen);
int rndis_filter_set_device_mac(struct netvsc_device *ndev,
@@ -524,6 +522,8 @@ struct nvsp_2_vsc_capability {
u64 ieee8021q:1;
u64 correlation_id:1;
u64 teaming:1;
+ u64 vsubnetid:1;
+ u64 rsc:1;
};
};
} __packed;
@@ -826,7 +826,7 @@ struct nvsp_message {
#define NETVSC_SUPPORTED_HW_FEATURES (NETIF_F_RXCSUM | NETIF_F_IP_CSUM | \
NETIF_F_TSO | NETIF_F_IPV6_CSUM | \
- NETIF_F_TSO6)
+ NETIF_F_TSO6 | NETIF_F_LRO)
#define VRSS_SEND_TAB_SIZE 16 /* must be power of 2 */
#define VRSS_CHANNEL_MAX 64
@@ -852,6 +852,18 @@ struct multi_recv_comp {
u32 next; /* next entry for writing */
};
+#define NVSP_RSC_MAX 562 /* Max #RSC frags in a vmbus xfer page pkt */
+
+struct nvsc_rsc {
+ const struct ndis_pkt_8021q_info *vlan;
+ const struct ndis_tcp_ip_checksum_info *csum_info;
+ u8 is_last; /* last RNDIS msg in a vmtransfer_page */
+ u32 cnt; /* #fragments in an RSC packet */
+ u32 pktlen; /* Full packet length */
+ void *data[NVSP_RSC_MAX];
+ u32 len[NVSP_RSC_MAX];
+};
+
struct netvsc_stats {
u64 packets;
u64 bytes;
@@ -955,6 +967,7 @@ struct netvsc_channel {
struct multi_send_data msd;
struct multi_recv_comp mrc;
atomic_t queue_sends;
+ struct nvsc_rsc rsc;
struct netvsc_stats tx_stats;
struct netvsc_stats rx_stats;
@@ -1136,7 +1149,8 @@ struct rndis_oobd {
/* Packet extension field contents associated with a Data message. */
struct rndis_per_packet_info {
u32 size;
- u32 type;
+ u32 type:31;
+ u32 internal:1;
u32 ppi_offset;
};
@@ -1157,6 +1171,25 @@ enum ndis_per_pkt_info_type {
MAX_PER_PKT_INFO
};
+enum rndis_per_pkt_info_interal_type {
+ RNDIS_PKTINFO_ID = 1,
+ /* Add more memebers here */
+
+ RNDIS_PKTINFO_MAX
+};
+
+#define RNDIS_PKTINFO_SUBALLOC BIT(0)
+#define RNDIS_PKTINFO_1ST_FRAG BIT(1)
+#define RNDIS_PKTINFO_LAST_FRAG BIT(2)
+
+#define RNDIS_PKTINFO_ID_V1 1
+
+struct rndis_pktinfo_id {
+ u8 ver;
+ u8 flag;
+ u16 pkt_id;
+};
+
struct ndis_pkt_8021q_info {
union {
struct {
diff --git a/drivers/net/hyperv/netvsc.c b/drivers/net/hyperv/netvsc.c
index fe01e141c8f8..922054c1d544 100644
--- a/drivers/net/hyperv/netvsc.c
+++ b/drivers/net/hyperv/netvsc.c
@@ -542,6 +542,9 @@ static int negotiate_nvsp_ver(struct hv_device *device,
init_packet->msg.v2_msg.send_ndis_config.capability.teaming = 1;
}
+ if (nvsp_ver >= NVSP_PROTOCOL_VERSION_61)
+ init_packet->msg.v2_msg.send_ndis_config.capability.rsc = 1;
+
trace_nvsp_send(ndev, init_packet);
ret = vmbus_sendpacket(device->channel, init_packet,
@@ -1111,11 +1114,12 @@ static void enq_receive_complete(struct net_device *ndev,
static int netvsc_receive(struct net_device *ndev,
struct netvsc_device *net_device,
- struct vmbus_channel *channel,
+ struct netvsc_channel *nvchan,
const struct vmpacket_descriptor *desc,
const struct nvsp_message *nvsp)
{
struct net_device_context *net_device_ctx = netdev_priv(ndev);
+ struct vmbus_channel *channel = nvchan->channel;
const struct vmtransfer_page_packet_header *vmxferpage_packet
= container_of(desc, const struct vmtransfer_page_packet_header, d);
u16 q_idx = channel->offermsg.offer.sub_channel_index;
@@ -1150,6 +1154,7 @@ static int netvsc_receive(struct net_device *ndev,
int ret;
if (unlikely(offset + buflen > net_device->recv_buf_size)) {
+ nvchan->rsc.cnt = 0;
status = NVSP_STAT_FAIL;
netif_err(net_device_ctx, rx_err, ndev,
"Packet offset:%u + len:%u too big\n",
@@ -1160,11 +1165,13 @@ static int netvsc_receive(struct net_device *ndev,
data = recv_buf + offset;
+ nvchan->rsc.is_last = (i == count - 1);
+
trace_rndis_recv(ndev, q_idx, data);
/* Pass it to the upper layer */
ret = rndis_filter_receive(ndev, net_device,
- channel, data, buflen);
+ nvchan, data, buflen);
if (unlikely(ret != NVSP_STAT_SUCCESS))
status = NVSP_STAT_FAIL;
@@ -1223,12 +1230,13 @@ static void netvsc_receive_inband(struct net_device *ndev,
}
static int netvsc_process_raw_pkt(struct hv_device *device,
- struct vmbus_channel *channel,
+ struct netvsc_channel *nvchan,
struct netvsc_device *net_device,
struct net_device *ndev,
const struct vmpacket_descriptor *desc,
int budget)
{
+ struct vmbus_channel *channel = nvchan->channel;
const struct nvsp_message *nvmsg = hv_pkt_data(desc);
trace_nvsp_recv(ndev, channel, nvmsg);
@@ -1240,7 +1248,7 @@ static int netvsc_process_raw_pkt(struct hv_device *device,
break;
case VM_PKT_DATA_USING_XFER_PAGES:
- return netvsc_receive(ndev, net_device, channel,
+ return netvsc_receive(ndev, net_device, nvchan,
desc, nvmsg);
break;
@@ -1284,7 +1292,7 @@ int netvsc_poll(struct napi_struct *napi, int budget)
nvchan->desc = hv_pkt_iter_first(channel);
while (nvchan->desc && work_done < budget) {
- work_done += netvsc_process_raw_pkt(device, channel, net_device,
+ work_done += netvsc_process_raw_pkt(device, nvchan, net_device,
ndev, nvchan->desc, budget);
nvchan->desc = hv_pkt_iter_next(channel, nvchan->desc);
}
diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c
index 3af6d8d15233..f8c18370d9d1 100644
--- a/drivers/net/hyperv/netvsc_drv.c
+++ b/drivers/net/hyperv/netvsc_drv.c
@@ -744,14 +744,16 @@ void netvsc_linkstatus_callback(struct net_device *net,
}
static struct sk_buff *netvsc_alloc_recv_skb(struct net_device *net,
- struct napi_struct *napi,
- const struct ndis_tcp_ip_checksum_info *csum_info,
- const struct ndis_pkt_8021q_info *vlan,
- void *data, u32 buflen)
+ struct netvsc_channel *nvchan)
{
+ struct napi_struct *napi = &nvchan->napi;
+ const struct ndis_pkt_8021q_info *vlan = nvchan->rsc.vlan;
+ const struct ndis_tcp_ip_checksum_info *csum_info =
+ nvchan->rsc.csum_info;
struct sk_buff *skb;
+ int i;
- skb = napi_alloc_skb(napi, buflen);
+ skb = napi_alloc_skb(napi, nvchan->rsc.pktlen);
if (!skb)
return skb;
@@ -759,7 +761,8 @@ static struct sk_buff *netvsc_alloc_recv_skb(struct net_device *net,
* Copy to skb. This copy is needed here since the memory pointed by
* hv_netvsc_packet cannot be deallocated
*/
- skb_put_data(skb, data, buflen);
+ for (i = 0; i < nvchan->rsc.cnt; i++)
+ skb_put_data(skb, nvchan->rsc.data[i], nvchan->rsc.len[i]);
skb->protocol = eth_type_trans(skb, net);
@@ -792,14 +795,11 @@ static struct sk_buff *netvsc_alloc_recv_skb(struct net_device *net,
*/
int netvsc_recv_callback(struct net_device *net,
struct netvsc_device *net_device,
- struct vmbus_channel *channel,
- void *data, u32 len,
- const struct ndis_tcp_ip_checksum_info *csum_info,
- const struct ndis_pkt_8021q_info *vlan)
+ struct netvsc_channel *nvchan)
{
struct net_device_context *net_device_ctx = netdev_priv(net);
+ struct vmbus_channel *channel = nvchan->channel;
u16 q_idx = channel->offermsg.offer.sub_channel_index;
- struct netvsc_channel *nvchan = &net_device->chan_table[q_idx];
struct sk_buff *skb;
struct netvsc_stats *rx_stats;
@@ -807,8 +807,8 @@ int netvsc_recv_callback(struct net_device *net,
return NVSP_STAT_FAIL;
/* Allocate a skb - TODO direct I/O to pages? */
- skb = netvsc_alloc_recv_skb(net, &nvchan->napi,
- csum_info, vlan, data, len);
+ skb = netvsc_alloc_recv_skb(net, nvchan);
+
if (unlikely(!skb)) {
++net_device_ctx->eth_stats.rx_no_memory;
rcu_read_unlock();
@@ -825,7 +825,7 @@ int netvsc_recv_callback(struct net_device *net,
rx_stats = &nvchan->rx_stats;
u64_stats_update_begin(&rx_stats->syncp);
rx_stats->packets++;
- rx_stats->bytes += len;
+ rx_stats->bytes += nvchan->rsc.pktlen;
if (skb->pkt_type == PACKET_BROADCAST)
++rx_stats->broadcast;
diff --git a/drivers/net/hyperv/rndis_filter.c b/drivers/net/hyperv/rndis_filter.c
index 2a5209f23f29..7bc8ade5c2bd 100644
--- a/drivers/net/hyperv/rndis_filter.c
+++ b/drivers/net/hyperv/rndis_filter.c
@@ -342,7 +342,8 @@ static void rndis_filter_receive_response(struct net_device *ndev,
* Get the Per-Packet-Info with the specified type
* return NULL if not found.
*/
-static inline void *rndis_get_ppi(struct rndis_packet *rpkt, u32 type)
+static inline void *rndis_get_ppi(struct rndis_packet *rpkt,
+ u32 type, u8 internal)
{
struct rndis_per_packet_info *ppi;
int len;
@@ -355,7 +356,7 @@ static inline void *rndis_get_ppi(struct rndis_packet *rpkt, u32 type)
len = rpkt->per_pkt_info_len;
while (len > 0) {
- if (ppi->type == type)
+ if (ppi->type == type && ppi->internal == internal)
return (void *)((ulong)ppi + ppi->ppi_offset);
len -= ppi->size;
ppi = (struct rndis_per_packet_info *)((ulong)ppi + ppi->size);
@@ -364,17 +365,41 @@ static inline void *rndis_get_ppi(struct rndis_packet *rpkt, u32 type)
return NULL;
}
+static inline void rsc_add_data
+ (struct netvsc_channel *nvchan,
+ const struct ndis_pkt_8021q_info *vlan,
+ const struct ndis_tcp_ip_checksum_info *csum_info,
+ void *data, u32 len)
+{
+ u32 cnt = nvchan->rsc.cnt;
+
+ if (cnt) {
+ nvchan->rsc.pktlen += len;
+ } else {
+ nvchan->rsc.vlan = vlan;
+ nvchan->rsc.csum_info = csum_info;
+ nvchan->rsc.pktlen = len;
+ }
+
+ nvchan->rsc.data[cnt] = data;
+ nvchan->rsc.len[cnt] = len;
+ nvchan->rsc.cnt++;
+}
+
static int rndis_filter_receive_data(struct net_device *ndev,
struct netvsc_device *nvdev,
- struct vmbus_channel *channel,
+ struct netvsc_channel *nvchan,
struct rndis_message *msg,
u32 data_buflen)
{
struct rndis_packet *rndis_pkt = &msg->msg.pkt;
const struct ndis_tcp_ip_checksum_info *csum_info;
const struct ndis_pkt_8021q_info *vlan;
+ const struct rndis_pktinfo_id *pktinfo_id;
u32 data_offset;
void *data;
+ bool rsc_more = false;
+ int ret;
/* Remove the rndis header and pass it back up the stack */
data_offset = RNDIS_HEADER_SIZE + rndis_pkt->data_offset;
@@ -393,25 +418,59 @@ static int rndis_filter_receive_data(struct net_device *ndev,
return NVSP_STAT_FAIL;
}
- vlan = rndis_get_ppi(rndis_pkt, IEEE_8021Q_INFO);
+ vlan = rndis_get_ppi(rndis_pkt, IEEE_8021Q_INFO, 0);
+
+ csum_info = rndis_get_ppi(rndis_pkt, TCPIP_CHKSUM_PKTINFO, 0);
- csum_info = rndis_get_ppi(rndis_pkt, TCPIP_CHKSUM_PKTINFO);
+ pktinfo_id = rndis_get_ppi(rndis_pkt, RNDIS_PKTINFO_ID, 1);
data = (void *)msg + data_offset;
- /*
- * Remove the rndis trailer padding from rndis packet message
+ /* Identify RSC frags, drop erroneous packets */
+ if (pktinfo_id && (pktinfo_id->flag & RNDIS_PKTINFO_SUBALLOC)) {
+ if (pktinfo_id->flag & RNDIS_PKTINFO_1ST_FRAG)
+ nvchan->rsc.cnt = 0;
+ else if (nvchan->rsc.cnt == 0)
+ goto drop;
+
+ rsc_more = true;
+
+ if (pktinfo_id->flag & RNDIS_PKTINFO_LAST_FRAG)
+ rsc_more = false;
+
+ if (rsc_more && nvchan->rsc.is_last)
+ goto drop;
+ } else {
+ nvchan->rsc.cnt = 0;
+ }
+
+ if (unlikely(nvchan->rsc.cnt >= NVSP_RSC_MAX))
+ goto drop;
+
+ /* Put data into per channel structure.
+ * Also, remove the rndis trailer padding from rndis packet message
* rndis_pkt->data_len tell us the real data length, we only copy
* the data packet to the stack, without the rndis trailer padding
*/
- return netvsc_recv_callback(ndev, nvdev, channel,
- data, rndis_pkt->data_len,
- csum_info, vlan);
+ rsc_add_data(nvchan, vlan, csum_info, data, rndis_pkt->data_len);
+
+ if (rsc_more)
+ return NVSP_STAT_SUCCESS;
+
+ ret = netvsc_recv_callback(ndev, nvdev, nvchan);
+ nvchan->rsc.cnt = 0;
+
+ return ret;
+
+drop:
+ /* Drop incomplete packet */
+ nvchan->rsc.cnt = 0;
+ return NVSP_STAT_FAIL;
}
int rndis_filter_receive(struct net_device *ndev,
struct netvsc_device *net_dev,
- struct vmbus_channel *channel,
+ struct netvsc_channel *nvchan,
void *data, u32 buflen)
{
struct net_device_context *net_device_ctx = netdev_priv(ndev);
@@ -422,7 +481,7 @@ int rndis_filter_receive(struct net_device *ndev,
switch (rndis_msg->ndis_msg_type) {
case RNDIS_MSG_PACKET:
- return rndis_filter_receive_data(ndev, net_dev, channel,
+ return rndis_filter_receive_data(ndev, net_dev, nvchan,
rndis_msg, buflen);
case RNDIS_MSG_INIT_C:
case RNDIS_MSG_QUERY_C:
@@ -1184,6 +1243,13 @@ static int rndis_netdev_set_hwcaps(struct rndis_device *rndis_device,
}
}
+ if (hwcaps.rsc.ip4 && hwcaps.rsc.ip6) {
+ net->hw_features |= NETIF_F_LRO;
+
+ offloads.rsc_ip_v4 = NDIS_OFFLOAD_PARAMETERS_RSC_ENABLED;
+ offloads.rsc_ip_v6 = NDIS_OFFLOAD_PARAMETERS_RSC_ENABLED;
+ }
+
/* In case some hw_features disappeared we need to remove them from
* net->features list as they're no longer supported.
*/
--
2.18.0
^ permalink raw reply related
* [PATCH net-next,0/3] hv_netvsc: Support LRO/RSC in the vSwitch
From: Haiyang Zhang @ 2018-09-20 17:06 UTC (permalink / raw)
To: davem, netdev
Cc: haiyangz, kys, sthemmin, olaf, vkuznets, devel, linux-kernel
From: Haiyang Zhang <haiyangz@microsoft.com>
The patch adds support for LRO/RSC in the vSwitch feature. It reduces
the per packet processing overhead by coalescing multiple TCP segments
when possible. The feature is enabled by default on VMs running on
Windows Server 2019 and later.
The patch set also adds ethtool command handler and documents.
Haiyang Zhang (3):
hv_netvsc: Add support for LRO/RSC in the vSwitch
hv_netvsc: Add handler for LRO setting change
hv_netvsc: Update document for LRO/RSC support
Documentation/networking/netvsc.txt | 9 +++
drivers/net/hyperv/hyperv_net.h | 51 ++++++++++++---
drivers/net/hyperv/netvsc.c | 18 ++++--
drivers/net/hyperv/netvsc_drv.c | 58 ++++++++++++-----
drivers/net/hyperv/rndis_filter.c | 97 +++++++++++++++++++++++++----
5 files changed, 194 insertions(+), 39 deletions(-)
--
2.18.0
^ permalink raw reply
* [RFC PATCH iproute2-next 3/3] rdma: Add an option to rename IB device interface
From: Leon Romanovsky @ 2018-09-20 11:22 UTC (permalink / raw)
To: David Ahern; +Cc: Leon Romanovsky, netdev, RDMA mailing list, Stephen Hemminger
In-Reply-To: <20180920112232.9602-1-leon@kernel.org>
From: Leon Romanovsky <leonro@mellanox.com>
Enrich rdmatool with an option to rename IB devices,
the command interface follows Iproute2 convention:
"rdma dev set [OLD-DEVNAME] name NEW-DEVNAME"
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
rdma/dev.c | 35 +++++++++++++++++++++++++++++++++++
1 file changed, 35 insertions(+)
diff --git a/rdma/dev.c b/rdma/dev.c
index e2eafe47..760b7fb3 100644
--- a/rdma/dev.c
+++ b/rdma/dev.c
@@ -14,6 +14,7 @@
static int dev_help(struct rd *rd)
{
pr_out("Usage: %s dev show [DEV]\n", rd->filename);
+ pr_out(" %s dev set [DEV] name DEVNAME\n", rd->filename);
return 0;
}
@@ -240,17 +241,51 @@ static int dev_one_show(struct rd *rd)
return rd_exec_cmd(rd, cmds, "parameter");
}
+static int dev_set_name(struct rd *rd)
+{
+ uint32_t seq;
+
+ if (rd_no_arg(rd)) {
+ pr_err("Please provide device new name.\n");
+ return -EINVAL;
+ }
+
+ rd_prepare_msg(rd, RDMA_NLDEV_CMD_SET,
+ &seq, (NLM_F_REQUEST | NLM_F_ACK));
+ mnl_attr_put_u32(rd->nlh, RDMA_NLDEV_ATTR_DEV_INDEX, rd->dev_idx);
+ mnl_attr_put_strz(rd->nlh, RDMA_NLDEV_ATTR_DEV_NAME, rd_argv(rd));
+
+ return rd_send_msg(rd);
+}
+
+static int dev_one_set(struct rd *rd)
+{
+ const struct rd_cmd cmds[] = {
+ { NULL, dev_help},
+ { "name", dev_set_name},
+ { 0 }
+ };
+
+ return rd_exec_cmd(rd, cmds, "parameter");
+}
+
static int dev_show(struct rd *rd)
{
return rd_exec_dev(rd, dev_one_show);
}
+static int dev_set(struct rd *rd)
+{
+ return rd_exec_require_dev(rd, dev_one_set);
+}
+
int cmd_dev(struct rd *rd)
{
const struct rd_cmd cmds[] = {
{ NULL, dev_show },
{ "show", dev_show },
{ "list", dev_show },
+ { "set", dev_set },
{ "help", dev_help },
{ 0 }
};
^ permalink raw reply related
* [RFC PATCH iproute2-next 2/3] rdma: Introduce command execution helper with required device name
From: Leon Romanovsky @ 2018-09-20 11:22 UTC (permalink / raw)
To: David Ahern; +Cc: Leon Romanovsky, netdev, RDMA mailing list, Stephen Hemminger
In-Reply-To: <20180920112232.9602-1-leon@kernel.org>
From: Leon Romanovsky <leonro@mellanox.com>
In contradiction to various show commands, the set command explicitly
requires to use device name as an argument. Provide new command
execution helper which enforces it.
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
rdma/rdma.h | 1 +
rdma/utils.c | 10 ++++++++++
2 files changed, 11 insertions(+)
diff --git a/rdma/rdma.h b/rdma/rdma.h
index d4b7ba19..dde9e128 100644
--- a/rdma/rdma.h
+++ b/rdma/rdma.h
@@ -90,6 +90,7 @@ int cmd_link(struct rd *rd);
int cmd_res(struct rd *rd);
int rd_exec_cmd(struct rd *rd, const struct rd_cmd *c, const char *str);
int rd_exec_dev(struct rd *rd, int (*cb)(struct rd *rd));
+int rd_exec_require_dev(struct rd *rd, int (*cb)(struct rd *rd));
int rd_exec_link(struct rd *rd, int (*cb)(struct rd *rd), bool strict_port);
void rd_free(struct rd *rd);
int rd_set_arg_to_devname(struct rd *rd);
diff --git a/rdma/utils.c b/rdma/utils.c
index 4840bf22..61f4aeb1 100644
--- a/rdma/utils.c
+++ b/rdma/utils.c
@@ -577,6 +577,16 @@ out:
return ret;
}
+int rd_exec_require_dev(struct rd *rd, int (*cb)(struct rd *rd))
+{
+ if (rd_no_arg(rd)) {
+ pr_err("Please provide device name.\n");
+ return -EINVAL;
+ }
+
+ return rd_exec_dev(rd, cb);
+}
+
int rd_exec_cmd(struct rd *rd, const struct rd_cmd *cmds, const char *str)
{
const struct rd_cmd *c;
^ permalink raw reply related
* [RFC PATCH iproute2-next 1/3] rdma: Update kernel include file to support IB device renaming
From: Leon Romanovsky @ 2018-09-20 11:22 UTC (permalink / raw)
To: David Ahern; +Cc: Leon Romanovsky, netdev, RDMA mailing list, Stephen Hemminger
In-Reply-To: <20180920112232.9602-1-leon@kernel.org>
From: Leon Romanovsky <leonro@mellanox.com>
Bring kernel header file changes upto commit 908b44308eda
("RDMA/nldev: Allow IB device rename through RDMA netlink")
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
rdma/include/uapi/rdma/rdma_netlink.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/rdma/include/uapi/rdma/rdma_netlink.h b/rdma/include/uapi/rdma/rdma_netlink.h
index 6513fb89..e2228c09 100644
--- a/rdma/include/uapi/rdma/rdma_netlink.h
+++ b/rdma/include/uapi/rdma/rdma_netlink.h
@@ -227,8 +227,9 @@ enum rdma_nldev_command {
RDMA_NLDEV_CMD_UNSPEC,
RDMA_NLDEV_CMD_GET, /* can dump */
+ RDMA_NLDEV_CMD_SET,
- /* 2 - 4 are free to use */
+ /* 3 - 4 are free to use */
RDMA_NLDEV_CMD_PORT_GET = 5, /* can dump */
^ permalink raw reply related
* [RFC PATCH iproute2-next 0/3] rdma: IB device rename
From: Leon Romanovsky @ 2018-09-20 11:22 UTC (permalink / raw)
To: David Ahern; +Cc: Leon Romanovsky, netdev, RDMA mailing list, Stephen Hemminger
From: Leon Romanovsky <leonro@mellanox.com>
Hi,
This is comprehensive part of kernel series posted earlier. The kernel
part is not accepted yet, so first patch will have different commit
message with different commit SHA1. This is why it is marked as RFC.
An example:
[leonro@server /]$ lspci |grep -i Ether
00:08.0 Ethernet controller: Red Hat, Inc. Virtio network device
00:09.0 Ethernet controller: Mellanox Technologies MT27700 Family [ConnectX-4]
[leonro@server /]$ sudo rdma dev
1: mlx5_0: node_type ca fw 3.8.9999 node_guid 5254:00c0:fe12:3455 sys_image_guid 5254:00c0:fe12:3455
[leonro@server /]$ sudo rdma dev set mlx5_0 name hfi1_0
[leonro@server /]$ sudo rdma dev
1: hfi1_0: node_type ca fw 3.8.9999 node_guid 5254:00c0:fe12:3455 sys_image_guid 5254:00c0:fe12:3455
Thanks
Leon Romanovsky (3):
rdma: Update kernel include file to support IB device renaming
rdma: Introduce command execution helper with required device name
rdma: Add an option to rename IB device interface
rdma/dev.c | 35 +++++++++++++++++++++++++++++++++++
rdma/include/uapi/rdma/rdma_netlink.h | 3 ++-
rdma/rdma.h | 1 +
rdma/utils.c | 10 ++++++++++
4 files changed, 48 insertions(+), 1 deletion(-)
^ permalink raw reply
* [PATCH rdma-next 3/5] RDMA: Convert IB drivers to name allocation routine
From: Leon Romanovsky @ 2018-09-20 11:22 UTC (permalink / raw)
To: Doug Ledford, Jason Gunthorpe
Cc: Leon Romanovsky, RDMA mailing list, linux-s390, Ursula Braun,
David S. Miller, netdev, Selvin Xavier, Steve Wise, Lijun Ou,
Shiraz Saleem, Ariel Elior, Christian Benvenuti, Adit Ranadive,
Dennis Dalessandro
In-Reply-To: <20180920112202.9181-1-leon@kernel.org>
From: Leon Romanovsky <leonro@mellanox.com>
Move internal implementation of device name to the IB/core.
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
drivers/infiniband/core/device.c | 5 -----
drivers/infiniband/hw/bnxt_re/main.c | 6 +++++-
drivers/infiniband/hw/cxgb3/iwch_provider.c | 5 ++++-
drivers/infiniband/hw/cxgb4/provider.c | 5 ++++-
drivers/infiniband/hw/hns/hns_roce_main.c | 4 +++-
drivers/infiniband/hw/i40iw/i40iw_verbs.c | 7 ++++++-
drivers/infiniband/hw/mlx4/main.c | 7 ++++++-
drivers/infiniband/hw/mlx5/main.c | 4 +++-
drivers/infiniband/hw/mthca/mthca_provider.c | 5 ++++-
drivers/infiniband/hw/nes/nes_verbs.c | 6 +++++-
drivers/infiniband/hw/ocrdma/ocrdma_main.c | 7 ++++++-
drivers/infiniband/hw/qedr/main.c | 4 +++-
drivers/infiniband/hw/usnic/usnic_ib_main.c | 5 ++++-
drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c | 9 ++++++---
drivers/infiniband/sw/rxe/rxe_verbs.c | 5 ++++-
15 files changed, 63 insertions(+), 21 deletions(-)
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 3270cde6d806..a9bc2a3f490c 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -518,11 +518,6 @@ int ib_register_device(struct ib_device *device,
mutex_lock(&device_mutex);
- if (strchr(device->name, '%')) {
- ret = alloc_name(device->name);
- if (ret)
- goto out;
-
if (ib_device_check_mandatory(device)) {
ret = -EINVAL;
goto out;
diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index 20b9f31052bf..5255e5ffd6f4 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -575,11 +575,15 @@ static void bnxt_re_unregister_ib(struct bnxt_re_dev *rdev)
static int bnxt_re_register_ib(struct bnxt_re_dev *rdev)
{
struct ib_device *ibdev = &rdev->ibdev;
+ int ret;
/* ib device init */
ibdev->owner = THIS_MODULE;
ibdev->node_type = RDMA_NODE_IB_CA;
- strlcpy(ibdev->name, "bnxt_re%d", IB_DEVICE_NAME_MAX);
+ ret = ib_device_alloc_name(ibdev, "bnxt_re%d");
+ if (ret)
+ return ret;
+
strlcpy(ibdev->node_desc, BNXT_RE_DESC " HCA",
strlen(BNXT_RE_DESC) + 5);
ibdev->phys_port_cnt = 1;
diff --git a/drivers/infiniband/hw/cxgb3/iwch_provider.c b/drivers/infiniband/hw/cxgb3/iwch_provider.c
index 1b9ff21aa1d5..3680d80036eb 100644
--- a/drivers/infiniband/hw/cxgb3/iwch_provider.c
+++ b/drivers/infiniband/hw/cxgb3/iwch_provider.c
@@ -1319,7 +1319,10 @@ int iwch_register_device(struct iwch_dev *dev)
int i;
pr_debug("%s iwch_dev %p\n", __func__, dev);
- strlcpy(dev->ibdev.name, "cxgb3_%d", IB_DEVICE_NAME_MAX);
+ ret = ib_device_alloc_name(&dev->ibdev, "cxgb3_%d");
+ if (ret)
+ return ret;
+
memset(&dev->ibdev.node_guid, 0, sizeof(dev->ibdev.node_guid));
memcpy(&dev->ibdev.node_guid, dev->rdev.t3cdev_p->lldev->dev_addr, 6);
dev->ibdev.owner = THIS_MODULE;
diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c
index 4eda6872e617..37d40cbdf0dc 100644
--- a/drivers/infiniband/hw/cxgb4/provider.c
+++ b/drivers/infiniband/hw/cxgb4/provider.c
@@ -535,7 +535,10 @@ void c4iw_register_device(struct work_struct *work)
struct c4iw_dev *dev = ctx->dev;
pr_debug("c4iw_dev %p\n", dev);
- strlcpy(dev->ibdev.name, "cxgb4_%d", IB_DEVICE_NAME_MAX);
+ ret = ib_device_alloc_name(&dev->ibdev, "cxgb4_%d");
+ if (ret)
+ return;
+
memset(&dev->ibdev.node_guid, 0, sizeof(dev->ibdev.node_guid));
memcpy(&dev->ibdev.node_guid, dev->rdev.lldi.ports[0]->dev_addr, 6);
dev->ibdev.owner = THIS_MODULE;
diff --git a/drivers/infiniband/hw/hns/hns_roce_main.c b/drivers/infiniband/hw/hns/hns_roce_main.c
index 6edb547baee8..e16f72692c83 100644
--- a/drivers/infiniband/hw/hns/hns_roce_main.c
+++ b/drivers/infiniband/hw/hns/hns_roce_main.c
@@ -449,7 +449,9 @@ static int hns_roce_register_device(struct hns_roce_dev *hr_dev)
spin_lock_init(&iboe->lock);
ib_dev = &hr_dev->ib_dev;
- strlcpy(ib_dev->name, "hns_%d", IB_DEVICE_NAME_MAX);
+ ret = ib_device_alloc_name(ib_dev, "hns_%d");
+ if (ret)
+ return ret;
ib_dev->owner = THIS_MODULE;
ib_dev->node_type = RDMA_NODE_IB_CA;
diff --git a/drivers/infiniband/hw/i40iw/i40iw_verbs.c b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
index e2e6c74a7452..829be11da4e7 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_verbs.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_verbs.c
@@ -2746,13 +2746,18 @@ static struct i40iw_ib_device *i40iw_init_rdma_device(struct i40iw_device *iwdev
struct i40iw_ib_device *iwibdev;
struct net_device *netdev = iwdev->netdev;
struct pci_dev *pcidev = (struct pci_dev *)iwdev->hw.dev_context;
+ int ret;
iwibdev = (struct i40iw_ib_device *)ib_alloc_device(sizeof(*iwibdev));
if (!iwibdev) {
i40iw_pr_err("iwdev == NULL\n");
return NULL;
}
- strlcpy(iwibdev->ibdev.name, "i40iw%d", IB_DEVICE_NAME_MAX);
+
+ ret = ib_device_alloc_name(&iwibdev->ibdev, "i40iw%d");
+ if (ret)
+ return NULL;
+
iwibdev->ibdev.owner = THIS_MODULE;
iwdev->iwibdev = iwibdev;
iwibdev->iwdev = iwdev;
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index bf3cdb88aaf5..8edbc50d298a 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -2540,7 +2540,10 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
ibdev->dev = dev;
ibdev->bond_next_port = 0;
- strlcpy(ibdev->ib_dev.name, "mlx4_%d", IB_DEVICE_NAME_MAX);
+ err = ib_device_alloc_name(&ibdev->ib_dev, "mlx4_%d");
+ if (err)
+ goto err_name;
+
ibdev->ib_dev.owner = THIS_MODULE;
ibdev->ib_dev.node_type = RDMA_NODE_IB_CA;
ibdev->ib_dev.local_dma_lkey = dev->caps.reserved_lkey;
@@ -2882,6 +2885,8 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
err_map:
mlx4_ib_free_eqs(dev, ibdev);
+
+err_name:
iounmap(ibdev->uar_map);
err_uar:
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index 73782100750d..a4a578ba72ee 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -5698,8 +5698,10 @@ int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev)
name = "mlx5_%d";
else
name = "mlx5_bond_%d";
+ err = ib_device_alloc_name(&dev->ib_dev, name);
+ if (err)
+ goto err_mp;
- strlcpy(dev->ib_dev.name, name, IB_DEVICE_NAME_MAX);
dev->ib_dev.owner = THIS_MODULE;
dev->ib_dev.node_type = RDMA_NODE_IB_CA;
dev->ib_dev.local_dma_lkey = 0 /* not supported for now */;
diff --git a/drivers/infiniband/hw/mthca/mthca_provider.c b/drivers/infiniband/hw/mthca/mthca_provider.c
index 0d3473b4596e..67ff3ecdf8d5 100644
--- a/drivers/infiniband/hw/mthca/mthca_provider.c
+++ b/drivers/infiniband/hw/mthca/mthca_provider.c
@@ -1198,7 +1198,10 @@ int mthca_register_device(struct mthca_dev *dev)
if (ret)
return ret;
- strlcpy(dev->ib_dev.name, "mthca%d", IB_DEVICE_NAME_MAX);
+ ret = ib_device_alloc_name(&dev->ib_dev, "mthca%d");
+ if (ret)
+ return ret;
+
dev->ib_dev.owner = THIS_MODULE;
dev->ib_dev.uverbs_abi_ver = MTHCA_UVERBS_ABI_VERSION;
diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c
index 6940c7215961..7394d50b02f5 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.c
+++ b/drivers/infiniband/hw/nes/nes_verbs.c
@@ -3635,12 +3635,16 @@ struct nes_ib_device *nes_init_ofa_device(struct net_device *netdev)
struct nes_ib_device *nesibdev;
struct nes_vnic *nesvnic = netdev_priv(netdev);
struct nes_device *nesdev = nesvnic->nesdev;
+ int ret;
nesibdev = (struct nes_ib_device *)ib_alloc_device(sizeof(struct nes_ib_device));
if (nesibdev == NULL) {
return NULL;
}
- strlcpy(nesibdev->ibdev.name, "nes%d", IB_DEVICE_NAME_MAX);
+ ret = ib_device_alloc_name(&nesibdev->ibdev, "nes%d");
+ if (ret)
+ return NULL;
+
nesibdev->ibdev.owner = THIS_MODULE;
nesibdev->ibdev.node_type = RDMA_NODE_RNIC;
diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_main.c b/drivers/infiniband/hw/ocrdma/ocrdma_main.c
index 7832ee3e0c84..34f7d8f31ee1 100644
--- a/drivers/infiniband/hw/ocrdma/ocrdma_main.c
+++ b/drivers/infiniband/hw/ocrdma/ocrdma_main.c
@@ -116,7 +116,12 @@ static void get_dev_fw_str(struct ib_device *device, char *str)
static int ocrdma_register_device(struct ocrdma_dev *dev)
{
- strlcpy(dev->ibdev.name, "ocrdma%d", IB_DEVICE_NAME_MAX);
+ int ret;
+
+ ret = ib_device_alloc_name(&dev->ibdev, "ocrdma%d");
+ if (ret)
+ return ret;
+
ocrdma_get_guid(dev, (u8 *)&dev->ibdev.node_guid);
BUILD_BUG_ON(sizeof(OCRDMA_NODE_DESC) > IB_DEVICE_NODE_DESC_MAX);
memcpy(dev->ibdev.node_desc, OCRDMA_NODE_DESC,
diff --git a/drivers/infiniband/hw/qedr/main.c b/drivers/infiniband/hw/qedr/main.c
index a0af6d424aed..43e14f6b1c6d 100644
--- a/drivers/infiniband/hw/qedr/main.c
+++ b/drivers/infiniband/hw/qedr/main.c
@@ -170,7 +170,9 @@ static int qedr_register_device(struct qedr_dev *dev)
{
int rc;
- strlcpy(dev->ibdev.name, "qedr%d", IB_DEVICE_NAME_MAX);
+ rc = ib_device_alloc_name(&dev->ibdev, "qedr%d");
+ if (rc)
+ return rc;
dev->ibdev.node_guid = dev->attr.node_guid;
memcpy(dev->ibdev.node_desc, QEDR_NODE_DESC, sizeof(QEDR_NODE_DESC));
diff --git a/drivers/infiniband/hw/usnic/usnic_ib_main.c b/drivers/infiniband/hw/usnic/usnic_ib_main.c
index f0538a460328..15f1509cac8d 100644
--- a/drivers/infiniband/hw/usnic/usnic_ib_main.c
+++ b/drivers/infiniband/hw/usnic/usnic_ib_main.c
@@ -335,6 +335,7 @@ static void *usnic_ib_device_add(struct pci_dev *dev)
union ib_gid gid;
struct in_device *ind;
struct net_device *netdev;
+ int ret;
usnic_dbg("\n");
netdev = pci_get_drvdata(dev);
@@ -364,7 +365,9 @@ static void *usnic_ib_device_add(struct pci_dev *dev)
us_ibdev->ib_dev.num_comp_vectors = USNIC_IB_NUM_COMP_VECTORS;
us_ibdev->ib_dev.dev.parent = &dev->dev;
us_ibdev->ib_dev.uverbs_abi_ver = USNIC_UVERBS_ABI_VERSION;
- strlcpy(us_ibdev->ib_dev.name, "usnic_%d", IB_DEVICE_NAME_MAX);
+ ret = ib_device_alloc_name(&us_ibdev->ib_dev, "usnic_%d");
+ if (ret)
+ goto err_fwd_dealloc;
us_ibdev->ib_dev.uverbs_cmd_mask =
(1ull << IB_USER_VERBS_CMD_GET_CONTEXT) |
diff --git a/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c
index a5719899f49a..afe84d55bb52 100644
--- a/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c
+++ b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c
@@ -159,10 +159,13 @@ static struct net_device *pvrdma_get_netdev(struct ib_device *ibdev,
static int pvrdma_register_device(struct pvrdma_dev *dev)
{
- int ret = -1;
int i = 0;
+ int ret;
+
+ ret = ib_device_alloc_name(&dev->ib_dev, "vmw_pvrdma%d");
+ if (ret)
+ return ret;
- strlcpy(dev->ib_dev.name, "vmw_pvrdma%d", IB_DEVICE_NAME_MAX);
dev->ib_dev.node_guid = dev->dsr->caps.node_guid;
dev->sys_image_guid = dev->dsr->caps.sys_image_guid;
dev->flags = 0;
@@ -235,7 +238,7 @@ static int pvrdma_register_device(struct pvrdma_dev *dev)
dev->cq_tbl = kcalloc(dev->dsr->caps.max_cq, sizeof(struct pvrdma_cq *),
GFP_KERNEL);
if (!dev->cq_tbl)
- return ret;
+ return -ENOMEM;
spin_lock_init(&dev->cq_tbl_lock);
dev->qp_tbl = kcalloc(dev->dsr->caps.max_qp, sizeof(struct pvrdma_qp *),
diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.c b/drivers/infiniband/sw/rxe/rxe_verbs.c
index f5b1e0ad6142..2b017bfeecda 100644
--- a/drivers/infiniband/sw/rxe/rxe_verbs.c
+++ b/drivers/infiniband/sw/rxe/rxe_verbs.c
@@ -1159,7 +1159,10 @@ int rxe_register_device(struct rxe_dev *rxe)
struct ib_device *dev = &rxe->ib_dev;
struct crypto_shash *tfm;
- strlcpy(dev->name, "rxe%d", IB_DEVICE_NAME_MAX);
+ err = ib_device_alloc_name(dev, "rxe%d");
+ if (err)
+ return err;
+
strlcpy(dev->node_desc, "rxe", sizeof(dev->node_desc));
dev->owner = THIS_MODULE;
^ permalink raw reply related
* [PATCH rdma-next 5/5] RDMA/nldev: Allow IB device rename through RDMA netlink
From: Leon Romanovsky @ 2018-09-20 11:22 UTC (permalink / raw)
To: Doug Ledford, Jason Gunthorpe
Cc: Leon Romanovsky, RDMA mailing list, linux-s390, Ursula Braun,
David S. Miller, netdev, Selvin Xavier, Steve Wise, Lijun Ou,
Shiraz Saleem, Ariel Elior, Christian Benvenuti, Adit Ranadive,
Dennis Dalessandro
In-Reply-To: <20180920112202.9181-1-leon@kernel.org>
From: Leon Romanovsky <leonro@mellanox.com>
Provide an option to rename IB device name through RDMA netlink and
limit it to users with ADMIN capability only.
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
drivers/infiniband/core/nldev.c | 33 +++++++++++++++++++++++++++++++++
include/uapi/rdma/rdma_netlink.h | 3 ++-
2 files changed, 35 insertions(+), 1 deletion(-)
diff --git a/drivers/infiniband/core/nldev.c b/drivers/infiniband/core/nldev.c
index 0385ab438320..64e1d12e9678 100644
--- a/drivers/infiniband/core/nldev.c
+++ b/drivers/infiniband/core/nldev.c
@@ -645,6 +645,35 @@ static int nldev_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
return err;
}
+static int nldev_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
+ struct netlink_ext_ack *extack)
+{
+ struct nlattr *tb[RDMA_NLDEV_ATTR_MAX];
+ struct ib_device *device;
+ u32 index;
+ int err;
+
+ err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1, nldev_policy,
+ extack);
+ if (err || !tb[RDMA_NLDEV_ATTR_DEV_INDEX])
+ return -EINVAL;
+
+ index = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]);
+ device = ib_device_get_by_index(index);
+ if (!device)
+ return -EINVAL;
+
+ if (tb[RDMA_NLDEV_ATTR_DEV_NAME]) {
+ char name[IB_DEVICE_NAME_MAX] = {};
+
+ nla_strlcpy(name, tb[RDMA_NLDEV_ATTR_DEV_NAME],
+ IB_DEVICE_NAME_MAX);
+ return ib_device_rename(device, name);
+ }
+
+ return 0;
+}
+
static int _nldev_get_dumpit(struct ib_device *device,
struct sk_buff *skb,
struct netlink_callback *cb,
@@ -1077,6 +1106,10 @@ static const struct rdma_nl_cbs nldev_cb_table[RDMA_NLDEV_NUM_OPS] = {
.doit = nldev_get_doit,
.dump = nldev_get_dumpit,
},
+ [RDMA_NLDEV_CMD_SET] = {
+ .doit = nldev_set_doit,
+ .flags = RDMA_NL_ADMIN_PERM,
+ },
[RDMA_NLDEV_CMD_PORT_GET] = {
.doit = nldev_port_get_doit,
.dump = nldev_port_get_dumpit,
diff --git a/include/uapi/rdma/rdma_netlink.h b/include/uapi/rdma/rdma_netlink.h
index edba6351ac13..f9c41bf59efc 100644
--- a/include/uapi/rdma/rdma_netlink.h
+++ b/include/uapi/rdma/rdma_netlink.h
@@ -227,8 +227,9 @@ enum rdma_nldev_command {
RDMA_NLDEV_CMD_UNSPEC,
RDMA_NLDEV_CMD_GET, /* can dump */
+ RDMA_NLDEV_CMD_SET,
- /* 2 - 4 are free to use */
+ /* 3 - 4 are free to use */
RDMA_NLDEV_CMD_PORT_GET = 5, /* can dump */
^ permalink raw reply related
* [PATCH rdma-next 4/5] RDMA/core: Implement IB device rename function
From: Leon Romanovsky @ 2018-09-20 11:22 UTC (permalink / raw)
To: Doug Ledford, Jason Gunthorpe
Cc: Leon Romanovsky, RDMA mailing list, linux-s390, Ursula Braun,
David S. Miller, netdev, Selvin Xavier, Steve Wise, Lijun Ou,
Shiraz Saleem, Ariel Elior, Christian Benvenuti, Adit Ranadive,
Dennis Dalessandro
In-Reply-To: <20180920112202.9181-1-leon@kernel.org>
From: Leon Romanovsky <leonro@mellanox.com>
Generic implementation of IB device rename function.
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
drivers/infiniband/core/core_priv.h | 1 +
drivers/infiniband/core/device.c | 23 +++++++++++++++++++++++
2 files changed, 24 insertions(+)
diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h
index d7399d5b1cb6..c5881756b799 100644
--- a/drivers/infiniband/core/core_priv.h
+++ b/drivers/infiniband/core/core_priv.h
@@ -87,6 +87,7 @@ int ib_device_register_sysfs(struct ib_device *device,
int (*port_callback)(struct ib_device *,
u8, struct kobject *));
void ib_device_unregister_sysfs(struct ib_device *device);
+int ib_device_rename(struct ib_device *ibdev, const char *name);
typedef void (*roce_netdev_callback)(struct ib_device *device, u8 port,
struct net_device *idev, void *cookie);
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index a9bc2a3f490c..36b79cd6b620 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -178,6 +178,29 @@ void ib_device_get_name(struct ib_device *ibdev, char *name)
}
EXPORT_SYMBOL(ib_device_get_name);
+int ib_device_rename(struct ib_device *ibdev, const char *name)
+{
+ struct ib_device *device;
+ int ret = 0;
+
+ if (!strcmp(name, ibdev->name))
+ return ret;
+
+ mutex_lock(&device_mutex);
+ list_for_each_entry(device, &device_list, core_list) {
+ if (!strcmp(name, device->name)) {
+ ret = -EEXIST;
+ goto out;
+ }
+ }
+
+ strlcpy(ibdev->name, name, IB_DEVICE_NAME_MAX);
+ ret = device_rename(&ibdev->dev, name);
+out:
+ mutex_unlock(&device_mutex);
+ return ret;
+}
+
static int alloc_name(char *name)
{
unsigned long *inuse;
^ permalink raw reply related
* [PATCH rdma-next 2/5] net/smc: Use IB device index instead of name
From: Leon Romanovsky @ 2018-09-20 11:21 UTC (permalink / raw)
To: Doug Ledford, Jason Gunthorpe
Cc: Leon Romanovsky, RDMA mailing list, linux-s390, Ursula Braun,
David S. Miller, netdev, Selvin Xavier, Steve Wise, Lijun Ou,
Shiraz Saleem, Ariel Elior, Christian Benvenuti, Adit Ranadive,
Dennis Dalessandro
In-Reply-To: <20180920112202.9181-1-leon@kernel.org>
From: Leon Romanovsky <leonro@mellanox.com>
IB device name is not stable and will be possible to rename in the
following patches, update SMC code to use IB index as a stable
identification.
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
net/smc/smc_diag.c | 6 +++---
net/smc/smc_pnet.c | 27 ++++++++++++++++-----------
2 files changed, 19 insertions(+), 14 deletions(-)
diff --git a/net/smc/smc_diag.c b/net/smc/smc_diag.c
index dbf64a93d68a..ba8b5a5671ec 100644
--- a/net/smc/smc_diag.c
+++ b/net/smc/smc_diag.c
@@ -156,9 +156,9 @@ static int __smc_diag_dump(struct sock *sk, struct sk_buff *skb,
.lnk[0].link_id = smc->conn.lgr->lnk[0].link_id,
};
- memcpy(linfo.lnk[0].ibname,
- smc->conn.lgr->lnk[0].smcibdev->ibdev->name,
- sizeof(smc->conn.lgr->lnk[0].smcibdev->ibdev->name));
+ ib_device_get_name(smc->conn.lgr->lnk[0].smcibdev->ibdev,
+ linfo.lnk[0].ibname);
+
smc_gid_be16_convert(linfo.lnk[0].gid,
smc->conn.lgr->lnk[0].gid);
smc_gid_be16_convert(linfo.lnk[0].peer_gid,
diff --git a/net/smc/smc_pnet.c b/net/smc/smc_pnet.c
index 01c6ce042a1c..67aec5ce6112 100644
--- a/net/smc/smc_pnet.c
+++ b/net/smc/smc_pnet.c
@@ -70,15 +70,14 @@ struct smc_pnetentry {
u8 ib_port;
};
-/* Check if two RDMA device entries are identical. Use device name and port
+/* Check if two RDMA device entries are identical. Use device index and port
* number for comparison.
*/
-static bool smc_pnet_same_ibname(struct smc_pnetentry *pnetelem, char *ibname,
- u8 ibport)
+static bool smc_pnet_same_ibindex(struct smc_pnetentry *pnetelem, u32 ibindex,
+ u8 ibport)
{
return pnetelem->ib_port == ibport &&
- !strncmp(pnetelem->smcibdev->ibdev->name, ibname,
- sizeof(pnetelem->smcibdev->ibdev->name));
+ pnetelem->smcibdev->ibdev->index == ibindex;
}
/* Find a pnetid in the pnet table.
@@ -179,9 +178,9 @@ static int smc_pnet_enter(struct smc_pnetentry *new_pnetelem)
sizeof(new_pnetelem->pnet_name)) ||
!strncmp(pnetelem->ndev->name, new_pnetelem->ndev->name,
sizeof(new_pnetelem->ndev->name)) ||
- smc_pnet_same_ibname(pnetelem,
- new_pnetelem->smcibdev->ibdev->name,
- new_pnetelem->ib_port)) {
+ smc_pnet_same_ibindex(pnetelem,
+ new_pnetelem->smcibdev->ibdev->index,
+ new_pnetelem->ib_port)) {
dev_put(pnetelem->ndev);
goto found;
}
@@ -227,10 +226,11 @@ static struct smc_ib_device *smc_pnet_find_ib(char *ib_name)
spin_lock(&smc_ib_devices.lock);
list_for_each_entry(ibdev, &smc_ib_devices.list, list) {
- if (!strncmp(ibdev->ibdev->name, ib_name,
- sizeof(ibdev->ibdev->name))) {
+ char name[IB_DEVICE_NAME_MAX] = {};
+
+ ib_device_get_name(ibdev->ibdev, name);
+ if (!strncmp(name, ib_name, IB_DEVICE_NAME_MAX))
goto out;
- }
}
ibdev = NULL;
out:
@@ -267,6 +267,11 @@ static int smc_pnet_fill_entry(struct net *net, struct smc_pnetentry *pnetelem,
goto error;
rc = -EINVAL;
+ /* NOTE !!!: Sadly enough, but this is part of ABI.
+ * From day one, the accesses are performed with device names and not
+ * device indexes for both ETH and IB. It means that this function isn't
+ * reliable after device renaming.
+ */
if (!tb[SMC_PNETID_IBNAME])
goto error;
rc = -ENOENT;
^ permalink raw reply related
* [PATCH rdma-next 1/5] RDMA/core: Provide getter and setter to access IB device name
From: Leon Romanovsky @ 2018-09-20 11:21 UTC (permalink / raw)
To: Doug Ledford, Jason Gunthorpe
Cc: Leon Romanovsky, RDMA mailing list, linux-s390, Ursula Braun,
David S. Miller, netdev, Selvin Xavier, Steve Wise, Lijun Ou,
Shiraz Saleem, Ariel Elior, Christian Benvenuti, Adit Ranadive,
Dennis Dalessandro
In-Reply-To: <20180920112202.9181-1-leon@kernel.org>
From: Leon Romanovsky <leonro@mellanox.com>
Prepare IB device name field to rename operation by ensuring that all
accesses to it are protected with lock and users don't see part of name.
The protection is done with global device_lock because it is used in
allocation and deallocation phases. At this stage, this lock is not
busy and easily can be moved to be per-device, once it will be needed.
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
---
drivers/infiniband/core/device.c | 24 +++++++++++++++++++++++-
include/rdma/ib_verbs.h | 8 +++++++-
2 files changed, 30 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 5a680a88aa87..3270cde6d806 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -170,6 +170,14 @@ static struct ib_device *__ib_device_get_by_name(const char *name)
return NULL;
}
+void ib_device_get_name(struct ib_device *ibdev, char *name)
+{
+ down_read(&lists_rwsem);
+ strlcpy(name, ibdev->name, IB_DEVICE_NAME_MAX);
+ up_read(&lists_rwsem);
+}
+EXPORT_SYMBOL(ib_device_get_name);
+
static int alloc_name(char *name)
{
unsigned long *inuse;
@@ -202,6 +210,21 @@ static int alloc_name(char *name)
return 0;
}
+int ib_device_alloc_name(struct ib_device *ibdev, const char *pattern)
+{
+ int ret = 0;
+
+ mutex_lock(&device_mutex);
+ strlcpy(ibdev->name, pattern, IB_DEVICE_NAME_MAX);
+ if (strchr(ibdev->name, '%'))
+ ret = alloc_name(ibdev->name);
+
+ mutex_unlock(&device_mutex);
+
+ return ret;
+}
+EXPORT_SYMBOL(ib_device_alloc_name);
+
static void ib_device_release(struct device *device)
{
struct ib_device *dev = container_of(device, struct ib_device, dev);
@@ -499,7 +522,6 @@ int ib_register_device(struct ib_device *device,
ret = alloc_name(device->name);
if (ret)
goto out;
- }
if (ib_device_check_mandatory(device)) {
ret = -EINVAL;
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index e764ed1f6025..66660e7b9854 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -2260,6 +2260,11 @@ struct ib_device {
/* Do not access @dma_device directly from ULP nor from HW drivers. */
struct device *dma_device;
+ /*
+ * Do not access @name directly,
+ * use ib_device_get_name()/ib_device_alloc_name()
+ * and don't assume that it can't change after access.
+ */
char name[IB_DEVICE_NAME_MAX];
struct list_head event_handler_list;
@@ -2638,7 +2643,8 @@ struct ib_device *ib_alloc_device(size_t size);
void ib_dealloc_device(struct ib_device *device);
void ib_get_device_fw_str(struct ib_device *device, char *str);
-
+int ib_device_alloc_name(struct ib_device *ibdev, const char *pattern);
+void ib_device_get_name(struct ib_device *ibdev, char *name);
int ib_register_device(struct ib_device *device,
int (*port_callback)(struct ib_device *,
u8, struct kobject *));
^ permalink raw reply related
* [PATCH rdma-next 0/5] IB device rename support
From: Leon Romanovsky @ 2018-09-20 11:21 UTC (permalink / raw)
To: Doug Ledford, Jason Gunthorpe
Cc: Leon Romanovsky, RDMA mailing list, linux-s390, Ursula Braun,
David S. Miller, netdev, Selvin Xavier, Steve Wise, Lijun Ou,
Shiraz Saleem, Ariel Elior, Christian Benvenuti, Adit Ranadive,
Dennis Dalessandro
From: Leon Romanovsky <leonro@mellanox.com>
Hi,
This series introduce long-waiting feature - "IB device rename".
Such feature gives and option to rename user visible IB device name from
vendor specific name (e.g. mlx5_0) to anything else.
The user space component through rdmatool will follow this series.
[leonro@server /]$ lspci |grep -i Ether
00:08.0 Ethernet controller: Red Hat, Inc. Virtio network device
00:09.0 Ethernet controller: Mellanox Technologies MT27700 Family [ConnectX-4]
[leonro@server /]$ sudo rdma dev
1: mlx5_0: node_type ca fw 3.8.9999 node_guid 5254:00c0:fe12:3455 sys_image_guid 5254:00c0:fe12:3455
[leonro@server /]$ sudo rdma dev set mlx5_0 name hfi1_0
[leonro@server /]$ sudo rdma dev
1: hfi1_0: node_type ca fw 3.8.9999 node_guid 5254:00c0:fe12:3455 sys_image_guid 5254:00c0:fe12:3455
First patch introduces getter/setter to access names, i didn't convert
all drivers to stop using name directly, because they don't base their
decision on "name", and use this print only and can print truncated name
if renaming is done at the same time as logging.
Second patch updates SMC to use IB device index instead of name.
Third patch globally converts all drivers to new allocation name
routine.
Forth and fifth patches are actually implement and exports through RDMA
netlink the rename routines.
It uses exported by device_rename() function, despite the comment from
2010, which warns about downsides of this function, the netdev is still
uses, so we will use too.
There is one patch/series which was dropped from this submission -
conversion of SElinux from being IB device name to be IB device index
based. It simply needs more special care and more testing.
This series was tested with mlx5 devices with/without traffic and with
non-modified rdma-core.
Dennis,
I didn't touch hfi1, but I'm not sure if it is needed.
Thanks
Leon Romanovsky (5):
RDMA/core: Provide getter and setter to access IB device name
net/smc: Use IB device index instead of name
RDMA: Convert IB drivers to name allocation routine
RDMA/core: Implement IB device rename function
RDMA/nldev: Allow IB device rename through RDMA netlink
drivers/infiniband/core/core_priv.h | 1 +
drivers/infiniband/core/device.c | 52 +++++++++++++++++++++++---
drivers/infiniband/core/nldev.c | 33 ++++++++++++++++
drivers/infiniband/hw/bnxt_re/main.c | 6 ++-
drivers/infiniband/hw/cxgb3/iwch_provider.c | 5 ++-
drivers/infiniband/hw/cxgb4/provider.c | 5 ++-
drivers/infiniband/hw/hns/hns_roce_main.c | 4 +-
drivers/infiniband/hw/i40iw/i40iw_verbs.c | 7 +++-
drivers/infiniband/hw/mlx4/main.c | 7 +++-
drivers/infiniband/hw/mlx5/main.c | 4 +-
drivers/infiniband/hw/mthca/mthca_provider.c | 5 ++-
drivers/infiniband/hw/nes/nes_verbs.c | 6 ++-
drivers/infiniband/hw/ocrdma/ocrdma_main.c | 7 +++-
drivers/infiniband/hw/qedr/main.c | 4 +-
drivers/infiniband/hw/usnic/usnic_ib_main.c | 5 ++-
drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c | 9 +++--
drivers/infiniband/sw/rxe/rxe_verbs.c | 5 ++-
include/rdma/ib_verbs.h | 8 +++-
include/uapi/rdma/rdma_netlink.h | 3 +-
net/smc/smc_diag.c | 6 +--
net/smc/smc_pnet.c | 27 +++++++------
21 files changed, 171 insertions(+), 38 deletions(-)
^ permalink raw reply
* Re: [PATCH net-next 00/22] net: fix return type of ndo_start_xmit function
From: Grygorii Strashko @ 2018-09-20 17:04 UTC (permalink / raw)
To: YueHaibing, davem, dmitry.tarnyagin, wg, mkl, michal.simek,
hsweeten, madalin.bucur, pantelis.antoniou, claudiu.manoil,
leoyang.li, linux, sammy, ralf, nico, steve.glendinning,
f.fainelli, w-kwok2, m-karicheri2, t.sailer, jreuter, kys,
haiyangz, wei.liu2, paul.durrant, arvid.brodin, pshelar
Cc: dev, linux-mips, xen-devel, netdev, linux-usb, linux-kernel,
linux-can, devel, linux-hams, linux-omap, linuxppc-dev,
linux-arm-kernel
In-Reply-To: <20180920123306.14772-1-yuehaibing@huawei.com>
On 09/20/2018 07:32 AM, YueHaibing wrote:
> The method ndo_start_xmit() is defined as returning an 'netdev_tx_t',
> which is a typedef for an enum type, so make sure the implementation in
> this driver has returns 'netdev_tx_t' value, and change the function
> return type to netdev_tx_t.
>
May be I missed smth, but it's acceptable to report standard error codes from
.xmit() callback as per dev_xmit_complete().
--
regards,
-grygorii
^ permalink raw reply
* Re: [PATCH] netfilter: nft_osf: use enum nft_data_types for nft_validate_register_store
From: Pablo Neira Ayuso @ 2018-09-20 16:24 UTC (permalink / raw)
To: Stefan Agner
Cc: ffmancera, kadlec, fw, davem, netfilter-devel, coreteam, netdev,
linux-kernel
In-Reply-To: <20180918052136.25978-1-stefan@agner.ch>
On Mon, Sep 17, 2018 at 10:21:36PM -0700, Stefan Agner wrote:
> The function nft_validate_register_store requires a struct of type
> struct nft_data_types. NFTA_DATA_VALUE is of type enum
> nft_verdict_attributes. Pass the correct enum type.
Applied, thanks.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox