* Re: use-after-free in usbnet
From: Alan Stern @ 2012-03-21 14:35 UTC (permalink / raw)
To: Ming Lei; +Cc: netdev, linux-usb, Fedora Kernel Team, Dave Jones
In-Reply-To: <CACVXFVPK9vj8oqjr1SFr2MgoQn-2XoDqyb95oW+7LCHT7020Ow@mail.gmail.com>
On Wed, 21 Mar 2012, Ming Lei wrote:
> Looks it is a general issue about usb_hcd_unlink_urb.
>
> Alan, how about increasing URB reference count before calling unlink1
> inside usb_hcd_unlink_urb to fix this kind of problem?
No, that won't fix the problem. The URB could complete and be
deallocated even before usb_hcd_unlink_urb() is called, so nothing that
function can do will prevent the error.
It is the caller's responsibility to make sure that the URB does not
get freed before usb_unlink_urb() or usb_kill_urb() returns. We
probably should mention this in the kerneldoc...
Alan Stern
^ permalink raw reply
* Re: [PATCH] net: phy: smsc: energy saving does not work without interrupts
From: Wolfram Sang @ 2012-03-21 14:28 UTC (permalink / raw)
To: netdev; +Cc: David Miller, Giuseppe Cavallaro
In-Reply-To: <1332331707-2350-1-git-send-email-w.sang@pengutronix.de>
[-- Attachment #1: Type: text/plain, Size: 2434 bytes --]
On Wed, Mar 21, 2012 at 01:08:27PM +0100, Wolfram Sang wrote:
> We discovered link detection failures using this PHY with the Freescale FEC
> which does not have interrupt support yet. Disabling the energy saving mode
> makes all go well. It should be possible to make this mode work without
> interrupts by adding some code dealing with the ENERGYON flag. Since adding
> interrupt support to the FEC driver seems to be the better solution for us, I
> want to at least make sure the PHY works reliably always. So, disable energy
> saving if there is no interrupt support.
>
> Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
> Cc: Giuseppe Cavallaro <peppe.cavallaro@st.com>
> ---
>
> Giuseppe: Did you test with interrupts? Do you agree?
>
> If Giuseppe is fine with the patch, this should probably be tagged for stable.
>
> drivers/net/phy/smsc.c | 22 ++++++++++++----------
> 1 files changed, 12 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/net/phy/smsc.c b/drivers/net/phy/smsc.c
> index fc3e7e9..6dcf57a 100644
> --- a/drivers/net/phy/smsc.c
> +++ b/drivers/net/phy/smsc.c
> @@ -43,16 +43,18 @@ static int smsc_phy_ack_interrupt(struct phy_device *phydev)
>
> static int smsc_phy_config_init(struct phy_device *phydev)
> {
> - int rc = phy_read(phydev, MII_LAN83C185_CTRL_STATUS);
> - if (rc < 0)
> - return rc;
> -
> - /* Enable energy detect mode for this SMSC Transceivers */
> - rc = phy_write(phydev, MII_LAN83C185_CTRL_STATUS,
> - rc | MII_LAN83C185_EDPWRDOWN);
> - if (rc < 0)
> - return rc;
> -
> + /* TODO: add code to make EDPWRDOWN work without interrupts */
> + if (PHY_INTERRUPT_ENABLED == phydev->interrupts) {
Gnarg, that should probably be:
+ if (phydev->irq != PHY_POLL) ...
since interrupts maybe not enabled at this point? Will resend, but will
wait some more for comments.
> + int rc = phy_read(phydev, MII_LAN83C185_CTRL_STATUS);
> + if (rc < 0)
> + return rc;
> +
> + /* Enable energy detect mode for this SMSC Transceivers */
> + rc = phy_write(phydev, MII_LAN83C185_CTRL_STATUS,
> + rc | MII_LAN83C185_EDPWRDOWN);
> + if (rc < 0)
> + return rc;
> + }
> return smsc_phy_ack_interrupt (phydev);
> }
>
> --
> 1.7.9.1
>
--
Pengutronix e.K. | Wolfram Sang |
Industrial Linux Solutions | http://www.pengutronix.de/ |
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]
^ permalink raw reply
* Re: [PATCH V1 5/6] net/mlx4_en: sk_prio <=> UP for untagged traffic
From: John Fastabend @ 2012-03-21 14:20 UTC (permalink / raw)
To: Amir Vadai; +Cc: David S. Miller, netdev, Roland Dreier, Oren Duer, Amir Vadai
In-Reply-To: <1332321903-11704-6-git-send-email-amirv@mellanox.com>
On 3/21/2012 2:25 AM, Amir Vadai wrote:
> From: Amir Vadai <amirv@mellanox.co.il>
>
> Since vlan egress map is only good for tagged traffic, need to have other
> mapping to be used by untagged traffic.
> For that, the driver uses sch_mqprio mapping. This mapping could be set by
> using tc tool from iproute2 package.
> Mapped UP will be used by the HW for QoS purposes, but won't go out on the
> wire.
>
> Signed-off-by: Amir Vadai <amirv@mellanox.com>
> ---
[...]
> diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
> index 445a771..f228728 100644
> --- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
> +++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
> @@ -570,15 +570,15 @@ static void build_inline_wqe(struct mlx4_en_tx_desc *tx_desc, struct sk_buff *sk
>
> u16 mlx4_en_select_queue(struct net_device *dev, struct sk_buff *skb)
> {
> - u16 vlan_tag = 0;
> + int up = -1;
>
> - /* If we support per priority flow control and the packet contains
> - * a vlan tag, send the packet to the TX ring assigned to that priority
> - */
> - if (vlan_tx_tag_present(skb)) {
> - vlan_tag = vlan_tx_tag_get(skb);
> - return MLX4_EN_NUM_TX_RINGS + (vlan_tag >> 13);
> - }
> + if (vlan_tx_tag_present(skb))
> + up = (vlan_tx_tag_get(skb) >> 13);
> + else if (dev->num_tc)
> + up = netdev_get_prio_tc_map(dev, skb->priority);
> +
> + if (up >= 0)
> + return MLX4_EN_NUM_TX_RINGS + up;
I expected the else case covered by the netdev_set_tc_queue() setup above? Did
I miss something.
>
> return __skb_tx_hash(dev, skb, MLX4_EN_NUM_TX_RINGS);
> }
^ permalink raw reply
* Re: [PATCH V1 4/6] net/mlx4_en: Set max rate-limit for a TC
From: Amir Vadai @ 2012-03-21 13:43 UTC (permalink / raw)
To: Eric Dumazet
Cc: David S. Miller, netdev, Roland Dreier, Oren Duer, Amir Vadai
In-Reply-To: <1332324776.25742.14.camel@edumazet-laptop>
On 03/21/2012 12:12 PM, Eric Dumazet wrote:
> Le mercredi 21 mars 2012 à 11:25 +0200, Amir Vadai a écrit :
>> From: Amir Vadai<amirv@mellanox.co.il>
>>
>> Set max rate-limit using sysfs file /sys/class/net/<interface>/qos/ratelimit
>>
>> To set, enter a space separated list of values in units of 100Mbps. For
>> example to set ratelimit of 5G to TC0 and 10G for the reset on eth2 issue:
>> echo 50 100 100 100 100 100 100 100 100> /sys/class/net/eth2/qos/ratelimit
>>
>
> Oh thats not very user friendly.
>
> I suggest one value per file (Its probably a generic /sys rule anyway)
> and not using 100Mps for the unit, even if it is the base granularity of
> this hardware.
>
>
> /sys/class/net/eth2/qos/0/ratelimit
> ...
> /sys/class/net/eth2/qos/7/ratelimit
>
>
Ok, will change the structure to be:
/sys/class/net/eth2/qos/tc0/ratelimit
...
/sys/class/net/eth2/qos/tc7/ratelimit
And use units of 1Mbps
- Amir
^ permalink raw reply
* Re: [PATCH V1 6/6] IB/rdma_cm: TOS <=> UP mapping for IBoE
From: Amir Vadai @ 2012-03-21 13:26 UTC (permalink / raw)
To: Eric Dumazet
Cc: David S. Miller, netdev, Roland Dreier, Oren Duer, Amir Vadai,
Sean Hefty
In-Reply-To: <1332324196.25742.8.camel@edumazet-laptop>
On 03/21/2012 12:03 PM, Eric Dumazet wrote:
> Le mercredi 21 mars 2012 à 11:25 +0200, Amir Vadai a écrit :
>> From: Amir Vadai<amirv@mellanox.co.il>
>>
>> Both tagged traffic and untagged traffic use tc tool mapping.
>> Treat RDMA TOS same as IP TOS when mapping to SL
>>
>> Since IP TOS to priority mapping is not exported, had to borrow the code from
>> net/ipv4/route.c
>>
>> Signed-off-by: Amir Vadai<amirv@mellanox.co.il>
>> CC: Sean Hefty<sean.hefty@intel.com>
>>
>> ---
>>
>> We can export IP TOS to priority mapping if the networking maintainers will say
>> where they want it.
>
> Yes please do so. This makes no sense to duplicate ip_tos2prio[]
>
Ok, I will add a patch to export ip_tos2prio, and will use in the code
rt_tos2priority().
- Amir
^ permalink raw reply
* Re: [PATCH] Fix bug of update IPv4 PMTU when received ICMP Fragmentation Needed message
From: Ranjeeth @ 2012-03-21 11:40 UTC (permalink / raw)
To: netdev
In-Reply-To: <4663A561.7070300@cn.fujitsu.com>
Hello,
I have a new query regarding the jumbo frames support in ICMP.
I am supporting jumbo frame support for self tx/rx.
So I need to handle the icmp reply and request for size of 9000.
i was able to ping upto max 3000 size. more than that icmp echo response is not
generated????
why is this???
any limitation?
^ permalink raw reply
* [PATCH] net: phy: smsc: energy saving does not work without interrupts
From: Wolfram Sang @ 2012-03-21 12:08 UTC (permalink / raw)
To: netdev; +Cc: David Miller, Wolfram Sang, Giuseppe Cavallaro
We discovered link detection failures using this PHY with the Freescale FEC
which does not have interrupt support yet. Disabling the energy saving mode
makes all go well. It should be possible to make this mode work without
interrupts by adding some code dealing with the ENERGYON flag. Since adding
interrupt support to the FEC driver seems to be the better solution for us, I
want to at least make sure the PHY works reliably always. So, disable energy
saving if there is no interrupt support.
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Cc: Giuseppe Cavallaro <peppe.cavallaro@st.com>
---
Giuseppe: Did you test with interrupts? Do you agree?
If Giuseppe is fine with the patch, this should probably be tagged for stable.
drivers/net/phy/smsc.c | 22 ++++++++++++----------
1 files changed, 12 insertions(+), 10 deletions(-)
diff --git a/drivers/net/phy/smsc.c b/drivers/net/phy/smsc.c
index fc3e7e9..6dcf57a 100644
--- a/drivers/net/phy/smsc.c
+++ b/drivers/net/phy/smsc.c
@@ -43,16 +43,18 @@ static int smsc_phy_ack_interrupt(struct phy_device *phydev)
static int smsc_phy_config_init(struct phy_device *phydev)
{
- int rc = phy_read(phydev, MII_LAN83C185_CTRL_STATUS);
- if (rc < 0)
- return rc;
-
- /* Enable energy detect mode for this SMSC Transceivers */
- rc = phy_write(phydev, MII_LAN83C185_CTRL_STATUS,
- rc | MII_LAN83C185_EDPWRDOWN);
- if (rc < 0)
- return rc;
-
+ /* TODO: add code to make EDPWRDOWN work without interrupts */
+ if (PHY_INTERRUPT_ENABLED == phydev->interrupts) {
+ int rc = phy_read(phydev, MII_LAN83C185_CTRL_STATUS);
+ if (rc < 0)
+ return rc;
+
+ /* Enable energy detect mode for this SMSC Transceivers */
+ rc = phy_write(phydev, MII_LAN83C185_CTRL_STATUS,
+ rc | MII_LAN83C185_EDPWRDOWN);
+ if (rc < 0)
+ return rc;
+ }
return smsc_phy_ack_interrupt (phydev);
}
--
1.7.9.1
^ permalink raw reply related
* Re: [PATCH 1/1] net: phylib: add error checking if no phy found
From: Dong Aisheng @ 2012-03-21 11:45 UTC (permalink / raw)
To: Florian Fainelli
Cc: Dong Aisheng-B29396, netdev@vger.kernel.org,
linux-kernel@vger.kernel.org,
linux-arm-kernel@lists.infradead.org, davem@davemloft.net,
Tabi Timur-B04825, rdunlap@xenotime.net, kernel@pengutronix.de
In-Reply-To: <4F5F5ABF.5030103@openwrt.org>
On Tue, Mar 13, 2012 at 10:33:35PM +0800, Florian Fainelli wrote:
> Hi,
>
> Le 03/13/12 11:58, Dong Aisheng a écrit :
> > From: Dong Aisheng<dong.aisheng@linaro.org>
> >
> > Signed-off-by: Dong Aisheng<dong.aisheng@linaro.org>
> > ---
> > drivers/net/phy/mdio_bus.c | 11 ++++++++++-
> > 1 files changed, 10 insertions(+), 1 deletions(-)
> >
> > diff --git a/drivers/net/phy/mdio_bus.c b/drivers/net/phy/mdio_bus.c
> > index 88cc5db..d8e044d 100644
> > --- a/drivers/net/phy/mdio_bus.c
> > +++ b/drivers/net/phy/mdio_bus.c
> > @@ -101,6 +101,7 @@ static struct class mdio_bus_class = {
> > int mdiobus_register(struct mii_bus *bus)
> > {
> > int i, err;
> > + u32 phynum = 0;
> >
> > if (NULL == bus || NULL == bus->name ||
> > NULL == bus->read ||
> > @@ -134,12 +135,20 @@ int mdiobus_register(struct mii_bus *bus)
> > if (IS_ERR(phydev)) {
> > err = PTR_ERR(phydev);
> > goto error;
> > + } else if (phydev != NULL) {
> > + phynum++;
> > }
> > }
> > }
> >
> > + if (!phynum) {
> > + err = -ENODEV;
> > + printk(KERN_ERR "%s: no phys found\n", bus->name);
> > + goto error;
> > + }
>
> I am not sure if this is a good idea to change it, especially if:
> - the MII bus probing is also done later (quite some drivers do this)
> - if we want to attach a fixed PHY because we could not probe one at
> registration time
>
You're correct.
It seems keep it as before is the right solution right now.
Thanks for the info.
Regards
Dong Aisheng
^ permalink raw reply
* [REGRESSION][PATCH] Fix an old sky2 WOL regression
From: Knut Petersen @ 2012-03-21 11:40 UTC (permalink / raw)
To: Linus Torvalds
Cc: Andrew Morton, Stephen Hemminger, David S. Miller, arekm, Jared,
dilieto, linux-kernel, netdev
[-- Attachment #1: Type: text/plain, Size: 1455 bytes --]
Sky2 Wake on LAN is broken since February 2010 on a number of systems.
Yes. More than two years.
We know about the problem and the cause since October 2010
(Bugzilla bug #19492). It´s commit 87b09f1f25cd1e01d7c50bf423c7fe33027d7511.
Stephen, David: You signed off that commit.
Andrew: You called it a regression in October 2010.
It has been proposed to revert the commit that caused the problem.
Nothing happened.
I proposed to re-establish the old code for dmi_match()ed systems.
Without success.
Now it is proposed to re-establish the old code as a configuration option.
If nothing happens again I will propose a module parameter ;-)
Stephen, I don´t want to be a pain in the neck, and it is not my intention
to offend you by my "attitude". But I simply cannot understand why this
know regression is not fixed. The bit we talk about is documented,
and in fact it was set for a number of kernel versions unconditionally.
Nobody complained about ruined hardware or minor problems.
The systems affected are old enough that no manufacturer cares about
them, but they are still quite usable for a lot of jobs (kernel 3.3 compile
time here is below 15 minutes).
If there is a problem in the kernel and if we do know an easy solution,
that solution should be commited to the kernel, no matter what is written
in some random documentation, no matter if we could blame some
BIOS authors. That´s the way Linux works - at least I thought so.
cu,
Knut
[-- Attachment #2: 0001-Fix-an-old-sky2-WOL-regression.patch --]
[-- Type: text/x-patch, Size: 2758 bytes --]
>From 7fae2839363277835a0ac1f3f75a788aa299e0a1 Mon Sep 17 00:00:00 2001
From: Knut Petersen <Knut_Petersen@t-online.de>
Date: Wed, 21 Mar 2012 07:42:20 +0100
Subject: [PATCH] Fix an old sky2 WOL regression
Commit 87b09f1f25cd1e01d7c50bf423c7fe33027d7511 removed
code to enable legacy power management mode during wake
on lan initialization. According to the documentation
this was the right thing to do, but it definitely broke
wake on lan on a number of systems.
The regression and the cause of the problem has been
reported more than _18_ months ago, see Bugzilla bug
19492. We should fix it now.
This patch introduces a kernel configuration option that
allows to select inclusion of code to set legacy PM mode
during wake on lan init.
Reported-by: Arkadiusz Miskiewicz <arekm@maven.pl>
Signed-off-by: Knut Petersen <Knut_Petersen@t-online.de>
---
drivers/net/ethernet/marvell/Kconfig | 10 ++++++++++
drivers/net/ethernet/marvell/sky2.c | 11 ++++++++++-
2 files changed, 20 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/marvell/Kconfig b/drivers/net/ethernet/marvell/Kconfig
index 0029934..4f97020 100644
--- a/drivers/net/ethernet/marvell/Kconfig
+++ b/drivers/net/ethernet/marvell/Kconfig
@@ -98,6 +98,16 @@ config SKY2
To compile this driver as a module, choose M here: the module
will be called sky2. This is recommended.
+config SKY2_PME_LEGACY
+ bool "WOL support for broken BIOSes"
+ depends on SKY2 && DEBUG_FS
+ ---help---
+ If this option is set, Yukon 2 legacy power management mode will
+ be enabled during wake on lan setup. Some BIOS implementations
+ need this for actual WOL support.
+
+ If unsure, say N.
+
config SKY2_DEBUG
bool "Debugging interface"
depends on SKY2 && DEBUG_FS
diff --git a/drivers/net/ethernet/marvell/sky2.c b/drivers/net/ethernet/marvell/sky2.c
index 760c2b1..832589c 100644
--- a/drivers/net/ethernet/marvell/sky2.c
+++ b/drivers/net/ethernet/marvell/sky2.c
@@ -814,7 +814,9 @@ static void sky2_wol_init(struct sky2_port *sky2)
unsigned port = sky2->port;
enum flow_control save_mode;
u16 ctrl;
-
+#ifdef CONFIG_SKY2_PME_LEGACY
+ u32 reg1;
+#endif
/* Bring hardware out of reset */
sky2_write16(hw, B0_CTST, CS_RST_CLR);
sky2_write16(hw, SK_REG(port, GMAC_LINK_CTRL), GMLC_RST_CLR);
@@ -867,6 +869,13 @@ static void sky2_wol_init(struct sky2_port *sky2)
/* Disable PiG firmware */
sky2_write16(hw, B0_CTST, Y2_HW_WOL_OFF);
+#ifdef CONFIG_SKY2_PME_LEGACY
+ /* Needed by some broken BIOSes*/
+ reg1 = sky2_pci_read32(hw, PCI_DEV_REG1);
+ reg1 |= PCI_Y2_PME_LEGACY;
+ sky2_pci_write32(hw, PCI_DEV_REG1, reg1);
+#endif
+
/* block receiver */
sky2_write8(hw, SK_REG(port, RX_GMF_CTRL_T), GMF_RST_SET);
sky2_read32(hw, B0_CTST);
--
1.7.9.2
^ permalink raw reply related
* Re: [PATCH 1/1] net: phylib: remove the length limitation of mii bus id
From: Dong Aisheng @ 2012-03-21 11:26 UTC (permalink / raw)
To: Florian Fainelli
Cc: Dong Aisheng, Shawn Guo, Dong Aisheng-B29396,
netdev@vger.kernel.org, davem@davemloft.net,
linux-kernel@vger.kernel.org,
linux-arm-kernel@lists.infradead.org
In-Reply-To: <4F688F01.40604@openwrt.org>
On Tue, Mar 20, 2012 at 10:06:57PM +0800, Florian Fainelli wrote:
> Hi,
>
> Le 03/20/12 11:53, Dong Aisheng a écrit :
> > On Tue, Mar 20, 2012 at 1:32 AM, Shawn Guo<shawn.guo@linaro.org> wrote:
> >> On 20 March 2012 12:23, Dong Aisheng<b29396@freescale.com> wrote:
> >>> From: Dong Aisheng<dong.aisheng@linaro.org>
> >>>
> >>> When convert to dt, the length of old mii bus id (17 bytes) is not
> >>> sufficent to use.
> >>> For example, the bus id could be 800f0000.ethernet-1:00 in DT.
> >>>
> >>> This patch removes the bus id length limitation by changing the
> >>> bus id to a const char pionter and user could dynamically set the
> >>> bus id via kasprintf function call.
> >>>
> >>> Since then no users use MII_BUS_ID_SIZE any more, just remove it.
> >>>
> >>> Signed-off-by: Dong Aisheng<dong.aisheng@linaro.org>
> >>> ---
> >>> The simplest way may just change MII_BUS_ID_SIZE to a more bigger size,
> >>> but i'm not sure that's gonna be accepted.
> >>
> >> The simplest fix has been applied on -next tree as below.
> >>
> > Oh, i missed it, seems my patch covers that change.
> > But that patch only fixed wrong phy_name buffer length issue.
> > My patch is totally remove the length limitation of bus id and phy_name
> > or the bus id name will be truncated if it's longer than MII_BUS_ID_SIZE
> > which is not a comfortable limitation.
>
> Then just increase the bus id size locally for your driver. Drivers that
> can live with a static allocation are better of using it.
>
That's the common part, can not change it locally by driver.
Or we can change MII_BUS_ID_SIZE to a more big size in common code,
maybe 30 bytes, but when is the end?
> Your patch completely missed testing the return value of kasprintf(),
> and its return value can be NULL, please fix this at least.
>
Yes, thanks for the reminder.
I will fix it.
Regards
Dong Aisheng
^ permalink raw reply
* Re: [PATCH] ath9k: fix a memory leak in ath_rx_tasklet()
From: Eric Dumazet @ 2012-03-21 10:21 UTC (permalink / raw)
To: John W. Linville
Cc: David Miller, Jouni Malinen, Felix Fietkau, Trond Wuellner,
Grant Grundler, Paul Stewart, netdev
In-Reply-To: <20120316140010.GB2563@tuxdriver.com>
Le vendredi 16 mars 2012 à 10:00 -0400, John W. Linville a écrit :
> On Thu, Mar 15, 2012 at 01:43:29PM -0700, Eric Dumazet wrote:
> > commit 0d95521ea7 (ath9k: use split rx buffers to get rid of order-1 skb
> > allocations) added in memory leak in error path.
> >
> > sc->rx.frag should be cleared after the pskb_expand_head() call, or else
> > we jump to requeue_drop_frag and leak an skb.
> >
> > Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
> > Cc: Jouni Malinen <jouni@qca.qualcomm.com>
> > Cc: Felix Fietkau <nbd@openwrt.org>
> > Cc: John W. Linville <linville@tuxdriver.com>
> > Cc: Trond Wuellner <trond@chromium.org>
> > Cc: Grant Grundler <grundler@chromium.org>
> > Cc: Paul Stewart <pstew@chromium.org>
> > Cc: David Miller <davem@davemloft.net>
>
> Acked-by: John W. Linville <linville@tuxdriver.com>
>
> Dave, will you pick this up yourself? Or should I take it around the bend?
I dont know, please make sure its not lost anyway :
I have an upcoming patch on top on this one to reduce ath9k skb truesize
by 50% and get better tcp receives.
^ permalink raw reply
* Re: [PATCH V1 4/6] net/mlx4_en: Set max rate-limit for a TC
From: Eric Dumazet @ 2012-03-21 10:12 UTC (permalink / raw)
To: Amir Vadai; +Cc: David S. Miller, netdev, Roland Dreier, Oren Duer, Amir Vadai
In-Reply-To: <1332321903-11704-5-git-send-email-amirv@mellanox.com>
Le mercredi 21 mars 2012 à 11:25 +0200, Amir Vadai a écrit :
> From: Amir Vadai <amirv@mellanox.co.il>
>
> Set max rate-limit using sysfs file /sys/class/net/<interface>/qos/ratelimit
>
> To set, enter a space separated list of values in units of 100Mbps. For
> example to set ratelimit of 5G to TC0 and 10G for the reset on eth2 issue:
> echo 50 100 100 100 100 100 100 100 100 > /sys/class/net/eth2/qos/ratelimit
>
Oh thats not very user friendly.
I suggest one value per file (Its probably a generic /sys rule anyway)
and not using 100Mps for the unit, even if it is the base granularity of
this hardware.
/sys/class/net/eth2/qos/0/ratelimit
...
/sys/class/net/eth2/qos/7/ratelimit
^ permalink raw reply
* Re: [PATCH V1 6/6] IB/rdma_cm: TOS <=> UP mapping for IBoE
From: Eric Dumazet @ 2012-03-21 10:03 UTC (permalink / raw)
To: Amir Vadai
Cc: David S. Miller, netdev, Roland Dreier, Oren Duer, Amir Vadai,
Sean Hefty
In-Reply-To: <1332321903-11704-7-git-send-email-amirv@mellanox.com>
Le mercredi 21 mars 2012 à 11:25 +0200, Amir Vadai a écrit :
> From: Amir Vadai <amirv@mellanox.co.il>
>
> Both tagged traffic and untagged traffic use tc tool mapping.
> Treat RDMA TOS same as IP TOS when mapping to SL
>
> Since IP TOS to priority mapping is not exported, had to borrow the code from
> net/ipv4/route.c
>
> Signed-off-by: Amir Vadai <amirv@mellanox.co.il>
> CC: Sean Hefty <sean.hefty@intel.com>
>
> ---
>
> We can export IP TOS to priority mapping if the networking maintainers will say
> where they want it.
Yes please do so. This makes no sense to duplicate ip_tos2prio[]
^ permalink raw reply
* [PATCH 09/11] net/ariadne: Improve debug prints
From: Geert Uytterhoeven @ 2012-03-21 9:51 UTC (permalink / raw)
To: linux-m68k; +Cc: linux-kernel, Geert Uytterhoeven, David S. Miller, netdev
In-Reply-To: <1332323515-7314-1-git-send-email-geert@linux-m68k.org>
Remove casts and use proper printf()-style format specifiers instead.
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: David S. Miller <davem@davemloft.net>
Cc: netdev@vger.kernel.org
---
drivers/net/ethernet/amd/ariadne.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/amd/ariadne.c b/drivers/net/ethernet/amd/ariadne.c
index eb18e1f..2e7a1c5 100644
--- a/drivers/net/ethernet/amd/ariadne.c
+++ b/drivers/net/ethernet/amd/ariadne.c
@@ -213,10 +213,10 @@ static int ariadne_rx(struct net_device *dev)
(const void *)priv->rx_buff[entry],
pkt_len);
skb->protocol = eth_type_trans(skb, dev);
- netdev_dbg(dev, "RX pkt type 0x%04x from %pM to %pM data 0x%08x len %d\n",
+ netdev_dbg(dev, "RX pkt type 0x%04x from %pM to %pM data %p len %u\n",
((u_short *)skb->data)[6],
skb->data + 6, skb->data,
- (int)skb->data, (int)skb->len);
+ skb->data, skb->len);
netif_rx(skb);
dev->stats.rx_packets++;
@@ -566,10 +566,10 @@ static netdev_tx_t ariadne_start_xmit(struct sk_buff *skb,
/* Fill in a Tx ring entry */
- netdev_dbg(dev, "TX pkt type 0x%04x from %pM to %pM data 0x%08x len %d\n",
+ netdev_dbg(dev, "TX pkt type 0x%04x from %pM to %pM data %p len %u\n",
((u_short *)skb->data)[6],
skb->data + 6, skb->data,
- (int)skb->data, (int)skb->len);
+ skb->data, skb->len);
local_irq_save(flags);
--
1.7.0.4
^ permalink raw reply related
* [PATCH 06/11] m68k/atari: Change VME irq numbers from unsigned long to unsigned int
From: Geert Uytterhoeven @ 2012-03-21 9:51 UTC (permalink / raw)
To: linux-m68k; +Cc: linux-kernel, Geert Uytterhoeven, netdev
In-Reply-To: <1332323515-7314-1-git-send-email-geert@linux-m68k.org>
Device interrupts numbers were changed to unsigned int in 1997, the year
IRQ_MACHSPEC was killed as well.
Also kill a related cast while we're at it.
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: netdev@vger.kernel.org
---
arch/m68k/atari/ataints.c | 4 ++--
arch/m68k/include/asm/atariints.h | 4 ++--
drivers/net/ethernet/amd/atarilance.c | 11 ++++-------
3 files changed, 8 insertions(+), 11 deletions(-)
diff --git a/arch/m68k/atari/ataints.c b/arch/m68k/atari/ataints.c
index 8048e1b..54807d5 100644
--- a/arch/m68k/atari/ataints.c
+++ b/arch/m68k/atari/ataints.c
@@ -207,7 +207,7 @@ void __init atari_init_IRQ(void)
* hardware with a programmable int vector (probably a VME board).
*/
-unsigned long atari_register_vme_int(void)
+unsigned int atari_register_vme_int(void)
{
int i;
@@ -224,7 +224,7 @@ unsigned long atari_register_vme_int(void)
EXPORT_SYMBOL(atari_register_vme_int);
-void atari_unregister_vme_int(unsigned long irq)
+void atari_unregister_vme_int(unsigned int irq)
{
if (irq >= VME_SOURCE_BASE && irq < VME_SOURCE_BASE + VME_MAX_SOURCES) {
irq -= VME_SOURCE_BASE;
diff --git a/arch/m68k/include/asm/atariints.h b/arch/m68k/include/asm/atariints.h
index 656bbbf..5fc13bd 100644
--- a/arch/m68k/include/asm/atariints.h
+++ b/arch/m68k/include/asm/atariints.h
@@ -198,7 +198,7 @@ static inline int atari_irq_pending( unsigned irq )
return( get_mfp_bit( irq, MFP_PENDING ) );
}
-unsigned long atari_register_vme_int( void );
-void atari_unregister_vme_int( unsigned long );
+unsigned int atari_register_vme_int(void);
+void atari_unregister_vme_int(unsigned int);
#endif /* linux/atariints.h */
diff --git a/drivers/net/ethernet/amd/atarilance.c b/drivers/net/ethernet/amd/atarilance.c
index 15bfa28..20a4f6d 100644
--- a/drivers/net/ethernet/amd/atarilance.c
+++ b/drivers/net/ethernet/amd/atarilance.c
@@ -558,21 +558,18 @@ static unsigned long __init lance_probe1( struct net_device *dev,
printk( "Lance: request for irq %d failed\n", IRQ_AUTO_5 );
return 0;
}
- dev->irq = (unsigned short)IRQ_AUTO_5;
+ dev->irq = IRQ_AUTO_5;
}
else {
- /* For VME-RieblCards, request a free VME int;
- * (This must be unsigned long, since dev->irq is short and the
- * IRQ_MACHSPEC bit would be cut off...)
- */
- unsigned long irq = atari_register_vme_int();
+ /* For VME-RieblCards, request a free VME int */
+ unsigned int irq = atari_register_vme_int();
if (!irq) {
printk( "Lance: request for VME interrupt failed\n" );
return 0;
}
if (request_irq(irq, lance_interrupt, IRQ_TYPE_PRIO,
"Riebl-VME Ethernet", dev)) {
- printk( "Lance: request for irq %ld failed\n", irq );
+ printk( "Lance: request for irq %u failed\n", irq );
return 0;
}
dev->irq = irq;
--
1.7.0.4
^ permalink raw reply related
* [PATCH V1 3/6] net/mlx4_en: DCB QoS support
From: Amir Vadai @ 2012-03-21 9:25 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Roland Dreier, Oren Duer, Amir Vadai
In-Reply-To: <1332321903-11704-1-git-send-email-amirv@mellanox.com>
From: Amir Vadai <amirv@mellanox.co.il>
Set TSA, promised BW and PFC using IEEE 802.1qaz netlink commands.
Signed-off-by: Amir Vadai <amirv@mellanox.co.il>
---
drivers/net/ethernet/mellanox/mlx4/Kconfig | 12 ++
drivers/net/ethernet/mellanox/mlx4/Makefile | 1 +
drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c | 203 ++++++++++++++++++++++++
drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 16 ++
drivers/net/ethernet/mellanox/mlx4/mlx4_en.h | 21 +++
5 files changed, 253 insertions(+), 0 deletions(-)
create mode 100644 drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
diff --git a/drivers/net/ethernet/mellanox/mlx4/Kconfig b/drivers/net/ethernet/mellanox/mlx4/Kconfig
index 1bb9353..5f027f9 100644
--- a/drivers/net/ethernet/mellanox/mlx4/Kconfig
+++ b/drivers/net/ethernet/mellanox/mlx4/Kconfig
@@ -11,6 +11,18 @@ config MLX4_EN
This driver supports Mellanox Technologies ConnectX Ethernet
devices.
+config MLX4_EN_DCB
+ bool "Data Center Bridging (DCB) Support"
+ default y
+ depends on MLX4_EN && DCB
+ ---help---
+ Say Y here if you want to use Data Center Bridging (DCB) in the
+ driver.
+ If set to N, will not be able to configure QoS and ratelimit attributes.
+ This flag is depended on the kernel's DCB support.
+
+ If unsure, set to Y
+
config MLX4_CORE
tristate
depends on PCI
diff --git a/drivers/net/ethernet/mellanox/mlx4/Makefile b/drivers/net/ethernet/mellanox/mlx4/Makefile
index 4a40ab9..293127d 100644
--- a/drivers/net/ethernet/mellanox/mlx4/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx4/Makefile
@@ -7,3 +7,4 @@ obj-$(CONFIG_MLX4_EN) += mlx4_en.o
mlx4_en-y := en_main.o en_tx.o en_rx.o en_ethtool.o en_port.o en_cq.o \
en_resources.o en_netdev.o en_selftest.o
+mlx4_en-$(CONFIG_MLX4_EN_DCB) += en_dcb_nl.o
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
new file mode 100644
index 0000000..24ca87c
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
@@ -0,0 +1,203 @@
+/*
+ * Copyright (c) 2011 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/dcbnl.h>
+
+#include "mlx4_en.h"
+
+static int mlx4_en_dcbnl_ieee_getets(struct net_device *dev,
+ struct ieee_ets *ets)
+{
+ struct mlx4_en_priv *priv = netdev_priv(dev);
+ struct ieee_ets *my_ets = priv->mlx4_en_ieee_ets;
+
+ /* No IEEE PFC settings available */
+ if (!my_ets)
+ return -EINVAL;
+
+ ets->ets_cap = IEEE_8021QAZ_MAX_TCS;
+ ets->cbs = my_ets->cbs;
+ memcpy(ets->tc_tx_bw, my_ets->tc_tx_bw, sizeof(ets->tc_tx_bw));
+ memcpy(ets->tc_tsa, my_ets->tc_tsa, sizeof(ets->tc_tsa));
+ memcpy(ets->prio_tc, my_ets->prio_tc, sizeof(ets->prio_tc));
+
+ return 0;
+}
+
+static int mlx4_en_ets_validate(struct mlx4_en_priv *priv, struct ieee_ets *ets)
+{
+ int i;
+ int total_ets_bw = 0;
+ int has_ets_tc = 0;
+
+ for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
+ if (ets->prio_tc[i] > MLX4_EN_NUM_UP) {
+ en_err(priv, "Bad priority in UP <=> TC mapping. "
+ "TC: %d, UP: %d\n", i, ets->prio_tc[i]);
+ return -EINVAL;
+ }
+
+ switch (ets->tc_tsa[i]) {
+ case IEEE_8021QAZ_TSA_STRICT:
+ break;
+ case IEEE_8021QAZ_TSA_ETS:
+ has_ets_tc = 1;
+ total_ets_bw += ets->tc_tx_bw[i];
+ break;
+ default:
+ en_err(priv, "TC[%d]: Not supported TSA: %d\n",
+ i, ets->tc_tsa[i]);
+ return -ENOTSUPP;
+ }
+ }
+
+ if (has_ets_tc && total_ets_bw != MLX4_EN_BW_MAX) {
+ en_err(priv, "Bad ETS BW sum: %d. Should be exactly 100%%\n",
+ total_ets_bw);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static int
+mlx4_en_dcbnl_ieee_setets(struct net_device *dev, struct ieee_ets *ets)
+{
+ struct mlx4_en_priv *priv = netdev_priv(dev);
+ struct mlx4_en_dev *mdev = priv->mdev;
+ int num_strict = 0;
+ int i, err;
+ __u8 tc_tx_bw[IEEE_8021QAZ_MAX_TCS] = { 0 };
+ __u8 pg[IEEE_8021QAZ_MAX_TCS] = { 0 };
+
+ if (ets) {
+ err = mlx4_en_ets_validate(priv, ets);
+ if (err)
+ return err;
+ } else
+ ets = priv->mlx4_en_ieee_ets;
+
+ /* higher TC means higher priority => lower pg */
+ for (i = IEEE_8021QAZ_MAX_TCS - 1; i >= 0; i--) {
+ switch (ets->tc_tsa[i]) {
+ case IEEE_8021QAZ_TSA_STRICT:
+ pg[i] = num_strict++;
+ tc_tx_bw[i] = MLX4_EN_BW_MAX;
+ break;
+ case IEEE_8021QAZ_TSA_ETS:
+ pg[i] = MLX4_EN_TC_ETS;
+ tc_tx_bw[i] = ets->tc_tx_bw[i] ?: MLX4_EN_BW_MIN;
+ break;
+ }
+ }
+
+ err = mlx4_SET_PORT_PRIO2TC(mdev->dev, priv->port, ets->prio_tc);
+ if (err)
+ return err;
+
+ err = mlx4_SET_PORT_SCHEDULER(mdev->dev, priv->port, tc_tx_bw, pg,
+ NULL);
+ if (err)
+ return err;
+
+ if (ets != priv->mlx4_en_ieee_ets)
+ memcpy(priv->mlx4_en_ieee_ets, ets,
+ sizeof(*priv->mlx4_en_ieee_ets));
+
+ return 0;
+}
+
+static int mlx4_en_dcbnl_ieee_getpfc(struct net_device *dev,
+ struct ieee_pfc *pfc)
+{
+ struct mlx4_en_priv *priv = netdev_priv(dev);
+
+ pfc->pfc_cap = IEEE_8021QAZ_MAX_TCS;
+ pfc->pfc_en = priv->prof->tx_ppp;
+
+ return 0;
+}
+
+static int mlx4_en_dcbnl_ieee_setpfc(struct net_device *dev,
+ struct ieee_pfc *pfc)
+{
+ struct mlx4_en_priv *priv = netdev_priv(dev);
+ struct mlx4_en_dev *mdev = priv->mdev;
+ int err;
+
+ en_dbg(DRV, priv, "cap: 0x%x en: 0x%x mbc: 0x%x delay: %d\n",
+ pfc->pfc_cap,
+ pfc->pfc_en,
+ pfc->mbc,
+ pfc->delay);
+
+ priv->prof->rx_pause = priv->prof->tx_pause = !!pfc->pfc_en;
+ priv->prof->rx_ppp = priv->prof->tx_ppp = pfc->pfc_en;
+
+ err = mlx4_SET_PORT_general(mdev->dev, priv->port,
+ priv->rx_skb_size + ETH_FCS_LEN,
+ priv->prof->tx_pause,
+ priv->prof->tx_ppp,
+ priv->prof->rx_pause,
+ priv->prof->rx_ppp);
+ if (err)
+ en_err(priv, "Failed setting pause params\n");
+
+ return err;
+}
+
+static u8 mlx4_en_dcbnl_getdcbx(struct net_device *dev)
+{
+ return DCB_CAP_DCBX_VER_IEEE;
+}
+
+static u8 mlx4_en_dcbnl_setdcbx(struct net_device *dev, u8 mode)
+{
+ if ((mode & DCB_CAP_DCBX_LLD_MANAGED) ||
+ (mode & DCB_CAP_DCBX_VER_CEE) ||
+ !(mode & DCB_CAP_DCBX_VER_IEEE) ||
+ !(mode & DCB_CAP_DCBX_HOST))
+ return 1;
+
+ return 0;
+}
+
+const struct dcbnl_rtnl_ops mlx4_en_dcbnl_ops = {
+ .ieee_getets = mlx4_en_dcbnl_ieee_getets,
+ .ieee_setets = mlx4_en_dcbnl_ieee_setets,
+ .ieee_getpfc = mlx4_en_dcbnl_ieee_getpfc,
+ .ieee_setpfc = mlx4_en_dcbnl_ieee_setpfc,
+
+ .getdcbx = mlx4_en_dcbnl_getdcbx,
+ .setdcbx = mlx4_en_dcbnl_setdcbx,
+};
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 2322622..9b456ae 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -967,6 +967,11 @@ void mlx4_en_destroy_netdev(struct net_device *dev)
mutex_unlock(&mdev->state_lock);
mlx4_en_free_resources(priv);
+
+#ifdef CONFIG_MLX4_EN_DCB
+ vfree(priv->mlx4_en_ieee_ets);
+#endif
+
free_netdev(dev);
}
@@ -1080,6 +1085,17 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
INIT_WORK(&priv->watchdog_task, mlx4_en_restart);
INIT_WORK(&priv->linkstate_task, mlx4_en_linkstate);
INIT_DELAYED_WORK(&priv->stats_task, mlx4_en_do_get_stats);
+#ifdef CONFIG_MLX4_EN_DCB
+ if (!mlx4_is_slave(priv->mdev->dev)) {
+ dev->dcbnl_ops = &mlx4_en_dcbnl_ops;
+
+ priv->mlx4_en_ieee_ets = vzalloc(sizeof(struct ieee_ets));
+ if (!priv->mlx4_en_ieee_ets) {
+ err = -ENOMEM;
+ goto out;
+ }
+ }
+#endif
/* Query for default mac and max mtu */
priv->max_mtu = mdev->dev->caps.eth_mtu_cap[priv->port];
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index 5bd7c2a..fa09792 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -40,6 +40,9 @@
#include <linux/mutex.h>
#include <linux/netdevice.h>
#include <linux/if_vlan.h>
+#ifdef CONFIG_MLX4_EN_DCB
+#include <linux/dcbnl.h>
+#endif
#include <linux/mlx4/device.h>
#include <linux/mlx4/qp.h>
@@ -110,6 +113,7 @@ enum {
#define MLX4_EN_NUM_TX_RINGS 8
#define MLX4_EN_NUM_PPP_RINGS 8
#define MAX_TX_RINGS (MLX4_EN_NUM_TX_RINGS + MLX4_EN_NUM_PPP_RINGS)
+#define MLX4_EN_NUM_UP 8
#define MLX4_EN_DEF_TX_RING_SIZE 512
#define MLX4_EN_DEF_RX_RING_SIZE 1024
@@ -410,6 +414,15 @@ struct mlx4_en_frag_info {
};
+#ifdef CONFIG_MLX4_EN_DCB
+/* Minimal TC BW - setting to 0 will block traffic */
+#define MLX4_EN_BW_MIN 1
+#define MLX4_EN_BW_MAX 100 /* Utilize 100% of the line */
+
+#define MLX4_EN_TC_ETS 7
+
+#endif
+
struct mlx4_en_priv {
struct mlx4_en_dev *mdev;
struct mlx4_en_port_profile *prof;
@@ -483,6 +496,10 @@ struct mlx4_en_priv {
int vids[128];
bool wol;
struct device *ddev;
+
+#ifdef CONFIG_MLX4_EN_DCB
+ struct ieee_ets *mlx4_en_ieee_ets;
+#endif
};
enum mlx4_en_wol {
@@ -557,6 +574,10 @@ int mlx4_SET_VLAN_FLTR(struct mlx4_dev *dev, struct mlx4_en_priv *priv);
int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset);
int mlx4_en_QUERY_PORT(struct mlx4_en_dev *mdev, u8 port);
+#ifdef CONFIG_MLX4_EN_DCB
+extern const struct dcbnl_rtnl_ops mlx4_en_dcbnl_ops;
+#endif
+
#define MLX4_EN_NUM_SELF_TEST 5
void mlx4_en_ex_selftest(struct net_device *dev, u32 *flags, u64 *buf);
u64 mlx4_en_mac_to_u64(u8 *addr);
--
1.7.8.2
^ permalink raw reply related
* [PATCH V1 6/6] IB/rdma_cm: TOS <=> UP mapping for IBoE
From: Amir Vadai @ 2012-03-21 9:25 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Roland Dreier, Oren Duer, Amir Vadai, Sean Hefty
In-Reply-To: <1332321903-11704-1-git-send-email-amirv@mellanox.com>
From: Amir Vadai <amirv@mellanox.co.il>
Both tagged traffic and untagged traffic use tc tool mapping.
Treat RDMA TOS same as IP TOS when mapping to SL
Since IP TOS to priority mapping is not exported, had to borrow the code from
net/ipv4/route.c
Signed-off-by: Amir Vadai <amirv@mellanox.co.il>
CC: Sean Hefty <sean.hefty@intel.com>
---
We can export IP TOS to priority mapping if the networking maintainers will say
where they want it.
---
drivers/infiniband/core/cma.c | 35 ++++++++++++++++++++++++++++++++++-
1 files changed, 34 insertions(+), 1 deletions(-)
diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index e3e470f..c0eeb2c 100644
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -42,6 +42,7 @@
#include <linux/inetdevice.h>
#include <linux/slab.h>
#include <linux/module.h>
+#include <linux/pkt_sched.h>
#include <net/tcp.h>
#include <net/ipv6.h>
@@ -1781,6 +1782,32 @@ static int cma_resolve_iw_route(struct rdma_id_private *id_priv, int timeout_ms)
return 0;
}
+#define ECN_OR_COST(class) TC_PRIO_##class
+
+static const __u8 tos2prio[16] = {
+ TC_PRIO_BESTEFFORT,
+ ECN_OR_COST(BESTEFFORT),
+ TC_PRIO_BESTEFFORT,
+ ECN_OR_COST(BESTEFFORT),
+ TC_PRIO_BULK,
+ ECN_OR_COST(BULK),
+ TC_PRIO_BULK,
+ ECN_OR_COST(BULK),
+ TC_PRIO_INTERACTIVE,
+ ECN_OR_COST(INTERACTIVE),
+ TC_PRIO_INTERACTIVE,
+ ECN_OR_COST(INTERACTIVE),
+ TC_PRIO_INTERACTIVE_BULK,
+ ECN_OR_COST(INTERACTIVE_BULK),
+ TC_PRIO_INTERACTIVE_BULK,
+ ECN_OR_COST(INTERACTIVE_BULK)
+};
+
+static inline char cma_tos2priority(u8 tos)
+{
+ return tos2prio[IPTOS_TOS(tos)>>1];
+}
+
static int cma_resolve_iboe_route(struct rdma_id_private *id_priv)
{
struct rdma_route *route = &id_priv->id.route;
@@ -1826,7 +1853,13 @@ static int cma_resolve_iboe_route(struct rdma_id_private *id_priv)
route->path_rec->reversible = 1;
route->path_rec->pkey = cpu_to_be16(0xffff);
route->path_rec->mtu_selector = IB_SA_EQ;
- route->path_rec->sl = id_priv->tos >> 5;
+ if (ndev->priv_flags & IFF_802_1Q_VLAN)
+ route->path_rec->sl =
+ netdev_get_prio_tc_map(vlan_dev_real_dev(ndev),
+ cma_tos2priority(id_priv->tos));
+ else
+ route->path_rec->sl = netdev_get_prio_tc_map(ndev,
+ cma_tos2priority(id_priv->tos));
route->path_rec->mtu = iboe_get_mtu(ndev->mtu);
route->path_rec->rate_selector = IB_SA_EQ;
--
1.7.8.2
^ permalink raw reply related
* [PATCH V1 4/6] net/mlx4_en: Set max rate-limit for a TC
From: Amir Vadai @ 2012-03-21 9:25 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Roland Dreier, Oren Duer, Amir Vadai
In-Reply-To: <1332321903-11704-1-git-send-email-amirv@mellanox.com>
From: Amir Vadai <amirv@mellanox.co.il>
Set max rate-limit using sysfs file /sys/class/net/<interface>/qos/ratelimit
To set, enter a space separated list of values in units of 100Mbps. For
example to set ratelimit of 5G to TC0 and 10G for the reset on eth2 issue:
echo 50 100 100 100 100 100 100 100 100 > /sys/class/net/eth2/qos/ratelimit
Signed-off-by: Amir Vadai <amirv@mellanox.co.il>
---
We used sysfs since max bw isn't part of the ETS / DCBX NL support, and we're
open to other suggestions to add generic support for max bw, e.g add call to
the DCBX NL API.
drivers/net/ethernet/mellanox/mlx4/Makefile | 2 +-
drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c | 14 +++-
drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 13 +++
drivers/net/ethernet/mellanox/mlx4/en_sysfs.c | 120 ++++++++++++++++++++++++
drivers/net/ethernet/mellanox/mlx4/mlx4_en.h | 5 +
5 files changed, 152 insertions(+), 2 deletions(-)
create mode 100644 drivers/net/ethernet/mellanox/mlx4/en_sysfs.c
diff --git a/drivers/net/ethernet/mellanox/mlx4/Makefile b/drivers/net/ethernet/mellanox/mlx4/Makefile
index 293127d..8e5bd88 100644
--- a/drivers/net/ethernet/mellanox/mlx4/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx4/Makefile
@@ -7,4 +7,4 @@ obj-$(CONFIG_MLX4_EN) += mlx4_en.o
mlx4_en-y := en_main.o en_tx.o en_rx.o en_ethtool.o en_port.o en_cq.o \
en_resources.o en_netdev.o en_selftest.o
-mlx4_en-$(CONFIG_MLX4_EN_DCB) += en_dcb_nl.o
+mlx4_en-$(CONFIG_MLX4_EN_DCB) += en_dcb_nl.o en_sysfs.o
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
index 24ca87c..eb0964f 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
@@ -126,7 +126,7 @@ mlx4_en_dcbnl_ieee_setets(struct net_device *dev, struct ieee_ets *ets)
return err;
err = mlx4_SET_PORT_SCHEDULER(mdev->dev, priv->port, tc_tx_bw, pg,
- NULL);
+ priv->ratelimit);
if (err)
return err;
@@ -192,6 +192,18 @@ static u8 mlx4_en_dcbnl_setdcbx(struct net_device *dev, u8 mode)
return 0;
}
+int mlx4_en_dcbnl_set_ratelimit(struct mlx4_en_priv *priv, u16 *ratelimit)
+{
+ memcpy(priv->ratelimit, ratelimit, sizeof(priv->ratelimit));
+
+ if (mlx4_en_dcbnl_ieee_setets(priv->dev, NULL)) {
+ en_err(priv, "Error setting ets in HW\n");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
const struct dcbnl_rtnl_ops mlx4_en_dcbnl_ops = {
.ieee_getets = mlx4_en_dcbnl_ieee_getets,
.ieee_setets = mlx4_en_dcbnl_ieee_setets,
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 9b456ae..f45d544 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -810,6 +810,18 @@ static void mlx4_en_restart(struct work_struct *work)
mutex_unlock(&mdev->state_lock);
}
+static int mlx4_en_init(struct net_device *dev)
+{
+#ifdef CONFIG_MLX4_EN_DCB
+ struct mlx4_en_priv *priv = netdev_priv(dev);
+
+ if (!mlx4_is_slave(priv->mdev->dev))
+ mlx4_en_prepare_sysfs_group(priv);
+#endif
+
+ return 0;
+}
+
static void mlx4_en_clear_stats(struct net_device *dev)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
@@ -1026,6 +1038,7 @@ static int mlx4_en_set_features(struct net_device *netdev,
}
static const struct net_device_ops mlx4_netdev_ops = {
+ .ndo_init = mlx4_en_init,
.ndo_open = mlx4_en_open,
.ndo_stop = mlx4_en_close,
.ndo_start_xmit = mlx4_en_xmit,
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_sysfs.c b/drivers/net/ethernet/mellanox/mlx4/en_sysfs.c
new file mode 100644
index 0000000..fdada20
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx4/en_sysfs.c
@@ -0,0 +1,120 @@
+/*
+ * Copyright (c) 2011 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/device.h>
+#include <linux/netdevice.h>
+
+#include "mlx4_en.h"
+
+#define to_en_priv(cd) ((struct mlx4_en_priv *)(netdev_priv(to_net_dev(cd))))
+
+static ssize_t mlx4_en_show_ratelimit(struct device *d,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct mlx4_en_priv *priv = to_en_priv(d);
+ int i;
+ int len = 0;
+
+ for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++)
+ len += sprintf(buf + len, "%d ", priv->ratelimit[i]);
+ len += sprintf(buf + len, "\n");
+
+ return len;
+}
+
+static ssize_t mlx4_en_store_ratelimit(struct device *d,
+ struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ int ret = count;
+ struct mlx4_en_priv *priv = to_en_priv(d);
+ char save;
+ int i = 0;
+ u16 ratelimit[IEEE_8021QAZ_MAX_TCS] = { 0 };
+
+ do {
+ int len;
+ int new_value;
+
+ if (i >= IEEE_8021QAZ_MAX_TCS)
+ goto bad_elem_count;
+
+ len = strcspn(buf, " ");
+
+ /* nul-terminate and parse */
+ save = buf[len];
+ ((char *)buf)[len] = '\0';
+
+ if (sscanf(buf, "%d", &new_value) != 1 ||
+ new_value > 100 || new_value < 0) {
+ en_err(priv, "bad ratelimit value: '%s'\n", buf);
+ ret = -EINVAL;
+ goto out;
+ }
+ ratelimit[i] = new_value;
+
+ buf += len+1;
+ i++;
+ } while (save == ' ');
+
+ if (i != IEEE_8021QAZ_MAX_TCS)
+ goto bad_elem_count;
+
+ mlx4_en_dcbnl_set_ratelimit(priv, &ratelimit[0]);
+
+out:
+ return ret;
+
+bad_elem_count:
+ en_err(priv, "bad number of elemets in ratelimit array\n");
+ return -EINVAL;
+}
+
+static DEVICE_ATTR(ratelimit, S_IRUGO | S_IWUSR,
+ mlx4_en_show_ratelimit, mlx4_en_store_ratelimit);
+
+static struct attribute *mlx4_en_qos_attrs[] = {
+ &dev_attr_ratelimit.attr,
+ NULL,
+};
+
+static struct attribute_group qos_group = {
+ .name = "qos",
+ .attrs = mlx4_en_qos_attrs,
+};
+
+void mlx4_en_prepare_sysfs_group(struct mlx4_en_priv *priv)
+{
+ priv->dev->sysfs_groups[0] = &qos_group;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index fa09792..32b447d 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -499,6 +499,7 @@ struct mlx4_en_priv {
#ifdef CONFIG_MLX4_EN_DCB
struct ieee_ets *mlx4_en_ieee_ets;
+ u16 ratelimit[IEEE_8021QAZ_MAX_TCS];
#endif
};
@@ -576,6 +577,10 @@ int mlx4_en_QUERY_PORT(struct mlx4_en_dev *mdev, u8 port);
#ifdef CONFIG_MLX4_EN_DCB
extern const struct dcbnl_rtnl_ops mlx4_en_dcbnl_ops;
+
+void mlx4_en_prepare_sysfs_group(struct mlx4_en_priv *priv);
+
+int mlx4_en_dcbnl_set_ratelimit(struct mlx4_en_priv *priv, u16 *ratelimit);
#endif
#define MLX4_EN_NUM_SELF_TEST 5
--
1.7.8.2
^ permalink raw reply related
* [PATCH V1 5/6] net/mlx4_en: sk_prio <=> UP for untagged traffic
From: Amir Vadai @ 2012-03-21 9:25 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Roland Dreier, Oren Duer, Amir Vadai, Amir Vadai
In-Reply-To: <1332321903-11704-1-git-send-email-amirv@mellanox.com>
From: Amir Vadai <amirv@mellanox.co.il>
Since vlan egress map is only good for tagged traffic, need to have other
mapping to be used by untagged traffic.
For that, the driver uses sch_mqprio mapping. This mapping could be set by
using tc tool from iproute2 package.
Mapped UP will be used by the HW for QoS purposes, but won't go out on the
wire.
Signed-off-by: Amir Vadai <amirv@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 15 +++++++++++++++
drivers/net/ethernet/mellanox/mlx4/en_tx.c | 16 ++++++++--------
2 files changed, 23 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index f45d544..22c4da6 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -45,6 +45,14 @@
#include "mlx4_en.h"
#include "en_port.h"
+static int mlx4_en_setup_tc(struct net_device *dev, u8 up)
+{
+ if (up != MLX4_EN_NUM_UP)
+ return -EINVAL;
+
+ return 0;
+}
+
static int mlx4_en_vlan_rx_add_vid(struct net_device *dev, unsigned short vid)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
@@ -1055,6 +1063,7 @@ static const struct net_device_ops mlx4_netdev_ops = {
.ndo_poll_controller = mlx4_en_netpoll,
#endif
.ndo_set_features = mlx4_en_set_features,
+ .ndo_setup_tc = mlx4_en_setup_tc,
};
int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
@@ -1143,6 +1152,12 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
netif_set_real_num_tx_queues(dev, priv->tx_ring_num);
netif_set_real_num_rx_queues(dev, priv->rx_ring_num);
+ netdev_set_num_tc(dev, MLX4_EN_NUM_UP);
+
+ /* Partition Tx queues evenly amongst UP's */
+ for (i = 0; i < MLX4_EN_NUM_UP; i++)
+ netdev_set_tc_queue(dev, i, 1, MLX4_EN_NUM_TX_RINGS + i);
+
SET_ETHTOOL_OPS(dev, &mlx4_en_ethtool_ops);
/* Set defualt MAC */
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index 445a771..f228728 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -570,15 +570,15 @@ static void build_inline_wqe(struct mlx4_en_tx_desc *tx_desc, struct sk_buff *sk
u16 mlx4_en_select_queue(struct net_device *dev, struct sk_buff *skb)
{
- u16 vlan_tag = 0;
+ int up = -1;
- /* If we support per priority flow control and the packet contains
- * a vlan tag, send the packet to the TX ring assigned to that priority
- */
- if (vlan_tx_tag_present(skb)) {
- vlan_tag = vlan_tx_tag_get(skb);
- return MLX4_EN_NUM_TX_RINGS + (vlan_tag >> 13);
- }
+ if (vlan_tx_tag_present(skb))
+ up = (vlan_tx_tag_get(skb) >> 13);
+ else if (dev->num_tc)
+ up = netdev_get_prio_tc_map(dev, skb->priority);
+
+ if (up >= 0)
+ return MLX4_EN_NUM_TX_RINGS + up;
return __skb_tx_hash(dev, skb, MLX4_EN_NUM_TX_RINGS);
}
--
1.7.8.2
^ permalink raw reply related
* [PATCH V1 2/6] net/mlx4_core: set port QoS attributes
From: Amir Vadai @ 2012-03-21 9:24 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Roland Dreier, Oren Duer, Amir Vadai
In-Reply-To: <1332321903-11704-1-git-send-email-amirv@mellanox.com>
From: Amir Vadai <amirv@mellanox.co.il>
Adding QoS firmware commands:
- mlx4_en_SET_PORT_PRIO2TC - set UP <=> TC
- mlx4_en_SET_PORT_SCHEDULER - set promised BW, max BW and PG number
Signed-off-by: Amir Vadai <amirv@mellanox.co.il>
---
drivers/net/ethernet/mellanox/mlx4/en_port.h | 2 +
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 20 ++++++++
drivers/net/ethernet/mellanox/mlx4/port.c | 62 ++++++++++++++++++++++++++
include/linux/mlx4/cmd.h | 4 ++
include/linux/mlx4/device.h | 3 +
5 files changed, 91 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_port.h b/drivers/net/ethernet/mellanox/mlx4/en_port.h
index 6934fd7..745090b 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_port.h
+++ b/drivers/net/ethernet/mellanox/mlx4/en_port.h
@@ -39,6 +39,8 @@
#define SET_PORT_PROMISC_SHIFT 31
#define SET_PORT_MC_PROMISC_SHIFT 30
+#define MLX4_EN_NUM_TC 8
+
#define VLAN_FLTR_SIZE 128
struct mlx4_set_vlan_fltr_mbox {
__be32 entry[VLAN_FLTR_SIZE];
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index 5da51b9..33bf39d 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -53,6 +53,26 @@
#define DRV_VERSION "1.1"
#define DRV_RELDATE "Dec, 2011"
+#define MLX4_NUM_UP 8
+#define MLX4_NUM_TC 8
+#define MLX4_RATELIMIT_UNITS 3 /* 100 Mbps */
+#define MLX4_RATELIMIT_DEFAULT 0xffff
+
+struct mlx4_set_port_prio2tc_context {
+ u8 prio2tc[4];
+};
+
+struct mlx4_port_scheduler_tc_cfg_be {
+ __be16 pg;
+ __be16 bw_precentage;
+ __be16 max_bw_units; /* 3-100Mbps, 4-1Gbps, other values - reserved */
+ __be16 max_bw_value;
+};
+
+struct mlx4_set_port_scheduler_context {
+ struct mlx4_port_scheduler_tc_cfg_be tc[MLX4_NUM_TC];
+};
+
enum {
MLX4_HCR_BASE = 0x80680,
MLX4_HCR_SIZE = 0x0001c,
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
index 98e7762..762bbda 100644
--- a/drivers/net/ethernet/mellanox/mlx4/port.c
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -858,6 +858,68 @@ int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
}
EXPORT_SYMBOL(mlx4_SET_PORT_qpn_calc);
+int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc)
+{
+ struct mlx4_cmd_mailbox *mailbox;
+ struct mlx4_set_port_prio2tc_context *context;
+ int err;
+ u32 in_mod;
+ int i;
+
+ mailbox = mlx4_alloc_cmd_mailbox(dev);
+ if (IS_ERR(mailbox))
+ return PTR_ERR(mailbox);
+ context = mailbox->buf;
+ memset(context, 0, sizeof *context);
+
+ for (i = 0; i < MLX4_NUM_UP; i += 2)
+ context->prio2tc[i >> 1] = prio2tc[i] << 4 | prio2tc[i + 1];
+
+ in_mod = MLX4_SET_PORT_PRIO2TC << 8 | port;
+ err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+ MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+ mlx4_free_cmd_mailbox(dev, mailbox);
+ return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_PRIO2TC);
+
+int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
+ u8 *pg, u16 *ratelimit)
+{
+ struct mlx4_cmd_mailbox *mailbox;
+ struct mlx4_set_port_scheduler_context *context;
+ int err;
+ u32 in_mod;
+ int i;
+
+ mailbox = mlx4_alloc_cmd_mailbox(dev);
+ if (IS_ERR(mailbox))
+ return PTR_ERR(mailbox);
+ context = mailbox->buf;
+ memset(context, 0, sizeof *context);
+
+ for (i = 0; i < MLX4_NUM_TC; i++) {
+ struct mlx4_port_scheduler_tc_cfg_be *tc = &context->tc[i];
+ u16 r = ratelimit && ratelimit[i] ? ratelimit[i] :
+ MLX4_RATELIMIT_DEFAULT;
+
+ tc->pg = htons(pg[i]);
+ tc->bw_precentage = htons(tc_tx_bw[i]);
+
+ tc->max_bw_units = htons(MLX4_RATELIMIT_UNITS);
+ tc->max_bw_value = htons(r);
+ }
+
+ in_mod = MLX4_SET_PORT_SCHEDULER << 8 | port;
+ err = mlx4_cmd(dev, mailbox->dma, in_mod, 1, MLX4_CMD_SET_PORT,
+ MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
+
+ mlx4_free_cmd_mailbox(dev, mailbox);
+ return err;
+}
+EXPORT_SYMBOL(mlx4_SET_PORT_SCHEDULER);
+
int mlx4_SET_MCAST_FLTR_wrapper(struct mlx4_dev *dev, int slave,
struct mlx4_vhcr *vhcr,
struct mlx4_cmd_mailbox *inbox,
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index 9958ff2..1f3860a 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -150,6 +150,10 @@ enum {
/* statistics commands */
MLX4_CMD_QUERY_IF_STAT = 0X54,
MLX4_CMD_SET_IF_STAT = 0X55,
+
+ /* set port opcode modifiers */
+ MLX4_SET_PORT_PRIO2TC = 0x8,
+ MLX4_SET_PORT_SCHEDULER = 0x9,
};
enum {
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 44d8144..20c706f 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -626,6 +626,9 @@ int mlx4_SET_PORT_general(struct mlx4_dev *dev, u8 port, int mtu,
u8 pptx, u8 pfctx, u8 pprx, u8 pfcrx);
int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
u8 promisc);
+int mlx4_SET_PORT_PRIO2TC(struct mlx4_dev *dev, u8 port, u8 *prio2tc);
+int mlx4_SET_PORT_SCHEDULER(struct mlx4_dev *dev, u8 port, u8 *tc_tx_bw,
+ u8 *pg, u16 *ratelimit);
int mlx4_find_cached_vlan(struct mlx4_dev *dev, u8 port, u16 vid, int *idx);
int mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan, int *index);
void mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, int index);
--
1.7.8.2
^ permalink raw reply related
* [PATCH V1 0/6] net/mlx4_en: DCB QoS support
From: Amir Vadai @ 2012-03-21 9:24 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Roland Dreier, Oren Duer, Amir Vadai
DCBX version 802.1qaz is supported.
User Priority (UP) is set in QP context instead of in WQE (QP Work Queue
Element), which means that all traffic from a queue will have the same UP.
UP is also set for untagged traffic to be able to classify such traffic too.
Mapping from sk_prio to User Priority is done by sch_mqprio mapping. Although
confusingly sch_mqprio maps sk_prio to something called TC, it is not related
to DCBX's TC, and is interpreted by mlx4_en driver as UP.
The Current HW based QoS mechanism which was introduced in commit 4f57c087de9
"net: implement mechanism for HW based QOS" is in orientation to ETS traffic
class. Patch 7/8 introduces an approach which allow to use this mechanism also
with hardware who has queues per user priority (UP). After the change,
__skb_tx_hash() will direct a flow to a tx ring from a range of tx rings. This
range is defined by the caller function by the specific HW. If TC based queues,
the range is by TC number and for UP based queues, the range is by UP.
---
Changes from V0:
- Removed patches 6 and 7 who deal with the interaction between the kernel HW
QoS constructs to the queue selection logic are still under discussion with
John and some changes might be needed there.
Amir Vadai (6):
net/mlx4_en: Force user priority by QP attribute
net/mlx4_core: set port QoS attributes
net/mlx4_en: DCB QoS support
net/mlx4_en: Set max rate-limit for a TC
net/mlx4_en: sk_prio <=> UP for untagged traffic
IB/rdma_cm: TOS <=> UP mapping for IBoE
drivers/infiniband/core/cma.c | 35 ++++-
drivers/net/ethernet/mellanox/mlx4/Kconfig | 12 ++
drivers/net/ethernet/mellanox/mlx4/Makefile | 1 +
drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c | 215 +++++++++++++++++++++
drivers/net/ethernet/mellanox/mlx4/en_main.c | 2 +-
drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 47 +++++-
drivers/net/ethernet/mellanox/mlx4/en_port.h | 2 +
drivers/net/ethernet/mellanox/mlx4/en_resources.c | 6 +-
drivers/net/ethernet/mellanox/mlx4/en_rx.c | 4 +-
drivers/net/ethernet/mellanox/mlx4/en_sysfs.c | 120 ++++++++++++
drivers/net/ethernet/mellanox/mlx4/en_tx.c | 23 +--
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 20 ++
drivers/net/ethernet/mellanox/mlx4/mlx4_en.h | 32 +++-
drivers/net/ethernet/mellanox/mlx4/port.c | 62 ++++++
include/linux/mlx4/cmd.h | 4 +
include/linux/mlx4/device.h | 3 +
include/linux/mlx4/qp.h | 3 +-
17 files changed, 569 insertions(+), 22 deletions(-)
create mode 100644 drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
create mode 100644 drivers/net/ethernet/mellanox/mlx4/en_sysfs.c
--
1.7.8.2
^ permalink raw reply
* [PATCH V1 1/6] net/mlx4_en: Force user priority by QP attribute
From: Amir Vadai @ 2012-03-21 9:24 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Roland Dreier, Oren Duer, Amir Vadai, Amir Vadai
In-Reply-To: <1332321903-11704-1-git-send-email-amirv@mellanox.com>
From: Amir Vadai <amirv@mellanox.co.il>
Instead of relying on HW to change schedule queue by UP, schedule
queue is fixed for a tx_ring, and UP in WQE is ignored in this aspect. This
resolves two issues with untagged traffic:
1. untagged traffic has no UP in packet which is needed for QoS. The change
above allows setting the schedule queue (and by that the UP) of such a stream.
2. BlueFlame uses the same field used by vlan tag. So forcing UP from QPC
allows using BF for untagged but prioritized traffic.
In old firmware that force UP is not supported, untagged traffic will not subject to
QoS.
Because UP is set by QP, need to always have a tx ring per UP, even if pfcrx
module paramter is false.
Signed-off-by: Amir Vadai <amirv@mellanox.co.il>
Signed-off-by: Amir Vadai <amirv@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/en_main.c | 2 +-
drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 3 ++-
drivers/net/ethernet/mellanox/mlx4/en_resources.c | 6 +++++-
drivers/net/ethernet/mellanox/mlx4/en_rx.c | 4 ++--
drivers/net/ethernet/mellanox/mlx4/en_tx.c | 9 ++++-----
drivers/net/ethernet/mellanox/mlx4/mlx4_en.h | 6 +++---
include/linux/mlx4/qp.h | 3 ++-
7 files changed, 19 insertions(+), 14 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_main.c b/drivers/net/ethernet/mellanox/mlx4/en_main.c
index 2097a7d..346fdb2 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_main.c
@@ -114,7 +114,7 @@ static int mlx4_en_get_profile(struct mlx4_en_dev *mdev)
params->prof[i].tx_ring_size = MLX4_EN_DEF_TX_RING_SIZE;
params->prof[i].rx_ring_size = MLX4_EN_DEF_RX_RING_SIZE;
params->prof[i].tx_ring_num = MLX4_EN_NUM_TX_RINGS +
- (!!pfcrx) * MLX4_EN_NUM_PPP_RINGS;
+ MLX4_EN_NUM_PPP_RINGS;
params->prof[i].rss_rings = 0;
}
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 31b455a..2322622 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -650,7 +650,8 @@ int mlx4_en_start_port(struct net_device *dev)
/* Configure ring */
tx_ring = &priv->tx_ring[i];
- err = mlx4_en_activate_tx_ring(priv, tx_ring, cq->mcq.cqn);
+ err = mlx4_en_activate_tx_ring(priv, tx_ring, cq->mcq.cqn,
+ max(0, i - MLX4_EN_NUM_TX_RINGS));
if (err) {
en_err(priv, "Failed allocating Tx ring\n");
mlx4_en_deactivate_cq(priv, cq);
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_resources.c b/drivers/net/ethernet/mellanox/mlx4/en_resources.c
index bcbc54c..10c24c7 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_resources.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_resources.c
@@ -39,7 +39,7 @@
void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
int is_tx, int rss, int qpn, int cqn,
- struct mlx4_qp_context *context)
+ int user_prio, struct mlx4_qp_context *context)
{
struct mlx4_en_dev *mdev = priv->mdev;
@@ -57,6 +57,10 @@ void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
context->local_qpn = cpu_to_be32(qpn);
context->pri_path.ackto = 1 & 0x07;
context->pri_path.sched_queue = 0x83 | (priv->port - 1) << 6;
+ if (user_prio >= 0) {
+ context->pri_path.sched_queue |= user_prio << 3;
+ context->pri_path.feup = 1 << 6;
+ }
context->pri_path.counter_index = 0xff;
context->cqn_send = cpu_to_be32(cqn);
context->cqn_recv = cpu_to_be32(cqn);
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
index 9adbd53..d49a7ac 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -823,7 +823,7 @@ static int mlx4_en_config_rss_qp(struct mlx4_en_priv *priv, int qpn,
memset(context, 0, sizeof *context);
mlx4_en_fill_qp_context(priv, ring->actual_size, ring->stride, 0, 0,
- qpn, ring->cqn, context);
+ qpn, ring->cqn, -1, context);
context->db_rec_addr = cpu_to_be64(ring->wqres.db.dma);
/* Cancel FCS removal if FW allows */
@@ -890,7 +890,7 @@ int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv)
}
rss_map->indir_qp.event = mlx4_en_sqp_event;
mlx4_en_fill_qp_context(priv, 0, 0, 0, 1, priv->base_qpn,
- priv->rx_ring[0].cqn, &context);
+ priv->rx_ring[0].cqn, -1, &context);
if (!priv->prof->rss_rings || priv->prof->rss_rings > priv->rx_ring_num)
rss_rings = priv->rx_ring_num;
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index 1796824..445a771 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -156,7 +156,7 @@ void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv,
int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
struct mlx4_en_tx_ring *ring,
- int cq)
+ int cq, int user_prio)
{
struct mlx4_en_dev *mdev = priv->mdev;
int err;
@@ -174,7 +174,7 @@ int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
ring->doorbell_qpn = ring->qp.qpn << 8;
mlx4_en_fill_qp_context(priv, ring->size, ring->stride, 1, 0, ring->qpn,
- ring->cqn, &ring->context);
+ ring->cqn, user_prio, &ring->context);
if (ring->bf_enabled)
ring->context.usr_page = cpu_to_be32(ring->bf.uar->index);
@@ -570,18 +570,17 @@ static void build_inline_wqe(struct mlx4_en_tx_desc *tx_desc, struct sk_buff *sk
u16 mlx4_en_select_queue(struct net_device *dev, struct sk_buff *skb)
{
- struct mlx4_en_priv *priv = netdev_priv(dev);
u16 vlan_tag = 0;
/* If we support per priority flow control and the packet contains
* a vlan tag, send the packet to the TX ring assigned to that priority
*/
- if (priv->prof->rx_ppp && vlan_tx_tag_present(skb)) {
+ if (vlan_tx_tag_present(skb)) {
vlan_tag = vlan_tx_tag_get(skb);
return MLX4_EN_NUM_TX_RINGS + (vlan_tag >> 13);
}
- return skb_tx_hash(dev, skb);
+ return __skb_tx_hash(dev, skb, MLX4_EN_NUM_TX_RINGS);
}
static void mlx4_bf_copy(void __iomem *dst, unsigned long *src, unsigned bytecnt)
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index 9e2b911..5bd7c2a 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -521,7 +521,7 @@ int mlx4_en_create_tx_ring(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ri
void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ring);
int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
struct mlx4_en_tx_ring *ring,
- int cq);
+ int cq, int user_prio);
void mlx4_en_deactivate_tx_ring(struct mlx4_en_priv *priv,
struct mlx4_en_tx_ring *ring);
@@ -539,8 +539,8 @@ int mlx4_en_process_rx_cq(struct net_device *dev,
int budget);
int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget);
void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
- int is_tx, int rss, int qpn, int cqn,
- struct mlx4_qp_context *context);
+ int is_tx, int rss, int qpn, int cqn, int user_prio,
+ struct mlx4_qp_context *context);
void mlx4_en_sqp_event(struct mlx4_qp *qp, enum mlx4_event event);
int mlx4_en_map_buffer(struct mlx4_buf *buf);
void mlx4_en_unmap_buffer(struct mlx4_buf *buf);
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index 091f9e7..96005d7 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -139,7 +139,8 @@ struct mlx4_qp_path {
u8 rgid[16];
u8 sched_queue;
u8 vlan_index;
- u8 reserved3[2];
+ u8 feup;
+ u8 reserved3;
u8 reserved4[2];
u8 dmac[6];
};
--
1.7.8.2
^ permalink raw reply related
* [PATCH 5/9] ipvs: use adaptive pause in master thread
From: Simon Horman @ 2012-03-21 8:56 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Simon Horman
In-Reply-To: <1332320185-27157-1-git-send-email-horms@verge.net.au>
From: Julian Anastasov <ja@ssi.bg>
High rate of sync messages in master can lead to
overflowing the socket buffer and dropping the messages.
Instead of using fixed pause try to adapt to the sync
rate, so that we do not wakeup too often while at the same
time staying below the socket buffer limit.
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Tested-by: Aleksey Chudov <aleksey.chudov@gmail.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
net/netfilter/ipvs/ip_vs_sync.c | 60 +++++++++++++++++++++++++++++---------
1 files changed, 46 insertions(+), 14 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_sync.c b/net/netfilter/ipvs/ip_vs_sync.c
index 0e36679..9201c43 100644
--- a/net/netfilter/ipvs/ip_vs_sync.c
+++ b/net/netfilter/ipvs/ip_vs_sync.c
@@ -1392,18 +1392,22 @@ ip_vs_send_async(struct socket *sock, const char *buffer, const size_t length)
return len;
}
-static void
+static int
ip_vs_send_sync_msg(struct socket *sock, struct ip_vs_sync_mesg *msg)
{
int msize;
+ int ret;
msize = msg->size;
/* Put size in network byte order */
msg->size = htons(msg->size);
- if (ip_vs_send_async(sock, (char *)msg, msize) != msize)
- pr_err("ip_vs_send_async error\n");
+ ret = ip_vs_send_async(sock, (char *)msg, msize);
+ if (ret >= 0 || ret == -EAGAIN)
+ return ret;
+ pr_err("ip_vs_send_async error %d\n", ret);
+ return 0;
}
static int
@@ -1428,33 +1432,61 @@ ip_vs_receive(struct socket *sock, char *buffer, const size_t buflen)
return len;
}
+/* Get next buffer to send */
+static inline struct ip_vs_sync_buff *
+next_sync_buff(struct netns_ipvs *ipvs)
+{
+ struct ip_vs_sync_buff *sb;
+
+ sb = sb_dequeue(ipvs);
+ if (sb)
+ return sb;
+ /* Do not delay entries in buffer for more than 2 seconds */
+ return get_curr_sync_buff(ipvs, 2 * HZ);
+}
static int sync_thread_master(void *data)
{
struct ip_vs_sync_thread_data *tinfo = data;
struct netns_ipvs *ipvs = net_ipvs(tinfo->net);
- struct ip_vs_sync_buff *sb;
+ struct ip_vs_sync_buff *sb = NULL;
+ int pause = HZ / 10;
pr_info("sync thread started: state = MASTER, mcast_ifn = %s, "
"syncid = %d\n",
ipvs->master_mcast_ifn, ipvs->master_syncid);
while (!kthread_should_stop()) {
- while ((sb = sb_dequeue(ipvs))) {
- ip_vs_send_sync_msg(tinfo->sock, sb->mesg);
- ip_vs_sync_buff_release(sb);
- }
+ int count = 0;
- /* check if entries stay in ipvs->sync_buff for 2 seconds */
- sb = get_curr_sync_buff(ipvs, 2 * HZ);
- if (sb) {
- ip_vs_send_sync_msg(tinfo->sock, sb->mesg);
- ip_vs_sync_buff_release(sb);
+ for (;;) {
+ if (!sb) {
+ sb = next_sync_buff(ipvs);
+ if (!sb)
+ break;
+ }
+ if (ip_vs_send_sync_msg(tinfo->sock, sb->mesg) >= 0 ||
+ pause == 1) {
+ ip_vs_sync_buff_release(sb);
+ sb = NULL;
+ count++;
+ } else {
+ pause = max(1, (pause >> 1));
+ break;
+ }
}
- schedule_timeout_interruptible(HZ);
+ /* Max packets to send at once */
+ if (count > 200)
+ pause = max(1, (pause - HZ / 20));
+ else if (count < 20)
+ pause = min(HZ / 4, (pause + HZ / 20));
+ schedule_timeout_interruptible(sb ? 1 : pause);
}
+ if (sb)
+ ip_vs_sync_buff_release(sb);
+
/* clean up the sync_buff queue */
while ((sb = sb_dequeue(ipvs)))
ip_vs_sync_buff_release(sb);
--
1.7.6.3
^ permalink raw reply related
* [PATCH 6/9] ipvs: reduce sync rate with time thresholds
From: Simon Horman @ 2012-03-21 8:56 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Simon Horman
In-Reply-To: <1332320185-27157-1-git-send-email-horms@verge.net.au>
From: Julian Anastasov <ja@ssi.bg>
Add two new sysctl vars to control the sync rate with the
main idea to reduce the rate for connection templates because
currently it depends on the packet rate for controlled connections.
This mechanism should be useful also for normal connections
with high traffic.
sync_refresh_period: in seconds, difference in reported connection
timer that triggers new sync message. It can be used to
avoid sync messages for the specified period (or half of
the connection timeout if it is lower) if connection state
is not changed from last sync.
sync_retries: integer, 0..3, defines sync retries with period of
sync_refresh_period/8. Useful to protect against loss of
sync messages.
Allow sysctl_sync_threshold to be used with
sysctl_sync_period=0, so that only single sync message is sent
if sync_refresh_period is also 0.
Add new field "sync_endtime" in connection structure to
hold the reported time when connection expires. The 2 lowest
bits will represent the retry count.
As the sysctl_sync_period now can be 0 use ACCESS_ONCE to
avoid division by zero.
Special thanks to Aleksey Chudov for being patient with me,
for his extensive reports and helping in all tests.
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Tested-by: Aleksey Chudov <aleksey.chudov@gmail.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
include/net/ip_vs.h | 36 ++++++++++--
net/netfilter/ipvs/ip_vs_conn.c | 7 ++-
net/netfilter/ipvs/ip_vs_core.c | 30 +---------
net/netfilter/ipvs/ip_vs_ctl.c | 25 ++++++++-
net/netfilter/ipvs/ip_vs_sync.c | 121 +++++++++++++++++++++++++++++++++-----
5 files changed, 168 insertions(+), 51 deletions(-)
diff --git a/include/net/ip_vs.h b/include/net/ip_vs.h
index ebe517f..8bee5d0 100644
--- a/include/net/ip_vs.h
+++ b/include/net/ip_vs.h
@@ -504,6 +504,7 @@ struct ip_vs_conn {
* state transition triggerd
* synchronization
*/
+ unsigned long sync_endtime; /* jiffies + sent_retries */
/* Control members */
struct ip_vs_conn *control; /* Master control connection */
@@ -873,6 +874,8 @@ struct netns_ipvs {
int sysctl_expire_nodest_conn;
int sysctl_expire_quiescent_template;
int sysctl_sync_threshold[2];
+ unsigned int sysctl_sync_refresh_period;
+ int sysctl_sync_retries;
int sysctl_nat_icmp_send;
/* ip_vs_lblc */
@@ -908,9 +911,12 @@ struct netns_ipvs {
struct net *net; /* Needed by timer routines */
};
-#define DEFAULT_SYNC_THRESHOLD 3
-#define DEFAULT_SYNC_PERIOD 50
-#define DEFAULT_SYNC_VER 1
+#define DEFAULT_SYNC_THRESHOLD 3
+#define DEFAULT_SYNC_PERIOD 50
+#define DEFAULT_SYNC_REFRESH_PERIOD (0U * HZ)
+#define DEFAULT_SYNC_RETRIES 0
+#define DEFAULT_SYNC_VER 1
+#define IPVS_SYNC_FLUSH_TIME (HZ * 2)
#ifdef CONFIG_SYSCTL
@@ -921,7 +927,17 @@ static inline int sysctl_sync_threshold(struct netns_ipvs *ipvs)
static inline int sysctl_sync_period(struct netns_ipvs *ipvs)
{
- return ipvs->sysctl_sync_threshold[1];
+ return ACCESS_ONCE(ipvs->sysctl_sync_threshold[1]);
+}
+
+static inline unsigned int sysctl_sync_refresh_period(struct netns_ipvs *ipvs)
+{
+ return ACCESS_ONCE(ipvs->sysctl_sync_refresh_period);
+}
+
+static inline int sysctl_sync_retries(struct netns_ipvs *ipvs)
+{
+ return ipvs->sysctl_sync_retries;
}
static inline int sysctl_sync_ver(struct netns_ipvs *ipvs)
@@ -941,6 +957,16 @@ static inline int sysctl_sync_period(struct netns_ipvs *ipvs)
return DEFAULT_SYNC_PERIOD;
}
+static inline unsigned int sysctl_sync_refresh_period(struct netns_ipvs *ipvs)
+{
+ return DEFAULT_SYNC_REFRESH_PERIOD;
+}
+
+static inline int sysctl_sync_retries(struct netns_ipvs *ipvs)
+{
+ return DEFAULT_SYNC_RETRIES & 3;
+}
+
static inline int sysctl_sync_ver(struct netns_ipvs *ipvs)
{
return DEFAULT_SYNC_VER;
@@ -1218,7 +1244,7 @@ extern struct ip_vs_dest *ip_vs_try_bind_dest(struct ip_vs_conn *cp);
extern int start_sync_thread(struct net *net, int state, char *mcast_ifn,
__u8 syncid);
extern int stop_sync_thread(struct net *net, int state);
-extern void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp);
+extern void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp, int pkts);
/*
diff --git a/net/netfilter/ipvs/ip_vs_conn.c b/net/netfilter/ipvs/ip_vs_conn.c
index a7c6bc8..126971c 100644
--- a/net/netfilter/ipvs/ip_vs_conn.c
+++ b/net/netfilter/ipvs/ip_vs_conn.c
@@ -761,7 +761,8 @@ int ip_vs_check_template(struct ip_vs_conn *ct)
static void ip_vs_conn_expire(unsigned long data)
{
struct ip_vs_conn *cp = (struct ip_vs_conn *)data;
- struct netns_ipvs *ipvs = net_ipvs(ip_vs_conn_net(cp));
+ struct net *net = ip_vs_conn_net(cp);
+ struct netns_ipvs *ipvs = net_ipvs(net);
cp->timeout = 60*HZ;
@@ -826,6 +827,9 @@ static void ip_vs_conn_expire(unsigned long data)
atomic_read(&cp->refcnt)-1,
atomic_read(&cp->n_control));
+ if (ipvs->sync_state & IP_VS_STATE_MASTER)
+ ip_vs_sync_conn(net, cp, sysctl_sync_threshold(ipvs));
+
ip_vs_conn_put(cp);
}
@@ -899,6 +903,7 @@ ip_vs_conn_new(const struct ip_vs_conn_param *p,
/* Set its state and timeout */
cp->state = 0;
cp->timeout = 3*HZ;
+ cp->sync_endtime = jiffies & ~3UL;
/* Bind its packet transmitter */
#ifdef CONFIG_IP_VS_IPV6
diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
index 2555816..b3686d8 100644
--- a/net/netfilter/ipvs/ip_vs_core.c
+++ b/net/netfilter/ipvs/ip_vs_core.c
@@ -1613,34 +1613,8 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb, int af)
else
pkts = atomic_add_return(1, &cp->in_pkts);
- if ((ipvs->sync_state & IP_VS_STATE_MASTER) &&
- cp->protocol == IPPROTO_SCTP) {
- if ((cp->state == IP_VS_SCTP_S_ESTABLISHED &&
- (pkts % sysctl_sync_period(ipvs)
- == sysctl_sync_threshold(ipvs))) ||
- (cp->old_state != cp->state &&
- ((cp->state == IP_VS_SCTP_S_CLOSED) ||
- (cp->state == IP_VS_SCTP_S_SHUT_ACK_CLI) ||
- (cp->state == IP_VS_SCTP_S_SHUT_ACK_SER)))) {
- ip_vs_sync_conn(net, cp);
- goto out;
- }
- }
-
- /* Keep this block last: TCP and others with pp->num_states <= 1 */
- else if ((ipvs->sync_state & IP_VS_STATE_MASTER) &&
- (((cp->protocol != IPPROTO_TCP ||
- cp->state == IP_VS_TCP_S_ESTABLISHED) &&
- (pkts % sysctl_sync_period(ipvs)
- == sysctl_sync_threshold(ipvs))) ||
- ((cp->protocol == IPPROTO_TCP) && (cp->old_state != cp->state) &&
- ((cp->state == IP_VS_TCP_S_FIN_WAIT) ||
- (cp->state == IP_VS_TCP_S_CLOSE) ||
- (cp->state == IP_VS_TCP_S_CLOSE_WAIT) ||
- (cp->state == IP_VS_TCP_S_TIME_WAIT)))))
- ip_vs_sync_conn(net, cp);
-out:
- cp->old_state = cp->state;
+ if (ipvs->sync_state & IP_VS_STATE_MASTER)
+ ip_vs_sync_conn(net, cp, pkts);
ip_vs_conn_put(cp);
return ret;
diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c
index b3afe18..da922c7 100644
--- a/net/netfilter/ipvs/ip_vs_ctl.c
+++ b/net/netfilter/ipvs/ip_vs_ctl.c
@@ -1599,6 +1599,10 @@ static int ip_vs_zero_all(struct net *net)
}
#ifdef CONFIG_SYSCTL
+
+static int zero;
+static int three = 3;
+
static int
proc_do_defense_mode(ctl_table *table, int write,
void __user *buffer, size_t *lenp, loff_t *ppos)
@@ -1632,7 +1636,8 @@ proc_do_sync_threshold(ctl_table *table, int write,
memcpy(val, valp, sizeof(val));
rc = proc_dointvec(table, write, buffer, lenp, ppos);
- if (write && (valp[0] < 0 || valp[1] < 0 || valp[0] >= valp[1])) {
+ if (write && (valp[0] < 0 || valp[1] < 0 ||
+ (valp[0] >= valp[1] && valp[1]))) {
/* Restore the correct value */
memcpy(valp, val, sizeof(val));
}
@@ -1743,6 +1748,20 @@ static struct ctl_table vs_vars[] = {
.proc_handler = proc_do_sync_threshold,
},
{
+ .procname = "sync_refresh_period",
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = proc_dointvec_jiffies,
+ },
+ {
+ .procname = "sync_retries",
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = proc_dointvec_minmax,
+ .extra1 = &zero,
+ .extra2 = &three,
+ },
+ {
.procname = "nat_icmp_send",
.maxlen = sizeof(int),
.mode = 0644,
@@ -3661,6 +3680,10 @@ int __net_init ip_vs_control_net_init_sysctl(struct net *net)
ipvs->sysctl_sync_threshold[1] = DEFAULT_SYNC_PERIOD;
tbl[idx].data = &ipvs->sysctl_sync_threshold;
tbl[idx++].maxlen = sizeof(ipvs->sysctl_sync_threshold);
+ ipvs->sysctl_sync_refresh_period = DEFAULT_SYNC_REFRESH_PERIOD;
+ tbl[idx++].data = &ipvs->sysctl_sync_refresh_period;
+ ipvs->sysctl_sync_retries = clamp_t(int, DEFAULT_SYNC_RETRIES, 0, 3);
+ tbl[idx++].data = &ipvs->sysctl_sync_retries;
tbl[idx++].data = &ipvs->sysctl_nat_icmp_send;
diff --git a/net/netfilter/ipvs/ip_vs_sync.c b/net/netfilter/ipvs/ip_vs_sync.c
index 9201c43..8ee8a57 100644
--- a/net/netfilter/ipvs/ip_vs_sync.c
+++ b/net/netfilter/ipvs/ip_vs_sync.c
@@ -442,11 +442,94 @@ ip_vs_sync_buff_create_v0(struct netns_ipvs *ipvs)
return sb;
}
+/* Check if conn should be synced.
+ * pkts: conn packets, use sysctl_sync_threshold to avoid packet check
+ * - (1) sync_refresh_period: reduce sync rate. Additionally, retry
+ * sync_retries times with period of sync_refresh_period/8
+ * - (2) if both sync_refresh_period and sync_period are 0 send sync only
+ * for state changes or only once when pkts matches sync_threshold
+ * - (3) templates: rate can be reduced only with sync_refresh_period or
+ * with (2)
+ */
+static int ip_vs_sync_conn_needed(struct netns_ipvs *ipvs,
+ struct ip_vs_conn *cp, int pkts)
+{
+ unsigned long orig = ACCESS_ONCE(cp->sync_endtime);
+ unsigned long now = jiffies;
+ unsigned long n = (now + cp->timeout) & ~3UL;
+ unsigned int sync_refresh_period;
+ int sync_period;
+ int force;
+
+ /* Check if we sync in current state */
+ if (unlikely(cp->flags & IP_VS_CONN_F_TEMPLATE))
+ force = 0;
+ else if (likely(cp->protocol == IPPROTO_TCP)) {
+ if (!((1 << cp->state) &
+ ((1 << IP_VS_TCP_S_ESTABLISHED) |
+ (1 << IP_VS_TCP_S_FIN_WAIT) |
+ (1 << IP_VS_TCP_S_CLOSE) |
+ (1 << IP_VS_TCP_S_CLOSE_WAIT) |
+ (1 << IP_VS_TCP_S_TIME_WAIT))))
+ return 0;
+ force = cp->state != cp->old_state;
+ if (force && cp->state != IP_VS_TCP_S_ESTABLISHED)
+ goto set;
+ } else if (unlikely(cp->protocol == IPPROTO_SCTP)) {
+ if (!((1 << cp->state) &
+ ((1 << IP_VS_SCTP_S_ESTABLISHED) |
+ (1 << IP_VS_SCTP_S_CLOSED) |
+ (1 << IP_VS_SCTP_S_SHUT_ACK_CLI) |
+ (1 << IP_VS_SCTP_S_SHUT_ACK_SER))))
+ return 0;
+ force = cp->state != cp->old_state;
+ if (force && cp->state != IP_VS_SCTP_S_ESTABLISHED)
+ goto set;
+ } else {
+ /* UDP or another protocol with single state */
+ force = 0;
+ }
+
+ sync_refresh_period = sysctl_sync_refresh_period(ipvs);
+ if (sync_refresh_period > 0) {
+ long diff = n - orig;
+ long min_diff = max(cp->timeout >> 1, 10UL * HZ);
+
+ /* Avoid sync if difference is below sync_refresh_period
+ * and below the half timeout.
+ */
+ if (abs(diff) < min_t(long, sync_refresh_period, min_diff)) {
+ int retries = orig & 3;
+
+ if (retries >= sysctl_sync_retries(ipvs))
+ return 0;
+ if (time_before(now, orig - cp->timeout +
+ (sync_refresh_period >> 3)))
+ return 0;
+ n |= retries + 1;
+ }
+ }
+ sync_period = sysctl_sync_period(ipvs);
+ if (sync_period > 0) {
+ if (!(cp->flags & IP_VS_CONN_F_TEMPLATE) &&
+ pkts % sync_period != sysctl_sync_threshold(ipvs))
+ return 0;
+ } else if (sync_refresh_period <= 0 &&
+ pkts != sysctl_sync_threshold(ipvs))
+ return 0;
+
+set:
+ cp->old_state = cp->state;
+ n = cmpxchg(&cp->sync_endtime, orig, n);
+ return n == orig || force;
+}
+
/*
* Version 0 , could be switched in by sys_ctl.
* Add an ip_vs_conn information into the current sync_buff.
*/
-void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp)
+static void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp,
+ int pkts)
{
struct netns_ipvs *ipvs = net_ipvs(net);
struct ip_vs_sync_mesg_v0 *m;
@@ -459,6 +542,9 @@ void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp)
if (cp->flags & IP_VS_CONN_F_ONE_PACKET)
return;
+ if (!ip_vs_sync_conn_needed(ipvs, cp, pkts))
+ return;
+
spin_lock(&ipvs->sync_buff_lock);
if (!ipvs->sync_buff) {
ipvs->sync_buff =
@@ -504,8 +590,14 @@ void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp)
spin_unlock(&ipvs->sync_buff_lock);
/* synchronize its controller if it has */
- if (cp->control)
- ip_vs_sync_conn(net, cp->control);
+ cp = cp->control;
+ if (cp) {
+ if (cp->flags & IP_VS_CONN_F_TEMPLATE)
+ pkts = atomic_add_return(1, &cp->in_pkts);
+ else
+ pkts = sysctl_sync_threshold(ipvs);
+ ip_vs_sync_conn(net, cp->control, pkts);
+ }
}
/*
@@ -513,7 +605,7 @@ void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp)
* Called by ip_vs_in.
* Sending Version 1 messages
*/
-void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp)
+void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp, int pkts)
{
struct netns_ipvs *ipvs = net_ipvs(net);
struct ip_vs_sync_mesg *m;
@@ -523,13 +615,16 @@ void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp)
/* Handle old version of the protocol */
if (sysctl_sync_ver(ipvs) == 0) {
- ip_vs_sync_conn_v0(net, cp);
+ ip_vs_sync_conn_v0(net, cp, pkts);
return;
}
/* Do not sync ONE PACKET */
if (cp->flags & IP_VS_CONN_F_ONE_PACKET)
goto control;
sloop:
+ if (!ip_vs_sync_conn_needed(ipvs, cp, pkts))
+ goto control;
+
/* Sanity checks */
pe_name_len = 0;
if (cp->pe_data_len) {
@@ -644,16 +739,10 @@ control:
cp = cp->control;
if (!cp)
return;
- /*
- * Reduce sync rate for templates
- * i.e only increment in_pkts for Templates.
- */
- if (cp->flags & IP_VS_CONN_F_TEMPLATE) {
- int pkts = atomic_add_return(1, &cp->in_pkts);
-
- if (pkts % sysctl_sync_period(ipvs) != 1)
- return;
- }
+ if (cp->flags & IP_VS_CONN_F_TEMPLATE)
+ pkts = atomic_add_return(1, &cp->in_pkts);
+ else
+ pkts = sysctl_sync_threshold(ipvs);
goto sloop;
}
@@ -1442,7 +1531,7 @@ next_sync_buff(struct netns_ipvs *ipvs)
if (sb)
return sb;
/* Do not delay entries in buffer for more than 2 seconds */
- return get_curr_sync_buff(ipvs, 2 * HZ);
+ return get_curr_sync_buff(ipvs, IPVS_SYNC_FLUSH_TIME);
}
static int sync_thread_master(void *data)
--
1.7.6.3
^ permalink raw reply related
* [PATCH 2/9] ipvs: remove check for IP_VS_CONN_F_SYNC from ip_vs_bind_dest
From: Simon Horman @ 2012-03-21 8:56 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Simon Horman
In-Reply-To: <1332320185-27157-1-git-send-email-horms@verge.net.au>
From: Julian Anastasov <ja@ssi.bg>
As the IP_VS_CONN_F_INACTIVE bit is properly set
in cp->flags for all kind of connections we do not need to
add special checks for synced connections when updating
the activeconns/inactconns counters for first time. Now
logic will look just like in ip_vs_unbind_dest.
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
net/netfilter/ipvs/ip_vs_conn.c | 9 ++++-----
1 files changed, 4 insertions(+), 5 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_conn.c b/net/netfilter/ipvs/ip_vs_conn.c
index 6413d58..18f58a08 100644
--- a/net/netfilter/ipvs/ip_vs_conn.c
+++ b/net/netfilter/ipvs/ip_vs_conn.c
@@ -585,11 +585,10 @@ ip_vs_bind_dest(struct ip_vs_conn *cp, struct ip_vs_dest *dest)
/* Update the connection counters */
if (!(cp->flags & IP_VS_CONN_F_TEMPLATE)) {
- /* It is a normal connection, so increase the inactive
- connection counter because it is in TCP SYNRECV
- state (inactive) or other protocol inacive state */
- if ((cp->flags & IP_VS_CONN_F_SYNC) &&
- (!(cp->flags & IP_VS_CONN_F_INACTIVE)))
+ /* It is a normal connection, so modify the counters
+ * according to the flags, later the protocol can
+ * update them on state change */
+ if (!(cp->flags & IP_VS_CONN_F_INACTIVE))
atomic_inc(&dest->activeconns);
else
atomic_inc(&dest->inactconns);
--
1.7.6.3
^ permalink raw reply related
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