* Re: [PATCH 0/2] cxgb4 FR_NSMR_TPTE_WR support
From: Doug Ledford @ 2016-10-03 14:50 UTC (permalink / raw)
To: Steve Wise; +Cc: netdev, linux-rdma, 'David Miller'
In-Reply-To: <023401d2190b$03b00e50$0b102af0$@opengridcomputing.com>
[-- Attachment #1.1: Type: text/plain, Size: 710 bytes --]
On 9/27/2016 6:03 PM, Steve Wise wrote:
>>
>>>
>>>>
>>>> From: Steve Wise <swise@opengridcomputing.com>
>>>> Date: Fri, 16 Sep 2016 07:54:55 -0700
>>>>
>>>>> This series enables a new work request to optimize small REG_MR
>>>>> operations. This is intended for 4.9. If everyone agrees, I suggest
>>>>> Doug take both the cxgb4 and iw_cxgb4 patches through his tree.
>>>>
>>>> I'm assuming this mean that I do _not_ apply these to my tree.
>>>
>>> Yes, if you're ok with that.
>>
>> I am.
>
> Doug, please include this for 4.9 if it looks good to you. Let me know.
>
> Thanks,
>
> Steve.
>
Applied, thanks.
--
Doug Ledford <dledford@redhat.com>
GPG Key ID: 0E572FDD
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]
^ permalink raw reply
* Re: [PATCH] phy: micrel.c: Support ksz9031 energy-detect power-down mode
From: Andrew Lunn @ 2016-10-03 14:35 UTC (permalink / raw)
To: Mike Looijmans
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, devicetree-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA,
f.fainelli-Re5JQEeQqe8AvxtiuMwx3w
In-Reply-To: <1475499870-11465-1-git-send-email-mike.looijmans-Oq418RWZeHk@public.gmane.org>
> + Boolean:
> +
> + - enable-edpd : Not related to timing. Specify this property to enable
> + energy-detect power-down mode in the PHY.
The Broadcom PHYs unconditionally enable this feature.
The SMSC driver has an option: smsc,disable-energy-detect because it
seems to have a rare hardware bug when it is enabled.
And you would like to make it off by default.
Could we avoid having three drivers doing three different things?
Could you have it unconditionally on, and see if anybody reports it
being broken?
Andrew
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH] ptp: Fix resource leak in case of error
From: Richard Cochran @ 2016-10-03 14:04 UTC (permalink / raw)
To: Christophe JAILLET; +Cc: netdev, linux-kernel, kernel-janitors
In-Reply-To: <1475391856-15046-1-git-send-email-christophe.jaillet@wanadoo.fr>
On Sun, Oct 02, 2016 at 09:04:16AM +0200, Christophe JAILLET wrote:
> A call to 'ida_simple_remove()' is missing in the error handling path.
>
> This as been spotted with the following coccinelle script which tries to
> detect missing 'ida_simple_remove()' call in error handling paths.
Acked-by: Richard Cochran <richardcochran@gmail.com>
^ permalink raw reply
* Get a safe loan within 48/hours
From: info @ 2016-10-01 13:04 UTC (permalink / raw)
To: Recipients
I am Mr.Mathew Raymond I give out affordable loan at 2% interest rate, Contact us with your (1)Names: (2) Country: (3)Amount needed as loan: (4) Loan duration: (5)Cell-Phone number: Visit our Via Email: mathewraymond1@gmail.com
Regard
Mr.Mathew Raymond
Managing Director
Mathew Raymond Loan Firm.
^ permalink raw reply
* Re: [PATCH net v2] net: pktgen: fix pkt_size
From: Greg @ 2016-10-03 13:56 UTC (permalink / raw)
To: Paolo Abeni
Cc: netdev, David S. Miller, Bogdan Hamciuc, Ben Greear,
Sergei Shtylyov
In-Reply-To: <1475345120.6330.21.camel@redhat.com>
On Sat, 2016-10-01 at 20:05 +0200, Paolo Abeni wrote:
> On Fri, 2016-09-30 at 08:39 -0700, Greg wrote:
> > On Fri, 2016-09-30 at 16:56 +0200, Paolo Abeni wrote:
> > > The commit 879c7220e828 ("net: pktgen: Observe needed_headroom
> > > of the device") increased the 'pkt_overhead' field value by
> > > LL_RESERVED_SPACE.
> > > As a side effect the generated packet size, computed as:
> > >
> > > /* Eth + IPh + UDPh + mpls */
> > > datalen = pkt_dev->cur_pkt_size - 14 - 20 - 8 -
> > > pkt_dev->pkt_overhead;
> > >
> > > is decreased by the same value.
> > > The above changed slightly the behavior of existing pktgen users,
> > > and made the procfs interface somewhat inconsistent.
> > > Fix it by restoring the previous pkt_overhead value and using
> > > LL_RESERVED_SPACE as extralen in skb allocation.
> > > Also, change pktgen_alloc_skb() to only partially reserve
> > > the headroom to allow the caller to prefetch from ll header
> > > start.
> > >
> > > v1 -> v2:
> > > - fixed some typos in the comments
> > >
> > > Fixes: 879c7220e828 ("net: pktgen: Observe needed_headroom of the device")
> > > Suggested-by: Ben Greear <greearb@candelatech.com>
> > > Signed-off-by: Paolo Abeni <pabeni@redhat.com>
> > > ---
> > > net/core/pktgen.c | 21 ++++++++++-----------
> > > 1 file changed, 10 insertions(+), 11 deletions(-)
> > >
> > > diff --git a/net/core/pktgen.c b/net/core/pktgen.c
> > > index bbd118b..5219a9e 100644
> > > --- a/net/core/pktgen.c
> > > +++ b/net/core/pktgen.c
> > > @@ -2286,7 +2286,7 @@ out:
> > >
> > > static inline void set_pkt_overhead(struct pktgen_dev *pkt_dev)
> > > {
> > > - pkt_dev->pkt_overhead = LL_RESERVED_SPACE(pkt_dev->odev);
> > > + pkt_dev->pkt_overhead = 0;
> > > pkt_dev->pkt_overhead += pkt_dev->nr_labels*sizeof(u32);
> > > pkt_dev->pkt_overhead += VLAN_TAG_SIZE(pkt_dev);
> > > pkt_dev->pkt_overhead += SVLAN_TAG_SIZE(pkt_dev);
> > > @@ -2777,13 +2777,13 @@ static void pktgen_finalize_skb(struct pktgen_dev *pkt_dev, struct sk_buff *skb,
> > > }
> > >
> > > static struct sk_buff *pktgen_alloc_skb(struct net_device *dev,
> > > - struct pktgen_dev *pkt_dev,
> > > - unsigned int extralen)
> > > + struct pktgen_dev *pkt_dev)
> > > {
> > > + unsigned int extralen = LL_RESERVED_SPACE(dev);
> > > struct sk_buff *skb = NULL;
> > > - unsigned int size = pkt_dev->cur_pkt_size + 64 + extralen +
> > > - pkt_dev->pkt_overhead;
> > > + unsigned int size;
> > >
> > > + size = pkt_dev->cur_pkt_size + 64 + extralen + pkt_dev->pkt_overhead;
> > > if (pkt_dev->flags & F_NODE) {
> > > int node = pkt_dev->node >= 0 ? pkt_dev->node : numa_node_id();
> > >
> > > @@ -2796,8 +2796,9 @@ static struct sk_buff *pktgen_alloc_skb(struct net_device *dev,
> > > skb = __netdev_alloc_skb(dev, size, GFP_NOWAIT);
> > > }
> > >
> > > + /* the caller pre-fetches from skb->data and reserves for the mac hdr */
> > > if (likely(skb))
> > > - skb_reserve(skb, LL_RESERVED_SPACE(dev));
> > > + skb_reserve(skb, extralen - 16);
> >
> > Is the 16 here the same as HD_DATA_MOD?
> >
> > Magic numbers...
>
> This magic numbers comes from the current code (the extralen argument),
> this patch just move it around. I can send a v3 using a define for it,
> if really needed, but it seems more a separated clean-up.
Sure, sounds good. Looks like Dave applied it.
- Greg
>
> Thank you,
>
> Paolo
>
^ permalink raw reply
* [PATCH] phy: micrel.c: Support ksz9031 energy-detect power-down mode
From: Mike Looijmans @ 2016-10-03 13:04 UTC (permalink / raw)
To: netdev; +Cc: devicetree, linux-kernel, f.fainelli, Mike Looijmans
Set bit 0 in register 1C.23 to enable the EDPD feature of the
KSZ9031 PHY. This reduces power consumption when the link is
down. To use this, set "enable-edpd" in the devicetree.
Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl>
---
.../devicetree/bindings/net/micrel-ksz90x1.txt | 6 ++++++
drivers/net/phy/micrel.c | 19 +++++++++++++++++++
2 files changed, 25 insertions(+)
diff --git a/Documentation/devicetree/bindings/net/micrel-ksz90x1.txt b/Documentation/devicetree/bindings/net/micrel-ksz90x1.txt
index f9c32ad..fbec6f8 100644
--- a/Documentation/devicetree/bindings/net/micrel-ksz90x1.txt
+++ b/Documentation/devicetree/bindings/net/micrel-ksz90x1.txt
@@ -56,6 +56,11 @@ KSZ9031:
- txd2-skew-ps : Skew control of TX data 2 pad
- txd3-skew-ps : Skew control of TX data 3 pad
+ Boolean:
+
+ - enable-edpd : Not related to timing. Specify this property to enable
+ energy-detect power-down mode in the PHY.
+
Examples:
mdio {
@@ -64,6 +69,7 @@ Examples:
rxdv-skew-ps = <0>;
txc-skew-ps = <3000>;
txen-skew-ps = <0>;
+ enable-edpd;
reg = <0>;
};
};
diff --git a/drivers/net/phy/micrel.c b/drivers/net/phy/micrel.c
index 4516c8a..0a365b4 100644
--- a/drivers/net/phy/micrel.c
+++ b/drivers/net/phy/micrel.c
@@ -409,6 +409,10 @@ static int ksz9021_config_init(struct phy_device *phydev)
#define MII_KSZ9031RN_TX_DATA_PAD_SKEW 6
#define MII_KSZ9031RN_CLK_PAD_SKEW 8
+/* MMD Address 0x1C */
+#define MII_KSZ9031RN_EDPD 0x23
+#define MII_KSZ9031RN_EDPD_ENABLE BIT(0)
+
static int ksz9031_extended_write(struct phy_device *phydev,
u8 mode, u32 dev_addr, u32 regnum, u16 val)
{
@@ -480,6 +484,18 @@ static int ksz9031_center_flp_timing(struct phy_device *phydev)
return genphy_restart_aneg(phydev);
}
+/* Enable energy-detect power-down mode */
+static int ksz9031_enable_edpd(struct phy_device *phydev)
+{
+ int reg;
+
+ reg = ksz9031_extended_read(phydev, OP_DATA, 0x1C, MII_KSZ9031RN_EDPD);
+ if (reg < 0)
+ return reg;
+ return ksz9031_extended_write(phydev, OP_DATA, 0x1C, MII_KSZ9031RN_EDPD,
+ reg | MII_KSZ9031RN_EDPD_ENABLE);
+}
+
static int ksz9031_config_init(struct phy_device *phydev)
{
const struct device *dev = &phydev->mdio.dev;
@@ -522,6 +538,9 @@ static int ksz9031_config_init(struct phy_device *phydev)
ksz9031_of_load_skew_values(phydev, of_node,
MII_KSZ9031RN_TX_DATA_PAD_SKEW, 4,
tx_data_skews, 4);
+
+ if (of_property_read_bool(of_node, "enable-edpd"))
+ ksz9031_enable_edpd(phydev);
}
return ksz9031_center_flp_timing(phydev);
--
1.9.1
^ permalink raw reply related
* Re: [v2] mwifiex: report wakeup for wowlan
From: Kalle Valo @ 2016-10-03 13:04 UTC (permalink / raw)
To: Rajat Jain
Cc: Amitkumar Karwar, Nishant Sarmukadam,
linux-wireless-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA, Rajat Jain, Wei-Ning Huang,
Brian Norris, Eric Caruso, rajatxjain-Re5JQEeQqe8AvxtiuMwx3w
In-Reply-To: <1475027104-17423-1-git-send-email-rajatja-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>
Rajat Jain <rajatja-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org> wrote:
> Enable notifying wakeup source to the PM core in case of
> a wake on wireless LAN event.
>
> Signed-off-by: Wei-Ning Huang <wnhuang-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>
> Signed-off-by: Rajat Jain <rajatja-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>
> Tested-by: Wei-Ning Huang <wnhuang-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
> Reviewed-by: Eric Caruso <ejcaruso-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
> Acked-by: Amitkumar Karwar <akarwar-eYqpPyKDWXRBDgjK7y7TUQ@public.gmane.org>
The commit log doesn't give any background info. Does this fix a bug or
why is it needed?
Patch set to Changes Requested.
--
https://patchwork.kernel.org/patch/9353051/
Documentation about submitting wireless patches and checking status
from patchwork:
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
^ permalink raw reply
* Re: [net-next PATCH] net: netlink messages for HW addr programming
From: Patrick Ruddy @ 2016-10-03 10:42 UTC (permalink / raw)
To: jiri@resnulli.us
Cc: stephen@networkplumber.org, netdev@vger.kernel.org,
davem@davemloft.net, Luca Boccassi, alexander.h.duyck@intel.com,
roopa@cumulusnetworks.com, Sven-Thorsten Dietrich
In-Reply-To: <20160920054959.GB1843@nanopsycho.orion>
On Tue, 2016-09-20 at 07:49 +0200, Jiri Pirko wrote:
> Tue, Sep 20, 2016 at 07:31:27AM CEST, roopa@cumulusnetworks.com wrote:
> >On 9/19/16, 7:46 AM, Patrick Ruddy wrote:
> >> On Sun, 2016-09-18 at 07:51 -0700, Roopa Prabhu wrote:
> >>> On 9/15/16, 9:48 AM, Patrick Ruddy wrote:
> >>>> Add RTM_NEWADDR and RTM_DELADDR netlink messages with family
> >>>> AF_UNSPEC to indicate interest in specific unicast and multicast
> >>>> hardware addresses. These messages are sent when addresses are
> >>>> added or deleted from the appropriate interface driver.
> >>>> Added AF_UNSPEC GETADDR function to allow the netlink notifications
> >>>> to be replayed to avoid loss of state due to application start
> >>>> ordering or restart.
> >>>>
> >>>> Signed-off-by: Patrick Ruddy <pruddy@brocade.com>
> >>>> ---
> >>> RTM_NEWADDR and RTM_DELADDR are not used to add these entries to the kernel.
> >>> so, it seems a bit wrong to use RTM_NEWADDR and RTM_DELADDR to notify them to
> >>> userspace and also to request a special dump of these addresses.
> >>>
> >>> This could just be a new nested netlink attribute in the existing link dump ?
> >> Hi Roopa
> >>
> >> Thanks for the review. I did initially code this using NEW/DEL/GET_LINK
> >> messages but was asked to change to to ADDR messages by Stephen
> >> Hemminger (cc'd).
> >>
> >> However I agree that these addresses fall between the LINK and ADDR
> >> areas so I'm happy to change this if we can reach some consensus on the
> >> format.
> >>
> >ok, thanks for the history. yes, they do lie in a weird spot.
>
> They are l2 addresses, they should be threated accordingly. Am I missing
> something?
In looking to rework this I remembered something. One of the plus sides
of using the ADDR messages to associate addresses with the device is
that we can use the NEW/DELADDR messages to signify the addition and
deletion of these addresses individually as they come and go, which they
do all the time e.g with IGMP joins and leaves. If we embed these
addresses within the LINK message we cannot use NEWLINK and DELLINK to
add and remove individual addresses as that would signify the link
coming and going which is not the case. SETLINK might work but then the
onus is on the netlink client to be able to work out the address
additions/deletions based on it's own state.
>
> >the general convention for other rtnl registrations seems to be
> >AF_UNSPEC family means include all supported families. thats where this seems a bit odd.
> >
> >On the other hand, one reason I see where using RTM_*ADDR will be useful for this is if we wanted
> >to provide a way to add these uc and mc address via ip addr add in the future.
> >ip addr add <lladdr> dev eth0
> >
> >Does this patch allow that in the future ?
>
> This shoul go under ip link I believe. "ip addr" is for l3.
>
>
> >
> >also, will these l2 addresses now show up in 'ip addr show' output ?.
> >
I think the us ones will display in ip addr show and the mutlicast ones
in ip maddr show - this is the case today (i.e. without my changes)
thanks
-pr
^ permalink raw reply
* Re: pull-request: mac80211-next 2016-09-30
From: Johannes Berg @ 2016-10-03 10:07 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-wireless
In-Reply-To: <20161003.015401.1915647209236695770.davem@davemloft.net>
On Mon, 2016-10-03 at 01:54 -0400, David Miller wrote:
>
> Could you please look into this and send me a new pull request with
> the conflicts resolved?
>
Sure, will do (tomorrow, we have a holiday today).
Would you prefer I send you the merge resolution for your tree, or just
pull in net for a new pull request?
johannes
^ permalink raw reply
* Re: [PATCH v4 net-next] net: phy: Add Edge-rate driver for Microsemi PHYs.
From: Raju Lakkaraju @ 2016-10-03 9:23 UTC (permalink / raw)
To: Andrew Lunn
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, devicetree-u79uwXL29TY76Z2rM5mHXA,
f.fainelli-Re5JQEeQqe8AvxtiuMwx3w,
Allan.Nielsen-dzo6w/eZyo2tG0bUXCXiUA
In-Reply-To: <20161003075656.GC18219-g2DYL2Zd6BY@public.gmane.org>
Hi Andrew,
On Mon, Oct 03, 2016 at 09:56:56AM +0200, Andrew Lunn wrote:
> EXTERNAL EMAIL
>
>
> > > > + vsc8531_0: ethernet-phy@0 {
> > > > + compatible = "ethernet-phy-id0007.0570";
> > > > + vsc8531,vddmac = /bits/ 16 <MSCC_VDDMAC_3300>;
> > > > + vsc8531,edge-slowdown = /bits/ 8 <17>;
> > >
> > > No, real values please:
> > >
> > > vsc8531,vddmac = <2000>;
> > > vsc8531,edge-slowdown = <21>;
> > >
> > > The driver should then do the maths to figure out the nearest magic
> > > value to write to the register, or complain the settings are out of
> > > range with an -EINVAL.
>
> I think you missed my point. Notice my example does not use values
> from the table. Is there a reason not to use the PHY with 2000mv? Or
> does it break when you don't use one of the 4 listed voltages?
>
As per PHY data sheet, It should use one of the voltage in list.
i.e. 3.3V (3300 mV) or 2.5V (2500 mV) or 1.8V (1800 mV) or 1.5V (1500).
Other voltages are not accepted. Drivr considers default as 3.3V (3300 mV).
> I was wanting you to implement some formula, which given the voltage
> and percentage slowdown, returns the magic number. Hence the comment:
>
Based on Voltage selection and percentage slowdown, Driver pick the
magic number from Edge rate table.
> > > FYI: No floating point maths are allowed in the kernel.
>
> You need to do integer arithmetic, which is why i suggested mV, not V.
>
I understood your suggestion and used voltage in milli Volts.
> Andrew
Thanks,
Raju.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* [PATCH 1/3] net: smc91x: isolate u16 writes alignment workaround
From: Robert Jarzmik @ 2016-10-03 9:05 UTC (permalink / raw)
To: Rob Herring, Mark Rutland, Nicolas Pitre,
Russell King - ARM Linux, Arnd Bergmann
Cc: netdev, devicetree, linux-kernel, Robert Jarzmik
Writes to u16 has a special handling on 3 PXA platforms, where the
hardware wiring forces these writes to be u32 aligned.
This patch isolates this handling for PXA platforms as before, but
enables this "workaround" to be set up dynamically, which will be the
case in device-tree build types.
This patch was tested on 2 PXA platforms : mainstone, which relies on
the workaround, and lubbock, which doesn't.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
---
drivers/net/ethernet/smsc/smc91x.c | 6 ++-
drivers/net/ethernet/smsc/smc91x.h | 80 ++++++++++++++++++++++----------------
2 files changed, 52 insertions(+), 34 deletions(-)
diff --git a/drivers/net/ethernet/smsc/smc91x.c b/drivers/net/ethernet/smsc/smc91x.c
index 77ad2a3f59db..b1f74e06d98e 100644
--- a/drivers/net/ethernet/smsc/smc91x.c
+++ b/drivers/net/ethernet/smsc/smc91x.c
@@ -602,7 +602,8 @@ static void smc_hardware_send_pkt(unsigned long data)
SMC_PUSH_DATA(lp, buf, len & ~1);
/* Send final ctl word with the last byte if there is one */
- SMC_outw(((len & 1) ? (0x2000 | buf[len-1]) : 0), ioaddr, DATA_REG(lp));
+ SMC_outw(lp, ((len & 1) ? (0x2000 | buf[len-1]) : 0), ioaddr,
+ DATA_REG(lp));
/*
* If THROTTLE_TX_PKTS is set, we stop the queue here. This will
@@ -2276,6 +2277,9 @@ static int smc_drv_probe(struct platform_device *pdev)
memcpy(&lp->cfg, pd, sizeof(lp->cfg));
lp->io_shift = SMC91X_IO_SHIFT(lp->cfg.flags);
}
+ lp->half_word_align4 =
+ machine_is_mainstone() || machine_is_stargate2() ||
+ machine_is_pxa_idp();
#if IS_BUILTIN(CONFIG_OF)
match = of_match_device(of_match_ptr(smc91x_match), &pdev->dev);
diff --git a/drivers/net/ethernet/smsc/smc91x.h b/drivers/net/ethernet/smsc/smc91x.h
index 1a55c7976df0..2b7752db8635 100644
--- a/drivers/net/ethernet/smsc/smc91x.h
+++ b/drivers/net/ethernet/smsc/smc91x.h
@@ -66,10 +66,10 @@
#define SMC_IRQ_FLAGS (-1) /* from resource */
/* We actually can't write halfwords properly if not word aligned */
-static inline void SMC_outw(u16 val, void __iomem *ioaddr, int reg)
+static inline void _SMC_outw_align4(u16 val, void __iomem *ioaddr, int reg,
+ bool use_align4_workaround)
{
- if ((machine_is_mainstone() || machine_is_stargate2() ||
- machine_is_pxa_idp()) && reg & 2) {
+ if (use_align4_workaround) {
unsigned int v = val << 16;
v |= readl(ioaddr + (reg & ~2)) & 0xffff;
writel(v, ioaddr + (reg & ~2));
@@ -78,6 +78,12 @@ static inline void SMC_outw(u16 val, void __iomem *ioaddr, int reg)
}
}
+#define SMC_outw(lp, v, a, r) \
+ _SMC_outw_align4((v), (a), (r), \
+ IS_BUILTIN(CONFIG_ARCH_PXA) && ((r) & 2) && \
+ lp->half_word_align4)
+
+
#elif defined(CONFIG_SH_SH4202_MICRODEV)
#define SMC_CAN_USE_8BIT 0
@@ -88,7 +94,8 @@ static inline void SMC_outw(u16 val, void __iomem *ioaddr, int reg)
#define SMC_inw(a, r) inw((a) + (r) - 0xa0000000)
#define SMC_inl(a, r) inl((a) + (r) - 0xa0000000)
#define SMC_outb(v, a, r) outb(v, (a) + (r) - 0xa0000000)
-#define SMC_outw(v, a, r) outw(v, (a) + (r) - 0xa0000000)
+#define _SMC_outw(v, a, r) outw(v, (a) + (r) - 0xa0000000)
+#define SMC_outw(lp, v, a, r) _SMC_outw((v), (a), (r))
#define SMC_outl(v, a, r) outl(v, (a) + (r) - 0xa0000000)
#define SMC_insl(a, r, p, l) insl((a) + (r) - 0xa0000000, p, l)
#define SMC_outsl(a, r, p, l) outsl((a) + (r) - 0xa0000000, p, l)
@@ -106,7 +113,8 @@ static inline void SMC_outw(u16 val, void __iomem *ioaddr, int reg)
#define SMC_inb(a, r) inb(((u32)a) + (r))
#define SMC_inw(a, r) inw(((u32)a) + (r))
#define SMC_outb(v, a, r) outb(v, ((u32)a) + (r))
-#define SMC_outw(v, a, r) outw(v, ((u32)a) + (r))
+#define _SMC_outw(v, a, r) outw(v, ((u32)a) + (r))
+#define SMC_outw(lp, v, a, r) _SMC_outw((v), (a), (r))
#define SMC_insw(a, r, p, l) insw(((u32)a) + (r), p, l)
#define SMC_outsw(a, r, p, l) outsw(((u32)a) + (r), p, l)
@@ -134,7 +142,8 @@ static inline void SMC_outw(u16 val, void __iomem *ioaddr, int reg)
#define SMC_inw(a, r) readw((a) + (r))
#define SMC_inl(a, r) readl((a) + (r))
#define SMC_outb(v, a, r) writeb(v, (a) + (r))
-#define SMC_outw(v, a, r) writew(v, (a) + (r))
+#define _SMC_outw(v, a, r) writew(v, (a) + (r))
+#define SMC_outw(lp, v, a, r) _SMC_outw((v), (a), (r))
#define SMC_outl(v, a, r) writel(v, (a) + (r))
#define SMC_insw(a, r, p, l) readsw((a) + (r), p, l)
#define SMC_outsw(a, r, p, l) writesw((a) + (r), p, l)
@@ -166,7 +175,8 @@ static inline void mcf_outsw(void *a, unsigned char *p, int l)
}
#define SMC_inw(a, r) _swapw(readw((a) + (r)))
-#define SMC_outw(v, a, r) writew(_swapw(v), (a) + (r))
+#define _SMC_outw(v, a, r) writew(_swapw(v), (a) + (r))
+#define SMC_outw(lp, v, a, r) _SMC_outw((v), (a), (r))
#define SMC_insw(a, r, p, l) mcf_insw(a + r, p, l)
#define SMC_outsw(a, r, p, l) mcf_outsw(a + r, p, l)
@@ -200,7 +210,8 @@ static inline void mcf_outsw(void *a, unsigned char *p, int l)
#define SMC_inw(a, r) ioread16((a) + (r))
#define SMC_inl(a, r) ioread32((a) + (r))
#define SMC_outb(v, a, r) iowrite8(v, (a) + (r))
-#define SMC_outw(v, a, r) iowrite16(v, (a) + (r))
+#define _SMC_outw(v, a, r) iowrite16(v, (a) + (r))
+#define SMC_outw(lp, v, a, r) _SMC_outw((v), (a), (r))
#define SMC_outl(v, a, r) iowrite32(v, (a) + (r))
#define SMC_insw(a, r, p, l) ioread16_rep((a) + (r), p, l)
#define SMC_outsw(a, r, p, l) iowrite16_rep((a) + (r), p, l)
@@ -262,6 +273,8 @@ struct smc_local {
/* the low address lines on some platforms aren't connected... */
int io_shift;
+ /* on some platforms a u16 write must be 4-bytes aligned */
+ bool half_word_align4;
struct smc91x_platdata cfg;
};
@@ -420,12 +433,13 @@ smc_pxa_dma_insw(void __iomem *ioaddr, struct smc_local *lp, int reg, int dma,
* Any 16-bit access is performed with two 8-bit accesses if the hardware
* can't do it directly. Most registers are 16-bit so those are mandatory.
*/
-#define SMC_outw(x, ioaddr, reg) \
+#define _SMC_outw(x, ioaddr, reg) \
do { \
unsigned int __val16 = (x); \
SMC_outb( __val16, ioaddr, reg ); \
SMC_outb( __val16 >> 8, ioaddr, reg + (1 << SMC_IO_SHIFT));\
} while (0)
+#define SMC_outw(lp, v, a, r) _SMC_outw((v), (a), (r))
#define SMC_inw(ioaddr, reg) \
({ \
unsigned int __val16; \
@@ -882,7 +896,7 @@ static const char * chip_ids[ 16 ] = {
else if (SMC_8BIT(lp)) \
SMC_outb(x, ioaddr, PN_REG(lp)); \
else \
- SMC_outw(x, ioaddr, PN_REG(lp)); \
+ SMC_outw(lp, x, ioaddr, PN_REG(lp)); \
} while (0)
#define SMC_GET_AR(lp) \
@@ -910,7 +924,7 @@ static const char * chip_ids[ 16 ] = {
int __mask; \
local_irq_save(__flags); \
__mask = SMC_inw(ioaddr, INT_REG(lp)) & ~0xff; \
- SMC_outw(__mask | (x), ioaddr, INT_REG(lp)); \
+ SMC_outw(lp, __mask | (x), ioaddr, INT_REG(lp)); \
local_irq_restore(__flags); \
} \
} while (0)
@@ -924,7 +938,7 @@ static const char * chip_ids[ 16 ] = {
if (SMC_8BIT(lp)) \
SMC_outb(x, ioaddr, IM_REG(lp)); \
else \
- SMC_outw((x) << 8, ioaddr, INT_REG(lp)); \
+ SMC_outw(lp, (x) << 8, ioaddr, INT_REG(lp)); \
} while (0)
#define SMC_CURRENT_BANK(lp) SMC_inw(ioaddr, BANK_SELECT)
@@ -934,22 +948,22 @@ static const char * chip_ids[ 16 ] = {
if (SMC_MUST_ALIGN_WRITE(lp)) \
SMC_outl((x)<<16, ioaddr, 12<<SMC_IO_SHIFT); \
else \
- SMC_outw(x, ioaddr, BANK_SELECT); \
+ SMC_outw(lp, x, ioaddr, BANK_SELECT); \
} while (0)
#define SMC_GET_BASE(lp) SMC_inw(ioaddr, BASE_REG(lp))
-#define SMC_SET_BASE(lp, x) SMC_outw(x, ioaddr, BASE_REG(lp))
+#define SMC_SET_BASE(lp, x) SMC_outw(lp, x, ioaddr, BASE_REG(lp))
#define SMC_GET_CONFIG(lp) SMC_inw(ioaddr, CONFIG_REG(lp))
-#define SMC_SET_CONFIG(lp, x) SMC_outw(x, ioaddr, CONFIG_REG(lp))
+#define SMC_SET_CONFIG(lp, x) SMC_outw(lp, x, ioaddr, CONFIG_REG(lp))
#define SMC_GET_COUNTER(lp) SMC_inw(ioaddr, COUNTER_REG(lp))
#define SMC_GET_CTL(lp) SMC_inw(ioaddr, CTL_REG(lp))
-#define SMC_SET_CTL(lp, x) SMC_outw(x, ioaddr, CTL_REG(lp))
+#define SMC_SET_CTL(lp, x) SMC_outw(lp, x, ioaddr, CTL_REG(lp))
#define SMC_GET_MII(lp) SMC_inw(ioaddr, MII_REG(lp))
@@ -960,18 +974,18 @@ static const char * chip_ids[ 16 ] = {
if (SMC_MUST_ALIGN_WRITE(lp)) \
SMC_outl((x)<<16, ioaddr, SMC_REG(lp, 8, 1)); \
else \
- SMC_outw(x, ioaddr, GP_REG(lp)); \
+ SMC_outw(lp, x, ioaddr, GP_REG(lp)); \
} while (0)
-#define SMC_SET_MII(lp, x) SMC_outw(x, ioaddr, MII_REG(lp))
+#define SMC_SET_MII(lp, x) SMC_outw(lp, x, ioaddr, MII_REG(lp))
#define SMC_GET_MIR(lp) SMC_inw(ioaddr, MIR_REG(lp))
-#define SMC_SET_MIR(lp, x) SMC_outw(x, ioaddr, MIR_REG(lp))
+#define SMC_SET_MIR(lp, x) SMC_outw(lp, x, ioaddr, MIR_REG(lp))
#define SMC_GET_MMU_CMD(lp) SMC_inw(ioaddr, MMU_CMD_REG(lp))
-#define SMC_SET_MMU_CMD(lp, x) SMC_outw(x, ioaddr, MMU_CMD_REG(lp))
+#define SMC_SET_MMU_CMD(lp, x) SMC_outw(lp, x, ioaddr, MMU_CMD_REG(lp))
#define SMC_GET_FIFO(lp) SMC_inw(ioaddr, FIFO_REG(lp))
@@ -982,14 +996,14 @@ static const char * chip_ids[ 16 ] = {
if (SMC_MUST_ALIGN_WRITE(lp)) \
SMC_outl((x)<<16, ioaddr, SMC_REG(lp, 4, 2)); \
else \
- SMC_outw(x, ioaddr, PTR_REG(lp)); \
+ SMC_outw(lp, x, ioaddr, PTR_REG(lp)); \
} while (0)
#define SMC_GET_EPH_STATUS(lp) SMC_inw(ioaddr, EPH_STATUS_REG(lp))
#define SMC_GET_RCR(lp) SMC_inw(ioaddr, RCR_REG(lp))
-#define SMC_SET_RCR(lp, x) SMC_outw(x, ioaddr, RCR_REG(lp))
+#define SMC_SET_RCR(lp, x) SMC_outw(lp, x, ioaddr, RCR_REG(lp))
#define SMC_GET_REV(lp) SMC_inw(ioaddr, REV_REG(lp))
@@ -1000,12 +1014,12 @@ static const char * chip_ids[ 16 ] = {
if (SMC_MUST_ALIGN_WRITE(lp)) \
SMC_outl((x)<<16, ioaddr, SMC_REG(lp, 8, 0)); \
else \
- SMC_outw(x, ioaddr, RPC_REG(lp)); \
+ SMC_outw(lp, x, ioaddr, RPC_REG(lp)); \
} while (0)
#define SMC_GET_TCR(lp) SMC_inw(ioaddr, TCR_REG(lp))
-#define SMC_SET_TCR(lp, x) SMC_outw(x, ioaddr, TCR_REG(lp))
+#define SMC_SET_TCR(lp, x) SMC_outw(lp, x, ioaddr, TCR_REG(lp))
#ifndef SMC_GET_MAC_ADDR
#define SMC_GET_MAC_ADDR(lp, addr) \
@@ -1022,18 +1036,18 @@ static const char * chip_ids[ 16 ] = {
#define SMC_SET_MAC_ADDR(lp, addr) \
do { \
- SMC_outw(addr[0]|(addr[1] << 8), ioaddr, ADDR0_REG(lp)); \
- SMC_outw(addr[2]|(addr[3] << 8), ioaddr, ADDR1_REG(lp)); \
- SMC_outw(addr[4]|(addr[5] << 8), ioaddr, ADDR2_REG(lp)); \
+ SMC_outw(lp, addr[0]|(addr[1] << 8), ioaddr, ADDR0_REG(lp)); \
+ SMC_outw(lp, addr[2]|(addr[3] << 8), ioaddr, ADDR1_REG(lp)); \
+ SMC_outw(lp, addr[4]|(addr[5] << 8), ioaddr, ADDR2_REG(lp)); \
} while (0)
#define SMC_SET_MCAST(lp, x) \
do { \
const unsigned char *mt = (x); \
- SMC_outw(mt[0] | (mt[1] << 8), ioaddr, MCAST_REG1(lp)); \
- SMC_outw(mt[2] | (mt[3] << 8), ioaddr, MCAST_REG2(lp)); \
- SMC_outw(mt[4] | (mt[5] << 8), ioaddr, MCAST_REG3(lp)); \
- SMC_outw(mt[6] | (mt[7] << 8), ioaddr, MCAST_REG4(lp)); \
+ SMC_outw(lp, mt[0] | (mt[1] << 8), ioaddr, MCAST_REG1(lp)); \
+ SMC_outw(lp, mt[2] | (mt[3] << 8), ioaddr, MCAST_REG2(lp)); \
+ SMC_outw(lp, mt[4] | (mt[5] << 8), ioaddr, MCAST_REG3(lp)); \
+ SMC_outw(lp, mt[6] | (mt[7] << 8), ioaddr, MCAST_REG4(lp)); \
} while (0)
#define SMC_PUT_PKT_HDR(lp, status, length) \
@@ -1042,8 +1056,8 @@ static const char * chip_ids[ 16 ] = {
SMC_outl((status) | (length)<<16, ioaddr, \
DATA_REG(lp)); \
else { \
- SMC_outw(status, ioaddr, DATA_REG(lp)); \
- SMC_outw(length, ioaddr, DATA_REG(lp)); \
+ SMC_outw(lp, status, ioaddr, DATA_REG(lp)); \
+ SMC_outw(lp, length, ioaddr, DATA_REG(lp)); \
} \
} while (0)
--
2.1.4
^ permalink raw reply related
* [PATCH 3/3] net: smsc911x: add u16 workaround for pxa platforms
From: Robert Jarzmik @ 2016-10-03 9:05 UTC (permalink / raw)
To: Rob Herring, Mark Rutland, Nicolas Pitre,
Russell King - ARM Linux, Arnd Bergmann
Cc: netdev, devicetree, linux-kernel, Robert Jarzmik
In-Reply-To: <1475485553-18747-1-git-send-email-robert.jarzmik@free.fr>
Add a workaround for mainstone, idp and stargate2 boards, for u16 writes
which must be aligned on 32 bits addresses.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
---
Documentation/devicetree/bindings/net/smsc911x.txt | 2 ++
1 file changed, 2 insertions(+)
diff --git a/Documentation/devicetree/bindings/net/smsc911x.txt b/Documentation/devicetree/bindings/net/smsc911x.txt
index 3fed3c124411..224965b7453c 100644
--- a/Documentation/devicetree/bindings/net/smsc911x.txt
+++ b/Documentation/devicetree/bindings/net/smsc911x.txt
@@ -13,6 +13,8 @@ Optional properties:
- reg-io-width : Specify the size (in bytes) of the IO accesses that
should be performed on the device. Valid value for SMSC LAN is
2 or 4. If it's omitted or invalid, the size would be 2.
+- reg-u16-align4 : Boolean, put in place the workaround the force all
+ u16 writes to be 32 bits aligned
- smsc,irq-active-high : Indicates the IRQ polarity is active-high
- smsc,irq-push-pull : Indicates the IRQ type is push-pull
- smsc,force-internal-phy : Forces SMSC LAN controller to use
--
2.1.4
^ permalink raw reply related
* [PATCH 2/3] net: smc91x: take into account half-word workaround
From: Robert Jarzmik @ 2016-10-03 9:05 UTC (permalink / raw)
To: Rob Herring, Mark Rutland, Nicolas Pitre,
Russell King - ARM Linux, Arnd Bergmann
Cc: netdev-u79uwXL29TY76Z2rM5mHXA, devicetree-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA, Robert Jarzmik
In-Reply-To: <1475485553-18747-1-git-send-email-robert.jarzmik-GANU6spQydw@public.gmane.org>
For device-tree builds, platforms such as mainstone, idp and stargate2
must have their u16 writes all aligned on 32 bit boundaries. This is
already enabled in platform data builds, and this patch adds it to
device-tree builds.
Signed-off-by: Robert Jarzmik <robert.jarzmik-GANU6spQydw@public.gmane.org>
---
drivers/net/ethernet/smsc/smc91x.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ethernet/smsc/smc91x.c b/drivers/net/ethernet/smsc/smc91x.c
index b1f74e06d98e..b081d5ca88ce 100644
--- a/drivers/net/ethernet/smsc/smc91x.c
+++ b/drivers/net/ethernet/smsc/smc91x.c
@@ -2323,6 +2323,8 @@ static int smc_drv_probe(struct platform_device *pdev)
if (!device_property_read_u32(&pdev->dev, "reg-shift",
&val))
lp->io_shift = val;
+ lp->half_word_align4 =
+ device_property_read_bool(&pdev->dev, "reg-u16-align4");
}
#endif
--
2.1.4
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related
* [PATCH] bluetooth.h: __ variants of u8 and friends are not neccessary inside kernel
From: Pavel Machek @ 2016-10-03 8:56 UTC (permalink / raw)
To: trivial, marcel, gustavo, johan.hedberg, davem, linux-bluetooth,
netdev, linux-kernel
[-- Attachment #1: Type: text/plain, Size: 2842 bytes --]
bluetooth.h is not part of user API, so __ variants are not neccessary
here.
Signed-off-by: Pavel Machek <pavel@ucw.cz>
diff --git a/include/net/bluetooth/bluetooth.h b/include/net/bluetooth/bluetooth.h
index bfd1590..aea0371 100644
--- a/include/net/bluetooth/bluetooth.h
+++ b/include/net/bluetooth/bluetooth.h
@@ -60,8 +60,8 @@
#define BT_SECURITY 4
struct bt_security {
- __u8 level;
- __u8 key_size;
+ u8 level;
+ u8 key_size;
};
#define BT_SECURITY_SDP 0
#define BT_SECURITY_LOW 1
@@ -78,7 +78,7 @@ struct bt_security {
#define BT_POWER 9
struct bt_power {
- __u8 force_active;
+ u8 force_active;
};
#define BT_POWER_FORCE_ACTIVE_OFF 0
#define BT_POWER_FORCE_ACTIVE_ON 1
@@ -112,7 +112,7 @@ struct bt_power {
#define BT_VOICE 11
struct bt_voice {
- __u16 setting;
+ u16 setting;
};
#define BT_VOICE_TRANSPARENT 0x0003
@@ -188,7 +188,7 @@ static inline const char *state_to_string(int state)
/* BD Address */
typedef struct {
- __u8 b[6];
+ u8 b[6];
} __packed bdaddr_t;
/* BD Address type */
@@ -196,7 +196,7 @@ typedef struct {
#define BDADDR_LE_PUBLIC 0x01
#define BDADDR_LE_RANDOM 0x02
-static inline bool bdaddr_type_is_valid(__u8 type)
+static inline bool bdaddr_type_is_valid(u8 type)
{
switch (type) {
case BDADDR_BREDR:
@@ -208,7 +208,7 @@ static inline bool bdaddr_type_is_valid(__u8 type)
return false;
}
-static inline bool bdaddr_type_is_le(__u8 type)
+static inline bool bdaddr_type_is_le(u8 type)
{
switch (type) {
case BDADDR_LE_PUBLIC:
@@ -278,15 +278,16 @@ struct sock *bt_accept_dequeue(struct sock *parent, struct socket *newsock);
/* Skb helpers */
struct l2cap_ctrl {
- __u8 sframe:1,
+ u8 sframe:1,
poll:1,
final:1,
fcs:1,
sar:2,
super:2;
- __u16 reqseq;
- __u16 txseq;
- __u8 retries;
+
+ u16 reqseq;
+ u16 txseq;
+ u8 retries;
__le16 psm;
bdaddr_t bdaddr;
struct l2cap_chan *chan;
@@ -302,7 +303,7 @@ typedef void (*hci_req_complete_skb_t)(struct hci_dev *hdev, u8 status,
#define HCI_REQ_SKB BIT(1)
struct hci_ctrl {
- __u16 opcode;
+ u16 opcode;
u8 req_flags;
u8 req_event;
union {
@@ -312,10 +313,10 @@ struct hci_ctrl {
};
struct bt_skb_cb {
- __u8 pkt_type;
- __u8 force_active;
- __u16 expect;
- __u8 incoming:1;
+ u8 pkt_type;
+ u8 force_active;
+ u16 expect;
+ u8 incoming:1;
union {
struct l2cap_ctrl l2cap;
struct hci_ctrl hci;
@@ -365,7 +366,7 @@ out:
return NULL;
}
-int bt_to_errno(__u16 code);
+int bt_to_errno(u16 code);
void hci_sock_set_flag(struct sock *sk, int nr);
void hci_sock_clear_flag(struct sock *sk, int nr);
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]
^ permalink raw reply related
* Re;一般贸易退税&東莞/深圳/廣州/中山/香港物流
From: Paul Tsang @ 2016-10-03 8:14 UTC (permalink / raw)
To: netdev
Dear Sir/Miss
只要贵司拥有一般纳税人资质和能开17%的增值税专用发票，我司便能为贵司做退税；
1.出口退税条件：①真正有生产经营能力的企业而不是空壳企业②需代办出口退税的企业③保证能开增值税专用发票的企业④保证企业开的增值税专用发票的合法，合理而不是虚开增值税专用发票
2.货物出口；工厂须提供产品名称，商品编码，规格，型号，包装方式，单价，总价及装箱单.
3.结汇：①货物出口后两个月内（不能超过两个月）必须进行结汇核销。工厂通知客户（外国）及时结汇，以便我司及时到外汇管理局进行外汇核销工作，否则会影响我司的运作及出口货物的退税工作。②我司收到汇款后按银行当日银行牌价，24小时内汇至工厂账户
4.开票：①我司汇出货款后，将〈汇款单〉、〈开票明细表〉、〈购货合同〉EMAIL至工厂财务部②工厂根据〈开票明细表〉开出17%的〈增值税专用发票〉，抬头为我司，发票的商品名称、规格、单位、单价、总价、税金均必须准确无误，发票的价税合计金额与汇款单金额一致。③工厂必须保证开的发票是合理、合法的。④工厂须将开好的增值税专用发票，盖好章的购货合同（一式二份）、盖好章的送货单三份资料给我司。
5.付税款：①待我司经过国税认证成功2个工作日以后将税款汇给工厂。②待我司收到国税的退税款2个工作日将税款汇给工厂③付的退税款是实际退税款扣除我司的代理手续费.
退税额计算方法如下
退税额=（增值税发票金额）/（1+增值税率）*出口退税率。
例如：货值100万人民币，增值税率为17%，退税率为13%，则：
1000000/1.17*0.13=111111.11元人民币
以下報價為常見物流交貨報價，如有需要，歡迎來電詢價
1）散貨出口香港報價；
广东江门/佛山/中山散货速递香港自提價；RMB0.8/KG或RMB200/CBM
广东清远/河源/惠州散货速递香港自提價；RMB0.8/KG或RMB200/CBM
广东汕头散货速递香港自提價；RMB0.9/KG或RMB210/CBM
湖南/江西/湖北散货速递香港自提價；RMB1/KG或RMB230/CBM
浙江/江苏/上海散货速递香港自提價；RMB1.2/KG或RMB250/CBM
河南/山东/安徽散货速递香港自提價；RMB2/KG或RMB280/CBM
京津翼/内蒙古各省散货速递香港自提價；RMB2.5/KG或RMB300/CBM
注：以上費用不包含香港送客、入仓收費.
散貨香港派送收費標準:
A：九龍區： RMB180 ；B:新界區： RMB200 C: 香港區： RMB300
散貨香港入仓收費標準:
香港入仓费；RMB350（登记费、停车费等杂费实报实销）
2）散（杂）貨交深圳盐田港、中外运、八达仓、金运达等外運倉報價；最低消費；RMB800/單 （报关、登记费等仓库杂费实报实销）
注；1.必须携带所属船公司或货运代理之落货纸(Shipping order)一式五份及报关文件（如无落货纸，本公司可拒绝收货）
2.深圳交货会产生哪些费用？
1）报关费：RMB350/单；
2）车辆入场登记费；吨车：RMB70/车 货柜车：RMB130/柜；
3）杂费（biding meterial fee）：RMB5/车 RMB15/柜；
4）代换商检：RMB150/单
3）香港普貨，中檔貨到內地各省市報價；
快件包稅報關入口特點:A.手續簡單,海關監管條件簡單,不用單證不用批文,一些電子設備可以不需3C證明也可入口. B.速度快,次日到貨,适合緊急貨物通關. C.如果貨量較大,可以采用每天分批報關方式. D.快件包稅報關不提供稅票,沒法進行稅票抵扣.
A 类
檔，白纸，纸品，纸袋，纸箱，纸盒，名片，日历，色卡，吊牌，记事簿，小册子，宣传单张，说明书，目录，标贴，海报，胶纸，贴纸，胶条，塑料袋，胶膜，胶扣具，塑料管，胶壳，胶盖，螺丝，水管，铁片，铁条，铁盖，铁壳，铁扣，铁架，鞋底及配件，织带，缎带，胶根带，布标，松紧带，编织带，月历，绳子。
￥9/kg
B 类
塑料接头，塑料喷头，铁芯，铁针，磁铁，弹簧，钢片，钢尺，钢珠，铝片，锡线，钢线，布料，纱线，线球，花边，饰扣，拉链，胶布，海绵，压克力制品，电话线插头，书册，软木，洋烛，墙纸，发泡胶，钛线，木板，铁座，丝带，刷子，泥胶，砂纸，伞配件，塑料玩具。
￥11/kg
C 类
卤轮，滚轮，滑轮，电线，建材，过滤网，干花，鼓纸，砂轮，石材，灯壳，灯罩，发饰品，低价洒水头， PVC 皮，磨石，油石，蕾丝，磨轮，漆包线，焊枝，塑料插头，五金插头，胶粒，铜材，铝材，印刷胶片。
￥13/kg
D 类
牛皮，羊皮，激光纸，皮光纸，烫金纸，轴心，喇叭网，表带，灯泡，灯管，染料，色粉，颜料，胶水，树脂，工业用皮带，陶瓷制品，玻璃制品，雕刻品，洗水丝，文具，绒毛玩具，工业用线圈，背包，尼龙袋，机械油封，袜子，手套，非银制餐具，刀片，刀具，非真皮手袋，帽子，五金工具，开关制，皮带，卡座，口罩，护具，陶瓷片，五金锁，压喉，表壳，表盖，手动工具，机壳，水龙头，铁轴，木家私，清洁剂，机械轴承，烙铁咀，空白线路板，脚踏车零件
￥15/kg
E 类
低价值运动用品，电筒，低价值灯饰，电制，衣服，鞋样，模具，圣诞灯饰，灯具，地毯，计算机接线，皮家私，滤蕊，灯箱，空墨盒，空碳鼓，保温杯，镜片，喷枪，油墨，厨具，电路开关制，研磨膏，焊锡膏，润滑油，相机用胶壳，胶按钮，钻咀，电器用轴承，电箱，端子，普通电池，切纸机。
￥16/kg
F 类
加热棒，低值变压器，电表，耳筒，耳机配件，气压配件，泵，空压表，活塞环，电热管，工业用电源开关，低值机械零配件，气动配件，低值充电器，散热扇零件，丝巾，领带，针车配件，压力表，油压表，气压表，发热线，车针，机械阀门，太阳膜， LED 小电灯，鱼竿，滑板车，电毯，辗刀和零件。
￥21/kg
G 类
二极管，三极管，电容，电阻，电动起子，温度计 ( 电子 / 液体 ) ，低值乐器，风扇，碳粉，电动工具，传声机，麦克风，计数机，电子数字板，暖风机， D/V/CD 外壳，干手机，耳咪线，投币器配件，游戏机控制杆，过滤器，眼镜，电磁阀，时钟，医疗卫生品，磨沙机，假发，继电器，电话机。
￥36/kg
H 类
低值液晶显示片，温控器，烟雾感应器，音箱，定时器，低值缝纫机，助听器，探头，机器用气缸，电子密码锁，感应器，变频器，变速器，健身器材，工业用变压器，冶具，卡片阅读机，钟表配件，手表配件，舞台射灯。
￥41/kg
I 类
打印机，扫瞄仪，计算机机箱，显示器，路由器，集线器，接线盒，转换器，遥控器，传真机，烧录器， IC ，线路板，频率器，计算机配件，医疗器材，消防器材及配件，钮扣电池，菲琳，相机，乐器，功放。
￥53-83/kg
BEST REGARDS
曾生
HAOCHING INT'L LOGISTICS LTD
(M.P)86-13430873117
(TEL)852-31763891
(微信)13430873117
(QQ)2926442692
(SKYPE)HAOCHENGHK1987
(E-MAIL)HAOCHENGHK1987@163.COM
^ permalink raw reply
* Re: [PATCH v4 net-next] net: phy: Add Edge-rate driver for Microsemi PHYs.
From: Andrew Lunn @ 2016-10-03 7:56 UTC (permalink / raw)
To: Raju Lakkaraju; +Cc: netdev, devicetree, f.fainelli, Allan.Nielsen
In-Reply-To: <20161003072043.GA24259@microsemi.com>
> > > + vsc8531_0: ethernet-phy@0 {
> > > + compatible = "ethernet-phy-id0007.0570";
> > > + vsc8531,vddmac = /bits/ 16 <MSCC_VDDMAC_3300>;
> > > + vsc8531,edge-slowdown = /bits/ 8 <17>;
> >
> > No, real values please:
> >
> > vsc8531,vddmac = <2000>;
> > vsc8531,edge-slowdown = <21>;
> >
> > The driver should then do the maths to figure out the nearest magic
> > value to write to the register, or complain the settings are out of
> > range with an -EINVAL.
I think you missed my point. Notice my example does not use values
from the table. Is there a reason not to use the PHY with 2000mv? Or
does it break when you don't use one of the 4 listed voltages?
I was wanting you to implement some formula, which given the voltage
and percentage slowdown, returns the magic number. Hence the comment:
> > FYI: No floating point maths are allowed in the kernel.
You need to do integer arithmetic, which is why i suggested mV, not V.
Andrew
^ permalink raw reply
* [PATCH net-next 1/7] xen-netback: separate guest side rx code into separate module
From: Paul Durrant @ 2016-10-03 7:31 UTC (permalink / raw)
To: netdev, xen-devel; +Cc: Paul Durrant, Wei Liu
In-Reply-To: <1475479872-23717-1-git-send-email-paul.durrant@citrix.com>
The netback source module has become very large and somewhat confusing.
This patch simply moves all code related to the backend to frontend (i.e
guest side rx) data-path into a separate rx source module.
This patch contains no functional change, it is code movement and
minimal changes to avoid patch style-check issues.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
Cc: Wei Liu <wei.liu2@citrix.com>
---
drivers/net/xen-netback/Makefile | 2 +-
drivers/net/xen-netback/netback.c | 754 ------------------------------------
drivers/net/xen-netback/rx.c | 789 ++++++++++++++++++++++++++++++++++++++
3 files changed, 790 insertions(+), 755 deletions(-)
create mode 100644 drivers/net/xen-netback/rx.c
diff --git a/drivers/net/xen-netback/Makefile b/drivers/net/xen-netback/Makefile
index 11e02be..d49798a 100644
--- a/drivers/net/xen-netback/Makefile
+++ b/drivers/net/xen-netback/Makefile
@@ -1,3 +1,3 @@
obj-$(CONFIG_XEN_NETDEV_BACKEND) := xen-netback.o
-xen-netback-y := netback.o xenbus.o interface.o hash.o
+xen-netback-y := netback.o xenbus.o interface.o hash.o rx.o
diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c
index edbae0b..1f9d92e 100644
--- a/drivers/net/xen-netback/netback.c
+++ b/drivers/net/xen-netback/netback.c
@@ -106,13 +106,6 @@ static void push_tx_responses(struct xenvif_queue *queue);
static inline int tx_work_todo(struct xenvif_queue *queue);
-static struct xen_netif_rx_response *make_rx_response(struct xenvif_queue *queue,
- u16 id,
- s8 st,
- u16 offset,
- u16 size,
- u16 flags);
-
static inline unsigned long idx_to_pfn(struct xenvif_queue *queue,
u16 idx)
{
@@ -155,571 +148,11 @@ static inline pending_ring_idx_t pending_index(unsigned i)
return i & (MAX_PENDING_REQS-1);
}
-static bool xenvif_rx_ring_slots_available(struct xenvif_queue *queue)
-{
- RING_IDX prod, cons;
- struct sk_buff *skb;
- int needed;
-
- skb = skb_peek(&queue->rx_queue);
- if (!skb)
- return false;
-
- needed = DIV_ROUND_UP(skb->len, XEN_PAGE_SIZE);
- if (skb_is_gso(skb))
- needed++;
- if (skb->sw_hash)
- needed++;
-
- do {
- prod = queue->rx.sring->req_prod;
- cons = queue->rx.req_cons;
-
- if (prod - cons >= needed)
- return true;
-
- queue->rx.sring->req_event = prod + 1;
-
- /* Make sure event is visible before we check prod
- * again.
- */
- mb();
- } while (queue->rx.sring->req_prod != prod);
-
- return false;
-}
-
-void xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb)
-{
- unsigned long flags;
-
- spin_lock_irqsave(&queue->rx_queue.lock, flags);
-
- __skb_queue_tail(&queue->rx_queue, skb);
-
- queue->rx_queue_len += skb->len;
- if (queue->rx_queue_len > queue->rx_queue_max)
- netif_tx_stop_queue(netdev_get_tx_queue(queue->vif->dev, queue->id));
-
- spin_unlock_irqrestore(&queue->rx_queue.lock, flags);
-}
-
-static struct sk_buff *xenvif_rx_dequeue(struct xenvif_queue *queue)
-{
- struct sk_buff *skb;
-
- spin_lock_irq(&queue->rx_queue.lock);
-
- skb = __skb_dequeue(&queue->rx_queue);
- if (skb)
- queue->rx_queue_len -= skb->len;
-
- spin_unlock_irq(&queue->rx_queue.lock);
-
- return skb;
-}
-
-static void xenvif_rx_queue_maybe_wake(struct xenvif_queue *queue)
-{
- spin_lock_irq(&queue->rx_queue.lock);
-
- if (queue->rx_queue_len < queue->rx_queue_max)
- netif_tx_wake_queue(netdev_get_tx_queue(queue->vif->dev, queue->id));
-
- spin_unlock_irq(&queue->rx_queue.lock);
-}
-
-
-static void xenvif_rx_queue_purge(struct xenvif_queue *queue)
-{
- struct sk_buff *skb;
- while ((skb = xenvif_rx_dequeue(queue)) != NULL)
- kfree_skb(skb);
-}
-
-static void xenvif_rx_queue_drop_expired(struct xenvif_queue *queue)
-{
- struct sk_buff *skb;
-
- for(;;) {
- skb = skb_peek(&queue->rx_queue);
- if (!skb)
- break;
- if (time_before(jiffies, XENVIF_RX_CB(skb)->expires))
- break;
- xenvif_rx_dequeue(queue);
- kfree_skb(skb);
- }
-}
-
-struct netrx_pending_operations {
- unsigned copy_prod, copy_cons;
- unsigned meta_prod, meta_cons;
- struct gnttab_copy *copy;
- struct xenvif_rx_meta *meta;
- int copy_off;
- grant_ref_t copy_gref;
-};
-
-static struct xenvif_rx_meta *get_next_rx_buffer(struct xenvif_queue *queue,
- struct netrx_pending_operations *npo)
-{
- struct xenvif_rx_meta *meta;
- struct xen_netif_rx_request req;
-
- RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
-
- meta = npo->meta + npo->meta_prod++;
- meta->gso_type = XEN_NETIF_GSO_TYPE_NONE;
- meta->gso_size = 0;
- meta->size = 0;
- meta->id = req.id;
-
- npo->copy_off = 0;
- npo->copy_gref = req.gref;
-
- return meta;
-}
-
-struct gop_frag_copy {
- struct xenvif_queue *queue;
- struct netrx_pending_operations *npo;
- struct xenvif_rx_meta *meta;
- int head;
- int gso_type;
- int protocol;
- int hash_present;
-
- struct page *page;
-};
-
-static void xenvif_setup_copy_gop(unsigned long gfn,
- unsigned int offset,
- unsigned int *len,
- struct gop_frag_copy *info)
-{
- struct gnttab_copy *copy_gop;
- struct xen_page_foreign *foreign;
- /* Convenient aliases */
- struct xenvif_queue *queue = info->queue;
- struct netrx_pending_operations *npo = info->npo;
- struct page *page = info->page;
-
- BUG_ON(npo->copy_off > MAX_BUFFER_OFFSET);
-
- if (npo->copy_off == MAX_BUFFER_OFFSET)
- info->meta = get_next_rx_buffer(queue, npo);
-
- if (npo->copy_off + *len > MAX_BUFFER_OFFSET)
- *len = MAX_BUFFER_OFFSET - npo->copy_off;
-
- copy_gop = npo->copy + npo->copy_prod++;
- copy_gop->flags = GNTCOPY_dest_gref;
- copy_gop->len = *len;
-
- foreign = xen_page_foreign(page);
- if (foreign) {
- copy_gop->source.domid = foreign->domid;
- copy_gop->source.u.ref = foreign->gref;
- copy_gop->flags |= GNTCOPY_source_gref;
- } else {
- copy_gop->source.domid = DOMID_SELF;
- copy_gop->source.u.gmfn = gfn;
- }
- copy_gop->source.offset = offset;
-
- copy_gop->dest.domid = queue->vif->domid;
- copy_gop->dest.offset = npo->copy_off;
- copy_gop->dest.u.ref = npo->copy_gref;
-
- npo->copy_off += *len;
- info->meta->size += *len;
-
- if (!info->head)
- return;
-
- /* Leave a gap for the GSO descriptor. */
- if ((1 << info->gso_type) & queue->vif->gso_mask)
- queue->rx.req_cons++;
-
- /* Leave a gap for the hash extra segment. */
- if (info->hash_present)
- queue->rx.req_cons++;
-
- info->head = 0; /* There must be something in this buffer now */
-}
-
-static void xenvif_gop_frag_copy_grant(unsigned long gfn,
- unsigned offset,
- unsigned int len,
- void *data)
-{
- unsigned int bytes;
-
- while (len) {
- bytes = len;
- xenvif_setup_copy_gop(gfn, offset, &bytes, data);
- offset += bytes;
- len -= bytes;
- }
-}
-
-/*
- * Set up the grant operations for this fragment. If it's a flipping
- * interface, we also set up the unmap request from here.
- */
-static void xenvif_gop_frag_copy(struct xenvif_queue *queue, struct sk_buff *skb,
- struct netrx_pending_operations *npo,
- struct page *page, unsigned long size,
- unsigned long offset, int *head)
-{
- struct gop_frag_copy info = {
- .queue = queue,
- .npo = npo,
- .head = *head,
- .gso_type = XEN_NETIF_GSO_TYPE_NONE,
- /* xenvif_set_skb_hash() will have either set a s/w
- * hash or cleared the hash depending on
- * whether the the frontend wants a hash for this skb.
- */
- .hash_present = skb->sw_hash,
- };
- unsigned long bytes;
-
- if (skb_is_gso(skb)) {
- if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
- info.gso_type = XEN_NETIF_GSO_TYPE_TCPV4;
- else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
- info.gso_type = XEN_NETIF_GSO_TYPE_TCPV6;
- }
-
- /* Data must not cross a page boundary. */
- BUG_ON(size + offset > PAGE_SIZE<<compound_order(page));
-
- info.meta = npo->meta + npo->meta_prod - 1;
-
- /* Skip unused frames from start of page */
- page += offset >> PAGE_SHIFT;
- offset &= ~PAGE_MASK;
-
- while (size > 0) {
- BUG_ON(offset >= PAGE_SIZE);
-
- bytes = PAGE_SIZE - offset;
- if (bytes > size)
- bytes = size;
-
- info.page = page;
- gnttab_foreach_grant_in_range(page, offset, bytes,
- xenvif_gop_frag_copy_grant,
- &info);
- size -= bytes;
- offset = 0;
-
- /* Next page */
- if (size) {
- BUG_ON(!PageCompound(page));
- page++;
- }
- }
-
- *head = info.head;
-}
-
-/*
- * Prepare an SKB to be transmitted to the frontend.
- *
- * This function is responsible for allocating grant operations, meta
- * structures, etc.
- *
- * It returns the number of meta structures consumed. The number of
- * ring slots used is always equal to the number of meta slots used
- * plus the number of GSO descriptors used. Currently, we use either
- * zero GSO descriptors (for non-GSO packets) or one descriptor (for
- * frontend-side LRO).
- */
-static int xenvif_gop_skb(struct sk_buff *skb,
- struct netrx_pending_operations *npo,
- struct xenvif_queue *queue)
-{
- struct xenvif *vif = netdev_priv(skb->dev);
- int nr_frags = skb_shinfo(skb)->nr_frags;
- int i;
- struct xen_netif_rx_request req;
- struct xenvif_rx_meta *meta;
- unsigned char *data;
- int head = 1;
- int old_meta_prod;
- int gso_type;
-
- old_meta_prod = npo->meta_prod;
-
- gso_type = XEN_NETIF_GSO_TYPE_NONE;
- if (skb_is_gso(skb)) {
- if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
- gso_type = XEN_NETIF_GSO_TYPE_TCPV4;
- else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
- gso_type = XEN_NETIF_GSO_TYPE_TCPV6;
- }
-
- /* Set up a GSO prefix descriptor, if necessary */
- if ((1 << gso_type) & vif->gso_prefix_mask) {
- RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
- meta = npo->meta + npo->meta_prod++;
- meta->gso_type = gso_type;
- meta->gso_size = skb_shinfo(skb)->gso_size;
- meta->size = 0;
- meta->id = req.id;
- }
-
- RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
- meta = npo->meta + npo->meta_prod++;
-
- if ((1 << gso_type) & vif->gso_mask) {
- meta->gso_type = gso_type;
- meta->gso_size = skb_shinfo(skb)->gso_size;
- } else {
- meta->gso_type = XEN_NETIF_GSO_TYPE_NONE;
- meta->gso_size = 0;
- }
-
- meta->size = 0;
- meta->id = req.id;
- npo->copy_off = 0;
- npo->copy_gref = req.gref;
-
- data = skb->data;
- while (data < skb_tail_pointer(skb)) {
- unsigned int offset = offset_in_page(data);
- unsigned int len = PAGE_SIZE - offset;
-
- if (data + len > skb_tail_pointer(skb))
- len = skb_tail_pointer(skb) - data;
-
- xenvif_gop_frag_copy(queue, skb, npo,
- virt_to_page(data), len, offset, &head);
- data += len;
- }
-
- for (i = 0; i < nr_frags; i++) {
- xenvif_gop_frag_copy(queue, skb, npo,
- skb_frag_page(&skb_shinfo(skb)->frags[i]),
- skb_frag_size(&skb_shinfo(skb)->frags[i]),
- skb_shinfo(skb)->frags[i].page_offset,
- &head);
- }
-
- return npo->meta_prod - old_meta_prod;
-}
-
-/*
- * This is a twin to xenvif_gop_skb. Assume that xenvif_gop_skb was
- * used to set up the operations on the top of
- * netrx_pending_operations, which have since been done. Check that
- * they didn't give any errors and advance over them.
- */
-static int xenvif_check_gop(struct xenvif *vif, int nr_meta_slots,
- struct netrx_pending_operations *npo)
-{
- struct gnttab_copy *copy_op;
- int status = XEN_NETIF_RSP_OKAY;
- int i;
-
- for (i = 0; i < nr_meta_slots; i++) {
- copy_op = npo->copy + npo->copy_cons++;
- if (copy_op->status != GNTST_okay) {
- netdev_dbg(vif->dev,
- "Bad status %d from copy to DOM%d.\n",
- copy_op->status, vif->domid);
- status = XEN_NETIF_RSP_ERROR;
- }
- }
-
- return status;
-}
-
-static void xenvif_add_frag_responses(struct xenvif_queue *queue, int status,
- struct xenvif_rx_meta *meta,
- int nr_meta_slots)
-{
- int i;
- unsigned long offset;
-
- /* No fragments used */
- if (nr_meta_slots <= 1)
- return;
-
- nr_meta_slots--;
-
- for (i = 0; i < nr_meta_slots; i++) {
- int flags;
- if (i == nr_meta_slots - 1)
- flags = 0;
- else
- flags = XEN_NETRXF_more_data;
-
- offset = 0;
- make_rx_response(queue, meta[i].id, status, offset,
- meta[i].size, flags);
- }
-}
-
void xenvif_kick_thread(struct xenvif_queue *queue)
{
wake_up(&queue->wq);
}
-static void xenvif_rx_action(struct xenvif_queue *queue)
-{
- struct xenvif *vif = queue->vif;
- s8 status;
- u16 flags;
- struct xen_netif_rx_response *resp;
- struct sk_buff_head rxq;
- struct sk_buff *skb;
- LIST_HEAD(notify);
- int ret;
- unsigned long offset;
- bool need_to_notify = false;
-
- struct netrx_pending_operations npo = {
- .copy = queue->grant_copy_op,
- .meta = queue->meta,
- };
-
- skb_queue_head_init(&rxq);
-
- while (xenvif_rx_ring_slots_available(queue)
- && (skb = xenvif_rx_dequeue(queue)) != NULL) {
- queue->last_rx_time = jiffies;
-
- XENVIF_RX_CB(skb)->meta_slots_used = xenvif_gop_skb(skb, &npo, queue);
-
- __skb_queue_tail(&rxq, skb);
- }
-
- BUG_ON(npo.meta_prod > ARRAY_SIZE(queue->meta));
-
- if (!npo.copy_prod)
- goto done;
-
- BUG_ON(npo.copy_prod > MAX_GRANT_COPY_OPS);
- gnttab_batch_copy(queue->grant_copy_op, npo.copy_prod);
-
- while ((skb = __skb_dequeue(&rxq)) != NULL) {
- struct xen_netif_extra_info *extra = NULL;
-
- if ((1 << queue->meta[npo.meta_cons].gso_type) &
- vif->gso_prefix_mask) {
- resp = RING_GET_RESPONSE(&queue->rx,
- queue->rx.rsp_prod_pvt++);
-
- resp->flags = XEN_NETRXF_gso_prefix | XEN_NETRXF_more_data;
-
- resp->offset = queue->meta[npo.meta_cons].gso_size;
- resp->id = queue->meta[npo.meta_cons].id;
- resp->status = XENVIF_RX_CB(skb)->meta_slots_used;
-
- npo.meta_cons++;
- XENVIF_RX_CB(skb)->meta_slots_used--;
- }
-
-
- queue->stats.tx_bytes += skb->len;
- queue->stats.tx_packets++;
-
- status = xenvif_check_gop(vif,
- XENVIF_RX_CB(skb)->meta_slots_used,
- &npo);
-
- if (XENVIF_RX_CB(skb)->meta_slots_used == 1)
- flags = 0;
- else
- flags = XEN_NETRXF_more_data;
-
- if (skb->ip_summed == CHECKSUM_PARTIAL) /* local packet? */
- flags |= XEN_NETRXF_csum_blank | XEN_NETRXF_data_validated;
- else if (skb->ip_summed == CHECKSUM_UNNECESSARY)
- /* remote but checksummed. */
- flags |= XEN_NETRXF_data_validated;
-
- offset = 0;
- resp = make_rx_response(queue, queue->meta[npo.meta_cons].id,
- status, offset,
- queue->meta[npo.meta_cons].size,
- flags);
-
- if ((1 << queue->meta[npo.meta_cons].gso_type) &
- vif->gso_mask) {
- extra = (struct xen_netif_extra_info *)
- RING_GET_RESPONSE(&queue->rx,
- queue->rx.rsp_prod_pvt++);
-
- resp->flags |= XEN_NETRXF_extra_info;
-
- extra->u.gso.type = queue->meta[npo.meta_cons].gso_type;
- extra->u.gso.size = queue->meta[npo.meta_cons].gso_size;
- extra->u.gso.pad = 0;
- extra->u.gso.features = 0;
-
- extra->type = XEN_NETIF_EXTRA_TYPE_GSO;
- extra->flags = 0;
- }
-
- if (skb->sw_hash) {
- /* Since the skb got here via xenvif_select_queue()
- * we know that the hash has been re-calculated
- * according to a configuration set by the frontend
- * and therefore we know that it is legitimate to
- * pass it to the frontend.
- */
- if (resp->flags & XEN_NETRXF_extra_info)
- extra->flags |= XEN_NETIF_EXTRA_FLAG_MORE;
- else
- resp->flags |= XEN_NETRXF_extra_info;
-
- extra = (struct xen_netif_extra_info *)
- RING_GET_RESPONSE(&queue->rx,
- queue->rx.rsp_prod_pvt++);
-
- extra->u.hash.algorithm =
- XEN_NETIF_CTRL_HASH_ALGORITHM_TOEPLITZ;
-
- if (skb->l4_hash)
- extra->u.hash.type =
- skb->protocol == htons(ETH_P_IP) ?
- _XEN_NETIF_CTRL_HASH_TYPE_IPV4_TCP :
- _XEN_NETIF_CTRL_HASH_TYPE_IPV6_TCP;
- else
- extra->u.hash.type =
- skb->protocol == htons(ETH_P_IP) ?
- _XEN_NETIF_CTRL_HASH_TYPE_IPV4 :
- _XEN_NETIF_CTRL_HASH_TYPE_IPV6;
-
- *(uint32_t *)extra->u.hash.value =
- skb_get_hash_raw(skb);
-
- extra->type = XEN_NETIF_EXTRA_TYPE_HASH;
- extra->flags = 0;
- }
-
- xenvif_add_frag_responses(queue, status,
- queue->meta + npo.meta_cons + 1,
- XENVIF_RX_CB(skb)->meta_slots_used);
-
- RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->rx, ret);
-
- need_to_notify |= !!ret;
-
- npo.meta_cons += XENVIF_RX_CB(skb)->meta_slots_used;
- dev_kfree_skb(skb);
- }
-
-done:
- if (need_to_notify)
- notify_remote_via_irq(queue->rx_irq);
-}
-
void xenvif_napi_schedule_or_enable_events(struct xenvif_queue *queue)
{
int more_to_do;
@@ -1951,29 +1384,6 @@ static void push_tx_responses(struct xenvif_queue *queue)
notify_remote_via_irq(queue->tx_irq);
}
-static struct xen_netif_rx_response *make_rx_response(struct xenvif_queue *queue,
- u16 id,
- s8 st,
- u16 offset,
- u16 size,
- u16 flags)
-{
- RING_IDX i = queue->rx.rsp_prod_pvt;
- struct xen_netif_rx_response *resp;
-
- resp = RING_GET_RESPONSE(&queue->rx, i);
- resp->offset = offset;
- resp->flags = flags;
- resp->id = id;
- resp->status = (s16)size;
- if (st < 0)
- resp->status = (s16)st;
-
- queue->rx.rsp_prod_pvt = ++i;
-
- return resp;
-}
-
void xenvif_idx_unmap(struct xenvif_queue *queue, u16 pending_idx)
{
int ret;
@@ -2055,170 +1465,6 @@ err:
return err;
}
-static void xenvif_queue_carrier_off(struct xenvif_queue *queue)
-{
- struct xenvif *vif = queue->vif;
-
- queue->stalled = true;
-
- /* At least one queue has stalled? Disable the carrier. */
- spin_lock(&vif->lock);
- if (vif->stalled_queues++ == 0) {
- netdev_info(vif->dev, "Guest Rx stalled");
- netif_carrier_off(vif->dev);
- }
- spin_unlock(&vif->lock);
-}
-
-static void xenvif_queue_carrier_on(struct xenvif_queue *queue)
-{
- struct xenvif *vif = queue->vif;
-
- queue->last_rx_time = jiffies; /* Reset Rx stall detection. */
- queue->stalled = false;
-
- /* All queues are ready? Enable the carrier. */
- spin_lock(&vif->lock);
- if (--vif->stalled_queues == 0) {
- netdev_info(vif->dev, "Guest Rx ready");
- netif_carrier_on(vif->dev);
- }
- spin_unlock(&vif->lock);
-}
-
-static bool xenvif_rx_queue_stalled(struct xenvif_queue *queue)
-{
- RING_IDX prod, cons;
-
- prod = queue->rx.sring->req_prod;
- cons = queue->rx.req_cons;
-
- return !queue->stalled && prod - cons < 1
- && time_after(jiffies,
- queue->last_rx_time + queue->vif->stall_timeout);
-}
-
-static bool xenvif_rx_queue_ready(struct xenvif_queue *queue)
-{
- RING_IDX prod, cons;
-
- prod = queue->rx.sring->req_prod;
- cons = queue->rx.req_cons;
-
- return queue->stalled && prod - cons >= 1;
-}
-
-static bool xenvif_have_rx_work(struct xenvif_queue *queue)
-{
- return xenvif_rx_ring_slots_available(queue)
- || (queue->vif->stall_timeout &&
- (xenvif_rx_queue_stalled(queue)
- || xenvif_rx_queue_ready(queue)))
- || kthread_should_stop()
- || queue->vif->disabled;
-}
-
-static long xenvif_rx_queue_timeout(struct xenvif_queue *queue)
-{
- struct sk_buff *skb;
- long timeout;
-
- skb = skb_peek(&queue->rx_queue);
- if (!skb)
- return MAX_SCHEDULE_TIMEOUT;
-
- timeout = XENVIF_RX_CB(skb)->expires - jiffies;
- return timeout < 0 ? 0 : timeout;
-}
-
-/* Wait until the guest Rx thread has work.
- *
- * The timeout needs to be adjusted based on the current head of the
- * queue (and not just the head at the beginning). In particular, if
- * the queue is initially empty an infinite timeout is used and this
- * needs to be reduced when a skb is queued.
- *
- * This cannot be done with wait_event_timeout() because it only
- * calculates the timeout once.
- */
-static void xenvif_wait_for_rx_work(struct xenvif_queue *queue)
-{
- DEFINE_WAIT(wait);
-
- if (xenvif_have_rx_work(queue))
- return;
-
- for (;;) {
- long ret;
-
- prepare_to_wait(&queue->wq, &wait, TASK_INTERRUPTIBLE);
- if (xenvif_have_rx_work(queue))
- break;
- ret = schedule_timeout(xenvif_rx_queue_timeout(queue));
- if (!ret)
- break;
- }
- finish_wait(&queue->wq, &wait);
-}
-
-int xenvif_kthread_guest_rx(void *data)
-{
- struct xenvif_queue *queue = data;
- struct xenvif *vif = queue->vif;
-
- if (!vif->stall_timeout)
- xenvif_queue_carrier_on(queue);
-
- for (;;) {
- xenvif_wait_for_rx_work(queue);
-
- if (kthread_should_stop())
- break;
-
- /* This frontend is found to be rogue, disable it in
- * kthread context. Currently this is only set when
- * netback finds out frontend sends malformed packet,
- * but we cannot disable the interface in softirq
- * context so we defer it here, if this thread is
- * associated with queue 0.
- */
- if (unlikely(vif->disabled && queue->id == 0)) {
- xenvif_carrier_off(vif);
- break;
- }
-
- if (!skb_queue_empty(&queue->rx_queue))
- xenvif_rx_action(queue);
-
- /* If the guest hasn't provided any Rx slots for a
- * while it's probably not responsive, drop the
- * carrier so packets are dropped earlier.
- */
- if (vif->stall_timeout) {
- if (xenvif_rx_queue_stalled(queue))
- xenvif_queue_carrier_off(queue);
- else if (xenvif_rx_queue_ready(queue))
- xenvif_queue_carrier_on(queue);
- }
-
- /* Queued packets may have foreign pages from other
- * domains. These cannot be queued indefinitely as
- * this would starve guests of grant refs and transmit
- * slots.
- */
- xenvif_rx_queue_drop_expired(queue);
-
- xenvif_rx_queue_maybe_wake(queue);
-
- cond_resched();
- }
-
- /* Bin any remaining skbs */
- xenvif_rx_queue_purge(queue);
-
- return 0;
-}
-
static bool xenvif_dealloc_kthread_should_stop(struct xenvif_queue *queue)
{
/* Dealloc thread must remain running until all inflight
diff --git a/drivers/net/xen-netback/rx.c b/drivers/net/xen-netback/rx.c
new file mode 100644
index 0000000..03836aa
--- /dev/null
+++ b/drivers/net/xen-netback/rx.c
@@ -0,0 +1,789 @@
+/*
+ * Copyright (c) 2016 Citrix Systems Inc.
+ * Copyright (c) 2002-2005, K A Fraser
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * 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 "common.h"
+
+#include <linux/kthread.h>
+
+#include <xen/xen.h>
+#include <xen/events.h>
+
+static bool xenvif_rx_ring_slots_available(struct xenvif_queue *queue)
+{
+ RING_IDX prod, cons;
+ struct sk_buff *skb;
+ int needed;
+
+ skb = skb_peek(&queue->rx_queue);
+ if (!skb)
+ return false;
+
+ needed = DIV_ROUND_UP(skb->len, XEN_PAGE_SIZE);
+ if (skb_is_gso(skb))
+ needed++;
+ if (skb->sw_hash)
+ needed++;
+
+ do {
+ prod = queue->rx.sring->req_prod;
+ cons = queue->rx.req_cons;
+
+ if (prod - cons >= needed)
+ return true;
+
+ queue->rx.sring->req_event = prod + 1;
+
+ /* Make sure event is visible before we check prod
+ * again.
+ */
+ mb();
+ } while (queue->rx.sring->req_prod != prod);
+
+ return false;
+}
+
+void xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&queue->rx_queue.lock, flags);
+
+ __skb_queue_tail(&queue->rx_queue, skb);
+
+ queue->rx_queue_len += skb->len;
+ if (queue->rx_queue_len > queue->rx_queue_max) {
+ struct net_device *dev = queue->vif->dev;
+
+ netif_tx_stop_queue(netdev_get_tx_queue(dev, queue->id));
+ }
+
+ spin_unlock_irqrestore(&queue->rx_queue.lock, flags);
+}
+
+static struct sk_buff *xenvif_rx_dequeue(struct xenvif_queue *queue)
+{
+ struct sk_buff *skb;
+
+ spin_lock_irq(&queue->rx_queue.lock);
+
+ skb = __skb_dequeue(&queue->rx_queue);
+ if (skb)
+ queue->rx_queue_len -= skb->len;
+
+ spin_unlock_irq(&queue->rx_queue.lock);
+
+ return skb;
+}
+
+static void xenvif_rx_queue_maybe_wake(struct xenvif_queue *queue)
+{
+ spin_lock_irq(&queue->rx_queue.lock);
+
+ if (queue->rx_queue_len < queue->rx_queue_max) {
+ struct net_device *dev = queue->vif->dev;
+
+ netif_tx_wake_queue(netdev_get_tx_queue(dev, queue->id));
+ }
+
+ spin_unlock_irq(&queue->rx_queue.lock);
+}
+
+static void xenvif_rx_queue_purge(struct xenvif_queue *queue)
+{
+ struct sk_buff *skb;
+
+ while ((skb = xenvif_rx_dequeue(queue)) != NULL)
+ kfree_skb(skb);
+}
+
+static void xenvif_rx_queue_drop_expired(struct xenvif_queue *queue)
+{
+ struct sk_buff *skb;
+
+ for (;;) {
+ skb = skb_peek(&queue->rx_queue);
+ if (!skb)
+ break;
+ if (time_before(jiffies, XENVIF_RX_CB(skb)->expires))
+ break;
+ xenvif_rx_dequeue(queue);
+ kfree_skb(skb);
+ }
+}
+
+struct netrx_pending_operations {
+ unsigned int copy_prod, copy_cons;
+ unsigned int meta_prod, meta_cons;
+ struct gnttab_copy *copy;
+ struct xenvif_rx_meta *meta;
+ int copy_off;
+ grant_ref_t copy_gref;
+};
+
+static struct xenvif_rx_meta *get_next_rx_buffer(
+ struct xenvif_queue *queue,
+ struct netrx_pending_operations *npo)
+{
+ struct xenvif_rx_meta *meta;
+ struct xen_netif_rx_request req;
+
+ RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
+
+ meta = npo->meta + npo->meta_prod++;
+ meta->gso_type = XEN_NETIF_GSO_TYPE_NONE;
+ meta->gso_size = 0;
+ meta->size = 0;
+ meta->id = req.id;
+
+ npo->copy_off = 0;
+ npo->copy_gref = req.gref;
+
+ return meta;
+}
+
+struct gop_frag_copy {
+ struct xenvif_queue *queue;
+ struct netrx_pending_operations *npo;
+ struct xenvif_rx_meta *meta;
+ int head;
+ int gso_type;
+ int protocol;
+ int hash_present;
+
+ struct page *page;
+};
+
+static void xenvif_setup_copy_gop(unsigned long gfn,
+ unsigned int offset,
+ unsigned int *len,
+ struct gop_frag_copy *info)
+{
+ struct gnttab_copy *copy_gop;
+ struct xen_page_foreign *foreign;
+ /* Convenient aliases */
+ struct xenvif_queue *queue = info->queue;
+ struct netrx_pending_operations *npo = info->npo;
+ struct page *page = info->page;
+
+ WARN_ON(npo->copy_off > MAX_BUFFER_OFFSET);
+
+ if (npo->copy_off == MAX_BUFFER_OFFSET)
+ info->meta = get_next_rx_buffer(queue, npo);
+
+ if (npo->copy_off + *len > MAX_BUFFER_OFFSET)
+ *len = MAX_BUFFER_OFFSET - npo->copy_off;
+
+ copy_gop = npo->copy + npo->copy_prod++;
+ copy_gop->flags = GNTCOPY_dest_gref;
+ copy_gop->len = *len;
+
+ foreign = xen_page_foreign(page);
+ if (foreign) {
+ copy_gop->source.domid = foreign->domid;
+ copy_gop->source.u.ref = foreign->gref;
+ copy_gop->flags |= GNTCOPY_source_gref;
+ } else {
+ copy_gop->source.domid = DOMID_SELF;
+ copy_gop->source.u.gmfn = gfn;
+ }
+ copy_gop->source.offset = offset;
+
+ copy_gop->dest.domid = queue->vif->domid;
+ copy_gop->dest.offset = npo->copy_off;
+ copy_gop->dest.u.ref = npo->copy_gref;
+
+ npo->copy_off += *len;
+ info->meta->size += *len;
+
+ if (!info->head)
+ return;
+
+ /* Leave a gap for the GSO descriptor. */
+ if ((1 << info->gso_type) & queue->vif->gso_mask)
+ queue->rx.req_cons++;
+
+ /* Leave a gap for the hash extra segment. */
+ if (info->hash_present)
+ queue->rx.req_cons++;
+
+ info->head = 0; /* There must be something in this buffer now */
+}
+
+static void xenvif_gop_frag_copy_grant(unsigned long gfn,
+ unsigned int offset,
+ unsigned int len,
+ void *data)
+{
+ unsigned int bytes;
+
+ while (len) {
+ bytes = len;
+ xenvif_setup_copy_gop(gfn, offset, &bytes, data);
+ offset += bytes;
+ len -= bytes;
+ }
+}
+
+/* Set up the grant operations for this fragment. If it's a flipping
+ * interface, we also set up the unmap request from here.
+ */
+static void xenvif_gop_frag_copy(struct xenvif_queue *queue,
+ struct sk_buff *skb,
+ struct netrx_pending_operations *npo,
+ struct page *page, unsigned long size,
+ unsigned long offset, int *head)
+{
+ struct gop_frag_copy info = {
+ .queue = queue,
+ .npo = npo,
+ .head = *head,
+ .gso_type = XEN_NETIF_GSO_TYPE_NONE,
+ /* xenvif_set_skb_hash() will have either set a s/w
+ * hash or cleared the hash depending on
+ * whether the the frontend wants a hash for this skb.
+ */
+ .hash_present = skb->sw_hash,
+ };
+ unsigned long bytes;
+
+ if (skb_is_gso(skb)) {
+ if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
+ info.gso_type = XEN_NETIF_GSO_TYPE_TCPV4;
+ else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
+ info.gso_type = XEN_NETIF_GSO_TYPE_TCPV6;
+ }
+
+ /* Data must not cross a page boundary. */
+ WARN_ON(size + offset > (PAGE_SIZE << compound_order(page)));
+
+ info.meta = npo->meta + npo->meta_prod - 1;
+
+ /* Skip unused frames from start of page */
+ page += offset >> PAGE_SHIFT;
+ offset &= ~PAGE_MASK;
+
+ while (size > 0) {
+ WARN_ON(offset >= PAGE_SIZE);
+
+ bytes = PAGE_SIZE - offset;
+ if (bytes > size)
+ bytes = size;
+
+ info.page = page;
+ gnttab_foreach_grant_in_range(page, offset, bytes,
+ xenvif_gop_frag_copy_grant,
+ &info);
+ size -= bytes;
+ offset = 0;
+
+ /* Next page */
+ if (size) {
+ WARN_ON(!PageCompound(page));
+ page++;
+ }
+ }
+
+ *head = info.head;
+}
+
+/* Prepare an SKB to be transmitted to the frontend.
+ *
+ * This function is responsible for allocating grant operations, meta
+ * structures, etc.
+ *
+ * It returns the number of meta structures consumed. The number of
+ * ring slots used is always equal to the number of meta slots used
+ * plus the number of GSO descriptors used. Currently, we use either
+ * zero GSO descriptors (for non-GSO packets) or one descriptor (for
+ * frontend-side LRO).
+ */
+static int xenvif_gop_skb(struct sk_buff *skb,
+ struct netrx_pending_operations *npo,
+ struct xenvif_queue *queue)
+{
+ struct xenvif *vif = netdev_priv(skb->dev);
+ int nr_frags = skb_shinfo(skb)->nr_frags;
+ int i;
+ struct xen_netif_rx_request req;
+ struct xenvif_rx_meta *meta;
+ unsigned char *data;
+ int head = 1;
+ int old_meta_prod;
+ int gso_type;
+
+ old_meta_prod = npo->meta_prod;
+
+ gso_type = XEN_NETIF_GSO_TYPE_NONE;
+ if (skb_is_gso(skb)) {
+ if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
+ gso_type = XEN_NETIF_GSO_TYPE_TCPV4;
+ else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
+ gso_type = XEN_NETIF_GSO_TYPE_TCPV6;
+ }
+
+ /* Set up a GSO prefix descriptor, if necessary */
+ if ((1 << gso_type) & vif->gso_prefix_mask) {
+ RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
+ meta = npo->meta + npo->meta_prod++;
+ meta->gso_type = gso_type;
+ meta->gso_size = skb_shinfo(skb)->gso_size;
+ meta->size = 0;
+ meta->id = req.id;
+ }
+
+ RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
+ meta = npo->meta + npo->meta_prod++;
+
+ if ((1 << gso_type) & vif->gso_mask) {
+ meta->gso_type = gso_type;
+ meta->gso_size = skb_shinfo(skb)->gso_size;
+ } else {
+ meta->gso_type = XEN_NETIF_GSO_TYPE_NONE;
+ meta->gso_size = 0;
+ }
+
+ meta->size = 0;
+ meta->id = req.id;
+ npo->copy_off = 0;
+ npo->copy_gref = req.gref;
+
+ data = skb->data;
+ while (data < skb_tail_pointer(skb)) {
+ unsigned int offset = offset_in_page(data);
+ unsigned int len = PAGE_SIZE - offset;
+
+ if (data + len > skb_tail_pointer(skb))
+ len = skb_tail_pointer(skb) - data;
+
+ xenvif_gop_frag_copy(queue, skb, npo,
+ virt_to_page(data), len, offset, &head);
+ data += len;
+ }
+
+ for (i = 0; i < nr_frags; i++) {
+ xenvif_gop_frag_copy(queue, skb, npo,
+ skb_frag_page(&skb_shinfo(skb)->frags[i]),
+ skb_frag_size(&skb_shinfo(skb)->frags[i]),
+ skb_shinfo(skb)->frags[i].page_offset,
+ &head);
+ }
+
+ return npo->meta_prod - old_meta_prod;
+}
+
+/* This is a twin to xenvif_gop_skb. Assume that xenvif_gop_skb was
+ * used to set up the operations on the top of
+ * netrx_pending_operations, which have since been done. Check that
+ * they didn't give any errors and advance over them.
+ */
+static int xenvif_check_gop(struct xenvif *vif, int nr_meta_slots,
+ struct netrx_pending_operations *npo)
+{
+ struct gnttab_copy *copy_op;
+ int status = XEN_NETIF_RSP_OKAY;
+ int i;
+
+ for (i = 0; i < nr_meta_slots; i++) {
+ copy_op = npo->copy + npo->copy_cons++;
+ if (copy_op->status != GNTST_okay) {
+ netdev_dbg(vif->dev,
+ "Bad status %d from copy to DOM%d.\n",
+ copy_op->status, vif->domid);
+ status = XEN_NETIF_RSP_ERROR;
+ }
+ }
+
+ return status;
+}
+
+static struct xen_netif_rx_response *make_rx_response(
+ struct xenvif_queue *queue, u16 id, s8 st, u16 offset, u16 size,
+ u16 flags)
+{
+ RING_IDX i = queue->rx.rsp_prod_pvt;
+ struct xen_netif_rx_response *resp;
+
+ resp = RING_GET_RESPONSE(&queue->rx, i);
+ resp->offset = offset;
+ resp->flags = flags;
+ resp->id = id;
+ resp->status = (s16)size;
+ if (st < 0)
+ resp->status = (s16)st;
+
+ queue->rx.rsp_prod_pvt = ++i;
+
+ return resp;
+}
+
+static void xenvif_add_frag_responses(struct xenvif_queue *queue,
+ int status,
+ struct xenvif_rx_meta *meta,
+ int nr_meta_slots)
+{
+ int i;
+ unsigned long offset;
+
+ /* No fragments used */
+ if (nr_meta_slots <= 1)
+ return;
+
+ nr_meta_slots--;
+
+ for (i = 0; i < nr_meta_slots; i++) {
+ int flags;
+
+ if (i == nr_meta_slots - 1)
+ flags = 0;
+ else
+ flags = XEN_NETRXF_more_data;
+
+ offset = 0;
+ make_rx_response(queue, meta[i].id, status, offset,
+ meta[i].size, flags);
+ }
+}
+
+static void xenvif_rx_action(struct xenvif_queue *queue)
+{
+ struct xenvif *vif = queue->vif;
+ s8 status;
+ u16 flags;
+ struct xen_netif_rx_response *resp;
+ struct sk_buff_head rxq;
+ struct sk_buff *skb;
+ LIST_HEAD(notify);
+ int ret;
+ unsigned long offset;
+ bool need_to_notify = false;
+
+ struct netrx_pending_operations npo = {
+ .copy = queue->grant_copy_op,
+ .meta = queue->meta,
+ };
+
+ skb_queue_head_init(&rxq);
+
+ while (xenvif_rx_ring_slots_available(queue) &&
+ (skb = xenvif_rx_dequeue(queue)) != NULL) {
+ queue->last_rx_time = jiffies;
+
+ XENVIF_RX_CB(skb)->meta_slots_used =
+ xenvif_gop_skb(skb, &npo, queue);
+
+ __skb_queue_tail(&rxq, skb);
+ }
+
+ WARN_ON(npo.meta_prod > ARRAY_SIZE(queue->meta));
+
+ if (!npo.copy_prod)
+ goto done;
+
+ WARN_ON(npo.copy_prod > MAX_GRANT_COPY_OPS);
+ gnttab_batch_copy(queue->grant_copy_op, npo.copy_prod);
+
+ while ((skb = __skb_dequeue(&rxq)) != NULL) {
+ struct xen_netif_extra_info *extra = NULL;
+
+ if ((1 << queue->meta[npo.meta_cons].gso_type) &
+ vif->gso_prefix_mask) {
+ resp = RING_GET_RESPONSE(&queue->rx,
+ queue->rx.rsp_prod_pvt++);
+
+ resp->flags = XEN_NETRXF_gso_prefix |
+ XEN_NETRXF_more_data;
+
+ resp->offset = queue->meta[npo.meta_cons].gso_size;
+ resp->id = queue->meta[npo.meta_cons].id;
+ resp->status = XENVIF_RX_CB(skb)->meta_slots_used;
+
+ npo.meta_cons++;
+ XENVIF_RX_CB(skb)->meta_slots_used--;
+ }
+
+ queue->stats.tx_bytes += skb->len;
+ queue->stats.tx_packets++;
+
+ status = xenvif_check_gop(vif,
+ XENVIF_RX_CB(skb)->meta_slots_used,
+ &npo);
+
+ if (XENVIF_RX_CB(skb)->meta_slots_used == 1)
+ flags = 0;
+ else
+ flags = XEN_NETRXF_more_data;
+
+ if (skb->ip_summed == CHECKSUM_PARTIAL) /* local packet? */
+ flags |= XEN_NETRXF_csum_blank |
+ XEN_NETRXF_data_validated;
+ else if (skb->ip_summed == CHECKSUM_UNNECESSARY)
+ /* remote but checksummed. */
+ flags |= XEN_NETRXF_data_validated;
+
+ offset = 0;
+ resp = make_rx_response(queue, queue->meta[npo.meta_cons].id,
+ status, offset,
+ queue->meta[npo.meta_cons].size,
+ flags);
+
+ if ((1 << queue->meta[npo.meta_cons].gso_type) &
+ vif->gso_mask) {
+ extra = (struct xen_netif_extra_info *)
+ RING_GET_RESPONSE(&queue->rx,
+ queue->rx.rsp_prod_pvt++);
+
+ resp->flags |= XEN_NETRXF_extra_info;
+
+ extra->u.gso.type = queue->meta[npo.meta_cons].gso_type;
+ extra->u.gso.size = queue->meta[npo.meta_cons].gso_size;
+ extra->u.gso.pad = 0;
+ extra->u.gso.features = 0;
+
+ extra->type = XEN_NETIF_EXTRA_TYPE_GSO;
+ extra->flags = 0;
+ }
+
+ if (skb->sw_hash) {
+ /* Since the skb got here via xenvif_select_queue()
+ * we know that the hash has been re-calculated
+ * according to a configuration set by the frontend
+ * and therefore we know that it is legitimate to
+ * pass it to the frontend.
+ */
+ if (resp->flags & XEN_NETRXF_extra_info)
+ extra->flags |= XEN_NETIF_EXTRA_FLAG_MORE;
+ else
+ resp->flags |= XEN_NETRXF_extra_info;
+
+ extra = (struct xen_netif_extra_info *)
+ RING_GET_RESPONSE(&queue->rx,
+ queue->rx.rsp_prod_pvt++);
+
+ extra->u.hash.algorithm =
+ XEN_NETIF_CTRL_HASH_ALGORITHM_TOEPLITZ;
+
+ if (skb->l4_hash)
+ extra->u.hash.type =
+ skb->protocol == htons(ETH_P_IP) ?
+ _XEN_NETIF_CTRL_HASH_TYPE_IPV4_TCP :
+ _XEN_NETIF_CTRL_HASH_TYPE_IPV6_TCP;
+ else
+ extra->u.hash.type =
+ skb->protocol == htons(ETH_P_IP) ?
+ _XEN_NETIF_CTRL_HASH_TYPE_IPV4 :
+ _XEN_NETIF_CTRL_HASH_TYPE_IPV6;
+
+ *(uint32_t *)extra->u.hash.value =
+ skb_get_hash_raw(skb);
+
+ extra->type = XEN_NETIF_EXTRA_TYPE_HASH;
+ extra->flags = 0;
+ }
+
+ xenvif_add_frag_responses(queue, status,
+ queue->meta + npo.meta_cons + 1,
+ XENVIF_RX_CB(skb)->meta_slots_used);
+
+ RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->rx, ret);
+
+ need_to_notify |= !!ret;
+
+ npo.meta_cons += XENVIF_RX_CB(skb)->meta_slots_used;
+ dev_kfree_skb(skb);
+ }
+
+done:
+ if (need_to_notify)
+ notify_remote_via_irq(queue->rx_irq);
+}
+
+static bool xenvif_rx_queue_stalled(struct xenvif_queue *queue)
+{
+ RING_IDX prod, cons;
+
+ prod = queue->rx.sring->req_prod;
+ cons = queue->rx.req_cons;
+
+ return !queue->stalled &&
+ prod - cons < 1 &&
+ time_after(jiffies,
+ queue->last_rx_time + queue->vif->stall_timeout);
+}
+
+static bool xenvif_rx_queue_ready(struct xenvif_queue *queue)
+{
+ RING_IDX prod, cons;
+
+ prod = queue->rx.sring->req_prod;
+ cons = queue->rx.req_cons;
+
+ return queue->stalled && prod - cons >= 1;
+}
+
+static bool xenvif_have_rx_work(struct xenvif_queue *queue)
+{
+ return xenvif_rx_ring_slots_available(queue) ||
+ (queue->vif->stall_timeout &&
+ (xenvif_rx_queue_stalled(queue) ||
+ xenvif_rx_queue_ready(queue))) ||
+ kthread_should_stop() ||
+ queue->vif->disabled;
+}
+
+static long xenvif_rx_queue_timeout(struct xenvif_queue *queue)
+{
+ struct sk_buff *skb;
+ long timeout;
+
+ skb = skb_peek(&queue->rx_queue);
+ if (!skb)
+ return MAX_SCHEDULE_TIMEOUT;
+
+ timeout = XENVIF_RX_CB(skb)->expires - jiffies;
+ return timeout < 0 ? 0 : timeout;
+}
+
+/* Wait until the guest Rx thread has work.
+ *
+ * The timeout needs to be adjusted based on the current head of the
+ * queue (and not just the head at the beginning). In particular, if
+ * the queue is initially empty an infinite timeout is used and this
+ * needs to be reduced when a skb is queued.
+ *
+ * This cannot be done with wait_event_timeout() because it only
+ * calculates the timeout once.
+ */
+static void xenvif_wait_for_rx_work(struct xenvif_queue *queue)
+{
+ DEFINE_WAIT(wait);
+
+ if (xenvif_have_rx_work(queue))
+ return;
+
+ for (;;) {
+ long ret;
+
+ prepare_to_wait(&queue->wq, &wait, TASK_INTERRUPTIBLE);
+ if (xenvif_have_rx_work(queue))
+ break;
+ ret = schedule_timeout(xenvif_rx_queue_timeout(queue));
+ if (!ret)
+ break;
+ }
+ finish_wait(&queue->wq, &wait);
+}
+
+static void xenvif_queue_carrier_off(struct xenvif_queue *queue)
+{
+ struct xenvif *vif = queue->vif;
+
+ queue->stalled = true;
+
+ /* At least one queue has stalled? Disable the carrier. */
+ spin_lock(&vif->lock);
+ if (vif->stalled_queues++ == 0) {
+ netdev_info(vif->dev, "Guest Rx stalled");
+ netif_carrier_off(vif->dev);
+ }
+ spin_unlock(&vif->lock);
+}
+
+static void xenvif_queue_carrier_on(struct xenvif_queue *queue)
+{
+ struct xenvif *vif = queue->vif;
+
+ queue->last_rx_time = jiffies; /* Reset Rx stall detection. */
+ queue->stalled = false;
+
+ /* All queues are ready? Enable the carrier. */
+ spin_lock(&vif->lock);
+ if (--vif->stalled_queues == 0) {
+ netdev_info(vif->dev, "Guest Rx ready");
+ netif_carrier_on(vif->dev);
+ }
+ spin_unlock(&vif->lock);
+}
+
+int xenvif_kthread_guest_rx(void *data)
+{
+ struct xenvif_queue *queue = data;
+ struct xenvif *vif = queue->vif;
+
+ if (!vif->stall_timeout)
+ xenvif_queue_carrier_on(queue);
+
+ for (;;) {
+ xenvif_wait_for_rx_work(queue);
+
+ if (kthread_should_stop())
+ break;
+
+ /* This frontend is found to be rogue, disable it in
+ * kthread context. Currently this is only set when
+ * netback finds out frontend sends malformed packet,
+ * but we cannot disable the interface in softirq
+ * context so we defer it here, if this thread is
+ * associated with queue 0.
+ */
+ if (unlikely(vif->disabled && queue->id == 0)) {
+ xenvif_carrier_off(vif);
+ break;
+ }
+
+ if (!skb_queue_empty(&queue->rx_queue))
+ xenvif_rx_action(queue);
+
+ /* If the guest hasn't provided any Rx slots for a
+ * while it's probably not responsive, drop the
+ * carrier so packets are dropped earlier.
+ */
+ if (vif->stall_timeout) {
+ if (xenvif_rx_queue_stalled(queue))
+ xenvif_queue_carrier_off(queue);
+ else if (xenvif_rx_queue_ready(queue))
+ xenvif_queue_carrier_on(queue);
+ }
+
+ /* Queued packets may have foreign pages from other
+ * domains. These cannot be queued indefinitely as
+ * this would starve guests of grant refs and transmit
+ * slots.
+ */
+ xenvif_rx_queue_drop_expired(queue);
+
+ xenvif_rx_queue_maybe_wake(queue);
+
+ cond_resched();
+ }
+
+ /* Bin any remaining skbs */
+ xenvif_rx_queue_purge(queue);
+
+ return 0;
+}
--
2.1.4
^ permalink raw reply related
* [PATCH net-next 7/7] xen/netback: add fraglist support for to-guest rx
From: Paul Durrant @ 2016-10-03 7:31 UTC (permalink / raw)
To: netdev, xen-devel; +Cc: Ross Lagerwall, Paul Durrant, Wei Liu
In-Reply-To: <1475479872-23717-1-git-send-email-paul.durrant@citrix.com>
From: Ross Lagerwall <ross.lagerwall@citrix.com>
This allows full 64K skbuffs (with 1500 mtu ethernet, composed of 45
fragments) to be handled by netback for to-guest rx.
Signed-off-by: Ross Lagerwall <ross.lagerwall@citrix.com>
[re-based]
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
Cc: Wei Liu <wei.liu2@citrix.com>
---
drivers/net/xen-netback/interface.c | 2 +-
drivers/net/xen-netback/rx.c | 38 ++++++++++++++++++++++++++++---------
2 files changed, 30 insertions(+), 10 deletions(-)
diff --git a/drivers/net/xen-netback/interface.c b/drivers/net/xen-netback/interface.c
index 1a009e7..8fef4fe 100644
--- a/drivers/net/xen-netback/interface.c
+++ b/drivers/net/xen-netback/interface.c
@@ -476,7 +476,7 @@ struct xenvif *xenvif_alloc(struct device *parent, domid_t domid,
dev->netdev_ops = &xenvif_netdev_ops;
dev->hw_features = NETIF_F_SG |
NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
- NETIF_F_TSO | NETIF_F_TSO6;
+ NETIF_F_TSO | NETIF_F_TSO6 | NETIF_F_FRAGLIST;
dev->features = dev->hw_features | NETIF_F_RXCSUM;
dev->ethtool_ops = &xenvif_ethtool_ops;
diff --git a/drivers/net/xen-netback/rx.c b/drivers/net/xen-netback/rx.c
index 8c8c5b5..8e9ade6 100644
--- a/drivers/net/xen-netback/rx.c
+++ b/drivers/net/xen-netback/rx.c
@@ -215,7 +215,8 @@ static unsigned int xenvif_gso_type(struct sk_buff *skb)
struct xenvif_pkt_state {
struct sk_buff *skb;
size_t remaining_len;
- int frag; /* frag == -1 => skb->head */
+ struct sk_buff *frag_iter;
+ int frag; /* frag == -1 => frag_iter->head */
unsigned int frag_offset;
struct xen_netif_extra_info extras[XEN_NETIF_EXTRA_TYPE_MAX - 1];
unsigned int extra_count;
@@ -237,6 +238,7 @@ static void xenvif_rx_next_skb(struct xenvif_queue *queue,
memset(pkt, 0, sizeof(struct xenvif_pkt_state));
pkt->skb = skb;
+ pkt->frag_iter = skb;
pkt->remaining_len = skb->len;
pkt->frag = -1;
@@ -293,20 +295,40 @@ static void xenvif_rx_complete(struct xenvif_queue *queue,
__skb_queue_tail(queue->rx_copy.completed, pkt->skb);
}
+static void xenvif_rx_next_frag(struct xenvif_pkt_state *pkt)
+{
+ struct sk_buff *frag_iter = pkt->frag_iter;
+ unsigned int nr_frags = skb_shinfo(frag_iter)->nr_frags;
+
+ pkt->frag++;
+ pkt->frag_offset = 0;
+
+ if (pkt->frag >= nr_frags) {
+ if (frag_iter == pkt->skb)
+ pkt->frag_iter = skb_shinfo(frag_iter)->frag_list;
+ else
+ pkt->frag_iter = frag_iter->next;
+
+ pkt->frag = -1;
+ }
+}
+
static void xenvif_rx_next_chunk(struct xenvif_queue *queue,
struct xenvif_pkt_state *pkt,
unsigned int offset, void **data,
size_t *len)
{
- struct sk_buff *skb = pkt->skb;
+ struct sk_buff *frag_iter = pkt->frag_iter;
void *frag_data;
size_t frag_len, chunk_len;
+ BUG_ON(!frag_iter);
+
if (pkt->frag == -1) {
- frag_data = skb->data;
- frag_len = skb_headlen(skb);
+ frag_data = frag_iter->data;
+ frag_len = skb_headlen(frag_iter);
} else {
- skb_frag_t *frag = &skb_shinfo(skb)->frags[pkt->frag];
+ skb_frag_t *frag = &skb_shinfo(frag_iter)->frags[pkt->frag];
frag_data = skb_frag_address(frag);
frag_len = skb_frag_size(frag);
@@ -322,10 +344,8 @@ static void xenvif_rx_next_chunk(struct xenvif_queue *queue,
pkt->frag_offset += chunk_len;
/* Advance to next frag? */
- if (frag_len == chunk_len) {
- pkt->frag++;
- pkt->frag_offset = 0;
- }
+ if (frag_len == chunk_len)
+ xenvif_rx_next_frag(pkt);
*data = frag_data;
*len = chunk_len;
--
2.1.4
^ permalink raw reply related
* [PATCH net-next 0/7] xen-netback: guest rx side refactor
From: Paul Durrant @ 2016-10-03 7:31 UTC (permalink / raw)
To: netdev, xen-devel; +Cc: Paul Durrant
This series refactors the guest rx side of xen-netback:
- The code is moved into its own source module.
- The prefix variant of GSO handling is retired (since it is no longer
in common use, and alternatives exist).
- The code is then simplified and modifications made to improve
performance.
David Vrabel (4):
xen-netback: refactor guest rx
xen-netback: immediately wake tx queue when guest rx queue has space
xen-netback: process guest rx packets in batches
xen-netback: batch copies for multiple to-guest rx packets
Paul Durrant (2):
xen-netback: separate guest side rx code into separate module
xen-netback: retire guest rx side prefix GSO feature
Ross Lagerwall (1):
xen/netback: add fraglist support for to-guest rx
drivers/net/xen-netback/Makefile | 2 +-
drivers/net/xen-netback/common.h | 25 +-
drivers/net/xen-netback/interface.c | 6 +-
drivers/net/xen-netback/netback.c | 754 ------------------------------------
drivers/net/xen-netback/rx.c | 628 ++++++++++++++++++++++++++++++
drivers/net/xen-netback/xenbus.c | 21 -
6 files changed, 643 insertions(+), 793 deletions(-)
create mode 100644 drivers/net/xen-netback/rx.c
--
2.1.4
^ permalink raw reply
* [PATCH net-next 6/7] xen-netback: batch copies for multiple to-guest rx packets
From: Paul Durrant @ 2016-10-03 7:31 UTC (permalink / raw)
To: netdev, xen-devel; +Cc: David Vrabel, Paul Durrant, Wei Liu
In-Reply-To: <1475479872-23717-1-git-send-email-paul.durrant@citrix.com>
From: David Vrabel <david.vrabel@citrix.com>
Instead of flushing the copy ops when an packet is complete, complete
packets when their copy ops are done. This improves performance by
reducing the number of grant copy hypercalls.
Latency is still limited by the relatively small size of the copy
batch.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
[re-based]
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
Cc: Wei Liu <wei.liu2@citrix.com>
---
drivers/net/xen-netback/common.h | 1 +
drivers/net/xen-netback/rx.c | 27 +++++++++++++++++----------
2 files changed, 18 insertions(+), 10 deletions(-)
diff --git a/drivers/net/xen-netback/common.h b/drivers/net/xen-netback/common.h
index adef482..5d40603 100644
--- a/drivers/net/xen-netback/common.h
+++ b/drivers/net/xen-netback/common.h
@@ -132,6 +132,7 @@ struct xenvif_copy_state {
struct gnttab_copy op[COPY_BATCH_SIZE];
RING_IDX idx[COPY_BATCH_SIZE];
unsigned int num;
+ struct sk_buff_head *completed;
};
struct xenvif_queue { /* Per-queue data for xenvif */
diff --git a/drivers/net/xen-netback/rx.c b/drivers/net/xen-netback/rx.c
index ae822b8..8c8c5b5 100644
--- a/drivers/net/xen-netback/rx.c
+++ b/drivers/net/xen-netback/rx.c
@@ -133,6 +133,7 @@ static void xenvif_rx_queue_drop_expired(struct xenvif_queue *queue)
static void xenvif_rx_copy_flush(struct xenvif_queue *queue)
{
unsigned int i;
+ int notify;
gnttab_batch_copy(queue->rx_copy.op, queue->rx_copy.num);
@@ -154,6 +155,13 @@ static void xenvif_rx_copy_flush(struct xenvif_queue *queue)
}
queue->rx_copy.num = 0;
+
+ /* Push responses for all completed packets. */
+ RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->rx, notify);
+ if (notify)
+ notify_remote_via_irq(queue->rx_irq);
+
+ __skb_queue_purge(queue->rx_copy.completed);
}
static void xenvif_rx_copy_add(struct xenvif_queue *queue,
@@ -279,18 +287,10 @@ static void xenvif_rx_next_skb(struct xenvif_queue *queue,
static void xenvif_rx_complete(struct xenvif_queue *queue,
struct xenvif_pkt_state *pkt)
{
- int notify;
-
- /* Complete any outstanding copy ops for this skb. */
- xenvif_rx_copy_flush(queue);
-
- /* Push responses and notify. */
+ /* All responses are ready to be pushed. */
queue->rx.rsp_prod_pvt = queue->rx.req_cons;
- RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->rx, notify);
- if (notify)
- notify_remote_via_irq(queue->rx_irq);
- dev_kfree_skb(pkt->skb);
+ __skb_queue_tail(queue->rx_copy.completed, pkt->skb);
}
static void xenvif_rx_next_chunk(struct xenvif_queue *queue,
@@ -429,13 +429,20 @@ void xenvif_rx_skb(struct xenvif_queue *queue)
void xenvif_rx_action(struct xenvif_queue *queue)
{
+ struct sk_buff_head completed_skbs;
unsigned int work_done = 0;
+ __skb_queue_head_init(&completed_skbs);
+ queue->rx_copy.completed = &completed_skbs;
+
while (xenvif_rx_ring_slots_available(queue) &&
work_done < RX_BATCH_SIZE) {
xenvif_rx_skb(queue);
work_done++;
}
+
+ /* Flush any pending copies and complete all skbs. */
+ xenvif_rx_copy_flush(queue);
}
static bool xenvif_rx_queue_stalled(struct xenvif_queue *queue)
--
2.1.4
^ permalink raw reply related
* [PATCH net-next 3/7] xen-netback: refactor guest rx
From: Paul Durrant @ 2016-10-03 7:31 UTC (permalink / raw)
To: netdev, xen-devel; +Cc: David Vrabel, Paul Durrant, Wei Liu
In-Reply-To: <1475479872-23717-1-git-send-email-paul.durrant@citrix.com>
From: David Vrabel <david.vrabel@citrix.com>
Refactor the to-guest (rx) path to:
1. Push responses for completed skbs earlier, reducing latency.
2. Reduce the per-queue memory overhead by greatly reducing the
maximum number of grant copy ops in each hypercall (from 4352 to
64). Each struct xenvif_queue is now only 44 kB instead of 220 kB.
3. Make the code more maintainable.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
[re-based]
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
Cc: Wei Liu <wei.liu2@citrix.com>
---
drivers/net/xen-netback/common.h | 23 +-
drivers/net/xen-netback/rx.c | 654 +++++++++++++++------------------------
2 files changed, 254 insertions(+), 423 deletions(-)
diff --git a/drivers/net/xen-netback/common.h b/drivers/net/xen-netback/common.h
index e16004a..adef482 100644
--- a/drivers/net/xen-netback/common.h
+++ b/drivers/net/xen-netback/common.h
@@ -91,13 +91,6 @@ struct xenvif_rx_meta {
*/
#define MAX_XEN_SKB_FRAGS (65536 / XEN_PAGE_SIZE + 1)
-/* It's possible for an skb to have a maximal number of frags
- * but still be less than MAX_BUFFER_OFFSET in size. Thus the
- * worst-case number of copy operations is MAX_XEN_SKB_FRAGS per
- * ring slot.
- */
-#define MAX_GRANT_COPY_OPS (MAX_XEN_SKB_FRAGS * XEN_NETIF_RX_RING_SIZE)
-
#define NETBACK_INVALID_HANDLE -1
/* To avoid confusion, we define XEN_NETBK_LEGACY_SLOTS_MAX indicating
@@ -133,6 +126,14 @@ struct xenvif_stats {
unsigned long tx_frag_overflow;
};
+#define COPY_BATCH_SIZE 64
+
+struct xenvif_copy_state {
+ struct gnttab_copy op[COPY_BATCH_SIZE];
+ RING_IDX idx[COPY_BATCH_SIZE];
+ unsigned int num;
+};
+
struct xenvif_queue { /* Per-queue data for xenvif */
unsigned int id; /* Queue ID, 0-based */
char name[QUEUE_NAME_SIZE]; /* DEVNAME-qN */
@@ -189,12 +190,7 @@ struct xenvif_queue { /* Per-queue data for xenvif */
unsigned long last_rx_time;
bool stalled;
- struct gnttab_copy grant_copy_op[MAX_GRANT_COPY_OPS];
-
- /* We create one meta structure per ring request we consume, so
- * the maximum number is the same as the ring size.
- */
- struct xenvif_rx_meta meta[XEN_NETIF_RX_RING_SIZE];
+ struct xenvif_copy_state rx_copy;
/* Transmit shaping: allow 'credit_bytes' every 'credit_usec'. */
unsigned long credit_bytes;
@@ -360,6 +356,7 @@ int xenvif_dealloc_kthread(void *data);
int xenvif_ctrl_kthread(void *data);
+void xenvif_rx_action(struct xenvif_queue *queue);
void xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb);
void xenvif_carrier_on(struct xenvif *vif);
diff --git a/drivers/net/xen-netback/rx.c b/drivers/net/xen-netback/rx.c
index 6bd7d6e..b0ce4c6 100644
--- a/drivers/net/xen-netback/rx.c
+++ b/drivers/net/xen-netback/rx.c
@@ -26,7 +26,6 @@
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
-
#include "common.h"
#include <linux/kthread.h>
@@ -137,464 +136,299 @@ static void xenvif_rx_queue_drop_expired(struct xenvif_queue *queue)
}
}
-struct netrx_pending_operations {
- unsigned int copy_prod, copy_cons;
- unsigned int meta_prod, meta_cons;
- struct gnttab_copy *copy;
- struct xenvif_rx_meta *meta;
- int copy_off;
- grant_ref_t copy_gref;
-};
-
-static struct xenvif_rx_meta *get_next_rx_buffer(
- struct xenvif_queue *queue,
- struct netrx_pending_operations *npo)
+static void xenvif_rx_copy_flush(struct xenvif_queue *queue)
{
- struct xenvif_rx_meta *meta;
- struct xen_netif_rx_request req;
+ unsigned int i;
- RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
+ gnttab_batch_copy(queue->rx_copy.op, queue->rx_copy.num);
- meta = npo->meta + npo->meta_prod++;
- meta->gso_type = XEN_NETIF_GSO_TYPE_NONE;
- meta->gso_size = 0;
- meta->size = 0;
- meta->id = req.id;
+ for (i = 0; i < queue->rx_copy.num; i++) {
+ struct gnttab_copy *op;
- npo->copy_off = 0;
- npo->copy_gref = req.gref;
+ op = &queue->rx_copy.op[i];
- return meta;
+ /* If the copy failed, overwrite the status field in
+ * the corresponding response.
+ */
+ if (unlikely(op->status != GNTST_okay)) {
+ struct xen_netif_rx_response *rsp;
+
+ rsp = RING_GET_RESPONSE(&queue->rx,
+ queue->rx_copy.idx[i]);
+ rsp->status = op->status;
+ }
+ }
+
+ queue->rx_copy.num = 0;
}
-struct gop_frag_copy {
- struct xenvif_queue *queue;
- struct netrx_pending_operations *npo;
- struct xenvif_rx_meta *meta;
- int head;
- int gso_type;
- int protocol;
- int hash_present;
-
- struct page *page;
-};
-
-static void xenvif_setup_copy_gop(unsigned long gfn,
- unsigned int offset,
- unsigned int *len,
- struct gop_frag_copy *info)
+static void xenvif_rx_copy_add(struct xenvif_queue *queue,
+ struct xen_netif_rx_request *req,
+ unsigned int offset, void *data, size_t len)
{
- struct gnttab_copy *copy_gop;
+ struct gnttab_copy *op;
+ struct page *page;
struct xen_page_foreign *foreign;
- /* Convenient aliases */
- struct xenvif_queue *queue = info->queue;
- struct netrx_pending_operations *npo = info->npo;
- struct page *page = info->page;
- WARN_ON(npo->copy_off > MAX_BUFFER_OFFSET);
+ if (queue->rx_copy.num == COPY_BATCH_SIZE)
+ xenvif_rx_copy_flush(queue);
- if (npo->copy_off == MAX_BUFFER_OFFSET)
- info->meta = get_next_rx_buffer(queue, npo);
+ op = &queue->rx_copy.op[queue->rx_copy.num];
- if (npo->copy_off + *len > MAX_BUFFER_OFFSET)
- *len = MAX_BUFFER_OFFSET - npo->copy_off;
+ page = virt_to_page(data);
- copy_gop = npo->copy + npo->copy_prod++;
- copy_gop->flags = GNTCOPY_dest_gref;
- copy_gop->len = *len;
+ op->flags = GNTCOPY_dest_gref;
foreign = xen_page_foreign(page);
if (foreign) {
- copy_gop->source.domid = foreign->domid;
- copy_gop->source.u.ref = foreign->gref;
- copy_gop->flags |= GNTCOPY_source_gref;
+ op->source.domid = foreign->domid;
+ op->source.u.ref = foreign->gref;
+ op->flags |= GNTCOPY_source_gref;
} else {
- copy_gop->source.domid = DOMID_SELF;
- copy_gop->source.u.gmfn = gfn;
+ op->source.u.gmfn = virt_to_gfn(data);
+ op->source.domid = DOMID_SELF;
}
- copy_gop->source.offset = offset;
- copy_gop->dest.domid = queue->vif->domid;
- copy_gop->dest.offset = npo->copy_off;
- copy_gop->dest.u.ref = npo->copy_gref;
+ op->source.offset = xen_offset_in_page(data);
+ op->dest.u.ref = req->gref;
+ op->dest.domid = queue->vif->domid;
+ op->dest.offset = offset;
+ op->len = len;
- npo->copy_off += *len;
- info->meta->size += *len;
-
- if (!info->head)
- return;
-
- /* Leave a gap for the GSO descriptor. */
- if ((1 << info->gso_type) & queue->vif->gso_mask)
- queue->rx.req_cons++;
-
- /* Leave a gap for the hash extra segment. */
- if (info->hash_present)
- queue->rx.req_cons++;
-
- info->head = 0; /* There must be something in this buffer now */
+ queue->rx_copy.idx[queue->rx_copy.num] = queue->rx.req_cons;
+ queue->rx_copy.num++;
}
-static void xenvif_gop_frag_copy_grant(unsigned long gfn,
- unsigned int offset,
- unsigned int len,
- void *data)
+static unsigned int xenvif_gso_type(struct sk_buff *skb)
{
- unsigned int bytes;
-
- while (len) {
- bytes = len;
- xenvif_setup_copy_gop(gfn, offset, &bytes, data);
- offset += bytes;
- len -= bytes;
- }
-}
-
-/* Set up the grant operations for this fragment. If it's a flipping
- * interface, we also set up the unmap request from here.
- */
-static void xenvif_gop_frag_copy(struct xenvif_queue *queue,
- struct sk_buff *skb,
- struct netrx_pending_operations *npo,
- struct page *page, unsigned long size,
- unsigned long offset, int *head)
-{
- struct gop_frag_copy info = {
- .queue = queue,
- .npo = npo,
- .head = *head,
- .gso_type = XEN_NETIF_GSO_TYPE_NONE,
- /* xenvif_set_skb_hash() will have either set a s/w
- * hash or cleared the hash depending on
- * whether the the frontend wants a hash for this skb.
- */
- .hash_present = skb->sw_hash,
- };
- unsigned long bytes;
-
if (skb_is_gso(skb)) {
if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
- info.gso_type = XEN_NETIF_GSO_TYPE_TCPV4;
- else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
- info.gso_type = XEN_NETIF_GSO_TYPE_TCPV6;
- }
-
- /* Data must not cross a page boundary. */
- WARN_ON(size + offset > (PAGE_SIZE << compound_order(page)));
-
- info.meta = npo->meta + npo->meta_prod - 1;
-
- /* Skip unused frames from start of page */
- page += offset >> PAGE_SHIFT;
- offset &= ~PAGE_MASK;
-
- while (size > 0) {
- WARN_ON(offset >= PAGE_SIZE);
-
- bytes = PAGE_SIZE - offset;
- if (bytes > size)
- bytes = size;
-
- info.page = page;
- gnttab_foreach_grant_in_range(page, offset, bytes,
- xenvif_gop_frag_copy_grant,
- &info);
- size -= bytes;
- offset = 0;
-
- /* Next page */
- if (size) {
- WARN_ON(!PageCompound(page));
- page++;
- }
- }
-
- *head = info.head;
-}
-
-/* Prepare an SKB to be transmitted to the frontend.
- *
- * This function is responsible for allocating grant operations, meta
- * structures, etc.
- *
- * It returns the number of meta structures consumed. The number of
- * ring slots used is always equal to the number of meta slots used
- * plus the number of GSO descriptors used. Currently, we use either
- * zero GSO descriptors (for non-GSO packets) or one descriptor (for
- * frontend-side LRO).
- */
-static int xenvif_gop_skb(struct sk_buff *skb,
- struct netrx_pending_operations *npo,
- struct xenvif_queue *queue)
-{
- struct xenvif *vif = netdev_priv(skb->dev);
- int nr_frags = skb_shinfo(skb)->nr_frags;
- int i;
- struct xen_netif_rx_request req;
- struct xenvif_rx_meta *meta;
- unsigned char *data;
- int head = 1;
- int old_meta_prod;
- int gso_type;
-
- old_meta_prod = npo->meta_prod;
-
- gso_type = XEN_NETIF_GSO_TYPE_NONE;
- if (skb_is_gso(skb)) {
- if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
- gso_type = XEN_NETIF_GSO_TYPE_TCPV4;
- else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
- gso_type = XEN_NETIF_GSO_TYPE_TCPV6;
- }
-
- RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
- meta = npo->meta + npo->meta_prod++;
-
- if ((1 << gso_type) & vif->gso_mask) {
- meta->gso_type = gso_type;
- meta->gso_size = skb_shinfo(skb)->gso_size;
- } else {
- meta->gso_type = XEN_NETIF_GSO_TYPE_NONE;
- meta->gso_size = 0;
- }
-
- meta->size = 0;
- meta->id = req.id;
- npo->copy_off = 0;
- npo->copy_gref = req.gref;
-
- data = skb->data;
- while (data < skb_tail_pointer(skb)) {
- unsigned int offset = offset_in_page(data);
- unsigned int len = PAGE_SIZE - offset;
-
- if (data + len > skb_tail_pointer(skb))
- len = skb_tail_pointer(skb) - data;
-
- xenvif_gop_frag_copy(queue, skb, npo,
- virt_to_page(data), len, offset, &head);
- data += len;
- }
-
- for (i = 0; i < nr_frags; i++) {
- xenvif_gop_frag_copy(queue, skb, npo,
- skb_frag_page(&skb_shinfo(skb)->frags[i]),
- skb_frag_size(&skb_shinfo(skb)->frags[i]),
- skb_shinfo(skb)->frags[i].page_offset,
- &head);
- }
-
- return npo->meta_prod - old_meta_prod;
-}
-
-/* This is a twin to xenvif_gop_skb. Assume that xenvif_gop_skb was
- * used to set up the operations on the top of
- * netrx_pending_operations, which have since been done. Check that
- * they didn't give any errors and advance over them.
- */
-static int xenvif_check_gop(struct xenvif *vif, int nr_meta_slots,
- struct netrx_pending_operations *npo)
-{
- struct gnttab_copy *copy_op;
- int status = XEN_NETIF_RSP_OKAY;
- int i;
-
- for (i = 0; i < nr_meta_slots; i++) {
- copy_op = npo->copy + npo->copy_cons++;
- if (copy_op->status != GNTST_okay) {
- netdev_dbg(vif->dev,
- "Bad status %d from copy to DOM%d.\n",
- copy_op->status, vif->domid);
- status = XEN_NETIF_RSP_ERROR;
- }
- }
-
- return status;
-}
-
-static struct xen_netif_rx_response *make_rx_response(
- struct xenvif_queue *queue, u16 id, s8 st, u16 offset, u16 size,
- u16 flags)
-{
- RING_IDX i = queue->rx.rsp_prod_pvt;
- struct xen_netif_rx_response *resp;
-
- resp = RING_GET_RESPONSE(&queue->rx, i);
- resp->offset = offset;
- resp->flags = flags;
- resp->id = id;
- resp->status = (s16)size;
- if (st < 0)
- resp->status = (s16)st;
-
- queue->rx.rsp_prod_pvt = ++i;
-
- return resp;
-}
-
-static void xenvif_add_frag_responses(struct xenvif_queue *queue,
- int status,
- struct xenvif_rx_meta *meta,
- int nr_meta_slots)
-{
- int i;
- unsigned long offset;
-
- /* No fragments used */
- if (nr_meta_slots <= 1)
- return;
-
- nr_meta_slots--;
-
- for (i = 0; i < nr_meta_slots; i++) {
- int flags;
-
- if (i == nr_meta_slots - 1)
- flags = 0;
+ return XEN_NETIF_GSO_TYPE_TCPV4;
else
- flags = XEN_NETRXF_more_data;
-
- offset = 0;
- make_rx_response(queue, meta[i].id, status, offset,
- meta[i].size, flags);
+ return XEN_NETIF_GSO_TYPE_TCPV6;
}
+ return XEN_NETIF_GSO_TYPE_NONE;
}
-static void xenvif_rx_action(struct xenvif_queue *queue)
-{
- struct xenvif *vif = queue->vif;
- s8 status;
- u16 flags;
- struct xen_netif_rx_response *resp;
- struct sk_buff_head rxq;
+struct xenvif_pkt_state {
struct sk_buff *skb;
- LIST_HEAD(notify);
- int ret;
- unsigned long offset;
- bool need_to_notify = false;
+ size_t remaining_len;
+ int frag; /* frag == -1 => skb->head */
+ unsigned int frag_offset;
+ struct xen_netif_extra_info extras[XEN_NETIF_EXTRA_TYPE_MAX - 1];
+ unsigned int extra_count;
+ unsigned int slot;
+};
- struct netrx_pending_operations npo = {
- .copy = queue->grant_copy_op,
- .meta = queue->meta,
- };
+static void xenvif_rx_next_skb(struct xenvif_queue *queue,
+ struct xenvif_pkt_state *pkt)
+{
+ struct sk_buff *skb;
+ unsigned int gso_type;
- skb_queue_head_init(&rxq);
+ skb = xenvif_rx_dequeue(queue);
- while (xenvif_rx_ring_slots_available(queue) &&
- (skb = xenvif_rx_dequeue(queue)) != NULL) {
- queue->last_rx_time = jiffies;
+ queue->stats.tx_bytes += skb->len;
+ queue->stats.tx_packets++;
- XENVIF_RX_CB(skb)->meta_slots_used =
- xenvif_gop_skb(skb, &npo, queue);
+ /* Reset packet state. */
+ memset(pkt, 0, sizeof(struct xenvif_pkt_state));
- __skb_queue_tail(&rxq, skb);
+ pkt->skb = skb;
+ pkt->remaining_len = skb->len;
+ pkt->frag = -1;
+
+ gso_type = xenvif_gso_type(skb);
+ if ((1 << gso_type) & queue->vif->gso_mask) {
+ struct xen_netif_extra_info *extra;
+
+ extra = &pkt->extras[XEN_NETIF_EXTRA_TYPE_GSO - 1];
+
+ extra->u.gso.type = gso_type;
+ extra->u.gso.size = skb_shinfo(skb)->gso_size;
+ extra->u.gso.pad = 0;
+ extra->u.gso.features = 0;
+ extra->type = XEN_NETIF_EXTRA_TYPE_GSO;
+ extra->flags = 0;
+
+ pkt->extra_count++;
}
- WARN_ON(npo.meta_prod > ARRAY_SIZE(queue->meta));
+ if (skb->sw_hash) {
+ struct xen_netif_extra_info *extra;
- if (!npo.copy_prod)
- goto done;
+ extra = &pkt->extras[XEN_NETIF_EXTRA_TYPE_HASH - 1];
- WARN_ON(npo.copy_prod > MAX_GRANT_COPY_OPS);
- gnttab_batch_copy(queue->grant_copy_op, npo.copy_prod);
+ extra->u.hash.algorithm =
+ XEN_NETIF_CTRL_HASH_ALGORITHM_TOEPLITZ;
- while ((skb = __skb_dequeue(&rxq)) != NULL) {
- struct xen_netif_extra_info *extra = NULL;
-
- queue->stats.tx_bytes += skb->len;
- queue->stats.tx_packets++;
-
- status = xenvif_check_gop(vif,
- XENVIF_RX_CB(skb)->meta_slots_used,
- &npo);
-
- if (XENVIF_RX_CB(skb)->meta_slots_used == 1)
- flags = 0;
+ if (skb->l4_hash)
+ extra->u.hash.type =
+ skb->protocol == htons(ETH_P_IP) ?
+ _XEN_NETIF_CTRL_HASH_TYPE_IPV4_TCP :
+ _XEN_NETIF_CTRL_HASH_TYPE_IPV6_TCP;
else
- flags = XEN_NETRXF_more_data;
+ extra->u.hash.type =
+ skb->protocol == htons(ETH_P_IP) ?
+ _XEN_NETIF_CTRL_HASH_TYPE_IPV4 :
+ _XEN_NETIF_CTRL_HASH_TYPE_IPV6;
- if (skb->ip_summed == CHECKSUM_PARTIAL) /* local packet? */
+ *(uint32_t *)extra->u.hash.value = skb_get_hash_raw(skb);
+
+ extra->type = XEN_NETIF_EXTRA_TYPE_HASH;
+ extra->flags = 0;
+
+ pkt->extra_count++;
+ }
+}
+
+static void xenvif_rx_complete(struct xenvif_queue *queue,
+ struct xenvif_pkt_state *pkt)
+{
+ int notify;
+
+ /* Complete any outstanding copy ops for this skb. */
+ xenvif_rx_copy_flush(queue);
+
+ /* Push responses and notify. */
+ queue->rx.rsp_prod_pvt = queue->rx.req_cons;
+ RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->rx, notify);
+ if (notify)
+ notify_remote_via_irq(queue->rx_irq);
+
+ dev_kfree_skb(pkt->skb);
+}
+
+static void xenvif_rx_next_chunk(struct xenvif_queue *queue,
+ struct xenvif_pkt_state *pkt,
+ unsigned int offset, void **data,
+ size_t *len)
+{
+ struct sk_buff *skb = pkt->skb;
+ void *frag_data;
+ size_t frag_len, chunk_len;
+
+ if (pkt->frag == -1) {
+ frag_data = skb->data;
+ frag_len = skb_headlen(skb);
+ } else {
+ skb_frag_t *frag = &skb_shinfo(skb)->frags[pkt->frag];
+
+ frag_data = skb_frag_address(frag);
+ frag_len = skb_frag_size(frag);
+ }
+
+ frag_data += pkt->frag_offset;
+ frag_len -= pkt->frag_offset;
+
+ chunk_len = min(frag_len, XEN_PAGE_SIZE - offset);
+ chunk_len = min(chunk_len,
+ XEN_PAGE_SIZE - xen_offset_in_page(frag_data));
+
+ pkt->frag_offset += chunk_len;
+
+ /* Advance to next frag? */
+ if (frag_len == chunk_len) {
+ pkt->frag++;
+ pkt->frag_offset = 0;
+ }
+
+ *data = frag_data;
+ *len = chunk_len;
+}
+
+static void xenvif_rx_data_slot(struct xenvif_queue *queue,
+ struct xenvif_pkt_state *pkt,
+ struct xen_netif_rx_request *req,
+ struct xen_netif_rx_response *rsp)
+{
+ unsigned int offset = 0;
+ unsigned int flags;
+
+ do {
+ size_t len;
+ void *data;
+
+ xenvif_rx_next_chunk(queue, pkt, offset, &data, &len);
+ xenvif_rx_copy_add(queue, req, offset, data, len);
+
+ offset += len;
+ pkt->remaining_len -= len;
+
+ } while (offset < XEN_PAGE_SIZE && pkt->remaining_len > 0);
+
+ if (pkt->remaining_len > 0)
+ flags = XEN_NETRXF_more_data;
+ else
+ flags = 0;
+
+ if (pkt->slot == 0) {
+ struct sk_buff *skb = pkt->skb;
+
+ if (skb->ip_summed == CHECKSUM_PARTIAL)
flags |= XEN_NETRXF_csum_blank |
XEN_NETRXF_data_validated;
else if (skb->ip_summed == CHECKSUM_UNNECESSARY)
- /* remote but checksummed. */
flags |= XEN_NETRXF_data_validated;
- offset = 0;
- resp = make_rx_response(queue, queue->meta[npo.meta_cons].id,
- status, offset,
- queue->meta[npo.meta_cons].size,
- flags);
-
- if ((1 << queue->meta[npo.meta_cons].gso_type) &
- vif->gso_mask) {
- extra = (struct xen_netif_extra_info *)
- RING_GET_RESPONSE(&queue->rx,
- queue->rx.rsp_prod_pvt++);
-
- resp->flags |= XEN_NETRXF_extra_info;
-
- extra->u.gso.type = queue->meta[npo.meta_cons].gso_type;
- extra->u.gso.size = queue->meta[npo.meta_cons].gso_size;
- extra->u.gso.pad = 0;
- extra->u.gso.features = 0;
-
- extra->type = XEN_NETIF_EXTRA_TYPE_GSO;
- extra->flags = 0;
- }
-
- if (skb->sw_hash) {
- /* Since the skb got here via xenvif_select_queue()
- * we know that the hash has been re-calculated
- * according to a configuration set by the frontend
- * and therefore we know that it is legitimate to
- * pass it to the frontend.
- */
- if (resp->flags & XEN_NETRXF_extra_info)
- extra->flags |= XEN_NETIF_EXTRA_FLAG_MORE;
- else
- resp->flags |= XEN_NETRXF_extra_info;
-
- extra = (struct xen_netif_extra_info *)
- RING_GET_RESPONSE(&queue->rx,
- queue->rx.rsp_prod_pvt++);
-
- extra->u.hash.algorithm =
- XEN_NETIF_CTRL_HASH_ALGORITHM_TOEPLITZ;
-
- if (skb->l4_hash)
- extra->u.hash.type =
- skb->protocol == htons(ETH_P_IP) ?
- _XEN_NETIF_CTRL_HASH_TYPE_IPV4_TCP :
- _XEN_NETIF_CTRL_HASH_TYPE_IPV6_TCP;
- else
- extra->u.hash.type =
- skb->protocol == htons(ETH_P_IP) ?
- _XEN_NETIF_CTRL_HASH_TYPE_IPV4 :
- _XEN_NETIF_CTRL_HASH_TYPE_IPV6;
-
- *(uint32_t *)extra->u.hash.value =
- skb_get_hash_raw(skb);
-
- extra->type = XEN_NETIF_EXTRA_TYPE_HASH;
- extra->flags = 0;
- }
-
- xenvif_add_frag_responses(queue, status,
- queue->meta + npo.meta_cons + 1,
- XENVIF_RX_CB(skb)->meta_slots_used);
-
- RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&queue->rx, ret);
-
- need_to_notify |= !!ret;
-
- npo.meta_cons += XENVIF_RX_CB(skb)->meta_slots_used;
- dev_kfree_skb(skb);
+ if (pkt->extra_count != 0)
+ flags |= XEN_NETRXF_extra_info;
}
-done:
- if (need_to_notify)
- notify_remote_via_irq(queue->rx_irq);
+ rsp->offset = 0;
+ rsp->flags = flags;
+ rsp->id = req->id;
+ rsp->status = (s16)offset;
+}
+
+static void xenvif_rx_extra_slot(struct xenvif_queue *queue,
+ struct xenvif_pkt_state *pkt,
+ struct xen_netif_rx_request *req,
+ struct xen_netif_rx_response *rsp)
+{
+ struct xen_netif_extra_info *extra = (void *)rsp;
+ unsigned int i;
+
+ pkt->extra_count--;
+
+ for (i = 0; i < ARRAY_SIZE(pkt->extras); i++) {
+ if (pkt->extras[i].type) {
+ *extra = pkt->extras[i];
+
+ if (pkt->extra_count != 0)
+ extra->flags |= XEN_NETIF_EXTRA_FLAG_MORE;
+
+ pkt->extras[i].type = 0;
+ return;
+ }
+ }
+ BUG();
+}
+
+void xenvif_rx_action(struct xenvif_queue *queue)
+{
+ struct xenvif_pkt_state pkt;
+
+ xenvif_rx_next_skb(queue, &pkt);
+
+ do {
+ struct xen_netif_rx_request *req;
+ struct xen_netif_rx_response *rsp;
+
+ req = RING_GET_REQUEST(&queue->rx, queue->rx.req_cons);
+ rsp = RING_GET_RESPONSE(&queue->rx, queue->rx.req_cons);
+
+ /* Extras must go after the first data slot */
+ if (pkt.slot != 0 && pkt.extra_count != 0)
+ xenvif_rx_extra_slot(queue, &pkt, req, rsp);
+ else
+ xenvif_rx_data_slot(queue, &pkt, req, rsp);
+
+ queue->rx.req_cons++;
+ pkt.slot++;
+ } while (pkt.remaining_len > 0 || pkt.extra_count != 0);
+
+ xenvif_rx_complete(queue, &pkt);
}
static bool xenvif_rx_queue_stalled(struct xenvif_queue *queue)
--
2.1.4
^ permalink raw reply related
* [PATCH net-next 2/7] xen-netback: retire guest rx side prefix GSO feature
From: Paul Durrant @ 2016-10-03 7:31 UTC (permalink / raw)
To: netdev, xen-devel; +Cc: Paul Durrant, Wei Liu
In-Reply-To: <1475479872-23717-1-git-send-email-paul.durrant@citrix.com>
As far as I am aware only very old Windows network frontends make use of
this style of passing GSO packets from backend to frontend. These
frontends can easily be replaced by the freely available Xen Project
Windows PV network frontend, which uses the 'default' mechanism for
passing GSO packets, which is also used by all Linux frontends.
NOTE: Removal of this feature will not cause breakage in old Windows
frontends. They simply will no longer receive GSO packets - the
packets instead being fragmented in the backend.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
Cc: Wei Liu <wei.liu2@citrix.com>
---
drivers/net/xen-netback/common.h | 1 -
drivers/net/xen-netback/interface.c | 4 ++--
drivers/net/xen-netback/rx.c | 26 --------------------------
drivers/net/xen-netback/xenbus.c | 21 ---------------------
4 files changed, 2 insertions(+), 50 deletions(-)
diff --git a/drivers/net/xen-netback/common.h b/drivers/net/xen-netback/common.h
index 3a56268..e16004a 100644
--- a/drivers/net/xen-netback/common.h
+++ b/drivers/net/xen-netback/common.h
@@ -260,7 +260,6 @@ struct xenvif {
/* Frontend feature information. */
int gso_mask;
- int gso_prefix_mask;
u8 can_sg:1;
u8 ip_csum:1;
diff --git a/drivers/net/xen-netback/interface.c b/drivers/net/xen-netback/interface.c
index 83deeeb..1a009e7 100644
--- a/drivers/net/xen-netback/interface.c
+++ b/drivers/net/xen-netback/interface.c
@@ -328,9 +328,9 @@ static netdev_features_t xenvif_fix_features(struct net_device *dev,
if (!vif->can_sg)
features &= ~NETIF_F_SG;
- if (~(vif->gso_mask | vif->gso_prefix_mask) & GSO_BIT(TCPV4))
+ if (~(vif->gso_mask) & GSO_BIT(TCPV4))
features &= ~NETIF_F_TSO;
- if (~(vif->gso_mask | vif->gso_prefix_mask) & GSO_BIT(TCPV6))
+ if (~(vif->gso_mask) & GSO_BIT(TCPV6))
features &= ~NETIF_F_TSO6;
if (!vif->ip_csum)
features &= ~NETIF_F_IP_CSUM;
diff --git a/drivers/net/xen-netback/rx.c b/drivers/net/xen-netback/rx.c
index 03836aa..6bd7d6e 100644
--- a/drivers/net/xen-netback/rx.c
+++ b/drivers/net/xen-netback/rx.c
@@ -347,16 +347,6 @@ static int xenvif_gop_skb(struct sk_buff *skb,
gso_type = XEN_NETIF_GSO_TYPE_TCPV6;
}
- /* Set up a GSO prefix descriptor, if necessary */
- if ((1 << gso_type) & vif->gso_prefix_mask) {
- RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
- meta = npo->meta + npo->meta_prod++;
- meta->gso_type = gso_type;
- meta->gso_size = skb_shinfo(skb)->gso_size;
- meta->size = 0;
- meta->id = req.id;
- }
-
RING_COPY_REQUEST(&queue->rx, queue->rx.req_cons++, &req);
meta = npo->meta + npo->meta_prod++;
@@ -511,22 +501,6 @@ static void xenvif_rx_action(struct xenvif_queue *queue)
while ((skb = __skb_dequeue(&rxq)) != NULL) {
struct xen_netif_extra_info *extra = NULL;
- if ((1 << queue->meta[npo.meta_cons].gso_type) &
- vif->gso_prefix_mask) {
- resp = RING_GET_RESPONSE(&queue->rx,
- queue->rx.rsp_prod_pvt++);
-
- resp->flags = XEN_NETRXF_gso_prefix |
- XEN_NETRXF_more_data;
-
- resp->offset = queue->meta[npo.meta_cons].gso_size;
- resp->id = queue->meta[npo.meta_cons].id;
- resp->status = XENVIF_RX_CB(skb)->meta_slots_used;
-
- npo.meta_cons++;
- XENVIF_RX_CB(skb)->meta_slots_used--;
- }
-
queue->stats.tx_bytes += skb->len;
queue->stats.tx_packets++;
diff --git a/drivers/net/xen-netback/xenbus.c b/drivers/net/xen-netback/xenbus.c
index bacf6e0..6c57b02 100644
--- a/drivers/net/xen-netback/xenbus.c
+++ b/drivers/net/xen-netback/xenbus.c
@@ -1154,7 +1154,6 @@ static int read_xenbus_vif_flags(struct backend_info *be)
vif->can_sg = !!val;
vif->gso_mask = 0;
- vif->gso_prefix_mask = 0;
if (xenbus_scanf(XBT_NIL, dev->otherend, "feature-gso-tcpv4",
"%d", &val) < 0)
@@ -1162,32 +1161,12 @@ static int read_xenbus_vif_flags(struct backend_info *be)
if (val)
vif->gso_mask |= GSO_BIT(TCPV4);
- if (xenbus_scanf(XBT_NIL, dev->otherend, "feature-gso-tcpv4-prefix",
- "%d", &val) < 0)
- val = 0;
- if (val)
- vif->gso_prefix_mask |= GSO_BIT(TCPV4);
-
if (xenbus_scanf(XBT_NIL, dev->otherend, "feature-gso-tcpv6",
"%d", &val) < 0)
val = 0;
if (val)
vif->gso_mask |= GSO_BIT(TCPV6);
- if (xenbus_scanf(XBT_NIL, dev->otherend, "feature-gso-tcpv6-prefix",
- "%d", &val) < 0)
- val = 0;
- if (val)
- vif->gso_prefix_mask |= GSO_BIT(TCPV6);
-
- if (vif->gso_mask & vif->gso_prefix_mask) {
- xenbus_dev_fatal(dev, err,
- "%s: gso and gso prefix flags are not "
- "mutually exclusive",
- dev->otherend);
- return -EOPNOTSUPP;
- }
-
if (xenbus_scanf(XBT_NIL, dev->otherend, "feature-no-csum-offload",
"%d", &val) < 0)
val = 0;
--
2.1.4
^ permalink raw reply related
* [PATCH net-next 5/7] xen-netback: process guest rx packets in batches
From: Paul Durrant @ 2016-10-03 7:31 UTC (permalink / raw)
To: netdev, xen-devel; +Cc: David Vrabel, Paul Durrant, Wei Liu
In-Reply-To: <1475479872-23717-1-git-send-email-paul.durrant@citrix.com>
From: David Vrabel <david.vrabel@citrix.com>
Instead of only placing one skb on the guest rx ring at a time, process
a batch of up-to 64. This improves performance by ~10% in some tests.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
[re-based]
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
Cc: Wei Liu <wei.liu2@citrix.com>
---
drivers/net/xen-netback/rx.c | 15 ++++++++++++++-
1 file changed, 14 insertions(+), 1 deletion(-)
diff --git a/drivers/net/xen-netback/rx.c b/drivers/net/xen-netback/rx.c
index 9548709..ae822b8 100644
--- a/drivers/net/xen-netback/rx.c
+++ b/drivers/net/xen-netback/rx.c
@@ -399,7 +399,7 @@ static void xenvif_rx_extra_slot(struct xenvif_queue *queue,
BUG();
}
-void xenvif_rx_action(struct xenvif_queue *queue)
+void xenvif_rx_skb(struct xenvif_queue *queue)
{
struct xenvif_pkt_state pkt;
@@ -425,6 +425,19 @@ void xenvif_rx_action(struct xenvif_queue *queue)
xenvif_rx_complete(queue, &pkt);
}
+#define RX_BATCH_SIZE 64
+
+void xenvif_rx_action(struct xenvif_queue *queue)
+{
+ unsigned int work_done = 0;
+
+ while (xenvif_rx_ring_slots_available(queue) &&
+ work_done < RX_BATCH_SIZE) {
+ xenvif_rx_skb(queue);
+ work_done++;
+ }
+}
+
static bool xenvif_rx_queue_stalled(struct xenvif_queue *queue)
{
RING_IDX prod, cons;
--
2.1.4
^ permalink raw reply related
* [PATCH net-next 4/7] xen-netback: immediately wake tx queue when guest rx queue has space
From: Paul Durrant @ 2016-10-03 7:31 UTC (permalink / raw)
To: netdev, xen-devel; +Cc: David Vrabel, Paul Durrant, Wei Liu
In-Reply-To: <1475479872-23717-1-git-send-email-paul.durrant@citrix.com>
From: David Vrabel <david.vrabel@citrix.com>
When an skb is removed from the guest rx queue, immediately wake the
tx queue, instead of after processing them.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
[re-based]
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
Cc: Wei Liu <wei.liu2@citrix.com>
---
drivers/net/xen-netback/rx.c | 24 ++++++++----------------
1 file changed, 8 insertions(+), 16 deletions(-)
diff --git a/drivers/net/xen-netback/rx.c b/drivers/net/xen-netback/rx.c
index b0ce4c6..9548709 100644
--- a/drivers/net/xen-netback/rx.c
+++ b/drivers/net/xen-netback/rx.c
@@ -92,27 +92,21 @@ static struct sk_buff *xenvif_rx_dequeue(struct xenvif_queue *queue)
spin_lock_irq(&queue->rx_queue.lock);
skb = __skb_dequeue(&queue->rx_queue);
- if (skb)
+ if (skb) {
queue->rx_queue_len -= skb->len;
+ if (queue->rx_queue_len < queue->rx_queue_max) {
+ struct netdev_queue *txq;
+
+ txq = netdev_get_tx_queue(queue->vif->dev, queue->id);
+ netif_tx_wake_queue(txq);
+ }
+ }
spin_unlock_irq(&queue->rx_queue.lock);
return skb;
}
-static void xenvif_rx_queue_maybe_wake(struct xenvif_queue *queue)
-{
- spin_lock_irq(&queue->rx_queue.lock);
-
- if (queue->rx_queue_len < queue->rx_queue_max) {
- struct net_device *dev = queue->vif->dev;
-
- netif_tx_wake_queue(netdev_get_tx_queue(dev, queue->id));
- }
-
- spin_unlock_irq(&queue->rx_queue.lock);
-}
-
static void xenvif_rx_queue_purge(struct xenvif_queue *queue)
{
struct sk_buff *skb;
@@ -585,8 +579,6 @@ int xenvif_kthread_guest_rx(void *data)
*/
xenvif_rx_queue_drop_expired(queue);
- xenvif_rx_queue_maybe_wake(queue);
-
cond_resched();
}
--
2.1.4
^ permalink raw reply related
* [net-next] i40e: avoid NULL pointer dereference and recursive errors on early PCI error
From: Jeff Kirsher @ 2016-10-03 7:31 UTC (permalink / raw)
To: davem
Cc: Guilherme G Piccoli, netdev, nhorman, sassmann, jogreene,
guru.anbalagane, stable, Jeff Kirsher
From: Guilherme G Piccoli <gpiccoli@linux.vnet.ibm.com>
Although rare, it's possible to hit PCI error early on device
probe, meaning possibly some structs are not entirely initialized,
and some might even be completely uninitialized, leading to NULL
pointer dereference.
The i40e driver currently presents a "bad" behavior if device hits
such early PCI error: firstly, the struct i40e_pf might not be
attached to pci_dev yet, leading to a NULL pointer dereference on
access to pf->state.
Even checking if the struct is NULL and avoiding the access in that
case isn't enough, since the driver cannot recover from PCI error
that early; in our experiments we saw multiple failures on kernel
log, like:
[549.664] i40e 0007:01:00.1: Initial pf_reset failed: -15
[549.664] i40e: probe of 0007:01:00.1 failed with error -15
[...]
[871.644] i40e 0007:01:00.1: The driver for the device stopped because the
device firmware failed to init. Try updating your NVM image.
[871.644] i40e: probe of 0007:01:00.1 failed with error -32
[...]
[872.516] i40e 0007:01:00.0: ARQ: Unknown event 0x0000 ignored
Between the first probe failure (error -15) and the second (error -32)
another PCI error happened due to the first bad probe. Also, driver
started to flood console with those ARQ event messages.
This patch will prevent these issues by allowing error recovery
mechanism to remove the failed device from the system instead of
trying to recover from early PCI errors during device probe.
CC: <stable@vger.kernel.org>
Signed-off-by: Guilherme G Piccoli <gpiccoli@linux.vnet.ibm.com>
Acked-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_main.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 8176596..efc0a36 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -11382,6 +11382,12 @@ static pci_ers_result_t i40e_pci_error_detected(struct pci_dev *pdev,
dev_info(&pdev->dev, "%s: error %d\n", __func__, error);
+ if (!pf) {
+ dev_info(&pdev->dev,
+ "Cannot recover - error happened during device probe\n");
+ return PCI_ERS_RESULT_DISCONNECT;
+ }
+
/* shutdown all operations */
if (!test_bit(__I40E_SUSPENDED, &pf->state)) {
rtnl_lock();
--
2.7.4
^ 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