* Re: [PATCHv2 bluetooth-next 07/10] ipv6: introduce neighbour discovery ops
From: Alexander Aring @ 2016-05-06 9:47 UTC (permalink / raw)
To: Stefan Schmidt, linux-wpan-u79uwXL29TY76Z2rM5mHXA
Cc: kernel-bIcnvbaLZ9MEGnE8C9+IrQ, marcel-kz+m5ild9QBg9hUCZPvPmw,
jukka.rissanen-VuQAYsv1563Yd54FQh9/CA,
hannes-tFNcAqjVMyqKXQKiL6tip0B+6BGkLq7r,
mcr-SWp7JaYWvAQV+D8aMU/kSg, werner-SEdMjqphH88wryQfseakQg,
linux-bluetooth-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA, David S . Miller, Alexey Kuznetsov,
James Morris, Hideaki YOSHIFUJI, Patrick McHardy
In-Reply-To: <5729E9AB.5000301-JPH+aEBZ4P+UEJcrhfAQsw@public.gmane.org>
Hi,
On 05/04/2016 02:23 PM, Stefan Schmidt wrote:
> Hello.
>
> On 20/04/16 10:19, Alexander Aring wrote:
>> This patch introduces neighbour discovery ops callback structure. The
>> structure contains at first receive and transmit handling for NS/NA and
>> userspace option field functionality.
>>
>> These callback offers 6lowpan different handling, such as 802.15.4 short
>> address handling or RFC6775 (Neighbor Discovery Optimization for IPv6 over
>> 6LoWPANs).
>>
>> Cc: David S. Miller<davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org>
>> Cc: Alexey Kuznetsov<kuznet-v/Mj1YrvjDBInbfyfbPRSQ@public.gmane.org>
>> Cc: James Morris<jmorris-gx6/JNMH7DfYtjvyW6yDsg@public.gmane.org>
>> Cc: Hideaki YOSHIFUJI<yoshfuji-VfPWfsRibaP+Ru+s062T9g@public.gmane.org>
>> Cc: Patrick McHardy<kaber-dcUjhNyLwpNeoWH0uzbU5w@public.gmane.org>
>> Signed-off-by: Alexander Aring<aar-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
>> ---
>> include/linux/netdevice.h | 3 ++
>> include/net/ndisc.h | 96 +++++++++++++++++++++++++++++++++++++++++++----
>> net/ipv6/addrconf.c | 1 +
>> net/ipv6/ndisc.c | 71 ++++++++++++++++++++++++-----------
>> net/ipv6/route.c | 2 +-
>> 5 files changed, 144 insertions(+), 29 deletions(-)
>>
>> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
>> index 0052c42..bc60033 100644
>> --- a/include/linux/netdevice.h
>> +++ b/include/linux/netdevice.h
>> @@ -1677,6 +1677,9 @@ struct net_device {
>> #ifdef CONFIG_NET_L3_MASTER_DEV
>> const struct l3mdev_ops *l3mdev_ops;
>> #endif
>> +#if IS_ENABLED(CONFIG_IPV6)
>> + const struct ndisc_ops *ndisc_ops;
>> +#endif
>> const struct header_ops *header_ops;
>> diff --git a/include/net/ndisc.h b/include/net/ndisc.h
>> index aac868e..14ed016 100644
>> --- a/include/net/ndisc.h
>> +++ b/include/net/ndisc.h
>> @@ -110,7 +110,8 @@ struct ndisc_options {
>> #define NDISC_OPT_SPACE(len) (((len)+2+7)&~7)
>> -struct ndisc_options *ndisc_parse_options(u8 *opt, int opt_len,
>> +struct ndisc_options *ndisc_parse_options(const struct net_device *dev,
>> + u8 *opt, int opt_len,
>> struct ndisc_options *ndopts);
>> /*
>> @@ -173,6 +174,93 @@ static inline struct neighbour *__ipv6_neigh_lookup(struct net_device *dev, cons
>> return n;
>> }
>> +static inline int __ip6_ndisc_is_useropt(struct nd_opt_hdr *opt)
>
> Name it __ipv6... instead of __ip6...?
I had ipv6 before, but then I saw ip6... prefixes functionality in
ndisc.c and changed it to ip6...., but both seems to be used.
See "ip6_nd_hdr".
>> +{
>> + return opt->nd_opt_type == ND_OPT_RDNSS ||
>> + opt->nd_opt_type == ND_OPT_DNSSL;
>> +}
>> +
>> +#if IS_ENABLED(CONFIG_IPV6)
>> +struct ndisc_ops {
>> + int (*is_useropt)(struct nd_opt_hdr *opt);
>> + void (*send_na)(struct net_device *dev,
>> + const struct in6_addr *daddr,
>> + const struct in6_addr *solicited_addr,
>> + bool router, bool solicited,
>> + bool override, bool inc_opt);
>> + void (*recv_na)(struct sk_buff *skb);
>> + void (*send_ns)(struct net_device *dev,
>> + const struct in6_addr *solicit,
>> + const struct in6_addr *daddr,
>> + const struct in6_addr *saddr);
>> + void (*recv_ns)(struct sk_buff *skb);
>> +};
>> +
>> +static inline int ndisc_is_useropt(const struct net_device *dev,
>> + struct nd_opt_hdr *opt)
>> +{
>> + if (likely(dev->ndisc_ops->is_useropt))
>> + return dev->ndisc_ops->is_useropt(opt);
>> + else
>> + return 0;
>> +}
>> +
>> +static inline void ndisc_send_na(struct net_device *dev,
>> + const struct in6_addr *daddr,
>> + const struct in6_addr *solicited_addr,
>> + bool router, bool solicited, bool override,
>> + bool inc_opt)
>> +{
>> + if (likely(dev->ndisc_ops->send_na))
>> + dev->ndisc_ops->send_na(dev, daddr, solicited_addr, router,
>> + solicited, override, inc_opt);
>> +}
>> +
>> +static inline void ndisc_recv_na(struct sk_buff *skb)
>> +{
>> + if (likely(skb->dev->ndisc_ops->recv_na))
>> + skb->dev->ndisc_ops->recv_na(skb);
>> +}
>> +
>> +static inline void ndisc_send_ns(struct net_device *dev,
>> + const struct in6_addr *solicit,
>> + const struct in6_addr *daddr,
>> + const struct in6_addr *saddr)
>> +{
>> + if (likely(dev->ndisc_ops->send_ns))
>> + dev->ndisc_ops->send_ns(dev, solicit, daddr, saddr);
>> +}
>> +
>> +static inline void ndisc_recv_ns(struct sk_buff *skb)
>> +{
>> + if (likely(skb->dev->ndisc_ops->recv_ns))
>> + skb->dev->ndisc_ops->recv_ns(skb);
>> +}
>> +#else
>> +static inline int ndisc_is_useropt(const struct net_device *dev,
>> + struct nd_opt_hdr *opt)
>> +{
>> + return 0;
>> +}
>> +
>> +static inline void ndisc_send_na(struct net_device *dev,
>> + const struct in6_addr *daddr,
>> + const struct in6_addr *solicited_addr,
>> + bool router, bool solicited, bool override,
>> + bool inc_opt) { }
>> +
>> +static inline void ndisc_recv_na(struct sk_buff *skb) { }
>> +
>> +static inline void ndisc_send_ns(struct net_device *dev,
>> + const struct in6_addr *solicit,
>> + const struct in6_addr *daddr,
>> + const struct in6_addr *saddr) { }
>> +
>> +static inline void ndisc_recv_ns(struct sk_buff *skb) { }
>> +#endif
>> +
>> +void ip6_register_ndisc_ops(struct net_device *dev);
>> +
>> int ndisc_init(void);
>> int ndisc_late_init(void);
>> @@ -181,14 +269,8 @@ void ndisc_cleanup(void);
>> int ndisc_rcv(struct sk_buff *skb);
>> -void ndisc_send_ns(struct net_device *dev, const struct in6_addr *solicit,
>> - const struct in6_addr *daddr, const struct in6_addr *saddr);
>> -
>> void ndisc_send_rs(struct net_device *dev,
>> const struct in6_addr *saddr, const struct in6_addr *daddr);
>> -void ndisc_send_na(struct net_device *dev, const struct in6_addr *daddr,
>> - const struct in6_addr *solicited_addr,
>> - bool router, bool solicited, bool override, bool inc_opt);
>> void ndisc_send_redirect(struct sk_buff *skb, const struct in6_addr *target);
>> diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
>> index 54e18c2..a2ef04b 100644
>> --- a/net/ipv6/addrconf.c
>> +++ b/net/ipv6/addrconf.c
>> @@ -3266,6 +3266,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
>> idev = ipv6_add_dev(dev);
>> if (IS_ERR(idev))
>> return notifier_from_errno(PTR_ERR(idev));
>> + ip6_register_ndisc_ops(dev);
>> }
>> break;
>> diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
>> index 176c7c4..297080a 100644
>> --- a/net/ipv6/ndisc.c
>> +++ b/net/ipv6/ndisc.c
>> @@ -185,24 +185,25 @@ static struct nd_opt_hdr *ndisc_next_option(struct nd_opt_hdr *cur,
>> return cur <= end && cur->nd_opt_type == type ? cur : NULL;
>> }
>> -static inline int ndisc_is_useropt(struct nd_opt_hdr *opt)
>> +static inline int ip6_ndisc_is_useropt(struct nd_opt_hdr *opt)
>> {
>> - return opt->nd_opt_type == ND_OPT_RDNSS ||
>> - opt->nd_opt_type == ND_OPT_DNSSL;
>> + return __ip6_ndisc_is_useropt(opt);
>
> Why putting this check into a different function? It looks like a not needed redirection.
because these IPv6 userspace options are valid in 6LoWPAN-ND and IPv6-ND.
In 6LoWPAN we do:
static inline int lowpan_ndisc_is_useropt(struct nd_opt_hdr *opt)
{
return __ip6_ndisc_is_useropt(opt) || opt->nd_opt_type == ND_OPT_6CO;
}
I don't believe that _any_ IPv6 implementation in userspace can do something
with 6CO option fields. For this reason ARPHRD_6LOWPAN should make them
valid for userspace only. (btw: the radvd maintainer wants to add some
patches for 6CO processing support).
- Alex
^ permalink raw reply
* OpenWRT wrong adjustment of fq_codel defaults (Was: fq_codel_drop vs a udp flood)
From: Jesper Dangaard Brouer @ 2016-05-06 9:42 UTC (permalink / raw)
To: Eric Dumazet, Felix Fietkau, Dave Taht
Cc: make-wifi-fast, zajec5, ath10k, netdev@vger.kernel.org,
codel@lists.bufferbloat.net, Jonathan Morton, Roman Yeryomin
In-Reply-To: <1462476207.13075.20.camel@edumazet-glaptop3.roam.corp.google.com>
Hi Felix,
This is an important fix for OpenWRT, please read!
OpenWRT changed the default fq_codel sch->limit from 10240 to 1024,
without also adjusting q->flows_cnt. Eric explains below that you must
also adjust the buckets (q->flows_cnt) for this not to break. (Just
adjust it to 128)
Problematic OpenWRT commit in question:
http://git.openwrt.org/?p=openwrt.git;a=patch;h=12cd6578084e
12cd6578084e ("kernel: revert fq_codel quantum override to prevent it from causing too much cpu load with higher speed (#21326)")
I also highly recommend you cherry-pick this very recent commit:
net-next: 9d18562a2278 ("fq_codel: add batch ability to fq_codel_drop()")
https://git.kernel.org/davem/net-next/c/9d18562a227
This should fix very high CPU usage in-case fq_codel goes into drop mode.
The problem is that drop mode was considered rare, and implementation
wise it was chosen to be more expensive (to save cycles on normal mode).
Unfortunately is it easy to trigger with an UDP flood. Drop mode is
especially expensive for smaller devices, as it scans a 4K big array,
thus 64 cache misses for small devices!
The fix is to allow drop-mode to bulk-drop more packets when entering
drop-mode (default 64 bulk drop). That way we don't suddenly
experience a significantly higher processing cost per packet, but
instead can amortize this.
To Eric, should we recommend OpenWRT to adjust default (max) 64 bulk
drop, given we also recommend bucket size to be 128 ? (thus the amount
of memory to scan is less, but their CPU is also much smaller).
--Jesper
On Thu, 05 May 2016 12:23:27 -0700 Eric Dumazet <eric.dumazet@gmail.com> wrote:
> On Thu, 2016-05-05 at 19:25 +0300, Roman Yeryomin wrote:
> > On 5 May 2016 at 19:12, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> > > On Thu, 2016-05-05 at 17:53 +0300, Roman Yeryomin wrote:
> > >
> > >>
> > >> qdisc fq_codel 0: dev eth0 root refcnt 2 limit 1024p flows 1024
> > >> quantum 1514 target 5.0ms interval 100.0ms ecn
> > >> Sent 12306 bytes 128 pkt (dropped 0, overlimits 0 requeues 0)
> > >> backlog 0b 0p requeues 0
> > >> maxpacket 0 drop_overlimit 0 new_flow_count 0 ecn_mark 0
> > >> new_flows_len 0 old_flows_len 0
> > >
> > >
> > > Limit of 1024 packets and 1024 flows is not wise I think.
> > >
> > > (If all buckets are in use, each bucket has a virtual queue of 1 packet,
> > > which is almost the same than having no queue at all)
> > >
> > > I suggest to have at least 8 packets per bucket, to let Codel have a
> > > chance to trigger.
> > >
> > > So you could either reduce number of buckets to 128 (if memory is
> > > tight), or increase limit to 8192.
> >
> > Will try, but what I've posted is default, I didn't change/configure that.
>
> fq_codel has a default of 10240 packets and 1024 buckets.
>
> http://lxr.free-electrons.com/source/net/sched/sch_fq_codel.c#L413
>
> If someone changed that in the linux variant you use, he probably should
> explain the rationale.
--
Best regards,
Jesper Dangaard Brouer
MSc.CS, Principal Kernel Engineer at Red Hat
Author of http://www.iptv-analyzer.org
LinkedIn: http://www.linkedin.com/in/brouer
_______________________________________________
Codel mailing list
Codel@lists.bufferbloat.net
https://lists.bufferbloat.net/listinfo/codel
^ permalink raw reply
* Re: [PATCHv2 bluetooth-next 07/10] ipv6: introduce neighbour discovery ops
From: Alexander Aring @ 2016-05-06 9:36 UTC (permalink / raw)
To: Stefan Schmidt, Hannes Frederic Sowa,
linux-wpan-u79uwXL29TY76Z2rM5mHXA
Cc: kernel-bIcnvbaLZ9MEGnE8C9+IrQ, marcel-kz+m5ild9QBg9hUCZPvPmw,
jukka.rissanen-VuQAYsv1563Yd54FQh9/CA, mcr-SWp7JaYWvAQV+D8aMU/kSg,
werner-SEdMjqphH88wryQfseakQg,
linux-bluetooth-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA, David S . Miller, Alexey Kuznetsov,
James Morris, Hideaki YOSHIFUJI, Patrick McHardy
In-Reply-To: <5728EB3F.2060808-JPH+aEBZ4P+UEJcrhfAQsw@public.gmane.org>
Hi,
On 05/03/2016 08:17 PM, Stefan Schmidt wrote:
> Hello.
>
> On 02/05/16 21:36, Hannes Frederic Sowa wrote:
>> On 20.04.2016 10:19, Alexander Aring wrote:
>>> This patch introduces neighbour discovery ops callback structure. The
>>> structure contains at first receive and transmit handling for NS/NA and
>>> userspace option field functionality.
>>>
>>> These callback offers 6lowpan different handling, such as 802.15.4 short
>>> address handling or RFC6775 (Neighbor Discovery Optimization for IPv6 over
>>> 6LoWPANs).
>>>
>>> Cc: David S. Miller<davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org>
>>> Cc: Alexey Kuznetsov<kuznet-v/Mj1YrvjDBInbfyfbPRSQ@public.gmane.org>
>>> Cc: James Morris<jmorris-gx6/JNMH7DfYtjvyW6yDsg@public.gmane.org>
>>> Cc: Hideaki YOSHIFUJI<yoshfuji-VfPWfsRibaP+Ru+s062T9g@public.gmane.org>
>>> Cc: Patrick McHardy<kaber-dcUjhNyLwpNeoWH0uzbU5w@public.gmane.org>
>>> Signed-off-by: Alexander Aring<aar-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
>>> ---
>>> include/linux/netdevice.h | 3 ++
>>> include/net/ndisc.h | 96 +++++++++++++++++++++++++++++++++++++++++++----
>>> net/ipv6/addrconf.c | 1 +
>>> net/ipv6/ndisc.c | 71 ++++++++++++++++++++++++-----------
>>> net/ipv6/route.c | 2 +-
>>> 5 files changed, 144 insertions(+), 29 deletions(-)
>>>
>>> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
>>> index 0052c42..bc60033 100644
>>> --- a/include/linux/netdevice.h
>>> +++ b/include/linux/netdevice.h
>>> @@ -1677,6 +1677,9 @@ struct net_device {
>>> #ifdef CONFIG_NET_L3_MASTER_DEV
>>> const struct l3mdev_ops *l3mdev_ops;
>>> #endif
>>> +#if IS_ENABLED(CONFIG_IPV6)
>>> + const struct ndisc_ops *ndisc_ops;
>>> +#endif
>>> const struct header_ops *header_ops;
>>> diff --git a/include/net/ndisc.h b/include/net/ndisc.h
>>> index aac868e..14ed016 100644
>>> --- a/include/net/ndisc.h
>>> +++ b/include/net/ndisc.h
>>> @@ -110,7 +110,8 @@ struct ndisc_options {
>>> #define NDISC_OPT_SPACE(len) (((len)+2+7)&~7)
>>> -struct ndisc_options *ndisc_parse_options(u8 *opt, int opt_len,
>>> +struct ndisc_options *ndisc_parse_options(const struct net_device *dev,
>>> + u8 *opt, int opt_len,
>>> struct ndisc_options *ndopts);
>>> /*
>>> @@ -173,6 +174,93 @@ static inline struct neighbour *__ipv6_neigh_lookup(struct net_device *dev, cons
>>> return n;
>>> }
>>> +static inline int __ip6_ndisc_is_useropt(struct nd_opt_hdr *opt)
>>> +{
>>> + return opt->nd_opt_type == ND_OPT_RDNSS ||
>>> + opt->nd_opt_type == ND_OPT_DNSSL;
>>> +}
>>> +
>>> +#if IS_ENABLED(CONFIG_IPV6)
>>> +struct ndisc_ops {
>>> + int (*is_useropt)(struct nd_opt_hdr *opt);
>>> + void (*send_na)(struct net_device *dev,
>>> + const struct in6_addr *daddr,
>>> + const struct in6_addr *solicited_addr,
>>> + bool router, bool solicited,
>>> + bool override, bool inc_opt);
>>> + void (*recv_na)(struct sk_buff *skb);
>>> + void (*send_ns)(struct net_device *dev,
>>> + const struct in6_addr *solicit,
>>> + const struct in6_addr *daddr,
>>> + const struct in6_addr *saddr);
>>> + void (*recv_ns)(struct sk_buff *skb);
>>> +};
>>> +
>>> +static inline int ndisc_is_useropt(const struct net_device *dev,
>>> + struct nd_opt_hdr *opt)
>>> +{
>>> + if (likely(dev->ndisc_ops->is_useropt))
>>> + return dev->ndisc_ops->is_useropt(opt);
>>> + else
>>> + return 0;
>>> +}
>>> +
>>> +static inline void ndisc_send_na(struct net_device *dev,
>>> + const struct in6_addr *daddr,
>>> + const struct in6_addr *solicited_addr,
>>> + bool router, bool solicited, bool override,
>>> + bool inc_opt)
>>> +{
>>> + if (likely(dev->ndisc_ops->send_na))
>>> + dev->ndisc_ops->send_na(dev, daddr, solicited_addr, router,
>>> + solicited, override, inc_opt);
>>> +}
>>> +
>>> +static inline void ndisc_recv_na(struct sk_buff *skb)
>>> +{
>>> + if (likely(skb->dev->ndisc_ops->recv_na))
>>> + skb->dev->ndisc_ops->recv_na(skb);
>>> +}
>>> +
>>> +static inline void ndisc_send_ns(struct net_device *dev,
>>> + const struct in6_addr *solicit,
>>> + const struct in6_addr *daddr,
>>> + const struct in6_addr *saddr)
>>> +{
>>> + if (likely(dev->ndisc_ops->send_ns))
>>> + dev->ndisc_ops->send_ns(dev, solicit, daddr, saddr);
>>> +}
>>> +
>>> +static inline void ndisc_recv_ns(struct sk_buff *skb)
>>> +{
>>> + if (likely(skb->dev->ndisc_ops->recv_ns))
>>> + skb->dev->ndisc_ops->recv_ns(skb);
>>> +}
>>> +#else
>>> +static inline int ndisc_is_useropt(const struct net_device *dev,
>>> + struct nd_opt_hdr *opt)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static inline void ndisc_send_na(struct net_device *dev,
>>> + const struct in6_addr *daddr,
>>> + const struct in6_addr *solicited_addr,
>>> + bool router, bool solicited, bool override,
>>> + bool inc_opt) { }
>>> +
>>> +static inline void ndisc_recv_na(struct sk_buff *skb) { }
>>> +
>>> +static inline void ndisc_send_ns(struct net_device *dev,
>>> + const struct in6_addr *solicit,
>>> + const struct in6_addr *daddr,
>>> + const struct in6_addr *saddr) { }
>>> +
>>> +static inline void ndisc_recv_ns(struct sk_buff *skb) { }
>>> +#endif
>> Do those empty functions actually make sense? I wonder a bit because
>> 6lowpan strictly depends on ipv6 and they should never be called without
>> IPv6, no?
>
> Agreed. 6LoWAPN is only an adaptation layer so we know that IPv6 must be enabled here. I would also argue for removing this ifdef and the empty functions.
>
I removed the empty function, removing ifdef is not possible. There are
still somewhere net/ipv6.h includes at do_mounts/networking filesystems.
I need to leave the ifdef there otherwise compiling will fail that
"ndisc->ops" are not available.
I tested it with several randconfigs.
- Alex
^ permalink raw reply
* Re: [PATCH v9 net-next 4/7] openvswitch: add layer 3 flow/port support
From: Jiri Benc @ 2016-05-06 9:35 UTC (permalink / raw)
To: Simon Horman; +Cc: netdev, dev, Lorand Jakab, Thomas Morin
In-Reply-To: <1462347393-22354-5-git-send-email-simon.horman@netronome.com>
On Wed, 4 May 2016 16:36:30 +0900, Simon Horman wrote:
> +static int push_eth(struct sk_buff *skb, struct sw_flow_key *key,
> + const struct ovs_action_push_eth *ethh)
> +{
> + int err;
> +
> + /* De-accelerate any hardware accelerated VLAN tag added to a previous
> + * Ethernet header */
> + err = skb_vlan_deaccel(skb);
> + if (unlikely(err))
> + return err;
> +
> + /* Add the new Ethernet header */
> + if (skb_cow_head(skb, ETH_HLEN) < 0)
> + return -ENOMEM;
> +
> + skb_push(skb, ETH_HLEN);
> + skb_reset_mac_header(skb);
> + skb_reset_mac_len(skb);
> +
> + ether_addr_copy(eth_hdr(skb)->h_source, ethh->addresses.eth_src);
> + ether_addr_copy(eth_hdr(skb)->h_dest, ethh->addresses.eth_dst);
> + eth_hdr(skb)->h_proto = ethh->eth_type;
This doesn't seem right. We know the packet type, it's skb->protocol.
We should fill in that.
In addition, we should check whether mac_len > 0 and in such case,
change skb->protocol to ETH_P_TEB first (and store that value in the
pushed eth header).
Similarly on pop_eth, we need to check skb->protocol and if it is
ETH_P_TEB, call eth_type_trans on the modified frame to set the new
skb->protocol correctly. It's probably not that simple, as we'd need a
version of eth_type_trans that doesn't need a net device.
Jiri
^ permalink raw reply
* Re: [PATCHv2 bluetooth-next 02/10] 6lowpan: add 802.15.4 short addr slaac
From: Alexander Aring @ 2016-05-06 9:26 UTC (permalink / raw)
To: Stefan Schmidt, linux-wpan
Cc: kernel, marcel, jukka.rissanen, hannes, mcr, werner,
linux-bluetooth, netdev, David S . Miller, Alexey Kuznetsov,
James Morris, Hideaki YOSHIFUJI, Patrick McHardy
In-Reply-To: <5728EB0A.5080108@osg.samsung.com>
Hi,
On 05/03/2016 08:16 PM, Stefan Schmidt wrote:
> Hello.
>
> On 20/04/16 10:19, Alexander Aring wrote:
>> This patch adds the autoconfiguration if a valid 802.15.4 short address
>> is available for 802.15.4 6LoWPAN interfaces.
>>
>> Cc: David S. Miller<davem@davemloft.net>
>> Cc: Alexey Kuznetsov<kuznet@ms2.inr.ac.ru>
>> Cc: James Morris<jmorris@namei.org>
>> Cc: Hideaki YOSHIFUJI<yoshfuji@linux-ipv6.org>
>> Cc: Patrick McHardy<kaber@trash.net>
>> Signed-off-by: Alexander Aring<aar@pengutronix.de>
>> ---
>> include/net/addrconf.h | 3 +++
>> net/6lowpan/core.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++
>> net/ipv6/addrconf.c | 5 +++--
>> 3 files changed, 52 insertions(+), 2 deletions(-)
>>
>> diff --git a/include/net/addrconf.h b/include/net/addrconf.h
>> index 730d856..b1774eb 100644
>> --- a/include/net/addrconf.h
>> +++ b/include/net/addrconf.h
>> @@ -94,6 +94,9 @@ int ipv6_rcv_saddr_equal(const struct sock *sk, const struct sock *sk2,
>> void addrconf_join_solict(struct net_device *dev, const struct in6_addr *addr);
>> void addrconf_leave_solict(struct inet6_dev *idev, const struct in6_addr *addr);
>> +void addrconf_add_linklocal(struct inet6_dev *idev,
>> + const struct in6_addr *addr, u32 flags);
>> +
>> static inline int addrconf_ifid_eui48(u8 *eui, struct net_device *dev)
>> {
>> if (dev->addr_len != ETH_ALEN)
>> diff --git a/net/6lowpan/core.c b/net/6lowpan/core.c
>> index 7a240b3..fbae31e 100644
>> --- a/net/6lowpan/core.c
>> +++ b/net/6lowpan/core.c
>> @@ -14,6 +14,7 @@
>> #include <linux/module.h>
>> #include <net/6lowpan.h>
>> +#include <net/addrconf.h>
>> #include "6lowpan_i.h"
>> @@ -72,16 +73,61 @@ void lowpan_unregister_netdev(struct net_device *dev)
>> }
>> EXPORT_SYMBOL(lowpan_unregister_netdev);
>> +static int addrconf_ifid_802154_6lowpan(u8 *eui, struct net_device *dev)
>> +{
>> + struct wpan_dev *wpan_dev = lowpan_802154_dev(dev)->wdev->ieee802154_ptr;
>> +
>> + /* Set short_addr autoconfiguration if short_addr is present only */
>> + if (!ieee802154_is_valid_src_short_addr(wpan_dev->short_addr))
>> + return -1;
>
> -EINVAL instead of -1?
>
The ipv6 implementation do that for the "dev->addr" slaac the same here.
I think the reason is because here, if this fails for any reason then
this is simple not an error, we just don't add a link-local for short
addr slaac here and don't abort completely the CHANGE/UP of interface.
IPv6 also use "-1" return value only if parsing fails, I would to see
that also for iphc stuff where we mixed a lot of "-EIO" and "-EINVAL",
anyway nobody will really care about that return value.
It just means: 0 parsing successful and -1 parsing failed.
- Alex
^ permalink raw reply
* Re: [PATCH v9 net-next 4/7] openvswitch: add layer 3 flow/port support
From: Jiri Benc @ 2016-05-06 9:25 UTC (permalink / raw)
To: Simon Horman
Cc: pravin shelar, Linux Kernel Network Developers, ovs dev,
Lorand Jakab, Thomas Morin
In-Reply-To: <20160506055705.GA9276@penelope.isobedori.kobe.vergenet.net>
On Fri, 6 May 2016 14:57:07 +0900, Simon Horman wrote:
> On Thu, May 05, 2016 at 10:37:08AM -0700, pravin shelar wrote:
> > On transmit side you are using mac_len to detect l3 packet, why not do
> > same while extracting the key?
I agree. The skb should be self-contained, i.e. it should be obvious
whether it has the MAC header set or not just from the skb itself at
any point in the packet processing. Otherwise, I'd expect things like
recirculation to break after push/pop of eth header.
> Unfortunately mac_len can't be relied on here, emprically it has the same
> value (28 in my tests) for both the TEB and layer3 case above.
That's strange, it looks like there's something setting the mac header
unconditionally in ovs code. We should find that place and change it.
The ARPHRD_NONE interfaces don't even set mac_header and mac_len, this
will need to be set by ovs upon getting frame from such interface.
> Perhaps that could be changed by futher enhancements in the tunneling code
> but I think things are symetric as they stand:
>
> * On recieve skb->protocol can be read to distinguish TEB and layer3 packets
> * On transmit skb->protocol should be set to distinguish TEB and layer3 packets
Yes, but you need to act upon this directly after receiving the
frame/just before sending the frame and set up an internal flag that
will be used throughout the code. That way, the packet can be handed
over to different parts of the code, recirculated, etc. without
worries. skb->mac_len is probably a good candidate for such flag.
Jiri
^ permalink raw reply
* [patch net 3/3] mlxsw: spectrum: Fix ordering in mlxsw_sp_fini
From: Jiri Pirko @ 2016-05-06 9:17 UTC (permalink / raw)
To: netdev; +Cc: davem, idosch, eladr, yotamg, ogerlitz
In-Reply-To: <1462526244-11898-1-git-send-email-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
Fixes: 0f433fa0ec ("mlxsw: spectrum_buffers: Implement shared buffer configuration")
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 668b2f4..749cc27 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -2159,6 +2159,7 @@ static void mlxsw_sp_fini(void *priv)
struct mlxsw_sp *mlxsw_sp = priv;
mlxsw_sp_switchdev_fini(mlxsw_sp);
+ mlxsw_sp_buffers_fini(mlxsw_sp);
mlxsw_sp_traps_fini(mlxsw_sp);
mlxsw_sp_event_unregister(mlxsw_sp, MLXSW_TRAP_ID_PUDE);
mlxsw_sp_ports_remove(mlxsw_sp);
--
2.5.5
^ permalink raw reply related
* [patch net 2/3] mlxsw: spectrum: Add missing rollback in flood configuration
From: Jiri Pirko @ 2016-05-06 9:17 UTC (permalink / raw)
To: netdev; +Cc: davem, idosch, eladr, yotamg, ogerlitz
In-Reply-To: <1462526244-11898-1-git-send-email-jiri@resnulli.us>
From: Ido Schimmel <idosch@mellanox.com>
When we fail to set the flooding configuration for the broadcast and
unregistered multicast traffic, we should revert the flooding
configuration of the unknown unicast traffic.
Fixes: 0293038e0c36 ("mlxsw: spectrum: Add support for flood control")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index e1c74ef..9cd6f47 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -214,7 +214,15 @@ static int __mlxsw_sp_port_flood_set(struct mlxsw_sp_port *mlxsw_sp_port,
mlxsw_reg_sftr_pack(sftr_pl, MLXSW_SP_FLOOD_TABLE_BM, idx_begin,
table_type, range, local_port, set);
err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sftr), sftr_pl);
+ if (err)
+ goto err_flood_bm_set;
+ else
+ goto buffer_out;
+err_flood_bm_set:
+ mlxsw_reg_sftr_pack(sftr_pl, MLXSW_SP_FLOOD_TABLE_UC, idx_begin,
+ table_type, range, local_port, !set);
+ mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sftr), sftr_pl);
buffer_out:
kfree(sftr_pl);
return err;
--
2.5.5
^ permalink raw reply related
* [patch net 1/3] mlxsw: spectrum: Fix rollback order in LAG join failure
From: Jiri Pirko @ 2016-05-06 9:17 UTC (permalink / raw)
To: netdev; +Cc: davem, idosch, eladr, yotamg, ogerlitz
In-Reply-To: <1462526244-11898-1-git-send-email-jiri@resnulli.us>
From: Ido Schimmel <idosch@mellanox.com>
Make the leave procedure in the error path symmetric to the join
procedure and first remove the port from the collector before
potentially destroying the LAG.
Fixes: 0d65fc13042f ("mlxsw: spectrum: Implement LAG port join/leave")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 4afbc3e..668b2f4 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -2541,11 +2541,11 @@ static int mlxsw_sp_port_lag_join(struct mlxsw_sp_port *mlxsw_sp_port,
lag->ref_count++;
return 0;
+err_col_port_enable:
+ mlxsw_sp_lag_col_port_remove(mlxsw_sp_port, lag_id);
err_col_port_add:
if (!lag->ref_count)
mlxsw_sp_lag_destroy(mlxsw_sp, lag_id);
-err_col_port_enable:
- mlxsw_sp_lag_col_port_remove(mlxsw_sp_port, lag_id);
return err;
}
--
2.5.5
^ permalink raw reply related
* [patch net 0/3] mlxsw: Couple of fixes
From: Jiri Pirko @ 2016-05-06 9:17 UTC (permalink / raw)
To: netdev; +Cc: davem, idosch, eladr, yotamg, ogerlitz
From: Jiri Pirko <jiri@mellanox.com>
Ido Schimmel (2):
mlxsw: spectrum: Fix rollback order in LAG join failure
mlxsw: spectrum: Add missing rollback in flood configuration
Jiri Pirko (1):
mlxsw: spectrum: Fix ordering in mlxsw_sp_fini
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 5 +++--
drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c | 8 ++++++++
2 files changed, 11 insertions(+), 2 deletions(-)
--
2.5.5
^ permalink raw reply
* Re: [PATCH v9 net-next 7/7] openvswitch: use ipgre tunnel rather than gretap tunnel
From: Jiri Benc @ 2016-05-06 9:15 UTC (permalink / raw)
To: Simon Horman; +Cc: pravin shelar, Linux Kernel Network Developers, ovs dev
In-Reply-To: <20160506065400.GA15277@penelope.isobedori.kobe.vergenet.net>
On Fri, 6 May 2016 15:54:02 +0900, Simon Horman wrote:
> -int ovs_netdev_send_raw_tun(struct sk_buff *skb)
> -{
> - if (skb->mac_len)
> - skb->protocol = ntohs(ETH_P_TEB);
> + if (dev->type != ARPHRD_ETHER && skb->mac_len) {
> + skb->protocol = htons(ETH_P_TEB);
> + } else if (dev->type == ARPHRD_ETHER && !skb->mac_len) {
> + kfree_skb(skb);
> + return -EINVAL;
> + }
This was something I was missing in your patches (sorry, did not get to
the full review yet).
You'll also need to enable ARPHRD_NONE and ARPHRD_IPGRE interfaces in
ovs_netdev_link.
Jiri
^ permalink raw reply
* Re: [PATCH] cfg80211/nl80211: add wifi tx power mode switching support
From: Wei-Ning Huang @ 2016-05-06 8:19 UTC (permalink / raw)
To: Dan Williams
Cc: Linux-Wireless, LKML, Johannes Berg, Sameer Nanda, Todd Broch,
davem, netdev
In-Reply-To: <1462464478.23962.12.camel@redhat.com>
On Fri, May 6, 2016 at 12:07 AM, Dan Williams <dcbw@redhat.com> wrote:
>
> On Thu, 2016-05-05 at 14:44 +0800, Wei-Ning Huang wrote:
> > Recent new hardware has the ability to switch between tablet mode and
> > clamshell mode. To optimize WiFi performance, we want to be able to
> > use
> > different power table between modes. This patch adds a new netlink
> > message type and cfg80211_ops function to allow userspace to trigger
> > a
> > power mode switch for a given wireless interface.
> >
> > Signed-off-by: Wei-Ning Huang <wnhuang@chromium.org>
> > ---
> > include/net/cfg80211.h | 11 +++++++++++
> > include/uapi/linux/nl80211.h | 21 +++++++++++++++++++++
> > net/wireless/nl80211.c | 16 ++++++++++++++++
> > net/wireless/rdev-ops.h | 22 ++++++++++++++++++++++
> > net/wireless/trace.h | 20 ++++++++++++++++++++
> > 5 files changed, 90 insertions(+)
> >
> > diff --git a/include/net/cfg80211.h b/include/net/cfg80211.h
> > index 9e1b24c..aa77fa0 100644
> > --- a/include/net/cfg80211.h
> > +++ b/include/net/cfg80211.h
> > @@ -2370,6 +2370,12 @@ struct cfg80211_qos_map {
> > * @get_tx_power: store the current TX power into the dbm variable;
> > * return 0 if successful
> > *
> > + * @set_tx_power_mode: set the transmit power mode. Some device have
> > the ability
> > + * to transform between different mode such as clamshell and
> > tablet mode.
> > + * set_tx_power_mode allows setting of different TX power
> > mode at runtime.
> > + * @get_tx_power_mode: store the current TX power mode into the mode
> > variable;
> > + * return 0 if successful
> > + *
> > * @set_wds_peer: set the WDS peer for a WDS interface
> > *
> > * @rfkill_poll: polls the hw rfkill line, use cfg80211 reporting
> > @@ -2631,6 +2637,11 @@ struct cfg80211_ops {
> > int (*get_tx_power)(struct wiphy *wiphy, struct
> > wireless_dev *wdev,
> > int *dbm);
> >
> > + int (*set_tx_power_mode)(struct wiphy *wiphy,
> > + enum nl80211_tx_power_mode
> > mode);
> > + int (*get_tx_power_mode)(struct wiphy *wiphy,
> > + enum nl80211_tx_power_mode
> > *mode);
> > +
> > int (*set_wds_peer)(struct wiphy *wiphy, struct
> > net_device *dev,
> > const u8 *addr);
> >
> > diff --git a/include/uapi/linux/nl80211.h
> > b/include/uapi/linux/nl80211.h
> > index 5a30a75..9b1888a 100644
> > --- a/include/uapi/linux/nl80211.h
> > +++ b/include/uapi/linux/nl80211.h
> > @@ -1796,6 +1796,9 @@ enum nl80211_commands {
> > * connecting to a PCP, and in %NL80211_CMD_START_AP to start
> > * a PCP instead of AP. Relevant for DMG networks only.
> > *
> > + * @NL80211_ATTR_WIPHY_TX_POWER_MODE: Transmit power mode. See
> > + * &enum nl80211_tx_power_mode for possible values.
> > + *
> > * @NUM_NL80211_ATTR: total number of nl80211_attrs available
> > * @NL80211_ATTR_MAX: highest attribute number currently defined
> > * @__NL80211_ATTR_AFTER_LAST: internal use
> > @@ -2172,6 +2175,8 @@ enum nl80211_attrs {
> >
> > NL80211_ATTR_PBSS,
> >
> > + NL80211_ATTR_WIPHY_TX_POWER_MODE,
> > +
> > /* add attributes here, update the policy in nl80211.c */
> >
> > __NL80211_ATTR_AFTER_LAST,
> > @@ -3703,6 +3708,22 @@ enum nl80211_tx_power_setting {
> > };
> >
> > /**
> > + * enum nl80211_tx_power_mode - TX power mode setting
> > + * @NL80211_TX_POWER_LOW: general low TX power mode
> > + * @NL80211_TX_POWER_MEDIUM: general medium TX power mode
> > + * @NL80211_TX_POWER_HIGH: general high TX power mode
> > + * @NL80211_TX_POWER_CLAMSHELL: clamshell mode TX power mode
> > + * @NL80211_TX_POWER_TABLET: tablet mode TX power mode
> > + */
> > +enum nl80211_tx_power_mode {
> > + NL80211_TX_POWER_LOW,
> > + NL80211_TX_POWER_MEDIUM,
> > + NL80211_TX_POWER_HIGH,
> > + NL80211_TX_POWER_CLAMSHELL,
> > + NL80211_TX_POWER_TABLET,
>
> "clamshell" and "tablet" probably mean many different things to many
> different people with respect to whether or not they should do anything
> with power saving or wifi. I feel like a more generic interface is
> needed here.
We could probably drop those two CLAMSHELL and TABLET constant or
describing what they mean
in more detail?
>
> Could this be already done by:
> @NL80211_ATTR_WIPHY_TX_POWER_SETTING = NL80211_TX_POWER_LIMITED
> @NL80211_ATTR_WIPHY_TX_POWER_LEVEL = <limit for clamshell/tablet mode>
>
> and then the device would be able to change its TX power as it saw fit
> up to that limit set by your application which figures out whether it's
> in clamshell or tablet mode?
We usually want different power settings in different band/channel.
For example, we can have three different power settings
in 2.4Ghz, channels 36-64 & channels 100+ on 5Ghz. In this case, we
can not simply set a fixed number to the power level.
A power table is required to map channel/band to actual power limit.
For most of the driver, changing power table requires loading
a new set of calibration data from either devicetree or ACPI. Due to
this, a new message type is required to allow the driver to
switch between tables.
Wei-Ning
>
>
> Dan
>
> > +};
> > +
> > +/**
> > * enum nl80211_packet_pattern_attr - packet pattern attribute
> > * @__NL80211_PKTPAT_INVALID: invalid number for nested attribute
> > * @NL80211_PKTPAT_PATTERN: the pattern, values where the mask has
> > diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c
> > index 056a730..61b474d 100644
> > --- a/net/wireless/nl80211.c
> > +++ b/net/wireless/nl80211.c
> > @@ -402,6 +402,7 @@ static const struct nla_policy
> > nl80211_policy[NUM_NL80211_ATTR] = {
> > [NL80211_ATTR_SCHED_SCAN_DELAY] = { .type = NLA_U32 },
> > [NL80211_ATTR_REG_INDOOR] = { .type = NLA_FLAG },
> > [NL80211_ATTR_PBSS] = { .type = NLA_FLAG },
> > + [NL80211_ATTR_WIPHY_TX_POWER_MODE] = { .type = NLA_U32 },
> > };
> >
> > /* policy for the key attributes */
> > @@ -2218,6 +2219,21 @@ static int nl80211_set_wiphy(struct sk_buff
> > *skb, struct genl_info *info)
> > return result;
> > }
> >
> > + if (info->attrs[NL80211_ATTR_WIPHY_TX_POWER_MODE]) {
> > + enum nl80211_tx_power_mode mode;
> > + int idx = 0;
> > +
> > + if (!rdev->ops->set_tx_power_mode)
> > + return -EOPNOTSUPP;
> > +
> > + idx = NL80211_ATTR_WIPHY_TX_POWER_MODE;
> > + mode = nla_get_u32(info->attrs[idx]);
> > +
> > + result = rdev_set_tx_power_mode(rdev, mode);
> > + if (result)
> > + return result;
> > + }
> > +
> > if (info->attrs[NL80211_ATTR_WIPHY_ANTENNA_TX] &&
> > info->attrs[NL80211_ATTR_WIPHY_ANTENNA_RX]) {
> > u32 tx_ant, rx_ant;
> > diff --git a/net/wireless/rdev-ops.h b/net/wireless/rdev-ops.h
> > index 8ae0c04..c3a1035 100644
> > --- a/net/wireless/rdev-ops.h
> > +++ b/net/wireless/rdev-ops.h
> > @@ -552,6 +552,28 @@ static inline int rdev_get_tx_power(struct
> > cfg80211_registered_device *rdev,
> > return ret;
> > }
> >
> > +static inline int
> > +rdev_set_tx_power_mode(struct cfg80211_registered_device *rdev,
> > + enum nl80211_tx_power_mode mode)
> > +{
> > + int ret;
> > + trace_rdev_set_tx_power_mode(&rdev->wiphy, mode);
> > + ret = rdev->ops->set_tx_power_mode(&rdev->wiphy, mode);
> > + trace_rdev_return_int(&rdev->wiphy, ret);
> > + return ret;
> > +}
> > +
> > +static inline int
> > +rdev_get_tx_power_mode(struct cfg80211_registered_device *rdev,
> > + enum nl80211_tx_power_mode *mode)
> > +{
> > + int ret;
> > + trace_rdev_get_tx_power_mode(&rdev->wiphy);
> > + ret = rdev->ops->get_tx_power_mode(&rdev->wiphy, mode);
> > + trace_rdev_return_int_int(&rdev->wiphy, ret, *mode);
> > + return ret;
> > +}
> > +
> > static inline int rdev_set_wds_peer(struct
> > cfg80211_registered_device *rdev,
> > struct net_device *dev, const u8
> > *addr)
> > {
> > diff --git a/net/wireless/trace.h b/net/wireless/trace.h
> > index 09b242b..132c8c2 100644
> > --- a/net/wireless/trace.h
> > +++ b/net/wireless/trace.h
> > @@ -1420,6 +1420,26 @@ TRACE_EVENT(rdev_set_tx_power,
> > WIPHY_PR_ARG, WDEV_PR_ARG,__entry->type, __entry-
> > >mbm)
> > );
> >
> > +DEFINE_EVENT(wiphy_only_evt, rdev_get_tx_power_mode,
> > + TP_PROTO(struct wiphy *wiphy),
> > + TP_ARGS(wiphy)
> > +);
> > +
> > +TRACE_EVENT(rdev_set_tx_power_mode,
> > + TP_PROTO(struct wiphy *wiphy, enum nl80211_tx_power_mode
> > mode),
> > + TP_ARGS(wiphy, mode),
> > + TP_STRUCT__entry(
> > + WIPHY_ENTRY
> > + __field(enum nl80211_tx_power_mode, mode)
> > + ),
> > + TP_fast_assign(
> > + WIPHY_ASSIGN;
> > + __entry->mode = mode;
> > + ),
> > + TP_printk(WIPHY_PR_FMT ", mode: %d",
> > + WIPHY_PR_ARG, __entry->mode)
> > +);
> > +
> > TRACE_EVENT(rdev_return_int_int,
> > TP_PROTO(struct wiphy *wiphy, int func_ret, int func_fill),
> > TP_ARGS(wiphy, func_ret, func_fill),
--
Wei-Ning Huang, 黃偉寧 | Software Engineer, Google Inc., Taiwan |
wnhuang@google.com | Cell: +886 910-380678
^ permalink raw reply
* Re: [PATCH net-next] vxlan: if_arp: introduce ARPHRD_VXLANGPE
From: Simon Horman @ 2016-05-06 8:11 UTC (permalink / raw)
To: Thadeu Lima de Souza Cascardo; +Cc: Jiri Benc, netdev
In-Reply-To: <20160505195643.GA14919@indiana.gru.redhat.com>
On Thu, May 05, 2016 at 04:56:43PM -0300, Thadeu Lima de Souza Cascardo wrote:
> On Thu, May 05, 2016 at 09:31:41PM +0200, Jiri Benc wrote:
> > On Thu, 5 May 2016 13:36:44 -0300, Thadeu Lima de Souza Cascardo wrote:
> > > Use ARPHRD_VXLANGPE to identify VxLAN GPE interfaces. This is going to be used
> > > to allow GPE interfaces to be added as openvswitch ports.
> >
> > What's wrong with ARPHRD_NONE? I don't think we need a separate type
> > for VXLAN-GPE. Just use ARPHRD_NONE in ovs and things should work, for
> > all ARPHRD_NONE interfaces as a bonus.
> >
>
> That's fine for me. I looked quickly at the few devices using ARPHRD_NONE in
> upstream kernel, not sure if there are broken out-of-tree drivers out there. And
> should we care?
It seems unlikely to me that we should.
^ permalink raw reply
* [PATCH nf-next 1/1] ipvs: make drop_entry protection effective for SIP-pe
From: Simon Horman @ 2016-05-06 7:32 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Marco Angaroni, Simon Horman
In-Reply-To: <1462519959-22382-1-git-send-email-horms@verge.net.au>
From: Marco Angaroni <marcoangaroni@gmail.com>
DoS protection policy that deletes connections to avoid out of memory is
currently not effective for SIP-pe plus OPS-mode for two reasons:
1) connection templates (holding SIP call-id) are always skipped in
ip_vs_random_dropentry()
2) in_pkts counter (used by drop_entry algorithm) is not incremented
for connection templates
This patch addresses such problems with the following changes:
a) connection templates associated (via their dest) to virtual-services
configured in OPS mode are included in ip_vs_random_dropentry()
monitoring. This applies to SIP-pe over UDP (which requires OPS mode),
but is more general principle: when OPS is controlled by templates
memory can be used only by templates themselves, since OPS conns are
deleted after packet is forwarded.
b) OPS connections, if controlled by a template, cause increment of
in_pkts counter of their template. This is already happening but only
in case director is in master-slave mode (see ip_vs_sync_conn()).
Signed-off-by: Marco Angaroni <marcoangaroni@gmail.com>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
net/netfilter/ipvs/ip_vs_conn.c | 22 +++++++++++++++++++---
net/netfilter/ipvs/ip_vs_core.c | 8 +++++++-
2 files changed, 26 insertions(+), 4 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_conn.c b/net/netfilter/ipvs/ip_vs_conn.c
index 292365ffa4f0..2cb3c626cd43 100644
--- a/net/netfilter/ipvs/ip_vs_conn.c
+++ b/net/netfilter/ipvs/ip_vs_conn.c
@@ -1261,6 +1261,16 @@ static inline int todrop_entry(struct ip_vs_conn *cp)
return 1;
}
+static inline bool ip_vs_conn_ops_mode(struct ip_vs_conn *cp)
+{
+ struct ip_vs_service *svc;
+
+ if (!cp->dest)
+ return false;
+ svc = rcu_dereference(cp->dest->svc);
+ return svc && (svc->flags & IP_VS_SVC_F_ONEPACKET);
+}
+
/* Called from keventd and must protect itself from softirqs */
void ip_vs_random_dropentry(struct netns_ipvs *ipvs)
{
@@ -1275,11 +1285,16 @@ void ip_vs_random_dropentry(struct netns_ipvs *ipvs)
unsigned int hash = prandom_u32() & ip_vs_conn_tab_mask;
hlist_for_each_entry_rcu(cp, &ip_vs_conn_tab[hash], c_list) {
- if (cp->flags & IP_VS_CONN_F_TEMPLATE)
- /* connection template */
- continue;
if (cp->ipvs != ipvs)
continue;
+ if (cp->flags & IP_VS_CONN_F_TEMPLATE) {
+ if (atomic_read(&cp->n_control) ||
+ !ip_vs_conn_ops_mode(cp))
+ continue;
+ else
+ /* connection template of OPS */
+ goto try_drop;
+ }
if (cp->protocol == IPPROTO_TCP) {
switch(cp->state) {
case IP_VS_TCP_S_SYN_RECV:
@@ -1307,6 +1322,7 @@ void ip_vs_random_dropentry(struct netns_ipvs *ipvs)
continue;
}
} else {
+try_drop:
if (!todrop_entry(cp))
continue;
}
diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
index f3bac2e9a25a..1207f20d24e4 100644
--- a/net/netfilter/ipvs/ip_vs_core.c
+++ b/net/netfilter/ipvs/ip_vs_core.c
@@ -612,7 +612,10 @@ int ip_vs_leave(struct ip_vs_service *svc, struct sk_buff *skb,
ret = cp->packet_xmit(skb, cp, pd->pp, iph);
/* do not touch skb anymore */
- atomic_inc(&cp->in_pkts);
+ if ((cp->flags & IP_VS_CONN_F_ONE_PACKET) && cp->control)
+ atomic_inc(&cp->control->in_pkts);
+ else
+ atomic_inc(&cp->in_pkts);
ip_vs_conn_put(cp);
return ret;
}
@@ -1991,6 +1994,9 @@ ip_vs_in(struct netns_ipvs *ipvs, unsigned int hooknum, struct sk_buff *skb, int
if (ipvs->sync_state & IP_VS_STATE_MASTER)
ip_vs_sync_conn(ipvs, cp, pkts);
+ else if ((cp->flags & IP_VS_CONN_F_ONE_PACKET) && cp->control)
+ /* increment is done inside ip_vs_sync_conn too */
+ atomic_inc(&cp->control->in_pkts);
ip_vs_conn_put(cp);
return ret;
--
2.7.0.rc3.207.g0ac5344
^ permalink raw reply related
* [GIT PULL nf-next 0/1] Second Round of IPVS Updates for v4.7
From: Simon Horman @ 2016-05-06 7:32 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Simon Horman
Hi Pablo,
please consider these enhancements to the IPVS. They allow its
DoS mitigation strategy effective in conjunction with the SIP persistence
engine.
The following changes since commit cb39ad8b8ef224c544074962780bf763077d6141:
netfilter: nf_tables: allow set names up to 32 bytes (2016-05-05 16:39:51 +0200)
are available in the git repository at:
https://git.kernel.org/pub/scm/linux/kernel/git/horms/ipvs-next.git tags/ipvs2-for-v4.7
for you to fetch changes up to 698e2a8dca98e4de32f3f630e6d9cd93753c52e1:
ipvs: make drop_entry protection effective for SIP-pe (2016-05-06 16:26:23 +0900)
----------------------------------------------------------------
Marco Angaroni (1):
ipvs: make drop_entry protection effective for SIP-pe
net/netfilter/ipvs/ip_vs_conn.c | 22 +++++++++++++++++++---
net/netfilter/ipvs/ip_vs_core.c | 8 +++++++-
2 files changed, 26 insertions(+), 4 deletions(-)
Marco Angaroni (1):
ipvs: make drop_entry protection effective for SIP-pe
net/netfilter/ipvs/ip_vs_conn.c | 22 +++++++++++++++++++---
net/netfilter/ipvs/ip_vs_core.c | 8 +++++++-
2 files changed, 26 insertions(+), 4 deletions(-)
--
2.7.0.rc3.207.g0ac5344
^ permalink raw reply
* [net-next 10/11] i40e: fix misleading indentation
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem; +Cc: Arnd Bergmann, netdev, nhorman, sassmann, jogreene, Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Arnd Bergmann <arnd@arndb.de>
Newly added code in i40e_vc_config_promiscuous_mode_msg() is indented
in a way that gcc rightly complains about:
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c: In function 'i40e_vc_config_promiscuous_mode_msg':
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:1543:4: error: this 'if' clause does not guard... [-Werror=misleading-indentation]
if (f->vlan >= 0 && f->vlan <= I40E_MAX_VLANID)
^~
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:1550:5: note: ...this statement, but the latter is misleadingly indented as if it is guarded by the 'if'
aq_err = pf->hw.aq.asq_last_status;
>From the context, it looks like the aq_err assignment was meant to be
inside of the conditional expression, so I'm adding the appropriate
curly braces now.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 5676a8b9cd9a ("i40e: Add VF promiscuous mode driver support")
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 36aa33a..a9b04e7 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -1544,7 +1544,7 @@ static int i40e_vc_config_promiscuous_mode_msg(struct i40e_vf *vf,
} else if (i40e_getnum_vf_vsi_vlan_filters(vsi)) {
list_for_each_entry(f, &vsi->mac_filter_list, list) {
aq_ret = 0;
- if (f->vlan >= 0 && f->vlan <= I40E_MAX_VLANID)
+ if (f->vlan >= 0 && f->vlan <= I40E_MAX_VLANID) {
aq_ret =
i40e_aq_set_vsi_uc_promisc_on_vlan(hw,
vsi->seid,
@@ -1552,6 +1552,7 @@ static int i40e_vc_config_promiscuous_mode_msg(struct i40e_vf *vf,
f->vlan,
NULL);
aq_err = pf->hw.aq.asq_last_status;
+ }
if (aq_ret)
dev_err(&pf->pdev->dev,
"Could not add VLAN %d to Unicast promiscuous domain err %s aq_err %s\n",
--
2.5.5
^ permalink raw reply related
* [net-next 11/11] i40e: constify i40e_client_ops structure
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem
Cc: Julia Lawall, netdev, nhorman, sassmann, jogreene, Julia Lawall,
Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Julia Lawall <julia.lawall@lip6.fr>
The i40e_client_ops structure is never modified, so declare it as const.
Done with the help of Coccinelle.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/infiniband/hw/i40iw/i40iw_main.c | 2 +-
drivers/net/ethernet/intel/i40e/i40e_client.h | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/hw/i40iw/i40iw_main.c b/drivers/infiniband/hw/i40iw/i40iw_main.c
index 90e5af2..e41fae24 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_main.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_main.c
@@ -1863,7 +1863,7 @@ static enum i40iw_status_code i40iw_virtchnl_send(struct i40iw_sc_dev *dev,
}
/* client interface functions */
-static struct i40e_client_ops i40e_ops = {
+static const struct i40e_client_ops i40e_ops = {
.open = i40iw_open,
.close = i40iw_close,
.l2_param_change = i40iw_l2param_change,
diff --git a/drivers/net/ethernet/intel/i40e/i40e_client.h b/drivers/net/ethernet/intel/i40e/i40e_client.h
index bf6b453..a4601d9 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_client.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_client.h
@@ -217,7 +217,7 @@ struct i40e_client {
#define I40E_CLIENT_FLAGS_LAUNCH_ON_PROBE BIT(0)
#define I40E_TX_FLAGS_NOTIFY_OTHER_EVENTS BIT(2)
enum i40e_client_type type;
- struct i40e_client_ops *ops; /* client ops provided by the client */
+ const struct i40e_client_ops *ops; /* client ops provided by the client */
};
static inline bool i40e_client_is_registered(struct i40e_client *client)
--
2.5.5
^ permalink raw reply related
* [net-next 06/11] i40evf: refactor receive routine
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem; +Cc: Jesse Brandeburg, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
This is part 2 of the Rx refactor series, just including
changes to i40evf.
This refactor aligns the receive routine with the one in
ixgbe which was highly optimized. This reduces the code
we have to maintain and allows for (hopefully) more readable
and maintainable RX hot path.
In order to do this:
- consolidate the receive path into a single function that doesn't
use packet split but *does* use pages for Rx buffers.
- remove the old _1buf routine
- consolidate several routines into helper functions
- remove VF ethtool control over packet split
- remove priv_flags interface since it is unused
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40evf/i40e_txrx.c | 898 +++++++++++----------
drivers/net/ethernet/intel/i40evf/i40e_txrx.h | 37 +-
drivers/net/ethernet/intel/i40evf/i40evf.h | 5 -
drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c | 51 --
drivers/net/ethernet/intel/i40evf/i40evf_main.c | 3 +-
5 files changed, 481 insertions(+), 513 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index 61d4a7a..fd7dae46 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -496,7 +496,6 @@ err:
void i40evf_clean_rx_ring(struct i40e_ring *rx_ring)
{
struct device *dev = rx_ring->dev;
- struct i40e_rx_buffer *rx_bi;
unsigned long bi_size;
u16 i;
@@ -506,30 +505,20 @@ void i40evf_clean_rx_ring(struct i40e_ring *rx_ring)
/* Free all the Rx ring sk_buffs */
for (i = 0; i < rx_ring->count; i++) {
- rx_bi = &rx_ring->rx_bi[i];
- if (rx_bi->dma) {
- dma_unmap_single(dev,
- rx_bi->dma,
- rx_ring->rx_buf_len,
- DMA_FROM_DEVICE);
- rx_bi->dma = 0;
- }
+ struct i40e_rx_buffer *rx_bi = &rx_ring->rx_bi[i];
+
if (rx_bi->skb) {
dev_kfree_skb(rx_bi->skb);
rx_bi->skb = NULL;
}
- if (rx_bi->page) {
- if (rx_bi->page_dma) {
- dma_unmap_page(dev,
- rx_bi->page_dma,
- PAGE_SIZE,
- DMA_FROM_DEVICE);
- rx_bi->page_dma = 0;
- }
- __free_page(rx_bi->page);
- rx_bi->page = NULL;
- rx_bi->page_offset = 0;
- }
+ if (!rx_bi->page)
+ continue;
+
+ dma_unmap_page(dev, rx_bi->dma, PAGE_SIZE, DMA_FROM_DEVICE);
+ __free_pages(rx_bi->page, 0);
+
+ rx_bi->page = NULL;
+ rx_bi->page_offset = 0;
}
bi_size = sizeof(struct i40e_rx_buffer) * rx_ring->count;
@@ -538,6 +527,7 @@ void i40evf_clean_rx_ring(struct i40e_ring *rx_ring)
/* Zero out the descriptor ring */
memset(rx_ring->desc, 0, rx_ring->size);
+ rx_ring->next_to_alloc = 0;
rx_ring->next_to_clean = 0;
rx_ring->next_to_use = 0;
}
@@ -562,37 +552,6 @@ void i40evf_free_rx_resources(struct i40e_ring *rx_ring)
}
/**
- * i40evf_alloc_rx_headers - allocate rx header buffers
- * @rx_ring: ring to alloc buffers
- *
- * Allocate rx header buffers for the entire ring. As these are static,
- * this is only called when setting up a new ring.
- **/
-void i40evf_alloc_rx_headers(struct i40e_ring *rx_ring)
-{
- struct device *dev = rx_ring->dev;
- struct i40e_rx_buffer *rx_bi;
- dma_addr_t dma;
- void *buffer;
- int buf_size;
- int i;
-
- if (rx_ring->rx_bi[0].hdr_buf)
- return;
- /* Make sure the buffers don't cross cache line boundaries. */
- buf_size = ALIGN(rx_ring->rx_hdr_len, 256);
- buffer = dma_alloc_coherent(dev, buf_size * rx_ring->count,
- &dma, GFP_KERNEL);
- if (!buffer)
- return;
- for (i = 0; i < rx_ring->count; i++) {
- rx_bi = &rx_ring->rx_bi[i];
- rx_bi->dma = dma + (i * buf_size);
- rx_bi->hdr_buf = buffer + (i * buf_size);
- }
-}
-
-/**
* i40evf_setup_rx_descriptors - Allocate Rx descriptors
* @rx_ring: Rx descriptor ring (for a specific queue) to setup
*
@@ -613,9 +572,7 @@ int i40evf_setup_rx_descriptors(struct i40e_ring *rx_ring)
u64_stats_init(&rx_ring->syncp);
/* Round up to nearest 4K */
- rx_ring->size = ring_is_16byte_desc_enabled(rx_ring)
- ? rx_ring->count * sizeof(union i40e_16byte_rx_desc)
- : rx_ring->count * sizeof(union i40e_32byte_rx_desc);
+ rx_ring->size = rx_ring->count * sizeof(union i40e_32byte_rx_desc);
rx_ring->size = ALIGN(rx_ring->size, 4096);
rx_ring->desc = dma_alloc_coherent(dev, rx_ring->size,
&rx_ring->dma, GFP_KERNEL);
@@ -626,6 +583,7 @@ int i40evf_setup_rx_descriptors(struct i40e_ring *rx_ring)
goto err;
}
+ rx_ring->next_to_alloc = 0;
rx_ring->next_to_clean = 0;
rx_ring->next_to_use = 0;
@@ -644,6 +602,10 @@ err:
static inline void i40e_release_rx_desc(struct i40e_ring *rx_ring, u32 val)
{
rx_ring->next_to_use = val;
+
+ /* update next to alloc since we have filled the ring */
+ rx_ring->next_to_alloc = val;
+
/* Force memory writes to complete before letting h/w
* know there are new descriptors to fetch. (Only
* applicable for weak-ordered memory model archs,
@@ -654,160 +616,122 @@ static inline void i40e_release_rx_desc(struct i40e_ring *rx_ring, u32 val)
}
/**
- * i40evf_alloc_rx_buffers_ps - Replace used receive buffers; packet split
- * @rx_ring: ring to place buffers on
- * @cleaned_count: number of buffers to replace
+ * i40e_alloc_mapped_page - recycle or make a new page
+ * @rx_ring: ring to use
+ * @bi: rx_buffer struct to modify
*
- * Returns true if any errors on allocation
+ * Returns true if the page was successfully allocated or
+ * reused.
**/
-bool i40evf_alloc_rx_buffers_ps(struct i40e_ring *rx_ring, u16 cleaned_count)
+static bool i40e_alloc_mapped_page(struct i40e_ring *rx_ring,
+ struct i40e_rx_buffer *bi)
{
- u16 i = rx_ring->next_to_use;
- union i40e_rx_desc *rx_desc;
- struct i40e_rx_buffer *bi;
- const int current_node = numa_node_id();
+ struct page *page = bi->page;
+ dma_addr_t dma;
- /* do nothing if no valid netdev defined */
- if (!rx_ring->netdev || !cleaned_count)
- return false;
+ /* since we are recycling buffers we should seldom need to alloc */
+ if (likely(page)) {
+ rx_ring->rx_stats.page_reuse_count++;
+ return true;
+ }
- while (cleaned_count--) {
- rx_desc = I40E_RX_DESC(rx_ring, i);
- bi = &rx_ring->rx_bi[i];
+ /* alloc new page for storage */
+ page = dev_alloc_page();
+ if (unlikely(!page)) {
+ rx_ring->rx_stats.alloc_page_failed++;
+ return false;
+ }
- if (bi->skb) /* desc is in use */
- goto no_buffers;
+ /* map page for use */
+ dma = dma_map_page(rx_ring->dev, page, 0, PAGE_SIZE, DMA_FROM_DEVICE);
- /* If we've been moved to a different NUMA node, release the
- * page so we can get a new one on the current node.
+ /* if mapping failed free memory back to system since
+ * there isn't much point in holding memory we can't use
*/
- if (bi->page && page_to_nid(bi->page) != current_node) {
- dma_unmap_page(rx_ring->dev,
- bi->page_dma,
- PAGE_SIZE,
- DMA_FROM_DEVICE);
- __free_page(bi->page);
- bi->page = NULL;
- bi->page_dma = 0;
- rx_ring->rx_stats.realloc_count++;
- } else if (bi->page) {
- rx_ring->rx_stats.page_reuse_count++;
- }
-
- if (!bi->page) {
- bi->page = alloc_page(GFP_ATOMIC);
- if (!bi->page) {
- rx_ring->rx_stats.alloc_page_failed++;
- goto no_buffers;
- }
- bi->page_dma = dma_map_page(rx_ring->dev,
- bi->page,
- 0,
- PAGE_SIZE,
- DMA_FROM_DEVICE);
- if (dma_mapping_error(rx_ring->dev, bi->page_dma)) {
- rx_ring->rx_stats.alloc_page_failed++;
- __free_page(bi->page);
- bi->page = NULL;
- bi->page_dma = 0;
- bi->page_offset = 0;
- goto no_buffers;
- }
- bi->page_offset = 0;
- }
-
- /* Refresh the desc even if buffer_addrs didn't change
- * because each write-back erases this info.
- */
- rx_desc->read.pkt_addr =
- cpu_to_le64(bi->page_dma + bi->page_offset);
- rx_desc->read.hdr_addr = cpu_to_le64(bi->dma);
- i++;
- if (i == rx_ring->count)
- i = 0;
+ if (dma_mapping_error(rx_ring->dev, dma)) {
+ __free_pages(page, 0);
+ rx_ring->rx_stats.alloc_page_failed++;
+ return false;
}
- if (rx_ring->next_to_use != i)
- i40e_release_rx_desc(rx_ring, i);
+ bi->dma = dma;
+ bi->page = page;
+ bi->page_offset = 0;
- return false;
+ return true;
+}
-no_buffers:
- if (rx_ring->next_to_use != i)
- i40e_release_rx_desc(rx_ring, i);
+/**
+ * i40e_receive_skb - Send a completed packet up the stack
+ * @rx_ring: rx ring in play
+ * @skb: packet to send up
+ * @vlan_tag: vlan tag for packet
+ **/
+static void i40e_receive_skb(struct i40e_ring *rx_ring,
+ struct sk_buff *skb, u16 vlan_tag)
+{
+ struct i40e_q_vector *q_vector = rx_ring->q_vector;
- /* make sure to come back via polling to try again after
- * allocation failure
- */
- return true;
+ if ((rx_ring->netdev->features & NETIF_F_HW_VLAN_CTAG_RX) &&
+ (vlan_tag & VLAN_VID_MASK))
+ __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), vlan_tag);
+
+ napi_gro_receive(&q_vector->napi, skb);
}
/**
- * i40evf_alloc_rx_buffers_1buf - Replace used receive buffers; single buffer
+ * i40evf_alloc_rx_buffers - Replace used receive buffers
* @rx_ring: ring to place buffers on
* @cleaned_count: number of buffers to replace
*
- * Returns true if any errors on allocation
+ * Returns false if all allocations were successful, true if any fail
**/
-bool i40evf_alloc_rx_buffers_1buf(struct i40e_ring *rx_ring, u16 cleaned_count)
+bool i40evf_alloc_rx_buffers(struct i40e_ring *rx_ring, u16 cleaned_count)
{
- u16 i = rx_ring->next_to_use;
+ u16 ntu = rx_ring->next_to_use;
union i40e_rx_desc *rx_desc;
struct i40e_rx_buffer *bi;
- struct sk_buff *skb;
/* do nothing if no valid netdev defined */
if (!rx_ring->netdev || !cleaned_count)
return false;
- while (cleaned_count--) {
- rx_desc = I40E_RX_DESC(rx_ring, i);
- bi = &rx_ring->rx_bi[i];
- skb = bi->skb;
-
- if (!skb) {
- skb = __netdev_alloc_skb_ip_align(rx_ring->netdev,
- rx_ring->rx_buf_len,
- GFP_ATOMIC |
- __GFP_NOWARN);
- if (!skb) {
- rx_ring->rx_stats.alloc_buff_failed++;
- goto no_buffers;
- }
- /* initialize queue mapping */
- skb_record_rx_queue(skb, rx_ring->queue_index);
- bi->skb = skb;
- }
+ rx_desc = I40E_RX_DESC(rx_ring, ntu);
+ bi = &rx_ring->rx_bi[ntu];
- if (!bi->dma) {
- bi->dma = dma_map_single(rx_ring->dev,
- skb->data,
- rx_ring->rx_buf_len,
- DMA_FROM_DEVICE);
- if (dma_mapping_error(rx_ring->dev, bi->dma)) {
- rx_ring->rx_stats.alloc_buff_failed++;
- bi->dma = 0;
- dev_kfree_skb(bi->skb);
- bi->skb = NULL;
- goto no_buffers;
- }
- }
+ do {
+ if (!i40e_alloc_mapped_page(rx_ring, bi))
+ goto no_buffers;
- rx_desc->read.pkt_addr = cpu_to_le64(bi->dma);
+ /* Refresh the desc even if buffer_addrs didn't change
+ * because each write-back erases this info.
+ */
+ rx_desc->read.pkt_addr = cpu_to_le64(bi->dma + bi->page_offset);
rx_desc->read.hdr_addr = 0;
- i++;
- if (i == rx_ring->count)
- i = 0;
- }
- if (rx_ring->next_to_use != i)
- i40e_release_rx_desc(rx_ring, i);
+ rx_desc++;
+ bi++;
+ ntu++;
+ if (unlikely(ntu == rx_ring->count)) {
+ rx_desc = I40E_RX_DESC(rx_ring, 0);
+ bi = rx_ring->rx_bi;
+ ntu = 0;
+ }
+
+ /* clear the status bits for the next_to_use descriptor */
+ rx_desc->wb.qword1.status_error_len = 0;
+
+ cleaned_count--;
+ } while (cleaned_count);
+
+ if (rx_ring->next_to_use != ntu)
+ i40e_release_rx_desc(rx_ring, ntu);
return false;
no_buffers:
- if (rx_ring->next_to_use != i)
- i40e_release_rx_desc(rx_ring, i);
+ if (rx_ring->next_to_use != ntu)
+ i40e_release_rx_desc(rx_ring, ntu);
/* make sure to come back via polling to try again after
* allocation failure
@@ -816,42 +740,35 @@ no_buffers:
}
/**
- * i40e_receive_skb - Send a completed packet up the stack
- * @rx_ring: rx ring in play
- * @skb: packet to send up
- * @vlan_tag: vlan tag for packet
- **/
-static void i40e_receive_skb(struct i40e_ring *rx_ring,
- struct sk_buff *skb, u16 vlan_tag)
-{
- struct i40e_q_vector *q_vector = rx_ring->q_vector;
-
- if ((rx_ring->netdev->features & NETIF_F_HW_VLAN_CTAG_RX) &&
- (vlan_tag & VLAN_VID_MASK))
- __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), vlan_tag);
-
- napi_gro_receive(&q_vector->napi, skb);
-}
-
-/**
* i40e_rx_checksum - Indicate in skb if hw indicated a good cksum
* @vsi: the VSI we care about
* @skb: skb currently being received and modified
- * @rx_status: status value of last descriptor in packet
- * @rx_error: error value of last descriptor in packet
- * @rx_ptype: ptype value of last descriptor in packet
+ * @rx_desc: the receive descriptor
+ *
+ * skb->protocol must be set before this function is called
**/
static inline void i40e_rx_checksum(struct i40e_vsi *vsi,
struct sk_buff *skb,
- u32 rx_status,
- u32 rx_error,
- u16 rx_ptype)
+ union i40e_rx_desc *rx_desc)
{
- struct i40e_rx_ptype_decoded decoded = decode_rx_desc_ptype(rx_ptype);
+ struct i40e_rx_ptype_decoded decoded;
bool ipv4, ipv6, tunnel = false;
+ u32 rx_error, rx_status;
+ u8 ptype;
+ u64 qword;
+
+ qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
+ ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >> I40E_RXD_QW1_PTYPE_SHIFT;
+ rx_error = (qword & I40E_RXD_QW1_ERROR_MASK) >>
+ I40E_RXD_QW1_ERROR_SHIFT;
+ rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >>
+ I40E_RXD_QW1_STATUS_SHIFT;
+ decoded = decode_rx_desc_ptype(ptype);
skb->ip_summed = CHECKSUM_NONE;
+ skb_checksum_none_assert(skb);
+
/* Rx csum enabled and ip headers found? */
if (!(vsi->netdev->features & NETIF_F_RXCSUM))
return;
@@ -917,7 +834,7 @@ checksum_fail:
*
* Returns a hash type to be used by skb_set_hash
**/
-static inline enum pkt_hash_types i40e_ptype_to_htype(u8 ptype)
+static inline int i40e_ptype_to_htype(u8 ptype)
{
struct i40e_rx_ptype_decoded decoded = decode_rx_desc_ptype(ptype);
@@ -945,7 +862,7 @@ static inline void i40e_rx_hash(struct i40e_ring *ring,
u8 rx_ptype)
{
u32 hash;
- const __le64 rss_mask =
+ const __le64 rss_mask =
cpu_to_le64((u64)I40E_RX_DESC_FLTSTAT_RSS_HASH <<
I40E_RX_DESC_STATUS_FLTSTAT_SHIFT);
@@ -959,315 +876,411 @@ static inline void i40e_rx_hash(struct i40e_ring *ring,
}
/**
- * i40e_clean_rx_irq_ps - Reclaim resources after receive; packet split
- * @rx_ring: rx ring to clean
- * @budget: how many cleans we're allowed
+ * i40evf_process_skb_fields - Populate skb header fields from Rx descriptor
+ * @rx_ring: rx descriptor ring packet is being transacted on
+ * @rx_desc: pointer to the EOP Rx descriptor
+ * @skb: pointer to current skb being populated
+ * @rx_ptype: the packet type decoded by hardware
*
- * Returns true if there's any budget left (e.g. the clean is finished)
+ * This function checks the ring, descriptor, and packet information in
+ * order to populate the hash, checksum, VLAN, protocol, and
+ * other fields within the skb.
**/
-static int i40e_clean_rx_irq_ps(struct i40e_ring *rx_ring, const int budget)
+static inline
+void i40evf_process_skb_fields(struct i40e_ring *rx_ring,
+ union i40e_rx_desc *rx_desc, struct sk_buff *skb,
+ u8 rx_ptype)
{
- unsigned int total_rx_bytes = 0, total_rx_packets = 0;
- u16 rx_packet_len, rx_header_len, rx_sph, rx_hbo;
- u16 cleaned_count = I40E_DESC_UNUSED(rx_ring);
- struct i40e_vsi *vsi = rx_ring->vsi;
- u16 i = rx_ring->next_to_clean;
- union i40e_rx_desc *rx_desc;
- u32 rx_error, rx_status;
- bool failure = false;
- u8 rx_ptype;
- u64 qword;
- u32 copysize;
+ i40e_rx_hash(rx_ring, rx_desc, skb, rx_ptype);
- do {
- struct i40e_rx_buffer *rx_bi;
- struct sk_buff *skb;
- u16 vlan_tag;
- /* return some buffers to hardware, one at a time is too slow */
- if (cleaned_count >= I40E_RX_BUFFER_WRITE) {
- failure = failure ||
- i40evf_alloc_rx_buffers_ps(rx_ring,
- cleaned_count);
- cleaned_count = 0;
- }
+ /* modifies the skb - consumes the enet header */
+ skb->protocol = eth_type_trans(skb, rx_ring->netdev);
- i = rx_ring->next_to_clean;
- rx_desc = I40E_RX_DESC(rx_ring, i);
- qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
- rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >>
- I40E_RXD_QW1_STATUS_SHIFT;
+ i40e_rx_checksum(rx_ring->vsi, skb, rx_desc);
- if (!(rx_status & BIT(I40E_RX_DESC_STATUS_DD_SHIFT)))
- break;
+ skb_record_rx_queue(skb, rx_ring->queue_index);
+}
- /* This memory barrier is needed to keep us from reading
- * any other fields out of the rx_desc until we know the
- * DD bit is set.
- */
- dma_rmb();
- /* sync header buffer for reading */
- dma_sync_single_range_for_cpu(rx_ring->dev,
- rx_ring->rx_bi[0].dma,
- i * rx_ring->rx_hdr_len,
- rx_ring->rx_hdr_len,
- DMA_FROM_DEVICE);
- rx_bi = &rx_ring->rx_bi[i];
- skb = rx_bi->skb;
- if (likely(!skb)) {
- skb = __netdev_alloc_skb_ip_align(rx_ring->netdev,
- rx_ring->rx_hdr_len,
- GFP_ATOMIC |
- __GFP_NOWARN);
- if (!skb) {
- rx_ring->rx_stats.alloc_buff_failed++;
- failure = true;
- break;
- }
+/**
+ * i40e_pull_tail - i40e specific version of skb_pull_tail
+ * @rx_ring: rx descriptor ring packet is being transacted on
+ * @skb: pointer to current skb being adjusted
+ *
+ * This function is an i40e specific version of __pskb_pull_tail. The
+ * main difference between this version and the original function is that
+ * this function can make several assumptions about the state of things
+ * that allow for significant optimizations versus the standard function.
+ * As a result we can do things like drop a frag and maintain an accurate
+ * truesize for the skb.
+ */
+static void i40e_pull_tail(struct i40e_ring *rx_ring, struct sk_buff *skb)
+{
+ struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[0];
+ unsigned char *va;
+ unsigned int pull_len;
- /* initialize queue mapping */
- skb_record_rx_queue(skb, rx_ring->queue_index);
- /* we are reusing so sync this buffer for CPU use */
- dma_sync_single_range_for_cpu(rx_ring->dev,
- rx_ring->rx_bi[0].dma,
- i * rx_ring->rx_hdr_len,
- rx_ring->rx_hdr_len,
- DMA_FROM_DEVICE);
- }
- rx_packet_len = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
- I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
- rx_header_len = (qword & I40E_RXD_QW1_LENGTH_HBUF_MASK) >>
- I40E_RXD_QW1_LENGTH_HBUF_SHIFT;
- rx_sph = (qword & I40E_RXD_QW1_LENGTH_SPH_MASK) >>
- I40E_RXD_QW1_LENGTH_SPH_SHIFT;
-
- rx_error = (qword & I40E_RXD_QW1_ERROR_MASK) >>
- I40E_RXD_QW1_ERROR_SHIFT;
- rx_hbo = rx_error & BIT(I40E_RX_DESC_ERROR_HBO_SHIFT);
- rx_error &= ~BIT(I40E_RX_DESC_ERROR_HBO_SHIFT);
+ /* it is valid to use page_address instead of kmap since we are
+ * working with pages allocated out of the lomem pool per
+ * alloc_page(GFP_ATOMIC)
+ */
+ va = skb_frag_address(frag);
- rx_ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >>
- I40E_RXD_QW1_PTYPE_SHIFT;
- /* sync half-page for reading */
- dma_sync_single_range_for_cpu(rx_ring->dev,
- rx_bi->page_dma,
- rx_bi->page_offset,
- PAGE_SIZE / 2,
- DMA_FROM_DEVICE);
- prefetch(page_address(rx_bi->page) + rx_bi->page_offset);
- rx_bi->skb = NULL;
- cleaned_count++;
- copysize = 0;
- if (rx_hbo || rx_sph) {
- int len;
-
- if (rx_hbo)
- len = I40E_RX_HDR_SIZE;
- else
- len = rx_header_len;
- memcpy(__skb_put(skb, len), rx_bi->hdr_buf, len);
- } else if (skb->len == 0) {
- int len;
- unsigned char *va = page_address(rx_bi->page) +
- rx_bi->page_offset;
-
- len = min(rx_packet_len, rx_ring->rx_hdr_len);
- memcpy(__skb_put(skb, len), va, len);
- copysize = len;
- rx_packet_len -= len;
- }
- /* Get the rest of the data if this was a header split */
- if (rx_packet_len) {
- skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags,
- rx_bi->page,
- rx_bi->page_offset + copysize,
- rx_packet_len, I40E_RXBUFFER_2048);
-
- /* If the page count is more than 2, then both halves
- * of the page are used and we need to free it. Do it
- * here instead of in the alloc code. Otherwise one
- * of the half-pages might be released between now and
- * then, and we wouldn't know which one to use.
- * Don't call get_page and free_page since those are
- * both expensive atomic operations that just change
- * the refcount in opposite directions. Just give the
- * page to the stack; he can have our refcount.
- */
- if (page_count(rx_bi->page) > 2) {
- dma_unmap_page(rx_ring->dev,
- rx_bi->page_dma,
- PAGE_SIZE,
- DMA_FROM_DEVICE);
- rx_bi->page = NULL;
- rx_bi->page_dma = 0;
- rx_ring->rx_stats.realloc_count++;
- } else {
- get_page(rx_bi->page);
- /* switch to the other half-page here; the
- * allocation code programs the right addr
- * into HW. If we haven't used this half-page,
- * the address won't be changed, and HW can
- * just use it next time through.
- */
- rx_bi->page_offset ^= PAGE_SIZE / 2;
- }
+ /* we need the header to contain the greater of either ETH_HLEN or
+ * 60 bytes if the skb->len is less than 60 for skb_pad.
+ */
+ pull_len = eth_get_headlen(va, I40E_RX_HDR_SIZE);
- }
- I40E_RX_INCREMENT(rx_ring, i);
+ /* align pull length to size of long to optimize memcpy performance */
+ skb_copy_to_linear_data(skb, va, ALIGN(pull_len, sizeof(long)));
- if (unlikely(
- !(rx_status & BIT(I40E_RX_DESC_STATUS_EOF_SHIFT)))) {
- struct i40e_rx_buffer *next_buffer;
+ /* update all of the pointers */
+ skb_frag_size_sub(frag, pull_len);
+ frag->page_offset += pull_len;
+ skb->data_len -= pull_len;
+ skb->tail += pull_len;
+}
- next_buffer = &rx_ring->rx_bi[i];
- next_buffer->skb = skb;
- rx_ring->rx_stats.non_eop_descs++;
- continue;
- }
+/**
+ * i40e_cleanup_headers - Correct empty headers
+ * @rx_ring: rx descriptor ring packet is being transacted on
+ * @skb: pointer to current skb being fixed
+ *
+ * Also address the case where we are pulling data in on pages only
+ * and as such no data is present in the skb header.
+ *
+ * In addition if skb is not at least 60 bytes we need to pad it so that
+ * it is large enough to qualify as a valid Ethernet frame.
+ *
+ * Returns true if an error was encountered and skb was freed.
+ **/
+static bool i40e_cleanup_headers(struct i40e_ring *rx_ring, struct sk_buff *skb)
+{
+ /* place header in linear portion of buffer */
+ if (skb_is_nonlinear(skb))
+ i40e_pull_tail(rx_ring, skb);
- /* ERR_MASK will only have valid bits if EOP set */
- if (unlikely(rx_error & BIT(I40E_RX_DESC_ERROR_RXE_SHIFT))) {
- dev_kfree_skb_any(skb);
- continue;
- }
+ /* if eth_skb_pad returns an error the skb was freed */
+ if (eth_skb_pad(skb))
+ return true;
- i40e_rx_hash(rx_ring, rx_desc, skb, rx_ptype);
+ return false;
+}
- /* probably a little skewed due to removing CRC */
- total_rx_bytes += skb->len;
- total_rx_packets++;
+/**
+ * i40e_reuse_rx_page - page flip buffer and store it back on the ring
+ * @rx_ring: rx descriptor ring to store buffers on
+ * @old_buff: donor buffer to have page reused
+ *
+ * Synchronizes page for reuse by the adapter
+ **/
+static void i40e_reuse_rx_page(struct i40e_ring *rx_ring,
+ struct i40e_rx_buffer *old_buff)
+{
+ struct i40e_rx_buffer *new_buff;
+ u16 nta = rx_ring->next_to_alloc;
- skb->protocol = eth_type_trans(skb, rx_ring->netdev);
+ new_buff = &rx_ring->rx_bi[nta];
- i40e_rx_checksum(vsi, skb, rx_status, rx_error, rx_ptype);
+ /* update, and store next to alloc */
+ nta++;
+ rx_ring->next_to_alloc = (nta < rx_ring->count) ? nta : 0;
- vlan_tag = rx_status & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)
- ? le16_to_cpu(rx_desc->wb.qword0.lo_dword.l2tag1)
- : 0;
-#ifdef I40E_FCOE
- if (unlikely(
- i40e_rx_is_fcoe(rx_ptype) &&
- !i40e_fcoe_handle_offload(rx_ring, rx_desc, skb))) {
- dev_kfree_skb_any(skb);
- continue;
- }
+ /* transfer page from old buffer to new buffer */
+ *new_buff = *old_buff;
+}
+
+/**
+ * i40e_page_is_reserved - check if reuse is possible
+ * @page: page struct to check
+ */
+static inline bool i40e_page_is_reserved(struct page *page)
+{
+ return (page_to_nid(page) != numa_mem_id()) || page_is_pfmemalloc(page);
+}
+
+/**
+ * i40e_add_rx_frag - Add contents of Rx buffer to sk_buff
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @rx_buffer: buffer containing page to add
+ * @rx_desc: descriptor containing length of buffer written by hardware
+ * @skb: sk_buff to place the data into
+ *
+ * This function will add the data contained in rx_buffer->page to the skb.
+ * This is done either through a direct copy if the data in the buffer is
+ * less than the skb header size, otherwise it will just attach the page as
+ * a frag to the skb.
+ *
+ * The function will then update the page offset if necessary and return
+ * true if the buffer can be reused by the adapter.
+ **/
+static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
+ struct i40e_rx_buffer *rx_buffer,
+ union i40e_rx_desc *rx_desc,
+ struct sk_buff *skb)
+{
+ struct page *page = rx_buffer->page;
+ u64 qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
+ unsigned int size = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
+ I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
+#if (PAGE_SIZE < 8192)
+ unsigned int truesize = I40E_RXBUFFER_2048;
+#else
+ unsigned int truesize = ALIGN(size, L1_CACHE_BYTES);
+ unsigned int last_offset = PAGE_SIZE - I40E_RXBUFFER_2048;
#endif
- i40e_receive_skb(rx_ring, skb, vlan_tag);
- rx_desc->wb.qword1.status_error_len = 0;
+ /* will the data fit in the skb we allocated? if so, just
+ * copy it as it is pretty small anyway
+ */
+ if ((size <= I40E_RX_HDR_SIZE) && !skb_is_nonlinear(skb)) {
+ unsigned char *va = page_address(page) + rx_buffer->page_offset;
- } while (likely(total_rx_packets < budget));
+ memcpy(__skb_put(skb, size), va, ALIGN(size, sizeof(long)));
- u64_stats_update_begin(&rx_ring->syncp);
- rx_ring->stats.packets += total_rx_packets;
- rx_ring->stats.bytes += total_rx_bytes;
- u64_stats_update_end(&rx_ring->syncp);
- rx_ring->q_vector->rx.total_packets += total_rx_packets;
- rx_ring->q_vector->rx.total_bytes += total_rx_bytes;
+ /* page is not reserved, we can reuse buffer as-is */
+ if (likely(!i40e_page_is_reserved(page)))
+ return true;
- return failure ? budget : total_rx_packets;
+ /* this page cannot be reused so discard it */
+ __free_pages(page, 0);
+ return false;
+ }
+
+ skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, page,
+ rx_buffer->page_offset, size, truesize);
+
+ /* avoid re-using remote pages */
+ if (unlikely(i40e_page_is_reserved(page)))
+ return false;
+
+#if (PAGE_SIZE < 8192)
+ /* if we are only owner of page we can reuse it */
+ if (unlikely(page_count(page) != 1))
+ return false;
+
+ /* flip page offset to other buffer */
+ rx_buffer->page_offset ^= truesize;
+#else
+ /* move offset up to the next cache line */
+ rx_buffer->page_offset += truesize;
+
+ if (rx_buffer->page_offset > last_offset)
+ return false;
+#endif
+
+ /* Even if we own the page, we are not allowed to use atomic_set()
+ * This would break get_page_unless_zero() users.
+ */
+ get_page(rx_buffer->page);
+
+ return true;
}
/**
- * i40e_clean_rx_irq_1buf - Reclaim resources after receive; single buffer
- * @rx_ring: rx ring to clean
- * @budget: how many cleans we're allowed
+ * i40evf_fetch_rx_buffer - Allocate skb and populate it
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @rx_desc: descriptor containing info written by hardware
*
- * Returns number of packets cleaned
+ * This function allocates an skb on the fly, and populates it with the page
+ * data from the current receive descriptor, taking care to set up the skb
+ * correctly, as well as handling calling the page recycle function if
+ * necessary.
+ */
+static inline
+struct sk_buff *i40evf_fetch_rx_buffer(struct i40e_ring *rx_ring,
+ union i40e_rx_desc *rx_desc)
+{
+ struct i40e_rx_buffer *rx_buffer;
+ struct sk_buff *skb;
+ struct page *page;
+
+ rx_buffer = &rx_ring->rx_bi[rx_ring->next_to_clean];
+ page = rx_buffer->page;
+ prefetchw(page);
+
+ skb = rx_buffer->skb;
+
+ if (likely(!skb)) {
+ void *page_addr = page_address(page) + rx_buffer->page_offset;
+
+ /* prefetch first cache line of first page */
+ prefetch(page_addr);
+#if L1_CACHE_BYTES < 128
+ prefetch(page_addr + L1_CACHE_BYTES);
+#endif
+
+ /* allocate a skb to store the frags */
+ skb = __napi_alloc_skb(&rx_ring->q_vector->napi,
+ I40E_RX_HDR_SIZE,
+ GFP_ATOMIC | __GFP_NOWARN);
+ if (unlikely(!skb)) {
+ rx_ring->rx_stats.alloc_buff_failed++;
+ return NULL;
+ }
+
+ /* we will be copying header into skb->data in
+ * pskb_may_pull so it is in our interest to prefetch
+ * it now to avoid a possible cache miss
+ */
+ prefetchw(skb->data);
+ } else {
+ rx_buffer->skb = NULL;
+ }
+
+ /* we are reusing so sync this buffer for CPU use */
+ dma_sync_single_range_for_cpu(rx_ring->dev,
+ rx_buffer->dma,
+ rx_buffer->page_offset,
+ I40E_RXBUFFER_2048,
+ DMA_FROM_DEVICE);
+
+ /* pull page into skb */
+ if (i40e_add_rx_frag(rx_ring, rx_buffer, rx_desc, skb)) {
+ /* hand second half of page back to the ring */
+ i40e_reuse_rx_page(rx_ring, rx_buffer);
+ rx_ring->rx_stats.page_reuse_count++;
+ } else {
+ /* we are not reusing the buffer so unmap it */
+ dma_unmap_page(rx_ring->dev, rx_buffer->dma, PAGE_SIZE,
+ DMA_FROM_DEVICE);
+ }
+
+ /* clear contents of buffer_info */
+ rx_buffer->page = NULL;
+
+ return skb;
+}
+
+/**
+ * i40e_is_non_eop - process handling of non-EOP buffers
+ * @rx_ring: Rx ring being processed
+ * @rx_desc: Rx descriptor for current buffer
+ * @skb: Current socket buffer containing buffer in progress
+ *
+ * This function updates next to clean. If the buffer is an EOP buffer
+ * this function exits returning false, otherwise it will place the
+ * sk_buff in the next buffer to be chained and return true indicating
+ * that this is in fact a non-EOP buffer.
**/
-static int i40e_clean_rx_irq_1buf(struct i40e_ring *rx_ring, int budget)
+static bool i40e_is_non_eop(struct i40e_ring *rx_ring,
+ union i40e_rx_desc *rx_desc,
+ struct sk_buff *skb)
+{
+ u32 ntc = rx_ring->next_to_clean + 1;
+
+ /* fetch, update, and store next to clean */
+ ntc = (ntc < rx_ring->count) ? ntc : 0;
+ rx_ring->next_to_clean = ntc;
+
+ prefetch(I40E_RX_DESC(rx_ring, ntc));
+
+ /* if we are the last buffer then there is nothing else to do */
+#define I40E_RXD_EOF BIT(I40E_RX_DESC_STATUS_EOF_SHIFT)
+ if (likely(i40e_test_staterr(rx_desc, I40E_RXD_EOF)))
+ return false;
+
+ /* place skb in next buffer to be received */
+ rx_ring->rx_bi[ntc].skb = skb;
+ rx_ring->rx_stats.non_eop_descs++;
+
+ return true;
+}
+
+/**
+ * i40e_clean_rx_irq - Clean completed descriptors from Rx ring - bounce buf
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @budget: Total limit on number of packets to process
+ *
+ * This function provides a "bounce buffer" approach to Rx interrupt
+ * processing. The advantage to this is that on systems that have
+ * expensive overhead for IOMMU access this provides a means of avoiding
+ * it by maintaining the mapping of the page to the system.
+ *
+ * Returns amount of work completed
+ **/
+static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
{
unsigned int total_rx_bytes = 0, total_rx_packets = 0;
u16 cleaned_count = I40E_DESC_UNUSED(rx_ring);
- struct i40e_vsi *vsi = rx_ring->vsi;
- union i40e_rx_desc *rx_desc;
- u32 rx_error, rx_status;
- u16 rx_packet_len;
bool failure = false;
- u8 rx_ptype;
- u64 qword;
- u16 i;
- do {
- struct i40e_rx_buffer *rx_bi;
+ while (likely(total_rx_packets < budget)) {
+ union i40e_rx_desc *rx_desc;
struct sk_buff *skb;
+ u32 rx_status;
u16 vlan_tag;
+ u8 rx_ptype;
+ u64 qword;
+
/* return some buffers to hardware, one at a time is too slow */
if (cleaned_count >= I40E_RX_BUFFER_WRITE) {
failure = failure ||
- i40evf_alloc_rx_buffers_1buf(rx_ring,
- cleaned_count);
+ i40evf_alloc_rx_buffers(rx_ring, cleaned_count);
cleaned_count = 0;
}
- i = rx_ring->next_to_clean;
- rx_desc = I40E_RX_DESC(rx_ring, i);
+ rx_desc = I40E_RX_DESC(rx_ring, rx_ring->next_to_clean);
+
qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
+ rx_ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >>
+ I40E_RXD_QW1_PTYPE_SHIFT;
rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >>
- I40E_RXD_QW1_STATUS_SHIFT;
+ I40E_RXD_QW1_STATUS_SHIFT;
if (!(rx_status & BIT(I40E_RX_DESC_STATUS_DD_SHIFT)))
break;
+ /* status_error_len will always be zero for unused descriptors
+ * because it's cleared in cleanup, and overlaps with hdr_addr
+ * which is always zero because packet split isn't used, if the
+ * hardware wrote DD then it will be non-zero
+ */
+ if (!rx_desc->wb.qword1.status_error_len)
+ break;
+
/* This memory barrier is needed to keep us from reading
* any other fields out of the rx_desc until we know the
* DD bit is set.
*/
dma_rmb();
- rx_bi = &rx_ring->rx_bi[i];
- skb = rx_bi->skb;
- prefetch(skb->data);
-
- rx_packet_len = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
- I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
-
- rx_error = (qword & I40E_RXD_QW1_ERROR_MASK) >>
- I40E_RXD_QW1_ERROR_SHIFT;
- rx_error &= ~BIT(I40E_RX_DESC_ERROR_HBO_SHIFT);
+ skb = i40evf_fetch_rx_buffer(rx_ring, rx_desc);
+ if (!skb)
+ break;
- rx_ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >>
- I40E_RXD_QW1_PTYPE_SHIFT;
- rx_bi->skb = NULL;
cleaned_count++;
- /* Get the header and possibly the whole packet
- * If this is an skb from previous receive dma will be 0
- */
- skb_put(skb, rx_packet_len);
- dma_unmap_single(rx_ring->dev, rx_bi->dma, rx_ring->rx_buf_len,
- DMA_FROM_DEVICE);
- rx_bi->dma = 0;
-
- I40E_RX_INCREMENT(rx_ring, i);
-
- if (unlikely(
- !(rx_status & BIT(I40E_RX_DESC_STATUS_EOF_SHIFT)))) {
- rx_ring->rx_stats.non_eop_descs++;
+ if (i40e_is_non_eop(rx_ring, rx_desc, skb))
continue;
- }
- /* ERR_MASK will only have valid bits if EOP set */
- if (unlikely(rx_error & BIT(I40E_RX_DESC_ERROR_RXE_SHIFT))) {
+ /* ERR_MASK will only have valid bits if EOP set, and
+ * what we are doing here is actually checking
+ * I40E_RX_DESC_ERROR_RXE_SHIFT, since it is the zeroth bit in
+ * the error field
+ */
+ if (unlikely(i40e_test_staterr(rx_desc, BIT(I40E_RXD_QW1_ERROR_SHIFT)))) {
dev_kfree_skb_any(skb);
continue;
}
- i40e_rx_hash(rx_ring, rx_desc, skb, rx_ptype);
+ if (i40e_cleanup_headers(rx_ring, skb))
+ continue;
+
/* probably a little skewed due to removing CRC */
total_rx_bytes += skb->len;
- total_rx_packets++;
- skb->protocol = eth_type_trans(skb, rx_ring->netdev);
+ /* populate checksum, VLAN, and protocol */
+ i40evf_process_skb_fields(rx_ring, rx_desc, skb, rx_ptype);
- i40e_rx_checksum(vsi, skb, rx_status, rx_error, rx_ptype);
- vlan_tag = rx_status & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)
- ? le16_to_cpu(rx_desc->wb.qword0.lo_dword.l2tag1)
- : 0;
+ vlan_tag = (qword & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)) ?
+ le16_to_cpu(rx_desc->wb.qword0.lo_dword.l2tag1) : 0;
+
i40e_receive_skb(rx_ring, skb, vlan_tag);
- rx_desc->wb.qword1.status_error_len = 0;
- } while (likely(total_rx_packets < budget));
+ /* update budget accounting */
+ total_rx_packets++;
+ }
u64_stats_update_begin(&rx_ring->syncp);
rx_ring->stats.packets += total_rx_packets;
@@ -1276,6 +1289,7 @@ static int i40e_clean_rx_irq_1buf(struct i40e_ring *rx_ring, int budget)
rx_ring->q_vector->rx.total_packets += total_rx_packets;
rx_ring->q_vector->rx.total_bytes += total_rx_bytes;
+ /* guarantee a trip back through this routine if there was a failure */
return failure ? budget : total_rx_packets;
}
@@ -1417,9 +1431,7 @@ int i40evf_napi_poll(struct napi_struct *napi, int budget)
budget_per_ring = max(budget/q_vector->num_ringpairs, 1);
i40e_for_each_ring(ring, q_vector->rx) {
- int cleaned;
-
- cleaned = i40e_clean_rx_irq_1buf(ring, budget_per_ring);
+ int cleaned = i40e_clean_rx_irq(ring, budget_per_ring);
work_done += cleaned;
/* if we clean as many as budgeted, we must not be done */
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
index f24a97e..4ba302e 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
@@ -102,8 +102,8 @@ enum i40e_dyn_idx_t {
(((pf)->flags & I40E_FLAG_MULTIPLE_TCP_UDP_RSS_PCTYPE) ? \
I40E_DEFAULT_RSS_HENA_EXPANDED : I40E_DEFAULT_RSS_HENA)
-/* Supported Rx Buffer Sizes */
-#define I40E_RXBUFFER_512 512 /* Used for packet split */
+/* Supported Rx Buffer Sizes (a multiple of 128) */
+#define I40E_RXBUFFER_256 256
#define I40E_RXBUFFER_2048 2048
#define I40E_RXBUFFER_3072 3072 /* For FCoE MTU of 2158 */
#define I40E_RXBUFFER_4096 4096
@@ -114,9 +114,28 @@ enum i40e_dyn_idx_t {
* reserve 2 more, and skb_shared_info adds an additional 384 bytes more,
* this adds up to 512 bytes of extra data meaning the smallest allocation
* we could have is 1K.
- * i.e. RXBUFFER_512 --> size-1024 slab
+ * i.e. RXBUFFER_256 --> 960 byte skb (size-1024 slab)
+ * i.e. RXBUFFER_512 --> 1216 byte skb (size-2048 slab)
*/
-#define I40E_RX_HDR_SIZE I40E_RXBUFFER_512
+#define I40E_RX_HDR_SIZE I40E_RXBUFFER_256
+#define i40e_rx_desc i40e_32byte_rx_desc
+
+/**
+ * i40e_test_staterr - tests bits in Rx descriptor status and error fields
+ * @rx_desc: pointer to receive descriptor (in le64 format)
+ * @stat_err_bits: value to mask
+ *
+ * This function does some fast chicanery in order to return the
+ * value of the mask which is really only used for boolean tests.
+ * The status_error_len doesn't need to be shifted because it begins
+ * at offset zero.
+ */
+static inline bool i40e_test_staterr(union i40e_rx_desc *rx_desc,
+ const u64 stat_err_bits)
+{
+ return !!(rx_desc->wb.qword1.status_error_len &
+ cpu_to_le64(stat_err_bits));
+}
/* How many Rx Buffers do we bundle into one write to the hardware ? */
#define I40E_RX_BUFFER_WRITE 16 /* Must be power of 2 */
@@ -142,8 +161,6 @@ enum i40e_dyn_idx_t {
prefetch((n)); \
} while (0)
-#define i40e_rx_desc i40e_32byte_rx_desc
-
#define I40E_MAX_BUFFER_TXD 8
#define I40E_MIN_TX_LEN 17
@@ -212,10 +229,8 @@ struct i40e_tx_buffer {
struct i40e_rx_buffer {
struct sk_buff *skb;
- void *hdr_buf;
dma_addr_t dma;
struct page *page;
- dma_addr_t page_dma;
unsigned int page_offset;
};
@@ -271,7 +286,6 @@ struct i40e_ring {
u16 count; /* Number of descriptors */
u16 reg_idx; /* HW register index of the ring */
- u16 rx_hdr_len;
u16 rx_buf_len;
#define I40E_RX_DTYPE_NO_SPLIT 0
#define I40E_RX_DTYPE_HEADER_SPLIT 1
@@ -311,6 +325,7 @@ struct i40e_ring {
struct i40e_q_vector *q_vector; /* Backreference to associated vector */
struct rcu_head rcu; /* to avoid race on free */
+ u16 next_to_alloc;
} ____cacheline_internodealigned_in_smp;
enum i40e_latency_range {
@@ -334,9 +349,7 @@ struct i40e_ring_container {
#define i40e_for_each_ring(pos, head) \
for (pos = (head).ring; pos != NULL; pos = pos->next)
-bool i40evf_alloc_rx_buffers_ps(struct i40e_ring *rxr, u16 cleaned_count);
-bool i40evf_alloc_rx_buffers_1buf(struct i40e_ring *rxr, u16 cleaned_count);
-void i40evf_alloc_rx_headers(struct i40e_ring *rxr);
+bool i40evf_alloc_rx_buffers(struct i40e_ring *rxr, u16 cleaned_count);
netdev_tx_t i40evf_xmit_frame(struct sk_buff *skb, struct net_device *netdev);
void i40evf_clean_tx_ring(struct i40e_ring *tx_ring);
void i40evf_clean_rx_ring(struct i40e_ring *rx_ring);
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf.h b/drivers/net/ethernet/intel/i40evf/i40evf.h
index 83ccc58..fa044a9 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf.h
+++ b/drivers/net/ethernet/intel/i40evf/i40evf.h
@@ -80,9 +80,6 @@ struct i40e_vsi {
#define I40EVF_REQ_DESCRIPTOR_MULTIPLE 32
/* Supported Rx Buffer Sizes */
-#define I40EVF_RXBUFFER_64 64 /* Used for packet split */
-#define I40EVF_RXBUFFER_128 128 /* Used for packet split */
-#define I40EVF_RXBUFFER_256 256 /* Used for packet split */
#define I40EVF_RXBUFFER_2048 2048
#define I40EVF_MAX_RXBUFFER 16384 /* largest size for single descriptor */
#define I40EVF_MAX_AQ_BUF_SIZE 4096
@@ -208,7 +205,6 @@ struct i40evf_adapter {
u32 flags;
#define I40EVF_FLAG_RX_CSUM_ENABLED BIT(0)
-#define I40EVF_FLAG_RX_1BUF_CAPABLE BIT(1)
#define I40EVF_FLAG_IMIR_ENABLED BIT(5)
#define I40EVF_FLAG_MQ_CAPABLE BIT(6)
#define I40EVF_FLAG_NEED_LINK_UPDATE BIT(7)
@@ -293,7 +289,6 @@ struct i40evf_adapter {
/* Ethtool Private Flags */
-#define I40EVF_PRIV_FLAGS_PS BIT(0)
/* needed by i40evf_ethtool.c */
extern char i40evf_driver_name[];
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
index e972ebc..c9c202f 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
@@ -63,12 +63,6 @@ static const struct i40evf_stats i40evf_gstrings_stats[] = {
#define I40EVF_STATS_LEN(_dev) \
(I40EVF_GLOBAL_STATS_LEN + I40EVF_QUEUE_STATS_LEN(_dev))
-static const char i40evf_priv_flags_strings[][ETH_GSTRING_LEN] = {
- "packet-split",
-};
-
-#define I40EVF_PRIV_FLAGS_STR_LEN ARRAY_SIZE(i40evf_priv_flags_strings)
-
/**
* i40evf_get_settings - Get Link Speed and Duplex settings
* @netdev: network interface device structure
@@ -103,8 +97,6 @@ static int i40evf_get_sset_count(struct net_device *netdev, int sset)
{
if (sset == ETH_SS_STATS)
return I40EVF_STATS_LEN(netdev);
- else if (sset == ETH_SS_PRIV_FLAGS)
- return I40EVF_PRIV_FLAGS_STR_LEN;
else
return -EINVAL;
}
@@ -170,12 +162,6 @@ static void i40evf_get_strings(struct net_device *netdev, u32 sset, u8 *data)
snprintf(p, ETH_GSTRING_LEN, "rx-%u.bytes", i);
p += ETH_GSTRING_LEN;
}
- } else if (sset == ETH_SS_PRIV_FLAGS) {
- for (i = 0; i < I40EVF_PRIV_FLAGS_STR_LEN; i++) {
- memcpy(data, i40evf_priv_flags_strings[i],
- ETH_GSTRING_LEN);
- data += ETH_GSTRING_LEN;
- }
}
}
@@ -225,7 +211,6 @@ static void i40evf_get_drvinfo(struct net_device *netdev,
strlcpy(drvinfo->version, i40evf_driver_version, 32);
strlcpy(drvinfo->fw_version, "N/A", 4);
strlcpy(drvinfo->bus_info, pci_name(adapter->pdev), 32);
- drvinfo->n_priv_flags = I40EVF_PRIV_FLAGS_STR_LEN;
}
/**
@@ -515,40 +500,6 @@ static int i40evf_set_rxfh(struct net_device *netdev, const u32 *indir,
return i40evf_config_rss(adapter);
}
-/**
- * i40evf_get_priv_flags - report device private flags
- * @dev: network interface device structure
- *
- * The get string set count and the string set should be matched for each
- * flag returned. Add new strings for each flag to the i40e_priv_flags_strings
- * array.
- *
- * Returns a u32 bitmap of flags.
- **/
-static u32 i40evf_get_priv_flags(struct net_device *dev)
-{
- u32 ret_flags = 0;
-
- return ret_flags;
-}
-
-/**
- * i40evf_set_priv_flags - set private flags
- * @dev: network interface device structure
- * @flags: bit flags to be set
- **/
-static int i40evf_set_priv_flags(struct net_device *dev, u32 flags)
-{
- struct i40evf_adapter *adapter = netdev_priv(dev);
- bool reset_required = false;
-
- /* if needed, issue reset to cause things to take effect */
- if (reset_required)
- i40evf_schedule_reset(adapter);
-
- return 0;
-}
-
static const struct ethtool_ops i40evf_ethtool_ops = {
.get_settings = i40evf_get_settings,
.get_drvinfo = i40evf_get_drvinfo,
@@ -558,8 +509,6 @@ static const struct ethtool_ops i40evf_ethtool_ops = {
.get_strings = i40evf_get_strings,
.get_ethtool_stats = i40evf_get_ethtool_stats,
.get_sset_count = i40evf_get_sset_count,
- .get_priv_flags = i40evf_get_priv_flags,
- .set_priv_flags = i40evf_set_priv_flags,
.get_msglevel = i40evf_get_msglevel,
.set_msglevel = i40evf_set_msglevel,
.get_coalesce = i40evf_get_coalesce,
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 52408bc..870bad8 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -990,7 +990,7 @@ static void i40evf_configure(struct i40evf_adapter *adapter)
for (i = 0; i < adapter->num_active_queues; i++) {
struct i40e_ring *ring = &adapter->rx_rings[i];
- i40evf_alloc_rx_buffers_1buf(ring, ring->count);
+ i40evf_alloc_rx_buffers(ring, ring->count);
ring->next_to_use = ring->count - 1;
writel(ring->next_to_use, ring->tail);
}
@@ -2401,7 +2401,6 @@ static void i40evf_init_task(struct work_struct *work)
adapter->current_op = I40E_VIRTCHNL_OP_UNKNOWN;
adapter->flags |= I40EVF_FLAG_RX_CSUM_ENABLED;
- adapter->flags |= I40EVF_FLAG_RX_1BUF_CAPABLE;
netdev->netdev_ops = &i40evf_netdev_ops;
i40evf_set_ethtool_ops(netdev);
--
2.5.5
^ permalink raw reply related
* [net-next 09/11] i40e: Test memory before ethtool alloc succeeds
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem; +Cc: Jesse Brandeburg, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
When testing on systems with very limited amounts of RAM, a bug was
found where, while changing the number of descriptors using ethtool,
the driver didn't test the limits of system memory before permanently
assuming it would be able to get receive buffer memory.
Work around this issue by pre-allocation of the receive buffer
memory, in the "ghost" ring, which is then used during reinit
using the new ring length.
Change-Id: I92d7a5fb59a6c884b2efdd1ec652845f101c3359
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@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_ethtool.c | 34 +++++++++++++++++++++++---
1 file changed, 31 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index 6fd730ac..51a994d 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -1274,6 +1274,13 @@ static int i40e_set_ringparam(struct net_device *netdev,
}
for (i = 0; i < vsi->num_queue_pairs; i++) {
+ /* this is to allow wr32 to have something to write to
+ * during early allocation of Rx buffers
+ */
+ u32 __iomem faketail = 0;
+ struct i40e_ring *ring;
+ u16 unused;
+
/* clone ring and setup updated count */
rx_rings[i] = *vsi->rx_rings[i];
rx_rings[i].count = new_rx_count;
@@ -1282,12 +1289,22 @@ static int i40e_set_ringparam(struct net_device *netdev,
*/
rx_rings[i].desc = NULL;
rx_rings[i].rx_bi = NULL;
+ rx_rings[i].tail = (u8 __iomem *)&faketail;
err = i40e_setup_rx_descriptors(&rx_rings[i]);
+ if (err)
+ goto rx_unwind;
+
+ /* now allocate the Rx buffers to make sure the OS
+ * has enough memory, any failure here means abort
+ */
+ ring = &rx_rings[i];
+ unused = I40E_DESC_UNUSED(ring);
+ err = i40e_alloc_rx_buffers(ring, unused);
+rx_unwind:
if (err) {
- while (i) {
- i--;
+ do {
i40e_free_rx_resources(&rx_rings[i]);
- }
+ } while (i--);
kfree(rx_rings);
rx_rings = NULL;
@@ -1313,6 +1330,17 @@ static int i40e_set_ringparam(struct net_device *netdev,
if (rx_rings) {
for (i = 0; i < vsi->num_queue_pairs; i++) {
i40e_free_rx_resources(vsi->rx_rings[i]);
+ /* get the real tail offset */
+ rx_rings[i].tail = vsi->rx_rings[i]->tail;
+ /* this is to fake out the allocation routine
+ * into thinking it has to realloc everything
+ * but the recycling logic will let us re-use
+ * the buffers allocated above
+ */
+ rx_rings[i].next_to_use = 0;
+ rx_rings[i].next_to_clean = 0;
+ rx_rings[i].next_to_alloc = 0;
+ /* do a struct copy */
*vsi->rx_rings[i] = rx_rings[i];
}
kfree(rx_rings);
--
2.5.5
^ permalink raw reply related
* [net-next 04/11] i40e: Refactor receive routine
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem; +Cc: Jesse Brandeburg, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
This is part 1 of the Rx refactor series, just including
changes to i40e.
This refactor aligns the receive routine with the one in
ixgbe which was highly optimized. This reduces the code
we have to maintain and allows for (hopefully) more readable
and maintainable RX hot path.
In order to do this:
- consolidate the receive path into a single function that doesn't
use packet split but *does* use pages for Rx buffers.
- remove the old _1buf routine
- consolidate several routines into helper functions
- remove ethtool control over packet split
Change-ID: I5ca100721de65992aa0114f8b4bac844b84758e0
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@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.h | 1 -
drivers/net/ethernet/intel/i40e/i40e_debugfs.c | 9 +-
drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 1 -
drivers/net/ethernet/intel/i40e/i40e_main.c | 16 +-
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 770 ++++++++++++++++---------
drivers/net/ethernet/intel/i40e/i40e_txrx.h | 37 +-
6 files changed, 531 insertions(+), 303 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index ea6a69a..ebf423b 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -531,7 +531,6 @@ struct i40e_vsi {
u8 *rss_lut_user; /* User configured lookup table entries */
u16 max_frame;
- u16 rx_hdr_len;
u16 rx_buf_len;
u8 dtype;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
index c0a01e0..8ae30f7 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
@@ -268,9 +268,8 @@ static void i40e_dbg_dump_vsi_seid(struct i40e_pf *pf, int seid)
rx_ring->queue_index,
rx_ring->reg_idx);
dev_info(&pf->pdev->dev,
- " rx_rings[%i]: rx_hdr_len = %d, rx_buf_len = %d\n",
- i, rx_ring->rx_hdr_len,
- rx_ring->rx_buf_len);
+ " rx_rings[%i]: rx_buf_len = %d\n",
+ i, rx_ring->rx_buf_len);
dev_info(&pf->pdev->dev,
" rx_rings[%i]: next_to_use = %d, next_to_clean = %d, ring_active = %i\n",
i,
@@ -361,8 +360,8 @@ static void i40e_dbg_dump_vsi_seid(struct i40e_pf *pf, int seid)
" work_limit = %d\n",
vsi->work_limit);
dev_info(&pf->pdev->dev,
- " max_frame = %d, rx_hdr_len = %d, rx_buf_len = %d dtype = %d\n",
- vsi->max_frame, vsi->rx_hdr_len, vsi->rx_buf_len, vsi->dtype);
+ " max_frame = %d, rx_buf_len = %d dtype = %d\n",
+ vsi->max_frame, vsi->rx_buf_len, vsi->dtype);
dev_info(&pf->pdev->dev,
" num_q_vectors = %i, base_vector = %i\n",
vsi->num_q_vectors, vsi->base_vector);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index 858e169..6fd730ac 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -235,7 +235,6 @@ static const char i40e_priv_flags_strings[][ETH_GSTRING_LEN] = {
"LinkPolling",
"flow-director-atr",
"veb-stats",
- "packet-split",
"hw-atr-eviction",
};
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 84e8d4e..e466111 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -2855,10 +2855,8 @@ static int i40e_configure_rx_ring(struct i40e_ring *ring)
memset(&rx_ctx, 0, sizeof(rx_ctx));
ring->rx_buf_len = vsi->rx_buf_len;
- ring->rx_hdr_len = vsi->rx_hdr_len;
rx_ctx.dbuff = ring->rx_buf_len >> I40E_RXQ_CTX_DBUFF_SHIFT;
- rx_ctx.hbuff = ring->rx_hdr_len >> I40E_RXQ_CTX_HBUFF_SHIFT;
rx_ctx.base = (ring->dma / 128);
rx_ctx.qlen = ring->count;
@@ -2910,7 +2908,7 @@ static int i40e_configure_rx_ring(struct i40e_ring *ring)
ring->tail = hw->hw_addr + I40E_QRX_TAIL(pf_q);
writel(0, ring->tail);
- i40e_alloc_rx_buffers_1buf(ring, I40E_DESC_UNUSED(ring));
+ i40e_alloc_rx_buffers(ring, I40E_DESC_UNUSED(ring));
return 0;
}
@@ -2949,15 +2947,13 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
else
vsi->max_frame = I40E_RXBUFFER_2048;
- vsi->rx_hdr_len = 0;
- vsi->rx_buf_len = vsi->max_frame;
+ vsi->rx_buf_len = I40E_RXBUFFER_2048;
vsi->dtype = I40E_RX_DTYPE_NO_SPLIT;
#ifdef I40E_FCOE
/* setup rx buffer for FCoE */
if ((vsi->type == I40E_VSI_FCOE) &&
(vsi->back->flags & I40E_FLAG_FCOE_ENABLED)) {
- vsi->rx_hdr_len = 0;
vsi->rx_buf_len = I40E_RXBUFFER_3072;
vsi->max_frame = I40E_RXBUFFER_3072;
vsi->dtype = I40E_RX_DTYPE_NO_SPLIT;
@@ -2965,8 +2961,6 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
#endif /* I40E_FCOE */
/* round up for the chip's needs */
- vsi->rx_hdr_len = ALIGN(vsi->rx_hdr_len,
- BIT_ULL(I40E_RXQ_CTX_HBUFF_SHIFT));
vsi->rx_buf_len = ALIGN(vsi->rx_buf_len,
BIT_ULL(I40E_RXQ_CTX_DBUFF_SHIFT));
@@ -10661,11 +10655,9 @@ static void i40e_print_features(struct i40e_pf *pf)
#ifdef CONFIG_PCI_IOV
i += snprintf(&buf[i], REMAIN(i), " VFs: %d", pf->num_req_vfs);
#endif
- i += snprintf(&buf[i], REMAIN(i), " VSIs: %d QP: %d RX: %s",
+ i += snprintf(&buf[i], REMAIN(i), " VSIs: %d QP: %d",
pf->hw.func_caps.num_vsis,
- pf->vsi[pf->lan_vsi]->num_queue_pairs,
- "1BUF");
-
+ pf->vsi[pf->lan_vsi]->num_queue_pairs);
if (pf->flags & I40E_FLAG_RSS_ENABLED)
i += snprintf(&buf[i], REMAIN(i), " RSS");
if (pf->flags & I40E_FLAG_FD_ATR_ENABLED)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 450ecdd..b0edffe 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -1024,7 +1024,6 @@ err:
void i40e_clean_rx_ring(struct i40e_ring *rx_ring)
{
struct device *dev = rx_ring->dev;
- struct i40e_rx_buffer *rx_bi;
unsigned long bi_size;
u16 i;
@@ -1034,30 +1033,20 @@ void i40e_clean_rx_ring(struct i40e_ring *rx_ring)
/* Free all the Rx ring sk_buffs */
for (i = 0; i < rx_ring->count; i++) {
- rx_bi = &rx_ring->rx_bi[i];
- if (rx_bi->dma) {
- dma_unmap_single(dev,
- rx_bi->dma,
- rx_ring->rx_buf_len,
- DMA_FROM_DEVICE);
- rx_bi->dma = 0;
- }
+ struct i40e_rx_buffer *rx_bi = &rx_ring->rx_bi[i];
+
if (rx_bi->skb) {
dev_kfree_skb(rx_bi->skb);
rx_bi->skb = NULL;
}
- if (rx_bi->page) {
- if (rx_bi->page_dma) {
- dma_unmap_page(dev,
- rx_bi->page_dma,
- PAGE_SIZE,
- DMA_FROM_DEVICE);
- rx_bi->page_dma = 0;
- }
- __free_page(rx_bi->page);
- rx_bi->page = NULL;
- rx_bi->page_offset = 0;
- }
+ if (!rx_bi->page)
+ continue;
+
+ dma_unmap_page(dev, rx_bi->dma, PAGE_SIZE, DMA_FROM_DEVICE);
+ __free_pages(rx_bi->page, 0);
+
+ rx_bi->page = NULL;
+ rx_bi->page_offset = 0;
}
bi_size = sizeof(struct i40e_rx_buffer) * rx_ring->count;
@@ -1066,6 +1055,7 @@ void i40e_clean_rx_ring(struct i40e_ring *rx_ring)
/* Zero out the descriptor ring */
memset(rx_ring->desc, 0, rx_ring->size);
+ rx_ring->next_to_alloc = 0;
rx_ring->next_to_clean = 0;
rx_ring->next_to_use = 0;
}
@@ -1090,37 +1080,6 @@ void i40e_free_rx_resources(struct i40e_ring *rx_ring)
}
/**
- * i40e_alloc_rx_headers - allocate rx header buffers
- * @rx_ring: ring to alloc buffers
- *
- * Allocate rx header buffers for the entire ring. As these are static,
- * this is only called when setting up a new ring.
- **/
-void i40e_alloc_rx_headers(struct i40e_ring *rx_ring)
-{
- struct device *dev = rx_ring->dev;
- struct i40e_rx_buffer *rx_bi;
- dma_addr_t dma;
- void *buffer;
- int buf_size;
- int i;
-
- if (rx_ring->rx_bi[0].hdr_buf)
- return;
- /* Make sure the buffers don't cross cache line boundaries. */
- buf_size = ALIGN(rx_ring->rx_hdr_len, 256);
- buffer = dma_alloc_coherent(dev, buf_size * rx_ring->count,
- &dma, GFP_KERNEL);
- if (!buffer)
- return;
- for (i = 0; i < rx_ring->count; i++) {
- rx_bi = &rx_ring->rx_bi[i];
- rx_bi->dma = dma + (i * buf_size);
- rx_bi->hdr_buf = buffer + (i * buf_size);
- }
-}
-
-/**
* i40e_setup_rx_descriptors - Allocate Rx descriptors
* @rx_ring: Rx descriptor ring (for a specific queue) to setup
*
@@ -1141,9 +1100,7 @@ int i40e_setup_rx_descriptors(struct i40e_ring *rx_ring)
u64_stats_init(&rx_ring->syncp);
/* Round up to nearest 4K */
- rx_ring->size = ring_is_16byte_desc_enabled(rx_ring)
- ? rx_ring->count * sizeof(union i40e_16byte_rx_desc)
- : rx_ring->count * sizeof(union i40e_32byte_rx_desc);
+ rx_ring->size = rx_ring->count * sizeof(union i40e_32byte_rx_desc);
rx_ring->size = ALIGN(rx_ring->size, 4096);
rx_ring->desc = dma_alloc_coherent(dev, rx_ring->size,
&rx_ring->dma, GFP_KERNEL);
@@ -1154,6 +1111,7 @@ int i40e_setup_rx_descriptors(struct i40e_ring *rx_ring)
goto err;
}
+ rx_ring->next_to_alloc = 0;
rx_ring->next_to_clean = 0;
rx_ring->next_to_use = 0;
@@ -1172,6 +1130,10 @@ err:
static inline void i40e_release_rx_desc(struct i40e_ring *rx_ring, u32 val)
{
rx_ring->next_to_use = val;
+
+ /* update next to alloc since we have filled the ring */
+ rx_ring->next_to_alloc = val;
+
/* Force memory writes to complete before letting h/w
* know there are new descriptors to fetch. (Only
* applicable for weak-ordered memory model archs,
@@ -1182,160 +1144,122 @@ static inline void i40e_release_rx_desc(struct i40e_ring *rx_ring, u32 val)
}
/**
- * i40e_alloc_rx_buffers_ps - Replace used receive buffers; packet split
- * @rx_ring: ring to place buffers on
- * @cleaned_count: number of buffers to replace
+ * i40e_alloc_mapped_page - recycle or make a new page
+ * @rx_ring: ring to use
+ * @bi: rx_buffer struct to modify
*
- * Returns true if any errors on allocation
+ * Returns true if the page was successfully allocated or
+ * reused.
**/
-bool i40e_alloc_rx_buffers_ps(struct i40e_ring *rx_ring, u16 cleaned_count)
+static bool i40e_alloc_mapped_page(struct i40e_ring *rx_ring,
+ struct i40e_rx_buffer *bi)
{
- u16 i = rx_ring->next_to_use;
- union i40e_rx_desc *rx_desc;
- struct i40e_rx_buffer *bi;
- const int current_node = numa_node_id();
+ struct page *page = bi->page;
+ dma_addr_t dma;
- /* do nothing if no valid netdev defined */
- if (!rx_ring->netdev || !cleaned_count)
- return false;
+ /* since we are recycling buffers we should seldom need to alloc */
+ if (likely(page)) {
+ rx_ring->rx_stats.page_reuse_count++;
+ return true;
+ }
- while (cleaned_count--) {
- rx_desc = I40E_RX_DESC(rx_ring, i);
- bi = &rx_ring->rx_bi[i];
+ /* alloc new page for storage */
+ page = dev_alloc_page();
+ if (unlikely(!page)) {
+ rx_ring->rx_stats.alloc_page_failed++;
+ return false;
+ }
- if (bi->skb) /* desc is in use */
- goto no_buffers;
+ /* map page for use */
+ dma = dma_map_page(rx_ring->dev, page, 0, PAGE_SIZE, DMA_FROM_DEVICE);
- /* If we've been moved to a different NUMA node, release the
- * page so we can get a new one on the current node.
+ /* if mapping failed free memory back to system since
+ * there isn't much point in holding memory we can't use
*/
- if (bi->page && page_to_nid(bi->page) != current_node) {
- dma_unmap_page(rx_ring->dev,
- bi->page_dma,
- PAGE_SIZE,
- DMA_FROM_DEVICE);
- __free_page(bi->page);
- bi->page = NULL;
- bi->page_dma = 0;
- rx_ring->rx_stats.realloc_count++;
- } else if (bi->page) {
- rx_ring->rx_stats.page_reuse_count++;
- }
-
- if (!bi->page) {
- bi->page = alloc_page(GFP_ATOMIC);
- if (!bi->page) {
- rx_ring->rx_stats.alloc_page_failed++;
- goto no_buffers;
- }
- bi->page_dma = dma_map_page(rx_ring->dev,
- bi->page,
- 0,
- PAGE_SIZE,
- DMA_FROM_DEVICE);
- if (dma_mapping_error(rx_ring->dev, bi->page_dma)) {
- rx_ring->rx_stats.alloc_page_failed++;
- __free_page(bi->page);
- bi->page = NULL;
- bi->page_dma = 0;
- bi->page_offset = 0;
- goto no_buffers;
- }
- bi->page_offset = 0;
- }
-
- /* Refresh the desc even if buffer_addrs didn't change
- * because each write-back erases this info.
- */
- rx_desc->read.pkt_addr =
- cpu_to_le64(bi->page_dma + bi->page_offset);
- rx_desc->read.hdr_addr = cpu_to_le64(bi->dma);
- i++;
- if (i == rx_ring->count)
- i = 0;
+ if (dma_mapping_error(rx_ring->dev, dma)) {
+ __free_pages(page, 0);
+ rx_ring->rx_stats.alloc_page_failed++;
+ return false;
}
- if (rx_ring->next_to_use != i)
- i40e_release_rx_desc(rx_ring, i);
+ bi->dma = dma;
+ bi->page = page;
+ bi->page_offset = 0;
- return false;
+ return true;
+}
-no_buffers:
- if (rx_ring->next_to_use != i)
- i40e_release_rx_desc(rx_ring, i);
+/**
+ * i40e_receive_skb - Send a completed packet up the stack
+ * @rx_ring: rx ring in play
+ * @skb: packet to send up
+ * @vlan_tag: vlan tag for packet
+ **/
+static void i40e_receive_skb(struct i40e_ring *rx_ring,
+ struct sk_buff *skb, u16 vlan_tag)
+{
+ struct i40e_q_vector *q_vector = rx_ring->q_vector;
- /* make sure to come back via polling to try again after
- * allocation failure
- */
- return true;
+ if ((rx_ring->netdev->features & NETIF_F_HW_VLAN_CTAG_RX) &&
+ (vlan_tag & VLAN_VID_MASK))
+ __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), vlan_tag);
+
+ napi_gro_receive(&q_vector->napi, skb);
}
/**
- * i40e_alloc_rx_buffers_1buf - Replace used receive buffers; single buffer
+ * i40e_alloc_rx_buffers - Replace used receive buffers
* @rx_ring: ring to place buffers on
* @cleaned_count: number of buffers to replace
*
- * Returns true if any errors on allocation
+ * Returns false if all allocations were successful, true if any fail
**/
-bool i40e_alloc_rx_buffers_1buf(struct i40e_ring *rx_ring, u16 cleaned_count)
+bool i40e_alloc_rx_buffers(struct i40e_ring *rx_ring, u16 cleaned_count)
{
- u16 i = rx_ring->next_to_use;
+ u16 ntu = rx_ring->next_to_use;
union i40e_rx_desc *rx_desc;
struct i40e_rx_buffer *bi;
- struct sk_buff *skb;
/* do nothing if no valid netdev defined */
if (!rx_ring->netdev || !cleaned_count)
return false;
- while (cleaned_count--) {
- rx_desc = I40E_RX_DESC(rx_ring, i);
- bi = &rx_ring->rx_bi[i];
- skb = bi->skb;
-
- if (!skb) {
- skb = __netdev_alloc_skb_ip_align(rx_ring->netdev,
- rx_ring->rx_buf_len,
- GFP_ATOMIC |
- __GFP_NOWARN);
- if (!skb) {
- rx_ring->rx_stats.alloc_buff_failed++;
- goto no_buffers;
- }
- /* initialize queue mapping */
- skb_record_rx_queue(skb, rx_ring->queue_index);
- bi->skb = skb;
- }
+ rx_desc = I40E_RX_DESC(rx_ring, ntu);
+ bi = &rx_ring->rx_bi[ntu];
- if (!bi->dma) {
- bi->dma = dma_map_single(rx_ring->dev,
- skb->data,
- rx_ring->rx_buf_len,
- DMA_FROM_DEVICE);
- if (dma_mapping_error(rx_ring->dev, bi->dma)) {
- rx_ring->rx_stats.alloc_buff_failed++;
- bi->dma = 0;
- dev_kfree_skb(bi->skb);
- bi->skb = NULL;
- goto no_buffers;
- }
- }
+ do {
+ if (!i40e_alloc_mapped_page(rx_ring, bi))
+ goto no_buffers;
- rx_desc->read.pkt_addr = cpu_to_le64(bi->dma);
+ /* Refresh the desc even if buffer_addrs didn't change
+ * because each write-back erases this info.
+ */
+ rx_desc->read.pkt_addr = cpu_to_le64(bi->dma + bi->page_offset);
rx_desc->read.hdr_addr = 0;
- i++;
- if (i == rx_ring->count)
- i = 0;
- }
- if (rx_ring->next_to_use != i)
- i40e_release_rx_desc(rx_ring, i);
+ rx_desc++;
+ bi++;
+ ntu++;
+ if (unlikely(ntu == rx_ring->count)) {
+ rx_desc = I40E_RX_DESC(rx_ring, 0);
+ bi = rx_ring->rx_bi;
+ ntu = 0;
+ }
+
+ /* clear the status bits for the next_to_use descriptor */
+ rx_desc->wb.qword1.status_error_len = 0;
+
+ cleaned_count--;
+ } while (cleaned_count);
+
+ if (rx_ring->next_to_use != ntu)
+ i40e_release_rx_desc(rx_ring, ntu);
return false;
no_buffers:
- if (rx_ring->next_to_use != i)
- i40e_release_rx_desc(rx_ring, i);
+ if (rx_ring->next_to_use != ntu)
+ i40e_release_rx_desc(rx_ring, ntu);
/* make sure to come back via polling to try again after
* allocation failure
@@ -1344,42 +1268,35 @@ no_buffers:
}
/**
- * i40e_receive_skb - Send a completed packet up the stack
- * @rx_ring: rx ring in play
- * @skb: packet to send up
- * @vlan_tag: vlan tag for packet
- **/
-static void i40e_receive_skb(struct i40e_ring *rx_ring,
- struct sk_buff *skb, u16 vlan_tag)
-{
- struct i40e_q_vector *q_vector = rx_ring->q_vector;
-
- if ((rx_ring->netdev->features & NETIF_F_HW_VLAN_CTAG_RX) &&
- (vlan_tag & VLAN_VID_MASK))
- __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), vlan_tag);
-
- napi_gro_receive(&q_vector->napi, skb);
-}
-
-/**
* i40e_rx_checksum - Indicate in skb if hw indicated a good cksum
* @vsi: the VSI we care about
* @skb: skb currently being received and modified
- * @rx_status: status value of last descriptor in packet
- * @rx_error: error value of last descriptor in packet
- * @rx_ptype: ptype value of last descriptor in packet
+ * @rx_desc: the receive descriptor
+ *
+ * skb->protocol must be set before this function is called
**/
static inline void i40e_rx_checksum(struct i40e_vsi *vsi,
struct sk_buff *skb,
- u32 rx_status,
- u32 rx_error,
- u16 rx_ptype)
+ union i40e_rx_desc *rx_desc)
{
- struct i40e_rx_ptype_decoded decoded = decode_rx_desc_ptype(rx_ptype);
+ struct i40e_rx_ptype_decoded decoded;
bool ipv4, ipv6, tunnel = false;
+ u32 rx_error, rx_status;
+ u8 ptype;
+ u64 qword;
+
+ qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
+ ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >> I40E_RXD_QW1_PTYPE_SHIFT;
+ rx_error = (qword & I40E_RXD_QW1_ERROR_MASK) >>
+ I40E_RXD_QW1_ERROR_SHIFT;
+ rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >>
+ I40E_RXD_QW1_STATUS_SHIFT;
+ decoded = decode_rx_desc_ptype(ptype);
skb->ip_summed = CHECKSUM_NONE;
+ skb_checksum_none_assert(skb);
+
/* Rx csum enabled and ip headers found? */
if (!(vsi->netdev->features & NETIF_F_RXCSUM))
return;
@@ -1445,7 +1362,7 @@ checksum_fail:
*
* Returns a hash type to be used by skb_set_hash
**/
-static inline enum pkt_hash_types i40e_ptype_to_htype(u8 ptype)
+static inline int i40e_ptype_to_htype(u8 ptype)
{
struct i40e_rx_ptype_decoded decoded = decode_rx_desc_ptype(ptype);
@@ -1473,7 +1390,7 @@ static inline void i40e_rx_hash(struct i40e_ring *ring,
u8 rx_ptype)
{
u32 hash;
- const __le64 rss_mask =
+ const __le64 rss_mask =
cpu_to_le64((u64)I40E_RX_DESC_FLTSTAT_RSS_HASH <<
I40E_RX_DESC_STATUS_FLTSTAT_SHIFT);
@@ -1487,114 +1404,419 @@ static inline void i40e_rx_hash(struct i40e_ring *ring,
}
/**
- * i40e_clean_rx_irq_1buf - Reclaim resources after receive; single buffer
- * @rx_ring: rx ring to clean
- * @budget: how many cleans we're allowed
+ * i40e_process_skb_fields - Populate skb header fields from Rx descriptor
+ * @rx_ring: rx descriptor ring packet is being transacted on
+ * @rx_desc: pointer to the EOP Rx descriptor
+ * @skb: pointer to current skb being populated
+ * @rx_ptype: the packet type decoded by hardware
*
- * Returns number of packets cleaned
+ * This function checks the ring, descriptor, and packet information in
+ * order to populate the hash, checksum, VLAN, protocol, and
+ * other fields within the skb.
**/
-static int i40e_clean_rx_irq_1buf(struct i40e_ring *rx_ring, int budget)
+static inline
+void i40e_process_skb_fields(struct i40e_ring *rx_ring,
+ union i40e_rx_desc *rx_desc, struct sk_buff *skb,
+ u8 rx_ptype)
+{
+ u64 qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
+ u32 rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >>
+ I40E_RXD_QW1_STATUS_SHIFT;
+ u32 rsyn = (rx_status & I40E_RXD_QW1_STATUS_TSYNINDX_MASK) >>
+ I40E_RXD_QW1_STATUS_TSYNINDX_SHIFT;
+
+ if (unlikely(rsyn)) {
+ i40e_ptp_rx_hwtstamp(rx_ring->vsi->back, skb, rsyn);
+ rx_ring->last_rx_timestamp = jiffies;
+ }
+
+ i40e_rx_hash(rx_ring, rx_desc, skb, rx_ptype);
+
+ /* modifies the skb - consumes the enet header */
+ skb->protocol = eth_type_trans(skb, rx_ring->netdev);
+
+ i40e_rx_checksum(rx_ring->vsi, skb, rx_desc);
+
+ skb_record_rx_queue(skb, rx_ring->queue_index);
+}
+
+/**
+ * i40e_pull_tail - i40e specific version of skb_pull_tail
+ * @rx_ring: rx descriptor ring packet is being transacted on
+ * @skb: pointer to current skb being adjusted
+ *
+ * This function is an i40e specific version of __pskb_pull_tail. The
+ * main difference between this version and the original function is that
+ * this function can make several assumptions about the state of things
+ * that allow for significant optimizations versus the standard function.
+ * As a result we can do things like drop a frag and maintain an accurate
+ * truesize for the skb.
+ */
+static void i40e_pull_tail(struct i40e_ring *rx_ring, struct sk_buff *skb)
+{
+ struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[0];
+ unsigned char *va;
+ unsigned int pull_len;
+
+ /* it is valid to use page_address instead of kmap since we are
+ * working with pages allocated out of the lomem pool per
+ * alloc_page(GFP_ATOMIC)
+ */
+ va = skb_frag_address(frag);
+
+ /* we need the header to contain the greater of either ETH_HLEN or
+ * 60 bytes if the skb->len is less than 60 for skb_pad.
+ */
+ pull_len = eth_get_headlen(va, I40E_RX_HDR_SIZE);
+
+ /* align pull length to size of long to optimize memcpy performance */
+ skb_copy_to_linear_data(skb, va, ALIGN(pull_len, sizeof(long)));
+
+ /* update all of the pointers */
+ skb_frag_size_sub(frag, pull_len);
+ frag->page_offset += pull_len;
+ skb->data_len -= pull_len;
+ skb->tail += pull_len;
+}
+
+/**
+ * i40e_cleanup_headers - Correct empty headers
+ * @rx_ring: rx descriptor ring packet is being transacted on
+ * @skb: pointer to current skb being fixed
+ *
+ * Also address the case where we are pulling data in on pages only
+ * and as such no data is present in the skb header.
+ *
+ * In addition if skb is not at least 60 bytes we need to pad it so that
+ * it is large enough to qualify as a valid Ethernet frame.
+ *
+ * Returns true if an error was encountered and skb was freed.
+ **/
+static bool i40e_cleanup_headers(struct i40e_ring *rx_ring, struct sk_buff *skb)
+{
+ /* place header in linear portion of buffer */
+ if (skb_is_nonlinear(skb))
+ i40e_pull_tail(rx_ring, skb);
+
+ /* if eth_skb_pad returns an error the skb was freed */
+ if (eth_skb_pad(skb))
+ return true;
+
+ return false;
+}
+
+/**
+ * i40e_reuse_rx_page - page flip buffer and store it back on the ring
+ * @rx_ring: rx descriptor ring to store buffers on
+ * @old_buff: donor buffer to have page reused
+ *
+ * Synchronizes page for reuse by the adapter
+ **/
+static void i40e_reuse_rx_page(struct i40e_ring *rx_ring,
+ struct i40e_rx_buffer *old_buff)
+{
+ struct i40e_rx_buffer *new_buff;
+ u16 nta = rx_ring->next_to_alloc;
+
+ new_buff = &rx_ring->rx_bi[nta];
+
+ /* update, and store next to alloc */
+ nta++;
+ rx_ring->next_to_alloc = (nta < rx_ring->count) ? nta : 0;
+
+ /* transfer page from old buffer to new buffer */
+ *new_buff = *old_buff;
+}
+
+/**
+ * i40e_page_is_reserved - check if reuse is possible
+ * @page: page struct to check
+ */
+static inline bool i40e_page_is_reserved(struct page *page)
+{
+ return (page_to_nid(page) != numa_mem_id()) || page_is_pfmemalloc(page);
+}
+
+/**
+ * i40e_add_rx_frag - Add contents of Rx buffer to sk_buff
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @rx_buffer: buffer containing page to add
+ * @rx_desc: descriptor containing length of buffer written by hardware
+ * @skb: sk_buff to place the data into
+ *
+ * This function will add the data contained in rx_buffer->page to the skb.
+ * This is done either through a direct copy if the data in the buffer is
+ * less than the skb header size, otherwise it will just attach the page as
+ * a frag to the skb.
+ *
+ * The function will then update the page offset if necessary and return
+ * true if the buffer can be reused by the adapter.
+ **/
+static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
+ struct i40e_rx_buffer *rx_buffer,
+ union i40e_rx_desc *rx_desc,
+ struct sk_buff *skb)
+{
+ struct page *page = rx_buffer->page;
+ u64 qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
+ unsigned int size = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
+ I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
+#if (PAGE_SIZE < 8192)
+ unsigned int truesize = I40E_RXBUFFER_2048;
+#else
+ unsigned int truesize = ALIGN(size, L1_CACHE_BYTES);
+ unsigned int last_offset = PAGE_SIZE - I40E_RXBUFFER_2048;
+#endif
+
+ /* will the data fit in the skb we allocated? if so, just
+ * copy it as it is pretty small anyway
+ */
+ if ((size <= I40E_RX_HDR_SIZE) && !skb_is_nonlinear(skb)) {
+ unsigned char *va = page_address(page) + rx_buffer->page_offset;
+
+ memcpy(__skb_put(skb, size), va, ALIGN(size, sizeof(long)));
+
+ /* page is not reserved, we can reuse buffer as-is */
+ if (likely(!i40e_page_is_reserved(page)))
+ return true;
+
+ /* this page cannot be reused so discard it */
+ __free_pages(page, 0);
+ return false;
+ }
+
+ skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, page,
+ rx_buffer->page_offset, size, truesize);
+
+ /* avoid re-using remote pages */
+ if (unlikely(i40e_page_is_reserved(page)))
+ return false;
+
+#if (PAGE_SIZE < 8192)
+ /* if we are only owner of page we can reuse it */
+ if (unlikely(page_count(page) != 1))
+ return false;
+
+ /* flip page offset to other buffer */
+ rx_buffer->page_offset ^= truesize;
+#else
+ /* move offset up to the next cache line */
+ rx_buffer->page_offset += truesize;
+
+ if (rx_buffer->page_offset > last_offset)
+ return false;
+#endif
+
+ /* Even if we own the page, we are not allowed to use atomic_set()
+ * This would break get_page_unless_zero() users.
+ */
+ get_page(rx_buffer->page);
+
+ return true;
+}
+
+/**
+ * i40e_fetch_rx_buffer - Allocate skb and populate it
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @rx_desc: descriptor containing info written by hardware
+ *
+ * This function allocates an skb on the fly, and populates it with the page
+ * data from the current receive descriptor, taking care to set up the skb
+ * correctly, as well as handling calling the page recycle function if
+ * necessary.
+ */
+static inline
+struct sk_buff *i40e_fetch_rx_buffer(struct i40e_ring *rx_ring,
+ union i40e_rx_desc *rx_desc)
+{
+ struct i40e_rx_buffer *rx_buffer;
+ struct sk_buff *skb;
+ struct page *page;
+
+ rx_buffer = &rx_ring->rx_bi[rx_ring->next_to_clean];
+ page = rx_buffer->page;
+ prefetchw(page);
+
+ skb = rx_buffer->skb;
+
+ if (likely(!skb)) {
+ void *page_addr = page_address(page) + rx_buffer->page_offset;
+
+ /* prefetch first cache line of first page */
+ prefetch(page_addr);
+#if L1_CACHE_BYTES < 128
+ prefetch(page_addr + L1_CACHE_BYTES);
+#endif
+
+ /* allocate a skb to store the frags */
+ skb = __napi_alloc_skb(&rx_ring->q_vector->napi,
+ I40E_RX_HDR_SIZE,
+ GFP_ATOMIC | __GFP_NOWARN);
+ if (unlikely(!skb)) {
+ rx_ring->rx_stats.alloc_buff_failed++;
+ return NULL;
+ }
+
+ /* we will be copying header into skb->data in
+ * pskb_may_pull so it is in our interest to prefetch
+ * it now to avoid a possible cache miss
+ */
+ prefetchw(skb->data);
+ } else {
+ rx_buffer->skb = NULL;
+ }
+
+ /* we are reusing so sync this buffer for CPU use */
+ dma_sync_single_range_for_cpu(rx_ring->dev,
+ rx_buffer->dma,
+ rx_buffer->page_offset,
+ I40E_RXBUFFER_2048,
+ DMA_FROM_DEVICE);
+
+ /* pull page into skb */
+ if (i40e_add_rx_frag(rx_ring, rx_buffer, rx_desc, skb)) {
+ /* hand second half of page back to the ring */
+ i40e_reuse_rx_page(rx_ring, rx_buffer);
+ rx_ring->rx_stats.page_reuse_count++;
+ } else {
+ /* we are not reusing the buffer so unmap it */
+ dma_unmap_page(rx_ring->dev, rx_buffer->dma, PAGE_SIZE,
+ DMA_FROM_DEVICE);
+ }
+
+ /* clear contents of buffer_info */
+ rx_buffer->page = NULL;
+
+ return skb;
+}
+
+/**
+ * i40e_is_non_eop - process handling of non-EOP buffers
+ * @rx_ring: Rx ring being processed
+ * @rx_desc: Rx descriptor for current buffer
+ * @skb: Current socket buffer containing buffer in progress
+ *
+ * This function updates next to clean. If the buffer is an EOP buffer
+ * this function exits returning false, otherwise it will place the
+ * sk_buff in the next buffer to be chained and return true indicating
+ * that this is in fact a non-EOP buffer.
+ **/
+static bool i40e_is_non_eop(struct i40e_ring *rx_ring,
+ union i40e_rx_desc *rx_desc,
+ struct sk_buff *skb)
+{
+ u32 ntc = rx_ring->next_to_clean + 1;
+
+ /* fetch, update, and store next to clean */
+ ntc = (ntc < rx_ring->count) ? ntc : 0;
+ rx_ring->next_to_clean = ntc;
+
+ prefetch(I40E_RX_DESC(rx_ring, ntc));
+
+#define staterrlen rx_desc->wb.qword1.status_error_len
+ if (unlikely(i40e_rx_is_programming_status(le64_to_cpu(staterrlen)))) {
+ i40e_clean_programming_status(rx_ring, rx_desc);
+ rx_ring->rx_bi[ntc].skb = skb;
+ return true;
+ }
+ /* if we are the last buffer then there is nothing else to do */
+#define I40E_RXD_EOF BIT(I40E_RX_DESC_STATUS_EOF_SHIFT)
+ if (likely(i40e_test_staterr(rx_desc, I40E_RXD_EOF)))
+ return false;
+
+ /* place skb in next buffer to be received */
+ rx_ring->rx_bi[ntc].skb = skb;
+ rx_ring->rx_stats.non_eop_descs++;
+
+ return true;
+}
+
+/**
+ * i40e_clean_rx_irq - Clean completed descriptors from Rx ring - bounce buf
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @budget: Total limit on number of packets to process
+ *
+ * This function provides a "bounce buffer" approach to Rx interrupt
+ * processing. The advantage to this is that on systems that have
+ * expensive overhead for IOMMU access this provides a means of avoiding
+ * it by maintaining the mapping of the page to the system.
+ *
+ * Returns amount of work completed
+ **/
+static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
{
unsigned int total_rx_bytes = 0, total_rx_packets = 0;
u16 cleaned_count = I40E_DESC_UNUSED(rx_ring);
- struct i40e_vsi *vsi = rx_ring->vsi;
- union i40e_rx_desc *rx_desc;
- u32 rx_error, rx_status;
- u16 rx_packet_len;
bool failure = false;
- u8 rx_ptype;
- u64 qword;
- u16 i;
- do {
- struct i40e_rx_buffer *rx_bi;
+ while (likely(total_rx_packets < budget)) {
+ union i40e_rx_desc *rx_desc;
struct sk_buff *skb;
+ u32 rx_status;
u16 vlan_tag;
+ u8 rx_ptype;
+ u64 qword;
+
/* return some buffers to hardware, one at a time is too slow */
if (cleaned_count >= I40E_RX_BUFFER_WRITE) {
failure = failure ||
- i40e_alloc_rx_buffers_1buf(rx_ring,
- cleaned_count);
+ i40e_alloc_rx_buffers(rx_ring, cleaned_count);
cleaned_count = 0;
}
- i = rx_ring->next_to_clean;
- rx_desc = I40E_RX_DESC(rx_ring, i);
+ rx_desc = I40E_RX_DESC(rx_ring, rx_ring->next_to_clean);
+
qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
+ rx_ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >>
+ I40E_RXD_QW1_PTYPE_SHIFT;
rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >>
- I40E_RXD_QW1_STATUS_SHIFT;
+ I40E_RXD_QW1_STATUS_SHIFT;
if (!(rx_status & BIT(I40E_RX_DESC_STATUS_DD_SHIFT)))
break;
+ /* status_error_len will always be zero for unused descriptors
+ * because it's cleared in cleanup, and overlaps with hdr_addr
+ * which is always zero because packet split isn't used, if the
+ * hardware wrote DD then it will be non-zero
+ */
+ if (!rx_desc->wb.qword1.status_error_len)
+ break;
+
/* This memory barrier is needed to keep us from reading
* any other fields out of the rx_desc until we know the
* DD bit is set.
*/
dma_rmb();
- if (i40e_rx_is_programming_status(qword)) {
- i40e_clean_programming_status(rx_ring, rx_desc);
- I40E_RX_INCREMENT(rx_ring, i);
- continue;
- }
- rx_bi = &rx_ring->rx_bi[i];
- skb = rx_bi->skb;
- prefetch(skb->data);
-
- rx_packet_len = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
- I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
-
- rx_error = (qword & I40E_RXD_QW1_ERROR_MASK) >>
- I40E_RXD_QW1_ERROR_SHIFT;
- rx_error &= ~BIT(I40E_RX_DESC_ERROR_HBO_SHIFT);
+ skb = i40e_fetch_rx_buffer(rx_ring, rx_desc);
+ if (!skb)
+ break;
- rx_ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >>
- I40E_RXD_QW1_PTYPE_SHIFT;
- rx_bi->skb = NULL;
cleaned_count++;
- /* Get the header and possibly the whole packet
- * If this is an skb from previous receive dma will be 0
- */
- skb_put(skb, rx_packet_len);
- dma_unmap_single(rx_ring->dev, rx_bi->dma, rx_ring->rx_buf_len,
- DMA_FROM_DEVICE);
- rx_bi->dma = 0;
-
- I40E_RX_INCREMENT(rx_ring, i);
-
- if (unlikely(
- !(rx_status & BIT(I40E_RX_DESC_STATUS_EOF_SHIFT)))) {
- rx_ring->rx_stats.non_eop_descs++;
+ if (i40e_is_non_eop(rx_ring, rx_desc, skb))
continue;
- }
- /* ERR_MASK will only have valid bits if EOP set */
- if (unlikely(rx_error & BIT(I40E_RX_DESC_ERROR_RXE_SHIFT))) {
+ /* ERR_MASK will only have valid bits if EOP set, and
+ * what we are doing here is actually checking
+ * I40E_RX_DESC_ERROR_RXE_SHIFT, since it is the zeroth bit in
+ * the error field
+ */
+ if (unlikely(i40e_test_staterr(rx_desc, BIT(I40E_RXD_QW1_ERROR_SHIFT)))) {
dev_kfree_skb_any(skb);
continue;
}
- i40e_rx_hash(rx_ring, rx_desc, skb, rx_ptype);
- if (unlikely(rx_status & I40E_RXD_QW1_STATUS_TSYNVALID_MASK)) {
- i40e_ptp_rx_hwtstamp(vsi->back, skb, (rx_status &
- I40E_RXD_QW1_STATUS_TSYNINDX_MASK) >>
- I40E_RXD_QW1_STATUS_TSYNINDX_SHIFT);
- rx_ring->last_rx_timestamp = jiffies;
- }
+ if (i40e_cleanup_headers(rx_ring, skb))
+ continue;
/* probably a little skewed due to removing CRC */
total_rx_bytes += skb->len;
- total_rx_packets++;
- skb->protocol = eth_type_trans(skb, rx_ring->netdev);
+ /* populate checksum, VLAN, and protocol */
+ i40e_process_skb_fields(rx_ring, rx_desc, skb, rx_ptype);
- i40e_rx_checksum(vsi, skb, rx_status, rx_error, rx_ptype);
-
- vlan_tag = rx_status & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)
- ? le16_to_cpu(rx_desc->wb.qword0.lo_dword.l2tag1)
- : 0;
#ifdef I40E_FCOE
if (unlikely(
i40e_rx_is_fcoe(rx_ptype) &&
@@ -1603,10 +1825,15 @@ static int i40e_clean_rx_irq_1buf(struct i40e_ring *rx_ring, int budget)
continue;
}
#endif
+
+ vlan_tag = (qword & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)) ?
+ le16_to_cpu(rx_desc->wb.qword0.lo_dword.l2tag1) : 0;
+
i40e_receive_skb(rx_ring, skb, vlan_tag);
- rx_desc->wb.qword1.status_error_len = 0;
- } while (likely(total_rx_packets < budget));
+ /* update budget accounting */
+ total_rx_packets++;
+ }
u64_stats_update_begin(&rx_ring->syncp);
rx_ring->stats.packets += total_rx_packets;
@@ -1615,6 +1842,7 @@ static int i40e_clean_rx_irq_1buf(struct i40e_ring *rx_ring, int budget)
rx_ring->q_vector->rx.total_packets += total_rx_packets;
rx_ring->q_vector->rx.total_bytes += total_rx_bytes;
+ /* guarantee a trip back through this routine if there was a failure */
return failure ? budget : total_rx_packets;
}
@@ -1759,9 +1987,7 @@ int i40e_napi_poll(struct napi_struct *napi, int budget)
budget_per_ring = max(budget/q_vector->num_ringpairs, 1);
i40e_for_each_ring(ring, q_vector->rx) {
- int cleaned;
-
- cleaned = i40e_clean_rx_irq_1buf(ring, budget_per_ring);
+ int cleaned = i40e_clean_rx_irq(ring, budget_per_ring);
work_done += cleaned;
/* if we clean as many as budgeted, we must not be done */
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.h b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
index 03e21d9..37643e6 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
@@ -102,8 +102,8 @@ enum i40e_dyn_idx_t {
(((pf)->flags & I40E_FLAG_MULTIPLE_TCP_UDP_RSS_PCTYPE) ? \
I40E_DEFAULT_RSS_HENA_EXPANDED : I40E_DEFAULT_RSS_HENA)
-/* Supported Rx Buffer Sizes */
-#define I40E_RXBUFFER_512 512 /* Used for packet split */
+/* Supported Rx Buffer Sizes (a multiple of 128) */
+#define I40E_RXBUFFER_256 256
#define I40E_RXBUFFER_2048 2048
#define I40E_RXBUFFER_3072 3072 /* For FCoE MTU of 2158 */
#define I40E_RXBUFFER_4096 4096
@@ -114,9 +114,28 @@ enum i40e_dyn_idx_t {
* reserve 2 more, and skb_shared_info adds an additional 384 bytes more,
* this adds up to 512 bytes of extra data meaning the smallest allocation
* we could have is 1K.
- * i.e. RXBUFFER_512 --> size-1024 slab
+ * i.e. RXBUFFER_256 --> 960 byte skb (size-1024 slab)
+ * i.e. RXBUFFER_512 --> 1216 byte skb (size-2048 slab)
*/
-#define I40E_RX_HDR_SIZE I40E_RXBUFFER_512
+#define I40E_RX_HDR_SIZE I40E_RXBUFFER_256
+#define i40e_rx_desc i40e_32byte_rx_desc
+
+/**
+ * i40e_test_staterr - tests bits in Rx descriptor status and error fields
+ * @rx_desc: pointer to receive descriptor (in le64 format)
+ * @stat_err_bits: value to mask
+ *
+ * This function does some fast chicanery in order to return the
+ * value of the mask which is really only used for boolean tests.
+ * The status_error_len doesn't need to be shifted because it begins
+ * at offset zero.
+ */
+static inline bool i40e_test_staterr(union i40e_rx_desc *rx_desc,
+ const u64 stat_err_bits)
+{
+ return !!(rx_desc->wb.qword1.status_error_len &
+ cpu_to_le64(stat_err_bits));
+}
/* How many Rx Buffers do we bundle into one write to the hardware ? */
#define I40E_RX_BUFFER_WRITE 16 /* Must be power of 2 */
@@ -142,8 +161,6 @@ enum i40e_dyn_idx_t {
prefetch((n)); \
} while (0)
-#define i40e_rx_desc i40e_32byte_rx_desc
-
#define I40E_MAX_BUFFER_TXD 8
#define I40E_MIN_TX_LEN 17
@@ -213,10 +230,8 @@ struct i40e_tx_buffer {
struct i40e_rx_buffer {
struct sk_buff *skb;
- void *hdr_buf;
dma_addr_t dma;
struct page *page;
- dma_addr_t page_dma;
unsigned int page_offset;
};
@@ -280,7 +295,6 @@ struct i40e_ring {
u16 count; /* Number of descriptors */
u16 reg_idx; /* HW register index of the ring */
- u16 rx_hdr_len;
u16 rx_buf_len;
#define I40E_RX_DTYPE_NO_SPLIT 0
#define I40E_RX_DTYPE_HEADER_SPLIT 1
@@ -322,6 +336,7 @@ struct i40e_ring {
struct i40e_q_vector *q_vector; /* Backreference to associated vector */
struct rcu_head rcu; /* to avoid race on free */
+ u16 next_to_alloc;
} ____cacheline_internodealigned_in_smp;
enum i40e_latency_range {
@@ -345,9 +360,7 @@ struct i40e_ring_container {
#define i40e_for_each_ring(pos, head) \
for (pos = (head).ring; pos != NULL; pos = pos->next)
-bool i40e_alloc_rx_buffers_ps(struct i40e_ring *rxr, u16 cleaned_count);
-bool i40e_alloc_rx_buffers_1buf(struct i40e_ring *rxr, u16 cleaned_count);
-void i40e_alloc_rx_headers(struct i40e_ring *rxr);
+bool i40e_alloc_rx_buffers(struct i40e_ring *rxr, u16 cleaned_count);
netdev_tx_t i40e_lan_xmit_frame(struct sk_buff *skb, struct net_device *netdev);
void i40e_clean_tx_ring(struct i40e_ring *tx_ring);
void i40e_clean_rx_ring(struct i40e_ring *rx_ring);
--
2.5.5
^ permalink raw reply related
* [net-next 08/11] i40evf: Allocate Rx buffers properly
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem; +Cc: Mitch Williams, netdev, nhorman, sassmann, jogreene, Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Mitch Williams <mitch.a.williams@intel.com>
Allocate the correct number of RX buffers, and don't fiddle with
next_to_use. The common RX code handles all of this. This fixes a memory
leak of one page each time the driver is opened.
Change-Id: Id06eca353086e084921f047acad28c14745684ee
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40evf/i40evf_main.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 870bad8..b548dbe 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -990,9 +990,7 @@ static void i40evf_configure(struct i40evf_adapter *adapter)
for (i = 0; i < adapter->num_active_queues; i++) {
struct i40e_ring *ring = &adapter->rx_rings[i];
- i40evf_alloc_rx_buffers(ring, ring->count);
- ring->next_to_use = ring->count - 1;
- writel(ring->next_to_use, ring->tail);
+ i40evf_alloc_rx_buffers(ring, I40E_DESC_UNUSED(ring));
}
}
@@ -2768,7 +2766,6 @@ static void i40evf_remove(struct pci_dev *pdev)
iounmap(hw->hw_addr);
pci_release_regions(pdev);
-
i40evf_free_all_tx_resources(adapter);
i40evf_free_all_rx_resources(adapter);
i40evf_free_queues(adapter);
--
2.5.5
^ permalink raw reply related
* [net-next 05/11] i40evf: Drop packet split receive routine
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem; +Cc: Jesse Brandeburg, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
As part of preparation for the rx-refactor, remove the
packet split receive routine and ancillary code.
Some of the split related context set up code stays in
i40e_virtchnl_pf.c in case an older VF driver tries to load
and still wants to use packet split.
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@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_virtchnl_pf.c | 2 +-
drivers/net/ethernet/intel/i40evf/i40e_txrx.c | 21 +---------------
drivers/net/ethernet/intel/i40evf/i40e_txrx.h | 7 ------
drivers/net/ethernet/intel/i40evf/i40evf.h | 2 --
drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c | 14 -----------
drivers/net/ethernet/intel/i40evf/i40evf_main.c | 28 +---------------------
.../net/ethernet/intel/i40evf/i40evf_virtchnl.c | 4 ----
7 files changed, 3 insertions(+), 75 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 6b9db79..36aa33a 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -590,7 +590,7 @@ static int i40e_config_vsi_rx_queue(struct i40e_vf *vf, u16 vsi_id,
}
rx_ctx.hbuff = info->hdr_size >> I40E_RXQ_CTX_HBUFF_SHIFT;
- /* set splitalways mode 10b */
+ /* set split mode 10b */
rx_ctx.dtype = I40E_RX_DTYPE_HEADER_SPLIT;
}
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index a37a3f3..61d4a7a 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -504,22 +504,6 @@ void i40evf_clean_rx_ring(struct i40e_ring *rx_ring)
if (!rx_ring->rx_bi)
return;
- if (ring_is_ps_enabled(rx_ring)) {
- int bufsz = ALIGN(rx_ring->rx_hdr_len, 256) * rx_ring->count;
-
- rx_bi = &rx_ring->rx_bi[0];
- if (rx_bi->hdr_buf) {
- dma_free_coherent(dev,
- bufsz,
- rx_bi->hdr_buf,
- rx_bi->dma);
- for (i = 0; i < rx_ring->count; i++) {
- rx_bi = &rx_ring->rx_bi[i];
- rx_bi->dma = 0;
- rx_bi->hdr_buf = NULL;
- }
- }
- }
/* Free all the Rx ring sk_buffs */
for (i = 0; i < rx_ring->count; i++) {
rx_bi = &rx_ring->rx_bi[i];
@@ -1435,10 +1419,7 @@ int i40evf_napi_poll(struct napi_struct *napi, int budget)
i40e_for_each_ring(ring, q_vector->rx) {
int cleaned;
- if (ring_is_ps_enabled(ring))
- cleaned = i40e_clean_rx_irq_ps(ring, budget_per_ring);
- else
- cleaned = i40e_clean_rx_irq_1buf(ring, budget_per_ring);
+ cleaned = i40e_clean_rx_irq_1buf(ring, budget_per_ring);
work_done += cleaned;
/* if we clean as many as budgeted, we must not be done */
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
index 3b3f976..f24a97e 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
@@ -244,16 +244,9 @@ struct i40e_rx_queue_stats {
enum i40e_ring_state_t {
__I40E_TX_FDIR_INIT_DONE,
__I40E_TX_XPS_INIT_DONE,
- __I40E_RX_PS_ENABLED,
__I40E_RX_16BYTE_DESC_ENABLED,
};
-#define ring_is_ps_enabled(ring) \
- test_bit(__I40E_RX_PS_ENABLED, &(ring)->state)
-#define set_ring_ps_enabled(ring) \
- set_bit(__I40E_RX_PS_ENABLED, &(ring)->state)
-#define clear_ring_ps_enabled(ring) \
- clear_bit(__I40E_RX_PS_ENABLED, &(ring)->state)
#define ring_is_16byte_desc_enabled(ring) \
test_bit(__I40E_RX_16BYTE_DESC_ENABLED, &(ring)->state)
#define set_ring_16byte_desc_enabled(ring) \
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf.h b/drivers/net/ethernet/intel/i40evf/i40evf.h
index 25afabf..83ccc58 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf.h
+++ b/drivers/net/ethernet/intel/i40evf/i40evf.h
@@ -209,8 +209,6 @@ struct i40evf_adapter {
u32 flags;
#define I40EVF_FLAG_RX_CSUM_ENABLED BIT(0)
#define I40EVF_FLAG_RX_1BUF_CAPABLE BIT(1)
-#define I40EVF_FLAG_RX_PS_CAPABLE BIT(2)
-#define I40EVF_FLAG_RX_PS_ENABLED BIT(3)
#define I40EVF_FLAG_IMIR_ENABLED BIT(5)
#define I40EVF_FLAG_MQ_CAPABLE BIT(6)
#define I40EVF_FLAG_NEED_LINK_UPDATE BIT(7)
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
index 5a48ee0..e972ebc 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
@@ -527,12 +527,8 @@ static int i40evf_set_rxfh(struct net_device *netdev, const u32 *indir,
**/
static u32 i40evf_get_priv_flags(struct net_device *dev)
{
- struct i40evf_adapter *adapter = netdev_priv(dev);
u32 ret_flags = 0;
- ret_flags |= adapter->flags & I40EVF_FLAG_RX_PS_ENABLED ?
- I40EVF_PRIV_FLAGS_PS : 0;
-
return ret_flags;
}
@@ -546,16 +542,6 @@ static int i40evf_set_priv_flags(struct net_device *dev, u32 flags)
struct i40evf_adapter *adapter = netdev_priv(dev);
bool reset_required = false;
- if ((flags & I40EVF_PRIV_FLAGS_PS) &&
- !(adapter->flags & I40EVF_FLAG_RX_PS_ENABLED)) {
- adapter->flags |= I40EVF_FLAG_RX_PS_ENABLED;
- reset_required = true;
- } else if (!(flags & I40EVF_PRIV_FLAGS_PS) &&
- (adapter->flags & I40EVF_FLAG_RX_PS_ENABLED)) {
- adapter->flags &= ~I40EVF_FLAG_RX_PS_ENABLED;
- reset_required = true;
- }
-
/* if needed, issue reset to cause things to take effect */
if (reset_required)
i40evf_schedule_reset(adapter);
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 9f0bd7a..52408bc 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -641,28 +641,11 @@ static void i40evf_configure_tx(struct i40evf_adapter *adapter)
static void i40evf_configure_rx(struct i40evf_adapter *adapter)
{
struct i40e_hw *hw = &adapter->hw;
- struct net_device *netdev = adapter->netdev;
- int max_frame = netdev->mtu + ETH_HLEN + ETH_FCS_LEN;
int i;
- int rx_buf_len;
-
-
- /* Set the RX buffer length according to the mode */
- if (adapter->flags & I40EVF_FLAG_RX_PS_ENABLED ||
- netdev->mtu <= ETH_DATA_LEN)
- rx_buf_len = I40EVF_RXBUFFER_2048;
- else
- rx_buf_len = ALIGN(max_frame, 1024);
for (i = 0; i < adapter->num_active_queues; i++) {
adapter->rx_rings[i].tail = hw->hw_addr + I40E_QRX_TAIL1(i);
- adapter->rx_rings[i].rx_buf_len = rx_buf_len;
- if (adapter->flags & I40EVF_FLAG_RX_PS_ENABLED) {
- set_ring_ps_enabled(&adapter->rx_rings[i]);
- adapter->rx_rings[i].rx_hdr_len = I40E_RX_HDR_SIZE;
- } else {
- clear_ring_ps_enabled(&adapter->rx_rings[i]);
- }
+ adapter->rx_rings[i].rx_buf_len = I40EVF_RXBUFFER_2048;
}
}
@@ -1007,12 +990,7 @@ static void i40evf_configure(struct i40evf_adapter *adapter)
for (i = 0; i < adapter->num_active_queues; i++) {
struct i40e_ring *ring = &adapter->rx_rings[i];
- if (adapter->flags & I40EVF_FLAG_RX_PS_ENABLED) {
- i40evf_alloc_rx_headers(ring);
- i40evf_alloc_rx_buffers_ps(ring, ring->count);
- } else {
i40evf_alloc_rx_buffers_1buf(ring, ring->count);
- }
ring->next_to_use = ring->count - 1;
writel(ring->next_to_use, ring->tail);
}
@@ -2424,10 +2402,6 @@ static void i40evf_init_task(struct work_struct *work)
adapter->flags |= I40EVF_FLAG_RX_CSUM_ENABLED;
adapter->flags |= I40EVF_FLAG_RX_1BUF_CAPABLE;
- adapter->flags |= I40EVF_FLAG_RX_PS_CAPABLE;
-
- /* Default to single buffer rx, can be changed through ethtool. */
- adapter->flags &= ~I40EVF_FLAG_RX_PS_ENABLED;
netdev->netdev_ops = &i40evf_netdev_ops;
i40evf_set_ethtool_ops(netdev);
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c b/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c
index ba7fbc0..c5d33a2 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c
@@ -270,10 +270,6 @@ void i40evf_configure_queues(struct i40evf_adapter *adapter)
vqpi->rxq.max_pkt_size = adapter->netdev->mtu
+ ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN;
vqpi->rxq.databuffer_size = adapter->rx_rings[i].rx_buf_len;
- if (adapter->flags & I40EVF_FLAG_RX_PS_ENABLED) {
- vqpi->rxq.splithdr_enabled = true;
- vqpi->rxq.hdr_size = I40E_RX_HDR_SIZE;
- }
vqpi++;
}
--
2.5.5
^ permalink raw reply related
* [net-next 07/11] i40e/i40evf: Remove unused hardware receive descriptor code
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem; +Cc: Jesse Brandeburg, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
The hardware supports a 16 byte descriptor for receive, but the
driver was never using it in production. There was no performance
benefit to the real driver of 16 byte descriptors, so drop a whole
lot of complexity while getting rid of the code.
Also since the previous patch made us use no-split mode all the
time, drop any support in the driver for any other value in dtype
and assume it is always zero (aka no-split).
Hooray for code removal!
Change-ID: I2257e902e4dad84a07b94db6d2e6f4ce69b27bc0
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@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.h | 7 +------
drivers/net/ethernet/intel/i40e/i40e_debugfs.c | 16 +---------------
drivers/net/ethernet/intel/i40e/i40e_main.c | 18 +++++-------------
drivers/net/ethernet/intel/i40e/i40e_txrx.h | 24 ++++++++++--------------
drivers/net/ethernet/intel/i40evf/i40e_txrx.h | 24 ++++++++++--------------
5 files changed, 27 insertions(+), 62 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index ebf423b..2a6a5d3 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -122,10 +122,7 @@
#define XSTRINGIFY(bar) STRINGIFY(bar)
#define I40E_RX_DESC(R, i) \
- ((ring_is_16byte_desc_enabled(R)) \
- ? (union i40e_32byte_rx_desc *) \
- (&(((union i40e_16byte_rx_desc *)((R)->desc))[i])) \
- : (&(((union i40e_32byte_rx_desc *)((R)->desc))[i])))
+ (&(((union i40e_32byte_rx_desc *)((R)->desc))[i]))
#define I40E_TX_DESC(R, i) \
(&(((struct i40e_tx_desc *)((R)->desc))[i]))
#define I40E_TX_CTXTDESC(R, i) \
@@ -327,7 +324,6 @@ struct i40e_pf {
#ifdef I40E_FCOE
#define I40E_FLAG_FCOE_ENABLED BIT_ULL(11)
#endif /* I40E_FCOE */
-#define I40E_FLAG_16BYTE_RX_DESC_ENABLED BIT_ULL(13)
#define I40E_FLAG_CLEAN_ADMINQ BIT_ULL(14)
#define I40E_FLAG_FILTER_SYNC BIT_ULL(15)
#define I40E_FLAG_SERVICE_CLIENT_REQUESTED BIT_ULL(16)
@@ -532,7 +528,6 @@ struct i40e_vsi {
u16 max_frame;
u16 rx_buf_len;
- u8 dtype;
/* List of q_vectors allocated to this VSI */
struct i40e_q_vector **q_vectors;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
index 8ae30f7..e6af8c8 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
@@ -361,7 +361,7 @@ static void i40e_dbg_dump_vsi_seid(struct i40e_pf *pf, int seid)
vsi->work_limit);
dev_info(&pf->pdev->dev,
" max_frame = %d, rx_buf_len = %d dtype = %d\n",
- vsi->max_frame, vsi->rx_buf_len, vsi->dtype);
+ vsi->max_frame, vsi->rx_buf_len, 0);
dev_info(&pf->pdev->dev,
" num_q_vectors = %i, base_vector = %i\n",
vsi->num_q_vectors, vsi->base_vector);
@@ -586,13 +586,6 @@ static void i40e_dbg_dump_desc(int cnt, int vsi_seid, int ring_id, int desc_n,
" d[%03x] = 0x%016llx 0x%016llx\n",
i, txd->buffer_addr,
txd->cmd_type_offset_bsz);
- } else if (sizeof(union i40e_rx_desc) ==
- sizeof(union i40e_16byte_rx_desc)) {
- rxd = I40E_RX_DESC(ring, i);
- dev_info(&pf->pdev->dev,
- " d[%03x] = 0x%016llx 0x%016llx\n",
- i, rxd->read.pkt_addr,
- rxd->read.hdr_addr);
} else {
rxd = I40E_RX_DESC(ring, i);
dev_info(&pf->pdev->dev,
@@ -614,13 +607,6 @@ static void i40e_dbg_dump_desc(int cnt, int vsi_seid, int ring_id, int desc_n,
"vsi = %02i tx ring = %02i d[%03x] = 0x%016llx 0x%016llx\n",
vsi_seid, ring_id, desc_n,
txd->buffer_addr, txd->cmd_type_offset_bsz);
- } else if (sizeof(union i40e_rx_desc) ==
- sizeof(union i40e_16byte_rx_desc)) {
- rxd = I40E_RX_DESC(ring, desc_n);
- dev_info(&pf->pdev->dev,
- "vsi = %02i rx ring = %02i d[%03x] = 0x%016llx 0x%016llx\n",
- vsi_seid, ring_id, desc_n,
- rxd->read.pkt_addr, rxd->read.hdr_addr);
} else {
rxd = I40E_RX_DESC(ring, desc_n);
dev_info(&pf->pdev->dev,
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index e466111..46a3a67 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -2861,14 +2861,12 @@ static int i40e_configure_rx_ring(struct i40e_ring *ring)
rx_ctx.base = (ring->dma / 128);
rx_ctx.qlen = ring->count;
- if (vsi->back->flags & I40E_FLAG_16BYTE_RX_DESC_ENABLED) {
- set_ring_16byte_desc_enabled(ring);
- rx_ctx.dsize = 0;
- } else {
- rx_ctx.dsize = 1;
- }
+ /* use 32 byte descriptors */
+ rx_ctx.dsize = 1;
- rx_ctx.dtype = vsi->dtype;
+ /* descriptor type is always zero
+ * rx_ctx.dtype = 0;
+ */
rx_ctx.hsplit_0 = 0;
rx_ctx.rxmax = min_t(u16, vsi->max_frame, chain_len * ring->rx_buf_len);
@@ -2948,7 +2946,6 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
vsi->max_frame = I40E_RXBUFFER_2048;
vsi->rx_buf_len = I40E_RXBUFFER_2048;
- vsi->dtype = I40E_RX_DTYPE_NO_SPLIT;
#ifdef I40E_FCOE
/* setup rx buffer for FCoE */
@@ -2956,7 +2953,6 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
(vsi->back->flags & I40E_FLAG_FCOE_ENABLED)) {
vsi->rx_buf_len = I40E_RXBUFFER_3072;
vsi->max_frame = I40E_RXBUFFER_3072;
- vsi->dtype = I40E_RX_DTYPE_NO_SPLIT;
}
#endif /* I40E_FCOE */
@@ -7476,10 +7472,6 @@ static int i40e_alloc_rings(struct i40e_vsi *vsi)
rx_ring->count = vsi->num_desc;
rx_ring->size = 0;
rx_ring->dcb_tc = 0;
- if (pf->flags & I40E_FLAG_16BYTE_RX_DESC_ENABLED)
- set_ring_16byte_desc_enabled(rx_ring);
- else
- clear_ring_16byte_desc_enabled(rx_ring);
rx_ring->rx_itr_setting = pf->rx_itr_default;
vsi->rx_rings[i] = rx_ring;
}
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.h b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
index 37643e6..b78c810 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
@@ -260,15 +260,18 @@ struct i40e_rx_queue_stats {
enum i40e_ring_state_t {
__I40E_TX_FDIR_INIT_DONE,
__I40E_TX_XPS_INIT_DONE,
- __I40E_RX_16BYTE_DESC_ENABLED,
};
-#define ring_is_16byte_desc_enabled(ring) \
- test_bit(__I40E_RX_16BYTE_DESC_ENABLED, &(ring)->state)
-#define set_ring_16byte_desc_enabled(ring) \
- set_bit(__I40E_RX_16BYTE_DESC_ENABLED, &(ring)->state)
-#define clear_ring_16byte_desc_enabled(ring) \
- clear_bit(__I40E_RX_16BYTE_DESC_ENABLED, &(ring)->state)
+/* some useful defines for virtchannel interface, which
+ * is the only remaining user of header split
+ */
+#define I40E_RX_DTYPE_NO_SPLIT 0
+#define I40E_RX_DTYPE_HEADER_SPLIT 1
+#define I40E_RX_DTYPE_SPLIT_ALWAYS 2
+#define I40E_RX_SPLIT_L2 0x1
+#define I40E_RX_SPLIT_IP 0x2
+#define I40E_RX_SPLIT_TCP_UDP 0x4
+#define I40E_RX_SPLIT_SCTP 0x8
/* struct that defines a descriptor ring, associated with a VSI */
struct i40e_ring {
@@ -296,13 +299,6 @@ struct i40e_ring {
u16 count; /* Number of descriptors */
u16 reg_idx; /* HW register index of the ring */
u16 rx_buf_len;
-#define I40E_RX_DTYPE_NO_SPLIT 0
-#define I40E_RX_DTYPE_HEADER_SPLIT 1
-#define I40E_RX_DTYPE_SPLIT_ALWAYS 2
-#define I40E_RX_SPLIT_L2 0x1
-#define I40E_RX_SPLIT_IP 0x2
-#define I40E_RX_SPLIT_TCP_UDP 0x4
-#define I40E_RX_SPLIT_SCTP 0x8
/* used in interrupt processing */
u16 next_to_use;
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
index 4ba302e..0112277 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
@@ -259,15 +259,18 @@ struct i40e_rx_queue_stats {
enum i40e_ring_state_t {
__I40E_TX_FDIR_INIT_DONE,
__I40E_TX_XPS_INIT_DONE,
- __I40E_RX_16BYTE_DESC_ENABLED,
};
-#define ring_is_16byte_desc_enabled(ring) \
- test_bit(__I40E_RX_16BYTE_DESC_ENABLED, &(ring)->state)
-#define set_ring_16byte_desc_enabled(ring) \
- set_bit(__I40E_RX_16BYTE_DESC_ENABLED, &(ring)->state)
-#define clear_ring_16byte_desc_enabled(ring) \
- clear_bit(__I40E_RX_16BYTE_DESC_ENABLED, &(ring)->state)
+/* some useful defines for virtchannel interface, which
+ * is the only remaining user of header split
+ */
+#define I40E_RX_DTYPE_NO_SPLIT 0
+#define I40E_RX_DTYPE_HEADER_SPLIT 1
+#define I40E_RX_DTYPE_SPLIT_ALWAYS 2
+#define I40E_RX_SPLIT_L2 0x1
+#define I40E_RX_SPLIT_IP 0x2
+#define I40E_RX_SPLIT_TCP_UDP 0x4
+#define I40E_RX_SPLIT_SCTP 0x8
/* struct that defines a descriptor ring, associated with a VSI */
struct i40e_ring {
@@ -287,13 +290,6 @@ struct i40e_ring {
u16 count; /* Number of descriptors */
u16 reg_idx; /* HW register index of the ring */
u16 rx_buf_len;
-#define I40E_RX_DTYPE_NO_SPLIT 0
-#define I40E_RX_DTYPE_HEADER_SPLIT 1
-#define I40E_RX_DTYPE_SPLIT_ALWAYS 2
-#define I40E_RX_SPLIT_L2 0x1
-#define I40E_RX_SPLIT_IP 0x2
-#define I40E_RX_SPLIT_TCP_UDP 0x4
-#define I40E_RX_SPLIT_SCTP 0x8
/* used in interrupt processing */
u16 next_to_use;
--
2.5.5
^ permalink raw reply related
* [net-next 02/11] i40e: Drop packet split receive routine
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem; +Cc: Jesse Brandeburg, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
As part of preparation for the rx-refactor, remove the
packet split receive routine and ancillary code.
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@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.h | 3 -
drivers/net/ethernet/intel/i40e/i40e_debugfs.c | 4 +-
drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 19 --
drivers/net/ethernet/intel/i40e/i40e_main.c | 49 +----
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 245 +------------------------
drivers/net/ethernet/intel/i40e/i40e_txrx.h | 7 -
6 files changed, 10 insertions(+), 317 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 00c4738..ea6a69a 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -101,7 +101,6 @@
#define I40E_PRIV_FLAGS_LINKPOLL_FLAG BIT(1)
#define I40E_PRIV_FLAGS_FD_ATR BIT(2)
#define I40E_PRIV_FLAGS_VEB_STATS BIT(3)
-#define I40E_PRIV_FLAGS_PS BIT(4)
#define I40E_PRIV_FLAGS_HW_ATR_EVICT BIT(5)
#define I40E_NVM_VERSION_LO_SHIFT 0
@@ -320,8 +319,6 @@ struct i40e_pf {
#define I40E_FLAG_RX_CSUM_ENABLED BIT_ULL(1)
#define I40E_FLAG_MSI_ENABLED BIT_ULL(2)
#define I40E_FLAG_MSIX_ENABLED BIT_ULL(3)
-#define I40E_FLAG_RX_1BUF_ENABLED BIT_ULL(4)
-#define I40E_FLAG_RX_PS_ENABLED BIT_ULL(5)
#define I40E_FLAG_RSS_ENABLED BIT_ULL(6)
#define I40E_FLAG_VMDQ_ENABLED BIT_ULL(7)
#define I40E_FLAG_FDIR_REQUIRES_REINIT BIT_ULL(8)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
index 83dccf1..f119a74 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
@@ -273,8 +273,8 @@ static void i40e_dbg_dump_vsi_seid(struct i40e_pf *pf, int seid)
rx_ring->rx_buf_len,
rx_ring->dtype);
dev_info(&pf->pdev->dev,
- " rx_rings[%i]: hsplit = %d, next_to_use = %d, next_to_clean = %d, ring_active = %i\n",
- i, ring_is_ps_enabled(rx_ring),
+ " rx_rings[%i]: next_to_use = %d, next_to_clean = %d, ring_active = %i\n",
+ i,
rx_ring->next_to_use,
rx_ring->next_to_clean,
rx_ring->ring_active);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index 8e56c43..858e169 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -2829,8 +2829,6 @@ static u32 i40e_get_priv_flags(struct net_device *dev)
I40E_PRIV_FLAGS_FD_ATR : 0;
ret_flags |= pf->flags & I40E_FLAG_VEB_STATS_ENABLED ?
I40E_PRIV_FLAGS_VEB_STATS : 0;
- ret_flags |= pf->flags & I40E_FLAG_RX_PS_ENABLED ?
- I40E_PRIV_FLAGS_PS : 0;
ret_flags |= pf->auto_disable_flags & I40E_FLAG_HW_ATR_EVICT_CAPABLE ?
0 : I40E_PRIV_FLAGS_HW_ATR_EVICT;
@@ -2851,23 +2849,6 @@ static int i40e_set_priv_flags(struct net_device *dev, u32 flags)
/* NOTE: MFP is not settable */
- /* allow the user to control the method of receive
- * buffer DMA, whether the packet is split at header
- * boundaries into two separate buffers. In some cases
- * one routine or the other will perform better.
- */
- if ((flags & I40E_PRIV_FLAGS_PS) &&
- !(pf->flags & I40E_FLAG_RX_PS_ENABLED)) {
- pf->flags |= I40E_FLAG_RX_PS_ENABLED;
- pf->flags &= ~I40E_FLAG_RX_1BUF_ENABLED;
- reset_required = true;
- } else if (!(flags & I40E_PRIV_FLAGS_PS) &&
- (pf->flags & I40E_FLAG_RX_PS_ENABLED)) {
- pf->flags &= ~I40E_FLAG_RX_PS_ENABLED;
- pf->flags |= I40E_FLAG_RX_1BUF_ENABLED;
- reset_required = true;
- }
-
if (flags & I40E_PRIV_FLAGS_LINKPOLL_FLAG)
pf->flags |= I40E_FLAG_LINK_POLLING_ENABLED;
else
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index f6da6b7..84e8d4e 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -2871,18 +2871,9 @@ static int i40e_configure_rx_ring(struct i40e_ring *ring)
}
rx_ctx.dtype = vsi->dtype;
- if (vsi->dtype) {
- set_ring_ps_enabled(ring);
- rx_ctx.hsplit_0 = I40E_RX_SPLIT_L2 |
- I40E_RX_SPLIT_IP |
- I40E_RX_SPLIT_TCP_UDP |
- I40E_RX_SPLIT_SCTP;
- } else {
- rx_ctx.hsplit_0 = 0;
- }
+ rx_ctx.hsplit_0 = 0;
- rx_ctx.rxmax = min_t(u16, vsi->max_frame,
- (chain_len * ring->rx_buf_len));
+ rx_ctx.rxmax = min_t(u16, vsi->max_frame, chain_len * ring->rx_buf_len);
if (hw->revision_id == 0)
rx_ctx.lrxqthresh = 0;
else
@@ -2919,12 +2910,7 @@ static int i40e_configure_rx_ring(struct i40e_ring *ring)
ring->tail = hw->hw_addr + I40E_QRX_TAIL(pf_q);
writel(0, ring->tail);
- if (ring_is_ps_enabled(ring)) {
- i40e_alloc_rx_headers(ring);
- i40e_alloc_rx_buffers_ps(ring, I40E_DESC_UNUSED(ring));
- } else {
- i40e_alloc_rx_buffers_1buf(ring, I40E_DESC_UNUSED(ring));
- }
+ i40e_alloc_rx_buffers_1buf(ring, I40E_DESC_UNUSED(ring));
return 0;
}
@@ -2963,25 +2949,9 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
else
vsi->max_frame = I40E_RXBUFFER_2048;
- /* figure out correct receive buffer length */
- switch (vsi->back->flags & (I40E_FLAG_RX_1BUF_ENABLED |
- I40E_FLAG_RX_PS_ENABLED)) {
- case I40E_FLAG_RX_1BUF_ENABLED:
- vsi->rx_hdr_len = 0;
- vsi->rx_buf_len = vsi->max_frame;
- vsi->dtype = I40E_RX_DTYPE_NO_SPLIT;
- break;
- case I40E_FLAG_RX_PS_ENABLED:
- vsi->rx_hdr_len = I40E_RX_HDR_SIZE;
- vsi->rx_buf_len = I40E_RXBUFFER_2048;
- vsi->dtype = I40E_RX_DTYPE_HEADER_SPLIT;
- break;
- default:
- vsi->rx_hdr_len = I40E_RX_HDR_SIZE;
- vsi->rx_buf_len = I40E_RXBUFFER_2048;
- vsi->dtype = I40E_RX_DTYPE_SPLIT_ALWAYS;
- break;
- }
+ vsi->rx_hdr_len = 0;
+ vsi->rx_buf_len = vsi->max_frame;
+ vsi->dtype = I40E_RX_DTYPE_NO_SPLIT;
#ifdef I40E_FCOE
/* setup rx buffer for FCoE */
@@ -8460,11 +8430,6 @@ static int i40e_sw_init(struct i40e_pf *pf)
I40E_FLAG_MSI_ENABLED |
I40E_FLAG_MSIX_ENABLED;
- if (iommu_present(&pci_bus_type))
- pf->flags |= I40E_FLAG_RX_PS_ENABLED;
- else
- pf->flags |= I40E_FLAG_RX_1BUF_ENABLED;
-
/* Set default ITR */
pf->rx_itr_default = I40E_ITR_DYNAMIC | I40E_ITR_RX_DEF;
pf->tx_itr_default = I40E_ITR_DYNAMIC | I40E_ITR_TX_DEF;
@@ -10699,7 +10664,7 @@ static void i40e_print_features(struct i40e_pf *pf)
i += snprintf(&buf[i], REMAIN(i), " VSIs: %d QP: %d RX: %s",
pf->hw.func_caps.num_vsis,
pf->vsi[pf->lan_vsi]->num_queue_pairs,
- pf->flags & I40E_FLAG_RX_PS_ENABLED ? "PS" : "1BUF");
+ "1BUF");
if (pf->flags & I40E_FLAG_RSS_ENABLED)
i += snprintf(&buf[i], REMAIN(i), " RSS");
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index dab733c..450ecdd 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -1032,22 +1032,6 @@ void i40e_clean_rx_ring(struct i40e_ring *rx_ring)
if (!rx_ring->rx_bi)
return;
- if (ring_is_ps_enabled(rx_ring)) {
- int bufsz = ALIGN(rx_ring->rx_hdr_len, 256) * rx_ring->count;
-
- rx_bi = &rx_ring->rx_bi[0];
- if (rx_bi->hdr_buf) {
- dma_free_coherent(dev,
- bufsz,
- rx_bi->hdr_buf,
- rx_bi->dma);
- for (i = 0; i < rx_ring->count; i++) {
- rx_bi = &rx_ring->rx_bi[i];
- rx_bi->dma = 0;
- rx_bi->hdr_buf = NULL;
- }
- }
- }
/* Free all the Rx ring sk_buffs */
for (i = 0; i < rx_ring->count; i++) {
rx_bi = &rx_ring->rx_bi[i];
@@ -1503,230 +1487,6 @@ static inline void i40e_rx_hash(struct i40e_ring *ring,
}
/**
- * i40e_clean_rx_irq_ps - Reclaim resources after receive; packet split
- * @rx_ring: rx ring to clean
- * @budget: how many cleans we're allowed
- *
- * Returns true if there's any budget left (e.g. the clean is finished)
- **/
-static int i40e_clean_rx_irq_ps(struct i40e_ring *rx_ring, const int budget)
-{
- unsigned int total_rx_bytes = 0, total_rx_packets = 0;
- u16 rx_packet_len, rx_header_len, rx_sph, rx_hbo;
- u16 cleaned_count = I40E_DESC_UNUSED(rx_ring);
- struct i40e_vsi *vsi = rx_ring->vsi;
- u16 i = rx_ring->next_to_clean;
- union i40e_rx_desc *rx_desc;
- u32 rx_error, rx_status;
- bool failure = false;
- u8 rx_ptype;
- u64 qword;
- u32 copysize;
-
- if (budget <= 0)
- return 0;
-
- do {
- struct i40e_rx_buffer *rx_bi;
- struct sk_buff *skb;
- u16 vlan_tag;
- /* return some buffers to hardware, one at a time is too slow */
- if (cleaned_count >= I40E_RX_BUFFER_WRITE) {
- failure = failure ||
- i40e_alloc_rx_buffers_ps(rx_ring,
- cleaned_count);
- cleaned_count = 0;
- }
-
- i = rx_ring->next_to_clean;
- rx_desc = I40E_RX_DESC(rx_ring, i);
- qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
- rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >>
- I40E_RXD_QW1_STATUS_SHIFT;
-
- if (!(rx_status & BIT(I40E_RX_DESC_STATUS_DD_SHIFT)))
- break;
-
- /* This memory barrier is needed to keep us from reading
- * any other fields out of the rx_desc until we know the
- * DD bit is set.
- */
- dma_rmb();
- /* sync header buffer for reading */
- dma_sync_single_range_for_cpu(rx_ring->dev,
- rx_ring->rx_bi[0].dma,
- i * rx_ring->rx_hdr_len,
- rx_ring->rx_hdr_len,
- DMA_FROM_DEVICE);
- if (i40e_rx_is_programming_status(qword)) {
- i40e_clean_programming_status(rx_ring, rx_desc);
- I40E_RX_INCREMENT(rx_ring, i);
- continue;
- }
- rx_bi = &rx_ring->rx_bi[i];
- skb = rx_bi->skb;
- if (likely(!skb)) {
- skb = __netdev_alloc_skb_ip_align(rx_ring->netdev,
- rx_ring->rx_hdr_len,
- GFP_ATOMIC |
- __GFP_NOWARN);
- if (!skb) {
- rx_ring->rx_stats.alloc_buff_failed++;
- failure = true;
- break;
- }
-
- /* initialize queue mapping */
- skb_record_rx_queue(skb, rx_ring->queue_index);
- /* we are reusing so sync this buffer for CPU use */
- dma_sync_single_range_for_cpu(rx_ring->dev,
- rx_ring->rx_bi[0].dma,
- i * rx_ring->rx_hdr_len,
- rx_ring->rx_hdr_len,
- DMA_FROM_DEVICE);
- }
- rx_packet_len = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
- I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
- rx_header_len = (qword & I40E_RXD_QW1_LENGTH_HBUF_MASK) >>
- I40E_RXD_QW1_LENGTH_HBUF_SHIFT;
- rx_sph = (qword & I40E_RXD_QW1_LENGTH_SPH_MASK) >>
- I40E_RXD_QW1_LENGTH_SPH_SHIFT;
-
- rx_error = (qword & I40E_RXD_QW1_ERROR_MASK) >>
- I40E_RXD_QW1_ERROR_SHIFT;
- rx_hbo = rx_error & BIT(I40E_RX_DESC_ERROR_HBO_SHIFT);
- rx_error &= ~BIT(I40E_RX_DESC_ERROR_HBO_SHIFT);
-
- rx_ptype = (qword & I40E_RXD_QW1_PTYPE_MASK) >>
- I40E_RXD_QW1_PTYPE_SHIFT;
- /* sync half-page for reading */
- dma_sync_single_range_for_cpu(rx_ring->dev,
- rx_bi->page_dma,
- rx_bi->page_offset,
- PAGE_SIZE / 2,
- DMA_FROM_DEVICE);
- prefetch(page_address(rx_bi->page) + rx_bi->page_offset);
- rx_bi->skb = NULL;
- cleaned_count++;
- copysize = 0;
- if (rx_hbo || rx_sph) {
- int len;
-
- if (rx_hbo)
- len = I40E_RX_HDR_SIZE;
- else
- len = rx_header_len;
- memcpy(__skb_put(skb, len), rx_bi->hdr_buf, len);
- } else if (skb->len == 0) {
- int len;
- unsigned char *va = page_address(rx_bi->page) +
- rx_bi->page_offset;
-
- len = min(rx_packet_len, rx_ring->rx_hdr_len);
- memcpy(__skb_put(skb, len), va, len);
- copysize = len;
- rx_packet_len -= len;
- }
- /* Get the rest of the data if this was a header split */
- if (rx_packet_len) {
- skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags,
- rx_bi->page,
- rx_bi->page_offset + copysize,
- rx_packet_len, I40E_RXBUFFER_2048);
-
- /* If the page count is more than 2, then both halves
- * of the page are used and we need to free it. Do it
- * here instead of in the alloc code. Otherwise one
- * of the half-pages might be released between now and
- * then, and we wouldn't know which one to use.
- * Don't call get_page and free_page since those are
- * both expensive atomic operations that just change
- * the refcount in opposite directions. Just give the
- * page to the stack; he can have our refcount.
- */
- if (page_count(rx_bi->page) > 2) {
- dma_unmap_page(rx_ring->dev,
- rx_bi->page_dma,
- PAGE_SIZE,
- DMA_FROM_DEVICE);
- rx_bi->page = NULL;
- rx_bi->page_dma = 0;
- rx_ring->rx_stats.realloc_count++;
- } else {
- get_page(rx_bi->page);
- /* switch to the other half-page here; the
- * allocation code programs the right addr
- * into HW. If we haven't used this half-page,
- * the address won't be changed, and HW can
- * just use it next time through.
- */
- rx_bi->page_offset ^= PAGE_SIZE / 2;
- }
-
- }
- I40E_RX_INCREMENT(rx_ring, i);
-
- if (unlikely(
- !(rx_status & BIT(I40E_RX_DESC_STATUS_EOF_SHIFT)))) {
- struct i40e_rx_buffer *next_buffer;
-
- next_buffer = &rx_ring->rx_bi[i];
- next_buffer->skb = skb;
- rx_ring->rx_stats.non_eop_descs++;
- continue;
- }
-
- /* ERR_MASK will only have valid bits if EOP set */
- if (unlikely(rx_error & BIT(I40E_RX_DESC_ERROR_RXE_SHIFT))) {
- dev_kfree_skb_any(skb);
- continue;
- }
-
- i40e_rx_hash(rx_ring, rx_desc, skb, rx_ptype);
-
- if (unlikely(rx_status & I40E_RXD_QW1_STATUS_TSYNVALID_MASK)) {
- i40e_ptp_rx_hwtstamp(vsi->back, skb, (rx_status &
- I40E_RXD_QW1_STATUS_TSYNINDX_MASK) >>
- I40E_RXD_QW1_STATUS_TSYNINDX_SHIFT);
- rx_ring->last_rx_timestamp = jiffies;
- }
-
- /* probably a little skewed due to removing CRC */
- total_rx_bytes += skb->len;
- total_rx_packets++;
-
- skb->protocol = eth_type_trans(skb, rx_ring->netdev);
-
- i40e_rx_checksum(vsi, skb, rx_status, rx_error, rx_ptype);
-
- vlan_tag = rx_status & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)
- ? le16_to_cpu(rx_desc->wb.qword0.lo_dword.l2tag1)
- : 0;
-#ifdef I40E_FCOE
- if (unlikely(
- i40e_rx_is_fcoe(rx_ptype) &&
- !i40e_fcoe_handle_offload(rx_ring, rx_desc, skb))) {
- dev_kfree_skb_any(skb);
- continue;
- }
-#endif
- i40e_receive_skb(rx_ring, skb, vlan_tag);
-
- rx_desc->wb.qword1.status_error_len = 0;
-
- } while (likely(total_rx_packets < budget));
-
- u64_stats_update_begin(&rx_ring->syncp);
- rx_ring->stats.packets += total_rx_packets;
- rx_ring->stats.bytes += total_rx_bytes;
- u64_stats_update_end(&rx_ring->syncp);
- rx_ring->q_vector->rx.total_packets += total_rx_packets;
- rx_ring->q_vector->rx.total_bytes += total_rx_bytes;
-
- return failure ? budget : total_rx_packets;
-}
-
-/**
* i40e_clean_rx_irq_1buf - Reclaim resources after receive; single buffer
* @rx_ring: rx ring to clean
* @budget: how many cleans we're allowed
@@ -2001,10 +1761,7 @@ int i40e_napi_poll(struct napi_struct *napi, int budget)
i40e_for_each_ring(ring, q_vector->rx) {
int cleaned;
- if (ring_is_ps_enabled(ring))
- cleaned = i40e_clean_rx_irq_ps(ring, budget_per_ring);
- else
- cleaned = i40e_clean_rx_irq_1buf(ring, budget_per_ring);
+ cleaned = i40e_clean_rx_irq_1buf(ring, budget_per_ring);
work_done += cleaned;
/* if we clean as many as budgeted, we must not be done */
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.h b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
index 6b2b191..5a2d0fe 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
@@ -245,16 +245,9 @@ struct i40e_rx_queue_stats {
enum i40e_ring_state_t {
__I40E_TX_FDIR_INIT_DONE,
__I40E_TX_XPS_INIT_DONE,
- __I40E_RX_PS_ENABLED,
__I40E_RX_16BYTE_DESC_ENABLED,
};
-#define ring_is_ps_enabled(ring) \
- test_bit(__I40E_RX_PS_ENABLED, &(ring)->state)
-#define set_ring_ps_enabled(ring) \
- set_bit(__I40E_RX_PS_ENABLED, &(ring)->state)
-#define clear_ring_ps_enabled(ring) \
- clear_bit(__I40E_RX_PS_ENABLED, &(ring)->state)
#define ring_is_16byte_desc_enabled(ring) \
test_bit(__I40E_RX_16BYTE_DESC_ENABLED, &(ring)->state)
#define set_ring_16byte_desc_enabled(ring) \
--
2.5.5
^ permalink raw reply related
* [net-next 03/11] i40e/i40evf: Remove reference to ring->dtype
From: Jeff Kirsher @ 2016-05-06 7:03 UTC (permalink / raw)
To: davem; +Cc: Jesse Brandeburg, netdev, nhorman, sassmann, jogreene,
Jeff Kirsher
In-Reply-To: <1462518228-30527-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
As part of the rx-refactor, the dtype variable in the i40e_ring
struct is no longer used, so remove it.
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@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_debugfs.c | 8 ++------
drivers/net/ethernet/intel/i40e/i40e_txrx.h | 1 -
drivers/net/ethernet/intel/i40evf/i40e_txrx.h | 1 -
3 files changed, 2 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
index f119a74..c0a01e0 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
@@ -268,10 +268,9 @@ static void i40e_dbg_dump_vsi_seid(struct i40e_pf *pf, int seid)
rx_ring->queue_index,
rx_ring->reg_idx);
dev_info(&pf->pdev->dev,
- " rx_rings[%i]: rx_hdr_len = %d, rx_buf_len = %d, dtype = %d\n",
+ " rx_rings[%i]: rx_hdr_len = %d, rx_buf_len = %d\n",
i, rx_ring->rx_hdr_len,
- rx_ring->rx_buf_len,
- rx_ring->dtype);
+ rx_ring->rx_buf_len);
dev_info(&pf->pdev->dev,
" rx_rings[%i]: next_to_use = %d, next_to_clean = %d, ring_active = %i\n",
i,
@@ -326,9 +325,6 @@ static void i40e_dbg_dump_vsi_seid(struct i40e_pf *pf, int seid)
tx_ring->queue_index,
tx_ring->reg_idx);
dev_info(&pf->pdev->dev,
- " tx_rings[%i]: dtype = %d\n",
- i, tx_ring->dtype);
- dev_info(&pf->pdev->dev,
" tx_rings[%i]: next_to_use = %d, next_to_clean = %d, ring_active = %i\n",
i,
tx_ring->next_to_use,
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.h b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
index 5a2d0fe..03e21d9 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
@@ -282,7 +282,6 @@ struct i40e_ring {
u16 reg_idx; /* HW register index of the ring */
u16 rx_hdr_len;
u16 rx_buf_len;
- u8 dtype;
#define I40E_RX_DTYPE_NO_SPLIT 0
#define I40E_RX_DTYPE_HEADER_SPLIT 1
#define I40E_RX_DTYPE_SPLIT_ALWAYS 2
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
index 54b52e8..3b3f976 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
@@ -280,7 +280,6 @@ struct i40e_ring {
u16 reg_idx; /* HW register index of the ring */
u16 rx_hdr_len;
u16 rx_buf_len;
- u8 dtype;
#define I40E_RX_DTYPE_NO_SPLIT 0
#define I40E_RX_DTYPE_HEADER_SPLIT 1
#define I40E_RX_DTYPE_SPLIT_ALWAYS 2
--
2.5.5
^ 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