* [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 18:13 ` Marcelo Ricardo Leitner
2021-02-09 0:20 ` patchwork-bot+netdevbpf
2021-02-06 5:02 ` [net-next V2 02/17] net/mlx5e: E-Switch, Maintain vhca_id to vport_num mapping Saeed Mahameed
` (17 subsequent siblings)
18 siblings, 2 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski; +Cc: David S. Miller, netdev, Mark Bloch, Saeed Mahameed
From: Mark Bloch <mbloch@nvidia.com>
Setting the source port requires only the E-Switch and vport number.
Refactor the function to get those parameters instead of passing the full
attribute.
Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../mellanox/mlx5/core/eswitch_offloads.c | 19 ++++++++++++-------
1 file changed, 12 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 7f09f2bbf7c1..416ede2fe5d7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -257,7 +257,8 @@ mlx5_eswitch_set_rule_flow_source(struct mlx5_eswitch *esw,
static void
mlx5_eswitch_set_rule_source_port(struct mlx5_eswitch *esw,
struct mlx5_flow_spec *spec,
- struct mlx5_esw_flow_attr *attr)
+ struct mlx5_eswitch *src_esw,
+ u16 vport)
{
void *misc2;
void *misc;
@@ -268,8 +269,8 @@ mlx5_eswitch_set_rule_source_port(struct mlx5_eswitch *esw,
if (mlx5_eswitch_vport_match_metadata_enabled(esw)) {
misc2 = MLX5_ADDR_OF(fte_match_param, spec->match_value, misc_parameters_2);
MLX5_SET(fte_match_set_misc2, misc2, metadata_reg_c_0,
- mlx5_eswitch_get_vport_metadata_for_match(attr->in_mdev->priv.eswitch,
- attr->in_rep->vport));
+ mlx5_eswitch_get_vport_metadata_for_match(src_esw,
+ vport));
misc2 = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, misc_parameters_2);
MLX5_SET(fte_match_set_misc2, misc2, metadata_reg_c_0,
@@ -278,12 +279,12 @@ mlx5_eswitch_set_rule_source_port(struct mlx5_eswitch *esw,
spec->match_criteria_enable |= MLX5_MATCH_MISC_PARAMETERS_2;
} else {
misc = MLX5_ADDR_OF(fte_match_param, spec->match_value, misc_parameters);
- MLX5_SET(fte_match_set_misc, misc, source_port, attr->in_rep->vport);
+ MLX5_SET(fte_match_set_misc, misc, source_port, vport);
if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
MLX5_SET(fte_match_set_misc, misc,
source_eswitch_owner_vhca_id,
- MLX5_CAP_GEN(attr->in_mdev, vhca_id));
+ MLX5_CAP_GEN(src_esw->dev, vhca_id));
misc = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, misc_parameters);
MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_port);
@@ -407,7 +408,9 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
fdb = attr->ft;
if (!(attr->flags & MLX5_ESW_ATTR_FLAG_NO_IN_PORT))
- mlx5_eswitch_set_rule_source_port(esw, spec, esw_attr);
+ mlx5_eswitch_set_rule_source_port(esw, spec,
+ esw_attr->in_mdev->priv.eswitch,
+ esw_attr->in_rep->vport);
}
if (IS_ERR(fdb)) {
rule = ERR_CAST(fdb);
@@ -487,7 +490,9 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
dest[i].ft = fwd_fdb;
i++;
- mlx5_eswitch_set_rule_source_port(esw, spec, esw_attr);
+ mlx5_eswitch_set_rule_source_port(esw, spec,
+ esw_attr->in_mdev->priv.eswitch,
+ esw_attr->in_rep->vport);
if (attr->outer_match_level != MLX5_MATCH_NONE)
spec->match_criteria_enable |= MLX5_MATCH_OUTER_HEADERS;
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-06 5:02 ` [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port Saeed Mahameed
@ 2021-02-06 18:13 ` Marcelo Ricardo Leitner
2021-02-08 8:21 ` Vlad Buslov
2021-02-09 0:20 ` patchwork-bot+netdevbpf
1 sibling, 1 reply; 39+ messages in thread
From: Marcelo Ricardo Leitner @ 2021-02-06 18:13 UTC (permalink / raw)
To: Saeed Mahameed
Cc: Jakub Kicinski, David S. Miller, netdev, Mark Bloch,
Saeed Mahameed, Vlad Buslov
Hi,
I didn't receive the cover letter, so I'm replying on this one. :-)
This is nice. One thing is not clear to me yet. From the samples on
the cover letter:
$ tc -s filter show dev enp8s0f0_1 ingress
filter protocol ip pref 4 flower chain 0
filter protocol ip pref 4 flower chain 0 handle 0x1
dst_mac 0a:40:bd:30:89:99
src_mac ca:2e:a7:3f:f5:0f
eth_type ipv4
ip_tos 0/0x3
ip_flags nofrag
in_hw in_hw_count 1
action order 1: tunnel_key set
src_ip 7.7.7.5
dst_ip 7.7.7.1
...
$ tc -s filter show dev vxlan_sys_4789 ingress
filter protocol ip pref 4 flower chain 0
filter protocol ip pref 4 flower chain 0 handle 0x1
dst_mac ca:2e:a7:3f:f5:0f
src_mac 0a:40:bd:30:89:99
eth_type ipv4
enc_dst_ip 7.7.7.5
enc_src_ip 7.7.7.1
enc_key_id 98
enc_dst_port 4789
enc_tos 0
...
These operations imply that 7.7.7.5 is configured on some interface on
the host. Most likely the VF representor itself, as that aids with ARP
resolution. Is that so?
Thanks,
Marcelo
^ permalink raw reply [flat|nested] 39+ messages in thread* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-06 18:13 ` Marcelo Ricardo Leitner
@ 2021-02-08 8:21 ` Vlad Buslov
2021-02-08 13:25 ` Marcelo Ricardo Leitner
2021-02-08 20:22 ` Jakub Kicinski
0 siblings, 2 replies; 39+ messages in thread
From: Vlad Buslov @ 2021-02-08 8:21 UTC (permalink / raw)
To: Marcelo Ricardo Leitner
Cc: Saeed Mahameed, Jakub Kicinski, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed
On Sat 06 Feb 2021 at 20:13, Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> wrote:
> Hi,
>
> I didn't receive the cover letter, so I'm replying on this one. :-)
>
> This is nice. One thing is not clear to me yet. From the samples on
> the cover letter:
>
> $ tc -s filter show dev enp8s0f0_1 ingress
> filter protocol ip pref 4 flower chain 0
> filter protocol ip pref 4 flower chain 0 handle 0x1
> dst_mac 0a:40:bd:30:89:99
> src_mac ca:2e:a7:3f:f5:0f
> eth_type ipv4
> ip_tos 0/0x3
> ip_flags nofrag
> in_hw in_hw_count 1
> action order 1: tunnel_key set
> src_ip 7.7.7.5
> dst_ip 7.7.7.1
> ...
>
> $ tc -s filter show dev vxlan_sys_4789 ingress
> filter protocol ip pref 4 flower chain 0
> filter protocol ip pref 4 flower chain 0 handle 0x1
> dst_mac ca:2e:a7:3f:f5:0f
> src_mac 0a:40:bd:30:89:99
> eth_type ipv4
> enc_dst_ip 7.7.7.5
> enc_src_ip 7.7.7.1
> enc_key_id 98
> enc_dst_port 4789
> enc_tos 0
> ...
>
> These operations imply that 7.7.7.5 is configured on some interface on
> the host. Most likely the VF representor itself, as that aids with ARP
> resolution. Is that so?
>
> Thanks,
> Marcelo
Hi Marcelo,
The tunnel endpoint IP address is configured on VF that is represented
by enp8s0f0_0 representor in example rules. The VF is on host.
Regards,
Vlad
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-08 8:21 ` Vlad Buslov
@ 2021-02-08 13:25 ` Marcelo Ricardo Leitner
2021-02-08 13:31 ` Vlad Buslov
2021-02-08 20:22 ` Jakub Kicinski
1 sibling, 1 reply; 39+ messages in thread
From: Marcelo Ricardo Leitner @ 2021-02-08 13:25 UTC (permalink / raw)
To: Vlad Buslov
Cc: Saeed Mahameed, Jakub Kicinski, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed
On Mon, Feb 08, 2021 at 10:21:21AM +0200, Vlad Buslov wrote:
>
> On Sat 06 Feb 2021 at 20:13, Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> wrote:
> > Hi,
> >
> > I didn't receive the cover letter, so I'm replying on this one. :-)
> >
> > This is nice. One thing is not clear to me yet. From the samples on
> > the cover letter:
> >
> > $ tc -s filter show dev enp8s0f0_1 ingress
> > filter protocol ip pref 4 flower chain 0
> > filter protocol ip pref 4 flower chain 0 handle 0x1
> > dst_mac 0a:40:bd:30:89:99
> > src_mac ca:2e:a7:3f:f5:0f
> > eth_type ipv4
> > ip_tos 0/0x3
> > ip_flags nofrag
> > in_hw in_hw_count 1
> > action order 1: tunnel_key set
> > src_ip 7.7.7.5
> > dst_ip 7.7.7.1
> > ...
> >
> > $ tc -s filter show dev vxlan_sys_4789 ingress
> > filter protocol ip pref 4 flower chain 0
> > filter protocol ip pref 4 flower chain 0 handle 0x1
> > dst_mac ca:2e:a7:3f:f5:0f
> > src_mac 0a:40:bd:30:89:99
> > eth_type ipv4
> > enc_dst_ip 7.7.7.5
> > enc_src_ip 7.7.7.1
> > enc_key_id 98
> > enc_dst_port 4789
> > enc_tos 0
> > ...
> >
> > These operations imply that 7.7.7.5 is configured on some interface on
> > the host. Most likely the VF representor itself, as that aids with ARP
> > resolution. Is that so?
> >
> > Thanks,
> > Marcelo
>
> Hi Marcelo,
>
> The tunnel endpoint IP address is configured on VF that is represented
> by enp8s0f0_0 representor in example rules. The VF is on host.
That's interesting and odd. The VF would be isolated by a netns and
not be visible by whoever is administrating the VF representor. Some
cooperation between the two entities (host and container, say) is
needed then, right? Because the host needs to know the endpoint IP
address that the container will be using, and vice-versa. If so, why
not offload the tunnel actions via the VF itself and avoid this need
for cooperation? Container privileges maybe?
Thx,
Marcelo
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-08 13:25 ` Marcelo Ricardo Leitner
@ 2021-02-08 13:31 ` Vlad Buslov
2021-02-08 13:42 ` Marcelo Ricardo Leitner
0 siblings, 1 reply; 39+ messages in thread
From: Vlad Buslov @ 2021-02-08 13:31 UTC (permalink / raw)
To: Marcelo Ricardo Leitner
Cc: Saeed Mahameed, Jakub Kicinski, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed
On Mon 08 Feb 2021 at 15:25, Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> wrote:
> On Mon, Feb 08, 2021 at 10:21:21AM +0200, Vlad Buslov wrote:
>>
>> On Sat 06 Feb 2021 at 20:13, Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> wrote:
>> > Hi,
>> >
>> > I didn't receive the cover letter, so I'm replying on this one. :-)
>> >
>> > This is nice. One thing is not clear to me yet. From the samples on
>> > the cover letter:
>> >
>> > $ tc -s filter show dev enp8s0f0_1 ingress
>> > filter protocol ip pref 4 flower chain 0
>> > filter protocol ip pref 4 flower chain 0 handle 0x1
>> > dst_mac 0a:40:bd:30:89:99
>> > src_mac ca:2e:a7:3f:f5:0f
>> > eth_type ipv4
>> > ip_tos 0/0x3
>> > ip_flags nofrag
>> > in_hw in_hw_count 1
>> > action order 1: tunnel_key set
>> > src_ip 7.7.7.5
>> > dst_ip 7.7.7.1
>> > ...
>> >
>> > $ tc -s filter show dev vxlan_sys_4789 ingress
>> > filter protocol ip pref 4 flower chain 0
>> > filter protocol ip pref 4 flower chain 0 handle 0x1
>> > dst_mac ca:2e:a7:3f:f5:0f
>> > src_mac 0a:40:bd:30:89:99
>> > eth_type ipv4
>> > enc_dst_ip 7.7.7.5
>> > enc_src_ip 7.7.7.1
>> > enc_key_id 98
>> > enc_dst_port 4789
>> > enc_tos 0
>> > ...
>> >
>> > These operations imply that 7.7.7.5 is configured on some interface on
>> > the host. Most likely the VF representor itself, as that aids with ARP
>> > resolution. Is that so?
>> >
>> > Thanks,
>> > Marcelo
>>
>> Hi Marcelo,
>>
>> The tunnel endpoint IP address is configured on VF that is represented
>> by enp8s0f0_0 representor in example rules. The VF is on host.
>
> That's interesting and odd. The VF would be isolated by a netns and
> not be visible by whoever is administrating the VF representor. Some
> cooperation between the two entities (host and container, say) is
> needed then, right? Because the host needs to know the endpoint IP
> address that the container will be using, and vice-versa. If so, why
> not offload the tunnel actions via the VF itself and avoid this need
> for cooperation? Container privileges maybe?
>
> Thx,
> Marcelo
As I wrote in previous email, tunnel endpoint VF is on host (not in
namespace/container, VM, etc.).
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-08 13:31 ` Vlad Buslov
@ 2021-02-08 13:42 ` Marcelo Ricardo Leitner
0 siblings, 0 replies; 39+ messages in thread
From: Marcelo Ricardo Leitner @ 2021-02-08 13:42 UTC (permalink / raw)
To: Vlad Buslov
Cc: Saeed Mahameed, Jakub Kicinski, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed
On Mon, Feb 08, 2021 at 03:31:50PM +0200, Vlad Buslov wrote:
>
> On Mon 08 Feb 2021 at 15:25, Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> wrote:
> > On Mon, Feb 08, 2021 at 10:21:21AM +0200, Vlad Buslov wrote:
> >>
> >> On Sat 06 Feb 2021 at 20:13, Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> wrote:
> >> > Hi,
> >> >
> >> > I didn't receive the cover letter, so I'm replying on this one. :-)
> >> >
> >> > This is nice. One thing is not clear to me yet. From the samples on
> >> > the cover letter:
> >> >
> >> > $ tc -s filter show dev enp8s0f0_1 ingress
> >> > filter protocol ip pref 4 flower chain 0
> >> > filter protocol ip pref 4 flower chain 0 handle 0x1
> >> > dst_mac 0a:40:bd:30:89:99
> >> > src_mac ca:2e:a7:3f:f5:0f
> >> > eth_type ipv4
> >> > ip_tos 0/0x3
> >> > ip_flags nofrag
> >> > in_hw in_hw_count 1
> >> > action order 1: tunnel_key set
> >> > src_ip 7.7.7.5
> >> > dst_ip 7.7.7.1
> >> > ...
> >> >
> >> > $ tc -s filter show dev vxlan_sys_4789 ingress
> >> > filter protocol ip pref 4 flower chain 0
> >> > filter protocol ip pref 4 flower chain 0 handle 0x1
> >> > dst_mac ca:2e:a7:3f:f5:0f
> >> > src_mac 0a:40:bd:30:89:99
> >> > eth_type ipv4
> >> > enc_dst_ip 7.7.7.5
> >> > enc_src_ip 7.7.7.1
> >> > enc_key_id 98
> >> > enc_dst_port 4789
> >> > enc_tos 0
> >> > ...
> >> >
> >> > These operations imply that 7.7.7.5 is configured on some interface on
> >> > the host. Most likely the VF representor itself, as that aids with ARP
> >> > resolution. Is that so?
> >> >
> >> > Thanks,
> >> > Marcelo
> >>
> >> Hi Marcelo,
> >>
> >> The tunnel endpoint IP address is configured on VF that is represented
> >> by enp8s0f0_0 representor in example rules. The VF is on host.
> >
> > That's interesting and odd. The VF would be isolated by a netns and
> > not be visible by whoever is administrating the VF representor. Some
> > cooperation between the two entities (host and container, say) is
> > needed then, right? Because the host needs to know the endpoint IP
> > address that the container will be using, and vice-versa. If so, why
> > not offload the tunnel actions via the VF itself and avoid this need
> > for cooperation? Container privileges maybe?
> >
> > Thx,
> > Marcelo
>
> As I wrote in previous email, tunnel endpoint VF is on host (not in
> namespace/container, VM, etc.).
Right. I assumed it was just for simplicity of testing. Okay, I think
I can see some use cases for this. Thanks.
Cheers,
Marcelo
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-08 8:21 ` Vlad Buslov
2021-02-08 13:25 ` Marcelo Ricardo Leitner
@ 2021-02-08 20:22 ` Jakub Kicinski
2021-02-09 14:22 ` Vlad Buslov
1 sibling, 1 reply; 39+ messages in thread
From: Jakub Kicinski @ 2021-02-08 20:22 UTC (permalink / raw)
To: Vlad Buslov
Cc: Marcelo Ricardo Leitner, Saeed Mahameed, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed
On Mon, 8 Feb 2021 10:21:21 +0200 Vlad Buslov wrote:
> > These operations imply that 7.7.7.5 is configured on some interface on
> > the host. Most likely the VF representor itself, as that aids with ARP
> > resolution. Is that so?
>
> Hi Marcelo,
>
> The tunnel endpoint IP address is configured on VF that is represented
> by enp8s0f0_0 representor in example rules. The VF is on host.
This is very confusing, are you saying that the 7.7.7.5 is configured
both on VF and VFrep? Could you provide a full picture of the config
with IP addresses and routing?
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-08 20:22 ` Jakub Kicinski
@ 2021-02-09 14:22 ` Vlad Buslov
2021-02-09 16:10 ` Or Gerlitz
2021-02-09 18:05 ` Jakub Kicinski
0 siblings, 2 replies; 39+ messages in thread
From: Vlad Buslov @ 2021-02-09 14:22 UTC (permalink / raw)
To: Jakub Kicinski
Cc: Marcelo Ricardo Leitner, Saeed Mahameed, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed
On Mon 08 Feb 2021 at 22:22, Jakub Kicinski <kuba@kernel.org> wrote:
> On Mon, 8 Feb 2021 10:21:21 +0200 Vlad Buslov wrote:
>> > These operations imply that 7.7.7.5 is configured on some interface on
>> > the host. Most likely the VF representor itself, as that aids with ARP
>> > resolution. Is that so?
>>
>> Hi Marcelo,
>>
>> The tunnel endpoint IP address is configured on VF that is represented
>> by enp8s0f0_0 representor in example rules. The VF is on host.
>
> This is very confusing, are you saying that the 7.7.7.5 is configured
> both on VF and VFrep? Could you provide a full picture of the config
> with IP addresses and routing?
Hi Jakub,
No, tunnel IP is configured on VF. That particular VF is in host
namespace. When mlx5 resolves tunneling the code checks if tunnel
endpoint IP address is on such mlx5 VF, since the VF is in same
namespace as eswitch manager (e.g. on host) and route returned by
ip_route_output_key() is resolved through rt->dst.dev==tunVF device.
After establishing that tunnel is on VF the goal is to process two
resulting TC rules (in both directions) fully in hardware without
exposing the packet on tunneling device or tunnel VF in sw, which is
implemented with all the infrastructure from this series.
So, to summarize with IP addresses from TC examples presented in cover letter,
we have underlay network 7.7.7.0/24 in host namespace with tunnel endpoint IP
address on VF:
$ ip a show dev enp8s0f0v0
1537: enp8s0f0v0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
link/ether 52:e5:6d:f2:00:69 brd ff:ff:ff:ff:ff:ff
altname enp8s0f0np0v0
inet 7.7.7.5/24 scope global enp8s0f0v0
valid_lft forever preferred_lft forever
inet6 fe80::50e5:6dff:fef2:69/64 scope link
valid_lft forever preferred_lft forever
Like all VFs in switchdev model the tunnel VF is controlled through representor
that doesn't have any IP address assigned:
$ ip a show dev enp8s0f0_0
1534: enp8s0f0_0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq master ovs-system state UP group default qlen 1000
link/ether 96:98:b1:59:aa:5e brd ff:ff:ff:ff:ff:ff
altname enp8s0f0npf0vf0
inet6 fe80::9498:b1ff:fe59:aa5e/64 scope link
valid_lft forever preferred_lft forever
User VFs have IP addresses from overlay network (5.5.5.0/24 in my tests) and are
in namespaces/VMs, while only their representors are on host attached to same
v-switch bridge with tunnel VF represetor:
$ sudo ip netns exec ns0 ip a show dev enp8s0f0v1
1538: enp8s0f0v1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
link/ether 9e:cf:b5:69:84:d1 brd ff:ff:ff:ff:ff:ff
altname enp8s0f0np0v1
inet 5.5.5.5/24 scope global enp8s0f0v1
valid_lft forever preferred_lft forever
inet6 fe80::9ccf:b5ff:fe69:84d1/64 scope link
valid_lft forever preferred_lft forever
$ ip a show dev enp8s0f0_1
1535: enp8s0f0_1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq master ovs-system state UP group default qlen 1000
link/ether 06:96:1e:23:df:a4 brd ff:ff:ff:ff:ff:ff
altname enp8s0f0npf0vf1
OVS bridge ports:
$ sudo ovs-vsctl list-ports ovs-br
enp8s0f0
enp8s0f0_0
enp8s0f0_1
enp8s0f0_2
vxlan0
The TC rules from cover letter are installed by OVS configured according to
description above when running iperf traffic from namespaced VF enp8s0f0v1 to
another machine connected over uplink port:
$ sudo ip netns exec ns0 iperf3 -c 5.5.5.1 -t 10000
Connecting to host 5.5.5.1, port 5201
[ 5] local 5.5.5.5 port 34486 connected to 5.5.5.1 port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 158 MBytes 1.32 Gbits/sec 41 771 KBytes
Hope this clarifies things and sorry for confusion!
Regards,
Vlad
^ permalink raw reply [flat|nested] 39+ messages in thread* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-09 14:22 ` Vlad Buslov
@ 2021-02-09 16:10 ` Or Gerlitz
2021-02-10 13:56 ` Marcelo Ricardo Leitner
2021-02-09 18:05 ` Jakub Kicinski
1 sibling, 1 reply; 39+ messages in thread
From: Or Gerlitz @ 2021-02-09 16:10 UTC (permalink / raw)
To: Vlad Buslov
Cc: Jakub Kicinski, Marcelo Ricardo Leitner, Saeed Mahameed,
David S. Miller, Linux Netdev List, Mark Bloch, Saeed Mahameed
On Tue, Feb 9, 2021 at 4:26 PM Vlad Buslov <vladbu@nvidia.com> wrote:
> On Mon 08 Feb 2021 at 22:22, Jakub Kicinski <kuba@kernel.org> wrote:
> > On Mon, 8 Feb 2021 10:21:21 +0200 Vlad Buslov wrote:
> >> > These operations imply that 7.7.7.5 is configured on some interface on
> >> > the host. Most likely the VF representor itself, as that aids with ARP
> >> > resolution. Is that so?
> >> The tunnel endpoint IP address is configured on VF that is represented
> >> by enp8s0f0_0 representor in example rules. The VF is on host.
> > This is very confusing, are you saying that the 7.7.7.5 is configured
> > both on VF and VFrep? Could you provide a full picture of the config
> > with IP addresses and routing?
> No, tunnel IP is configured on VF. That particular VF is in host [..]
What's the motivation for that? isn't that introducing 3x slow down?
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-09 16:10 ` Or Gerlitz
@ 2021-02-10 13:56 ` Marcelo Ricardo Leitner
2021-02-10 16:44 ` Vlad Buslov
0 siblings, 1 reply; 39+ messages in thread
From: Marcelo Ricardo Leitner @ 2021-02-10 13:56 UTC (permalink / raw)
To: Or Gerlitz
Cc: Vlad Buslov, Jakub Kicinski, Saeed Mahameed, David S. Miller,
Linux Netdev List, Mark Bloch, Saeed Mahameed
On Tue, Feb 09, 2021 at 06:10:59PM +0200, Or Gerlitz wrote:
> On Tue, Feb 9, 2021 at 4:26 PM Vlad Buslov <vladbu@nvidia.com> wrote:
> > On Mon 08 Feb 2021 at 22:22, Jakub Kicinski <kuba@kernel.org> wrote:
> > > On Mon, 8 Feb 2021 10:21:21 +0200 Vlad Buslov wrote:
>
> > >> > These operations imply that 7.7.7.5 is configured on some interface on
> > >> > the host. Most likely the VF representor itself, as that aids with ARP
> > >> > resolution. Is that so?
>
> > >> The tunnel endpoint IP address is configured on VF that is represented
> > >> by enp8s0f0_0 representor in example rules. The VF is on host.
>
> > > This is very confusing, are you saying that the 7.7.7.5 is configured
> > > both on VF and VFrep? Could you provide a full picture of the config
> > > with IP addresses and routing?
>
> > No, tunnel IP is configured on VF. That particular VF is in host [..]
>
> What's the motivation for that? isn't that introducing 3x slow down?
Vlad please correct me if I'm wrong.
I think this boils down to not using the uplink representor as a real
interface. This way, the host can make use of 7.7.7.5 for other stuff
as well without passing (heavy) traffic through representor ports,
which are not meant for it.
So the host can have the IP 7.7.7.5 and also decapsulate vxlan traffic
on it, which wouldn't be possible/recommended otherwise.
Another moment that this gets visible is with VF LAG. When we bond the
uplink representors, add an IP to it and do vxlan decap, that IP is
meant only for the decap process and shouldn't be used for heavier
traffic as its passing through representor ports.
Then, tc config for decap need to be done on VF0rep and not on VF0
itself because that would be a security problem: one VF (which could
be on a netns) could steer packets to another VF at will.
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-10 13:56 ` Marcelo Ricardo Leitner
@ 2021-02-10 16:44 ` Vlad Buslov
0 siblings, 0 replies; 39+ messages in thread
From: Vlad Buslov @ 2021-02-10 16:44 UTC (permalink / raw)
To: Marcelo Ricardo Leitner
Cc: Or Gerlitz, Jakub Kicinski, Saeed Mahameed, David S. Miller,
Linux Netdev List, Mark Bloch, Saeed Mahameed
On Wed 10 Feb 2021 at 15:56, Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> wrote:
> On Tue, Feb 09, 2021 at 06:10:59PM +0200, Or Gerlitz wrote:
>> On Tue, Feb 9, 2021 at 4:26 PM Vlad Buslov <vladbu@nvidia.com> wrote:
>> > On Mon 08 Feb 2021 at 22:22, Jakub Kicinski <kuba@kernel.org> wrote:
>> > > On Mon, 8 Feb 2021 10:21:21 +0200 Vlad Buslov wrote:
>>
>> > >> > These operations imply that 7.7.7.5 is configured on some interface on
>> > >> > the host. Most likely the VF representor itself, as that aids with ARP
>> > >> > resolution. Is that so?
>>
>> > >> The tunnel endpoint IP address is configured on VF that is represented
>> > >> by enp8s0f0_0 representor in example rules. The VF is on host.
>>
>> > > This is very confusing, are you saying that the 7.7.7.5 is configured
>> > > both on VF and VFrep? Could you provide a full picture of the config
>> > > with IP addresses and routing?
>>
>> > No, tunnel IP is configured on VF. That particular VF is in host [..]
>>
>> What's the motivation for that? isn't that introducing 3x slow down?
>
> Vlad please correct me if I'm wrong.
>
> I think this boils down to not using the uplink representor as a real
> interface. This way, the host can make use of 7.7.7.5 for other stuff
> as well without passing (heavy) traffic through representor ports,
> which are not meant for it.
>
> So the host can have the IP 7.7.7.5 and also decapsulate vxlan traffic
> on it, which wouldn't be possible/recommended otherwise.
>
> Another moment that this gets visible is with VF LAG. When we bond the
> uplink representors, add an IP to it and do vxlan decap, that IP is
> meant only for the decap process and shouldn't be used for heavier
> traffic as its passing through representor ports.
>
> Then, tc config for decap need to be done on VF0rep and not on VF0
> itself because that would be a security problem: one VF (which could
> be on a netns) could steer packets to another VF at will.
While on-host VF (the one with IP 7.7.7.5 in my examples) is intended to
be used for unencapsulated control traffic as well, we don't expect
significant bandwidth of such traffic, so traffic-load on representor
wasn't the main motivation. I didn't want to go into the details in
cover letter because they are mostly OVS-specific and this series is a
groundwork for features to come.
So the main motivation is to be able to apply policy on both on underlay
network (UL) and overlay network (tunnel netdev). As that will allow us
to subject overlay and underlay traffic to different set of OVS rules,
for example underlay traffic may be subject to vlan encap/decap,
security policy or any other flow rule that the user may define.
Hope this also answers some of Or's questions from this thread.
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-09 14:22 ` Vlad Buslov
2021-02-09 16:10 ` Or Gerlitz
@ 2021-02-09 18:05 ` Jakub Kicinski
2021-02-09 19:17 ` Vlad Buslov
1 sibling, 1 reply; 39+ messages in thread
From: Jakub Kicinski @ 2021-02-09 18:05 UTC (permalink / raw)
To: Vlad Buslov
Cc: Marcelo Ricardo Leitner, Saeed Mahameed, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed, Or Gerlitz
On Tue, 9 Feb 2021 16:22:26 +0200 Vlad Buslov wrote:
> No, tunnel IP is configured on VF. That particular VF is in host
> namespace. When mlx5 resolves tunneling the code checks if tunnel
> endpoint IP address is on such mlx5 VF, since the VF is in same
> namespace as eswitch manager (e.g. on host) and route returned by
> ip_route_output_key() is resolved through rt->dst.dev==tunVF device.
> After establishing that tunnel is on VF the goal is to process two
> resulting TC rules (in both directions) fully in hardware without
> exposing the packet on tunneling device or tunnel VF in sw, which is
> implemented with all the infrastructure from this series.
>
> So, to summarize with IP addresses from TC examples presented in cover letter,
> we have underlay network 7.7.7.0/24 in host namespace with tunnel endpoint IP
> address on VF:
>
> $ ip a show dev enp8s0f0v0
> 1537: enp8s0f0v0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
> link/ether 52:e5:6d:f2:00:69 brd ff:ff:ff:ff:ff:ff
> altname enp8s0f0np0v0
> inet 7.7.7.5/24 scope global enp8s0f0v0
> valid_lft forever preferred_lft forever
> inet6 fe80::50e5:6dff:fef2:69/64 scope link
> valid_lft forever preferred_lft forever
Isn't this 100% the wrong way around. Disable the offloads. Does the
traffic hit the VF encapsulated?
IIUC SW will do this:
PHY port
|
device | ,-----.
-----------|------------|-------|----------
kernel | | |
(UL/PF) (VFr) (VF)
| | |
[TC ing]>redir -` V
And the packet never hits encap.
^ permalink raw reply [flat|nested] 39+ messages in thread* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-09 18:05 ` Jakub Kicinski
@ 2021-02-09 19:17 ` Vlad Buslov
2021-02-09 19:50 ` Jakub Kicinski
0 siblings, 1 reply; 39+ messages in thread
From: Vlad Buslov @ 2021-02-09 19:17 UTC (permalink / raw)
To: Jakub Kicinski
Cc: Marcelo Ricardo Leitner, Saeed Mahameed, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed, Or Gerlitz
On Tue 09 Feb 2021 at 20:05, Jakub Kicinski <kuba@kernel.org> wrote:
> On Tue, 9 Feb 2021 16:22:26 +0200 Vlad Buslov wrote:
>> No, tunnel IP is configured on VF. That particular VF is in host
>> namespace. When mlx5 resolves tunneling the code checks if tunnel
>> endpoint IP address is on such mlx5 VF, since the VF is in same
>> namespace as eswitch manager (e.g. on host) and route returned by
>> ip_route_output_key() is resolved through rt->dst.dev==tunVF device.
>> After establishing that tunnel is on VF the goal is to process two
>> resulting TC rules (in both directions) fully in hardware without
>> exposing the packet on tunneling device or tunnel VF in sw, which is
>> implemented with all the infrastructure from this series.
>>
>> So, to summarize with IP addresses from TC examples presented in cover letter,
>> we have underlay network 7.7.7.0/24 in host namespace with tunnel endpoint IP
>> address on VF:
>>
>> $ ip a show dev enp8s0f0v0
>> 1537: enp8s0f0v0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
>> link/ether 52:e5:6d:f2:00:69 brd ff:ff:ff:ff:ff:ff
>> altname enp8s0f0np0v0
>> inet 7.7.7.5/24 scope global enp8s0f0v0
>> valid_lft forever preferred_lft forever
>> inet6 fe80::50e5:6dff:fef2:69/64 scope link
>> valid_lft forever preferred_lft forever
>
> Isn't this 100% the wrong way around. Disable the offloads. Does the
> traffic hit the VF encapsulated?
>
> IIUC SW will do this:
>
> PHY port
> |
> device | ,-----.
> -----------|------------|-------|----------
> kernel | | |
> (UL/PF) (VFr) (VF)
> | | |
> [TC ing]>redir -` V
>
> And the packet never hits encap.
We can look at dumps on every stage (produced by running exactly the
same test with OVS option other_config:tc-policy=skip_hw):
1. Traffic arrives at UL with vxlan encapsulation
$ sudo tcpdump -ni enp8s0f0 -vvv -c 3
dropped privs to tcpdump
tcpdump: listening on enp8s0f0, link-type EN10MB (Ethernet), capture size 262144 bytes
21:01:28.619346 IP (tos 0x0, ttl 64, id 65187, offset 0, flags [none], proto UDP (17), length 102)
7.7.7.1.52277 > 7.7.7.5.vxlan: [udp sum ok] VXLAN, flags [I] (0x08), vni 98
IP (tos 0x0, ttl 64, id 43919, offset 0, flags [DF], proto TCP (6), length 52)
5.5.5.1.targus-getdata1 > 5.5.5.5.34538: Flags [.], cksum 0x467b (correct), seq 2194968387, ack 2680742983, win 24576, options [nop,nop,TS val 1092282319 ecr 348802330], length 0
21:01:28.619505 IP (tos 0x0, ttl 64, id 888, offset 0, flags [none], proto UDP (17), length 1500)
7.7.7.5.40092 > 7.7.7.1.vxlan: [no cksum] VXLAN, flags [I] (0x08), vni 98
IP (tos 0x0, ttl 64, id 6662, offset 0, flags [DF], proto TCP (6), length 1450)
5.5.5.5.34538 > 5.5.5.1.targus-getdata1: Flags [.], cksum 0x8025 (correct), seq 673837:675235, ack 0, win 502, options [nop,nop,TS val 348802333 ecr 1092282319], length 1398
21:01:28.619506 IP (tos 0x0, ttl 64, id 889, offset 0, flags [none], proto UDP (17), length 1500)
7.7.7.5.40092 > 7.7.7.1.vxlan: [no cksum] VXLAN, flags [I] (0x08), vni 98
IP (tos 0x0, ttl 64, id 6663, offset 0, flags [DF], proto TCP (6), length 1450)
5.5.5.5.34538 > 5.5.5.1.targus-getdata1: Flags [.], cksum 0x19d1 (correct), seq 675235:676633, ack 0, win 502, options [nop,nop,TS val 348802333 ecr 1092282319], length 1398
2. By TC rule traffic is redirected to tunnel VF that has IP address
7.7.7.5 (still encapsulated as there is no decap action attached to
filter on enp8s0f0):
$ sudo tcpdump -ni enp8s0f0v0 -vvv -c 3
dropped privs to tcpdump
tcpdump: listening on enp8s0f0v0, link-type EN10MB (Ethernet), capture size 262144 bytes
21:03:41.524244 IP (tos 0x0, ttl 64, id 48184, offset 0, flags [none], proto UDP (17), length 1500)
7.7.7.5.40092 > 7.7.7.1.vxlan: [no cksum] VXLAN, flags [I] (0x08), vni 98
IP (tos 0x0, ttl 64, id 52619, offset 0, flags [DF], proto TCP (6), length 1450)
5.5.5.5.34538 > 5.5.5.1.targus-getdata1: Flags [.], cksum 0xaddb (correct), seq 279895999:279897397, ack 2194968387, win 502, options [nop,nop,TS val 348935238 ecr 1092415214], length 1398
21:03:41.568055 IP (tos 0x0, ttl 64, id 701, offset 0, flags [none], proto UDP (17), length 102)
7.7.7.1.52277 > 7.7.7.5.vxlan: [udp sum ok] VXLAN, flags [I] (0x08), vni 98
IP (tos 0x0, ttl 64, id 44938, offset 0, flags [DF], proto TCP (6), length 52)
5.5.5.1.targus-getdata1 > 5.5.5.5.34538: Flags [.], cksum 0xc623 (correct), seq 1, ack 1398, win 24576, options [nop,nop,TS val 1092415267 ecr 348935238], length 0
21:03:41.568384 IP (tos 0x0, ttl 64, id 48191, offset 0, flags [none], proto UDP (17), length 1500)
7.7.7.5.40092 > 7.7.7.1.vxlan: [no cksum] VXLAN, flags [I] (0x08), vni 98
IP (tos 0x0, ttl 64, id 52620, offset 0, flags [DF], proto TCP (6), length 1450)
5.5.5.5.34538 > 5.5.5.1.targus-getdata1: Flags [.], cksum 0xe1b9 (correct), seq 1398:2796, ack 1, win 502, options [nop,nop,TS val 348935282 ecr 1092415267], length 1398
3. Traffic gets to tunnel device, where it gets decapsulated and
redirected to destination VF by TC rule on vxlan_sys_4789:
$ sudo tcpdump -ni vxlan_sys_4789 -vvv -c 3
dropped privs to tcpdump
tcpdump: listening on vxlan_sys_4789, link-type EN10MB (Ethernet), capture size 262144 bytes
21:07:39.836141 IP (tos 0x0, ttl 64, id 15565, offset 0, flags [DF], proto TCP (6), length 52)
5.5.5.1.targus-getdata1 > 5.5.5.5.34538: Flags [.], cksum 0xbe91 (correct), seq 2194968387, ack 4279285947, win 24576, options [nop,nop,TS val 1092653536 ecr 349173547], length 0
21:07:39.836202 IP (tos 0x0, ttl 64, id 50774, offset 0, flags [DF], proto TCP (6), length 64360)
5.5.5.5.34538 > 5.5.5.1.targus-getdata1: Flags [P.], cksum 0x0f6b (incorrect -> 0x1d69), seq 746533:810841, ack 0, win 502, options [nop,nop,TS val 349173550 ecr 1092653536], length 64308
21:07:39.836449 IP (tos 0x0, ttl 64, id 15566, offset 0, flags [DF], proto TCP (6), length 52)
5.5.5.1.targus-getdata1 > 5.5.5.5.34538: Flags [.], cksum 0x610f (correct), seq 0, ack 89473, win 24576, options [nop,nop,TS val 1092653536 ecr 349173548], length 0
4. Decapsulated payload appears on namespaced VF with IP address
5.5.5.5:
$ sudo ip netns exec ns0 tcpdump -ni enp8s0f0v1 -vvv -c 3
yp_bind_client_create_v3: RPC: Unable to send
dropped privs to tcpdump
tcpdump: listening on enp8s0f0v1, link-type EN10MB (Ethernet), capture size 262144 bytes
21:09:06.758107 IP (tos 0x0, ttl 64, id 27527, offset 0, flags [DF], proto TCP (6), length 32206)
5.5.5.5.34538 > 5.5.5.1.targus-getdata1: Flags [P.], cksum 0x91d0 (incorrect -> 0x2a2a), seq 1198920825:1198952979, ack 2194968387, win 502, options [nop,nop,TS val 349260472 ecr 1092740448], length 32154
21:09:06.758697 IP (tos 0x0, ttl 64, id 3008, offset 0, flags [DF], proto TCP (6), length 64)
5.5.5.1.targus-getdata1 > 5.5.5.5.34538: Flags [.], cksum 0x6a1a (correct), seq 1, ack 4294942132, win 24576, options [nop,nop,TS val 1092740458 ecr 349260463,nop,nop,sack 1 {0:32154}], length 0
21:09:06.758748 IP (tos 0x0, ttl 64, id 27550, offset 0, flags [DF], proto TCP (6), length 25216)
5.5.5.5.34538 > 5.5.5.1.targus-getdata1: Flags [P.], cksum 0x7682 (incorrect -> 0x7627), seq 4294942132:0, ack 1, win 502, options [nop,nop,TS val 349260473 ecr 1092740458], length 25164
As you can see from the dump Tx is symmetrical. And that is exactly the
behavior we are reproducing with offloads. So I guess correct diagram
would be:
PHY port
|
device | ,(vxlan)
-----------|------------|-------|----------
kernel | | |
(UL/PF) (VFr) (VF)
| | |
[TC ing]>redir -` V
Regards,
Vlad
^ permalink raw reply [flat|nested] 39+ messages in thread* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-09 19:17 ` Vlad Buslov
@ 2021-02-09 19:50 ` Jakub Kicinski
2021-02-10 11:25 ` Vlad Buslov
0 siblings, 1 reply; 39+ messages in thread
From: Jakub Kicinski @ 2021-02-09 19:50 UTC (permalink / raw)
To: Vlad Buslov
Cc: Marcelo Ricardo Leitner, Saeed Mahameed, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed, Or Gerlitz
On Tue, 9 Feb 2021 21:17:11 +0200 Vlad Buslov wrote:
> 4. Decapsulated payload appears on namespaced VF with IP address
> 5.5.5.5:
>
> $ sudo ip netns exec ns0 tcpdump -ni enp8s0f0v1 -vvv -c 3
So there are two VFs? Hm, completely missed that. Could you *please*
provide an ascii diagram for the entire flow? None of those dumps
you're showing gives us the high level picture, and it's quite hard
to follow which enpsfyxz interface is what.
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-09 19:50 ` Jakub Kicinski
@ 2021-02-10 11:25 ` Vlad Buslov
2021-02-10 19:43 ` Jakub Kicinski
0 siblings, 1 reply; 39+ messages in thread
From: Vlad Buslov @ 2021-02-10 11:25 UTC (permalink / raw)
To: Jakub Kicinski
Cc: Marcelo Ricardo Leitner, Saeed Mahameed, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed, Or Gerlitz
On Tue 09 Feb 2021 at 21:50, Jakub Kicinski <kuba@kernel.org> wrote:
> On Tue, 9 Feb 2021 21:17:11 +0200 Vlad Buslov wrote:
>> 4. Decapsulated payload appears on namespaced VF with IP address
>> 5.5.5.5:
>>
>> $ sudo ip netns exec ns0 tcpdump -ni enp8s0f0v1 -vvv -c 3
>
> So there are two VFs? Hm, completely missed that. Could you *please*
> provide an ascii diagram for the entire flow? None of those dumps
> you're showing gives us the high level picture, and it's quite hard
> to follow which enpsfyxz interface is what.
Sure. Here it is:
+-------------------------------------------------------------------------------------+
| |
| OVS br TC ingress TC ingress |
| +---------------------+ +---------------------+ |
| | TC ingress | | TC ingress | |
| | +-------------+ | | +-------------+ | |
| | | | | | | | | |
| +------v---+---+ +---v---+------+ +------v---+---+ +---v---+------+ |
+---+ +-----+ +-----+ +-----+ +---+
| UL rep | | VF0 rep | | vxlan | | VF1 rep |
+---+ +-----+ +-----+ +-----+ +---+
| +-------^------+ +-^------------+ +-----^--------+ +-------------^+ |
| | | | | |
| Kernel | | | | |
| | | +---------------+ +--------------------+ | |
| | | | |namespace | | |
| | | | | | | |
| | | +------v-------+ | +--------------+ | | |
+----------------------------+ +-----------------+ +--------+
| | | VF0 | | | VF1 | | |
| | | | | | | | |
| | +----^---------+ | +----------^---+ | |
| | | | | | |
| | | +--------------------+ |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
+-------------------------------------------------------------------------------------+
| | | | | | |
| Hardware | +------+ +-------+ |
| | |
| +-----+ |
+--------+ | +----------------------------------------------------------------------+
| v |
| |
+-----+
^ permalink raw reply [flat|nested] 39+ messages in thread* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-10 11:25 ` Vlad Buslov
@ 2021-02-10 19:43 ` Jakub Kicinski
0 siblings, 0 replies; 39+ messages in thread
From: Jakub Kicinski @ 2021-02-10 19:43 UTC (permalink / raw)
To: Vlad Buslov
Cc: Marcelo Ricardo Leitner, Saeed Mahameed, David S. Miller, netdev,
Mark Bloch, Saeed Mahameed, Or Gerlitz
On Wed, 10 Feb 2021 13:25:05 +0200 Vlad Buslov wrote:
> On Tue 09 Feb 2021 at 21:50, Jakub Kicinski <kuba@kernel.org> wrote:
> > On Tue, 9 Feb 2021 21:17:11 +0200 Vlad Buslov wrote:
> >> 4. Decapsulated payload appears on namespaced VF with IP address
> >> 5.5.5.5:
> >>
> >> $ sudo ip netns exec ns0 tcpdump -ni enp8s0f0v1 -vvv -c 3
> >
> > So there are two VFs? Hm, completely missed that. Could you *please*
> > provide an ascii diagram for the entire flow? None of those dumps
> > you're showing gives us the high level picture, and it's quite hard
> > to follow which enpsfyxz interface is what.
>
> Sure. Here it is:
Thanks a lot, that clarifies it!
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port
2021-02-06 5:02 ` [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port Saeed Mahameed
2021-02-06 18:13 ` Marcelo Ricardo Leitner
@ 2021-02-09 0:20 ` patchwork-bot+netdevbpf
1 sibling, 0 replies; 39+ messages in thread
From: patchwork-bot+netdevbpf @ 2021-02-09 0:20 UTC (permalink / raw)
To: Saeed Mahameed; +Cc: kuba, davem, netdev, mbloch, saeedm
Hello:
This series was applied to netdev/net-next.git (refs/heads/master):
On Fri, 5 Feb 2021 21:02:24 -0800 you wrote:
> From: Mark Bloch <mbloch@nvidia.com>
>
> Setting the source port requires only the E-Switch and vport number.
> Refactor the function to get those parameters instead of passing the full
> attribute.
>
> Signed-off-by: Mark Bloch <mbloch@nvidia.com>
> Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
>
> [...]
Here is the summary with links:
- [net-next,V2,01/17] net/mlx5: E-Switch, Refactor setting source port
https://git.kernel.org/netdev/net-next/c/b055ecf5827d
- [net-next,V2,02/17] net/mlx5e: E-Switch, Maintain vhca_id to vport_num mapping
https://git.kernel.org/netdev/net-next/c/84ae9c1f29c0
- [net-next,V2,03/17] net/mlx5e: Always set attr mdev pointer
https://git.kernel.org/netdev/net-next/c/275c21d6cbe2
- [net-next,V2,04/17] net/mlx5: E-Switch, Refactor rule offload forward action processing
https://git.kernel.org/netdev/net-next/c/9e51c0a62492
- [net-next,V2,05/17] net/mlx5e: VF tunnel TX traffic offloading
https://git.kernel.org/netdev/net-next/c/10742efc20a4
- [net-next,V2,06/17] net/mlx5e: Refactor tun routing helpers
https://git.kernel.org/netdev/net-next/c/6717986e15a0
- [net-next,V2,07/17] net/mlx5: E-Switch, Indirect table infrastructure
https://git.kernel.org/netdev/net-next/c/34ca65352ddf
- [net-next,V2,08/17] net/mlx5e: Remove redundant match on tunnel destination mac
https://git.kernel.org/netdev/net-next/c/4ad9116c84ed
- [net-next,V2,09/17] net/mlx5e: VF tunnel RX traffic offloading
https://git.kernel.org/netdev/net-next/c/a508728a4c8b
- [net-next,V2,10/17] net/mlx5e: Refactor reg_c1 usage
https://git.kernel.org/netdev/net-next/c/48d216e5596a
- [net-next,V2,11/17] net/mlx5e: Match recirculated packet miss in slow table using reg_c1
https://git.kernel.org/netdev/net-next/c/8e404fefa58b
- [net-next,V2,12/17] net/mlx5e: Extract tc tunnel encap/decap code to dedicated file
https://git.kernel.org/netdev/net-next/c/0d9f96471493
- [net-next,V2,13/17] net/mlx5e: Create route entry infrastructure
https://git.kernel.org/netdev/net-next/c/777bb800c696
- [net-next,V2,14/17] net/mlx5e: Refactor neigh update infrastructure
https://git.kernel.org/netdev/net-next/c/2221d954d984
- [net-next,V2,15/17] net/mlx5e: TC preparation refactoring for routing update event
https://git.kernel.org/netdev/net-next/c/c7b9038d8af6
- [net-next,V2,16/17] net/mlx5e: Rename some encap-specific API to generic names
https://git.kernel.org/netdev/net-next/c/021905f8067d
- [net-next,V2,17/17] net/mlx5e: Handle FIB events to update tunnel endpoint device
https://git.kernel.org/netdev/net-next/c/8914add2c9e5
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
^ permalink raw reply [flat|nested] 39+ messages in thread
* [net-next V2 02/17] net/mlx5e: E-Switch, Maintain vhca_id to vport_num mapping
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 03/17] net/mlx5e: Always set attr mdev pointer Saeed Mahameed
` (16 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Following patches in the series need to be able to map VF netdev to vport.
Since it is trivial to obtain vhca_id from netdev, maintain mapping from
vhca_id to vport_num inside eswitch offloads using xarray. Provide function
mlx5_eswitch_vhca_id_to_vport() to be used by TC code in following patches
to obtain the mapping.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../net/ethernet/mellanox/mlx5/core/eswitch.c | 20 +++++
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 6 ++
.../mellanox/mlx5/core/eswitch_offloads.c | 79 +++++++++++++++++++
.../ethernet/mellanox/mlx5/core/mlx5_core.h | 2 +
.../net/ethernet/mellanox/mlx5/core/vport.c | 12 +++
5 files changed, 119 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 820305b1664e..aba17835465b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1300,6 +1300,13 @@ int mlx5_esw_vport_enable(struct mlx5_eswitch *esw, u16 vport_num,
(!vport_num && mlx5_core_is_ecpf(esw->dev)))
vport->info.trusted = true;
+ if (!mlx5_esw_is_manager_vport(esw, vport->vport) &&
+ MLX5_CAP_GEN(esw->dev, vhca_resource_manager)) {
+ ret = mlx5_esw_vport_vhca_id_set(esw, vport_num);
+ if (ret)
+ goto err_vhca_mapping;
+ }
+
esw_vport_change_handle_locked(vport);
esw->enabled_vports++;
@@ -1307,6 +1314,11 @@ int mlx5_esw_vport_enable(struct mlx5_eswitch *esw, u16 vport_num,
done:
mutex_unlock(&esw->state_lock);
return ret;
+
+err_vhca_mapping:
+ esw_vport_cleanup(esw, vport);
+ mutex_unlock(&esw->state_lock);
+ return ret;
}
void mlx5_esw_vport_disable(struct mlx5_eswitch *esw, u16 vport_num)
@@ -1325,6 +1337,11 @@ void mlx5_esw_vport_disable(struct mlx5_eswitch *esw, u16 vport_num)
/* Disable events from this vport */
arm_vport_context_events_cmd(esw->dev, vport->vport, 0);
+
+ if (!mlx5_esw_is_manager_vport(esw, vport->vport) &&
+ MLX5_CAP_GEN(esw->dev, vhca_resource_manager))
+ mlx5_esw_vport_vhca_id_clear(esw, vport_num);
+
/* We don't assume VFs will cleanup after themselves.
* Calling vport change handler while vport is disabled will cleanup
* the vport resources.
@@ -1815,6 +1832,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
mlx5e_mod_hdr_tbl_init(&esw->offloads.mod_hdr);
atomic64_set(&esw->offloads.num_flows, 0);
ida_init(&esw->offloads.vport_metadata_ida);
+ xa_init_flags(&esw->offloads.vhca_map, XA_FLAGS_ALLOC);
mutex_init(&esw->state_lock);
mutex_init(&esw->mode_lock);
@@ -1854,6 +1872,8 @@ void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw)
esw_offloads_cleanup_reps(esw);
mutex_destroy(&esw->mode_lock);
mutex_destroy(&esw->state_lock);
+ WARN_ON(!xa_empty(&esw->offloads.vhca_map));
+ xa_destroy(&esw->offloads.vhca_map);
ida_destroy(&esw->offloads.vport_metadata_ida);
mlx5e_mod_hdr_tbl_destroy(&esw->offloads.mod_hdr);
mutex_destroy(&esw->offloads.encap_tbl_lock);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index 479d2ac2cd85..1a045e95bc68 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -36,6 +36,7 @@
#include <linux/if_ether.h>
#include <linux/if_link.h>
#include <linux/atomic.h>
+#include <linux/xarray.h>
#include <net/devlink.h>
#include <linux/mlx5/device.h>
#include <linux/mlx5/eswitch.h>
@@ -212,6 +213,7 @@ struct mlx5_esw_offload {
struct mod_hdr_tbl mod_hdr;
DECLARE_HASHTABLE(termtbl_tbl, 8);
struct mutex termtbl_mutex; /* protects termtbl hash */
+ struct xarray vhca_map;
const struct mlx5_eswitch_rep_ops *rep_ops[NUM_REP_TYPES];
u8 inline_mode;
atomic64_t num_flows;
@@ -734,6 +736,10 @@ int mlx5_esw_offloads_sf_vport_enable(struct mlx5_eswitch *esw, struct devlink_p
u16 vport_num, u32 sfnum);
void mlx5_esw_offloads_sf_vport_disable(struct mlx5_eswitch *esw, u16 vport_num);
+int mlx5_esw_vport_vhca_id_set(struct mlx5_eswitch *esw, u16 vport_num);
+void mlx5_esw_vport_vhca_id_clear(struct mlx5_eswitch *esw, u16 vport_num);
+int mlx5_eswitch_vhca_id_to_vport(struct mlx5_eswitch *esw, u16 vhca_id, u16 *vport_num);
+
/**
* mlx5_esw_event_info - Indicates eswitch mode changed/changing.
*
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 416ede2fe5d7..3085bdd14fbb 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -2872,3 +2872,82 @@ void mlx5_esw_offloads_sf_vport_disable(struct mlx5_eswitch *esw, u16 vport_num)
mlx5_esw_devlink_sf_port_unregister(esw, vport_num);
mlx5_esw_vport_disable(esw, vport_num);
}
+
+static int mlx5_esw_query_vport_vhca_id(struct mlx5_eswitch *esw, u16 vport_num, u16 *vhca_id)
+{
+ int query_out_sz = MLX5_ST_SZ_BYTES(query_hca_cap_out);
+ void *query_ctx;
+ void *hca_caps;
+ int err;
+
+ *vhca_id = 0;
+ if (mlx5_esw_is_manager_vport(esw, vport_num) ||
+ !MLX5_CAP_GEN(esw->dev, vhca_resource_manager))
+ return -EPERM;
+
+ query_ctx = kzalloc(query_out_sz, GFP_KERNEL);
+ if (!query_ctx)
+ return -ENOMEM;
+
+ err = mlx5_vport_get_other_func_cap(esw->dev, vport_num, query_ctx);
+ if (err)
+ goto out_free;
+
+ hca_caps = MLX5_ADDR_OF(query_hca_cap_out, query_ctx, capability);
+ *vhca_id = MLX5_GET(cmd_hca_cap, hca_caps, vhca_id);
+
+out_free:
+ kfree(query_ctx);
+ return err;
+}
+
+int mlx5_esw_vport_vhca_id_set(struct mlx5_eswitch *esw, u16 vport_num)
+{
+ u16 *old_entry, *vhca_map_entry, vhca_id;
+ int err;
+
+ err = mlx5_esw_query_vport_vhca_id(esw, vport_num, &vhca_id);
+ if (err) {
+ esw_warn(esw->dev, "Getting vhca_id for vport failed (vport=%u,err=%d)\n",
+ vport_num, err);
+ return err;
+ }
+
+ vhca_map_entry = kmalloc(sizeof(*vhca_map_entry), GFP_KERNEL);
+ if (!vhca_map_entry)
+ return -ENOMEM;
+
+ *vhca_map_entry = vport_num;
+ old_entry = xa_store(&esw->offloads.vhca_map, vhca_id, vhca_map_entry, GFP_KERNEL);
+ if (xa_is_err(old_entry)) {
+ kfree(vhca_map_entry);
+ return xa_err(old_entry);
+ }
+ kfree(old_entry);
+ return 0;
+}
+
+void mlx5_esw_vport_vhca_id_clear(struct mlx5_eswitch *esw, u16 vport_num)
+{
+ u16 *vhca_map_entry, vhca_id;
+ int err;
+
+ err = mlx5_esw_query_vport_vhca_id(esw, vport_num, &vhca_id);
+ if (err)
+ esw_warn(esw->dev, "Getting vhca_id for vport failed (vport=%hu,err=%d)\n",
+ vport_num, err);
+
+ vhca_map_entry = xa_erase(&esw->offloads.vhca_map, vhca_id);
+ kfree(vhca_map_entry);
+}
+
+int mlx5_eswitch_vhca_id_to_vport(struct mlx5_eswitch *esw, u16 vhca_id, u16 *vport_num)
+{
+ u16 *res = xa_load(&esw->offloads.vhca_map, vhca_id);
+
+ if (!res)
+ return -ENOENT;
+
+ *vport_num = *res;
+ return 0;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
index 3754ef98554f..efe403c7e354 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -270,5 +270,7 @@ void mlx5_mdev_uninit(struct mlx5_core_dev *dev);
void mlx5_unload_one(struct mlx5_core_dev *dev, bool cleanup);
int mlx5_load_one(struct mlx5_core_dev *dev, bool boot);
+int mlx5_vport_get_other_func_cap(struct mlx5_core_dev *dev, u16 function_id, void *out);
+
void mlx5_events_work_enqueue(struct mlx5_core_dev *dev, struct work_struct *work);
#endif /* __MLX5_CORE_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/vport.c b/drivers/net/ethernet/mellanox/mlx5/core/vport.c
index ba78e0660523..e05c5c0f3ae1 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/vport.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/vport.c
@@ -1164,3 +1164,15 @@ u16 mlx5_eswitch_get_total_vports(const struct mlx5_core_dev *dev)
return MLX5_SPECIAL_VPORTS(dev) + mlx5_core_max_vfs(dev) + mlx5_sf_max_functions(dev);
}
EXPORT_SYMBOL_GPL(mlx5_eswitch_get_total_vports);
+
+int mlx5_vport_get_other_func_cap(struct mlx5_core_dev *dev, u16 function_id, void *out)
+{
+ u16 opmod = (MLX5_CAP_GENERAL << 1) | (HCA_CAP_OPMOD_GET_MAX & 0x01);
+ u8 in[MLX5_ST_SZ_BYTES(query_hca_cap_in)] = {};
+
+ MLX5_SET(query_hca_cap_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_CAP);
+ MLX5_SET(query_hca_cap_in, in, op_mod, opmod);
+ MLX5_SET(query_hca_cap_in, in, function_id, function_id);
+ MLX5_SET(query_hca_cap_in, in, other_function, true);
+ return mlx5_cmd_exec_inout(dev, query_hca_cap, in, out);
+}
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 03/17] net/mlx5e: Always set attr mdev pointer
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 01/17] net/mlx5: E-Switch, Refactor setting source port Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 02/17] net/mlx5e: E-Switch, Maintain vhca_id to vport_num mapping Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 04/17] net/mlx5: E-Switch, Refactor rule offload forward action processing Saeed Mahameed
` (15 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Eswitch offloads extensions in following patches in the series require
attr->esw_attr->in_mdev pointer to always be set. This is already the case
for all code paths except mlx5_tc_ct_entry_add_rule() function. Fix the
function to assign mdev pointer with priv->mdev value.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
index 40aaa105b2fc..3fb75dcdc68d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
@@ -711,6 +711,8 @@ mlx5_tc_ct_entry_add_rule(struct mlx5_tc_ct_priv *ct_priv,
attr->outer_match_level = MLX5_MATCH_L4;
attr->counter = entry->counter->counter;
attr->flags |= MLX5_ESW_ATTR_FLAG_NO_IN_PORT;
+ if (ct_priv->ns_type == MLX5_FLOW_NAMESPACE_FDB)
+ attr->esw_attr->in_mdev = priv->mdev;
mlx5_tc_ct_set_tuple_match(netdev_priv(ct_priv->netdev), spec, flow_rule);
mlx5e_tc_match_to_reg_match(spec, ZONE_TO_REG, entry->tuple.zone, MLX5_CT_ZONE_MASK);
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 04/17] net/mlx5: E-Switch, Refactor rule offload forward action processing
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (2 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 03/17] net/mlx5e: Always set attr mdev pointer Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 05/17] net/mlx5e: VF tunnel TX traffic offloading Saeed Mahameed
` (14 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Following patches in the series extend forwarding functionality with VF
tunnel TX and RX handling. Extract action forwarding processing code into
dedicated functions to simplify further extensions:
- Handle every forwarding case with dedicated function instead of inline
code.
- Extract forwarding dest dispatch conditional into helper function
esw_setup_dests().
- Unify forwaring cleanup code in error path of
mlx5_eswitch_add_offloaded_rule() and in rule deletion code of
__mlx5_eswitch_del_rule() in new helper function esw_cleanup_dests() (dual
to new esw_setup_dests() helper).
This patch does not change functionality.
Co-developed-by: Dmytro Linkin <dlinkin@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../mellanox/mlx5/core/eswitch_offloads.c | 189 ++++++++++++------
1 file changed, 129 insertions(+), 60 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 3085bdd14fbb..335dc83d1bb9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -296,6 +296,124 @@ mlx5_eswitch_set_rule_source_port(struct mlx5_eswitch *esw,
}
}
+static void
+esw_setup_ft_dest(struct mlx5_flow_destination *dest,
+ struct mlx5_flow_act *flow_act,
+ struct mlx5_flow_attr *attr,
+ int i)
+{
+ flow_act->flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
+ dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+ dest[i].ft = attr->dest_ft;
+}
+
+static void
+esw_setup_slow_path_dest(struct mlx5_flow_destination *dest,
+ struct mlx5_flow_act *flow_act,
+ struct mlx5_fs_chains *chains,
+ int i)
+{
+ flow_act->flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
+ dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+ dest[i].ft = mlx5_chains_get_tc_end_ft(chains);
+}
+
+static int
+esw_setup_chain_dest(struct mlx5_flow_destination *dest,
+ struct mlx5_flow_act *flow_act,
+ struct mlx5_fs_chains *chains,
+ u32 chain, u32 prio, u32 level,
+ int i)
+{
+ struct mlx5_flow_table *ft;
+
+ flow_act->flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
+ ft = mlx5_chains_get_table(chains, chain, prio, level);
+ if (IS_ERR(ft))
+ return PTR_ERR(ft);
+
+ dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+ dest[i].ft = ft;
+ return 0;
+}
+
+static void
+esw_cleanup_chain_dest(struct mlx5_fs_chains *chains, u32 chain, u32 prio, u32 level)
+{
+ mlx5_chains_put_table(chains, chain, prio, level);
+}
+
+static void
+esw_setup_vport_dest(struct mlx5_flow_destination *dest, struct mlx5_flow_act *flow_act,
+ struct mlx5_eswitch *esw, struct mlx5_esw_flow_attr *esw_attr,
+ int attr_idx, int dest_idx, bool pkt_reformat)
+{
+ dest[dest_idx].type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
+ dest[dest_idx].vport.num = esw_attr->dests[attr_idx].rep->vport;
+ dest[dest_idx].vport.vhca_id =
+ MLX5_CAP_GEN(esw_attr->dests[attr_idx].mdev, vhca_id);
+ if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
+ dest[dest_idx].vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID;
+ if (esw_attr->dests[attr_idx].flags & MLX5_ESW_DEST_ENCAP) {
+ if (pkt_reformat) {
+ flow_act->action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
+ flow_act->pkt_reformat = esw_attr->dests[attr_idx].pkt_reformat;
+ }
+ dest[dest_idx].vport.flags |= MLX5_FLOW_DEST_VPORT_REFORMAT_ID;
+ dest[dest_idx].vport.pkt_reformat = esw_attr->dests[attr_idx].pkt_reformat;
+ }
+}
+
+static int
+esw_setup_vport_dests(struct mlx5_flow_destination *dest, struct mlx5_flow_act *flow_act,
+ struct mlx5_eswitch *esw, struct mlx5_esw_flow_attr *esw_attr,
+ int i)
+{
+ int j;
+
+ for (j = esw_attr->split_count; j < esw_attr->out_count; j++, i++)
+ esw_setup_vport_dest(dest, flow_act, esw, esw_attr, j, i, true);
+ return i;
+}
+
+static int
+esw_setup_dests(struct mlx5_flow_destination *dest,
+ struct mlx5_flow_act *flow_act,
+ struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ int *i)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+ struct mlx5_fs_chains *chains = esw_chains(esw);
+ int err = 0;
+
+ if (attr->dest_ft) {
+ esw_setup_ft_dest(dest, flow_act, attr, *i);
+ (*i)++;
+ } else if (attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH) {
+ esw_setup_slow_path_dest(dest, flow_act, chains, *i);
+ (*i)++;
+ } else if (attr->dest_chain) {
+ err = esw_setup_chain_dest(dest, flow_act, chains, attr->dest_chain,
+ 1, 0, *i);
+ (*i)++;
+ } else {
+ *i = esw_setup_vport_dests(dest, flow_act, esw, esw_attr, *i);
+ }
+
+ return err;
+}
+
+static void
+esw_cleanup_dests(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr)
+{
+ struct mlx5_fs_chains *chains = esw_chains(esw);
+
+ if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH) && attr->dest_chain)
+ esw_cleanup_chain_dest(chains, attr->dest_chain, 1, 0);
+}
+
struct mlx5_flow_handle *
mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
struct mlx5_flow_spec *spec,
@@ -309,7 +427,7 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
struct mlx5_vport_tbl_attr fwd_attr;
struct mlx5_flow_handle *rule;
struct mlx5_flow_table *fdb;
- int j, i = 0;
+ int i = 0;
if (esw->mode != MLX5_ESWITCH_OFFLOADS)
return ERR_PTR(-EOPNOTSUPP);
@@ -331,49 +449,12 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
}
if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
- struct mlx5_flow_table *ft;
-
- if (attr->dest_ft) {
- flow_act.flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
- dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
- dest[i].ft = attr->dest_ft;
- i++;
- } else if (attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH) {
- flow_act.flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
- dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
- dest[i].ft = mlx5_chains_get_tc_end_ft(chains);
- i++;
- } else if (attr->dest_chain) {
- flow_act.flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
- ft = mlx5_chains_get_table(chains, attr->dest_chain,
- 1, 0);
- if (IS_ERR(ft)) {
- rule = ERR_CAST(ft);
- goto err_create_goto_table;
- }
-
- dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
- dest[i].ft = ft;
- i++;
- } else {
- for (j = esw_attr->split_count; j < esw_attr->out_count; j++) {
- dest[i].type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
- dest[i].vport.num = esw_attr->dests[j].rep->vport;
- dest[i].vport.vhca_id =
- MLX5_CAP_GEN(esw_attr->dests[j].mdev, vhca_id);
- if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
- dest[i].vport.flags |=
- MLX5_FLOW_DEST_VPORT_VHCA_ID;
- if (esw_attr->dests[j].flags & MLX5_ESW_DEST_ENCAP) {
- flow_act.action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
- flow_act.pkt_reformat =
- esw_attr->dests[j].pkt_reformat;
- dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_REFORMAT_ID;
- dest[i].vport.pkt_reformat =
- esw_attr->dests[j].pkt_reformat;
- }
- i++;
- }
+ int err;
+
+ err = esw_setup_dests(dest, &flow_act, esw, attr, &i);
+ if (err) {
+ rule = ERR_PTR(err);
+ goto err_create_goto_table;
}
}
@@ -437,8 +518,7 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
else if (attr->chain || attr->prio)
mlx5_chains_put_table(chains, attr->chain, attr->prio, 0);
err_esw_get:
- if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH) && attr->dest_chain)
- mlx5_chains_put_table(chains, attr->dest_chain, 1, 0);
+ esw_cleanup_dests(esw, attr);
err_create_goto_table:
return rule;
}
@@ -474,18 +554,8 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
}
flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
- for (i = 0; i < esw_attr->split_count; i++) {
- dest[i].type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
- dest[i].vport.num = esw_attr->dests[i].rep->vport;
- dest[i].vport.vhca_id =
- MLX5_CAP_GEN(esw_attr->dests[i].mdev, vhca_id);
- if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
- dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_VHCA_ID;
- if (esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP) {
- dest[i].vport.flags |= MLX5_FLOW_DEST_VPORT_REFORMAT_ID;
- dest[i].vport.pkt_reformat = esw_attr->dests[i].pkt_reformat;
- }
- }
+ for (i = 0; i < esw_attr->split_count; i++)
+ esw_setup_vport_dest(dest, &flow_act, esw, esw_attr, i, i, false);
dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
dest[i].ft = fwd_fdb;
i++;
@@ -552,8 +622,7 @@ __mlx5_eswitch_del_rule(struct mlx5_eswitch *esw,
esw_vport_tbl_put(esw, &fwd_attr);
else if (attr->chain || attr->prio)
mlx5_chains_put_table(chains, attr->chain, attr->prio, 0);
- if (attr->dest_chain)
- mlx5_chains_put_table(chains, attr->dest_chain, 1, 0);
+ esw_cleanup_dests(esw, attr);
}
}
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 05/17] net/mlx5e: VF tunnel TX traffic offloading
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (3 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 04/17] net/mlx5: E-Switch, Refactor rule offload forward action processing Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 06/17] net/mlx5e: Refactor tun routing helpers Saeed Mahameed
` (13 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
When tunnel endpoint is on VF, driver still assumes that endpoint is on
uplink and incorrectly configures encap rule offload according to that
assumption. As a result, traffic is sent directly to the uplink and rules
installed on representor of tunnel endpoint VF are ignored.
Implement following changes to allow offloading tx traffic with tunnel
endpoint on VF:
- For tunneling flows perform route lookup on route and out devices pair.
If out device is uplink and route device is VF of same physical port, then
modify packet reg_c_0 metadata register (source port) with the value of VF
vport. Use eswitch vhca_id->vport mapping introduced in one of previous
patches in the series to obtain vport from route netdevice.
- Recirculate encapsulated packets to VF vport in order to apply any flow
rules installed on VF representor that match on encapsulated traffic.
Only enable support for this functionality when all following conditions
are true:
- Hardware advertises capability to preserve reg_c_0 value on packet
recirculation.
- Vport metadata matching is enabled.
- Termination tables are to be used by the flow.
Example TC rules for VF tunnel traffic:
1. Rule that redirects packets from UL to VF rep that has the tunnel
endpoint IP address:
$ tc -s filter show dev enp8s0f0 ingress
filter protocol ip pref 4 flower chain 0
filter protocol ip pref 4 flower chain 0 handle 0x1
dst_mac 16:c9:a0:2d:69:2c
src_mac 0c:42:a1:58:ab:e4
eth_type ipv4
ip_flags nofrag
in_hw in_hw_count 1
action order 1: mirred (Egress Redirect to device enp8s0f0_0) stolen
index 3 ref 1 bind 1 installed 377 sec used 0 sec
Action statistics:
Sent 114096 bytes 952 pkt (dropped 0, overlimits 0 requeues 0)
Sent software 0 bytes 0 pkt
Sent hardware 114096 bytes 952 pkt
backlog 0b 0p requeues 0
cookie 878fa48d8c423fc08c3b6ca599b50a97
no_percpu
used_hw_stats delayed
2. Rule that decapsulates the tunneled flow and redirects to destination VF
representor:
$ tc -s filter show dev vxlan_sys_4789 ingress
filter protocol ip pref 4 flower chain 0
filter protocol ip pref 4 flower chain 0 handle 0x1
dst_mac ca:2e:a7:3f:f5:0f
src_mac 0a:40:bd:30:89:99
eth_type ipv4
enc_dst_ip 7.7.7.5
enc_src_ip 7.7.7.1
enc_key_id 98
enc_dst_port 4789
enc_tos 0
ip_flags nofrag
in_hw in_hw_count 1
action order 1: tunnel_key unset pipe
index 2 ref 1 bind 1 installed 434 sec used 434 sec
Action statistics:
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
used_hw_stats delayed
action order 2: mirred (Egress Redirect to device enp8s0f0_1) stolen
index 4 ref 1 bind 1 installed 434 sec used 0 sec
Action statistics:
Sent 129936 bytes 1082 pkt (dropped 0, overlimits 0 requeues 0)
Sent software 0 bytes 0 pkt
Sent hardware 129936 bytes 1082 pkt
backlog 0b 0p requeues 0
cookie ac17cf398c4c69e4a5b2f7aabd1b88ff
no_percpu
used_hw_stats delayed
Co-developed-by: Dmytro Linkin <dlinkin@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../net/ethernet/mellanox/mlx5/core/en_tc.c | 87 +++++++++++++
.../net/ethernet/mellanox/mlx5/core/en_tc.h | 1 +
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 2 +
.../mellanox/mlx5/core/eswitch_offloads.c | 120 ++++++++++++++++--
include/linux/mlx5/eswitch.h | 2 +
5 files changed, 201 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 280ea1e1e039..43f1508a05b5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -165,6 +165,11 @@ struct mlx5e_tc_attr_to_reg_mapping mlx5e_tc_attr_to_reg_mappings[] = {
.moffset = 0,
.mlen = 2,
},
+ [VPORT_TO_REG] = {
+ .mfield = MLX5_ACTION_IN_FIELD_METADATA_REG_C_0,
+ .moffset = 2,
+ .mlen = 2,
+ },
[TUNNEL_TO_REG] = {
.mfield = MLX5_ACTION_IN_FIELD_METADATA_REG_C_1,
.moffset = 1,
@@ -1315,6 +1320,44 @@ static void remove_unready_flow(struct mlx5e_tc_flow *flow)
mutex_unlock(&uplink_priv->unready_flows_lock);
}
+static bool same_hw_devs(struct mlx5e_priv *priv, struct mlx5e_priv *peer_priv);
+
+static bool mlx5e_tc_is_vf_tunnel(struct net_device *out_dev, struct net_device *route_dev)
+{
+ struct mlx5_core_dev *out_mdev, *route_mdev;
+ struct mlx5e_priv *out_priv, *route_priv;
+
+ out_priv = netdev_priv(out_dev);
+ out_mdev = out_priv->mdev;
+ route_priv = netdev_priv(route_dev);
+ route_mdev = route_priv->mdev;
+
+ if (out_mdev->coredev_type != MLX5_COREDEV_PF ||
+ route_mdev->coredev_type != MLX5_COREDEV_VF)
+ return false;
+
+ return same_hw_devs(out_priv, route_priv);
+}
+
+static int mlx5e_tc_query_route_vport(struct net_device *out_dev, struct net_device *route_dev,
+ u16 *vport)
+{
+ struct mlx5e_priv *out_priv, *route_priv;
+ struct mlx5_core_dev *route_mdev;
+ struct mlx5_eswitch *esw;
+ u16 vhca_id;
+ int err;
+
+ out_priv = netdev_priv(out_dev);
+ esw = out_priv->mdev->priv.eswitch;
+ route_priv = netdev_priv(route_dev);
+ route_mdev = route_priv->mdev;
+
+ vhca_id = MLX5_CAP_GEN(route_mdev, vhca_id);
+ err = mlx5_eswitch_vhca_id_to_vport(esw, vhca_id, vport);
+ return err;
+}
+
static int
mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow,
@@ -3700,6 +3743,45 @@ static bool is_duplicated_encap_entry(struct mlx5e_priv *priv,
return false;
}
+static int mlx5e_set_vf_tunnel(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
+ struct net_device *out_dev,
+ int route_dev_ifindex,
+ int out_index)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+ struct net_device *route_dev;
+ u16 vport_num;
+ int err = 0;
+ u32 data;
+
+ route_dev = dev_get_by_index(dev_net(out_dev), route_dev_ifindex);
+
+ if (!route_dev || route_dev->netdev_ops != &mlx5e_netdev_ops ||
+ !mlx5e_tc_is_vf_tunnel(out_dev, route_dev))
+ goto out;
+
+ err = mlx5e_tc_query_route_vport(out_dev, route_dev, &vport_num);
+ if (err)
+ goto out;
+
+ attr->dest_chain = 0;
+ attr->action |= MLX5_FLOW_CONTEXT_ACTION_MOD_HDR;
+ esw_attr->dests[out_index].flags |= MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE;
+ data = mlx5_eswitch_get_vport_metadata_for_set(esw_attr->in_mdev->priv.eswitch,
+ vport_num);
+ err = mlx5e_tc_match_to_reg_set(esw->dev, mod_hdr_acts,
+ MLX5_FLOW_NAMESPACE_FDB, VPORT_TO_REG, data);
+ if (err)
+ goto out;
+
+out:
+ if (route_dev)
+ dev_put(route_dev);
+ return err;
+}
+
static int mlx5e_attach_encap(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow,
struct net_device *mirred_dev,
@@ -3791,6 +3873,11 @@ static int mlx5e_attach_encap(struct mlx5e_priv *priv,
e->compl_result = 1;
attach_flow:
+ err = mlx5e_set_vf_tunnel(esw, attr, &parse_attr->mod_hdr_acts, e->out_dev,
+ e->route_dev_ifindex, out_index);
+ if (err)
+ goto out_err;
+
flow->encaps[out_index].e = e;
list_add(&flow->encaps[out_index].list, &e->flows);
flow->encaps[out_index].index = out_index;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
index 4a2ce241522e..56d809904ea7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -167,6 +167,7 @@ void mlx5e_tc_reoffload_flows_work(struct work_struct *work);
enum mlx5e_tc_attr_to_reg {
CHAIN_TO_REG,
+ VPORT_TO_REG,
TUNNEL_TO_REG,
CTSTATE_TO_REG,
ZONE_TO_REG,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index 1a045e95bc68..1ab34751329e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -389,12 +389,14 @@ enum mlx5_flow_match_level {
enum {
MLX5_ESW_DEST_ENCAP = BIT(0),
MLX5_ESW_DEST_ENCAP_VALID = BIT(1),
+ MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE = BIT(2),
};
enum {
MLX5_ESW_ATTR_FLAG_VLAN_HANDLED = BIT(0),
MLX5_ESW_ATTR_FLAG_SLOW_PATH = BIT(1),
MLX5_ESW_ATTR_FLAG_NO_IN_PORT = BIT(2),
+ MLX5_ESW_ATTR_FLAG_SRC_REWRITE = BIT(3),
};
struct mlx5_esw_flow_attr {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 335dc83d1bb9..1b18f624e04a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -337,6 +337,65 @@ esw_setup_chain_dest(struct mlx5_flow_destination *dest,
return 0;
}
+static void esw_put_dest_tables_loop(struct mlx5_eswitch *esw, struct mlx5_flow_attr *attr,
+ int from, int to)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+ struct mlx5_fs_chains *chains = esw_chains(esw);
+ int i;
+
+ for (i = from; i < to; i++)
+ if (esw_attr->dests[i].flags & MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE)
+ mlx5_chains_put_table(chains, 0, 1, 0);
+}
+
+static bool
+esw_is_chain_src_port_rewrite(struct mlx5_eswitch *esw, struct mlx5_esw_flow_attr *esw_attr)
+{
+ int i;
+
+ for (i = esw_attr->split_count; i < esw_attr->out_count; i++)
+ if (esw_attr->dests[i].flags & MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE)
+ return true;
+ return false;
+}
+
+static int
+esw_setup_chain_src_port_rewrite(struct mlx5_flow_destination *dest,
+ struct mlx5_flow_act *flow_act,
+ struct mlx5_eswitch *esw,
+ struct mlx5_fs_chains *chains,
+ struct mlx5_flow_attr *attr,
+ int *i)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+ int j, err;
+
+ if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SRC_REWRITE))
+ return -EOPNOTSUPP;
+
+ for (j = esw_attr->split_count; j < esw_attr->out_count; j++, (*i)++) {
+ err = esw_setup_chain_dest(dest, flow_act, chains, attr->dest_chain, 1, 0, *i);
+ if (err)
+ goto err_setup_chain;
+ flow_act->action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
+ flow_act->pkt_reformat = esw_attr->dests[j].pkt_reformat;
+ }
+ return 0;
+
+err_setup_chain:
+ esw_put_dest_tables_loop(esw, attr, esw_attr->split_count, j);
+ return err;
+}
+
+static void esw_cleanup_chain_src_port_rewrite(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+
+ esw_put_dest_tables_loop(esw, attr, esw_attr->split_count, esw_attr->out_count);
+}
+
static void
esw_cleanup_chain_dest(struct mlx5_fs_chains *chains, u32 chain, u32 prio, u32 level)
{
@@ -381,12 +440,18 @@ esw_setup_dests(struct mlx5_flow_destination *dest,
struct mlx5_flow_act *flow_act,
struct mlx5_eswitch *esw,
struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec,
int *i)
{
struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
struct mlx5_fs_chains *chains = esw_chains(esw);
int err = 0;
+ if (!mlx5_eswitch_termtbl_required(esw, attr, flow_act, spec) &&
+ MLX5_CAP_GEN(esw_attr->in_mdev, reg_c_preserve) &&
+ mlx5_eswitch_vport_match_metadata_enabled(esw))
+ attr->flags |= MLX5_ESW_ATTR_FLAG_SRC_REWRITE;
+
if (attr->dest_ft) {
esw_setup_ft_dest(dest, flow_act, attr, *i);
(*i)++;
@@ -397,6 +462,8 @@ esw_setup_dests(struct mlx5_flow_destination *dest,
err = esw_setup_chain_dest(dest, flow_act, chains, attr->dest_chain,
1, 0, *i);
(*i)++;
+ } else if (esw_is_chain_src_port_rewrite(esw, esw_attr)) {
+ err = esw_setup_chain_src_port_rewrite(dest, flow_act, esw, chains, attr, i);
} else {
*i = esw_setup_vport_dests(dest, flow_act, esw, esw_attr, *i);
}
@@ -408,10 +475,15 @@ static void
esw_cleanup_dests(struct mlx5_eswitch *esw,
struct mlx5_flow_attr *attr)
{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
struct mlx5_fs_chains *chains = esw_chains(esw);
- if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH) && attr->dest_chain)
- esw_cleanup_chain_dest(chains, attr->dest_chain, 1, 0);
+ if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH)) {
+ if (attr->dest_chain)
+ esw_cleanup_chain_dest(chains, attr->dest_chain, 1, 0);
+ else if (esw_is_chain_src_port_rewrite(esw, esw_attr))
+ esw_cleanup_chain_src_port_rewrite(esw, attr);
+ }
}
struct mlx5_flow_handle *
@@ -448,10 +520,12 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
}
}
+ mlx5_eswitch_set_rule_flow_source(esw, spec, esw_attr);
+
if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
int err;
- err = esw_setup_dests(dest, &flow_act, esw, attr, &i);
+ err = esw_setup_dests(dest, &flow_act, esw, attr, spec, &i);
if (err) {
rule = ERR_PTR(err);
goto err_create_goto_table;
@@ -498,8 +572,6 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
goto err_esw_get;
}
- mlx5_eswitch_set_rule_flow_source(esw, spec, esw_attr);
-
if (mlx5_eswitch_termtbl_required(esw, attr, &flow_act, spec))
rule = mlx5_eswitch_add_termtbl_rule(esw, fdb, spec, esw_attr,
&flow_act, dest, i);
@@ -536,7 +608,7 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
struct mlx5_flow_table *fast_fdb;
struct mlx5_flow_table *fwd_fdb;
struct mlx5_flow_handle *rule;
- int i;
+ int i, err = 0;
fast_fdb = mlx5_chains_get_table(chains, attr->chain, attr->prio, 0);
if (IS_ERR(fast_fdb)) {
@@ -554,8 +626,18 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
}
flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
- for (i = 0; i < esw_attr->split_count; i++)
- esw_setup_vport_dest(dest, &flow_act, esw, esw_attr, i, i, false);
+ for (i = 0; i < esw_attr->split_count; i++) {
+ if (esw_is_chain_src_port_rewrite(esw, esw_attr))
+ err = esw_setup_chain_src_port_rewrite(dest, &flow_act, esw, chains, attr,
+ &i);
+ else
+ esw_setup_vport_dest(dest, &flow_act, esw, esw_attr, i, i, false);
+
+ if (err) {
+ rule = ERR_PTR(err);
+ goto err_chain_src_rewrite;
+ }
+ }
dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
dest[i].ft = fwd_fdb;
i++;
@@ -570,13 +652,16 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
flow_act.flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
rule = mlx5_add_flow_rules(fast_fdb, spec, &flow_act, dest, i);
- if (IS_ERR(rule))
- goto add_err;
+ if (IS_ERR(rule)) {
+ i = esw_attr->split_count;
+ goto err_chain_src_rewrite;
+ }
atomic64_inc(&esw->offloads.num_flows);
return rule;
-add_err:
+err_chain_src_rewrite:
+ esw_put_dest_tables_loop(esw, attr, 0, i);
esw_vport_tbl_put(esw, &fwd_attr);
err_get_fwd:
mlx5_chains_put_table(chains, attr->chain, attr->prio, 0);
@@ -617,6 +702,7 @@ __mlx5_eswitch_del_rule(struct mlx5_eswitch *esw,
if (fwd_rule) {
esw_vport_tbl_put(esw, &fwd_attr);
mlx5_chains_put_table(chains, attr->chain, attr->prio, 0);
+ esw_put_dest_tables_loop(esw, attr, 0, esw_attr->split_count);
} else {
if (split)
esw_vport_tbl_put(esw, &fwd_attr);
@@ -3020,3 +3106,15 @@ int mlx5_eswitch_vhca_id_to_vport(struct mlx5_eswitch *esw, u16 vhca_id, u16 *vp
*vport_num = *res;
return 0;
}
+
+u32 mlx5_eswitch_get_vport_metadata_for_set(struct mlx5_eswitch *esw,
+ u16 vport_num)
+{
+ struct mlx5_vport *vport = mlx5_eswitch_get_vport(esw, vport_num);
+
+ if (WARN_ON_ONCE(IS_ERR(vport)))
+ return 0;
+
+ return vport->metadata;
+}
+EXPORT_SYMBOL(mlx5_eswitch_get_vport_metadata_for_set);
diff --git a/include/linux/mlx5/eswitch.h b/include/linux/mlx5/eswitch.h
index 29fd832950e0..67e341274a22 100644
--- a/include/linux/mlx5/eswitch.h
+++ b/include/linux/mlx5/eswitch.h
@@ -96,6 +96,8 @@ static inline u32 mlx5_eswitch_get_vport_metadata_mask(void)
u32 mlx5_eswitch_get_vport_metadata_for_match(struct mlx5_eswitch *esw,
u16 vport_num);
+u32 mlx5_eswitch_get_vport_metadata_for_set(struct mlx5_eswitch *esw,
+ u16 vport_num);
u8 mlx5_eswitch_mode(struct mlx5_core_dev *dev);
#else /* CONFIG_MLX5_ESWITCH */
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 06/17] net/mlx5e: Refactor tun routing helpers
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (4 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 05/17] net/mlx5e: VF tunnel TX traffic offloading Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 07/17] net/mlx5: E-Switch, Indirect table infrastructure Saeed Mahameed
` (12 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Refactor tun routing helpers to use dedicated struct
mlx5e_tc_tun_route_attr instead of multiple output arguments. This
simplifies the callers (no need to keep track of bunch of output param
pointers) and allows to unify struct release code in new
mlx5e_tc_tun_route_attr_cleanup() helper instead of requiring callers to
manually release some of the output parameters that require it.
Simplify code by unifying error handling at the end of the function and
rearranging code. Remove redundant empty line.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../ethernet/mellanox/mlx5/core/en/tc_tun.c | 235 ++++++++++--------
1 file changed, 126 insertions(+), 109 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
index 90930e54b6f2..3e18ca200c86 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
@@ -10,6 +10,27 @@
#include "rep/tc.h"
#include "rep/neigh.h"
+struct mlx5e_tc_tun_route_attr {
+ struct net_device *out_dev;
+ struct net_device *route_dev;
+ union {
+ struct flowi4 fl4;
+ struct flowi6 fl6;
+ } fl;
+ struct neighbour *n;
+ u8 ttl;
+};
+
+#define TC_TUN_ROUTE_ATTR_INIT(name) struct mlx5e_tc_tun_route_attr name = {}
+
+static void mlx5e_tc_tun_route_attr_cleanup(struct mlx5e_tc_tun_route_attr *attr)
+{
+ if (attr->n)
+ neigh_release(attr->n);
+ if (attr->route_dev)
+ dev_put(attr->route_dev);
+}
+
struct mlx5e_tc_tunnel *mlx5e_get_tc_tun(struct net_device *tunnel_dev)
{
if (netif_is_vxlan(tunnel_dev))
@@ -79,12 +100,10 @@ static int get_route_and_out_devs(struct mlx5e_priv *priv,
static int mlx5e_route_lookup_ipv4_get(struct mlx5e_priv *priv,
struct net_device *mirred_dev,
- struct net_device **out_dev,
- struct net_device **route_dev,
- struct flowi4 *fl4,
- struct neighbour **out_n,
- u8 *out_ttl)
+ struct mlx5e_tc_tun_route_attr *attr)
{
+ struct net_device *route_dev;
+ struct net_device *out_dev;
struct neighbour *n;
struct rtable *rt;
@@ -97,46 +116,50 @@ static int mlx5e_route_lookup_ipv4_get(struct mlx5e_priv *priv,
struct mlx5_eswitch *esw = mdev->priv.eswitch;
uplink_dev = mlx5_eswitch_uplink_get_proto_dev(esw, REP_ETH);
- fl4->flowi4_oif = uplink_dev->ifindex;
+ attr->fl.fl4.flowi4_oif = uplink_dev->ifindex;
}
- rt = ip_route_output_key(dev_net(mirred_dev), fl4);
+ rt = ip_route_output_key(dev_net(mirred_dev), &attr->fl.fl4);
if (IS_ERR(rt))
return PTR_ERR(rt);
if (mlx5_lag_is_multipath(mdev) && rt->rt_gw_family != AF_INET) {
- ip_rt_put(rt);
- return -ENETUNREACH;
+ ret = -ENETUNREACH;
+ goto err_rt_release;
}
#else
return -EOPNOTSUPP;
#endif
- ret = get_route_and_out_devs(priv, rt->dst.dev, route_dev, out_dev);
- if (ret < 0) {
- ip_rt_put(rt);
- return ret;
- }
- dev_hold(*route_dev);
+ ret = get_route_and_out_devs(priv, rt->dst.dev, &route_dev, &out_dev);
+ if (ret < 0)
+ goto err_rt_release;
+ dev_hold(route_dev);
- if (!(*out_ttl))
- *out_ttl = ip4_dst_hoplimit(&rt->dst);
- n = dst_neigh_lookup(&rt->dst, &fl4->daddr);
- ip_rt_put(rt);
+ if (!attr->ttl)
+ attr->ttl = ip4_dst_hoplimit(&rt->dst);
+ n = dst_neigh_lookup(&rt->dst, &attr->fl.fl4.daddr);
if (!n) {
- dev_put(*route_dev);
- return -ENOMEM;
+ ret = -ENOMEM;
+ goto err_dev_release;
}
- *out_n = n;
+ ip_rt_put(rt);
+ attr->route_dev = route_dev;
+ attr->out_dev = out_dev;
+ attr->n = n;
return 0;
+
+err_dev_release:
+ dev_put(route_dev);
+err_rt_release:
+ ip_rt_put(rt);
+ return ret;
}
-static void mlx5e_route_lookup_ipv4_put(struct net_device *route_dev,
- struct neighbour *n)
+static void mlx5e_route_lookup_ipv4_put(struct mlx5e_tc_tun_route_attr *attr)
{
- neigh_release(n);
- dev_put(route_dev);
+ mlx5e_tc_tun_route_attr_cleanup(attr);
}
static const char *mlx5e_netdev_kind(struct net_device *dev)
@@ -188,28 +211,25 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
{
int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size);
const struct ip_tunnel_key *tun_key = &e->tun_info->key;
- struct net_device *out_dev, *route_dev;
- struct flowi4 fl4 = {};
- struct neighbour *n;
+ TC_TUN_ROUTE_ATTR_INIT(attr);
int ipv4_encap_size;
char *encap_header;
- u8 nud_state, ttl;
struct iphdr *ip;
+ u8 nud_state;
int err;
/* add the IP fields */
- fl4.flowi4_tos = tun_key->tos;
- fl4.daddr = tun_key->u.ipv4.dst;
- fl4.saddr = tun_key->u.ipv4.src;
- ttl = tun_key->ttl;
+ attr.fl.fl4.flowi4_tos = tun_key->tos;
+ attr.fl.fl4.daddr = tun_key->u.ipv4.dst;
+ attr.fl.fl4.saddr = tun_key->u.ipv4.src;
+ attr.ttl = tun_key->ttl;
- err = mlx5e_route_lookup_ipv4_get(priv, mirred_dev, &out_dev, &route_dev,
- &fl4, &n, &ttl);
+ err = mlx5e_route_lookup_ipv4_get(priv, mirred_dev, &attr);
if (err)
return err;
ipv4_encap_size =
- (is_vlan_dev(route_dev) ? VLAN_ETH_HLEN : ETH_HLEN) +
+ (is_vlan_dev(attr.route_dev) ? VLAN_ETH_HLEN : ETH_HLEN) +
sizeof(struct iphdr) +
e->tunnel->calc_hlen(e);
@@ -229,37 +249,37 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
/* used by mlx5e_detach_encap to lookup a neigh hash table
* entry in the neigh hash table when a user deletes a rule
*/
- e->m_neigh.dev = n->dev;
- e->m_neigh.family = n->ops->family;
- memcpy(&e->m_neigh.dst_ip, n->primary_key, n->tbl->key_len);
- e->out_dev = out_dev;
- e->route_dev_ifindex = route_dev->ifindex;
+ e->m_neigh.dev = attr.n->dev;
+ e->m_neigh.family = attr.n->ops->family;
+ memcpy(&e->m_neigh.dst_ip, attr.n->primary_key, attr.n->tbl->key_len);
+ e->out_dev = attr.out_dev;
+ e->route_dev_ifindex = attr.route_dev->ifindex;
/* It's important to add the neigh to the hash table before checking
* the neigh validity state. So if we'll get a notification, in case the
* neigh changes it's validity state, we would find the relevant neigh
* in the hash.
*/
- err = mlx5e_rep_encap_entry_attach(netdev_priv(out_dev), e);
+ err = mlx5e_rep_encap_entry_attach(netdev_priv(attr.out_dev), e);
if (err)
goto free_encap;
- read_lock_bh(&n->lock);
- nud_state = n->nud_state;
- ether_addr_copy(e->h_dest, n->ha);
- read_unlock_bh(&n->lock);
+ read_lock_bh(&attr.n->lock);
+ nud_state = attr.n->nud_state;
+ ether_addr_copy(e->h_dest, attr.n->ha);
+ read_unlock_bh(&attr.n->lock);
/* add ethernet header */
- ip = (struct iphdr *)gen_eth_tnl_hdr(encap_header, route_dev, e,
+ ip = (struct iphdr *)gen_eth_tnl_hdr(encap_header, attr.route_dev, e,
ETH_P_IP);
/* add ip header */
ip->tos = tun_key->tos;
ip->version = 0x4;
ip->ihl = 0x5;
- ip->ttl = ttl;
- ip->daddr = fl4.daddr;
- ip->saddr = fl4.saddr;
+ ip->ttl = attr.ttl;
+ ip->daddr = attr.fl.fl4.daddr;
+ ip->saddr = attr.fl.fl4.saddr;
/* add tunneling protocol header */
err = mlx5e_gen_ip_tunnel_header((char *)ip + sizeof(struct iphdr),
@@ -271,7 +291,7 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
e->encap_header = encap_header;
if (!(nud_state & NUD_VALID)) {
- neigh_event_send(n, NULL);
+ neigh_event_send(attr.n, NULL);
/* the encap entry will be made valid on neigh update event
* and not used before that.
*/
@@ -287,8 +307,8 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
}
e->flags |= MLX5_ENCAP_ENTRY_VALID;
- mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev));
- mlx5e_route_lookup_ipv4_put(route_dev, n);
+ mlx5e_rep_queue_neigh_stats_work(netdev_priv(attr.out_dev));
+ mlx5e_route_lookup_ipv4_put(&attr);
return err;
destroy_neigh_entry:
@@ -296,55 +316,56 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
free_encap:
kfree(encap_header);
release_neigh:
- mlx5e_route_lookup_ipv4_put(route_dev, n);
+ mlx5e_route_lookup_ipv4_put(&attr);
return err;
}
#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
static int mlx5e_route_lookup_ipv6_get(struct mlx5e_priv *priv,
struct net_device *mirred_dev,
- struct net_device **out_dev,
- struct net_device **route_dev,
- struct flowi6 *fl6,
- struct neighbour **out_n,
- u8 *out_ttl)
+ struct mlx5e_tc_tun_route_attr *attr)
{
+ struct net_device *route_dev;
+ struct net_device *out_dev;
struct dst_entry *dst;
struct neighbour *n;
-
int ret;
- dst = ipv6_stub->ipv6_dst_lookup_flow(dev_net(mirred_dev), NULL, fl6,
+ dst = ipv6_stub->ipv6_dst_lookup_flow(dev_net(mirred_dev), NULL, &attr->fl.fl6,
NULL);
if (IS_ERR(dst))
return PTR_ERR(dst);
- if (!(*out_ttl))
- *out_ttl = ip6_dst_hoplimit(dst);
+ if (!attr->ttl)
+ attr->ttl = ip6_dst_hoplimit(dst);
- ret = get_route_and_out_devs(priv, dst->dev, route_dev, out_dev);
- if (ret < 0) {
- dst_release(dst);
- return ret;
- }
+ ret = get_route_and_out_devs(priv, dst->dev, &route_dev, &out_dev);
+ if (ret < 0)
+ goto err_dst_release;
- dev_hold(*route_dev);
- n = dst_neigh_lookup(dst, &fl6->daddr);
- dst_release(dst);
+ dev_hold(route_dev);
+ n = dst_neigh_lookup(dst, &attr->fl.fl6.daddr);
if (!n) {
- dev_put(*route_dev);
- return -ENOMEM;
+ ret = -ENOMEM;
+ goto err_dev_release;
}
- *out_n = n;
+ dst_release(dst);
+ attr->out_dev = out_dev;
+ attr->route_dev = route_dev;
+ attr->n = n;
return 0;
+
+err_dev_release:
+ dev_put(route_dev);
+err_dst_release:
+ dst_release(dst);
+ return ret;
}
-static void mlx5e_route_lookup_ipv6_put(struct net_device *route_dev,
- struct neighbour *n)
+static void mlx5e_route_lookup_ipv6_put(struct mlx5e_tc_tun_route_attr *attr)
{
- neigh_release(n);
- dev_put(route_dev);
+ mlx5e_tc_tun_route_attr_cleanup(attr);
}
int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
@@ -353,28 +374,24 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
{
int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size);
const struct ip_tunnel_key *tun_key = &e->tun_info->key;
- struct net_device *out_dev, *route_dev;
- struct flowi6 fl6 = {};
+ TC_TUN_ROUTE_ATTR_INIT(attr);
struct ipv6hdr *ip6h;
- struct neighbour *n = NULL;
int ipv6_encap_size;
char *encap_header;
- u8 nud_state, ttl;
+ u8 nud_state;
int err;
- ttl = tun_key->ttl;
-
- fl6.flowlabel = ip6_make_flowinfo(RT_TOS(tun_key->tos), tun_key->label);
- fl6.daddr = tun_key->u.ipv6.dst;
- fl6.saddr = tun_key->u.ipv6.src;
+ attr.ttl = tun_key->ttl;
+ attr.fl.fl6.flowlabel = ip6_make_flowinfo(RT_TOS(tun_key->tos), tun_key->label);
+ attr.fl.fl6.daddr = tun_key->u.ipv6.dst;
+ attr.fl.fl6.saddr = tun_key->u.ipv6.src;
- err = mlx5e_route_lookup_ipv6_get(priv, mirred_dev, &out_dev, &route_dev,
- &fl6, &n, &ttl);
+ err = mlx5e_route_lookup_ipv6_get(priv, mirred_dev, &attr);
if (err)
return err;
ipv6_encap_size =
- (is_vlan_dev(route_dev) ? VLAN_ETH_HLEN : ETH_HLEN) +
+ (is_vlan_dev(attr.route_dev) ? VLAN_ETH_HLEN : ETH_HLEN) +
sizeof(struct ipv6hdr) +
e->tunnel->calc_hlen(e);
@@ -394,36 +411,36 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
/* used by mlx5e_detach_encap to lookup a neigh hash table
* entry in the neigh hash table when a user deletes a rule
*/
- e->m_neigh.dev = n->dev;
- e->m_neigh.family = n->ops->family;
- memcpy(&e->m_neigh.dst_ip, n->primary_key, n->tbl->key_len);
- e->out_dev = out_dev;
- e->route_dev_ifindex = route_dev->ifindex;
+ e->m_neigh.dev = attr.n->dev;
+ e->m_neigh.family = attr.n->ops->family;
+ memcpy(&e->m_neigh.dst_ip, attr.n->primary_key, attr.n->tbl->key_len);
+ e->out_dev = attr.out_dev;
+ e->route_dev_ifindex = attr.route_dev->ifindex;
/* It's importent to add the neigh to the hash table before checking
* the neigh validity state. So if we'll get a notification, in case the
* neigh changes it's validity state, we would find the relevant neigh
* in the hash.
*/
- err = mlx5e_rep_encap_entry_attach(netdev_priv(out_dev), e);
+ err = mlx5e_rep_encap_entry_attach(netdev_priv(attr.out_dev), e);
if (err)
goto free_encap;
- read_lock_bh(&n->lock);
- nud_state = n->nud_state;
- ether_addr_copy(e->h_dest, n->ha);
- read_unlock_bh(&n->lock);
+ read_lock_bh(&attr.n->lock);
+ nud_state = attr.n->nud_state;
+ ether_addr_copy(e->h_dest, attr.n->ha);
+ read_unlock_bh(&attr.n->lock);
/* add ethernet header */
- ip6h = (struct ipv6hdr *)gen_eth_tnl_hdr(encap_header, route_dev, e,
+ ip6h = (struct ipv6hdr *)gen_eth_tnl_hdr(encap_header, attr.route_dev, e,
ETH_P_IPV6);
/* add ip header */
ip6_flow_hdr(ip6h, tun_key->tos, 0);
/* the HW fills up ipv6 payload len */
- ip6h->hop_limit = ttl;
- ip6h->daddr = fl6.daddr;
- ip6h->saddr = fl6.saddr;
+ ip6h->hop_limit = attr.ttl;
+ ip6h->daddr = attr.fl.fl6.daddr;
+ ip6h->saddr = attr.fl.fl6.saddr;
/* add tunneling protocol header */
err = mlx5e_gen_ip_tunnel_header((char *)ip6h + sizeof(struct ipv6hdr),
@@ -435,7 +452,7 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
e->encap_header = encap_header;
if (!(nud_state & NUD_VALID)) {
- neigh_event_send(n, NULL);
+ neigh_event_send(attr.n, NULL);
/* the encap entry will be made valid on neigh update event
* and not used before that.
*/
@@ -452,8 +469,8 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
}
e->flags |= MLX5_ENCAP_ENTRY_VALID;
- mlx5e_rep_queue_neigh_stats_work(netdev_priv(out_dev));
- mlx5e_route_lookup_ipv6_put(route_dev, n);
+ mlx5e_rep_queue_neigh_stats_work(netdev_priv(attr.out_dev));
+ mlx5e_route_lookup_ipv6_put(&attr);
return err;
destroy_neigh_entry:
@@ -461,7 +478,7 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
free_encap:
kfree(encap_header);
release_neigh:
- mlx5e_route_lookup_ipv6_put(route_dev, n);
+ mlx5e_route_lookup_ipv6_put(&attr);
return err;
}
#endif
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 07/17] net/mlx5: E-Switch, Indirect table infrastructure
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (5 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 06/17] net/mlx5e: Refactor tun routing helpers Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 08/17] net/mlx5e: Remove redundant match on tunnel destination mac Saeed Mahameed
` (11 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Indirect table infrastructure is used to allow fully processing VF tunnel
traffic in hardware. Kernel software model uses two TC rules for such
traffic: UL rep to tunnel device, then tunnel VF rep to destination VF rep.
To implement such pipeline driver needs to program the hardware after
matching on UL rule to overwrite source vport from UL to tunnel VF and
recirculate the packet to the root table to allow matching on the rule
installed on tunnel VF. For this indirect table matches all encapsulated
traffic by tunnel parameters and all other IP traffic is sent to tunnel VF
by the miss rule.
Indirect table API overview:
- mlx5_esw_indir_table_{init|destroy}() - init and destroy opaque indirect
table object.
- mlx5_esw_indir_table_get() - get or create new table according to vport
id and IP version. Table has following pre-created groups: recirculation
group with match on ethertype and VNI (rules that match encapsulated
packets are installed to this group) and forward group with default/miss
rule that forwards to vport of tunnel endpoint VF (rule for regular
non-encapsulated packets).
- mlx5_esw_indir_table_put() - decrease reference to the indirect table and
matching rule (for encapsulated traffic).
- mlx5_esw_indir_table_needed() - check that in_port is an uplink port and
out_port is VF on the same eswitch, verify that the rule is for IP traffic
and source port rewrite functionality can be used.
- mlx5_esw_indir_table_decap_vport() - function returns decap vport of
flow attribute.
Co-developed-by: Dmytro Linkin <dlinkin@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../net/ethernet/mellanox/mlx5/core/Makefile | 1 +
.../net/ethernet/mellanox/mlx5/core/en_tc.h | 14 +
.../mellanox/mlx5/core/esw/indir_table.c | 508 ++++++++++++++++++
.../mellanox/mlx5/core/esw/indir_table.h | 76 +++
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 5 +
.../mellanox/mlx5/core/eswitch_offloads.c | 12 +
6 files changed, 616 insertions(+)
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.c
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.h
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
index 8809dd4de57e..f1ccfba60068 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -40,6 +40,7 @@ mlx5_core-$(CONFIG_MLX5_ESWITCH) += lag_mp.o lib/geneve.o lib/port_tun.o \
en_rep.o en/rep/bond.o en/mod_hdr.o
mlx5_core-$(CONFIG_MLX5_CLS_ACT) += en_tc.o en/rep/tc.o en/rep/neigh.o \
en/mapping.o lib/fs_chains.o en/tc_tun.o \
+ esw/indir_table.o \
en/tc_tun_vxlan.o en/tc_tun_gre.o en/tc_tun_geneve.o \
en/tc_tun_mplsoudp.o diag/en_tc_tracepoint.o
mlx5_core-$(CONFIG_MLX5_TC_CT) += en/tc_ct.o
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
index 56d809904ea7..852e0981343d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -76,6 +76,7 @@ struct mlx5_flow_attr {
struct mlx5_flow_table *dest_ft;
u8 inner_match_level;
u8 outer_match_level;
+ u8 ip_version;
u32 flags;
union {
struct mlx5_esw_flow_attr esw_attr[0];
@@ -83,6 +84,19 @@ struct mlx5_flow_attr {
};
};
+struct mlx5_rx_tun_attr {
+ u16 decap_vport;
+ union {
+ __be32 v4;
+ struct in6_addr v6;
+ } src_ip; /* Valid if decap_vport is not zero */
+ union {
+ __be32 v4;
+ struct in6_addr v6;
+ } dst_ip; /* Valid if decap_vport is not zero */
+ u32 vni;
+};
+
#define MLX5E_TC_TABLE_CHAIN_TAG_BITS 16
#define MLX5E_TC_TABLE_CHAIN_TAG_MASK GENMASK(MLX5E_TC_TABLE_CHAIN_TAG_BITS - 1, 0)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.c
new file mode 100644
index 000000000000..a0ebf40c9907
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.c
@@ -0,0 +1,508 @@
+// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
+/* Copyright (c) 2021 Mellanox Technologies. */
+
+#include <linux/etherdevice.h>
+#include <linux/idr.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/mlx5_ifc.h>
+#include <linux/mlx5/vport.h>
+#include <linux/mlx5/fs.h>
+#include "mlx5_core.h"
+#include "eswitch.h"
+#include "en.h"
+#include "en_tc.h"
+#include "fs_core.h"
+#include "esw/indir_table.h"
+#include "lib/fs_chains.h"
+
+#define MLX5_ESW_INDIR_TABLE_SIZE 128
+#define MLX5_ESW_INDIR_TABLE_RECIRC_IDX_MAX (MLX5_ESW_INDIR_TABLE_SIZE - 2)
+#define MLX5_ESW_INDIR_TABLE_FWD_IDX (MLX5_ESW_INDIR_TABLE_SIZE - 1)
+
+struct mlx5_esw_indir_table_rule {
+ struct list_head list;
+ struct mlx5_flow_handle *handle;
+ union {
+ __be32 v4;
+ struct in6_addr v6;
+ } dst_ip;
+ u32 vni;
+ struct mlx5_modify_hdr *mh;
+ refcount_t refcnt;
+};
+
+struct mlx5_esw_indir_table_entry {
+ struct hlist_node hlist;
+ struct mlx5_flow_table *ft;
+ struct mlx5_flow_group *recirc_grp;
+ struct mlx5_flow_group *fwd_grp;
+ struct mlx5_flow_handle *fwd_rule;
+ struct list_head recirc_rules;
+ int recirc_cnt;
+ int fwd_ref;
+
+ u16 vport;
+ u8 ip_version;
+};
+
+struct mlx5_esw_indir_table {
+ struct mutex lock; /* protects table */
+ DECLARE_HASHTABLE(table, 8);
+};
+
+struct mlx5_esw_indir_table *
+mlx5_esw_indir_table_init(void)
+{
+ struct mlx5_esw_indir_table *indir = kvzalloc(sizeof(*indir), GFP_KERNEL);
+
+ if (!indir)
+ return ERR_PTR(-ENOMEM);
+
+ mutex_init(&indir->lock);
+ hash_init(indir->table);
+ return indir;
+}
+
+void
+mlx5_esw_indir_table_destroy(struct mlx5_esw_indir_table *indir)
+{
+ mutex_destroy(&indir->lock);
+ kvfree(indir);
+}
+
+bool
+mlx5_esw_indir_table_needed(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ u16 vport_num,
+ struct mlx5_core_dev *dest_mdev)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+
+ /* Use indirect table for all IP traffic from UL to VF with vport
+ * destination when source rewrite flag is set.
+ */
+ return esw_attr->in_rep->vport == MLX5_VPORT_UPLINK &&
+ mlx5_eswitch_is_vf_vport(esw, vport_num) &&
+ esw->dev == dest_mdev &&
+ attr->ip_version &&
+ attr->flags & MLX5_ESW_ATTR_FLAG_SRC_REWRITE;
+}
+
+u16
+mlx5_esw_indir_table_decap_vport(struct mlx5_flow_attr *attr)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+
+ return esw_attr->rx_tun_attr ? esw_attr->rx_tun_attr->decap_vport : 0;
+}
+
+static struct mlx5_esw_indir_table_rule *
+mlx5_esw_indir_table_rule_lookup(struct mlx5_esw_indir_table_entry *e,
+ struct mlx5_esw_flow_attr *attr)
+{
+ struct mlx5_esw_indir_table_rule *rule;
+
+ list_for_each_entry(rule, &e->recirc_rules, list)
+ if (rule->vni == attr->rx_tun_attr->vni &&
+ !memcmp(&rule->dst_ip, &attr->rx_tun_attr->dst_ip,
+ sizeof(attr->rx_tun_attr->dst_ip)))
+ goto found;
+ return NULL;
+
+found:
+ refcount_inc(&rule->refcnt);
+ return rule;
+}
+
+static int mlx5_esw_indir_table_rule_get(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec,
+ struct mlx5_esw_indir_table_entry *e)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+ struct mlx5_fs_chains *chains = esw_chains(esw);
+ struct mlx5e_tc_mod_hdr_acts mod_acts = {};
+ struct mlx5_flow_destination dest = {};
+ struct mlx5_esw_indir_table_rule *rule;
+ struct mlx5_flow_act flow_act = {};
+ struct mlx5_flow_spec *rule_spec;
+ struct mlx5_flow_handle *handle;
+ int err = 0;
+ u32 data;
+
+ rule = mlx5_esw_indir_table_rule_lookup(e, esw_attr);
+ if (rule)
+ return 0;
+
+ if (e->recirc_cnt == MLX5_ESW_INDIR_TABLE_RECIRC_IDX_MAX)
+ return -EINVAL;
+
+ rule_spec = kvzalloc(sizeof(*rule_spec), GFP_KERNEL);
+ if (!rule_spec)
+ return -ENOMEM;
+
+ rule = kzalloc(sizeof(*rule), GFP_KERNEL);
+ if (!rule) {
+ err = -ENOMEM;
+ goto out;
+ }
+
+ rule_spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS |
+ MLX5_MATCH_MISC_PARAMETERS |
+ MLX5_MATCH_MISC_PARAMETERS_2;
+ if (MLX5_CAP_FLOWTABLE_NIC_RX(esw->dev, ft_field_support.outer_ip_version)) {
+ MLX5_SET(fte_match_param, rule_spec->match_criteria,
+ outer_headers.ip_version, 0xf);
+ MLX5_SET(fte_match_param, rule_spec->match_value, outer_headers.ip_version,
+ attr->ip_version);
+ } else if (attr->ip_version) {
+ MLX5_SET_TO_ONES(fte_match_param, rule_spec->match_criteria,
+ outer_headers.ethertype);
+ MLX5_SET(fte_match_param, rule_spec->match_value, outer_headers.ethertype,
+ (attr->ip_version == 4 ? ETH_P_IP : ETH_P_IPV6));
+ } else {
+ err = -EOPNOTSUPP;
+ goto err_mod_hdr;
+ }
+
+ if (attr->ip_version == 4) {
+ MLX5_SET_TO_ONES(fte_match_param, rule_spec->match_criteria,
+ outer_headers.dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+ MLX5_SET(fte_match_param, rule_spec->match_value,
+ outer_headers.dst_ipv4_dst_ipv6.ipv4_layout.ipv4,
+ ntohl(esw_attr->rx_tun_attr->dst_ip.v4));
+ } else if (attr->ip_version == 6) {
+ int len = sizeof(struct in6_addr);
+
+ memset(MLX5_ADDR_OF(fte_match_param, rule_spec->match_criteria,
+ outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+ 0xff, len);
+ memcpy(MLX5_ADDR_OF(fte_match_param, rule_spec->match_value,
+ outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+ &esw_attr->rx_tun_attr->dst_ip.v6, len);
+ }
+
+ MLX5_SET_TO_ONES(fte_match_param, rule_spec->match_criteria,
+ misc_parameters.vxlan_vni);
+ MLX5_SET(fte_match_param, rule_spec->match_value, misc_parameters.vxlan_vni,
+ MLX5_GET(fte_match_param, spec->match_value, misc_parameters.vxlan_vni));
+
+ MLX5_SET(fte_match_param, rule_spec->match_criteria,
+ misc_parameters_2.metadata_reg_c_0, mlx5_eswitch_get_vport_metadata_mask());
+ MLX5_SET(fte_match_param, rule_spec->match_value, misc_parameters_2.metadata_reg_c_0,
+ mlx5_eswitch_get_vport_metadata_for_match(esw_attr->in_mdev->priv.eswitch,
+ MLX5_VPORT_UPLINK));
+
+ /* Modify flow source to recirculate packet */
+ data = mlx5_eswitch_get_vport_metadata_for_set(esw, esw_attr->rx_tun_attr->decap_vport);
+ err = mlx5e_tc_match_to_reg_set(esw->dev, &mod_acts, MLX5_FLOW_NAMESPACE_FDB,
+ VPORT_TO_REG, data);
+ if (err)
+ goto err_mod_hdr;
+
+ flow_act.modify_hdr = mlx5_modify_header_alloc(esw->dev, MLX5_FLOW_NAMESPACE_FDB,
+ mod_acts.num_actions, mod_acts.actions);
+ if (IS_ERR(flow_act.modify_hdr)) {
+ err = PTR_ERR(flow_act.modify_hdr);
+ goto err_mod_hdr;
+ }
+
+ flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | MLX5_FLOW_CONTEXT_ACTION_MOD_HDR;
+ flow_act.flags = FLOW_ACT_IGNORE_FLOW_LEVEL | FLOW_ACT_NO_APPEND;
+ dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+ dest.ft = mlx5_chains_get_table(chains, 0, 1, 0);
+ if (!dest.ft) {
+ err = PTR_ERR(dest.ft);
+ goto err_table;
+ }
+ handle = mlx5_add_flow_rules(e->ft, rule_spec, &flow_act, &dest, 1);
+ if (IS_ERR(handle)) {
+ err = PTR_ERR(handle);
+ goto err_handle;
+ }
+
+ dealloc_mod_hdr_actions(&mod_acts);
+ rule->handle = handle;
+ rule->vni = esw_attr->rx_tun_attr->vni;
+ rule->mh = flow_act.modify_hdr;
+ memcpy(&rule->dst_ip, &esw_attr->rx_tun_attr->dst_ip,
+ sizeof(esw_attr->rx_tun_attr->dst_ip));
+ refcount_set(&rule->refcnt, 1);
+ list_add(&rule->list, &e->recirc_rules);
+ e->recirc_cnt++;
+ goto out;
+
+err_handle:
+ mlx5_chains_put_table(chains, 0, 1, 0);
+err_table:
+ mlx5_modify_header_dealloc(esw->dev, flow_act.modify_hdr);
+err_mod_hdr:
+ kfree(rule);
+out:
+ kfree(rule_spec);
+ return err;
+}
+
+static void mlx5_esw_indir_table_rule_put(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5_esw_indir_table_entry *e)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+ struct mlx5_fs_chains *chains = esw_chains(esw);
+ struct mlx5_esw_indir_table_rule *rule;
+
+ list_for_each_entry(rule, &e->recirc_rules, list)
+ if (rule->vni == esw_attr->rx_tun_attr->vni &&
+ !memcmp(&rule->dst_ip, &esw_attr->rx_tun_attr->dst_ip,
+ sizeof(esw_attr->rx_tun_attr->dst_ip)))
+ goto found;
+
+ return;
+
+found:
+ if (!refcount_dec_and_test(&rule->refcnt))
+ return;
+
+ mlx5_del_flow_rules(rule->handle);
+ mlx5_chains_put_table(chains, 0, 1, 0);
+ mlx5_modify_header_dealloc(esw->dev, rule->mh);
+ list_del(&rule->list);
+ kfree(rule);
+ e->recirc_cnt--;
+}
+
+static int mlx5_create_indir_recirc_group(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec,
+ struct mlx5_esw_indir_table_entry *e)
+{
+ int err = 0, inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+ u32 *in, *match;
+
+ in = kvzalloc(inlen, GFP_KERNEL);
+ if (!in)
+ return -ENOMEM;
+
+ MLX5_SET(create_flow_group_in, in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS |
+ MLX5_MATCH_MISC_PARAMETERS | MLX5_MATCH_MISC_PARAMETERS_2);
+ match = MLX5_ADDR_OF(create_flow_group_in, in, match_criteria);
+
+ if (MLX5_CAP_FLOWTABLE_NIC_RX(esw->dev, ft_field_support.outer_ip_version))
+ MLX5_SET(fte_match_param, match, outer_headers.ip_version, 0xf);
+ else
+ MLX5_SET_TO_ONES(fte_match_param, match, outer_headers.ethertype);
+
+ if (attr->ip_version == 4) {
+ MLX5_SET_TO_ONES(fte_match_param, match,
+ outer_headers.dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+ } else if (attr->ip_version == 6) {
+ memset(MLX5_ADDR_OF(fte_match_param, match,
+ outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+ 0xff, sizeof(struct in6_addr));
+ } else {
+ err = -EOPNOTSUPP;
+ goto out;
+ }
+
+ MLX5_SET_TO_ONES(fte_match_param, match, misc_parameters.vxlan_vni);
+ MLX5_SET(fte_match_param, match, misc_parameters_2.metadata_reg_c_0,
+ mlx5_eswitch_get_vport_metadata_mask());
+ MLX5_SET(create_flow_group_in, in, start_flow_index, 0);
+ MLX5_SET(create_flow_group_in, in, end_flow_index, MLX5_ESW_INDIR_TABLE_RECIRC_IDX_MAX);
+ e->recirc_grp = mlx5_create_flow_group(e->ft, in);
+ if (IS_ERR(e->recirc_grp)) {
+ err = PTR_ERR(e->recirc_grp);
+ goto out;
+ }
+
+ INIT_LIST_HEAD(&e->recirc_rules);
+ e->recirc_cnt = 0;
+
+out:
+ kfree(in);
+ return err;
+}
+
+static int mlx5_create_indir_fwd_group(struct mlx5_eswitch *esw,
+ struct mlx5_esw_indir_table_entry *e)
+{
+ int err = 0, inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+ struct mlx5_flow_destination dest = {};
+ struct mlx5_flow_act flow_act = {};
+ struct mlx5_flow_spec *spec;
+ u32 *in;
+
+ in = kvzalloc(inlen, GFP_KERNEL);
+ if (!in)
+ return -ENOMEM;
+
+ spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+ if (!spec) {
+ kfree(in);
+ return -ENOMEM;
+ }
+
+ /* Hold one entry */
+ MLX5_SET(create_flow_group_in, in, start_flow_index, MLX5_ESW_INDIR_TABLE_FWD_IDX);
+ MLX5_SET(create_flow_group_in, in, end_flow_index, MLX5_ESW_INDIR_TABLE_FWD_IDX);
+ e->fwd_grp = mlx5_create_flow_group(e->ft, in);
+ if (IS_ERR(e->fwd_grp)) {
+ err = PTR_ERR(e->fwd_grp);
+ goto err_out;
+ }
+
+ flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+ dest.type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
+ dest.vport.num = e->vport;
+ dest.vport.vhca_id = MLX5_CAP_GEN(esw->dev, vhca_id);
+ e->fwd_rule = mlx5_add_flow_rules(e->ft, spec, &flow_act, &dest, 1);
+ if (IS_ERR(e->fwd_rule)) {
+ mlx5_destroy_flow_group(e->fwd_grp);
+ err = PTR_ERR(e->fwd_rule);
+ }
+
+err_out:
+ kfree(spec);
+ kfree(in);
+ return err;
+}
+
+static struct mlx5_esw_indir_table_entry *
+mlx5_esw_indir_table_entry_create(struct mlx5_eswitch *esw, struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec, u16 vport, bool decap)
+{
+ struct mlx5_flow_table_attr ft_attr = {};
+ struct mlx5_flow_namespace *root_ns;
+ struct mlx5_esw_indir_table_entry *e;
+ struct mlx5_flow_table *ft;
+ int err = 0;
+
+ root_ns = mlx5_get_flow_namespace(esw->dev, MLX5_FLOW_NAMESPACE_FDB);
+ if (!root_ns)
+ return ERR_PTR(-ENOENT);
+
+ e = kzalloc(sizeof(*e), GFP_KERNEL);
+ if (!e)
+ return ERR_PTR(-ENOMEM);
+
+ ft_attr.prio = FDB_TC_OFFLOAD;
+ ft_attr.max_fte = MLX5_ESW_INDIR_TABLE_SIZE;
+ ft_attr.flags = MLX5_FLOW_TABLE_UNMANAGED;
+ ft_attr.level = 1;
+
+ ft = mlx5_create_flow_table(root_ns, &ft_attr);
+ if (IS_ERR(ft)) {
+ err = PTR_ERR(ft);
+ goto tbl_err;
+ }
+ e->ft = ft;
+ e->vport = vport;
+ e->ip_version = attr->ip_version;
+ e->fwd_ref = !decap;
+
+ err = mlx5_create_indir_recirc_group(esw, attr, spec, e);
+ if (err)
+ goto recirc_grp_err;
+
+ if (decap) {
+ err = mlx5_esw_indir_table_rule_get(esw, attr, spec, e);
+ if (err)
+ goto recirc_rule_err;
+ }
+
+ err = mlx5_create_indir_fwd_group(esw, e);
+ if (err)
+ goto fwd_grp_err;
+
+ hash_add(esw->fdb_table.offloads.indir->table, &e->hlist,
+ vport << 16 | attr->ip_version);
+
+ return e;
+
+fwd_grp_err:
+ if (decap)
+ mlx5_esw_indir_table_rule_put(esw, attr, e);
+recirc_rule_err:
+ mlx5_destroy_flow_group(e->recirc_grp);
+recirc_grp_err:
+ mlx5_destroy_flow_table(e->ft);
+tbl_err:
+ kfree(e);
+ return ERR_PTR(err);
+}
+
+static struct mlx5_esw_indir_table_entry *
+mlx5_esw_indir_table_entry_lookup(struct mlx5_eswitch *esw, u16 vport, u8 ip_version)
+{
+ struct mlx5_esw_indir_table_entry *e;
+ u32 key = vport << 16 | ip_version;
+
+ hash_for_each_possible(esw->fdb_table.offloads.indir->table, e, hlist, key)
+ if (e->vport == vport && e->ip_version == ip_version)
+ return e;
+
+ return NULL;
+}
+
+struct mlx5_flow_table *mlx5_esw_indir_table_get(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec,
+ u16 vport, bool decap)
+{
+ struct mlx5_esw_indir_table_entry *e;
+ int err;
+
+ mutex_lock(&esw->fdb_table.offloads.indir->lock);
+ e = mlx5_esw_indir_table_entry_lookup(esw, vport, attr->ip_version);
+ if (e) {
+ if (!decap) {
+ e->fwd_ref++;
+ } else {
+ err = mlx5_esw_indir_table_rule_get(esw, attr, spec, e);
+ if (err)
+ goto out_err;
+ }
+ } else {
+ e = mlx5_esw_indir_table_entry_create(esw, attr, spec, vport, decap);
+ if (IS_ERR(e)) {
+ err = PTR_ERR(e);
+ esw_warn(esw->dev, "Failed to create indirection table, err %d.\n", err);
+ goto out_err;
+ }
+ }
+ mutex_unlock(&esw->fdb_table.offloads.indir->lock);
+ return e->ft;
+
+out_err:
+ mutex_unlock(&esw->fdb_table.offloads.indir->lock);
+ return ERR_PTR(err);
+}
+
+void mlx5_esw_indir_table_put(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ u16 vport, bool decap)
+{
+ struct mlx5_esw_indir_table_entry *e;
+
+ mutex_lock(&esw->fdb_table.offloads.indir->lock);
+ e = mlx5_esw_indir_table_entry_lookup(esw, vport, attr->ip_version);
+ if (!e)
+ goto out;
+
+ if (!decap)
+ e->fwd_ref--;
+ else
+ mlx5_esw_indir_table_rule_put(esw, attr, e);
+
+ if (e->fwd_ref || e->recirc_cnt)
+ goto out;
+
+ hash_del(&e->hlist);
+ mlx5_destroy_flow_group(e->recirc_grp);
+ mlx5_del_flow_rules(e->fwd_rule);
+ mlx5_destroy_flow_group(e->fwd_grp);
+ mlx5_destroy_flow_table(e->ft);
+ kfree(e);
+out:
+ mutex_unlock(&esw->fdb_table.offloads.indir->lock);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.h b/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.h
new file mode 100644
index 000000000000..cb9eafd1b4ee
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.h
@@ -0,0 +1,76 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2021 Mellanox Technologies. */
+
+#ifndef __MLX5_ESW_FT_H__
+#define __MLX5_ESW_FT_H__
+
+#ifdef CONFIG_MLX5_CLS_ACT
+
+struct mlx5_esw_indir_table *
+mlx5_esw_indir_table_init(void);
+void
+mlx5_esw_indir_table_destroy(struct mlx5_esw_indir_table *indir);
+
+struct mlx5_flow_table *mlx5_esw_indir_table_get(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec,
+ u16 vport, bool decap);
+void mlx5_esw_indir_table_put(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ u16 vport, bool decap);
+
+bool
+mlx5_esw_indir_table_needed(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ u16 vport_num,
+ struct mlx5_core_dev *dest_mdev);
+
+u16
+mlx5_esw_indir_table_decap_vport(struct mlx5_flow_attr *attr);
+
+#else
+/* indir API stubs */
+struct mlx5_esw_indir_table *
+mlx5_esw_indir_table_init(void)
+{
+ return NULL;
+}
+
+void
+mlx5_esw_indir_table_destroy(struct mlx5_esw_indir_table *indir)
+{
+}
+
+static inline struct mlx5_flow_table *
+mlx5_esw_indir_table_get(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec,
+ u16 vport, bool decap)
+{
+ return ERR_PTR(-EOPNOTSUPP);
+}
+
+static inline void
+mlx5_esw_indir_table_put(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ u16 vport, bool decap)
+{
+}
+
+bool
+mlx5_esw_indir_table_needed(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ u16 vport_num,
+ struct mlx5_core_dev *dest_mdev)
+{
+ return false;
+}
+
+static inline u16
+mlx5_esw_indir_table_decap_vport(struct mlx5_flow_attr *attr)
+{
+ return 0;
+}
+#endif
+
+#endif /* __MLX5_ESW_FT_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index 1ab34751329e..c2361c5b824c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -161,6 +161,8 @@ struct mlx5_vport {
struct devlink_port *dl_port;
};
+struct mlx5_esw_indir_table;
+
struct mlx5_eswitch_fdb {
union {
struct legacy_fdb {
@@ -191,6 +193,8 @@ struct mlx5_eswitch_fdb {
struct mutex lock;
} vports;
+ struct mlx5_esw_indir_table *indir;
+
} offloads;
};
u32 flags;
@@ -418,6 +422,7 @@ struct mlx5_esw_flow_attr {
struct mlx5_core_dev *mdev;
struct mlx5_termtbl_handle *termtbl;
} dests[MLX5_MAX_FLOW_FWD_VPORTS];
+ struct mlx5_rx_tun_attr *rx_tun_attr;
struct mlx5_pkt_reformat *decap_pkt_reformat;
};
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 1b18f624e04a..da843eab5c07 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -38,6 +38,7 @@
#include <linux/mlx5/fs.h>
#include "mlx5_core.h"
#include "eswitch.h"
+#include "esw/indir_table.h"
#include "esw/acl/ofld.h"
#include "rdma.h"
#include "en.h"
@@ -2342,12 +2343,20 @@ static void esw_destroy_uplink_offloads_acl_tables(struct mlx5_eswitch *esw)
static int esw_offloads_steering_init(struct mlx5_eswitch *esw)
{
+ struct mlx5_esw_indir_table *indir;
int err;
memset(&esw->fdb_table.offloads, 0, sizeof(struct offloads_fdb));
mutex_init(&esw->fdb_table.offloads.vports.lock);
hash_init(esw->fdb_table.offloads.vports.table);
+ indir = mlx5_esw_indir_table_init();
+ if (IS_ERR(indir)) {
+ err = PTR_ERR(indir);
+ goto create_indir_err;
+ }
+ esw->fdb_table.offloads.indir = indir;
+
err = esw_create_uplink_offloads_acl_tables(esw);
if (err)
goto create_acl_err;
@@ -2379,6 +2388,8 @@ static int esw_offloads_steering_init(struct mlx5_eswitch *esw)
create_offloads_err:
esw_destroy_uplink_offloads_acl_tables(esw);
create_acl_err:
+ mlx5_esw_indir_table_destroy(esw->fdb_table.offloads.indir);
+create_indir_err:
mutex_destroy(&esw->fdb_table.offloads.vports.lock);
return err;
}
@@ -2390,6 +2401,7 @@ static void esw_offloads_steering_cleanup(struct mlx5_eswitch *esw)
esw_destroy_restore_table(esw);
esw_destroy_offloads_table(esw);
esw_destroy_uplink_offloads_acl_tables(esw);
+ mlx5_esw_indir_table_destroy(esw->fdb_table.offloads.indir);
mutex_destroy(&esw->fdb_table.offloads.vports.lock);
}
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 08/17] net/mlx5e: Remove redundant match on tunnel destination mac
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (6 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 07/17] net/mlx5: E-Switch, Indirect table infrastructure Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 09/17] net/mlx5e: VF tunnel RX traffic offloading Saeed Mahameed
` (10 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Remove hardcoded match on tunnel destination MAC address. Such match is no
longer required and would be wrong for stacked devices topology where
encapsulation destination MAC address will be the address of tunnel VF that
can change dynamically on route change (implemented in following patches in
the series).
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 8 --------
1 file changed, 8 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
index 3e18ca200c86..13aa98b82576 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
@@ -642,14 +642,6 @@ int mlx5e_tc_tun_parse(struct net_device *filter_dev,
}
}
- /* Enforce DMAC when offloading incoming tunneled flows.
- * Flow counters require a match on the DMAC.
- */
- MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, dmac_47_16);
- MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, dmac_15_0);
- ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
- dmac_47_16), priv->netdev->dev_addr);
-
/* let software handle IP fragments */
MLX5_SET(fte_match_set_lyr_2_4, headers_c, frag, 1);
MLX5_SET(fte_match_set_lyr_2_4, headers_v, frag, 0);
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 09/17] net/mlx5e: VF tunnel RX traffic offloading
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (7 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 08/17] net/mlx5e: Remove redundant match on tunnel destination mac Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 10/17] net/mlx5e: Refactor reg_c1 usage Saeed Mahameed
` (9 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
When tunnel endpoint is on VF the encapsulated RX traffic is exposed on the
representor of the VF without any further processing of rules installed on
the VF. Detect such case by checking if the device returned by route lookup
in decap rule handling code is a mlx5 VF and handle it with new redirection
tables API.
Example TC rules for VF tunnel traffic:
1. Rule that encapsulates the tunneled flow and redirects packets from
source VF rep to tunnel device:
$ tc -s filter show dev enp8s0f0_1 ingress
filter protocol ip pref 4 flower chain 0
filter protocol ip pref 4 flower chain 0 handle 0x1
dst_mac 0a:40:bd:30:89:99
src_mac ca:2e:a7:3f:f5:0f
eth_type ipv4
ip_tos 0/0x3
ip_flags nofrag
in_hw in_hw_count 1
action order 1: tunnel_key set
src_ip 7.7.7.5
dst_ip 7.7.7.1
key_id 98
dst_port 4789
nocsum
ttl 64 pipe
index 1 ref 1 bind 1 installed 411 sec used 411 sec
Action statistics:
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
no_percpu
used_hw_stats delayed
action order 2: mirred (Egress Redirect to device vxlan_sys_4789) stolen
index 1 ref 1 bind 1 installed 411 sec used 0 sec
Action statistics:
Sent 5615833 bytes 4028 pkt (dropped 0, overlimits 0 requeues 0)
Sent software 0 bytes 0 pkt
Sent hardware 5615833 bytes 4028 pkt
backlog 0b 0p requeues 0
cookie bb406d45d343bf7ade9690ae80c7cba4
no_percpu
used_hw_stats delayed
2. Rule that redirects from tunnel device to UL rep:
$ tc -s filter show dev vxlan_sys_4789 ingress
filter protocol ip pref 4 flower chain 0
filter protocol ip pref 4 flower chain 0 handle 0x1
dst_mac ca:2e:a7:3f:f5:0f
src_mac 0a:40:bd:30:89:99
eth_type ipv4
enc_dst_ip 7.7.7.5
enc_src_ip 7.7.7.1
enc_key_id 98
enc_dst_port 4789
enc_tos 0
ip_flags nofrag
in_hw in_hw_count 1
action order 1: tunnel_key unset pipe
index 2 ref 1 bind 1 installed 434 sec used 434 sec
Action statistics:
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
used_hw_stats delayed
action order 2: mirred (Egress Redirect to device enp8s0f0_1) stolen
index 4 ref 1 bind 1 installed 434 sec used 0 sec
Action statistics:
Sent 129936 bytes 1082 pkt (dropped 0, overlimits 0 requeues 0)
Sent software 0 bytes 0 pkt
Sent hardware 129936 bytes 1082 pkt
backlog 0b 0p requeues 0
cookie ac17cf398c4c69e4a5b2f7aabd1b88ff
no_percpu
used_hw_stats delayed
Co-developed-by: Dmytro Linkin <dlinkin@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../ethernet/mellanox/mlx5/core/en/tc_tun.c | 51 ++++++++
.../ethernet/mellanox/mlx5/core/en/tc_tun.h | 3 +
.../net/ethernet/mellanox/mlx5/core/en_tc.c | 102 ++++++++++++++-
.../net/ethernet/mellanox/mlx5/core/en_tc.h | 4 +
.../mellanox/mlx5/core/eswitch_offloads.c | 119 +++++++++++++++++-
5 files changed, 271 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
index 13aa98b82576..73deafe4e693 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
@@ -483,6 +483,57 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
}
#endif
+int mlx5e_tc_tun_route_lookup(struct mlx5e_priv *priv,
+ struct mlx5_flow_spec *spec,
+ struct mlx5_flow_attr *flow_attr)
+{
+ struct mlx5_esw_flow_attr *esw_attr = flow_attr->esw_attr;
+ TC_TUN_ROUTE_ATTR_INIT(attr);
+ u16 vport_num;
+ int err = 0;
+
+ if (flow_attr->ip_version == 4) {
+ /* Addresses are swapped for decap */
+ attr.fl.fl4.saddr = esw_attr->rx_tun_attr->dst_ip.v4;
+ attr.fl.fl4.daddr = esw_attr->rx_tun_attr->src_ip.v4;
+ err = mlx5e_route_lookup_ipv4_get(priv, priv->netdev, &attr);
+ }
+#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
+ else if (flow_attr->ip_version == 6) {
+ /* Addresses are swapped for decap */
+ attr.fl.fl6.saddr = esw_attr->rx_tun_attr->dst_ip.v6;
+ attr.fl.fl6.daddr = esw_attr->rx_tun_attr->src_ip.v6;
+ err = mlx5e_route_lookup_ipv6_get(priv, priv->netdev, &attr);
+ }
+#endif
+ else
+ return 0;
+
+ if (err)
+ return err;
+
+ if (attr.route_dev->netdev_ops != &mlx5e_netdev_ops ||
+ !mlx5e_tc_is_vf_tunnel(attr.out_dev, attr.route_dev))
+ goto out;
+
+ err = mlx5e_tc_query_route_vport(attr.out_dev, attr.route_dev, &vport_num);
+ if (err)
+ goto out;
+
+ esw_attr->rx_tun_attr->vni = MLX5_GET(fte_match_param, spec->match_value,
+ misc_parameters.vxlan_vni);
+ esw_attr->rx_tun_attr->decap_vport = vport_num;
+
+out:
+ if (flow_attr->ip_version == 4)
+ mlx5e_route_lookup_ipv4_put(&attr);
+#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
+ else if (flow_attr->ip_version == 6)
+ mlx5e_route_lookup_ipv6_put(&attr);
+#endif
+ return err;
+}
+
bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv,
struct net_device *netdev)
{
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h
index 704359df6095..9d6ee9405eaf 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h
@@ -70,6 +70,9 @@ mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
struct net_device *mirred_dev,
struct mlx5e_encap_entry *e) { return -EOPNOTSUPP; }
#endif
+int mlx5e_tc_tun_route_lookup(struct mlx5e_priv *priv,
+ struct mlx5_flow_spec *spec,
+ struct mlx5_flow_attr *attr);
bool mlx5e_tc_tun_device_to_offload(struct mlx5e_priv *priv,
struct net_device *netdev);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 43f1508a05b5..098f3efa5d4d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -1322,7 +1322,7 @@ static void remove_unready_flow(struct mlx5e_tc_flow *flow)
static bool same_hw_devs(struct mlx5e_priv *priv, struct mlx5e_priv *peer_priv);
-static bool mlx5e_tc_is_vf_tunnel(struct net_device *out_dev, struct net_device *route_dev)
+bool mlx5e_tc_is_vf_tunnel(struct net_device *out_dev, struct net_device *route_dev)
{
struct mlx5_core_dev *out_mdev, *route_mdev;
struct mlx5e_priv *out_priv, *route_priv;
@@ -1339,8 +1339,7 @@ static bool mlx5e_tc_is_vf_tunnel(struct net_device *out_dev, struct net_device
return same_hw_devs(out_priv, route_priv);
}
-static int mlx5e_tc_query_route_vport(struct net_device *out_dev, struct net_device *route_dev,
- u16 *vport)
+int mlx5e_tc_query_route_vport(struct net_device *out_dev, struct net_device *route_dev, u16 *vport)
{
struct mlx5e_priv *out_priv, *route_priv;
struct mlx5_core_dev *route_mdev;
@@ -1504,6 +1503,7 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv,
kfree(attr->parse_attr->tun_info[out_index]);
}
kvfree(attr->parse_attr);
+ kvfree(attr->esw_attr->rx_tun_attr);
mlx5_tc_ct_match_del(get_ct_priv(priv), &flow->attr->ct_attr);
@@ -2134,6 +2134,67 @@ void mlx5e_tc_set_ethertype(struct mlx5_core_dev *mdev,
}
}
+static u8 mlx5e_tc_get_ip_version(struct mlx5_flow_spec *spec, bool outer)
+{
+ void *headers_v;
+ u16 ethertype;
+ u8 ip_version;
+
+ if (outer)
+ headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, outer_headers);
+ else
+ headers_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, inner_headers);
+
+ ip_version = MLX5_GET(fte_match_set_lyr_2_4, headers_v, ip_version);
+ /* Return ip_version converted from ethertype anyway */
+ if (!ip_version) {
+ ethertype = MLX5_GET(fte_match_set_lyr_2_4, headers_v, ethertype);
+ if (ethertype == ETH_P_IP || ethertype == ETH_P_ARP)
+ ip_version = 4;
+ else if (ethertype == ETH_P_IPV6)
+ ip_version = 6;
+ }
+ return ip_version;
+}
+
+static int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
+ struct mlx5_flow_spec *spec)
+{
+ struct mlx5_esw_flow_attr *esw_attr = flow->attr->esw_attr;
+ struct mlx5_rx_tun_attr *tun_attr;
+ void *daddr, *saddr;
+ u8 ip_version;
+
+ tun_attr = kvzalloc(sizeof(*tun_attr), GFP_KERNEL);
+ if (!tun_attr)
+ return -ENOMEM;
+
+ esw_attr->rx_tun_attr = tun_attr;
+ ip_version = mlx5e_tc_get_ip_version(spec, true);
+
+ if (ip_version == 4) {
+ daddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+ outer_headers.dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+ saddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+ outer_headers.src_ipv4_src_ipv6.ipv4_layout.ipv4);
+ tun_attr->dst_ip.v4 = *(__be32 *)daddr;
+ tun_attr->src_ip.v4 = *(__be32 *)saddr;
+ }
+#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
+ else if (ip_version == 6) {
+ int ipv6_size = MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6);
+
+ daddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+ outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6);
+ saddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+ outer_headers.src_ipv4_src_ipv6.ipv6_layout.ipv6);
+ memcpy(&tun_attr->dst_ip.v6, daddr, ipv6_size);
+ memcpy(&tun_attr->src_ip.v6, saddr, ipv6_size);
+ }
+#endif
+ return 0;
+}
+
static int parse_tunnel_attr(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow,
struct mlx5_flow_spec *spec,
@@ -2142,6 +2203,7 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
u8 *match_level,
bool *match_inner)
{
+ struct mlx5e_tc_tunnel *tunnel = mlx5e_get_tc_tun(filter_dev);
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
struct netlink_ext_ack *extack = f->common.extack;
bool needs_mapping, sets_mapping;
@@ -2179,6 +2241,31 @@ static int parse_tunnel_attr(struct mlx5e_priv *priv,
*/
if (!netif_is_bareudp(filter_dev))
flow->attr->action |= MLX5_FLOW_CONTEXT_ACTION_DECAP;
+ err = mlx5e_tc_set_attr_rx_tun(flow, spec);
+ if (err)
+ return err;
+ } else if (tunnel && tunnel->tunnel_type == MLX5E_TC_TUNNEL_TYPE_VXLAN) {
+ struct mlx5_flow_spec *tmp_spec;
+
+ tmp_spec = kvzalloc(sizeof(*tmp_spec), GFP_KERNEL);
+ if (!tmp_spec) {
+ NL_SET_ERR_MSG_MOD(extack, "Failed to allocate memory for vxlan tmp spec");
+ netdev_warn(priv->netdev, "Failed to allocate memory for vxlan tmp spec");
+ return -ENOMEM;
+ }
+ memcpy(tmp_spec, spec, sizeof(*tmp_spec));
+
+ err = mlx5e_tc_tun_parse(filter_dev, priv, tmp_spec, f, match_level);
+ if (err) {
+ kvfree(tmp_spec);
+ NL_SET_ERR_MSG_MOD(extack, "Failed to parse tunnel attributes");
+ netdev_warn(priv->netdev, "Failed to parse tunnel attributes");
+ return err;
+ }
+ err = mlx5e_tc_set_attr_rx_tun(flow, tmp_spec);
+ kvfree(tmp_spec);
+ if (err)
+ return err;
}
if (!needs_mapping && !sets_mapping)
@@ -4473,6 +4560,15 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv,
}
}
+ if (decap && esw_attr->rx_tun_attr) {
+ err = mlx5e_tc_tun_route_lookup(priv, &parse_attr->spec, attr);
+ if (err)
+ return err;
+ }
+
+ /* always set IP version for indirect table handling */
+ attr->ip_version = mlx5e_tc_get_ip_version(&parse_attr->spec, true);
+
if (MLX5_CAP_GEN(esw->dev, prio_tag_required) &&
action & MLX5_FLOW_CONTEXT_ACTION_VLAN_POP) {
/* For prio tag mode, replace vlan pop with rewrite vlan prio
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
index 852e0981343d..ee0029192504 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -257,6 +257,10 @@ mlx5_tc_rule_delete(struct mlx5e_priv *priv,
struct mlx5_flow_handle *rule,
struct mlx5_flow_attr *attr);
+bool mlx5e_tc_is_vf_tunnel(struct net_device *out_dev, struct net_device *route_dev);
+int mlx5e_tc_query_route_vport(struct net_device *out_dev, struct net_device *route_dev,
+ u16 *vport);
+
#else /* CONFIG_MLX5_CLS_ACT */
static inline int mlx5e_tc_nic_init(struct mlx5e_priv *priv) { return 0; }
static inline void mlx5e_tc_nic_cleanup(struct mlx5e_priv *priv) {}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index da843eab5c07..a44728595420 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -40,6 +40,7 @@
#include "eswitch.h"
#include "esw/indir_table.h"
#include "esw/acl/ofld.h"
+#include "esw/indir_table.h"
#include "rdma.h"
#include "en.h"
#include "fs_core.h"
@@ -258,6 +259,7 @@ mlx5_eswitch_set_rule_flow_source(struct mlx5_eswitch *esw,
static void
mlx5_eswitch_set_rule_source_port(struct mlx5_eswitch *esw,
struct mlx5_flow_spec *spec,
+ struct mlx5_flow_attr *attr,
struct mlx5_eswitch *src_esw,
u16 vport)
{
@@ -268,6 +270,8 @@ mlx5_eswitch_set_rule_source_port(struct mlx5_eswitch *esw,
* VHCA in dual-port RoCE mode, and matching on source vport may fail.
*/
if (mlx5_eswitch_vport_match_metadata_enabled(esw)) {
+ if (mlx5_esw_indir_table_decap_vport(attr))
+ vport = mlx5_esw_indir_table_decap_vport(attr);
misc2 = MLX5_ADDR_OF(fte_match_param, spec->match_value, misc_parameters_2);
MLX5_SET(fte_match_set_misc2, misc2, metadata_reg_c_0,
mlx5_eswitch_get_vport_metadata_for_match(src_esw,
@@ -297,15 +301,46 @@ mlx5_eswitch_set_rule_source_port(struct mlx5_eswitch *esw,
}
}
+static int
+esw_setup_decap_indir(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec)
+{
+ struct mlx5_flow_table *ft;
+
+ if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SRC_REWRITE))
+ return -EOPNOTSUPP;
+
+ ft = mlx5_esw_indir_table_get(esw, attr, spec,
+ mlx5_esw_indir_table_decap_vport(attr), true);
+ return PTR_ERR_OR_ZERO(ft);
+}
+
static void
+esw_cleanup_decap_indir(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr)
+{
+ if (mlx5_esw_indir_table_decap_vport(attr))
+ mlx5_esw_indir_table_put(esw, attr,
+ mlx5_esw_indir_table_decap_vport(attr),
+ true);
+}
+
+static int
esw_setup_ft_dest(struct mlx5_flow_destination *dest,
struct mlx5_flow_act *flow_act,
+ struct mlx5_eswitch *esw,
struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec,
int i)
{
flow_act->flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
dest[i].ft = attr->dest_ft;
+
+ if (mlx5_esw_indir_table_decap_vport(attr))
+ return esw_setup_decap_indir(esw, attr, spec);
+ return 0;
}
static void
@@ -348,6 +383,10 @@ static void esw_put_dest_tables_loop(struct mlx5_eswitch *esw, struct mlx5_flow_
for (i = from; i < to; i++)
if (esw_attr->dests[i].flags & MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE)
mlx5_chains_put_table(chains, 0, 1, 0);
+ else if (mlx5_esw_indir_table_needed(esw, attr, esw_attr->dests[i].rep->vport,
+ esw_attr->dests[i].mdev))
+ mlx5_esw_indir_table_put(esw, attr, esw_attr->dests[i].rep->vport,
+ false);
}
static bool
@@ -397,6 +436,68 @@ static void esw_cleanup_chain_src_port_rewrite(struct mlx5_eswitch *esw,
esw_put_dest_tables_loop(esw, attr, esw_attr->split_count, esw_attr->out_count);
}
+static bool
+esw_is_indir_table(struct mlx5_eswitch *esw, struct mlx5_flow_attr *attr)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+ int i;
+
+ for (i = esw_attr->split_count; i < esw_attr->out_count; i++)
+ if (mlx5_esw_indir_table_needed(esw, attr, esw_attr->dests[i].rep->vport,
+ esw_attr->dests[i].mdev))
+ return true;
+ return false;
+}
+
+static int
+esw_setup_indir_table(struct mlx5_flow_destination *dest,
+ struct mlx5_flow_act *flow_act,
+ struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5_flow_spec *spec,
+ bool ignore_flow_lvl,
+ int *i)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+ int j, err;
+
+ if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SRC_REWRITE))
+ return -EOPNOTSUPP;
+
+ for (j = esw_attr->split_count; j < esw_attr->out_count; j++, (*i)++) {
+ if (ignore_flow_lvl)
+ flow_act->flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
+ dest[*i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+
+ dest[*i].ft = mlx5_esw_indir_table_get(esw, attr, spec,
+ esw_attr->dests[j].rep->vport, false);
+ if (IS_ERR(dest[*i].ft)) {
+ err = PTR_ERR(dest[*i].ft);
+ goto err_indir_tbl_get;
+ }
+ }
+
+ if (mlx5_esw_indir_table_decap_vport(attr)) {
+ err = esw_setup_decap_indir(esw, attr, spec);
+ if (err)
+ goto err_indir_tbl_get;
+ }
+
+ return 0;
+
+err_indir_tbl_get:
+ esw_put_dest_tables_loop(esw, attr, esw_attr->split_count, j);
+ return err;
+}
+
+static void esw_cleanup_indir_table(struct mlx5_eswitch *esw, struct mlx5_flow_attr *attr)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+
+ esw_put_dest_tables_loop(esw, attr, esw_attr->split_count, esw_attr->out_count);
+ esw_cleanup_decap_indir(esw, attr);
+}
+
static void
esw_cleanup_chain_dest(struct mlx5_fs_chains *chains, u32 chain, u32 prio, u32 level)
{
@@ -454,7 +555,7 @@ esw_setup_dests(struct mlx5_flow_destination *dest,
attr->flags |= MLX5_ESW_ATTR_FLAG_SRC_REWRITE;
if (attr->dest_ft) {
- esw_setup_ft_dest(dest, flow_act, attr, *i);
+ esw_setup_ft_dest(dest, flow_act, esw, attr, spec, *i);
(*i)++;
} else if (attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH) {
esw_setup_slow_path_dest(dest, flow_act, chains, *i);
@@ -463,6 +564,8 @@ esw_setup_dests(struct mlx5_flow_destination *dest,
err = esw_setup_chain_dest(dest, flow_act, chains, attr->dest_chain,
1, 0, *i);
(*i)++;
+ } else if (esw_is_indir_table(esw, attr)) {
+ err = esw_setup_indir_table(dest, flow_act, esw, attr, spec, true, i);
} else if (esw_is_chain_src_port_rewrite(esw, esw_attr)) {
err = esw_setup_chain_src_port_rewrite(dest, flow_act, esw, chains, attr, i);
} else {
@@ -479,9 +582,13 @@ esw_cleanup_dests(struct mlx5_eswitch *esw,
struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
struct mlx5_fs_chains *chains = esw_chains(esw);
- if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH)) {
+ if (attr->dest_ft) {
+ esw_cleanup_decap_indir(esw, attr);
+ } else if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH)) {
if (attr->dest_chain)
esw_cleanup_chain_dest(chains, attr->dest_chain, 1, 0);
+ else if (esw_is_indir_table(esw, attr))
+ esw_cleanup_indir_table(esw, attr);
else if (esw_is_chain_src_port_rewrite(esw, esw_attr))
esw_cleanup_chain_src_port_rewrite(esw, attr);
}
@@ -564,7 +671,7 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
fdb = attr->ft;
if (!(attr->flags & MLX5_ESW_ATTR_FLAG_NO_IN_PORT))
- mlx5_eswitch_set_rule_source_port(esw, spec,
+ mlx5_eswitch_set_rule_source_port(esw, spec, attr,
esw_attr->in_mdev->priv.eswitch,
esw_attr->in_rep->vport);
}
@@ -628,7 +735,9 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
for (i = 0; i < esw_attr->split_count; i++) {
- if (esw_is_chain_src_port_rewrite(esw, esw_attr))
+ if (esw_is_indir_table(esw, attr))
+ err = esw_setup_indir_table(dest, &flow_act, esw, attr, spec, false, &i);
+ else if (esw_is_chain_src_port_rewrite(esw, esw_attr))
err = esw_setup_chain_src_port_rewrite(dest, &flow_act, esw, chains, attr,
&i);
else
@@ -643,7 +752,7 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
dest[i].ft = fwd_fdb;
i++;
- mlx5_eswitch_set_rule_source_port(esw, spec,
+ mlx5_eswitch_set_rule_source_port(esw, spec, attr,
esw_attr->in_mdev->priv.eswitch,
esw_attr->in_rep->vport);
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 10/17] net/mlx5e: Refactor reg_c1 usage
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (8 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 09/17] net/mlx5e: VF tunnel RX traffic offloading Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 11/17] net/mlx5e: Match recirculated packet miss in slow table using reg_c1 Saeed Mahameed
` (8 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Following patch in series uses reg_c1 in eswitch code. To use reg_c1
helpers in both TC and eswitch code, refactor existing helpers according to
similar use case of reg_c0 and move the functionality into eswitch.h.
Calculate reg mappings length from new defines to ensure that they are
always in sync and only need to be changed in single place.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../ethernet/mellanox/mlx5/core/en/rep/tc.c | 4 ++--
.../ethernet/mellanox/mlx5/core/en/tc_ct.h | 6 ++----
.../net/ethernet/mellanox/mlx5/core/en_tc.c | 4 ++--
.../net/ethernet/mellanox/mlx5/core/en_tc.h | 2 +-
include/linux/mlx5/eswitch.h | 19 +++++++++++++++++++
5 files changed, 26 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
index 76177f7c5ec2..14bcebd4a0b6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
@@ -651,7 +651,7 @@ bool mlx5e_rep_tc_update_skb(struct mlx5_cqe64 *cqe,
tc_skb_ext->chain = chain;
- zone_restore_id = reg_c1 & ZONE_RESTORE_MAX;
+ zone_restore_id = reg_c1 & ESW_ZONE_ID_MASK;
uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
uplink_priv = &uplink_rpriv->uplink_priv;
@@ -660,7 +660,7 @@ bool mlx5e_rep_tc_update_skb(struct mlx5_cqe64 *cqe,
return false;
}
- tunnel_id = reg_c1 >> REG_MAPPING_SHIFT(TUNNEL_TO_REG);
+ tunnel_id = reg_c1 >> ESW_TUN_OFFSET;
return mlx5e_restore_tunnel(priv, skb, tc_priv, tunnel_id);
#endif /* CONFIG_NET_TC_SKB_EXT */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.h
index 6503b614337c..69e618d17071 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.h
@@ -73,7 +73,7 @@ struct mlx5_ct_attr {
#define zone_restore_to_reg_ct {\
.mfield = MLX5_ACTION_IN_FIELD_METADATA_REG_C_1,\
.moffset = 0,\
- .mlen = 1,\
+ .mlen = (ESW_ZONE_ID_BITS / 8),\
.soffset = MLX5_BYTE_OFF(fte_match_param,\
misc_parameters_2.metadata_reg_c_1) + 3,\
}
@@ -81,14 +81,12 @@ struct mlx5_ct_attr {
#define nic_zone_restore_to_reg_ct {\
.mfield = MLX5_ACTION_IN_FIELD_METADATA_REG_B,\
.moffset = 2,\
- .mlen = 1,\
+ .mlen = (ESW_ZONE_ID_BITS / 8),\
}
#define REG_MAPPING_MLEN(reg) (mlx5e_tc_attr_to_reg_mappings[reg].mlen)
#define REG_MAPPING_MOFFSET(reg) (mlx5e_tc_attr_to_reg_mappings[reg].moffset)
#define REG_MAPPING_SHIFT(reg) (REG_MAPPING_MOFFSET(reg) * 8)
-#define ZONE_RESTORE_BITS (REG_MAPPING_MLEN(ZONE_RESTORE_TO_REG) * 8)
-#define ZONE_RESTORE_MAX GENMASK(ZONE_RESTORE_BITS - 1, 0)
#if IS_ENABLED(CONFIG_MLX5_TC_CT)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 098f3efa5d4d..90db5a99879d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -173,7 +173,7 @@ struct mlx5e_tc_attr_to_reg_mapping mlx5e_tc_attr_to_reg_mappings[] = {
[TUNNEL_TO_REG] = {
.mfield = MLX5_ACTION_IN_FIELD_METADATA_REG_C_1,
.moffset = 1,
- .mlen = 3,
+ .mlen = ((ESW_TUN_OPTS_BITS + ESW_TUN_ID_BITS) / 8),
.soffset = MLX5_BYTE_OFF(fte_match_param,
misc_parameters_2.metadata_reg_c_1),
},
@@ -5649,7 +5649,7 @@ bool mlx5e_tc_update_skb(struct mlx5_cqe64 *cqe,
tc_skb_ext->chain = chain;
zone_restore_id = (reg_b >> REG_MAPPING_SHIFT(NIC_ZONE_RESTORE_TO_REG)) &
- ZONE_RESTORE_MAX;
+ ESW_ZONE_ID_MASK;
if (!mlx5e_tc_ct_restore_flow(tc->ct, skb,
zone_restore_id))
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
index ee0029192504..1e4ee02bfb1c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -302,7 +302,7 @@ static inline bool mlx5e_cqe_regb_chain(struct mlx5_cqe64 *cqe)
reg_b = be32_to_cpu(cqe->ft_metadata);
- if (reg_b >> (MLX5E_TC_TABLE_CHAIN_TAG_BITS + ZONE_RESTORE_BITS))
+ if (reg_b >> (MLX5E_TC_TABLE_CHAIN_TAG_BITS + ESW_ZONE_ID_BITS))
return false;
chain = reg_b & MLX5E_TC_TABLE_CHAIN_TAG_MASK;
diff --git a/include/linux/mlx5/eswitch.h b/include/linux/mlx5/eswitch.h
index 67e341274a22..3b20e84049c1 100644
--- a/include/linux/mlx5/eswitch.h
+++ b/include/linux/mlx5/eswitch.h
@@ -98,6 +98,25 @@ u32 mlx5_eswitch_get_vport_metadata_for_match(struct mlx5_eswitch *esw,
u16 vport_num);
u32 mlx5_eswitch_get_vport_metadata_for_set(struct mlx5_eswitch *esw,
u16 vport_num);
+
+/* Reg C1 usage:
+ * Reg C1 = < ESW_TUN_ID(12) | ESW_TUN_OPTS(12) | ESW_ZONE_ID(8) >
+ *
+ * Highest 12 bits of reg c1 is the encapsulation tunnel id, next 12 bits is
+ * encapsulation tunnel options, and the lowest 8 bits are used for zone id.
+ *
+ * Zone id is used to restore CT flow when packet misses on chain.
+ *
+ * Tunnel id and options are used together to restore the tunnel info metadata
+ * on miss and to support inner header rewrite by means of implicit chain 0
+ * flows.
+ */
+#define ESW_ZONE_ID_BITS 8
+#define ESW_TUN_OPTS_BITS 12
+#define ESW_TUN_ID_BITS 12
+#define ESW_TUN_OFFSET ESW_ZONE_ID_BITS
+#define ESW_ZONE_ID_MASK GENMASK(ESW_ZONE_ID_BITS - 1, 0)
+
u8 mlx5_eswitch_mode(struct mlx5_core_dev *dev);
#else /* CONFIG_MLX5_ESWITCH */
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 11/17] net/mlx5e: Match recirculated packet miss in slow table using reg_c1
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (9 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 10/17] net/mlx5e: Refactor reg_c1 usage Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 12/17] net/mlx5e: Extract tc tunnel encap/decap code to dedicated file Saeed Mahameed
` (7 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Previous patch in series that implements stack devices RX path implements
indirect table rules that match on tunnel VNI. After such rule is created
all tunnel traffic is recirculated to root table. However, recirculated
packet might not match on any rules installed in the table (for example,
when IP traffic follows ARP traffic). In that case packets appear on
representor of tunnel endpoint VF instead being redirected to the VF
itself.
Extend slow table with additional flow group that matches on reg_c0 (source
port value set by indirect tables implemented by previous patch in series)
and reg_c1 (special 0xFFF mark). When creating offloads fdb tables, install
one rule per VF vport to match on recirculated miss packets and redirect
them to appropriate VF vport. Modify indirect tables code to also rewrite
reg_c1 with special 0xFFF mark.
Implementation reuses reg_c1 tunnel id bits. This is safe to do because
recirculated packets are always matched before decapsulation.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../net/ethernet/mellanox/mlx5/core/en_tc.c | 3 +-
.../mellanox/mlx5/core/esw/indir_table.c | 17 ++-
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 2 +
.../mellanox/mlx5/core/eswitch_offloads.c | 119 +++++++++++++++++-
include/linux/mlx5/eswitch.h | 10 +-
5 files changed, 143 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 90db5a99879d..21568d1fc00f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -5515,7 +5515,8 @@ int mlx5e_tc_esw_init(struct rhashtable *tc_ht)
}
uplink_priv->tunnel_mapping = mapping;
- mapping = mapping_create(sz_enc_opts, ENC_OPTS_BITS_MASK, true);
+ /* 0xFFF is reserved for stack devices slow path table mark */
+ mapping = mapping_create(sz_enc_opts, ENC_OPTS_BITS_MASK - 1, true);
if (IS_ERR(mapping)) {
err = PTR_ERR(mapping);
goto err_enc_opts_mapping;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.c
index a0ebf40c9907..b7d00c4c7046 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/indir_table.c
@@ -162,7 +162,7 @@ static int mlx5_esw_indir_table_rule_get(struct mlx5_eswitch *esw,
(attr->ip_version == 4 ? ETH_P_IP : ETH_P_IPV6));
} else {
err = -EOPNOTSUPP;
- goto err_mod_hdr;
+ goto err_ethertype;
}
if (attr->ip_version == 4) {
@@ -198,13 +198,18 @@ static int mlx5_esw_indir_table_rule_get(struct mlx5_eswitch *esw,
err = mlx5e_tc_match_to_reg_set(esw->dev, &mod_acts, MLX5_FLOW_NAMESPACE_FDB,
VPORT_TO_REG, data);
if (err)
- goto err_mod_hdr;
+ goto err_mod_hdr_regc0;
+
+ err = mlx5e_tc_match_to_reg_set(esw->dev, &mod_acts, MLX5_FLOW_NAMESPACE_FDB,
+ TUNNEL_TO_REG, ESW_TUN_SLOW_TABLE_GOTO_VPORT);
+ if (err)
+ goto err_mod_hdr_regc1;
flow_act.modify_hdr = mlx5_modify_header_alloc(esw->dev, MLX5_FLOW_NAMESPACE_FDB,
mod_acts.num_actions, mod_acts.actions);
if (IS_ERR(flow_act.modify_hdr)) {
err = PTR_ERR(flow_act.modify_hdr);
- goto err_mod_hdr;
+ goto err_mod_hdr_alloc;
}
flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | MLX5_FLOW_CONTEXT_ACTION_MOD_HDR;
@@ -236,7 +241,11 @@ static int mlx5_esw_indir_table_rule_get(struct mlx5_eswitch *esw,
mlx5_chains_put_table(chains, 0, 1, 0);
err_table:
mlx5_modify_header_dealloc(esw->dev, flow_act.modify_hdr);
-err_mod_hdr:
+err_mod_hdr_alloc:
+err_mod_hdr_regc1:
+ dealloc_mod_hdr_actions(&mod_acts);
+err_mod_hdr_regc0:
+err_ethertype:
kfree(rule);
out:
kfree(rule_spec);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index c2361c5b824c..34a21ff95472 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -179,9 +179,11 @@ struct mlx5_eswitch_fdb {
struct mlx5_flow_namespace *ns;
struct mlx5_flow_table *slow_fdb;
struct mlx5_flow_group *send_to_vport_grp;
+ struct mlx5_flow_group *send_to_vport_meta_grp;
struct mlx5_flow_group *peer_miss_grp;
struct mlx5_flow_handle **peer_miss_rules;
struct mlx5_flow_group *miss_grp;
+ struct mlx5_flow_handle **send_to_vport_meta_rules;
struct mlx5_flow_handle *miss_rule_uni;
struct mlx5_flow_handle *miss_rule_multi;
int vlan_push_pop_refcount;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index a44728595420..94cb0217b4f3 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -1080,6 +1080,81 @@ void mlx5_eswitch_del_send_to_vport_rule(struct mlx5_flow_handle *rule)
mlx5_del_flow_rules(rule);
}
+static void mlx5_eswitch_del_send_to_vport_meta_rules(struct mlx5_eswitch *esw)
+{
+ struct mlx5_flow_handle **flows = esw->fdb_table.offloads.send_to_vport_meta_rules;
+ int i = 0, num_vfs = esw->esw_funcs.num_vfs, vport_num;
+
+ if (!num_vfs || !flows)
+ return;
+
+ mlx5_esw_for_each_vf_vport_num(esw, vport_num, num_vfs)
+ mlx5_del_flow_rules(flows[i++]);
+
+ kvfree(flows);
+}
+
+static int
+mlx5_eswitch_add_send_to_vport_meta_rules(struct mlx5_eswitch *esw)
+{
+ int num_vfs, vport_num, rule_idx = 0, err = 0;
+ struct mlx5_flow_destination dest = {};
+ struct mlx5_flow_act flow_act = {0};
+ struct mlx5_flow_handle *flow_rule;
+ struct mlx5_flow_handle **flows;
+ struct mlx5_flow_spec *spec;
+
+ num_vfs = esw->esw_funcs.num_vfs;
+ flows = kvzalloc(num_vfs * sizeof(*flows), GFP_KERNEL);
+ if (!flows)
+ return -ENOMEM;
+
+ spec = kvzalloc(sizeof(*spec), GFP_KERNEL);
+ if (!spec) {
+ err = -ENOMEM;
+ goto alloc_err;
+ }
+
+ MLX5_SET(fte_match_param, spec->match_criteria,
+ misc_parameters_2.metadata_reg_c_0, mlx5_eswitch_get_vport_metadata_mask());
+ MLX5_SET(fte_match_param, spec->match_criteria,
+ misc_parameters_2.metadata_reg_c_1, ESW_TUN_MASK);
+ MLX5_SET(fte_match_param, spec->match_value, misc_parameters_2.metadata_reg_c_1,
+ ESW_TUN_SLOW_TABLE_GOTO_VPORT_MARK);
+
+ spec->match_criteria_enable = MLX5_MATCH_MISC_PARAMETERS_2;
+ dest.type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
+ flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+
+ mlx5_esw_for_each_vf_vport_num(esw, vport_num, num_vfs) {
+ MLX5_SET(fte_match_param, spec->match_value, misc_parameters_2.metadata_reg_c_0,
+ mlx5_eswitch_get_vport_metadata_for_match(esw, vport_num));
+ dest.vport.num = vport_num;
+
+ flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb,
+ spec, &flow_act, &dest, 1);
+ if (IS_ERR(flow_rule)) {
+ err = PTR_ERR(flow_rule);
+ esw_warn(esw->dev, "FDB: Failed to add send to vport meta rule idx %d, err %ld\n",
+ rule_idx, PTR_ERR(flow_rule));
+ goto rule_err;
+ }
+ flows[rule_idx++] = flow_rule;
+ }
+
+ esw->fdb_table.offloads.send_to_vport_meta_rules = flows;
+ kvfree(spec);
+ return 0;
+
+rule_err:
+ while (--rule_idx >= 0)
+ mlx5_del_flow_rules(flows[rule_idx]);
+ kvfree(spec);
+alloc_err:
+ kvfree(flows);
+ return err;
+}
+
static bool mlx5_eswitch_reg_c1_loopback_supported(struct mlx5_eswitch *esw)
{
return MLX5_CAP_ESW_FLOWTABLE(esw->dev, fdb_to_vport_reg_c_id) &
@@ -1562,11 +1637,11 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw)
{
int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
struct mlx5_flow_table_attr ft_attr = {};
+ int num_vfs, table_size, ix, err = 0;
struct mlx5_core_dev *dev = esw->dev;
struct mlx5_flow_namespace *root_ns;
struct mlx5_flow_table *fdb = NULL;
u32 flags = 0, *flow_group_in;
- int table_size, ix, err = 0;
struct mlx5_flow_group *g;
void *match_criteria;
u8 *dmac;
@@ -1592,7 +1667,7 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw)
}
table_size = esw->total_vports * MAX_SQ_NVPORTS + MAX_PF_SQ +
- MLX5_ESW_MISS_FLOWS + esw->total_vports;
+ MLX5_ESW_MISS_FLOWS + esw->total_vports + esw->esw_funcs.num_vfs;
/* create the slow path fdb with encap set, so further table instances
* can be created at run time while VFs are probed if the FW allows that.
@@ -1640,6 +1715,38 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw)
}
esw->fdb_table.offloads.send_to_vport_grp = g;
+ /* meta send to vport */
+ memset(flow_group_in, 0, inlen);
+ MLX5_SET(create_flow_group_in, flow_group_in, match_criteria_enable,
+ MLX5_MATCH_MISC_PARAMETERS_2);
+
+ match_criteria = MLX5_ADDR_OF(create_flow_group_in, flow_group_in, match_criteria);
+
+ MLX5_SET(fte_match_param, match_criteria,
+ misc_parameters_2.metadata_reg_c_0, mlx5_eswitch_get_vport_metadata_mask());
+ MLX5_SET(fte_match_param, match_criteria,
+ misc_parameters_2.metadata_reg_c_1, ESW_TUN_MASK);
+
+ num_vfs = esw->esw_funcs.num_vfs;
+ if (num_vfs) {
+ MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, ix);
+ MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, ix + num_vfs - 1);
+ ix += num_vfs;
+
+ g = mlx5_create_flow_group(fdb, flow_group_in);
+ if (IS_ERR(g)) {
+ err = PTR_ERR(g);
+ esw_warn(dev, "Failed to create send-to-vport meta flow group err(%d)\n",
+ err);
+ goto send_vport_meta_err;
+ }
+ esw->fdb_table.offloads.send_to_vport_meta_grp = g;
+
+ err = mlx5_eswitch_add_send_to_vport_meta_rules(esw);
+ if (err)
+ goto meta_rule_err;
+ }
+
if (MLX5_CAP_ESW(esw->dev, merged_eswitch)) {
/* create peer esw miss group */
memset(flow_group_in, 0, inlen);
@@ -1707,6 +1814,11 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw)
if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
mlx5_destroy_flow_group(esw->fdb_table.offloads.peer_miss_grp);
peer_miss_err:
+ mlx5_eswitch_del_send_to_vport_meta_rules(esw);
+meta_rule_err:
+ if (esw->fdb_table.offloads.send_to_vport_meta_grp)
+ mlx5_destroy_flow_group(esw->fdb_table.offloads.send_to_vport_meta_grp);
+send_vport_meta_err:
mlx5_destroy_flow_group(esw->fdb_table.offloads.send_to_vport_grp);
send_vport_err:
esw_chains_destroy(esw, esw_chains(esw));
@@ -1728,7 +1840,10 @@ static void esw_destroy_offloads_fdb_tables(struct mlx5_eswitch *esw)
esw_debug(esw->dev, "Destroy offloads FDB Tables\n");
mlx5_del_flow_rules(esw->fdb_table.offloads.miss_rule_multi);
mlx5_del_flow_rules(esw->fdb_table.offloads.miss_rule_uni);
+ mlx5_eswitch_del_send_to_vport_meta_rules(esw);
mlx5_destroy_flow_group(esw->fdb_table.offloads.send_to_vport_grp);
+ if (esw->fdb_table.offloads.send_to_vport_meta_grp)
+ mlx5_destroy_flow_group(esw->fdb_table.offloads.send_to_vport_meta_grp);
if (MLX5_CAP_ESW(esw->dev, merged_eswitch))
mlx5_destroy_flow_group(esw->fdb_table.offloads.peer_miss_grp);
mlx5_destroy_flow_group(esw->fdb_table.offloads.miss_grp);
diff --git a/include/linux/mlx5/eswitch.h b/include/linux/mlx5/eswitch.h
index 3b20e84049c1..994c2c8cb4fd 100644
--- a/include/linux/mlx5/eswitch.h
+++ b/include/linux/mlx5/eswitch.h
@@ -114,8 +114,16 @@ u32 mlx5_eswitch_get_vport_metadata_for_set(struct mlx5_eswitch *esw,
#define ESW_ZONE_ID_BITS 8
#define ESW_TUN_OPTS_BITS 12
#define ESW_TUN_ID_BITS 12
-#define ESW_TUN_OFFSET ESW_ZONE_ID_BITS
+#define ESW_TUN_OPTS_OFFSET ESW_ZONE_ID_BITS
+#define ESW_TUN_OFFSET ESW_TUN_OPTS_OFFSET
#define ESW_ZONE_ID_MASK GENMASK(ESW_ZONE_ID_BITS - 1, 0)
+#define ESW_TUN_OPTS_MASK GENMASK(32 - ESW_TUN_ID_BITS - 1, ESW_TUN_OPTS_OFFSET)
+#define ESW_TUN_MASK GENMASK(31, ESW_TUN_OFFSET)
+#define ESW_TUN_ID_SLOW_TABLE_GOTO_VPORT 0 /* 0 is not a valid tunnel id */
+#define ESW_TUN_OPTS_SLOW_TABLE_GOTO_VPORT 0xFFF /* 0xFFF is a reserved mapping */
+#define ESW_TUN_SLOW_TABLE_GOTO_VPORT ((ESW_TUN_ID_SLOW_TABLE_GOTO_VPORT << ESW_TUN_OPTS_BITS) | \
+ ESW_TUN_OPTS_SLOW_TABLE_GOTO_VPORT)
+#define ESW_TUN_SLOW_TABLE_GOTO_VPORT_MARK ESW_TUN_OPTS_MASK
u8 mlx5_eswitch_mode(struct mlx5_core_dev *dev);
#else /* CONFIG_MLX5_ESWITCH */
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 12/17] net/mlx5e: Extract tc tunnel encap/decap code to dedicated file
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (10 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 11/17] net/mlx5e: Match recirculated packet miss in slow table using reg_c1 Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-09 20:35 ` Guenter Roeck
2021-02-06 5:02 ` [net-next V2 13/17] net/mlx5e: Create route entry infrastructure Saeed Mahameed
` (6 subsequent siblings)
18 siblings, 1 reply; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Following patches in series extend the extracted code with routing
infrastructure. To improve code modularity created a dedicated
tc_tun_encap.c source file and move encap/decap related code to the new
file. Export code that is used by both regular TC code and encap/decap code
into tc_priv.h (new header intended to be used only by TC module). Rename
some exported functions by adding "mlx5e_" prefix to their names.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../net/ethernet/mellanox/mlx5/core/Makefile | 2 +-
.../ethernet/mellanox/mlx5/core/en/tc_priv.h | 163 ++++
.../ethernet/mellanox/mlx5/core/en/tc_tun.c | 1 +
.../mellanox/mlx5/core/en/tc_tun_encap.c | 734 ++++++++++++++
.../mellanox/mlx5/core/en/tc_tun_encap.h | 30 +
.../net/ethernet/mellanox/mlx5/core/en_tc.c | 900 +-----------------
.../net/ethernet/mellanox/mlx5/core/en_tc.h | 2 +
7 files changed, 947 insertions(+), 885 deletions(-)
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
index f1ccfba60068..8cb2625472c3 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -40,7 +40,7 @@ mlx5_core-$(CONFIG_MLX5_ESWITCH) += lag_mp.o lib/geneve.o lib/port_tun.o \
en_rep.o en/rep/bond.o en/mod_hdr.o
mlx5_core-$(CONFIG_MLX5_CLS_ACT) += en_tc.o en/rep/tc.o en/rep/neigh.o \
en/mapping.o lib/fs_chains.o en/tc_tun.o \
- esw/indir_table.o \
+ esw/indir_table.o en/tc_tun_encap.o \
en/tc_tun_vxlan.o en/tc_tun_gre.o en/tc_tun_geneve.o \
en/tc_tun_mplsoudp.o diag/en_tc_tracepoint.o
mlx5_core-$(CONFIG_MLX5_TC_CT) += en/tc_ct.o
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
new file mode 100644
index 000000000000..e0ae24d9a740
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
@@ -0,0 +1,163 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2021 Mellanox Technologies. */
+
+#ifndef __MLX5_EN_TC_PRIV_H__
+#define __MLX5_EN_TC_PRIV_H__
+
+#include "en_tc.h"
+
+#define MLX5E_TC_FLOW_BASE (MLX5E_TC_FLAG_LAST_EXPORTED_BIT + 1)
+
+#define MLX5E_TC_MAX_SPLITS 1
+
+enum {
+ MLX5E_TC_FLOW_FLAG_INGRESS = MLX5E_TC_FLAG_INGRESS_BIT,
+ MLX5E_TC_FLOW_FLAG_EGRESS = MLX5E_TC_FLAG_EGRESS_BIT,
+ MLX5E_TC_FLOW_FLAG_ESWITCH = MLX5E_TC_FLAG_ESW_OFFLOAD_BIT,
+ MLX5E_TC_FLOW_FLAG_FT = MLX5E_TC_FLAG_FT_OFFLOAD_BIT,
+ MLX5E_TC_FLOW_FLAG_NIC = MLX5E_TC_FLAG_NIC_OFFLOAD_BIT,
+ MLX5E_TC_FLOW_FLAG_OFFLOADED = MLX5E_TC_FLOW_BASE,
+ MLX5E_TC_FLOW_FLAG_HAIRPIN = MLX5E_TC_FLOW_BASE + 1,
+ MLX5E_TC_FLOW_FLAG_HAIRPIN_RSS = MLX5E_TC_FLOW_BASE + 2,
+ MLX5E_TC_FLOW_FLAG_SLOW = MLX5E_TC_FLOW_BASE + 3,
+ MLX5E_TC_FLOW_FLAG_DUP = MLX5E_TC_FLOW_BASE + 4,
+ MLX5E_TC_FLOW_FLAG_NOT_READY = MLX5E_TC_FLOW_BASE + 5,
+ MLX5E_TC_FLOW_FLAG_DELETED = MLX5E_TC_FLOW_BASE + 6,
+ MLX5E_TC_FLOW_FLAG_CT = MLX5E_TC_FLOW_BASE + 7,
+ MLX5E_TC_FLOW_FLAG_L3_TO_L2_DECAP = MLX5E_TC_FLOW_BASE + 8,
+};
+
+struct mlx5e_tc_flow_parse_attr {
+ const struct ip_tunnel_info *tun_info[MLX5_MAX_FLOW_FWD_VPORTS];
+ struct net_device *filter_dev;
+ struct mlx5_flow_spec spec;
+ struct mlx5e_tc_mod_hdr_acts mod_hdr_acts;
+ int mirred_ifindex[MLX5_MAX_FLOW_FWD_VPORTS];
+ struct ethhdr eth;
+};
+
+/* Helper struct for accessing a struct containing list_head array.
+ * Containing struct
+ * |- Helper array
+ * [0] Helper item 0
+ * |- list_head item 0
+ * |- index (0)
+ * [1] Helper item 1
+ * |- list_head item 1
+ * |- index (1)
+ * To access the containing struct from one of the list_head items:
+ * 1. Get the helper item from the list_head item using
+ * helper item =
+ * container_of(list_head item, helper struct type, list_head field)
+ * 2. Get the contining struct from the helper item and its index in the array:
+ * containing struct =
+ * container_of(helper item, containing struct type, helper field[index])
+ */
+struct encap_flow_item {
+ struct mlx5e_encap_entry *e; /* attached encap instance */
+ struct list_head list;
+ int index;
+};
+
+struct mlx5e_tc_flow {
+ struct rhash_head node;
+ struct mlx5e_priv *priv;
+ u64 cookie;
+ unsigned long flags;
+ struct mlx5_flow_handle *rule[MLX5E_TC_MAX_SPLITS + 1];
+
+ /* flows sharing the same reformat object - currently mpls decap */
+ struct list_head l3_to_l2_reformat;
+ struct mlx5e_decap_entry *decap_reformat;
+
+ /* Flow can be associated with multiple encap IDs.
+ * The number of encaps is bounded by the number of supported
+ * destinations.
+ */
+ struct encap_flow_item encaps[MLX5_MAX_FLOW_FWD_VPORTS];
+ struct mlx5e_tc_flow *peer_flow;
+ struct mlx5e_mod_hdr_handle *mh; /* attached mod header instance */
+ struct mlx5e_hairpin_entry *hpe; /* attached hairpin instance */
+ struct list_head hairpin; /* flows sharing the same hairpin */
+ struct list_head peer; /* flows with peer flow */
+ struct list_head unready; /* flows not ready to be offloaded (e.g
+ * due to missing route)
+ */
+ struct net_device *orig_dev; /* netdev adding flow first */
+ int tmp_efi_index;
+ struct list_head tmp_list; /* temporary flow list used by neigh update */
+ refcount_t refcnt;
+ struct rcu_head rcu_head;
+ struct completion init_done;
+ int tunnel_id; /* the mapped tunnel id of this flow */
+ struct mlx5_flow_attr *attr;
+};
+
+u8 mlx5e_tc_get_ip_version(struct mlx5_flow_spec *spec, bool outer);
+
+struct mlx5_flow_handle *
+mlx5e_tc_offload_fdb_rules(struct mlx5_eswitch *esw,
+ struct mlx5e_tc_flow *flow,
+ struct mlx5_flow_spec *spec,
+ struct mlx5_flow_attr *attr);
+
+bool mlx5e_is_offloaded_flow(struct mlx5e_tc_flow *flow);
+
+static inline void __flow_flag_set(struct mlx5e_tc_flow *flow, unsigned long flag)
+{
+ /* Complete all memory stores before setting bit. */
+ smp_mb__before_atomic();
+ set_bit(flag, &flow->flags);
+}
+
+#define flow_flag_set(flow, flag) __flow_flag_set(flow, MLX5E_TC_FLOW_FLAG_##flag)
+
+static inline bool __flow_flag_test_and_set(struct mlx5e_tc_flow *flow,
+ unsigned long flag)
+{
+ /* test_and_set_bit() provides all necessary barriers */
+ return test_and_set_bit(flag, &flow->flags);
+}
+
+#define flow_flag_test_and_set(flow, flag) \
+ __flow_flag_test_and_set(flow, \
+ MLX5E_TC_FLOW_FLAG_##flag)
+
+static inline void __flow_flag_clear(struct mlx5e_tc_flow *flow, unsigned long flag)
+{
+ /* Complete all memory stores before clearing bit. */
+ smp_mb__before_atomic();
+ clear_bit(flag, &flow->flags);
+}
+
+#define flow_flag_clear(flow, flag) __flow_flag_clear(flow, \
+ MLX5E_TC_FLOW_FLAG_##flag)
+
+static inline bool __flow_flag_test(struct mlx5e_tc_flow *flow, unsigned long flag)
+{
+ bool ret = test_bit(flag, &flow->flags);
+
+ /* Read fields of flow structure only after checking flags. */
+ smp_mb__after_atomic();
+ return ret;
+}
+
+#define flow_flag_test(flow, flag) __flow_flag_test(flow, \
+ MLX5E_TC_FLOW_FLAG_##flag)
+
+void mlx5e_tc_unoffload_from_slow_path(struct mlx5_eswitch *esw,
+ struct mlx5e_tc_flow *flow);
+struct mlx5_flow_handle *
+mlx5e_tc_offload_to_slow_path(struct mlx5_eswitch *esw,
+ struct mlx5e_tc_flow *flow,
+ struct mlx5_flow_spec *spec);
+void mlx5e_tc_unoffload_fdb_rules(struct mlx5_eswitch *esw,
+ struct mlx5e_tc_flow *flow,
+ struct mlx5_flow_attr *attr);
+
+struct mlx5e_tc_flow *mlx5e_flow_get(struct mlx5e_tc_flow *flow);
+void mlx5e_flow_put(struct mlx5e_priv *priv, struct mlx5e_tc_flow *flow);
+
+struct mlx5_fc *mlx5e_tc_get_counter(struct mlx5e_tc_flow *flow);
+
+#endif /* __MLX5_EN_TC_PRIV_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
index 73deafe4e693..def2335c39e4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
@@ -6,6 +6,7 @@
#include <net/geneve.h>
#include <net/bareudp.h>
#include "en/tc_tun.h"
+#include "en/tc_priv.h"
#include "en_tc.h"
#include "rep/tc.h"
#include "rep/neigh.h"
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
new file mode 100644
index 000000000000..63652911d56e
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
@@ -0,0 +1,734 @@
+// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
+/* Copyright (c) 2021 Mellanox Technologies. */
+
+#include "tc_tun_encap.h"
+#include "en_tc.h"
+#include "tc_tun.h"
+#include "rep/tc.h"
+#include "diag/en_tc_tracepoint.h"
+
+int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
+ struct mlx5_flow_spec *spec)
+{
+ struct mlx5_esw_flow_attr *esw_attr = flow->attr->esw_attr;
+ struct mlx5_rx_tun_attr *tun_attr;
+ void *daddr, *saddr;
+ u8 ip_version;
+
+ tun_attr = kvzalloc(sizeof(*tun_attr), GFP_KERNEL);
+ if (!tun_attr)
+ return -ENOMEM;
+
+ esw_attr->rx_tun_attr = tun_attr;
+ ip_version = mlx5e_tc_get_ip_version(spec, true);
+
+ if (ip_version == 4) {
+ daddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+ outer_headers.dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+ saddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+ outer_headers.src_ipv4_src_ipv6.ipv4_layout.ipv4);
+ tun_attr->dst_ip.v4 = *(__be32 *)daddr;
+ tun_attr->src_ip.v4 = *(__be32 *)saddr;
+ }
+#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
+ else if (ip_version == 6) {
+ int ipv6_size = MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6);
+
+ daddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+ outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6);
+ saddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+ outer_headers.src_ipv4_src_ipv6.ipv6_layout.ipv6);
+ memcpy(&tun_attr->dst_ip.v6, daddr, ipv6_size);
+ memcpy(&tun_attr->src_ip.v6, saddr, ipv6_size);
+ }
+#endif
+ return 0;
+}
+
+void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv,
+ struct mlx5e_encap_entry *e,
+ struct list_head *flow_list)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5_esw_flow_attr *esw_attr;
+ struct mlx5_flow_handle *rule;
+ struct mlx5_flow_attr *attr;
+ struct mlx5_flow_spec *spec;
+ struct mlx5e_tc_flow *flow;
+ int err;
+
+ e->pkt_reformat = mlx5_packet_reformat_alloc(priv->mdev,
+ e->reformat_type,
+ e->encap_size, e->encap_header,
+ MLX5_FLOW_NAMESPACE_FDB);
+ if (IS_ERR(e->pkt_reformat)) {
+ mlx5_core_warn(priv->mdev, "Failed to offload cached encapsulation header, %lu\n",
+ PTR_ERR(e->pkt_reformat));
+ return;
+ }
+ e->flags |= MLX5_ENCAP_ENTRY_VALID;
+ mlx5e_rep_queue_neigh_stats_work(priv);
+
+ list_for_each_entry(flow, flow_list, tmp_list) {
+ bool all_flow_encaps_valid = true;
+ int i;
+
+ if (!mlx5e_is_offloaded_flow(flow))
+ continue;
+ attr = flow->attr;
+ esw_attr = attr->esw_attr;
+ spec = &attr->parse_attr->spec;
+
+ esw_attr->dests[flow->tmp_efi_index].pkt_reformat = e->pkt_reformat;
+ esw_attr->dests[flow->tmp_efi_index].flags |= MLX5_ESW_DEST_ENCAP_VALID;
+ /* Flow can be associated with multiple encap entries.
+ * Before offloading the flow verify that all of them have
+ * a valid neighbour.
+ */
+ for (i = 0; i < MLX5_MAX_FLOW_FWD_VPORTS; i++) {
+ if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP))
+ continue;
+ if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP_VALID)) {
+ all_flow_encaps_valid = false;
+ break;
+ }
+ }
+ /* Do not offload flows with unresolved neighbors */
+ if (!all_flow_encaps_valid)
+ continue;
+ /* update from slow path rule to encap rule */
+ rule = mlx5e_tc_offload_fdb_rules(esw, flow, spec, attr);
+ if (IS_ERR(rule)) {
+ err = PTR_ERR(rule);
+ mlx5_core_warn(priv->mdev, "Failed to update cached encapsulation flow, %d\n",
+ err);
+ continue;
+ }
+
+ mlx5e_tc_unoffload_from_slow_path(esw, flow);
+ flow->rule[0] = rule;
+ /* was unset when slow path rule removed */
+ flow_flag_set(flow, OFFLOADED);
+ }
+}
+
+void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv,
+ struct mlx5e_encap_entry *e,
+ struct list_head *flow_list)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5_esw_flow_attr *esw_attr;
+ struct mlx5_flow_handle *rule;
+ struct mlx5_flow_attr *attr;
+ struct mlx5_flow_spec *spec;
+ struct mlx5e_tc_flow *flow;
+ int err;
+
+ list_for_each_entry(flow, flow_list, tmp_list) {
+ if (!mlx5e_is_offloaded_flow(flow))
+ continue;
+ attr = flow->attr;
+ esw_attr = attr->esw_attr;
+ spec = &attr->parse_attr->spec;
+
+ /* update from encap rule to slow path rule */
+ rule = mlx5e_tc_offload_to_slow_path(esw, flow, spec);
+ /* mark the flow's encap dest as non-valid */
+ esw_attr->dests[flow->tmp_efi_index].flags &= ~MLX5_ESW_DEST_ENCAP_VALID;
+
+ if (IS_ERR(rule)) {
+ err = PTR_ERR(rule);
+ mlx5_core_warn(priv->mdev, "Failed to update slow path (encap) flow, %d\n",
+ err);
+ continue;
+ }
+
+ mlx5e_tc_unoffload_fdb_rules(esw, flow, attr);
+ flow->rule[0] = rule;
+ /* was unset when fast path rule removed */
+ flow_flag_set(flow, OFFLOADED);
+ }
+
+ /* we know that the encap is valid */
+ e->flags &= ~MLX5_ENCAP_ENTRY_VALID;
+ mlx5_packet_reformat_dealloc(priv->mdev, e->pkt_reformat);
+}
+
+/* Takes reference to all flows attached to encap and adds the flows to
+ * flow_list using 'tmp_list' list_head in mlx5e_tc_flow.
+ */
+void mlx5e_take_all_encap_flows(struct mlx5e_encap_entry *e, struct list_head *flow_list)
+{
+ struct encap_flow_item *efi;
+ struct mlx5e_tc_flow *flow;
+
+ list_for_each_entry(efi, &e->flows, list) {
+ flow = container_of(efi, struct mlx5e_tc_flow, encaps[efi->index]);
+ if (IS_ERR(mlx5e_flow_get(flow)))
+ continue;
+ wait_for_completion(&flow->init_done);
+
+ flow->tmp_efi_index = efi->index;
+ list_add(&flow->tmp_list, flow_list);
+ }
+}
+
+static struct mlx5e_encap_entry *
+mlx5e_get_next_valid_encap(struct mlx5e_neigh_hash_entry *nhe,
+ struct mlx5e_encap_entry *e)
+{
+ struct mlx5e_encap_entry *next = NULL;
+
+retry:
+ rcu_read_lock();
+
+ /* find encap with non-zero reference counter value */
+ for (next = e ?
+ list_next_or_null_rcu(&nhe->encap_list,
+ &e->encap_list,
+ struct mlx5e_encap_entry,
+ encap_list) :
+ list_first_or_null_rcu(&nhe->encap_list,
+ struct mlx5e_encap_entry,
+ encap_list);
+ next;
+ next = list_next_or_null_rcu(&nhe->encap_list,
+ &next->encap_list,
+ struct mlx5e_encap_entry,
+ encap_list))
+ if (mlx5e_encap_take(next))
+ break;
+
+ rcu_read_unlock();
+
+ /* release starting encap */
+ if (e)
+ mlx5e_encap_put(netdev_priv(e->out_dev), e);
+ if (!next)
+ return next;
+
+ /* wait for encap to be fully initialized */
+ wait_for_completion(&next->res_ready);
+ /* continue searching if encap entry is not in valid state after completion */
+ if (!(next->flags & MLX5_ENCAP_ENTRY_VALID)) {
+ e = next;
+ goto retry;
+ }
+
+ return next;
+}
+
+void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe)
+{
+ struct mlx5e_neigh *m_neigh = &nhe->m_neigh;
+ struct mlx5e_encap_entry *e = NULL;
+ struct mlx5e_tc_flow *flow;
+ struct mlx5_fc *counter;
+ struct neigh_table *tbl;
+ bool neigh_used = false;
+ struct neighbour *n;
+ u64 lastuse;
+
+ if (m_neigh->family == AF_INET)
+ tbl = &arp_tbl;
+#if IS_ENABLED(CONFIG_IPV6)
+ else if (m_neigh->family == AF_INET6)
+ tbl = ipv6_stub->nd_tbl;
+#endif
+ else
+ return;
+
+ /* mlx5e_get_next_valid_encap() releases previous encap before returning
+ * next one.
+ */
+ while ((e = mlx5e_get_next_valid_encap(nhe, e)) != NULL) {
+ struct mlx5e_priv *priv = netdev_priv(e->out_dev);
+ struct encap_flow_item *efi, *tmp;
+ struct mlx5_eswitch *esw;
+ LIST_HEAD(flow_list);
+
+ esw = priv->mdev->priv.eswitch;
+ mutex_lock(&esw->offloads.encap_tbl_lock);
+ list_for_each_entry_safe(efi, tmp, &e->flows, list) {
+ flow = container_of(efi, struct mlx5e_tc_flow,
+ encaps[efi->index]);
+ if (IS_ERR(mlx5e_flow_get(flow)))
+ continue;
+ list_add(&flow->tmp_list, &flow_list);
+
+ if (mlx5e_is_offloaded_flow(flow)) {
+ counter = mlx5e_tc_get_counter(flow);
+ lastuse = mlx5_fc_query_lastuse(counter);
+ if (time_after((unsigned long)lastuse, nhe->reported_lastuse)) {
+ neigh_used = true;
+ break;
+ }
+ }
+ }
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+
+ mlx5e_put_encap_flow_list(priv, &flow_list);
+ if (neigh_used) {
+ /* release current encap before breaking the loop */
+ mlx5e_encap_put(priv, e);
+ break;
+ }
+ }
+
+ trace_mlx5e_tc_update_neigh_used_value(nhe, neigh_used);
+
+ if (neigh_used) {
+ nhe->reported_lastuse = jiffies;
+
+ /* find the relevant neigh according to the cached device and
+ * dst ip pair
+ */
+ n = neigh_lookup(tbl, &m_neigh->dst_ip, m_neigh->dev);
+ if (!n)
+ return;
+
+ neigh_event_send(n, NULL);
+ neigh_release(n);
+ }
+}
+
+static void mlx5e_encap_dealloc(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e)
+{
+ WARN_ON(!list_empty(&e->flows));
+
+ if (e->compl_result > 0) {
+ mlx5e_rep_encap_entry_detach(netdev_priv(e->out_dev), e);
+
+ if (e->flags & MLX5_ENCAP_ENTRY_VALID)
+ mlx5_packet_reformat_dealloc(priv->mdev, e->pkt_reformat);
+ }
+
+ kfree(e->tun_info);
+ kfree(e->encap_header);
+ kfree_rcu(e, rcu);
+}
+
+static void mlx5e_decap_dealloc(struct mlx5e_priv *priv,
+ struct mlx5e_decap_entry *d)
+{
+ WARN_ON(!list_empty(&d->flows));
+
+ if (!d->compl_result)
+ mlx5_packet_reformat_dealloc(priv->mdev, d->pkt_reformat);
+
+ kfree_rcu(d, rcu);
+}
+
+void mlx5e_encap_put(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+
+ if (!refcount_dec_and_mutex_lock(&e->refcnt, &esw->offloads.encap_tbl_lock))
+ return;
+ hash_del_rcu(&e->encap_hlist);
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+
+ mlx5e_encap_dealloc(priv, e);
+}
+
+static void mlx5e_decap_put(struct mlx5e_priv *priv, struct mlx5e_decap_entry *d)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+
+ if (!refcount_dec_and_mutex_lock(&d->refcnt, &esw->offloads.decap_tbl_lock))
+ return;
+ hash_del_rcu(&d->hlist);
+ mutex_unlock(&esw->offloads.decap_tbl_lock);
+
+ mlx5e_decap_dealloc(priv, d);
+}
+
+void mlx5e_detach_encap(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow, int out_index)
+{
+ struct mlx5e_encap_entry *e = flow->encaps[out_index].e;
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+
+ /* flow wasn't fully initialized */
+ if (!e)
+ return;
+
+ mutex_lock(&esw->offloads.encap_tbl_lock);
+ list_del(&flow->encaps[out_index].list);
+ flow->encaps[out_index].e = NULL;
+ if (!refcount_dec_and_test(&e->refcnt)) {
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+ return;
+ }
+ hash_del_rcu(&e->encap_hlist);
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+
+ mlx5e_encap_dealloc(priv, e);
+}
+
+void mlx5e_detach_decap(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_decap_entry *d = flow->decap_reformat;
+
+ if (!d)
+ return;
+
+ mutex_lock(&esw->offloads.decap_tbl_lock);
+ list_del(&flow->l3_to_l2_reformat);
+ flow->decap_reformat = NULL;
+
+ if (!refcount_dec_and_test(&d->refcnt)) {
+ mutex_unlock(&esw->offloads.decap_tbl_lock);
+ return;
+ }
+ hash_del_rcu(&d->hlist);
+ mutex_unlock(&esw->offloads.decap_tbl_lock);
+
+ mlx5e_decap_dealloc(priv, d);
+}
+
+struct encap_key {
+ const struct ip_tunnel_key *ip_tun_key;
+ struct mlx5e_tc_tunnel *tc_tunnel;
+};
+
+static int cmp_encap_info(struct encap_key *a,
+ struct encap_key *b)
+{
+ return memcmp(a->ip_tun_key, b->ip_tun_key, sizeof(*a->ip_tun_key)) ||
+ a->tc_tunnel->tunnel_type != b->tc_tunnel->tunnel_type;
+}
+
+static int cmp_decap_info(struct mlx5e_decap_key *a,
+ struct mlx5e_decap_key *b)
+{
+ return memcmp(&a->key, &b->key, sizeof(b->key));
+}
+
+static int hash_encap_info(struct encap_key *key)
+{
+ return jhash(key->ip_tun_key, sizeof(*key->ip_tun_key),
+ key->tc_tunnel->tunnel_type);
+}
+
+static int hash_decap_info(struct mlx5e_decap_key *key)
+{
+ return jhash(&key->key, sizeof(key->key), 0);
+}
+
+bool mlx5e_encap_take(struct mlx5e_encap_entry *e)
+{
+ return refcount_inc_not_zero(&e->refcnt);
+}
+
+static bool mlx5e_decap_take(struct mlx5e_decap_entry *e)
+{
+ return refcount_inc_not_zero(&e->refcnt);
+}
+
+static struct mlx5e_encap_entry *
+mlx5e_encap_get(struct mlx5e_priv *priv, struct encap_key *key,
+ uintptr_t hash_key)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_encap_entry *e;
+ struct encap_key e_key;
+
+ hash_for_each_possible_rcu(esw->offloads.encap_tbl, e,
+ encap_hlist, hash_key) {
+ e_key.ip_tun_key = &e->tun_info->key;
+ e_key.tc_tunnel = e->tunnel;
+ if (!cmp_encap_info(&e_key, key) &&
+ mlx5e_encap_take(e))
+ return e;
+ }
+
+ return NULL;
+}
+
+static struct mlx5e_decap_entry *
+mlx5e_decap_get(struct mlx5e_priv *priv, struct mlx5e_decap_key *key,
+ uintptr_t hash_key)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_decap_key r_key;
+ struct mlx5e_decap_entry *e;
+
+ hash_for_each_possible_rcu(esw->offloads.decap_tbl, e,
+ hlist, hash_key) {
+ r_key = e->key;
+ if (!cmp_decap_info(&r_key, key) &&
+ mlx5e_decap_take(e))
+ return e;
+ }
+ return NULL;
+}
+
+struct ip_tunnel_info *mlx5e_dup_tun_info(const struct ip_tunnel_info *tun_info)
+{
+ size_t tun_size = sizeof(*tun_info) + tun_info->options_len;
+
+ return kmemdup(tun_info, tun_size, GFP_KERNEL);
+}
+
+static bool is_duplicated_encap_entry(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ int out_index,
+ struct mlx5e_encap_entry *e,
+ struct netlink_ext_ack *extack)
+{
+ int i;
+
+ for (i = 0; i < out_index; i++) {
+ if (flow->encaps[i].e != e)
+ continue;
+ NL_SET_ERR_MSG_MOD(extack, "can't duplicate encap action");
+ netdev_err(priv->netdev, "can't duplicate encap action\n");
+ return true;
+ }
+
+ return false;
+}
+
+static int mlx5e_set_vf_tunnel(struct mlx5_eswitch *esw,
+ struct mlx5_flow_attr *attr,
+ struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
+ struct net_device *out_dev,
+ int route_dev_ifindex,
+ int out_index)
+{
+ struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
+ struct net_device *route_dev;
+ u16 vport_num;
+ int err = 0;
+ u32 data;
+
+ route_dev = dev_get_by_index(dev_net(out_dev), route_dev_ifindex);
+
+ if (!route_dev || route_dev->netdev_ops != &mlx5e_netdev_ops ||
+ !mlx5e_tc_is_vf_tunnel(out_dev, route_dev))
+ goto out;
+
+ err = mlx5e_tc_query_route_vport(out_dev, route_dev, &vport_num);
+ if (err)
+ goto out;
+
+ attr->dest_chain = 0;
+ attr->action |= MLX5_FLOW_CONTEXT_ACTION_MOD_HDR;
+ esw_attr->dests[out_index].flags |= MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE;
+ data = mlx5_eswitch_get_vport_metadata_for_set(esw_attr->in_mdev->priv.eswitch,
+ vport_num);
+ err = mlx5e_tc_match_to_reg_set(esw->dev, mod_hdr_acts,
+ MLX5_FLOW_NAMESPACE_FDB, VPORT_TO_REG, data);
+ if (err)
+ goto out;
+
+out:
+ if (route_dev)
+ dev_put(route_dev);
+ return err;
+}
+
+int mlx5e_attach_encap(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ struct net_device *mirred_dev,
+ int out_index,
+ struct netlink_ext_ack *extack,
+ struct net_device **encap_dev,
+ bool *encap_valid)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_tc_flow_parse_attr *parse_attr;
+ struct mlx5_flow_attr *attr = flow->attr;
+ const struct ip_tunnel_info *tun_info;
+ struct encap_key key;
+ struct mlx5e_encap_entry *e;
+ unsigned short family;
+ uintptr_t hash_key;
+ int err = 0;
+
+ parse_attr = attr->parse_attr;
+ tun_info = parse_attr->tun_info[out_index];
+ family = ip_tunnel_info_af(tun_info);
+ key.ip_tun_key = &tun_info->key;
+ key.tc_tunnel = mlx5e_get_tc_tun(mirred_dev);
+ if (!key.tc_tunnel) {
+ NL_SET_ERR_MSG_MOD(extack, "Unsupported tunnel");
+ return -EOPNOTSUPP;
+ }
+
+ hash_key = hash_encap_info(&key);
+
+ mutex_lock(&esw->offloads.encap_tbl_lock);
+ e = mlx5e_encap_get(priv, &key, hash_key);
+
+ /* must verify if encap is valid or not */
+ if (e) {
+ /* Check that entry was not already attached to this flow */
+ if (is_duplicated_encap_entry(priv, flow, out_index, e, extack)) {
+ err = -EOPNOTSUPP;
+ goto out_err;
+ }
+
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+ wait_for_completion(&e->res_ready);
+
+ /* Protect against concurrent neigh update. */
+ mutex_lock(&esw->offloads.encap_tbl_lock);
+ if (e->compl_result < 0) {
+ err = -EREMOTEIO;
+ goto out_err;
+ }
+ goto attach_flow;
+ }
+
+ e = kzalloc(sizeof(*e), GFP_KERNEL);
+ if (!e) {
+ err = -ENOMEM;
+ goto out_err;
+ }
+
+ refcount_set(&e->refcnt, 1);
+ init_completion(&e->res_ready);
+
+ tun_info = mlx5e_dup_tun_info(tun_info);
+ if (!tun_info) {
+ err = -ENOMEM;
+ goto out_err_init;
+ }
+ e->tun_info = tun_info;
+ err = mlx5e_tc_tun_init_encap_attr(mirred_dev, priv, e, extack);
+ if (err)
+ goto out_err_init;
+
+ INIT_LIST_HEAD(&e->flows);
+ hash_add_rcu(esw->offloads.encap_tbl, &e->encap_hlist, hash_key);
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+
+ if (family == AF_INET)
+ err = mlx5e_tc_tun_create_header_ipv4(priv, mirred_dev, e);
+ else if (family == AF_INET6)
+ err = mlx5e_tc_tun_create_header_ipv6(priv, mirred_dev, e);
+
+ /* Protect against concurrent neigh update. */
+ mutex_lock(&esw->offloads.encap_tbl_lock);
+ complete_all(&e->res_ready);
+ if (err) {
+ e->compl_result = err;
+ goto out_err;
+ }
+ e->compl_result = 1;
+
+attach_flow:
+ err = mlx5e_set_vf_tunnel(esw, attr, &parse_attr->mod_hdr_acts, e->out_dev,
+ e->route_dev_ifindex, out_index);
+ if (err)
+ goto out_err;
+
+ flow->encaps[out_index].e = e;
+ list_add(&flow->encaps[out_index].list, &e->flows);
+ flow->encaps[out_index].index = out_index;
+ *encap_dev = e->out_dev;
+ if (e->flags & MLX5_ENCAP_ENTRY_VALID) {
+ attr->esw_attr->dests[out_index].pkt_reformat = e->pkt_reformat;
+ attr->esw_attr->dests[out_index].flags |= MLX5_ESW_DEST_ENCAP_VALID;
+ *encap_valid = true;
+ } else {
+ *encap_valid = false;
+ }
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+
+ return err;
+
+out_err:
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+ if (e)
+ mlx5e_encap_put(priv, e);
+ return err;
+
+out_err_init:
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+ kfree(tun_info);
+ kfree(e);
+ return err;
+}
+
+int mlx5e_attach_decap(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ struct netlink_ext_ack *extack)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5_esw_flow_attr *attr = flow->attr->esw_attr;
+ struct mlx5e_tc_flow_parse_attr *parse_attr;
+ struct mlx5e_decap_entry *d;
+ struct mlx5e_decap_key key;
+ uintptr_t hash_key;
+ int err = 0;
+
+ parse_attr = flow->attr->parse_attr;
+ if (sizeof(parse_attr->eth) > MLX5_CAP_ESW(priv->mdev, max_encap_header_size)) {
+ NL_SET_ERR_MSG_MOD(extack,
+ "encap header larger than max supported");
+ return -EOPNOTSUPP;
+ }
+
+ key.key = parse_attr->eth;
+ hash_key = hash_decap_info(&key);
+ mutex_lock(&esw->offloads.decap_tbl_lock);
+ d = mlx5e_decap_get(priv, &key, hash_key);
+ if (d) {
+ mutex_unlock(&esw->offloads.decap_tbl_lock);
+ wait_for_completion(&d->res_ready);
+ mutex_lock(&esw->offloads.decap_tbl_lock);
+ if (d->compl_result) {
+ err = -EREMOTEIO;
+ goto out_free;
+ }
+ goto found;
+ }
+
+ d = kzalloc(sizeof(*d), GFP_KERNEL);
+ if (!d) {
+ err = -ENOMEM;
+ goto out_err;
+ }
+
+ d->key = key;
+ refcount_set(&d->refcnt, 1);
+ init_completion(&d->res_ready);
+ INIT_LIST_HEAD(&d->flows);
+ hash_add_rcu(esw->offloads.decap_tbl, &d->hlist, hash_key);
+ mutex_unlock(&esw->offloads.decap_tbl_lock);
+
+ d->pkt_reformat = mlx5_packet_reformat_alloc(priv->mdev,
+ MLX5_REFORMAT_TYPE_L3_TUNNEL_TO_L2,
+ sizeof(parse_attr->eth),
+ &parse_attr->eth,
+ MLX5_FLOW_NAMESPACE_FDB);
+ if (IS_ERR(d->pkt_reformat)) {
+ err = PTR_ERR(d->pkt_reformat);
+ d->compl_result = err;
+ }
+ mutex_lock(&esw->offloads.decap_tbl_lock);
+ complete_all(&d->res_ready);
+ if (err)
+ goto out_free;
+
+found:
+ flow->decap_reformat = d;
+ attr->decap_pkt_reformat = d->pkt_reformat;
+ list_add(&flow->l3_to_l2_reformat, &d->flows);
+ mutex_unlock(&esw->offloads.decap_tbl_lock);
+ return 0;
+
+out_free:
+ mutex_unlock(&esw->offloads.decap_tbl_lock);
+ mlx5e_decap_put(priv, d);
+ return err;
+
+out_err:
+ mutex_unlock(&esw->offloads.decap_tbl_lock);
+ return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h
new file mode 100644
index 000000000000..81b9fef1cf2a
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2021 Mellanox Technologies. */
+
+#ifndef __MLX5_EN_TC_TUN_ENCAP_H__
+#define __MLX5_EN_TC_TUN_ENCAP_H__
+
+#include "tc_priv.h"
+
+void mlx5e_detach_encap(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow, int out_index);
+
+int mlx5e_attach_encap(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ struct net_device *mirred_dev,
+ int out_index,
+ struct netlink_ext_ack *extack,
+ struct net_device **encap_dev,
+ bool *encap_valid);
+int mlx5e_attach_decap(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ struct netlink_ext_ack *extack);
+void mlx5e_detach_decap(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow);
+
+struct ip_tunnel_info *mlx5e_dup_tun_info(const struct ip_tunnel_info *tun_info);
+
+int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
+ struct mlx5_flow_spec *spec);
+
+#endif /* __MLX5_EN_TC_TUN_ENCAP_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 21568d1fc00f..cfe340e23dfc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -63,6 +63,8 @@
#include "en/mapping.h"
#include "en/tc_ct.h"
#include "en/mod_hdr.h"
+#include "en/tc_priv.h"
+#include "en/tc_tun_encap.h"
#include "lib/devcom.h"
#include "lib/geneve.h"
#include "lib/fs_chains.h"
@@ -71,90 +73,6 @@
#define nic_chains(priv) ((priv)->fs.tc.chains)
#define MLX5_MH_ACT_SZ MLX5_UN_SZ_BYTES(set_add_copy_action_in_auto)
-#define MLX5E_TC_FLOW_BASE (MLX5E_TC_FLAG_LAST_EXPORTED_BIT + 1)
-
-enum {
- MLX5E_TC_FLOW_FLAG_INGRESS = MLX5E_TC_FLAG_INGRESS_BIT,
- MLX5E_TC_FLOW_FLAG_EGRESS = MLX5E_TC_FLAG_EGRESS_BIT,
- MLX5E_TC_FLOW_FLAG_ESWITCH = MLX5E_TC_FLAG_ESW_OFFLOAD_BIT,
- MLX5E_TC_FLOW_FLAG_FT = MLX5E_TC_FLAG_FT_OFFLOAD_BIT,
- MLX5E_TC_FLOW_FLAG_NIC = MLX5E_TC_FLAG_NIC_OFFLOAD_BIT,
- MLX5E_TC_FLOW_FLAG_OFFLOADED = MLX5E_TC_FLOW_BASE,
- MLX5E_TC_FLOW_FLAG_HAIRPIN = MLX5E_TC_FLOW_BASE + 1,
- MLX5E_TC_FLOW_FLAG_HAIRPIN_RSS = MLX5E_TC_FLOW_BASE + 2,
- MLX5E_TC_FLOW_FLAG_SLOW = MLX5E_TC_FLOW_BASE + 3,
- MLX5E_TC_FLOW_FLAG_DUP = MLX5E_TC_FLOW_BASE + 4,
- MLX5E_TC_FLOW_FLAG_NOT_READY = MLX5E_TC_FLOW_BASE + 5,
- MLX5E_TC_FLOW_FLAG_DELETED = MLX5E_TC_FLOW_BASE + 6,
- MLX5E_TC_FLOW_FLAG_CT = MLX5E_TC_FLOW_BASE + 7,
- MLX5E_TC_FLOW_FLAG_L3_TO_L2_DECAP = MLX5E_TC_FLOW_BASE + 8,
-};
-
-#define MLX5E_TC_MAX_SPLITS 1
-
-/* Helper struct for accessing a struct containing list_head array.
- * Containing struct
- * |- Helper array
- * [0] Helper item 0
- * |- list_head item 0
- * |- index (0)
- * [1] Helper item 1
- * |- list_head item 1
- * |- index (1)
- * To access the containing struct from one of the list_head items:
- * 1. Get the helper item from the list_head item using
- * helper item =
- * container_of(list_head item, helper struct type, list_head field)
- * 2. Get the contining struct from the helper item and its index in the array:
- * containing struct =
- * container_of(helper item, containing struct type, helper field[index])
- */
-struct encap_flow_item {
- struct mlx5e_encap_entry *e; /* attached encap instance */
- struct list_head list;
- int index;
-};
-
-struct mlx5e_tc_flow {
- struct rhash_head node;
- struct mlx5e_priv *priv;
- u64 cookie;
- unsigned long flags;
- struct mlx5_flow_handle *rule[MLX5E_TC_MAX_SPLITS + 1];
-
- /* flows sharing the same reformat object - currently mpls decap */
- struct list_head l3_to_l2_reformat;
- struct mlx5e_decap_entry *decap_reformat;
-
- /* Flow can be associated with multiple encap IDs.
- * The number of encaps is bounded by the number of supported
- * destinations.
- */
- struct encap_flow_item encaps[MLX5_MAX_FLOW_FWD_VPORTS];
- struct mlx5e_tc_flow *peer_flow;
- struct mlx5e_mod_hdr_handle *mh; /* attached mod header instance */
- struct mlx5e_hairpin_entry *hpe; /* attached hairpin instance */
- struct list_head hairpin; /* flows sharing the same hairpin */
- struct list_head peer; /* flows with peer flow */
- struct list_head unready; /* flows not ready to be offloaded (e.g due to missing route) */
- struct net_device *orig_dev; /* netdev adding flow first */
- int tmp_efi_index;
- struct list_head tmp_list; /* temporary flow list used by neigh update */
- refcount_t refcnt;
- struct rcu_head rcu_head;
- struct completion init_done;
- int tunnel_id; /* the mapped tunnel id of this flow */
- struct mlx5_flow_attr *attr;
-};
-
-struct mlx5e_tc_flow_parse_attr {
- const struct ip_tunnel_info *tun_info[MLX5_MAX_FLOW_FWD_VPORTS];
- struct net_device *filter_dev;
- struct mlx5_flow_spec spec;
- struct mlx5e_tc_mod_hdr_acts mod_hdr_acts;
- int mirred_ifindex[MLX5_MAX_FLOW_FWD_VPORTS];
- struct ethhdr eth;
-};
#define MLX5E_TC_TABLE_NUM_GROUPS 4
#define MLX5E_TC_TABLE_MAX_GROUP_SIZE BIT(18)
@@ -368,15 +286,14 @@ struct mlx5e_hairpin_entry {
static void mlx5e_tc_del_flow(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow);
-static struct mlx5e_tc_flow *mlx5e_flow_get(struct mlx5e_tc_flow *flow)
+struct mlx5e_tc_flow *mlx5e_flow_get(struct mlx5e_tc_flow *flow)
{
if (!flow || !refcount_inc_not_zero(&flow->refcnt))
return ERR_PTR(-EINVAL);
return flow;
}
-static void mlx5e_flow_put(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow)
+void mlx5e_flow_put(struct mlx5e_priv *priv, struct mlx5e_tc_flow *flow)
{
if (refcount_dec_and_test(&flow->refcnt)) {
mlx5e_tc_del_flow(priv, flow);
@@ -384,48 +301,6 @@ static void mlx5e_flow_put(struct mlx5e_priv *priv,
}
}
-static void __flow_flag_set(struct mlx5e_tc_flow *flow, unsigned long flag)
-{
- /* Complete all memory stores before setting bit. */
- smp_mb__before_atomic();
- set_bit(flag, &flow->flags);
-}
-
-#define flow_flag_set(flow, flag) __flow_flag_set(flow, MLX5E_TC_FLOW_FLAG_##flag)
-
-static bool __flow_flag_test_and_set(struct mlx5e_tc_flow *flow,
- unsigned long flag)
-{
- /* test_and_set_bit() provides all necessary barriers */
- return test_and_set_bit(flag, &flow->flags);
-}
-
-#define flow_flag_test_and_set(flow, flag) \
- __flow_flag_test_and_set(flow, \
- MLX5E_TC_FLOW_FLAG_##flag)
-
-static void __flow_flag_clear(struct mlx5e_tc_flow *flow, unsigned long flag)
-{
- /* Complete all memory stores before clearing bit. */
- smp_mb__before_atomic();
- clear_bit(flag, &flow->flags);
-}
-
-#define flow_flag_clear(flow, flag) __flow_flag_clear(flow, \
- MLX5E_TC_FLOW_FLAG_##flag)
-
-static bool __flow_flag_test(struct mlx5e_tc_flow *flow, unsigned long flag)
-{
- bool ret = test_bit(flag, &flow->flags);
-
- /* Read fields of flow structure only after checking flags. */
- smp_mb__after_atomic();
- return ret;
-}
-
-#define flow_flag_test(flow, flag) __flow_flag_test(flow, \
- MLX5E_TC_FLOW_FLAG_##flag)
-
bool mlx5e_is_eswitch_flow(struct mlx5e_tc_flow *flow)
{
return flow_flag_test(flow, ESWITCH);
@@ -436,7 +311,7 @@ static bool mlx5e_is_ft_flow(struct mlx5e_tc_flow *flow)
return flow_flag_test(flow, FT);
}
-static bool mlx5e_is_offloaded_flow(struct mlx5e_tc_flow *flow)
+bool mlx5e_is_offloaded_flow(struct mlx5e_tc_flow *flow)
{
return flow_flag_test(flow, OFFLOADED);
}
@@ -1151,23 +1026,7 @@ static void mlx5e_tc_del_nic_flow(struct mlx5e_priv *priv,
kfree(flow->attr);
}
-static void mlx5e_detach_encap(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow, int out_index);
-
-static int mlx5e_attach_encap(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow,
- struct net_device *mirred_dev,
- int out_index,
- struct netlink_ext_ack *extack,
- struct net_device **encap_dev,
- bool *encap_valid);
-static int mlx5e_attach_decap(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow,
- struct netlink_ext_ack *extack);
-static void mlx5e_detach_decap(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow);
-
-static struct mlx5_flow_handle *
+struct mlx5_flow_handle *
mlx5e_tc_offload_fdb_rules(struct mlx5_eswitch *esw,
struct mlx5e_tc_flow *flow,
struct mlx5_flow_spec *spec,
@@ -1202,10 +1061,9 @@ mlx5e_tc_offload_fdb_rules(struct mlx5_eswitch *esw,
return rule;
}
-static void
-mlx5e_tc_unoffload_fdb_rules(struct mlx5_eswitch *esw,
- struct mlx5e_tc_flow *flow,
- struct mlx5_flow_attr *attr)
+void mlx5e_tc_unoffload_fdb_rules(struct mlx5_eswitch *esw,
+ struct mlx5e_tc_flow *flow,
+ struct mlx5_flow_attr *attr)
{
flow_flag_clear(flow, OFFLOADED);
@@ -1224,7 +1082,7 @@ mlx5e_tc_unoffload_fdb_rules(struct mlx5_eswitch *esw,
mlx5_eswitch_del_offloaded_rule(esw, flow->rule[0], attr);
}
-static struct mlx5_flow_handle *
+struct mlx5_flow_handle *
mlx5e_tc_offload_to_slow_path(struct mlx5_eswitch *esw,
struct mlx5e_tc_flow *flow,
struct mlx5_flow_spec *spec)
@@ -1250,9 +1108,8 @@ mlx5e_tc_offload_to_slow_path(struct mlx5_eswitch *esw,
return rule;
}
-static void
-mlx5e_tc_unoffload_from_slow_path(struct mlx5_eswitch *esw,
- struct mlx5e_tc_flow *flow)
+void mlx5e_tc_unoffload_from_slow_path(struct mlx5_eswitch *esw,
+ struct mlx5e_tc_flow *flow)
{
struct mlx5_flow_attr *slow_attr;
@@ -1519,139 +1376,11 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv,
kfree(flow->attr);
}
-void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv,
- struct mlx5e_encap_entry *e,
- struct list_head *flow_list)
-{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
- struct mlx5_esw_flow_attr *esw_attr;
- struct mlx5_flow_handle *rule;
- struct mlx5_flow_attr *attr;
- struct mlx5_flow_spec *spec;
- struct mlx5e_tc_flow *flow;
- int err;
-
- e->pkt_reformat = mlx5_packet_reformat_alloc(priv->mdev,
- e->reformat_type,
- e->encap_size, e->encap_header,
- MLX5_FLOW_NAMESPACE_FDB);
- if (IS_ERR(e->pkt_reformat)) {
- mlx5_core_warn(priv->mdev, "Failed to offload cached encapsulation header, %lu\n",
- PTR_ERR(e->pkt_reformat));
- return;
- }
- e->flags |= MLX5_ENCAP_ENTRY_VALID;
- mlx5e_rep_queue_neigh_stats_work(priv);
-
- list_for_each_entry(flow, flow_list, tmp_list) {
- bool all_flow_encaps_valid = true;
- int i;
-
- if (!mlx5e_is_offloaded_flow(flow))
- continue;
- attr = flow->attr;
- esw_attr = attr->esw_attr;
- spec = &attr->parse_attr->spec;
-
- esw_attr->dests[flow->tmp_efi_index].pkt_reformat = e->pkt_reformat;
- esw_attr->dests[flow->tmp_efi_index].flags |= MLX5_ESW_DEST_ENCAP_VALID;
- /* Flow can be associated with multiple encap entries.
- * Before offloading the flow verify that all of them have
- * a valid neighbour.
- */
- for (i = 0; i < MLX5_MAX_FLOW_FWD_VPORTS; i++) {
- if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP))
- continue;
- if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP_VALID)) {
- all_flow_encaps_valid = false;
- break;
- }
- }
- /* Do not offload flows with unresolved neighbors */
- if (!all_flow_encaps_valid)
- continue;
- /* update from slow path rule to encap rule */
- rule = mlx5e_tc_offload_fdb_rules(esw, flow, spec, attr);
- if (IS_ERR(rule)) {
- err = PTR_ERR(rule);
- mlx5_core_warn(priv->mdev, "Failed to update cached encapsulation flow, %d\n",
- err);
- continue;
- }
-
- mlx5e_tc_unoffload_from_slow_path(esw, flow);
- flow->rule[0] = rule;
- /* was unset when slow path rule removed */
- flow_flag_set(flow, OFFLOADED);
- }
-}
-
-void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv,
- struct mlx5e_encap_entry *e,
- struct list_head *flow_list)
-{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
- struct mlx5_esw_flow_attr *esw_attr;
- struct mlx5_flow_handle *rule;
- struct mlx5_flow_attr *attr;
- struct mlx5_flow_spec *spec;
- struct mlx5e_tc_flow *flow;
- int err;
-
- list_for_each_entry(flow, flow_list, tmp_list) {
- if (!mlx5e_is_offloaded_flow(flow))
- continue;
- attr = flow->attr;
- esw_attr = attr->esw_attr;
- spec = &attr->parse_attr->spec;
-
- /* update from encap rule to slow path rule */
- rule = mlx5e_tc_offload_to_slow_path(esw, flow, spec);
- /* mark the flow's encap dest as non-valid */
- esw_attr->dests[flow->tmp_efi_index].flags &= ~MLX5_ESW_DEST_ENCAP_VALID;
-
- if (IS_ERR(rule)) {
- err = PTR_ERR(rule);
- mlx5_core_warn(priv->mdev, "Failed to update slow path (encap) flow, %d\n",
- err);
- continue;
- }
-
- mlx5e_tc_unoffload_fdb_rules(esw, flow, attr);
- flow->rule[0] = rule;
- /* was unset when fast path rule removed */
- flow_flag_set(flow, OFFLOADED);
- }
-
- /* we know that the encap is valid */
- e->flags &= ~MLX5_ENCAP_ENTRY_VALID;
- mlx5_packet_reformat_dealloc(priv->mdev, e->pkt_reformat);
-}
-
-static struct mlx5_fc *mlx5e_tc_get_counter(struct mlx5e_tc_flow *flow)
+struct mlx5_fc *mlx5e_tc_get_counter(struct mlx5e_tc_flow *flow)
{
return flow->attr->counter;
}
-/* Takes reference to all flows attached to encap and adds the flows to
- * flow_list using 'tmp_list' list_head in mlx5e_tc_flow.
- */
-void mlx5e_take_all_encap_flows(struct mlx5e_encap_entry *e, struct list_head *flow_list)
-{
- struct encap_flow_item *efi;
- struct mlx5e_tc_flow *flow;
-
- list_for_each_entry(efi, &e->flows, list) {
- flow = container_of(efi, struct mlx5e_tc_flow, encaps[efi->index]);
- if (IS_ERR(mlx5e_flow_get(flow)))
- continue;
- wait_for_completion(&flow->init_done);
-
- flow->tmp_efi_index = efi->index;
- list_add(&flow->tmp_list, flow_list);
- }
-}
-
/* Iterate over tmp_list of flows attached to flow_list head. */
void mlx5e_put_encap_flow_list(struct mlx5e_priv *priv, struct list_head *flow_list)
{
@@ -1661,222 +1390,6 @@ void mlx5e_put_encap_flow_list(struct mlx5e_priv *priv, struct list_head *flow_l
mlx5e_flow_put(priv, flow);
}
-static struct mlx5e_encap_entry *
-mlx5e_get_next_valid_encap(struct mlx5e_neigh_hash_entry *nhe,
- struct mlx5e_encap_entry *e)
-{
- struct mlx5e_encap_entry *next = NULL;
-
-retry:
- rcu_read_lock();
-
- /* find encap with non-zero reference counter value */
- for (next = e ?
- list_next_or_null_rcu(&nhe->encap_list,
- &e->encap_list,
- struct mlx5e_encap_entry,
- encap_list) :
- list_first_or_null_rcu(&nhe->encap_list,
- struct mlx5e_encap_entry,
- encap_list);
- next;
- next = list_next_or_null_rcu(&nhe->encap_list,
- &next->encap_list,
- struct mlx5e_encap_entry,
- encap_list))
- if (mlx5e_encap_take(next))
- break;
-
- rcu_read_unlock();
-
- /* release starting encap */
- if (e)
- mlx5e_encap_put(netdev_priv(e->out_dev), e);
- if (!next)
- return next;
-
- /* wait for encap to be fully initialized */
- wait_for_completion(&next->res_ready);
- /* continue searching if encap entry is not in valid state after completion */
- if (!(next->flags & MLX5_ENCAP_ENTRY_VALID)) {
- e = next;
- goto retry;
- }
-
- return next;
-}
-
-void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe)
-{
- struct mlx5e_neigh *m_neigh = &nhe->m_neigh;
- struct mlx5e_encap_entry *e = NULL;
- struct mlx5e_tc_flow *flow;
- struct mlx5_fc *counter;
- struct neigh_table *tbl;
- bool neigh_used = false;
- struct neighbour *n;
- u64 lastuse;
-
- if (m_neigh->family == AF_INET)
- tbl = &arp_tbl;
-#if IS_ENABLED(CONFIG_IPV6)
- else if (m_neigh->family == AF_INET6)
- tbl = ipv6_stub->nd_tbl;
-#endif
- else
- return;
-
- /* mlx5e_get_next_valid_encap() releases previous encap before returning
- * next one.
- */
- while ((e = mlx5e_get_next_valid_encap(nhe, e)) != NULL) {
- struct mlx5e_priv *priv = netdev_priv(e->out_dev);
- struct encap_flow_item *efi, *tmp;
- struct mlx5_eswitch *esw;
- LIST_HEAD(flow_list);
-
- esw = priv->mdev->priv.eswitch;
- mutex_lock(&esw->offloads.encap_tbl_lock);
- list_for_each_entry_safe(efi, tmp, &e->flows, list) {
- flow = container_of(efi, struct mlx5e_tc_flow,
- encaps[efi->index]);
- if (IS_ERR(mlx5e_flow_get(flow)))
- continue;
- list_add(&flow->tmp_list, &flow_list);
-
- if (mlx5e_is_offloaded_flow(flow)) {
- counter = mlx5e_tc_get_counter(flow);
- lastuse = mlx5_fc_query_lastuse(counter);
- if (time_after((unsigned long)lastuse, nhe->reported_lastuse)) {
- neigh_used = true;
- break;
- }
- }
- }
- mutex_unlock(&esw->offloads.encap_tbl_lock);
-
- mlx5e_put_encap_flow_list(priv, &flow_list);
- if (neigh_used) {
- /* release current encap before breaking the loop */
- mlx5e_encap_put(priv, e);
- break;
- }
- }
-
- trace_mlx5e_tc_update_neigh_used_value(nhe, neigh_used);
-
- if (neigh_used) {
- nhe->reported_lastuse = jiffies;
-
- /* find the relevant neigh according to the cached device and
- * dst ip pair
- */
- n = neigh_lookup(tbl, &m_neigh->dst_ip, m_neigh->dev);
- if (!n)
- return;
-
- neigh_event_send(n, NULL);
- neigh_release(n);
- }
-}
-
-static void mlx5e_encap_dealloc(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e)
-{
- WARN_ON(!list_empty(&e->flows));
-
- if (e->compl_result > 0) {
- mlx5e_rep_encap_entry_detach(netdev_priv(e->out_dev), e);
-
- if (e->flags & MLX5_ENCAP_ENTRY_VALID)
- mlx5_packet_reformat_dealloc(priv->mdev, e->pkt_reformat);
- }
-
- kfree(e->tun_info);
- kfree(e->encap_header);
- kfree_rcu(e, rcu);
-}
-
-static void mlx5e_decap_dealloc(struct mlx5e_priv *priv,
- struct mlx5e_decap_entry *d)
-{
- WARN_ON(!list_empty(&d->flows));
-
- if (!d->compl_result)
- mlx5_packet_reformat_dealloc(priv->mdev, d->pkt_reformat);
-
- kfree_rcu(d, rcu);
-}
-
-void mlx5e_encap_put(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e)
-{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
-
- if (!refcount_dec_and_mutex_lock(&e->refcnt, &esw->offloads.encap_tbl_lock))
- return;
- hash_del_rcu(&e->encap_hlist);
- mutex_unlock(&esw->offloads.encap_tbl_lock);
-
- mlx5e_encap_dealloc(priv, e);
-}
-
-static void mlx5e_decap_put(struct mlx5e_priv *priv, struct mlx5e_decap_entry *d)
-{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
-
- if (!refcount_dec_and_mutex_lock(&d->refcnt, &esw->offloads.decap_tbl_lock))
- return;
- hash_del_rcu(&d->hlist);
- mutex_unlock(&esw->offloads.decap_tbl_lock);
-
- mlx5e_decap_dealloc(priv, d);
-}
-
-static void mlx5e_detach_encap(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow, int out_index)
-{
- struct mlx5e_encap_entry *e = flow->encaps[out_index].e;
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
-
- /* flow wasn't fully initialized */
- if (!e)
- return;
-
- mutex_lock(&esw->offloads.encap_tbl_lock);
- list_del(&flow->encaps[out_index].list);
- flow->encaps[out_index].e = NULL;
- if (!refcount_dec_and_test(&e->refcnt)) {
- mutex_unlock(&esw->offloads.encap_tbl_lock);
- return;
- }
- hash_del_rcu(&e->encap_hlist);
- mutex_unlock(&esw->offloads.encap_tbl_lock);
-
- mlx5e_encap_dealloc(priv, e);
-}
-
-static void mlx5e_detach_decap(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow)
-{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
- struct mlx5e_decap_entry *d = flow->decap_reformat;
-
- if (!d)
- return;
-
- mutex_lock(&esw->offloads.decap_tbl_lock);
- list_del(&flow->l3_to_l2_reformat);
- flow->decap_reformat = NULL;
-
- if (!refcount_dec_and_test(&d->refcnt)) {
- mutex_unlock(&esw->offloads.decap_tbl_lock);
- return;
- }
- hash_del_rcu(&d->hlist);
- mutex_unlock(&esw->offloads.decap_tbl_lock);
-
- mlx5e_decap_dealloc(priv, d);
-}
-
static void __mlx5e_tc_del_fdb_peer_flow(struct mlx5e_tc_flow *flow)
{
struct mlx5_eswitch *esw = flow->priv->mdev->priv.eswitch;
@@ -2134,7 +1647,7 @@ void mlx5e_tc_set_ethertype(struct mlx5_core_dev *mdev,
}
}
-static u8 mlx5e_tc_get_ip_version(struct mlx5_flow_spec *spec, bool outer)
+u8 mlx5e_tc_get_ip_version(struct mlx5_flow_spec *spec, bool outer)
{
void *headers_v;
u16 ethertype;
@@ -2157,44 +1670,6 @@ static u8 mlx5e_tc_get_ip_version(struct mlx5_flow_spec *spec, bool outer)
return ip_version;
}
-static int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
- struct mlx5_flow_spec *spec)
-{
- struct mlx5_esw_flow_attr *esw_attr = flow->attr->esw_attr;
- struct mlx5_rx_tun_attr *tun_attr;
- void *daddr, *saddr;
- u8 ip_version;
-
- tun_attr = kvzalloc(sizeof(*tun_attr), GFP_KERNEL);
- if (!tun_attr)
- return -ENOMEM;
-
- esw_attr->rx_tun_attr = tun_attr;
- ip_version = mlx5e_tc_get_ip_version(spec, true);
-
- if (ip_version == 4) {
- daddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
- outer_headers.dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
- saddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
- outer_headers.src_ipv4_src_ipv6.ipv4_layout.ipv4);
- tun_attr->dst_ip.v4 = *(__be32 *)daddr;
- tun_attr->src_ip.v4 = *(__be32 *)saddr;
- }
-#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
- else if (ip_version == 6) {
- int ipv6_size = MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6);
-
- daddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
- outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6);
- saddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
- outer_headers.src_ipv4_src_ipv6.ipv6_layout.ipv6);
- memcpy(&tun_attr->dst_ip.v6, daddr, ipv6_size);
- memcpy(&tun_attr->src_ip.v6, saddr, ipv6_size);
- }
-#endif
- return 0;
-}
-
static int parse_tunnel_attr(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow,
struct mlx5_flow_spec *spec,
@@ -3714,35 +3189,6 @@ static int parse_tc_nic_actions(struct mlx5e_priv *priv,
return 0;
}
-struct encap_key {
- const struct ip_tunnel_key *ip_tun_key;
- struct mlx5e_tc_tunnel *tc_tunnel;
-};
-
-static inline int cmp_encap_info(struct encap_key *a,
- struct encap_key *b)
-{
- return memcmp(a->ip_tun_key, b->ip_tun_key, sizeof(*a->ip_tun_key)) ||
- a->tc_tunnel->tunnel_type != b->tc_tunnel->tunnel_type;
-}
-
-static inline int cmp_decap_info(struct mlx5e_decap_key *a,
- struct mlx5e_decap_key *b)
-{
- return memcmp(&a->key, &b->key, sizeof(b->key));
-}
-
-static inline int hash_encap_info(struct encap_key *key)
-{
- return jhash(key->ip_tun_key, sizeof(*key->ip_tun_key),
- key->tc_tunnel->tunnel_type);
-}
-
-static inline int hash_decap_info(struct mlx5e_decap_key *key)
-{
- return jhash(&key->key, sizeof(key->key), 0);
-}
-
static bool is_merged_eswitch_vfs(struct mlx5e_priv *priv,
struct net_device *peer_netdev)
{
@@ -3756,321 +3202,6 @@ static bool is_merged_eswitch_vfs(struct mlx5e_priv *priv,
same_hw_devs(priv, peer_priv));
}
-bool mlx5e_encap_take(struct mlx5e_encap_entry *e)
-{
- return refcount_inc_not_zero(&e->refcnt);
-}
-
-static bool mlx5e_decap_take(struct mlx5e_decap_entry *e)
-{
- return refcount_inc_not_zero(&e->refcnt);
-}
-
-static struct mlx5e_encap_entry *
-mlx5e_encap_get(struct mlx5e_priv *priv, struct encap_key *key,
- uintptr_t hash_key)
-{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
- struct mlx5e_encap_entry *e;
- struct encap_key e_key;
-
- hash_for_each_possible_rcu(esw->offloads.encap_tbl, e,
- encap_hlist, hash_key) {
- e_key.ip_tun_key = &e->tun_info->key;
- e_key.tc_tunnel = e->tunnel;
- if (!cmp_encap_info(&e_key, key) &&
- mlx5e_encap_take(e))
- return e;
- }
-
- return NULL;
-}
-
-static struct mlx5e_decap_entry *
-mlx5e_decap_get(struct mlx5e_priv *priv, struct mlx5e_decap_key *key,
- uintptr_t hash_key)
-{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
- struct mlx5e_decap_key r_key;
- struct mlx5e_decap_entry *e;
-
- hash_for_each_possible_rcu(esw->offloads.decap_tbl, e,
- hlist, hash_key) {
- r_key = e->key;
- if (!cmp_decap_info(&r_key, key) &&
- mlx5e_decap_take(e))
- return e;
- }
- return NULL;
-}
-
-static struct ip_tunnel_info *dup_tun_info(const struct ip_tunnel_info *tun_info)
-{
- size_t tun_size = sizeof(*tun_info) + tun_info->options_len;
-
- return kmemdup(tun_info, tun_size, GFP_KERNEL);
-}
-
-static bool is_duplicated_encap_entry(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow,
- int out_index,
- struct mlx5e_encap_entry *e,
- struct netlink_ext_ack *extack)
-{
- int i;
-
- for (i = 0; i < out_index; i++) {
- if (flow->encaps[i].e != e)
- continue;
- NL_SET_ERR_MSG_MOD(extack, "can't duplicate encap action");
- netdev_err(priv->netdev, "can't duplicate encap action\n");
- return true;
- }
-
- return false;
-}
-
-static int mlx5e_set_vf_tunnel(struct mlx5_eswitch *esw,
- struct mlx5_flow_attr *attr,
- struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
- struct net_device *out_dev,
- int route_dev_ifindex,
- int out_index)
-{
- struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
- struct net_device *route_dev;
- u16 vport_num;
- int err = 0;
- u32 data;
-
- route_dev = dev_get_by_index(dev_net(out_dev), route_dev_ifindex);
-
- if (!route_dev || route_dev->netdev_ops != &mlx5e_netdev_ops ||
- !mlx5e_tc_is_vf_tunnel(out_dev, route_dev))
- goto out;
-
- err = mlx5e_tc_query_route_vport(out_dev, route_dev, &vport_num);
- if (err)
- goto out;
-
- attr->dest_chain = 0;
- attr->action |= MLX5_FLOW_CONTEXT_ACTION_MOD_HDR;
- esw_attr->dests[out_index].flags |= MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE;
- data = mlx5_eswitch_get_vport_metadata_for_set(esw_attr->in_mdev->priv.eswitch,
- vport_num);
- err = mlx5e_tc_match_to_reg_set(esw->dev, mod_hdr_acts,
- MLX5_FLOW_NAMESPACE_FDB, VPORT_TO_REG, data);
- if (err)
- goto out;
-
-out:
- if (route_dev)
- dev_put(route_dev);
- return err;
-}
-
-static int mlx5e_attach_encap(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow,
- struct net_device *mirred_dev,
- int out_index,
- struct netlink_ext_ack *extack,
- struct net_device **encap_dev,
- bool *encap_valid)
-{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
- struct mlx5e_tc_flow_parse_attr *parse_attr;
- struct mlx5_flow_attr *attr = flow->attr;
- const struct ip_tunnel_info *tun_info;
- struct encap_key key;
- struct mlx5e_encap_entry *e;
- unsigned short family;
- uintptr_t hash_key;
- int err = 0;
-
- parse_attr = attr->parse_attr;
- tun_info = parse_attr->tun_info[out_index];
- family = ip_tunnel_info_af(tun_info);
- key.ip_tun_key = &tun_info->key;
- key.tc_tunnel = mlx5e_get_tc_tun(mirred_dev);
- if (!key.tc_tunnel) {
- NL_SET_ERR_MSG_MOD(extack, "Unsupported tunnel");
- return -EOPNOTSUPP;
- }
-
- hash_key = hash_encap_info(&key);
-
- mutex_lock(&esw->offloads.encap_tbl_lock);
- e = mlx5e_encap_get(priv, &key, hash_key);
-
- /* must verify if encap is valid or not */
- if (e) {
- /* Check that entry was not already attached to this flow */
- if (is_duplicated_encap_entry(priv, flow, out_index, e, extack)) {
- err = -EOPNOTSUPP;
- goto out_err;
- }
-
- mutex_unlock(&esw->offloads.encap_tbl_lock);
- wait_for_completion(&e->res_ready);
-
- /* Protect against concurrent neigh update. */
- mutex_lock(&esw->offloads.encap_tbl_lock);
- if (e->compl_result < 0) {
- err = -EREMOTEIO;
- goto out_err;
- }
- goto attach_flow;
- }
-
- e = kzalloc(sizeof(*e), GFP_KERNEL);
- if (!e) {
- err = -ENOMEM;
- goto out_err;
- }
-
- refcount_set(&e->refcnt, 1);
- init_completion(&e->res_ready);
-
- tun_info = dup_tun_info(tun_info);
- if (!tun_info) {
- err = -ENOMEM;
- goto out_err_init;
- }
- e->tun_info = tun_info;
- err = mlx5e_tc_tun_init_encap_attr(mirred_dev, priv, e, extack);
- if (err)
- goto out_err_init;
-
- INIT_LIST_HEAD(&e->flows);
- hash_add_rcu(esw->offloads.encap_tbl, &e->encap_hlist, hash_key);
- mutex_unlock(&esw->offloads.encap_tbl_lock);
-
- if (family == AF_INET)
- err = mlx5e_tc_tun_create_header_ipv4(priv, mirred_dev, e);
- else if (family == AF_INET6)
- err = mlx5e_tc_tun_create_header_ipv6(priv, mirred_dev, e);
-
- /* Protect against concurrent neigh update. */
- mutex_lock(&esw->offloads.encap_tbl_lock);
- complete_all(&e->res_ready);
- if (err) {
- e->compl_result = err;
- goto out_err;
- }
- e->compl_result = 1;
-
-attach_flow:
- err = mlx5e_set_vf_tunnel(esw, attr, &parse_attr->mod_hdr_acts, e->out_dev,
- e->route_dev_ifindex, out_index);
- if (err)
- goto out_err;
-
- flow->encaps[out_index].e = e;
- list_add(&flow->encaps[out_index].list, &e->flows);
- flow->encaps[out_index].index = out_index;
- *encap_dev = e->out_dev;
- if (e->flags & MLX5_ENCAP_ENTRY_VALID) {
- attr->esw_attr->dests[out_index].pkt_reformat = e->pkt_reformat;
- attr->esw_attr->dests[out_index].flags |= MLX5_ESW_DEST_ENCAP_VALID;
- *encap_valid = true;
- } else {
- *encap_valid = false;
- }
- mutex_unlock(&esw->offloads.encap_tbl_lock);
-
- return err;
-
-out_err:
- mutex_unlock(&esw->offloads.encap_tbl_lock);
- if (e)
- mlx5e_encap_put(priv, e);
- return err;
-
-out_err_init:
- mutex_unlock(&esw->offloads.encap_tbl_lock);
- kfree(tun_info);
- kfree(e);
- return err;
-}
-
-static int mlx5e_attach_decap(struct mlx5e_priv *priv,
- struct mlx5e_tc_flow *flow,
- struct netlink_ext_ack *extack)
-{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
- struct mlx5_esw_flow_attr *attr = flow->attr->esw_attr;
- struct mlx5e_tc_flow_parse_attr *parse_attr;
- struct mlx5e_decap_entry *d;
- struct mlx5e_decap_key key;
- uintptr_t hash_key;
- int err = 0;
-
- parse_attr = flow->attr->parse_attr;
- if (sizeof(parse_attr->eth) > MLX5_CAP_ESW(priv->mdev, max_encap_header_size)) {
- NL_SET_ERR_MSG_MOD(extack,
- "encap header larger than max supported");
- return -EOPNOTSUPP;
- }
-
- key.key = parse_attr->eth;
- hash_key = hash_decap_info(&key);
- mutex_lock(&esw->offloads.decap_tbl_lock);
- d = mlx5e_decap_get(priv, &key, hash_key);
- if (d) {
- mutex_unlock(&esw->offloads.decap_tbl_lock);
- wait_for_completion(&d->res_ready);
- mutex_lock(&esw->offloads.decap_tbl_lock);
- if (d->compl_result) {
- err = -EREMOTEIO;
- goto out_free;
- }
- goto found;
- }
-
- d = kzalloc(sizeof(*d), GFP_KERNEL);
- if (!d) {
- err = -ENOMEM;
- goto out_err;
- }
-
- d->key = key;
- refcount_set(&d->refcnt, 1);
- init_completion(&d->res_ready);
- INIT_LIST_HEAD(&d->flows);
- hash_add_rcu(esw->offloads.decap_tbl, &d->hlist, hash_key);
- mutex_unlock(&esw->offloads.decap_tbl_lock);
-
- d->pkt_reformat = mlx5_packet_reformat_alloc(priv->mdev,
- MLX5_REFORMAT_TYPE_L3_TUNNEL_TO_L2,
- sizeof(parse_attr->eth),
- &parse_attr->eth,
- MLX5_FLOW_NAMESPACE_FDB);
- if (IS_ERR(d->pkt_reformat)) {
- err = PTR_ERR(d->pkt_reformat);
- d->compl_result = err;
- }
- mutex_lock(&esw->offloads.decap_tbl_lock);
- complete_all(&d->res_ready);
- if (err)
- goto out_free;
-
-found:
- flow->decap_reformat = d;
- attr->decap_pkt_reformat = d->pkt_reformat;
- list_add(&flow->l3_to_l2_reformat, &d->flows);
- mutex_unlock(&esw->offloads.decap_tbl_lock);
- return 0;
-
-out_free:
- mutex_unlock(&esw->offloads.decap_tbl_lock);
- mlx5e_decap_put(priv, d);
- return err;
-
-out_err:
- mutex_unlock(&esw->offloads.decap_tbl_lock);
- return err;
-}
-
static int parse_tc_vlan_action(struct mlx5e_priv *priv,
const struct flow_action_entry *act,
struct mlx5_esw_flow_attr *attr,
@@ -4423,7 +3554,8 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv,
if (encap) {
parse_attr->mirred_ifindex[esw_attr->out_count] =
out_dev->ifindex;
- parse_attr->tun_info[esw_attr->out_count] = dup_tun_info(info);
+ parse_attr->tun_info[esw_attr->out_count] =
+ mlx5e_dup_tun_info(info);
if (!parse_attr->tun_info[esw_attr->out_count])
return -ENOMEM;
encap = false;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
index 1e4ee02bfb1c..5434bbb9a217 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -37,6 +37,8 @@
#include "en.h"
#include "eswitch.h"
#include "en/tc_ct.h"
+#include "en/tc_tun.h"
+#include "en_rep.h"
#define MLX5E_TC_FLOW_ID_MASK 0x0000ffff
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* Re: [net-next V2 12/17] net/mlx5e: Extract tc tunnel encap/decap code to dedicated file
2021-02-06 5:02 ` [net-next V2 12/17] net/mlx5e: Extract tc tunnel encap/decap code to dedicated file Saeed Mahameed
@ 2021-02-09 20:35 ` Guenter Roeck
0 siblings, 0 replies; 39+ messages in thread
From: Guenter Roeck @ 2021-02-09 20:35 UTC (permalink / raw)
To: Saeed Mahameed
Cc: Jakub Kicinski, David S. Miller, netdev, Vlad Buslov,
Dmytro Linkin, Roi Dayan, Saeed Mahameed
On Fri, Feb 05, 2021 at 09:02:35PM -0800, Saeed Mahameed wrote:
> From: Vlad Buslov <vladbu@nvidia.com>
>
> Following patches in series extend the extracted code with routing
> infrastructure. To improve code modularity created a dedicated
> tc_tun_encap.c source file and move encap/decap related code to the new
> file. Export code that is used by both regular TC code and encap/decap code
> into tc_priv.h (new header intended to be used only by TC module). Rename
> some exported functions by adding "mlx5e_" prefix to their names.
>
> Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
> Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
> Reviewed-by: Roi Dayan <roid@nvidia.com>
> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
When building s390:defconfig, this patch results in:
In file included from drivers/net/ethernet/mellanox/mlx5/core/en_tc.h:40,
from drivers/net/ethernet/mellanox/mlx5/core/en_main.c:45:
drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h:24:29: error: field 'match_level' has incomplete type
Bisect log attached.
Guenter
---
# bad: [a4bfd8d46ac357c12529e4eebb6c89502b03ecc9] Add linux-next specific files for 20210209
# good: [92bf22614b21a2706f4993b278017e437f7785b3] Linux 5.11-rc7
git bisect start 'HEAD' 'v5.11-rc7'
# bad: [a8eb921ba7e8e77d994a1c6c69c8ef08456ecf53] Merge remote-tracking branch 'crypto/master'
git bisect bad a8eb921ba7e8e77d994a1c6c69c8ef08456ecf53
# good: [b68df186dae8ae890df08059bb068b78252b053a] Merge remote-tracking branch 'hid/for-next'
git bisect good b68df186dae8ae890df08059bb068b78252b053a
# good: [1299616023a0db19be4ff5588db4fb61d8cd51f9] Merge tag 'mt76-for-kvalo-2021-01-29' of https://github.com/nbd168/wireless
git bisect good 1299616023a0db19be4ff5588db4fb61d8cd51f9
# good: [db9deabba7a9a6fded1aeda51dda1ce9b78f0711] Merge remote-tracking branch 'v4l-dvb-next/master'
git bisect good db9deabba7a9a6fded1aeda51dda1ce9b78f0711
# good: [ed19e2282476b1e0dc01c67fc19010ab791a7a46] Merge remote-tracking branch 'rdma/for-next'
git bisect good ed19e2282476b1e0dc01c67fc19010ab791a7a46
# bad: [ae3b49168f93b4432b852ca6f2f7323fead45e8d] Merge remote-tracking branch 'bluetooth/master'
git bisect bad ae3b49168f93b4432b852ca6f2f7323fead45e8d
# good: [215cb7d3823e798de327e3232e396434fab84f42] bpf/benchs/bench_ringbufs: Remove unneeded semicolon
git bisect good 215cb7d3823e798de327e3232e396434fab84f42
# good: [de71a6cb4bf24d8993b9ca90d1ddb131b60251a1] Bluetooth: btusb: Fix memory leak in btusb_mtk_wmt_recv
git bisect good de71a6cb4bf24d8993b9ca90d1ddb131b60251a1
# bad: [8914add2c9e5518f6a864936658bba5752510b39] net/mlx5e: Handle FIB events to update tunnel endpoint device
git bisect bad 8914add2c9e5518f6a864936658bba5752510b39
# good: [4ad9116c84ed3243f7b706f07646a995f3bca502] net/mlx5e: Remove redundant match on tunnel destination mac
git bisect good 4ad9116c84ed3243f7b706f07646a995f3bca502
# bad: [0d9f96471493d5483d116c137693f03604332a04] net/mlx5e: Extract tc tunnel encap/decap code to dedicated file
git bisect bad 0d9f96471493d5483d116c137693f03604332a04
# good: [48d216e5596a58e3cfa6d4548343f982c5921b79] net/mlx5e: Refactor reg_c1 usage
git bisect good 48d216e5596a58e3cfa6d4548343f982c5921b79
# good: [8e404fefa58b6138531e3d4b5647ee79f75ae9a8] net/mlx5e: Match recirculated packet miss in slow table using reg_c1
git bisect good 8e404fefa58b6138531e3d4b5647ee79f75ae9a8
# first bad commit: [0d9f96471493d5483d116c137693f03604332a04] net/mlx5e: Extract tc tunnel encap/decap code to dedicated file
^ permalink raw reply [flat|nested] 39+ messages in thread
* [net-next V2 13/17] net/mlx5e: Create route entry infrastructure
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (11 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 12/17] net/mlx5e: Extract tc tunnel encap/decap code to dedicated file Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 14/17] net/mlx5e: Refactor neigh update infrastructure Saeed Mahameed
` (5 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Implement dedicated route entry infrastructure to be used in following
patch by route update event. Both encap (indirectly through their
corresponding encap entries) and decap (directly) flows are attached to
routing entry. Since route update also requires updating encap (route
device MAC address is a source MAC address of tunnel encapsulation), same
encap_tbl_lock mutex is used for synchronization.
The new infrastructure looks similar to existing infrastructures for shared
encap, mod_hdr and hairpin entries:
- Per-eswitch hash table is used for quick entry lookup.
- Flows are attached to per-entry linked list and hold reference to entry
during their lifetime.
- Atomic reference counting and rcu mechanisms are used as synchronization
primitives for concurrent access.
The infrastructure also enables connection tracking on stacked devices
topology by attaching CT chain 0 flow on tunneling dev to decap route
entry.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../ethernet/mellanox/mlx5/core/en/tc_priv.h | 11 +
.../mellanox/mlx5/core/en/tc_tun_encap.c | 258 +++++++++++++++++-
.../mellanox/mlx5/core/en/tc_tun_encap.h | 5 +
.../net/ethernet/mellanox/mlx5/core/en_rep.h | 1 +
.../net/ethernet/mellanox/mlx5/core/en_tc.c | 24 +-
.../net/ethernet/mellanox/mlx5/core/eswitch.c | 1 +
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 1 +
7 files changed, 290 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
index e0ae24d9a740..14db9b5accb1 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
@@ -25,6 +25,7 @@ enum {
MLX5E_TC_FLOW_FLAG_DELETED = MLX5E_TC_FLOW_BASE + 6,
MLX5E_TC_FLOW_FLAG_CT = MLX5E_TC_FLOW_BASE + 7,
MLX5E_TC_FLOW_FLAG_L3_TO_L2_DECAP = MLX5E_TC_FLOW_BASE + 8,
+ MLX5E_TC_FLOW_FLAG_TUN_RX = MLX5E_TC_FLOW_BASE + 9,
};
struct mlx5e_tc_flow_parse_attr {
@@ -59,6 +60,11 @@ struct encap_flow_item {
int index;
};
+struct encap_route_flow_item {
+ struct mlx5e_route_entry *r; /* attached route instance */
+ int index;
+};
+
struct mlx5e_tc_flow {
struct rhash_head node;
struct mlx5e_priv *priv;
@@ -70,6 +76,11 @@ struct mlx5e_tc_flow {
struct list_head l3_to_l2_reformat;
struct mlx5e_decap_entry *decap_reformat;
+ /* flows sharing same route entry */
+ struct list_head decap_routes;
+ struct mlx5e_route_entry *decap_route;
+ struct encap_route_flow_item encap_routes[MLX5_MAX_FLOW_FWD_VPORTS];
+
/* Flow can be associated with multiple encap IDs.
* The number of encaps is bounded by the number of supported
* destinations.
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
index 63652911d56e..75fcffecfe3d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
@@ -7,6 +7,23 @@
#include "rep/tc.h"
#include "diag/en_tc_tracepoint.h"
+struct mlx5e_route_key {
+ int ip_version;
+ union {
+ __be32 v4;
+ struct in6_addr v6;
+ } endpoint_ip;
+};
+
+struct mlx5e_route_entry {
+ struct mlx5e_route_key key;
+ struct list_head encap_entries;
+ struct list_head decap_flows;
+ struct hlist_node hlist;
+ refcount_t refcnt;
+ struct rcu_head rcu;
+};
+
int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
struct mlx5_flow_spec *spec)
{
@@ -29,10 +46,13 @@ int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
outer_headers.src_ipv4_src_ipv6.ipv4_layout.ipv4);
tun_attr->dst_ip.v4 = *(__be32 *)daddr;
tun_attr->src_ip.v4 = *(__be32 *)saddr;
+ if (!tun_attr->dst_ip.v4 || !tun_attr->src_ip.v4)
+ return 0;
}
#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
else if (ip_version == 6) {
int ipv6_size = MLX5_FLD_SZ_BYTES(ipv6_layout, ipv6);
+ struct in6_addr zerov6 = {};
daddr = MLX5_ADDR_OF(fte_match_param, spec->match_value,
outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6);
@@ -40,8 +60,15 @@ int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
outer_headers.src_ipv4_src_ipv6.ipv6_layout.ipv6);
memcpy(&tun_attr->dst_ip.v6, daddr, ipv6_size);
memcpy(&tun_attr->src_ip.v6, saddr, ipv6_size);
+ if (!memcmp(&tun_attr->dst_ip.v6, &zerov6, sizeof(zerov6)) ||
+ !memcmp(&tun_attr->src_ip.v6, &zerov6, sizeof(zerov6)))
+ return 0;
}
#endif
+ /* Only set the flag if both src and dst ip addresses exist. They are
+ * required to establish routing.
+ */
+ flow_flag_set(flow, TUN_RX);
return 0;
}
@@ -325,6 +352,7 @@ void mlx5e_encap_put(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e)
if (!refcount_dec_and_mutex_lock(&e->refcnt, &esw->offloads.encap_tbl_lock))
return;
+ list_del(&e->route_list);
hash_del_rcu(&e->encap_hlist);
mutex_unlock(&esw->offloads.encap_tbl_lock);
@@ -343,12 +371,20 @@ static void mlx5e_decap_put(struct mlx5e_priv *priv, struct mlx5e_decap_entry *d
mlx5e_decap_dealloc(priv, d);
}
+static void mlx5e_detach_encap_route(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ int out_index);
+
void mlx5e_detach_encap(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow, int out_index)
{
struct mlx5e_encap_entry *e = flow->encaps[out_index].e;
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ if (flow->attr->esw_attr->dests[out_index].flags &
+ MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE)
+ mlx5e_detach_encap_route(priv, flow, out_index);
+
/* flow wasn't fully initialized */
if (!e)
return;
@@ -360,6 +396,7 @@ void mlx5e_detach_encap(struct mlx5e_priv *priv,
mutex_unlock(&esw->offloads.encap_tbl_lock);
return;
}
+ list_del(&e->route_list);
hash_del_rcu(&e->encap_hlist);
mutex_unlock(&esw->offloads.encap_tbl_lock);
@@ -531,6 +568,12 @@ static int mlx5e_set_vf_tunnel(struct mlx5_eswitch *esw,
return err;
}
+static int mlx5e_attach_encap_route(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ struct mlx5e_encap_entry *e,
+ bool new_encap_entry,
+ int out_index);
+
int mlx5e_attach_encap(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow,
struct net_device *mirred_dev,
@@ -545,6 +588,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
const struct ip_tunnel_info *tun_info;
struct encap_key key;
struct mlx5e_encap_entry *e;
+ bool entry_created = false;
unsigned short family;
uintptr_t hash_key;
int err = 0;
@@ -592,6 +636,8 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
refcount_set(&e->refcnt, 1);
init_completion(&e->res_ready);
+ entry_created = true;
+ INIT_LIST_HEAD(&e->route_list);
tun_info = mlx5e_dup_tun_info(tun_info);
if (!tun_info) {
@@ -622,8 +668,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
e->compl_result = 1;
attach_flow:
- err = mlx5e_set_vf_tunnel(esw, attr, &parse_attr->mod_hdr_acts, e->out_dev,
- e->route_dev_ifindex, out_index);
+ err = mlx5e_attach_encap_route(priv, flow, e, entry_created, out_index);
if (err)
goto out_err;
@@ -732,3 +777,212 @@ int mlx5e_attach_decap(struct mlx5e_priv *priv,
mutex_unlock(&esw->offloads.decap_tbl_lock);
return err;
}
+
+static int cmp_route_info(struct mlx5e_route_key *a,
+ struct mlx5e_route_key *b)
+{
+ if (a->ip_version == 4 && b->ip_version == 4)
+ return memcmp(&a->endpoint_ip.v4, &b->endpoint_ip.v4,
+ sizeof(a->endpoint_ip.v4));
+ else if (a->ip_version == 6 && b->ip_version == 6)
+ return memcmp(&a->endpoint_ip.v6, &b->endpoint_ip.v6,
+ sizeof(a->endpoint_ip.v6));
+ return 1;
+}
+
+static u32 hash_route_info(struct mlx5e_route_key *key)
+{
+ if (key->ip_version == 4)
+ return jhash(&key->endpoint_ip.v4, sizeof(key->endpoint_ip.v4), 0);
+ return jhash(&key->endpoint_ip.v6, sizeof(key->endpoint_ip.v6), 0);
+}
+
+static struct mlx5e_route_entry *
+mlx5e_route_get(struct mlx5e_priv *priv, struct mlx5e_route_key *key,
+ u32 hash_key)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_route_key r_key;
+ struct mlx5e_route_entry *r;
+
+ hash_for_each_possible(esw->offloads.route_tbl, r, hlist, hash_key) {
+ r_key = r->key;
+ if (!cmp_route_info(&r_key, key) &&
+ refcount_inc_not_zero(&r->refcnt))
+ return r;
+ }
+ return NULL;
+}
+
+static struct mlx5e_route_entry *
+mlx5e_route_get_create(struct mlx5e_priv *priv,
+ struct mlx5e_route_key *key)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_route_entry *r;
+ u32 hash_key;
+
+ hash_key = hash_route_info(key);
+ r = mlx5e_route_get(priv, key, hash_key);
+ if (r)
+ return r;
+
+ r = kzalloc(sizeof(*r), GFP_KERNEL);
+ if (!r)
+ return ERR_PTR(-ENOMEM);
+
+ r->key = *key;
+ refcount_set(&r->refcnt, 1);
+ INIT_LIST_HEAD(&r->decap_flows);
+ INIT_LIST_HEAD(&r->encap_entries);
+ hash_add(esw->offloads.route_tbl, &r->hlist, hash_key);
+ return r;
+}
+
+int mlx5e_attach_decap_route(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_tc_flow_parse_attr *parse_attr;
+ struct mlx5_flow_attr *attr = flow->attr;
+ struct mlx5_esw_flow_attr *esw_attr;
+ struct mlx5e_route_entry *r;
+ struct mlx5e_route_key key;
+ int err = 0;
+
+ esw_attr = attr->esw_attr;
+ parse_attr = attr->parse_attr;
+ mutex_lock(&esw->offloads.encap_tbl_lock);
+ if (!esw_attr->rx_tun_attr)
+ goto out;
+
+ err = mlx5e_tc_tun_route_lookup(priv, &parse_attr->spec, attr);
+ if (err || !esw_attr->rx_tun_attr->decap_vport)
+ goto out;
+
+ key.ip_version = attr->ip_version;
+ if (key.ip_version == 4)
+ key.endpoint_ip.v4 = esw_attr->rx_tun_attr->dst_ip.v4;
+ else
+ key.endpoint_ip.v6 = esw_attr->rx_tun_attr->dst_ip.v6;
+
+ r = mlx5e_route_get_create(priv, &key);
+ if (IS_ERR(r)) {
+ err = PTR_ERR(r);
+ goto out;
+ }
+
+ flow->decap_route = r;
+ list_add(&flow->decap_routes, &r->decap_flows);
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+ return 0;
+
+out:
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+ return err;
+}
+
+static int mlx5e_attach_encap_route(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ struct mlx5e_encap_entry *e,
+ bool new_encap_entry,
+ int out_index)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_tc_flow_parse_attr *parse_attr;
+ struct mlx5_flow_attr *attr = flow->attr;
+ const struct ip_tunnel_info *tun_info;
+ struct mlx5_esw_flow_attr *esw_attr;
+ struct mlx5e_route_entry *r;
+ struct mlx5e_route_key key;
+ unsigned short family;
+ int err = 0;
+
+ esw_attr = attr->esw_attr;
+ parse_attr = attr->parse_attr;
+ tun_info = parse_attr->tun_info[out_index];
+ family = ip_tunnel_info_af(tun_info);
+
+ if (family == AF_INET) {
+ key.endpoint_ip.v4 = tun_info->key.u.ipv4.src;
+ key.ip_version = 4;
+ } else if (family == AF_INET6) {
+ key.endpoint_ip.v6 = tun_info->key.u.ipv6.src;
+ key.ip_version = 6;
+ }
+
+ err = mlx5e_set_vf_tunnel(esw, attr, &parse_attr->mod_hdr_acts, e->out_dev,
+ e->route_dev_ifindex, out_index);
+ if (err || !(esw_attr->dests[out_index].flags &
+ MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE))
+ return err;
+
+ r = mlx5e_route_get_create(priv, &key);
+ if (IS_ERR(r))
+ return PTR_ERR(r);
+
+ flow->encap_routes[out_index].r = r;
+ if (new_encap_entry)
+ list_add(&e->route_list, &r->encap_entries);
+ flow->encap_routes[out_index].index = out_index;
+ return 0;
+}
+
+static void mlx5e_route_dealloc(struct mlx5e_priv *priv,
+ struct mlx5e_route_entry *r)
+{
+ WARN_ON(!list_empty(&r->decap_flows));
+ WARN_ON(!list_empty(&r->encap_entries));
+
+ kfree_rcu(r, rcu);
+}
+
+void mlx5e_detach_decap_route(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_route_entry *r = flow->decap_route;
+
+ if (!r)
+ return;
+
+ mutex_lock(&esw->offloads.encap_tbl_lock);
+ list_del(&flow->decap_routes);
+ flow->decap_route = NULL;
+
+ if (!refcount_dec_and_test(&r->refcnt)) {
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+ return;
+ }
+ hash_del_rcu(&r->hlist);
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+
+ mlx5e_route_dealloc(priv, r);
+}
+
+static void mlx5e_detach_encap_route(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow,
+ int out_index)
+{
+ struct mlx5e_route_entry *r = flow->encap_routes[out_index].r;
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_encap_entry *e, *tmp;
+
+ if (!r)
+ return;
+
+ mutex_lock(&esw->offloads.encap_tbl_lock);
+ flow->encap_routes[out_index].r = NULL;
+
+ if (!refcount_dec_and_test(&r->refcnt)) {
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+ return;
+ }
+ list_for_each_entry_safe(e, tmp, &r->encap_entries, route_list)
+ list_del_init(&e->route_list);
+ hash_del_rcu(&r->hlist);
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+
+ mlx5e_route_dealloc(priv, r);
+}
+
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h
index 81b9fef1cf2a..9939cff6e6c5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h
@@ -22,6 +22,11 @@ int mlx5e_attach_decap(struct mlx5e_priv *priv,
void mlx5e_detach_decap(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow);
+int mlx5e_attach_decap_route(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow);
+void mlx5e_detach_decap_route(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow *flow);
+
struct ip_tunnel_info *mlx5e_dup_tun_info(const struct ip_tunnel_info *tun_info);
int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
index 988195ab1c54..8e04126f088a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
@@ -181,6 +181,7 @@ struct mlx5e_encap_entry {
*/
struct hlist_node encap_hlist;
struct list_head flows;
+ struct list_head route_list;
struct mlx5_pkt_reformat *pkt_reformat;
const struct ip_tunnel_info *tun_info;
unsigned char h_dest[ETH_ALEN]; /* destination eth addr */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index cfe340e23dfc..2cc31f99db9b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -1251,6 +1251,12 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
return -EOPNOTSUPP;
}
+ if (flow_flag_test(flow, TUN_RX)) {
+ err = mlx5e_attach_decap_route(priv, flow);
+ if (err)
+ return err;
+ }
+
if (flow_flag_test(flow, L3_TO_L2_DECAP)) {
err = mlx5e_attach_decap(priv, flow, extack);
if (err)
@@ -1335,8 +1341,10 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv,
{
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
struct mlx5_flow_attr *attr = flow->attr;
+ struct mlx5_esw_flow_attr *esw_attr;
int out_index;
+ esw_attr = attr->esw_attr;
mlx5e_put_flow_tunnel_id(flow);
if (flow_flag_test(flow, NOT_READY))
@@ -1354,11 +1362,15 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv,
mlx5_eswitch_del_vlan_action(esw, attr);
- for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++)
- if (attr->esw_attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP) {
+ if (flow->decap_route)
+ mlx5e_detach_decap_route(priv, flow);
+
+ for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++) {
+ if (esw_attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP) {
mlx5e_detach_encap(priv, flow, out_index);
kfree(attr->parse_attr->tun_info[out_index]);
}
+ }
kvfree(attr->parse_attr);
kvfree(attr->esw_attr->rx_tun_attr);
@@ -1368,7 +1380,7 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv,
mlx5e_detach_mod_hdr(priv, flow);
if (attr->action & MLX5_FLOW_CONTEXT_ACTION_COUNT)
- mlx5_fc_destroy(attr->esw_attr->counter_dev, attr->counter);
+ mlx5_fc_destroy(esw_attr->counter_dev, attr->counter);
if (flow_flag_test(flow, L3_TO_L2_DECAP))
mlx5e_detach_decap(priv, flow);
@@ -3692,12 +3704,6 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv,
}
}
- if (decap && esw_attr->rx_tun_attr) {
- err = mlx5e_tc_tun_route_lookup(priv, &parse_attr->spec, attr);
- if (err)
- return err;
- }
-
/* always set IP version for indirect table handling */
attr->ip_version = mlx5e_tc_get_ip_version(&parse_attr->spec, true);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index aba17835465b..310c405e81d7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1830,6 +1830,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
mutex_init(&esw->offloads.decap_tbl_lock);
hash_init(esw->offloads.decap_tbl);
mlx5e_mod_hdr_tbl_init(&esw->offloads.mod_hdr);
+ hash_init(esw->offloads.route_tbl);
atomic64_set(&esw->offloads.num_flows, 0);
ida_init(&esw->offloads.vport_metadata_ida);
xa_init_flags(&esw->offloads.vhca_map, XA_FLAGS_ALLOC);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index 34a21ff95472..b90724e27960 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -214,6 +214,7 @@ struct mlx5_esw_offload {
struct mutex peer_mutex;
struct mutex encap_tbl_lock; /* protects encap_tbl */
DECLARE_HASHTABLE(encap_tbl, 8);
+ DECLARE_HASHTABLE(route_tbl, 8);
struct mutex decap_tbl_lock; /* protects decap_tbl */
DECLARE_HASHTABLE(decap_tbl, 8);
struct mod_hdr_tbl mod_hdr;
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 14/17] net/mlx5e: Refactor neigh update infrastructure
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (12 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 13/17] net/mlx5e: Create route entry infrastructure Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 15/17] net/mlx5e: TC preparation refactoring for routing update event Saeed Mahameed
` (4 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Following patches in series implements route update which can cause encap
entries to migrate between routing devices. Consecutively, their parent
nhe's need to be also transferable between devices instead of having neigh
device as a part of their immutable key. Move neigh device from struct
mlx5_neigh to struct mlx5e_neigh_hash_entry and check that nhe and neigh
devices are the same in workqueue neigh update handler.
Save neigh net_device that can change dynamically in dedicated nhe->dev
field. With FIB event handler that is implemented in following patches
changing nhe->dev, NETEVENT_DELAY_PROBE_TIME_UPDATE handler can
concurrently access the nhe entry when traversing neigh list under rcu read
lock. Processing stale values in that handler doesn't change the handler
logic, so just wrap all accesses to the dev pointer in {WRITE|READ}_ONCE()
helpers.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../mlx5/core/diag/en_rep_tracepoint.h | 4 ++--
.../mlx5/core/diag/en_tc_tracepoint.h | 4 ++--
.../mellanox/mlx5/core/en/rep/neigh.c | 16 +++++++++-----
.../mellanox/mlx5/core/en/rep/neigh.h | 3 ++-
.../ethernet/mellanox/mlx5/core/en/rep/tc.c | 8 ++++---
.../ethernet/mellanox/mlx5/core/en/rep/tc.h | 4 +++-
.../ethernet/mellanox/mlx5/core/en/tc_tun.c | 22 +++++++------------
.../mellanox/mlx5/core/en/tc_tun_encap.c | 2 +-
.../net/ethernet/mellanox/mlx5/core/en_rep.h | 3 +--
9 files changed, 35 insertions(+), 31 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/en_rep_tracepoint.h b/drivers/net/ethernet/mellanox/mlx5/core/diag/en_rep_tracepoint.h
index 1177860a2ee4..f15718db5d0e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/diag/en_rep_tracepoint.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/en_rep_tracepoint.h
@@ -15,7 +15,7 @@ TRACE_EVENT(mlx5e_rep_neigh_update,
TP_PROTO(const struct mlx5e_neigh_hash_entry *nhe, const u8 *ha,
bool neigh_connected),
TP_ARGS(nhe, ha, neigh_connected),
- TP_STRUCT__entry(__string(devname, nhe->m_neigh.dev->name)
+ TP_STRUCT__entry(__string(devname, nhe->neigh_dev->name)
__array(u8, ha, ETH_ALEN)
__array(u8, v4, 4)
__array(u8, v6, 16)
@@ -25,7 +25,7 @@ TRACE_EVENT(mlx5e_rep_neigh_update,
struct in6_addr *pin6;
__be32 *p32;
- __assign_str(devname, mn->dev->name);
+ __assign_str(devname, nhe->neigh_dev->name);
__entry->neigh_connected = neigh_connected;
memcpy(__entry->ha, ha, ETH_ALEN);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/en_tc_tracepoint.h b/drivers/net/ethernet/mellanox/mlx5/core/diag/en_tc_tracepoint.h
index d4e6cfaaade3..ac52ef37f38a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/diag/en_tc_tracepoint.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/en_tc_tracepoint.h
@@ -77,7 +77,7 @@ TRACE_EVENT(mlx5e_stats_flower,
TRACE_EVENT(mlx5e_tc_update_neigh_used_value,
TP_PROTO(const struct mlx5e_neigh_hash_entry *nhe, bool neigh_used),
TP_ARGS(nhe, neigh_used),
- TP_STRUCT__entry(__string(devname, nhe->m_neigh.dev->name)
+ TP_STRUCT__entry(__string(devname, nhe->neigh_dev->name)
__array(u8, v4, 4)
__array(u8, v6, 16)
__field(bool, neigh_used)
@@ -86,7 +86,7 @@ TRACE_EVENT(mlx5e_tc_update_neigh_used_value,
struct in6_addr *pin6;
__be32 *p32;
- __assign_str(devname, mn->dev->name);
+ __assign_str(devname, nhe->neigh_dev->name);
__entry->neigh_used = neigh_used;
p32 = (__be32 *)__entry->v4;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c
index 616ee585a985..be0ee03de721 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.c
@@ -129,10 +129,10 @@ static void mlx5e_rep_neigh_update(struct work_struct *work)
work);
struct mlx5e_neigh_hash_entry *nhe = update_work->nhe;
struct neighbour *n = update_work->n;
+ bool neigh_connected, same_dev;
struct mlx5e_encap_entry *e;
unsigned char ha[ETH_ALEN];
struct mlx5e_priv *priv;
- bool neigh_connected;
u8 nud_state, dead;
rtnl_lock();
@@ -146,12 +146,16 @@ static void mlx5e_rep_neigh_update(struct work_struct *work)
memcpy(ha, n->ha, ETH_ALEN);
nud_state = n->nud_state;
dead = n->dead;
+ same_dev = READ_ONCE(nhe->neigh_dev) == n->dev;
read_unlock_bh(&n->lock);
neigh_connected = (nud_state & NUD_VALID) && !dead;
trace_mlx5e_rep_neigh_update(nhe, ha, neigh_connected);
+ if (!same_dev)
+ goto out;
+
list_for_each_entry(e, &nhe->encap_list, encap_list) {
if (!mlx5e_encap_take(e))
continue;
@@ -160,6 +164,7 @@ static void mlx5e_rep_neigh_update(struct work_struct *work)
mlx5e_rep_update_flows(priv, e, neigh_connected, ha);
mlx5e_encap_put(priv, e);
}
+out:
rtnl_unlock();
mlx5e_release_neigh_update_work(update_work);
}
@@ -175,7 +180,6 @@ static struct neigh_update_work *mlx5e_alloc_neigh_update_work(struct mlx5e_priv
if (WARN_ON(!update_work))
return NULL;
- m_neigh.dev = n->dev;
m_neigh.family = n->ops->family;
memcpy(&m_neigh.dst_ip, n->primary_key, n->tbl->key_len);
@@ -246,7 +250,7 @@ static int mlx5e_rep_netevent_event(struct notifier_block *nb,
rcu_read_lock();
list_for_each_entry_rcu(nhe, &neigh_update->neigh_list,
neigh_list) {
- if (p->dev == nhe->m_neigh.dev) {
+ if (p->dev == READ_ONCE(nhe->neigh_dev)) {
found = true;
break;
}
@@ -369,7 +373,8 @@ mlx5e_rep_neigh_entry_lookup(struct mlx5e_priv *priv,
}
int mlx5e_rep_neigh_entry_create(struct mlx5e_priv *priv,
- struct mlx5e_encap_entry *e,
+ struct mlx5e_neigh *m_neigh,
+ struct net_device *neigh_dev,
struct mlx5e_neigh_hash_entry **nhe)
{
int err;
@@ -379,10 +384,11 @@ int mlx5e_rep_neigh_entry_create(struct mlx5e_priv *priv,
return -ENOMEM;
(*nhe)->priv = priv;
- memcpy(&(*nhe)->m_neigh, &e->m_neigh, sizeof(e->m_neigh));
+ memcpy(&(*nhe)->m_neigh, m_neigh, sizeof(*m_neigh));
spin_lock_init(&(*nhe)->encap_list_lock);
INIT_LIST_HEAD(&(*nhe)->encap_list);
refcount_set(&(*nhe)->refcnt, 1);
+ WRITE_ONCE((*nhe)->neigh_dev, neigh_dev);
err = mlx5e_rep_neigh_entry_insert(priv, *nhe);
if (err)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.h b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.h
index 32b239189c95..6fe0ab970943 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/neigh.h
@@ -16,7 +16,8 @@ struct mlx5e_neigh_hash_entry *
mlx5e_rep_neigh_entry_lookup(struct mlx5e_priv *priv,
struct mlx5e_neigh *m_neigh);
int mlx5e_rep_neigh_entry_create(struct mlx5e_priv *priv,
- struct mlx5e_encap_entry *e,
+ struct mlx5e_neigh *m_neigh,
+ struct net_device *neigh_dev,
struct mlx5e_neigh_hash_entry **nhe);
void mlx5e_rep_neigh_entry_release(struct mlx5e_neigh_hash_entry *nhe);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
index 14bcebd4a0b6..a7ba1c84371d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
@@ -26,7 +26,9 @@ struct mlx5e_rep_indr_block_priv {
};
int mlx5e_rep_encap_entry_attach(struct mlx5e_priv *priv,
- struct mlx5e_encap_entry *e)
+ struct mlx5e_encap_entry *e,
+ struct mlx5e_neigh *m_neigh,
+ struct net_device *neigh_dev)
{
struct mlx5e_rep_priv *rpriv = priv->ppriv;
struct mlx5_rep_uplink_priv *uplink_priv = &rpriv->uplink_priv;
@@ -39,9 +41,9 @@ int mlx5e_rep_encap_entry_attach(struct mlx5e_priv *priv,
return err;
mutex_lock(&rpriv->neigh_update.encap_lock);
- nhe = mlx5e_rep_neigh_entry_lookup(priv, &e->m_neigh);
+ nhe = mlx5e_rep_neigh_entry_lookup(priv, m_neigh);
if (!nhe) {
- err = mlx5e_rep_neigh_entry_create(priv, e, &nhe);
+ err = mlx5e_rep_neigh_entry_create(priv, m_neigh, neigh_dev, &nhe);
if (err) {
mutex_unlock(&rpriv->neigh_update.encap_lock);
mlx5_tun_entropy_refcount_dec(tun_entropy,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.h
index fdf9702c2d7d..d0661578467b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.h
@@ -27,7 +27,9 @@ void mlx5e_rep_update_flows(struct mlx5e_priv *priv,
unsigned char ha[ETH_ALEN]);
int mlx5e_rep_encap_entry_attach(struct mlx5e_priv *priv,
- struct mlx5e_encap_entry *e);
+ struct mlx5e_encap_entry *e,
+ struct mlx5e_neigh *m_neigh,
+ struct net_device *neigh_dev);
void mlx5e_rep_encap_entry_detach(struct mlx5e_priv *priv,
struct mlx5e_encap_entry *e);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
index def2335c39e4..0ad22f5709a1 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
@@ -212,6 +212,7 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
{
int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size);
const struct ip_tunnel_key *tun_key = &e->tun_info->key;
+ struct mlx5e_neigh m_neigh = {};
TC_TUN_ROUTE_ATTR_INIT(attr);
int ipv4_encap_size;
char *encap_header;
@@ -247,12 +248,8 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
goto release_neigh;
}
- /* used by mlx5e_detach_encap to lookup a neigh hash table
- * entry in the neigh hash table when a user deletes a rule
- */
- e->m_neigh.dev = attr.n->dev;
- e->m_neigh.family = attr.n->ops->family;
- memcpy(&e->m_neigh.dst_ip, attr.n->primary_key, attr.n->tbl->key_len);
+ m_neigh.family = attr.n->ops->family;
+ memcpy(&m_neigh.dst_ip, attr.n->primary_key, attr.n->tbl->key_len);
e->out_dev = attr.out_dev;
e->route_dev_ifindex = attr.route_dev->ifindex;
@@ -261,7 +258,7 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
* neigh changes it's validity state, we would find the relevant neigh
* in the hash.
*/
- err = mlx5e_rep_encap_entry_attach(netdev_priv(attr.out_dev), e);
+ err = mlx5e_rep_encap_entry_attach(netdev_priv(attr.out_dev), e, &m_neigh, attr.n->dev);
if (err)
goto free_encap;
@@ -375,6 +372,7 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
{
int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size);
const struct ip_tunnel_key *tun_key = &e->tun_info->key;
+ struct mlx5e_neigh m_neigh = {};
TC_TUN_ROUTE_ATTR_INIT(attr);
struct ipv6hdr *ip6h;
int ipv6_encap_size;
@@ -409,12 +407,8 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
goto release_neigh;
}
- /* used by mlx5e_detach_encap to lookup a neigh hash table
- * entry in the neigh hash table when a user deletes a rule
- */
- e->m_neigh.dev = attr.n->dev;
- e->m_neigh.family = attr.n->ops->family;
- memcpy(&e->m_neigh.dst_ip, attr.n->primary_key, attr.n->tbl->key_len);
+ m_neigh.family = attr.n->ops->family;
+ memcpy(&m_neigh.dst_ip, attr.n->primary_key, attr.n->tbl->key_len);
e->out_dev = attr.out_dev;
e->route_dev_ifindex = attr.route_dev->ifindex;
@@ -423,7 +417,7 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
* neigh changes it's validity state, we would find the relevant neigh
* in the hash.
*/
- err = mlx5e_rep_encap_entry_attach(netdev_priv(attr.out_dev), e);
+ err = mlx5e_rep_encap_entry_attach(netdev_priv(attr.out_dev), e, &m_neigh, attr.n->dev);
if (err)
goto free_encap;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
index 75fcffecfe3d..577216744a17 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
@@ -310,7 +310,7 @@ void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe)
/* find the relevant neigh according to the cached device and
* dst ip pair
*/
- n = neigh_lookup(tbl, &m_neigh->dst_ip, m_neigh->dev);
+ n = neigh_lookup(tbl, &m_neigh->dst_ip, READ_ONCE(nhe->neigh_dev));
if (!n)
return;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
index 8e04126f088a..e947921a2d5a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
@@ -110,7 +110,6 @@ struct mlx5e_rep_priv *mlx5e_rep_to_rep_priv(struct mlx5_eswitch_rep *rep)
}
struct mlx5e_neigh {
- struct net_device *dev;
union {
__be32 v4;
struct in6_addr v6;
@@ -122,6 +121,7 @@ struct mlx5e_neigh_hash_entry {
struct rhash_head rhash_node;
struct mlx5e_neigh m_neigh;
struct mlx5e_priv *priv;
+ struct net_device *neigh_dev;
/* Save the neigh hash entry in a list on the representor in
* addition to the hash table. In order to iterate easily over the
@@ -175,7 +175,6 @@ struct mlx5e_encap_entry {
struct mlx5e_neigh_hash_entry *nhe;
/* neigh hash entry list of encaps sharing the same neigh */
struct list_head encap_list;
- struct mlx5e_neigh m_neigh;
/* a node of the eswitch encap hash table which keeping all the encap
* entries
*/
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 15/17] net/mlx5e: TC preparation refactoring for routing update event
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (13 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 14/17] net/mlx5e: Refactor neigh update infrastructure Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 16/17] net/mlx5e: Rename some encap-specific API to generic names Saeed Mahameed
` (3 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Following patch in series implement routing update event which requires
ability to modify rule match_to_reg modify header actions dynamically
during rule lifetime. In order to accommodate such behavior, refactor and
extend TC infrastructure in following ways:
- Modify mod_hdr infrastructure to preserve its parse attribute for whole
rule lifetime, instead of deallocating it after rule creation.
- Extend match_to_reg infrastructure with new function
mlx5e_tc_match_to_reg_set_and_get_id() that returns mod_hdr action id that
can be used afterwards to update the action, and
mlx5e_tc_match_to_reg_mod_hdr_change() that can modify existing actions by
its id.
- Extend tun API with new functions mlx5e_tc_tun_update_header_ipv{4|6}()
that are used to updated existing encap entry tunnel header.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../ethernet/mellanox/mlx5/core/en/tc_ct.c | 1 -
.../ethernet/mellanox/mlx5/core/en/tc_tun.c | 198 ++++++++++++++++++
.../ethernet/mellanox/mlx5/core/en/tc_tun.h | 10 +
.../net/ethernet/mellanox/mlx5/core/en_tc.c | 73 ++++++-
.../net/ethernet/mellanox/mlx5/core/en_tc.h | 15 ++
5 files changed, 288 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
index 3fb75dcdc68d..0b503ebe59ec 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c
@@ -1763,7 +1763,6 @@ __mlx5_tc_ct_flow_offload_clear(struct mlx5_tc_ct_priv *ct_priv,
goto err_set_registers;
}
- dealloc_mod_hdr_actions(mod_acts);
pre_ct_attr->modify_hdr = mod_hdr;
pre_ct_attr->action |= MLX5_FLOW_CONTEXT_ACTION_MOD_HDR;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
index 0ad22f5709a1..f8075a604605 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c
@@ -318,6 +318,105 @@ int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
return err;
}
+int mlx5e_tc_tun_update_header_ipv4(struct mlx5e_priv *priv,
+ struct net_device *mirred_dev,
+ struct mlx5e_encap_entry *e)
+{
+ int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size);
+ const struct ip_tunnel_key *tun_key = &e->tun_info->key;
+ TC_TUN_ROUTE_ATTR_INIT(attr);
+ int ipv4_encap_size;
+ char *encap_header;
+ struct iphdr *ip;
+ u8 nud_state;
+ int err;
+
+ /* add the IP fields */
+ attr.fl.fl4.flowi4_tos = tun_key->tos;
+ attr.fl.fl4.daddr = tun_key->u.ipv4.dst;
+ attr.fl.fl4.saddr = tun_key->u.ipv4.src;
+ attr.ttl = tun_key->ttl;
+
+ err = mlx5e_route_lookup_ipv4_get(priv, mirred_dev, &attr);
+ if (err)
+ return err;
+
+ ipv4_encap_size =
+ (is_vlan_dev(attr.route_dev) ? VLAN_ETH_HLEN : ETH_HLEN) +
+ sizeof(struct iphdr) +
+ e->tunnel->calc_hlen(e);
+
+ if (max_encap_size < ipv4_encap_size) {
+ mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n",
+ ipv4_encap_size, max_encap_size);
+ err = -EOPNOTSUPP;
+ goto release_neigh;
+ }
+
+ encap_header = kzalloc(ipv4_encap_size, GFP_KERNEL);
+ if (!encap_header) {
+ err = -ENOMEM;
+ goto release_neigh;
+ }
+
+ e->route_dev_ifindex = attr.route_dev->ifindex;
+
+ read_lock_bh(&attr.n->lock);
+ nud_state = attr.n->nud_state;
+ ether_addr_copy(e->h_dest, attr.n->ha);
+ WRITE_ONCE(e->nhe->neigh_dev, attr.n->dev);
+ read_unlock_bh(&attr.n->lock);
+
+ /* add ethernet header */
+ ip = (struct iphdr *)gen_eth_tnl_hdr(encap_header, attr.route_dev, e,
+ ETH_P_IP);
+
+ /* add ip header */
+ ip->tos = tun_key->tos;
+ ip->version = 0x4;
+ ip->ihl = 0x5;
+ ip->ttl = attr.ttl;
+ ip->daddr = attr.fl.fl4.daddr;
+ ip->saddr = attr.fl.fl4.saddr;
+
+ /* add tunneling protocol header */
+ err = mlx5e_gen_ip_tunnel_header((char *)ip + sizeof(struct iphdr),
+ &ip->protocol, e);
+ if (err)
+ goto free_encap;
+
+ e->encap_size = ipv4_encap_size;
+ kfree(e->encap_header);
+ e->encap_header = encap_header;
+
+ if (!(nud_state & NUD_VALID)) {
+ neigh_event_send(attr.n, NULL);
+ /* the encap entry will be made valid on neigh update event
+ * and not used before that.
+ */
+ goto release_neigh;
+ }
+ e->pkt_reformat = mlx5_packet_reformat_alloc(priv->mdev,
+ e->reformat_type,
+ ipv4_encap_size, encap_header,
+ MLX5_FLOW_NAMESPACE_FDB);
+ if (IS_ERR(e->pkt_reformat)) {
+ err = PTR_ERR(e->pkt_reformat);
+ goto free_encap;
+ }
+
+ e->flags |= MLX5_ENCAP_ENTRY_VALID;
+ mlx5e_rep_queue_neigh_stats_work(netdev_priv(attr.out_dev));
+ mlx5e_route_lookup_ipv4_put(&attr);
+ return err;
+
+free_encap:
+ kfree(encap_header);
+release_neigh:
+ mlx5e_route_lookup_ipv4_put(&attr);
+ return err;
+}
+
#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
static int mlx5e_route_lookup_ipv6_get(struct mlx5e_priv *priv,
struct net_device *mirred_dev,
@@ -476,6 +575,105 @@ int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
mlx5e_route_lookup_ipv6_put(&attr);
return err;
}
+
+int mlx5e_tc_tun_update_header_ipv6(struct mlx5e_priv *priv,
+ struct net_device *mirred_dev,
+ struct mlx5e_encap_entry *e)
+{
+ int max_encap_size = MLX5_CAP_ESW(priv->mdev, max_encap_header_size);
+ const struct ip_tunnel_key *tun_key = &e->tun_info->key;
+ TC_TUN_ROUTE_ATTR_INIT(attr);
+ struct ipv6hdr *ip6h;
+ int ipv6_encap_size;
+ char *encap_header;
+ u8 nud_state;
+ int err;
+
+ attr.ttl = tun_key->ttl;
+
+ attr.fl.fl6.flowlabel = ip6_make_flowinfo(RT_TOS(tun_key->tos), tun_key->label);
+ attr.fl.fl6.daddr = tun_key->u.ipv6.dst;
+ attr.fl.fl6.saddr = tun_key->u.ipv6.src;
+
+ err = mlx5e_route_lookup_ipv6_get(priv, mirred_dev, &attr);
+ if (err)
+ return err;
+
+ ipv6_encap_size =
+ (is_vlan_dev(attr.route_dev) ? VLAN_ETH_HLEN : ETH_HLEN) +
+ sizeof(struct ipv6hdr) +
+ e->tunnel->calc_hlen(e);
+
+ if (max_encap_size < ipv6_encap_size) {
+ mlx5_core_warn(priv->mdev, "encap size %d too big, max supported is %d\n",
+ ipv6_encap_size, max_encap_size);
+ err = -EOPNOTSUPP;
+ goto release_neigh;
+ }
+
+ encap_header = kzalloc(ipv6_encap_size, GFP_KERNEL);
+ if (!encap_header) {
+ err = -ENOMEM;
+ goto release_neigh;
+ }
+
+ e->route_dev_ifindex = attr.route_dev->ifindex;
+
+ read_lock_bh(&attr.n->lock);
+ nud_state = attr.n->nud_state;
+ ether_addr_copy(e->h_dest, attr.n->ha);
+ WRITE_ONCE(e->nhe->neigh_dev, attr.n->dev);
+ read_unlock_bh(&attr.n->lock);
+
+ /* add ethernet header */
+ ip6h = (struct ipv6hdr *)gen_eth_tnl_hdr(encap_header, attr.route_dev, e,
+ ETH_P_IPV6);
+
+ /* add ip header */
+ ip6_flow_hdr(ip6h, tun_key->tos, 0);
+ /* the HW fills up ipv6 payload len */
+ ip6h->hop_limit = attr.ttl;
+ ip6h->daddr = attr.fl.fl6.daddr;
+ ip6h->saddr = attr.fl.fl6.saddr;
+
+ /* add tunneling protocol header */
+ err = mlx5e_gen_ip_tunnel_header((char *)ip6h + sizeof(struct ipv6hdr),
+ &ip6h->nexthdr, e);
+ if (err)
+ goto free_encap;
+
+ e->encap_size = ipv6_encap_size;
+ kfree(e->encap_header);
+ e->encap_header = encap_header;
+
+ if (!(nud_state & NUD_VALID)) {
+ neigh_event_send(attr.n, NULL);
+ /* the encap entry will be made valid on neigh update event
+ * and not used before that.
+ */
+ goto release_neigh;
+ }
+
+ e->pkt_reformat = mlx5_packet_reformat_alloc(priv->mdev,
+ e->reformat_type,
+ ipv6_encap_size, encap_header,
+ MLX5_FLOW_NAMESPACE_FDB);
+ if (IS_ERR(e->pkt_reformat)) {
+ err = PTR_ERR(e->pkt_reformat);
+ goto free_encap;
+ }
+
+ e->flags |= MLX5_ENCAP_ENTRY_VALID;
+ mlx5e_rep_queue_neigh_stats_work(netdev_priv(attr.out_dev));
+ mlx5e_route_lookup_ipv6_put(&attr);
+ return err;
+
+free_encap:
+ kfree(encap_header);
+release_neigh:
+ mlx5e_route_lookup_ipv6_put(&attr);
+ return err;
+}
#endif
int mlx5e_tc_tun_route_lookup(struct mlx5e_priv *priv,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h
index 9d6ee9405eaf..fa992e869044 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.h
@@ -59,16 +59,26 @@ int mlx5e_tc_tun_init_encap_attr(struct net_device *tunnel_dev,
int mlx5e_tc_tun_create_header_ipv4(struct mlx5e_priv *priv,
struct net_device *mirred_dev,
struct mlx5e_encap_entry *e);
+int mlx5e_tc_tun_update_header_ipv4(struct mlx5e_priv *priv,
+ struct net_device *mirred_dev,
+ struct mlx5e_encap_entry *e);
#if IS_ENABLED(CONFIG_INET) && IS_ENABLED(CONFIG_IPV6)
int mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
struct net_device *mirred_dev,
struct mlx5e_encap_entry *e);
+int mlx5e_tc_tun_update_header_ipv6(struct mlx5e_priv *priv,
+ struct net_device *mirred_dev,
+ struct mlx5e_encap_entry *e);
#else
static inline int
mlx5e_tc_tun_create_header_ipv6(struct mlx5e_priv *priv,
struct net_device *mirred_dev,
struct mlx5e_encap_entry *e) { return -EOPNOTSUPP; }
+int mlx5e_tc_tun_update_header_ipv6(struct mlx5e_priv *priv,
+ struct net_device *mirred_dev,
+ struct mlx5e_encap_entry *e)
+{ return -EOPNOTSUPP; }
#endif
int mlx5e_tc_tun_route_lookup(struct mlx5e_priv *priv,
struct mlx5_flow_spec *spec,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 2cc31f99db9b..e6150c7597f0 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -170,11 +170,11 @@ mlx5e_tc_match_to_reg_get_match(struct mlx5_flow_spec *spec,
}
int
-mlx5e_tc_match_to_reg_set(struct mlx5_core_dev *mdev,
- struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
- enum mlx5_flow_namespace_type ns,
- enum mlx5e_tc_attr_to_reg type,
- u32 data)
+mlx5e_tc_match_to_reg_set_and_get_id(struct mlx5_core_dev *mdev,
+ struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
+ enum mlx5_flow_namespace_type ns,
+ enum mlx5e_tc_attr_to_reg type,
+ u32 data)
{
int moffset = mlx5e_tc_attr_to_reg_mappings[type].moffset;
int mfield = mlx5e_tc_attr_to_reg_mappings[type].mfield;
@@ -198,9 +198,10 @@ mlx5e_tc_match_to_reg_set(struct mlx5_core_dev *mdev,
MLX5_SET(set_action_in, modact, offset, moffset * 8);
MLX5_SET(set_action_in, modact, length, mlen * 8);
MLX5_SET(set_action_in, modact, data, data);
+ err = mod_hdr_acts->num_actions;
mod_hdr_acts->num_actions++;
- return 0;
+ return err;
}
static struct mlx5_tc_ct_priv *
@@ -249,6 +250,41 @@ mlx5_tc_rule_delete(struct mlx5e_priv *priv,
mlx5e_del_offloaded_nic_rule(priv, rule, attr);
}
+int
+mlx5e_tc_match_to_reg_set(struct mlx5_core_dev *mdev,
+ struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
+ enum mlx5_flow_namespace_type ns,
+ enum mlx5e_tc_attr_to_reg type,
+ u32 data)
+{
+ int ret = mlx5e_tc_match_to_reg_set_and_get_id(mdev, mod_hdr_acts, ns, type, data);
+
+ return ret < 0 ? ret : 0;
+}
+
+void mlx5e_tc_match_to_reg_mod_hdr_change(struct mlx5_core_dev *mdev,
+ struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
+ enum mlx5e_tc_attr_to_reg type,
+ int act_id, u32 data)
+{
+ int moffset = mlx5e_tc_attr_to_reg_mappings[type].moffset;
+ int mfield = mlx5e_tc_attr_to_reg_mappings[type].mfield;
+ int mlen = mlx5e_tc_attr_to_reg_mappings[type].mlen;
+ char *modact;
+
+ modact = mod_hdr_acts->actions + (act_id * MLX5_MH_ACT_SZ);
+
+ /* Firmware has 5bit length field and 0 means 32bits */
+ if (mlen == 4)
+ mlen = 0;
+
+ MLX5_SET(set_action_in, modact, action_type, MLX5_ACTION_TYPE_SET);
+ MLX5_SET(set_action_in, modact, field, mfield);
+ MLX5_SET(set_action_in, modact, offset, moffset * 8);
+ MLX5_SET(set_action_in, modact, length, mlen * 8);
+ MLX5_SET(set_action_in, modact, data, data);
+}
+
struct mlx5e_hairpin {
struct mlx5_hairpin *pair;
@@ -1214,6 +1250,26 @@ int mlx5e_tc_query_route_vport(struct net_device *out_dev, struct net_device *ro
return err;
}
+int mlx5e_tc_add_flow_mod_hdr(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow_parse_attr *parse_attr,
+ struct mlx5e_tc_flow *flow)
+{
+ struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts = &parse_attr->mod_hdr_acts;
+ struct mlx5_modify_hdr *mod_hdr;
+
+ mod_hdr = mlx5_modify_header_alloc(priv->mdev,
+ get_flow_name_space(flow),
+ mod_hdr_acts->num_actions,
+ mod_hdr_acts->actions);
+ if (IS_ERR(mod_hdr))
+ return PTR_ERR(mod_hdr);
+
+ WARN_ON(flow->attr->modify_hdr);
+ flow->attr->modify_hdr = mod_hdr;
+
+ return 0;
+}
+
static int
mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow,
@@ -1293,7 +1349,6 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR &&
!(attr->ct_attr.ct_action & TCA_CT_ACT_CLEAR)) {
err = mlx5e_attach_mod_hdr(priv, flow, parse_attr);
- dealloc_mod_hdr_actions(&parse_attr->mod_hdr_acts);
if (err)
return err;
}
@@ -1376,8 +1431,10 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv,
mlx5_tc_ct_match_del(get_ct_priv(priv), &flow->attr->ct_attr);
- if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)
+ if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) {
+ dealloc_mod_hdr_actions(&attr->parse_attr->mod_hdr_acts);
mlx5e_detach_mod_hdr(priv, flow);
+ }
if (attr->action & MLX5_FLOW_CONTEXT_ACTION_COUNT)
mlx5_fc_destroy(esw_attr->counter_dev, attr->counter);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
index 5434bbb9a217..9042e64a96ca 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -214,6 +214,11 @@ int mlx5e_tc_match_to_reg_set(struct mlx5_core_dev *mdev,
enum mlx5e_tc_attr_to_reg type,
u32 data);
+void mlx5e_tc_match_to_reg_mod_hdr_change(struct mlx5_core_dev *mdev,
+ struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
+ enum mlx5e_tc_attr_to_reg type,
+ int act_id, u32 data);
+
void mlx5e_tc_match_to_reg_match(struct mlx5_flow_spec *spec,
enum mlx5e_tc_attr_to_reg type,
u32 data,
@@ -224,6 +229,16 @@ void mlx5e_tc_match_to_reg_get_match(struct mlx5_flow_spec *spec,
u32 *data,
u32 *mask);
+int mlx5e_tc_match_to_reg_set_and_get_id(struct mlx5_core_dev *mdev,
+ struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
+ enum mlx5_flow_namespace_type ns,
+ enum mlx5e_tc_attr_to_reg type,
+ u32 data);
+
+int mlx5e_tc_add_flow_mod_hdr(struct mlx5e_priv *priv,
+ struct mlx5e_tc_flow_parse_attr *parse_attr,
+ struct mlx5e_tc_flow *flow);
+
int alloc_mod_hdr_actions(struct mlx5_core_dev *mdev,
int namespace,
struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts);
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 16/17] net/mlx5e: Rename some encap-specific API to generic names
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (14 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 15/17] net/mlx5e: TC preparation refactoring for routing update event Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-06 5:02 ` [net-next V2 17/17] net/mlx5e: Handle FIB events to update tunnel endpoint device Saeed Mahameed
` (2 subsequent siblings)
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Some of the encap-specific functions and fields will also be used by route
update infrastructure in following patches. Rename them to generic names.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c | 2 +-
drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h | 2 +-
.../net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c | 10 +++++-----
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 2 +-
drivers/net/ethernet/mellanox/mlx5/core/en_tc.h | 2 +-
5 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
index a7ba1c84371d..065126370acd 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
@@ -124,7 +124,7 @@ void mlx5e_rep_update_flows(struct mlx5e_priv *priv,
}
unlock:
mutex_unlock(&esw->offloads.encap_tbl_lock);
- mlx5e_put_encap_flow_list(priv, &flow_list);
+ mlx5e_put_flow_list(priv, &flow_list);
}
static int
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
index 14db9b5accb1..4f35b486fb44 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
@@ -95,7 +95,7 @@ struct mlx5e_tc_flow {
* due to missing route)
*/
struct net_device *orig_dev; /* netdev adding flow first */
- int tmp_efi_index;
+ int tmp_entry_index;
struct list_head tmp_list; /* temporary flow list used by neigh update */
refcount_t refcnt;
struct rcu_head rcu_head;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
index 577216744a17..bc0e26f3fd4c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
@@ -106,8 +106,8 @@ void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv,
esw_attr = attr->esw_attr;
spec = &attr->parse_attr->spec;
- esw_attr->dests[flow->tmp_efi_index].pkt_reformat = e->pkt_reformat;
- esw_attr->dests[flow->tmp_efi_index].flags |= MLX5_ESW_DEST_ENCAP_VALID;
+ esw_attr->dests[flow->tmp_entry_index].pkt_reformat = e->pkt_reformat;
+ esw_attr->dests[flow->tmp_entry_index].flags |= MLX5_ESW_DEST_ENCAP_VALID;
/* Flow can be associated with multiple encap entries.
* Before offloading the flow verify that all of them have
* a valid neighbour.
@@ -161,7 +161,7 @@ void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv,
/* update from encap rule to slow path rule */
rule = mlx5e_tc_offload_to_slow_path(esw, flow, spec);
/* mark the flow's encap dest as non-valid */
- esw_attr->dests[flow->tmp_efi_index].flags &= ~MLX5_ESW_DEST_ENCAP_VALID;
+ esw_attr->dests[flow->tmp_entry_index].flags &= ~MLX5_ESW_DEST_ENCAP_VALID;
if (IS_ERR(rule)) {
err = PTR_ERR(rule);
@@ -195,7 +195,7 @@ void mlx5e_take_all_encap_flows(struct mlx5e_encap_entry *e, struct list_head *f
continue;
wait_for_completion(&flow->init_done);
- flow->tmp_efi_index = efi->index;
+ flow->tmp_entry_index = efi->index;
list_add(&flow->tmp_list, flow_list);
}
}
@@ -294,7 +294,7 @@ void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe)
}
mutex_unlock(&esw->offloads.encap_tbl_lock);
- mlx5e_put_encap_flow_list(priv, &flow_list);
+ mlx5e_put_flow_list(priv, &flow_list);
if (neigh_used) {
/* release current encap before breaking the loop */
mlx5e_encap_put(priv, e);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index e6150c7597f0..c5ecb9e4e767 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -1451,7 +1451,7 @@ struct mlx5_fc *mlx5e_tc_get_counter(struct mlx5e_tc_flow *flow)
}
/* Iterate over tmp_list of flows attached to flow_list head. */
-void mlx5e_put_encap_flow_list(struct mlx5e_priv *priv, struct list_head *flow_list)
+void mlx5e_put_flow_list(struct mlx5e_priv *priv, struct list_head *flow_list)
{
struct mlx5e_tc_flow *flow, *tmp;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
index 9042e64a96ca..89003ae7775a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -174,7 +174,7 @@ bool mlx5e_encap_take(struct mlx5e_encap_entry *e);
void mlx5e_encap_put(struct mlx5e_priv *priv, struct mlx5e_encap_entry *e);
void mlx5e_take_all_encap_flows(struct mlx5e_encap_entry *e, struct list_head *flow_list);
-void mlx5e_put_encap_flow_list(struct mlx5e_priv *priv, struct list_head *flow_list);
+void mlx5e_put_flow_list(struct mlx5e_priv *priv, struct list_head *flow_list);
struct mlx5e_neigh_hash_entry;
void mlx5e_tc_update_neigh_used_value(struct mlx5e_neigh_hash_entry *nhe);
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* [net-next V2 17/17] net/mlx5e: Handle FIB events to update tunnel endpoint device
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (15 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 16/17] net/mlx5e: Rename some encap-specific API to generic names Saeed Mahameed
@ 2021-02-06 5:02 ` Saeed Mahameed
2021-02-08 21:55 ` [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Or Gerlitz
2021-02-09 8:42 ` Or Gerlitz
18 siblings, 0 replies; 39+ messages in thread
From: Saeed Mahameed @ 2021-02-06 5:02 UTC (permalink / raw)
To: Jakub Kicinski
Cc: David S. Miller, netdev, Vlad Buslov, Dmytro Linkin, Roi Dayan,
Saeed Mahameed
From: Vlad Buslov <vladbu@nvidia.com>
Process FIB route update events to dynamically update the stack device
rules when tunnel routing changes. Use rtnl lock to prevent FIB event
handler from running concurrently with neigh update and neigh stats
workqueue tasks. Use encap_tbl_lock mutex to synchronize with TC rule
update path that doesn't use rtnl lock.
FIB event workflow for encap flows:
- Unoffload all flows attached to route encaps from slow or fast path
depending on encap destination endpoint neigh state.
- Update encap IP header according to new route dev.
- Update flows mod_hdr action that is responsible for overwriting reg_c0
source port bits to source port of new underlying VF of new route dev. This
step requires changing flow create/delete code to save flow parse attribute
mod_hdr_acts structure for whole flow lifetime instead of deallocating it
after flow creation. Refactor mod_hdr code to allow saving id of individual
mod_hdr actions and updating them with dedicated helper.
- Offload all flows to either slow or fast path depending on encap
destination endpoint neigh state.
FIB event workflow for decap flows:
- Unoffload all route flows from hardware. When last route flow is deleted
all indirect table rules for the route dev will also be deleted.
- Update flow attr decap_vport and destination MAC according to underlying
VF of new rote dev.
- Offload all route flows back to hardware creating new indirect table
rules according to updated flow attribute data.
Extract some neigh update code to helper functions to be used by both neigh
update and route update infrastructure.
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Dmytro Linkin <dlinkin@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../ethernet/mellanox/mlx5/core/en/tc_priv.h | 1 +
.../mellanox/mlx5/core/en/tc_tun_encap.c | 751 +++++++++++++++++-
.../mellanox/mlx5/core/en/tc_tun_encap.h | 3 +
.../net/ethernet/mellanox/mlx5/core/en_rep.h | 6 +
.../net/ethernet/mellanox/mlx5/core/en_tc.c | 76 +-
.../net/ethernet/mellanox/mlx5/core/eswitch.c | 1 -
.../net/ethernet/mellanox/mlx5/core/eswitch.h | 2 +-
7 files changed, 773 insertions(+), 67 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
index 4f35b486fb44..c223591ffc22 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
@@ -26,6 +26,7 @@ enum {
MLX5E_TC_FLOW_FLAG_CT = MLX5E_TC_FLOW_BASE + 7,
MLX5E_TC_FLOW_FLAG_L3_TO_L2_DECAP = MLX5E_TC_FLOW_BASE + 8,
MLX5E_TC_FLOW_FLAG_TUN_RX = MLX5E_TC_FLOW_BASE + 9,
+ MLX5E_TC_FLOW_FLAG_FAILED = MLX5E_TC_FLOW_BASE + 10,
};
struct mlx5e_tc_flow_parse_attr {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
index bc0e26f3fd4c..6a116335bb21 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
@@ -1,12 +1,17 @@
// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
/* Copyright (c) 2021 Mellanox Technologies. */
+#include <net/fib_notifier.h>
#include "tc_tun_encap.h"
#include "en_tc.h"
#include "tc_tun.h"
#include "rep/tc.h"
#include "diag/en_tc_tracepoint.h"
+enum {
+ MLX5E_ROUTE_ENTRY_VALID = BIT(0),
+};
+
struct mlx5e_route_key {
int ip_version;
union {
@@ -19,11 +24,26 @@ struct mlx5e_route_entry {
struct mlx5e_route_key key;
struct list_head encap_entries;
struct list_head decap_flows;
+ u32 flags;
struct hlist_node hlist;
refcount_t refcnt;
+ int tunnel_dev_index;
struct rcu_head rcu;
};
+struct mlx5e_tc_tun_encap {
+ struct mlx5e_priv *priv;
+ struct notifier_block fib_nb;
+ spinlock_t route_lock; /* protects route_tbl */
+ unsigned long route_tbl_last_update;
+ DECLARE_HASHTABLE(route_tbl, 8);
+};
+
+static bool mlx5e_route_entry_valid(struct mlx5e_route_entry *r)
+{
+ return r->flags & MLX5E_ROUTE_ENTRY_VALID;
+}
+
int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
struct mlx5_flow_spec *spec)
{
@@ -72,6 +92,27 @@ int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
return 0;
}
+static bool mlx5e_tc_flow_all_encaps_valid(struct mlx5_esw_flow_attr *esw_attr)
+{
+ bool all_flow_encaps_valid = true;
+ int i;
+
+ /* Flow can be associated with multiple encap entries.
+ * Before offloading the flow verify that all of them have
+ * a valid neighbour.
+ */
+ for (i = 0; i < MLX5_MAX_FLOW_FWD_VPORTS; i++) {
+ if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP))
+ continue;
+ if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP_VALID)) {
+ all_flow_encaps_valid = false;
+ break;
+ }
+ }
+
+ return all_flow_encaps_valid;
+}
+
void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv,
struct mlx5e_encap_entry *e,
struct list_head *flow_list)
@@ -84,6 +125,9 @@ void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow;
int err;
+ if (e->flags & MLX5_ENCAP_ENTRY_NO_ROUTE)
+ return;
+
e->pkt_reformat = mlx5_packet_reformat_alloc(priv->mdev,
e->reformat_type,
e->encap_size, e->encap_header,
@@ -97,9 +141,6 @@ void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv,
mlx5e_rep_queue_neigh_stats_work(priv);
list_for_each_entry(flow, flow_list, tmp_list) {
- bool all_flow_encaps_valid = true;
- int i;
-
if (!mlx5e_is_offloaded_flow(flow))
continue;
attr = flow->attr;
@@ -108,20 +149,9 @@ void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv,
esw_attr->dests[flow->tmp_entry_index].pkt_reformat = e->pkt_reformat;
esw_attr->dests[flow->tmp_entry_index].flags |= MLX5_ESW_DEST_ENCAP_VALID;
- /* Flow can be associated with multiple encap entries.
- * Before offloading the flow verify that all of them have
- * a valid neighbour.
- */
- for (i = 0; i < MLX5_MAX_FLOW_FWD_VPORTS; i++) {
- if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP))
- continue;
- if (!(esw_attr->dests[i].flags & MLX5_ESW_DEST_ENCAP_VALID)) {
- all_flow_encaps_valid = false;
- break;
- }
- }
+
/* Do not offload flows with unresolved neighbors */
- if (!all_flow_encaps_valid)
+ if (!mlx5e_tc_flow_all_encaps_valid(esw_attr))
continue;
/* update from slow path rule to encap rule */
rule = mlx5e_tc_offload_fdb_rules(esw, flow, spec, attr);
@@ -181,6 +211,18 @@ void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv,
mlx5_packet_reformat_dealloc(priv->mdev, e->pkt_reformat);
}
+static void mlx5e_take_tmp_flow(struct mlx5e_tc_flow *flow,
+ struct list_head *flow_list,
+ int index)
+{
+ if (IS_ERR(mlx5e_flow_get(flow)))
+ return;
+ wait_for_completion(&flow->init_done);
+
+ flow->tmp_entry_index = index;
+ list_add(&flow->tmp_list, flow_list);
+}
+
/* Takes reference to all flows attached to encap and adds the flows to
* flow_list using 'tmp_list' list_head in mlx5e_tc_flow.
*/
@@ -191,15 +233,22 @@ void mlx5e_take_all_encap_flows(struct mlx5e_encap_entry *e, struct list_head *f
list_for_each_entry(efi, &e->flows, list) {
flow = container_of(efi, struct mlx5e_tc_flow, encaps[efi->index]);
- if (IS_ERR(mlx5e_flow_get(flow)))
- continue;
- wait_for_completion(&flow->init_done);
-
- flow->tmp_entry_index = efi->index;
- list_add(&flow->tmp_list, flow_list);
+ mlx5e_take_tmp_flow(flow, flow_list, efi->index);
}
}
+/* Takes reference to all flows attached to route and adds the flows to
+ * flow_list using 'tmp_list' list_head in mlx5e_tc_flow.
+ */
+static void mlx5e_take_all_route_decap_flows(struct mlx5e_route_entry *r,
+ struct list_head *flow_list)
+{
+ struct mlx5e_tc_flow *flow;
+
+ list_for_each_entry(flow, &r->decap_flows, decap_routes)
+ mlx5e_take_tmp_flow(flow, flow_list, 0);
+}
+
static struct mlx5e_encap_entry *
mlx5e_get_next_valid_encap(struct mlx5e_neigh_hash_entry *nhe,
struct mlx5e_encap_entry *e)
@@ -557,21 +606,78 @@ static int mlx5e_set_vf_tunnel(struct mlx5_eswitch *esw,
esw_attr->dests[out_index].flags |= MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE;
data = mlx5_eswitch_get_vport_metadata_for_set(esw_attr->in_mdev->priv.eswitch,
vport_num);
- err = mlx5e_tc_match_to_reg_set(esw->dev, mod_hdr_acts,
- MLX5_FLOW_NAMESPACE_FDB, VPORT_TO_REG, data);
+ err = mlx5e_tc_match_to_reg_set_and_get_id(esw->dev, mod_hdr_acts,
+ MLX5_FLOW_NAMESPACE_FDB,
+ VPORT_TO_REG, data);
+ if (err >= 0) {
+ esw_attr->dests[out_index].src_port_rewrite_act_id = err;
+ err = 0;
+ }
+
+out:
+ if (route_dev)
+ dev_put(route_dev);
+ return err;
+}
+
+static int mlx5e_update_vf_tunnel(struct mlx5_eswitch *esw,
+ struct mlx5_esw_flow_attr *attr,
+ struct mlx5e_tc_mod_hdr_acts *mod_hdr_acts,
+ struct net_device *out_dev,
+ int route_dev_ifindex,
+ int out_index)
+{
+ int act_id = attr->dests[out_index].src_port_rewrite_act_id;
+ struct net_device *route_dev;
+ u16 vport_num;
+ int err = 0;
+ u32 data;
+
+ route_dev = dev_get_by_index(dev_net(out_dev), route_dev_ifindex);
+
+ if (!route_dev || route_dev->netdev_ops != &mlx5e_netdev_ops ||
+ !mlx5e_tc_is_vf_tunnel(out_dev, route_dev)) {
+ err = -ENODEV;
+ goto out;
+ }
+
+ err = mlx5e_tc_query_route_vport(out_dev, route_dev, &vport_num);
if (err)
goto out;
+ data = mlx5_eswitch_get_vport_metadata_for_set(attr->in_mdev->priv.eswitch,
+ vport_num);
+ mlx5e_tc_match_to_reg_mod_hdr_change(esw->dev, mod_hdr_acts, VPORT_TO_REG, act_id, data);
+
out:
if (route_dev)
dev_put(route_dev);
return err;
}
+static unsigned int mlx5e_route_tbl_get_last_update(struct mlx5e_priv *priv)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5_rep_uplink_priv *uplink_priv;
+ struct mlx5e_rep_priv *uplink_rpriv;
+ struct mlx5e_tc_tun_encap *encap;
+ unsigned int ret;
+
+ uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
+ uplink_priv = &uplink_rpriv->uplink_priv;
+ encap = uplink_priv->encap;
+
+ spin_lock_bh(&encap->route_lock);
+ ret = encap->route_tbl_last_update;
+ spin_unlock_bh(&encap->route_lock);
+ return ret;
+}
+
static int mlx5e_attach_encap_route(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow,
struct mlx5e_encap_entry *e,
bool new_encap_entry,
+ unsigned long tbl_time_before,
int out_index);
int mlx5e_attach_encap(struct mlx5e_priv *priv,
@@ -586,6 +692,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
struct mlx5e_tc_flow_parse_attr *parse_attr;
struct mlx5_flow_attr *attr = flow->attr;
const struct ip_tunnel_info *tun_info;
+ unsigned long tbl_time_before = 0;
struct encap_key key;
struct mlx5e_encap_entry *e;
bool entry_created = false;
@@ -651,6 +758,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
INIT_LIST_HEAD(&e->flows);
hash_add_rcu(esw->offloads.encap_tbl, &e->encap_hlist, hash_key);
+ tbl_time_before = mlx5e_route_tbl_get_last_update(priv);
mutex_unlock(&esw->offloads.encap_tbl_lock);
if (family == AF_INET)
@@ -668,7 +776,8 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
e->compl_result = 1;
attach_flow:
- err = mlx5e_attach_encap_route(priv, flow, e, entry_created, out_index);
+ err = mlx5e_attach_encap_route(priv, flow, e, entry_created, tbl_time_before,
+ out_index);
if (err)
goto out_err;
@@ -797,15 +906,48 @@ static u32 hash_route_info(struct mlx5e_route_key *key)
return jhash(&key->endpoint_ip.v6, sizeof(key->endpoint_ip.v6), 0);
}
+static void mlx5e_route_dealloc(struct mlx5e_priv *priv,
+ struct mlx5e_route_entry *r)
+{
+ WARN_ON(!list_empty(&r->decap_flows));
+ WARN_ON(!list_empty(&r->encap_entries));
+
+ kfree_rcu(r, rcu);
+}
+
+static void mlx5e_route_put(struct mlx5e_priv *priv, struct mlx5e_route_entry *r)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+
+ if (!refcount_dec_and_mutex_lock(&r->refcnt, &esw->offloads.encap_tbl_lock))
+ return;
+
+ hash_del_rcu(&r->hlist);
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+
+ mlx5e_route_dealloc(priv, r);
+}
+
+static void mlx5e_route_put_locked(struct mlx5e_priv *priv, struct mlx5e_route_entry *r)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+
+ lockdep_assert_held(&esw->offloads.encap_tbl_lock);
+
+ if (!refcount_dec_and_test(&r->refcnt))
+ return;
+ hash_del_rcu(&r->hlist);
+ mlx5e_route_dealloc(priv, r);
+}
+
static struct mlx5e_route_entry *
-mlx5e_route_get(struct mlx5e_priv *priv, struct mlx5e_route_key *key,
+mlx5e_route_get(struct mlx5e_tc_tun_encap *encap, struct mlx5e_route_key *key,
u32 hash_key)
{
- struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
struct mlx5e_route_key r_key;
struct mlx5e_route_entry *r;
- hash_for_each_possible(esw->offloads.route_tbl, r, hlist, hash_key) {
+ hash_for_each_possible(encap->route_tbl, r, hlist, hash_key) {
r_key = r->key;
if (!cmp_route_info(&r_key, key) &&
refcount_inc_not_zero(&r->refcnt))
@@ -816,33 +958,120 @@ mlx5e_route_get(struct mlx5e_priv *priv, struct mlx5e_route_key *key,
static struct mlx5e_route_entry *
mlx5e_route_get_create(struct mlx5e_priv *priv,
- struct mlx5e_route_key *key)
+ struct mlx5e_route_key *key,
+ int tunnel_dev_index,
+ unsigned long *route_tbl_change_time)
{
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5_rep_uplink_priv *uplink_priv;
+ struct mlx5e_rep_priv *uplink_rpriv;
+ struct mlx5e_tc_tun_encap *encap;
struct mlx5e_route_entry *r;
u32 hash_key;
+ uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
+ uplink_priv = &uplink_rpriv->uplink_priv;
+ encap = uplink_priv->encap;
+
hash_key = hash_route_info(key);
- r = mlx5e_route_get(priv, key, hash_key);
- if (r)
+ spin_lock_bh(&encap->route_lock);
+ r = mlx5e_route_get(encap, key, hash_key);
+ spin_unlock_bh(&encap->route_lock);
+ if (r) {
+ if (!mlx5e_route_entry_valid(r)) {
+ mlx5e_route_put_locked(priv, r);
+ return ERR_PTR(-EINVAL);
+ }
return r;
+ }
r = kzalloc(sizeof(*r), GFP_KERNEL);
if (!r)
return ERR_PTR(-ENOMEM);
r->key = *key;
+ r->flags |= MLX5E_ROUTE_ENTRY_VALID;
+ r->tunnel_dev_index = tunnel_dev_index;
refcount_set(&r->refcnt, 1);
INIT_LIST_HEAD(&r->decap_flows);
INIT_LIST_HEAD(&r->encap_entries);
- hash_add(esw->offloads.route_tbl, &r->hlist, hash_key);
+
+ spin_lock_bh(&encap->route_lock);
+ *route_tbl_change_time = encap->route_tbl_last_update;
+ hash_add(encap->route_tbl, &r->hlist, hash_key);
+ spin_unlock_bh(&encap->route_lock);
+
return r;
}
+static struct mlx5e_route_entry *
+mlx5e_route_lookup_for_update(struct mlx5e_tc_tun_encap *encap, struct mlx5e_route_key *key)
+{
+ u32 hash_key = hash_route_info(key);
+ struct mlx5e_route_entry *r;
+
+ spin_lock_bh(&encap->route_lock);
+ encap->route_tbl_last_update = jiffies;
+ r = mlx5e_route_get(encap, key, hash_key);
+ spin_unlock_bh(&encap->route_lock);
+
+ return r;
+}
+
+struct mlx5e_tc_fib_event_data {
+ struct work_struct work;
+ unsigned long event;
+ struct mlx5e_route_entry *r;
+ struct net_device *ul_dev;
+};
+
+static void mlx5e_tc_fib_event_work(struct work_struct *work);
+static struct mlx5e_tc_fib_event_data *
+mlx5e_tc_init_fib_work(unsigned long event, struct net_device *ul_dev, gfp_t flags)
+{
+ struct mlx5e_tc_fib_event_data *fib_work;
+
+ fib_work = kzalloc(sizeof(*fib_work), flags);
+ if (WARN_ON(!fib_work))
+ return NULL;
+
+ INIT_WORK(&fib_work->work, mlx5e_tc_fib_event_work);
+ fib_work->event = event;
+ fib_work->ul_dev = ul_dev;
+
+ return fib_work;
+}
+
+static int
+mlx5e_route_enqueue_update(struct mlx5e_priv *priv,
+ struct mlx5e_route_entry *r,
+ unsigned long event)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_tc_fib_event_data *fib_work;
+ struct mlx5e_rep_priv *uplink_rpriv;
+ struct net_device *ul_dev;
+
+ uplink_rpriv = mlx5_eswitch_get_uplink_priv(esw, REP_ETH);
+ ul_dev = uplink_rpriv->netdev;
+
+ fib_work = mlx5e_tc_init_fib_work(event, ul_dev, GFP_KERNEL);
+ if (!fib_work)
+ return -ENOMEM;
+
+ dev_hold(ul_dev);
+ refcount_inc(&r->refcnt);
+ fib_work->r = r;
+ queue_work(priv->wq, &fib_work->work);
+
+ return 0;
+}
+
int mlx5e_attach_decap_route(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow)
{
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ unsigned long tbl_time_before, tbl_time_after;
struct mlx5e_tc_flow_parse_attr *parse_attr;
struct mlx5_flow_attr *attr = flow->attr;
struct mlx5_esw_flow_attr *esw_attr;
@@ -856,6 +1085,8 @@ int mlx5e_attach_decap_route(struct mlx5e_priv *priv,
if (!esw_attr->rx_tun_attr)
goto out;
+ tbl_time_before = mlx5e_route_tbl_get_last_update(priv);
+ tbl_time_after = tbl_time_before;
err = mlx5e_tc_tun_route_lookup(priv, &parse_attr->spec, attr);
if (err || !esw_attr->rx_tun_attr->decap_vport)
goto out;
@@ -866,11 +1097,22 @@ int mlx5e_attach_decap_route(struct mlx5e_priv *priv,
else
key.endpoint_ip.v6 = esw_attr->rx_tun_attr->dst_ip.v6;
- r = mlx5e_route_get_create(priv, &key);
+ r = mlx5e_route_get_create(priv, &key, parse_attr->filter_dev->ifindex,
+ &tbl_time_after);
if (IS_ERR(r)) {
err = PTR_ERR(r);
goto out;
}
+ /* Routing changed concurrently. FIB event handler might have missed new
+ * entry, schedule update.
+ */
+ if (tbl_time_before != tbl_time_after) {
+ err = mlx5e_route_enqueue_update(priv, r, FIB_EVENT_ENTRY_REPLACE);
+ if (err) {
+ mlx5e_route_put_locked(priv, r);
+ goto out;
+ }
+ }
flow->decap_route = r;
list_add(&flow->decap_routes, &r->decap_flows);
@@ -886,9 +1128,11 @@ static int mlx5e_attach_encap_route(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow,
struct mlx5e_encap_entry *e,
bool new_encap_entry,
+ unsigned long tbl_time_before,
int out_index)
{
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ unsigned long tbl_time_after = tbl_time_before;
struct mlx5e_tc_flow_parse_attr *parse_attr;
struct mlx5_flow_attr *attr = flow->attr;
const struct ip_tunnel_info *tun_info;
@@ -917,9 +1161,20 @@ static int mlx5e_attach_encap_route(struct mlx5e_priv *priv,
MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE))
return err;
- r = mlx5e_route_get_create(priv, &key);
+ r = mlx5e_route_get_create(priv, &key, parse_attr->mirred_ifindex[out_index],
+ &tbl_time_after);
if (IS_ERR(r))
return PTR_ERR(r);
+ /* Routing changed concurrently. FIB event handler might have missed new
+ * entry, schedule update.
+ */
+ if (tbl_time_before != tbl_time_after) {
+ err = mlx5e_route_enqueue_update(priv, r, FIB_EVENT_ENTRY_REPLACE);
+ if (err) {
+ mlx5e_route_put_locked(priv, r);
+ return err;
+ }
+ }
flow->encap_routes[out_index].r = r;
if (new_encap_entry)
@@ -928,15 +1183,6 @@ static int mlx5e_attach_encap_route(struct mlx5e_priv *priv,
return 0;
}
-static void mlx5e_route_dealloc(struct mlx5e_priv *priv,
- struct mlx5e_route_entry *r)
-{
- WARN_ON(!list_empty(&r->decap_flows));
- WARN_ON(!list_empty(&r->encap_entries));
-
- kfree_rcu(r, rcu);
-}
-
void mlx5e_detach_decap_route(struct mlx5e_priv *priv,
struct mlx5e_tc_flow *flow)
{
@@ -986,3 +1232,422 @@ static void mlx5e_detach_encap_route(struct mlx5e_priv *priv,
mlx5e_route_dealloc(priv, r);
}
+static void mlx5e_invalidate_encap(struct mlx5e_priv *priv,
+ struct mlx5e_encap_entry *e,
+ struct list_head *encap_flows)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_tc_flow *flow;
+
+ list_for_each_entry(flow, encap_flows, tmp_list) {
+ struct mlx5_flow_attr *attr = flow->attr;
+ struct mlx5_esw_flow_attr *esw_attr;
+
+ if (!mlx5e_is_offloaded_flow(flow))
+ continue;
+ esw_attr = attr->esw_attr;
+
+ if (flow_flag_test(flow, SLOW))
+ mlx5e_tc_unoffload_from_slow_path(esw, flow);
+ else
+ mlx5e_tc_unoffload_fdb_rules(esw, flow, flow->attr);
+ mlx5_modify_header_dealloc(priv->mdev, attr->modify_hdr);
+ attr->modify_hdr = NULL;
+
+ esw_attr->dests[flow->tmp_entry_index].flags &=
+ ~MLX5_ESW_DEST_ENCAP_VALID;
+ esw_attr->dests[flow->tmp_entry_index].pkt_reformat = NULL;
+ }
+
+ e->flags |= MLX5_ENCAP_ENTRY_NO_ROUTE;
+ if (e->flags & MLX5_ENCAP_ENTRY_VALID) {
+ e->flags &= ~MLX5_ENCAP_ENTRY_VALID;
+ mlx5_packet_reformat_dealloc(priv->mdev, e->pkt_reformat);
+ e->pkt_reformat = NULL;
+ }
+}
+
+static void mlx5e_reoffload_encap(struct mlx5e_priv *priv,
+ struct net_device *tunnel_dev,
+ struct mlx5e_encap_entry *e,
+ struct list_head *encap_flows)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_tc_flow *flow;
+ int err;
+
+ err = ip_tunnel_info_af(e->tun_info) == AF_INET ?
+ mlx5e_tc_tun_update_header_ipv4(priv, tunnel_dev, e) :
+ mlx5e_tc_tun_update_header_ipv6(priv, tunnel_dev, e);
+ if (err)
+ mlx5_core_warn(priv->mdev, "Failed to update encap header, %d", err);
+ e->flags &= ~MLX5_ENCAP_ENTRY_NO_ROUTE;
+
+ list_for_each_entry(flow, encap_flows, tmp_list) {
+ struct mlx5e_tc_flow_parse_attr *parse_attr;
+ struct mlx5_flow_attr *attr = flow->attr;
+ struct mlx5_esw_flow_attr *esw_attr;
+ struct mlx5_flow_handle *rule;
+ struct mlx5_flow_spec *spec;
+
+ if (flow_flag_test(flow, FAILED))
+ continue;
+
+ esw_attr = attr->esw_attr;
+ parse_attr = attr->parse_attr;
+ spec = &parse_attr->spec;
+
+ err = mlx5e_update_vf_tunnel(esw, esw_attr, &parse_attr->mod_hdr_acts,
+ e->out_dev, e->route_dev_ifindex,
+ flow->tmp_entry_index);
+ if (err) {
+ mlx5_core_warn(priv->mdev, "Failed to update VF tunnel err=%d", err);
+ continue;
+ }
+
+ err = mlx5e_tc_add_flow_mod_hdr(priv, parse_attr, flow);
+ if (err) {
+ mlx5_core_warn(priv->mdev, "Failed to update flow mod_hdr err=%d",
+ err);
+ continue;
+ }
+
+ if (e->flags & MLX5_ENCAP_ENTRY_VALID) {
+ esw_attr->dests[flow->tmp_entry_index].pkt_reformat = e->pkt_reformat;
+ esw_attr->dests[flow->tmp_entry_index].flags |= MLX5_ESW_DEST_ENCAP_VALID;
+ if (!mlx5e_tc_flow_all_encaps_valid(esw_attr))
+ goto offload_to_slow_path;
+ /* update from slow path rule to encap rule */
+ rule = mlx5e_tc_offload_fdb_rules(esw, flow, spec, attr);
+ if (IS_ERR(rule)) {
+ err = PTR_ERR(rule);
+ mlx5_core_warn(priv->mdev, "Failed to update cached encapsulation flow, %d\n",
+ err);
+ } else {
+ flow->rule[0] = rule;
+ }
+ } else {
+offload_to_slow_path:
+ rule = mlx5e_tc_offload_to_slow_path(esw, flow, spec);
+ /* mark the flow's encap dest as non-valid */
+ esw_attr->dests[flow->tmp_entry_index].flags &=
+ ~MLX5_ESW_DEST_ENCAP_VALID;
+
+ if (IS_ERR(rule)) {
+ err = PTR_ERR(rule);
+ mlx5_core_warn(priv->mdev, "Failed to update slow path (encap) flow, %d\n",
+ err);
+ } else {
+ flow->rule[0] = rule;
+ }
+ }
+ flow_flag_set(flow, OFFLOADED);
+ }
+}
+
+static int mlx5e_update_route_encaps(struct mlx5e_priv *priv,
+ struct mlx5e_route_entry *r,
+ struct list_head *flow_list,
+ bool replace)
+{
+ struct net_device *tunnel_dev;
+ struct mlx5e_encap_entry *e;
+
+ tunnel_dev = __dev_get_by_index(dev_net(priv->netdev), r->tunnel_dev_index);
+ if (!tunnel_dev)
+ return -ENODEV;
+
+ list_for_each_entry(e, &r->encap_entries, route_list) {
+ LIST_HEAD(encap_flows);
+
+ mlx5e_take_all_encap_flows(e, &encap_flows);
+ if (list_empty(&encap_flows))
+ continue;
+
+ if (mlx5e_route_entry_valid(r))
+ mlx5e_invalidate_encap(priv, e, &encap_flows);
+
+ if (!replace) {
+ list_splice(&encap_flows, flow_list);
+ continue;
+ }
+
+ mlx5e_reoffload_encap(priv, tunnel_dev, e, &encap_flows);
+ list_splice(&encap_flows, flow_list);
+ }
+
+ return 0;
+}
+
+static void mlx5e_unoffload_flow_list(struct mlx5e_priv *priv,
+ struct list_head *flow_list)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_tc_flow *flow;
+
+ list_for_each_entry(flow, flow_list, tmp_list)
+ if (mlx5e_is_offloaded_flow(flow))
+ mlx5e_tc_unoffload_fdb_rules(esw, flow, flow->attr);
+}
+
+static void mlx5e_reoffload_decap(struct mlx5e_priv *priv,
+ struct list_head *decap_flows)
+{
+ struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
+ struct mlx5e_tc_flow *flow;
+
+ list_for_each_entry(flow, decap_flows, tmp_list) {
+ struct mlx5e_tc_flow_parse_attr *parse_attr;
+ struct mlx5_flow_attr *attr = flow->attr;
+ struct mlx5_flow_handle *rule;
+ struct mlx5_flow_spec *spec;
+ int err;
+
+ if (flow_flag_test(flow, FAILED))
+ continue;
+
+ parse_attr = attr->parse_attr;
+ spec = &parse_attr->spec;
+ err = mlx5e_tc_tun_route_lookup(priv, spec, attr);
+ if (err) {
+ mlx5_core_warn(priv->mdev, "Failed to lookup route for flow, %d\n",
+ err);
+ continue;
+ }
+
+ rule = mlx5e_tc_offload_fdb_rules(esw, flow, spec, attr);
+ if (IS_ERR(rule)) {
+ err = PTR_ERR(rule);
+ mlx5_core_warn(priv->mdev, "Failed to update cached decap flow, %d\n",
+ err);
+ } else {
+ flow->rule[0] = rule;
+ flow_flag_set(flow, OFFLOADED);
+ }
+ }
+}
+
+static int mlx5e_update_route_decap_flows(struct mlx5e_priv *priv,
+ struct mlx5e_route_entry *r,
+ struct list_head *flow_list,
+ bool replace)
+{
+ struct net_device *tunnel_dev;
+ LIST_HEAD(decap_flows);
+
+ tunnel_dev = __dev_get_by_index(dev_net(priv->netdev), r->tunnel_dev_index);
+ if (!tunnel_dev)
+ return -ENODEV;
+
+ mlx5e_take_all_route_decap_flows(r, &decap_flows);
+ if (mlx5e_route_entry_valid(r))
+ mlx5e_unoffload_flow_list(priv, &decap_flows);
+ if (replace)
+ mlx5e_reoffload_decap(priv, &decap_flows);
+
+ list_splice(&decap_flows, flow_list);
+
+ return 0;
+}
+
+static void mlx5e_tc_fib_event_work(struct work_struct *work)
+{
+ struct mlx5e_tc_fib_event_data *event_data =
+ container_of(work, struct mlx5e_tc_fib_event_data, work);
+ struct net_device *ul_dev = event_data->ul_dev;
+ struct mlx5e_priv *priv = netdev_priv(ul_dev);
+ struct mlx5e_route_entry *r = event_data->r;
+ struct mlx5_eswitch *esw;
+ LIST_HEAD(flow_list);
+ bool replace;
+ int err;
+
+ /* sync with concurrent neigh updates */
+ rtnl_lock();
+ esw = priv->mdev->priv.eswitch;
+ mutex_lock(&esw->offloads.encap_tbl_lock);
+ replace = event_data->event == FIB_EVENT_ENTRY_REPLACE;
+
+ if (!mlx5e_route_entry_valid(r) && !replace)
+ goto out;
+
+ err = mlx5e_update_route_encaps(priv, r, &flow_list, replace);
+ if (err)
+ mlx5_core_warn(priv->mdev, "Failed to update route encaps, %d\n",
+ err);
+
+ err = mlx5e_update_route_decap_flows(priv, r, &flow_list, replace);
+ if (err)
+ mlx5_core_warn(priv->mdev, "Failed to update route decap flows, %d\n",
+ err);
+
+ if (replace)
+ r->flags |= MLX5E_ROUTE_ENTRY_VALID;
+out:
+ mutex_unlock(&esw->offloads.encap_tbl_lock);
+ rtnl_unlock();
+
+ mlx5e_put_flow_list(priv, &flow_list);
+ mlx5e_route_put(priv, event_data->r);
+ dev_put(event_data->ul_dev);
+ kfree(event_data);
+}
+
+static struct mlx5e_tc_fib_event_data *
+mlx5e_init_fib_work_ipv4(struct mlx5e_priv *priv,
+ struct net_device *ul_dev,
+ struct mlx5e_tc_tun_encap *encap,
+ unsigned long event,
+ struct fib_notifier_info *info)
+{
+ struct fib_entry_notifier_info *fen_info;
+ struct mlx5e_tc_fib_event_data *fib_work;
+ struct mlx5e_route_entry *r;
+ struct mlx5e_route_key key;
+ struct net_device *fib_dev;
+
+ fen_info = container_of(info, struct fib_entry_notifier_info, info);
+ fib_dev = fib_info_nh(fen_info->fi, 0)->fib_nh_dev;
+ if (fib_dev->netdev_ops != &mlx5e_netdev_ops ||
+ fen_info->dst_len != 32)
+ return NULL;
+
+ fib_work = mlx5e_tc_init_fib_work(event, ul_dev, GFP_ATOMIC);
+ if (!fib_work)
+ return ERR_PTR(-ENOMEM);
+
+ key.endpoint_ip.v4 = htonl(fen_info->dst);
+ key.ip_version = 4;
+
+ /* Can't fail after this point because releasing reference to r
+ * requires obtaining sleeping mutex which we can't do in atomic
+ * context.
+ */
+ r = mlx5e_route_lookup_for_update(encap, &key);
+ if (!r)
+ goto out;
+ fib_work->r = r;
+ dev_hold(ul_dev);
+
+ return fib_work;
+
+out:
+ kfree(fib_work);
+ return NULL;
+}
+
+static struct mlx5e_tc_fib_event_data *
+mlx5e_init_fib_work_ipv6(struct mlx5e_priv *priv,
+ struct net_device *ul_dev,
+ struct mlx5e_tc_tun_encap *encap,
+ unsigned long event,
+ struct fib_notifier_info *info)
+{
+ struct fib6_entry_notifier_info *fen_info;
+ struct mlx5e_tc_fib_event_data *fib_work;
+ struct mlx5e_route_entry *r;
+ struct mlx5e_route_key key;
+ struct net_device *fib_dev;
+
+ fen_info = container_of(info, struct fib6_entry_notifier_info, info);
+ fib_dev = fib6_info_nh_dev(fen_info->rt);
+ if (fib_dev->netdev_ops != &mlx5e_netdev_ops ||
+ fen_info->rt->fib6_dst.plen != 128)
+ return NULL;
+
+ fib_work = mlx5e_tc_init_fib_work(event, ul_dev, GFP_ATOMIC);
+ if (!fib_work)
+ return ERR_PTR(-ENOMEM);
+
+ memcpy(&key.endpoint_ip.v6, &fen_info->rt->fib6_dst.addr,
+ sizeof(fen_info->rt->fib6_dst.addr));
+ key.ip_version = 6;
+
+ /* Can't fail after this point because releasing reference to r
+ * requires obtaining sleeping mutex which we can't do in atomic
+ * context.
+ */
+ r = mlx5e_route_lookup_for_update(encap, &key);
+ if (!r)
+ goto out;
+ fib_work->r = r;
+ dev_hold(ul_dev);
+
+ return fib_work;
+
+out:
+ kfree(fib_work);
+ return NULL;
+}
+
+static int mlx5e_tc_tun_fib_event(struct notifier_block *nb, unsigned long event, void *ptr)
+{
+ struct mlx5e_tc_fib_event_data *fib_work;
+ struct fib_notifier_info *info = ptr;
+ struct mlx5e_tc_tun_encap *encap;
+ struct net_device *ul_dev;
+ struct mlx5e_priv *priv;
+
+ encap = container_of(nb, struct mlx5e_tc_tun_encap, fib_nb);
+ priv = encap->priv;
+ ul_dev = priv->netdev;
+ priv = netdev_priv(ul_dev);
+
+ switch (event) {
+ case FIB_EVENT_ENTRY_REPLACE:
+ case FIB_EVENT_ENTRY_DEL:
+ if (info->family == AF_INET)
+ fib_work = mlx5e_init_fib_work_ipv4(priv, ul_dev, encap, event, info);
+ else if (info->family == AF_INET6)
+ fib_work = mlx5e_init_fib_work_ipv6(priv, ul_dev, encap, event, info);
+ else
+ return NOTIFY_DONE;
+
+ if (!IS_ERR_OR_NULL(fib_work)) {
+ queue_work(priv->wq, &fib_work->work);
+ } else if (IS_ERR(fib_work)) {
+ NL_SET_ERR_MSG_MOD(info->extack, "Failed to init fib work");
+ mlx5_core_warn(priv->mdev, "Failed to init fib work, %ld\n",
+ PTR_ERR(fib_work));
+ }
+
+ break;
+ default:
+ return NOTIFY_DONE;
+ }
+
+ return NOTIFY_DONE;
+}
+
+struct mlx5e_tc_tun_encap *mlx5e_tc_tun_init(struct mlx5e_priv *priv)
+{
+ struct mlx5e_tc_tun_encap *encap;
+ int err;
+
+ encap = kvzalloc(sizeof(*encap), GFP_KERNEL);
+ if (!encap)
+ return ERR_PTR(-ENOMEM);
+
+ encap->priv = priv;
+ encap->fib_nb.notifier_call = mlx5e_tc_tun_fib_event;
+ spin_lock_init(&encap->route_lock);
+ hash_init(encap->route_tbl);
+ err = register_fib_notifier(dev_net(priv->netdev), &encap->fib_nb,
+ NULL, NULL);
+ if (err) {
+ kvfree(encap);
+ return ERR_PTR(err);
+ }
+
+ return encap;
+}
+
+void mlx5e_tc_tun_cleanup(struct mlx5e_tc_tun_encap *encap)
+{
+ if (!encap)
+ return;
+
+ unregister_fib_notifier(dev_net(encap->priv->netdev), &encap->fib_nb);
+ flush_workqueue(encap->priv->wq); /* flush fib event works */
+ kvfree(encap);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h
index 9939cff6e6c5..3391504d9a08 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h
@@ -32,4 +32,7 @@ struct ip_tunnel_info *mlx5e_dup_tun_info(const struct ip_tunnel_info *tun_info)
int mlx5e_tc_set_attr_rx_tun(struct mlx5e_tc_flow *flow,
struct mlx5_flow_spec *spec);
+struct mlx5e_tc_tun_encap *mlx5e_tc_tun_init(struct mlx5e_priv *priv);
+void mlx5e_tc_tun_cleanup(struct mlx5e_tc_tun_encap *encap);
+
#endif /* __MLX5_EN_TC_TUN_ENCAP_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
index e947921a2d5a..d1696404cca9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
@@ -59,6 +59,8 @@ struct mlx5e_neigh_update_table {
struct mlx5_tc_ct_priv;
struct mlx5e_rep_bond;
+struct mlx5e_tc_tun_encap;
+
struct mlx5_rep_uplink_priv {
/* Filters DB - instantiated by the uplink representor and shared by
* the uplink's VFs
@@ -90,6 +92,9 @@ struct mlx5_rep_uplink_priv {
/* support eswitch vports bonding */
struct mlx5e_rep_bond *bond;
+
+ /* tc tunneling encapsulation private data */
+ struct mlx5e_tc_tun_encap *encap;
};
struct mlx5e_rep_priv {
@@ -153,6 +158,7 @@ enum {
/* set when the encap entry is successfully offloaded into HW */
MLX5_ENCAP_ENTRY_VALID = BIT(0),
MLX5_REFORMAT_DECAP = BIT(1),
+ MLX5_ENCAP_ENTRY_NO_ROUTE = BIT(2),
};
struct mlx5e_decap_key {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index c5ecb9e4e767..db142ee96510 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -1279,11 +1279,11 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
struct net_device *out_dev, *encap_dev = NULL;
struct mlx5e_tc_flow_parse_attr *parse_attr;
struct mlx5_flow_attr *attr = flow->attr;
+ bool vf_tun = false, encap_valid = true;
struct mlx5_esw_flow_attr *esw_attr;
struct mlx5_fc *counter = NULL;
struct mlx5e_rep_priv *rpriv;
struct mlx5e_priv *out_priv;
- bool encap_valid = true;
u32 max_prio, max_chain;
int err = 0;
int out_index;
@@ -1297,26 +1297,28 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
if (!mlx5e_is_ft_flow(flow) && attr->chain > max_chain) {
NL_SET_ERR_MSG_MOD(extack,
"Requested chain is out of supported range");
- return -EOPNOTSUPP;
+ err = -EOPNOTSUPP;
+ goto err_out;
}
max_prio = mlx5_chains_get_prio_range(esw_chains(esw));
if (attr->prio > max_prio) {
NL_SET_ERR_MSG_MOD(extack,
"Requested priority is out of supported range");
- return -EOPNOTSUPP;
+ err = -EOPNOTSUPP;
+ goto err_out;
}
if (flow_flag_test(flow, TUN_RX)) {
err = mlx5e_attach_decap_route(priv, flow);
if (err)
- return err;
+ goto err_out;
}
if (flow_flag_test(flow, L3_TO_L2_DECAP)) {
err = mlx5e_attach_decap(priv, flow, extack);
if (err)
- return err;
+ goto err_out;
}
parse_attr = attr->parse_attr;
@@ -1334,8 +1336,11 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
err = mlx5e_attach_encap(priv, flow, out_dev, out_index,
extack, &encap_dev, &encap_valid);
if (err)
- return err;
+ goto err_out;
+ if (esw_attr->dests[out_index].flags &
+ MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE)
+ vf_tun = true;
out_priv = netdev_priv(encap_dev);
rpriv = out_priv->ppriv;
esw_attr->dests[out_index].rep = rpriv->rep;
@@ -1344,19 +1349,27 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
err = mlx5_eswitch_add_vlan_action(esw, attr);
if (err)
- return err;
+ goto err_out;
if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR &&
!(attr->ct_attr.ct_action & TCA_CT_ACT_CLEAR)) {
- err = mlx5e_attach_mod_hdr(priv, flow, parse_attr);
- if (err)
- return err;
+ if (vf_tun) {
+ err = mlx5e_tc_add_flow_mod_hdr(priv, parse_attr, flow);
+ if (err)
+ goto err_out;
+ } else {
+ err = mlx5e_attach_mod_hdr(priv, flow, parse_attr);
+ if (err)
+ goto err_out;
+ }
}
if (attr->action & MLX5_FLOW_CONTEXT_ACTION_COUNT) {
counter = mlx5_fc_create(esw_attr->counter_dev, true);
- if (IS_ERR(counter))
- return PTR_ERR(counter);
+ if (IS_ERR(counter)) {
+ err = PTR_ERR(counter);
+ goto err_out;
+ }
attr->counter = counter;
}
@@ -1370,12 +1383,17 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
else
flow->rule[0] = mlx5e_tc_offload_fdb_rules(esw, flow, &parse_attr->spec, attr);
- if (IS_ERR(flow->rule[0]))
- return PTR_ERR(flow->rule[0]);
- else
- flow_flag_set(flow, OFFLOADED);
+ if (IS_ERR(flow->rule[0])) {
+ err = PTR_ERR(flow->rule[0]);
+ goto err_out;
+ }
+ flow_flag_set(flow, OFFLOADED);
return 0;
+
+err_out:
+ flow_flag_set(flow, FAILED);
+ return err;
}
static bool mlx5_flow_has_geneve_opt(struct mlx5e_tc_flow *flow)
@@ -1397,6 +1415,7 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv,
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
struct mlx5_flow_attr *attr = flow->attr;
struct mlx5_esw_flow_attr *esw_attr;
+ bool vf_tun = false;
int out_index;
esw_attr = attr->esw_attr;
@@ -1421,20 +1440,26 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv,
mlx5e_detach_decap_route(priv, flow);
for (out_index = 0; out_index < MLX5_MAX_FLOW_FWD_VPORTS; out_index++) {
+ if (esw_attr->dests[out_index].flags &
+ MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE)
+ vf_tun = true;
if (esw_attr->dests[out_index].flags & MLX5_ESW_DEST_ENCAP) {
mlx5e_detach_encap(priv, flow, out_index);
kfree(attr->parse_attr->tun_info[out_index]);
}
}
- kvfree(attr->parse_attr);
- kvfree(attr->esw_attr->rx_tun_attr);
mlx5_tc_ct_match_del(get_ct_priv(priv), &flow->attr->ct_attr);
if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) {
dealloc_mod_hdr_actions(&attr->parse_attr->mod_hdr_acts);
- mlx5e_detach_mod_hdr(priv, flow);
+ if (vf_tun && attr->modify_hdr)
+ mlx5_modify_header_dealloc(priv->mdev, attr->modify_hdr);
+ else
+ mlx5e_detach_mod_hdr(priv, flow);
}
+ kvfree(attr->parse_attr);
+ kvfree(attr->esw_attr->rx_tun_attr);
if (attr->action & MLX5_FLOW_CONTEXT_ACTION_COUNT)
mlx5_fc_destroy(esw_attr->counter_dev, attr->counter);
@@ -4044,7 +4069,6 @@ __mlx5e_add_fdb_flow(struct mlx5e_priv *priv,
return flow;
err_free:
- dealloc_mod_hdr_actions(&parse_attr->mod_hdr_acts);
mlx5e_flow_put(priv, flow);
out:
return ERR_PTR(err);
@@ -4189,6 +4213,7 @@ mlx5e_add_nic_flow(struct mlx5e_priv *priv,
return 0;
err_free:
+ flow_flag_set(flow, FAILED);
dealloc_mod_hdr_actions(&parse_attr->mod_hdr_acts);
mlx5e_flow_put(priv, flow);
out:
@@ -4724,8 +4749,14 @@ int mlx5e_tc_esw_init(struct rhashtable *tc_ht)
lockdep_set_class(&tc_ht->mutex, &tc_ht_lock_key);
+ uplink_priv->encap = mlx5e_tc_tun_init(priv);
+ if (IS_ERR(uplink_priv->encap))
+ goto err_register_fib_notifier;
+
return err;
+err_register_fib_notifier:
+ rhashtable_destroy(tc_ht);
err_ht_init:
mapping_destroy(uplink_priv->tunnel_enc_opts_mapping);
err_enc_opts_mapping:
@@ -4742,10 +4773,11 @@ void mlx5e_tc_esw_cleanup(struct rhashtable *tc_ht)
{
struct mlx5_rep_uplink_priv *uplink_priv;
- rhashtable_free_and_destroy(tc_ht, _mlx5e_tc_del_flow, NULL);
-
uplink_priv = container_of(tc_ht, struct mlx5_rep_uplink_priv, tc_ht);
+ rhashtable_free_and_destroy(tc_ht, _mlx5e_tc_del_flow, NULL);
+ mlx5e_tc_tun_cleanup(uplink_priv->encap);
+
mapping_destroy(uplink_priv->tunnel_enc_opts_mapping);
mapping_destroy(uplink_priv->tunnel_mapping);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 310c405e81d7..aba17835465b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1830,7 +1830,6 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
mutex_init(&esw->offloads.decap_tbl_lock);
hash_init(esw->offloads.decap_tbl);
mlx5e_mod_hdr_tbl_init(&esw->offloads.mod_hdr);
- hash_init(esw->offloads.route_tbl);
atomic64_set(&esw->offloads.num_flows, 0);
ida_init(&esw->offloads.vport_metadata_ida);
xa_init_flags(&esw->offloads.vhca_map, XA_FLAGS_ALLOC);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index b90724e27960..fdf5c8c05c1b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -214,7 +214,6 @@ struct mlx5_esw_offload {
struct mutex peer_mutex;
struct mutex encap_tbl_lock; /* protects encap_tbl */
DECLARE_HASHTABLE(encap_tbl, 8);
- DECLARE_HASHTABLE(route_tbl, 8);
struct mutex decap_tbl_lock; /* protects decap_tbl */
DECLARE_HASHTABLE(decap_tbl, 8);
struct mod_hdr_tbl mod_hdr;
@@ -424,6 +423,7 @@ struct mlx5_esw_flow_attr {
struct mlx5_pkt_reformat *pkt_reformat;
struct mlx5_core_dev *mdev;
struct mlx5_termtbl_handle *termtbl;
+ int src_port_rewrite_act_id;
} dests[MLX5_MAX_FLOW_FWD_VPORTS];
struct mlx5_rx_tun_attr *rx_tun_attr;
struct mlx5_pkt_reformat *decap_pkt_reformat;
--
2.29.2
^ permalink raw reply related [flat|nested] 39+ messages in thread* Re: [pull request][net-next V2 00/17] mlx5 updates 2021-02-04
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (16 preceding siblings ...)
2021-02-06 5:02 ` [net-next V2 17/17] net/mlx5e: Handle FIB events to update tunnel endpoint device Saeed Mahameed
@ 2021-02-08 21:55 ` Or Gerlitz
2021-02-09 8:42 ` Or Gerlitz
18 siblings, 0 replies; 39+ messages in thread
From: Or Gerlitz @ 2021-02-08 21:55 UTC (permalink / raw)
To: Saeed Mahameed
Cc: Jakub Kicinski, David S. Miller, Linux Netdev List, Saeed Mahameed
On Sat, Feb 6, 2021 at 7:10 AM Saeed Mahameed <saeed@kernel.org> wrote:
> From: Saeed Mahameed <saeedm@nvidia.com>
> This series adds the support for VF tunneling.
> Vlad Buslov says:
> =================
> Implement support for VF tunneling
> Abstract
> Currently, mlx5 only supports configuration with tunnel endpoint IP address on
> uplink representor. Remove implicit and explicit assumptions of tunnel always
> being terminated on uplink and implement necessary infrastructure for
> configuring tunnels on VF representors and updating rules on such tunnels
> according to routing changes.
>
> SW TC model
maybe before SW TC model, you can explain the SW model (TC is a vehicle
to implement the SW model).
SW model for VST and "classic" v-switch tunnel setup:
For example, in VST model, each virtio/vf/sf vport has a vlan such that
the v-switch tags packets going out "south" of the vport towards the
uplink, untags
packets going "north" from the uplink into the vport (and does nothing
for east-west traffic).
In a similar manner, in "classic" v-switch tunnel setup, each
virtio/vf/sf vport is somehow
associated with VNI/s marking the tenant/s it belongs to. Same tenant
east-west traffic
on the host doesn't go through any encap/decap. The v-switch adds the
relevant tunnel
MD to packets/skbs sent "southward" by the end-point and forwards it
to the VTEP which applies
encap and sends the packets to the wire. On RX, the VTEP decaps the
tunnel info from the packet,
adds it as MD to the skb and forwards the packet up into the stack
where the vsw hooks it, matches
on the MD + inner tuple and then forwards it to the relevant endpoint.
HW offloads for VST and "classic" v-switch tunnel setup:
more or less straight forward based on the above
> From TC perspective VF tunnel configuration requires two rules in both
> directions:
>
> TX rules
>
> 1. Rule that redirects packets from UL to VF rep that has the tunnel
> endpoint IP address:
> 2. Rule that decapsulates the tunneled flow and redirects to destination VF
> representor:
> RX rules
>
> 1. Rule that encapsulates the tunneled flow and redirects packets from
> source VF rep to tunnel device:
> 2. Rule that redirects from tunnel device to UL rep:
Sorry, I am not managing to follow and catch up a SW model from TC rules..
I think we need these two to begin with:
[1] Motivation for enhanced v-switch tunnel setup:
[2] SW model for enhanced v-switch tunnel setup:
> HW offloads model
a clear SW model before HW offloads model..
^ permalink raw reply [flat|nested] 39+ messages in thread* Re: [pull request][net-next V2 00/17] mlx5 updates 2021-02-04
2021-02-06 5:02 [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Saeed Mahameed
` (17 preceding siblings ...)
2021-02-08 21:55 ` [pull request][net-next V2 00/17] mlx5 updates 2021-02-04 Or Gerlitz
@ 2021-02-09 8:42 ` Or Gerlitz
2021-02-09 8:43 ` Or Gerlitz
2021-02-10 16:51 ` Vlad Buslov
18 siblings, 2 replies; 39+ messages in thread
From: Or Gerlitz @ 2021-02-09 8:42 UTC (permalink / raw)
To: Saeed Mahameed, Vlad Buslov, Marcelo Ricardo Leitner
Cc: Jakub Kicinski, Linux Netdev List
On Sat, Feb 6, 2021 at 7:10 AM Saeed Mahameed <saeed@kernel.org> wrote:
> Vlad Buslov says:
> Implement support for VF tunneling
> Currently, mlx5 only supports configuration with tunnel endpoint IP address on
> uplink representor. Remove implicit and explicit assumptions of tunnel always
> being terminated on uplink and implement necessary infrastructure for
> configuring tunnels on VF representors and updating rules on such tunnels
> according to routing changes.
> SW TC model
maybe before SW TC model, you can explain the vswitch SW model (TC is
a vehicle to implement the SW model).
SW model for VST and "classic" v-switch tunnel setup:
For example, in VST model, each virtio/vf/sf vport has a vlan
such that the v-switch tags packets going out "south" of the
vport towards the uplink, untags packets going "north" from
the uplink, matches on the vport tag and forwards them to
the vport (and does nothing for east-west traffic).
In a similar manner, in "classic" v-switch tunnel setup, each
virtio/vf/sf vport is somehow associated with VNI/s marking the
tenant/s it belongs to. Same tenant east-west traffic on the
host doesn't go through any encap/decap. The v-switch adds the
relevant tunnel MD to packets/skbs sent "southward" by the end-point
and forwards it to the VTEP which applies encap based on the MD (LWT
scheme) and sends the packets to the wire. On RX, the VTEP decaps
the tunnel info from the packet, adds it as MD to the skb and
forwards the packet up into the stack where the vsw hooks it, matches
on the MD + inner tuple and then forwards it to the relevant endpoint.
HW offloads for VST and "classic" v-switch tunnel setup:
more or less straight forward based on the above
> From TC perspective VF tunnel configuration requires two rules in both
> directions:
> TX rules
> 1. Rule that redirects packets from UL to VF rep that has the tunnel
> endpoint IP address:
> 2. Rule that decapsulates the tunneled flow and redirects to destination VF
> representor:
> RX rules
> 1. Rule that encapsulates the tunneled flow and redirects packets from
> source VF rep to tunnel device:
> 2. Rule that redirects from tunnel device to UL rep:
mmm it's kinda hard managing to follow and catch up a SW model from TC rules..
I think we need these two to begin with (in whatever order that works
better for you)
[1] Motivation for enhanced v-switch tunnel setup:
[2] SW model for enhanced v-switch tunnel setup:
> HW offloads model
a clear SW model before HW offloads model..
> 25 files changed, 3812 insertions(+), 1057 deletions(-)
for adding almost 4K LOCs
^ permalink raw reply [flat|nested] 39+ messages in thread* Re: [pull request][net-next V2 00/17] mlx5 updates 2021-02-04
2021-02-09 8:42 ` Or Gerlitz
@ 2021-02-09 8:43 ` Or Gerlitz
2021-02-10 16:51 ` Vlad Buslov
1 sibling, 0 replies; 39+ messages in thread
From: Or Gerlitz @ 2021-02-09 8:43 UTC (permalink / raw)
To: Saeed Mahameed, Vlad Buslov, Marcelo Ricardo Leitner
Cc: Jakub Kicinski, Linux Netdev List
On Tue, Feb 9, 2021 at 10:42 AM Or Gerlitz <gerlitz.or@gmail.com> wrote:
> On Sat, Feb 6, 2021 at 7:10 AM Saeed Mahameed <saeed@kernel.org> wrote:
> > Vlad Buslov says:
>
> > Implement support for VF tunneling
>
> > Currently, mlx5 only supports configuration with tunnel endpoint IP address on
> > uplink representor. Remove implicit and explicit assumptions of tunnel always
> > being terminated on uplink and implement necessary infrastructure for
> > configuring tunnels on VF representors and updating rules on such tunnels
> > according to routing changes.
>
> > SW TC model
>
> maybe before SW TC model, you can explain the vswitch SW model (TC is
> a vehicle to implement the SW model).
I thought my earlier post missed the list, so I reposted, but realized
now it didn't,
feel free to address either of the posts
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [pull request][net-next V2 00/17] mlx5 updates 2021-02-04
2021-02-09 8:42 ` Or Gerlitz
2021-02-09 8:43 ` Or Gerlitz
@ 2021-02-10 16:51 ` Vlad Buslov
1 sibling, 0 replies; 39+ messages in thread
From: Vlad Buslov @ 2021-02-10 16:51 UTC (permalink / raw)
To: Or Gerlitz
Cc: Saeed Mahameed, Vlad Buslov, Marcelo Ricardo Leitner,
Jakub Kicinski, Linux Netdev List
On Tue 09 Feb 2021 at 10:42, Or Gerlitz <gerlitz.or@gmail.com> wrote:
> On Sat, Feb 6, 2021 at 7:10 AM Saeed Mahameed <saeed@kernel.org> wrote:
>
>> Vlad Buslov says:
>
>> Implement support for VF tunneling
>
>> Currently, mlx5 only supports configuration with tunnel endpoint IP address on
>> uplink representor. Remove implicit and explicit assumptions of tunnel always
>> being terminated on uplink and implement necessary infrastructure for
>> configuring tunnels on VF representors and updating rules on such tunnels
>> according to routing changes.
>
>> SW TC model
>
> maybe before SW TC model, you can explain the vswitch SW model (TC is
> a vehicle to implement the SW model).
>
> SW model for VST and "classic" v-switch tunnel setup:
>
> For example, in VST model, each virtio/vf/sf vport has a vlan
> such that the v-switch tags packets going out "south" of the
> vport towards the uplink, untags packets going "north" from
> the uplink, matches on the vport tag and forwards them to
> the vport (and does nothing for east-west traffic).
>
> In a similar manner, in "classic" v-switch tunnel setup, each
> virtio/vf/sf vport is somehow associated with VNI/s marking the
> tenant/s it belongs to. Same tenant east-west traffic on the
> host doesn't go through any encap/decap. The v-switch adds the
> relevant tunnel MD to packets/skbs sent "southward" by the end-point
> and forwards it to the VTEP which applies encap based on the MD (LWT
> scheme) and sends the packets to the wire. On RX, the VTEP decaps
> the tunnel info from the packet, adds it as MD to the skb and
> forwards the packet up into the stack where the vsw hooks it, matches
> on the MD + inner tuple and then forwards it to the relevant endpoint.
Moving tunnel endpoint to VF doesn't change anything in this
high-level description.
>
> HW offloads for VST and "classic" v-switch tunnel setup:
>
> more or less straight forward based on the above
>
>> From TC perspective VF tunnel configuration requires two rules in both
>> directions:
>
>> TX rules
>> 1. Rule that redirects packets from UL to VF rep that has the tunnel
>> endpoint IP address:
>> 2. Rule that decapsulates the tunneled flow and redirects to destination VF
>> representor:
>
>> RX rules
>> 1. Rule that encapsulates the tunneled flow and redirects packets from
>> source VF rep to tunnel device:
>> 2. Rule that redirects from tunnel device to UL rep:
>
> mmm it's kinda hard managing to follow and catch up a SW model from TC rules..
>
> I think we need these two to begin with (in whatever order that works
> better for you)
>
> [1] Motivation for enhanced v-switch tunnel setup:
>
> [2] SW model for enhanced v-switch tunnel setup:
>
>> HW offloads model
>
> a clear SW model before HW offloads model..
Hope my replies to Jakub and Marcelo also address these.
>
>> 25 files changed, 3812 insertions(+), 1057 deletions(-)
>
> for adding almost 4K LOCs
^ permalink raw reply [flat|nested] 39+ messages in thread