mbox series

[iwl-next,v4,0/7] Add PFCP filter support

Message ID 20231207164911.14330-1-marcin.szycik@linux.intel.com (mailing list archive)
Headers show
Series Add PFCP filter support | expand

Message

Marcin Szycik Dec. 7, 2023, 4:49 p.m. UTC
Add support for creating PFCP filters in switchdev mode. Add pfcp module
that allows to create a PFCP-type netdev. The netdev then can be passed to
tc when creating a filter to indicate that PFCP filter should be created.

To add a PFCP filter, a special netdev must be created and passed to tc
command:

  ip link add pfcp0 type pfcp
  tc filter add dev eth0 ingress prio 1 flower pfcp_opts \
    1:12ab/ff:fffffffffffffff0 skip_hw action mirred egress redirect \
    dev pfcp0

Changes in iproute2 [1] are required to use pfcp_opts in tc.

ICE COMMS package is required as it contains PFCP profiles.

Part of this patchset modifies IP_TUNNEL_*_OPTs, which were previously
stored in a __be16. All possible values have already been used, making it
impossible to add new ones.

[1] https://lore.kernel.org/netdev/20230614091758.11180-1-marcin.szycik@linux.intel.com
---
This patchset should be applied on top of the "boys" tree [2], as it
depends on recent bitmap changes.

[2] https://github.com/norov/linux/commits/boys
---
v3: https://lore.kernel.org/intel-wired-lan/20230721071532.613888-1-marcin.szycik@linux.intel.com
v2: https://lore.kernel.org/intel-wired-lan/20230607112606.15899-1-marcin.szycik@linux.intel.com
v1: https://lore.kernel.org/intel-wired-lan/20230601131929.294667-1-marcin.szycik@linux.intel.com
---

Alexander Lobakin (3):
  ip_tunnel: use a separate struct to store tunnel params in the kernel
  ip_tunnel: convert __be16 tunnel flags to bitmaps
  lib/bitmap: add tests for IP tunnel flags conversion helpers

Marcin Szycik (2):
  ice: refactor ICE_TC_FLWR_FIELD_ENC_OPTS
  ice: Add support for PFCP hardware offload in switchdev

Michal Swiatkowski (1):
  pfcp: always set pfcp metadata

Wojciech Drewek (1):
  pfcp: add PFCP module

 drivers/net/Kconfig                           |  13 +
 drivers/net/Makefile                          |   1 +
 drivers/net/bareudp.c                         |  19 +-
 drivers/net/ethernet/intel/ice/ice_ddp.c      |   9 +
 .../net/ethernet/intel/ice/ice_flex_type.h    |   4 +-
 .../ethernet/intel/ice/ice_protocol_type.h    |  12 +
 drivers/net/ethernet/intel/ice/ice_switch.c   |  85 +++++
 drivers/net/ethernet/intel/ice/ice_switch.h   |   2 +
 drivers/net/ethernet/intel/ice/ice_tc_lib.c   |  68 +++-
 drivers/net/ethernet/intel/ice/ice_tc_lib.h   |   8 +-
 .../ethernet/mellanox/mlx5/core/en/tc_tun.h   |   2 +-
 .../mellanox/mlx5/core/en/tc_tun_encap.c      |   6 +-
 .../mellanox/mlx5/core/en/tc_tun_geneve.c     |  12 +-
 .../mellanox/mlx5/core/en/tc_tun_gre.c        |   8 +-
 .../mellanox/mlx5/core/en/tc_tun_vxlan.c      |   9 +-
 .../net/ethernet/mellanox/mlx5/core/en_tc.c   |  16 +-
 .../ethernet/mellanox/mlxsw/spectrum_ipip.c   |  56 ++--
 .../ethernet/mellanox/mlxsw/spectrum_ipip.h   |   2 +-
 .../ethernet/mellanox/mlxsw/spectrum_span.c   |  10 +-
 .../ethernet/netronome/nfp/flower/action.c    |  27 +-
 drivers/net/geneve.c                          |  44 ++-
 drivers/net/pfcp.c                            | 302 ++++++++++++++++++
 drivers/net/vxlan/vxlan_core.c                |  14 +-
 include/linux/netdevice.h                     |   7 +-
 include/net/dst_metadata.h                    |  10 +-
 include/net/flow_dissector.h                  |   2 +-
 include/net/gre.h                             |  70 ++--
 include/net/ip6_tunnel.h                      |   4 +-
 include/net/ip_tunnels.h                      | 139 ++++++--
 include/net/pfcp.h                            |  90 ++++++
 include/net/udp_tunnel.h                      |   4 +-
 include/uapi/linux/if_tunnel.h                |  36 +++
 include/uapi/linux/pkt_cls.h                  |  14 +
 lib/test_bitmap.c                             | 100 ++++++
 net/bridge/br_vlan_tunnel.c                   |   9 +-
 net/core/filter.c                             |  26 +-
 net/core/flow_dissector.c                     |  20 +-
 net/ipv4/fou_bpf.c                            |   2 +-
 net/ipv4/gre_demux.c                          |   2 +-
 net/ipv4/ip_gre.c                             | 144 +++++----
 net/ipv4/ip_tunnel.c                          | 109 +++++--
 net/ipv4/ip_tunnel_core.c                     |  82 +++--
 net/ipv4/ip_vti.c                             |  41 ++-
 net/ipv4/ipip.c                               |  33 +-
 net/ipv4/ipmr.c                               |   2 +-
 net/ipv4/udp_tunnel_core.c                    |   5 +-
 net/ipv6/addrconf.c                           |   3 +-
 net/ipv6/ip6_gre.c                            |  85 ++---
 net/ipv6/ip6_tunnel.c                         |  14 +-
 net/ipv6/sit.c                                |  38 ++-
 net/netfilter/ipvs/ip_vs_core.c               |   6 +-
 net/netfilter/ipvs/ip_vs_xmit.c               |  20 +-
 net/netfilter/nft_tunnel.c                    |  44 +--
 net/openvswitch/flow_netlink.c                |  61 ++--
 net/psample/psample.c                         |  26 +-
 net/sched/act_tunnel_key.c                    |  36 +--
 net/sched/cls_flower.c                        | 134 +++++++-
 57 files changed, 1652 insertions(+), 495 deletions(-)
 create mode 100644 drivers/net/pfcp.c
 create mode 100644 include/net/pfcp.h

Comments

Tony Nguyen Dec. 8, 2023, 9:34 p.m. UTC | #1
On 12/7/2023 8:49 AM, Marcin Szycik wrote:
> Add support for creating PFCP filters in switchdev mode. Add pfcp module
> that allows to create a PFCP-type netdev. The netdev then can be passed to
> tc when creating a filter to indicate that PFCP filter should be created.
> 
> To add a PFCP filter, a special netdev must be created and passed to tc
> command:
> 
>    ip link add pfcp0 type pfcp
>    tc filter add dev eth0 ingress prio 1 flower pfcp_opts \
>      1:12ab/ff:fffffffffffffff0 skip_hw action mirred egress redirect \
>      dev pfcp0
> 
> Changes in iproute2 [1] are required to use pfcp_opts in tc.
> 
> ICE COMMS package is required as it contains PFCP profiles.
> 
> Part of this patchset modifies IP_TUNNEL_*_OPTs, which were previously
> stored in a __be16. All possible values have already been used, making it
> impossible to add new ones.
> 
> [1] https://lore.kernel.org/netdev/20230614091758.11180-1-marcin.szycik@linux.intel.com
> ---
> This patchset should be applied on top of the "boys" tree [2], as it
> depends on recent bitmap changes.

Is this for comment only (RFC)? This doesn't seem to apply to iwl-next 
and if this based on, and has dependencies from, another tree, I can't 
apply them here.

> [2] https://github.com/norov/linux/commits/boys
> ---
> v3: https://lore.kernel.org/intel-wired-lan/20230721071532.613888-1-marcin.szycik@linux.intel.com
> v2: https://lore.kernel.org/intel-wired-lan/20230607112606.15899-1-marcin.szycik@linux.intel.com
> v1: https://lore.kernel.org/intel-wired-lan/20230601131929.294667-1-marcin.szycik@linux.intel.com
> ---
> 
> Alexander Lobakin (3):
>    ip_tunnel: use a separate struct to store tunnel params in the kernel
>    ip_tunnel: convert __be16 tunnel flags to bitmaps
>    lib/bitmap: add tests for IP tunnel flags conversion helpers
> 
> Marcin Szycik (2):
>    ice: refactor ICE_TC_FLWR_FIELD_ENC_OPTS
>    ice: Add support for PFCP hardware offload in switchdev
> 
> Michal Swiatkowski (1):
>    pfcp: always set pfcp metadata
> 
> Wojciech Drewek (1):
>    pfcp: add PFCP module
> 
>   drivers/net/Kconfig                           |  13 +
>   drivers/net/Makefile                          |   1 +
>   drivers/net/bareudp.c                         |  19 +-
>   drivers/net/ethernet/intel/ice/ice_ddp.c      |   9 +
>   .../net/ethernet/intel/ice/ice_flex_type.h    |   4 +-
>   .../ethernet/intel/ice/ice_protocol_type.h    |  12 +
>   drivers/net/ethernet/intel/ice/ice_switch.c   |  85 +++++
>   drivers/net/ethernet/intel/ice/ice_switch.h   |   2 +
>   drivers/net/ethernet/intel/ice/ice_tc_lib.c   |  68 +++-
>   drivers/net/ethernet/intel/ice/ice_tc_lib.h   |   8 +-
>   .../ethernet/mellanox/mlx5/core/en/tc_tun.h   |   2 +-
>   .../mellanox/mlx5/core/en/tc_tun_encap.c      |   6 +-
>   .../mellanox/mlx5/core/en/tc_tun_geneve.c     |  12 +-
>   .../mellanox/mlx5/core/en/tc_tun_gre.c        |   8 +-
>   .../mellanox/mlx5/core/en/tc_tun_vxlan.c      |   9 +-
>   .../net/ethernet/mellanox/mlx5/core/en_tc.c   |  16 +-
>   .../ethernet/mellanox/mlxsw/spectrum_ipip.c   |  56 ++--
>   .../ethernet/mellanox/mlxsw/spectrum_ipip.h   |   2 +-
>   .../ethernet/mellanox/mlxsw/spectrum_span.c   |  10 +-
>   .../ethernet/netronome/nfp/flower/action.c    |  27 +-
>   drivers/net/geneve.c                          |  44 ++-
>   drivers/net/pfcp.c                            | 302 ++++++++++++++++++
>   drivers/net/vxlan/vxlan_core.c                |  14 +-
>   include/linux/netdevice.h                     |   7 +-
>   include/net/dst_metadata.h                    |  10 +-
>   include/net/flow_dissector.h                  |   2 +-
>   include/net/gre.h                             |  70 ++--
>   include/net/ip6_tunnel.h                      |   4 +-
>   include/net/ip_tunnels.h                      | 139 ++++++--
>   include/net/pfcp.h                            |  90 ++++++
>   include/net/udp_tunnel.h                      |   4 +-
>   include/uapi/linux/if_tunnel.h                |  36 +++
>   include/uapi/linux/pkt_cls.h                  |  14 +
>   lib/test_bitmap.c                             | 100 ++++++
>   net/bridge/br_vlan_tunnel.c                   |   9 +-
>   net/core/filter.c                             |  26 +-
>   net/core/flow_dissector.c                     |  20 +-
>   net/ipv4/fou_bpf.c                            |   2 +-
>   net/ipv4/gre_demux.c                          |   2 +-
>   net/ipv4/ip_gre.c                             | 144 +++++----
>   net/ipv4/ip_tunnel.c                          | 109 +++++--
>   net/ipv4/ip_tunnel_core.c                     |  82 +++--
>   net/ipv4/ip_vti.c                             |  41 ++-
>   net/ipv4/ipip.c                               |  33 +-
>   net/ipv4/ipmr.c                               |   2 +-
>   net/ipv4/udp_tunnel_core.c                    |   5 +-
>   net/ipv6/addrconf.c                           |   3 +-
>   net/ipv6/ip6_gre.c                            |  85 ++---
>   net/ipv6/ip6_tunnel.c                         |  14 +-
>   net/ipv6/sit.c                                |  38 ++-
>   net/netfilter/ipvs/ip_vs_core.c               |   6 +-
>   net/netfilter/ipvs/ip_vs_xmit.c               |  20 +-
>   net/netfilter/nft_tunnel.c                    |  44 +--
>   net/openvswitch/flow_netlink.c                |  61 ++--
>   net/psample/psample.c                         |  26 +-
>   net/sched/act_tunnel_key.c                    |  36 +--
>   net/sched/cls_flower.c                        | 134 +++++++-
>   57 files changed, 1652 insertions(+), 495 deletions(-)
>   create mode 100644 drivers/net/pfcp.c
>   create mode 100644 include/net/pfcp.h
>
Alexander Lobakin Dec. 11, 2023, 12:38 p.m. UTC | #2
From: Tony Nguyen <anthony.l.nguyen@intel.com>
Date: Fri, 8 Dec 2023 13:34:10 -0800

> 
> 
> On 12/7/2023 8:49 AM, Marcin Szycik wrote:
>> Add support for creating PFCP filters in switchdev mode. Add pfcp module
>> that allows to create a PFCP-type netdev. The netdev then can be
>> passed to
>> tc when creating a filter to indicate that PFCP filter should be created.
>>
>> To add a PFCP filter, a special netdev must be created and passed to tc
>> command:
>>
>>    ip link add pfcp0 type pfcp
>>    tc filter add dev eth0 ingress prio 1 flower pfcp_opts \
>>      1:12ab/ff:fffffffffffffff0 skip_hw action mirred egress redirect \
>>      dev pfcp0
>>
>> Changes in iproute2 [1] are required to use pfcp_opts in tc.
>>
>> ICE COMMS package is required as it contains PFCP profiles.
>>
>> Part of this patchset modifies IP_TUNNEL_*_OPTs, which were previously
>> stored in a __be16. All possible values have already been used, making it
>> impossible to add new ones.
>>
>> [1]
>> https://lore.kernel.org/netdev/20230614091758.11180-1-marcin.szycik@linux.intel.com
>> ---
>> This patchset should be applied on top of the "boys" tree [2], as it
>> depends on recent bitmap changes.
> 
> Is this for comment only (RFC)? This doesn't seem to apply to iwl-next
> and if this based on, and has dependencies from, another tree, I can't
> apply them here.

It's not an RFC.
The series contains generic code changes and must go directly through
net-next. The dependency on the bitmap tree was discussed with Jakub and
Yury and we agreed that the netdev guys will pull it before applying
this one.

Thanks,
Olek
Tony Nguyen Dec. 11, 2023, 9:23 p.m. UTC | #3
On 12/11/2023 4:38 AM, Alexander Lobakin wrote:
> From: Tony Nguyen <anthony.l.nguyen@intel.com>
> Date: Fri, 8 Dec 2023 13:34:10 -0800
> 
>>
>>
>> On 12/7/2023 8:49 AM, Marcin Szycik wrote:
>>> Add support for creating PFCP filters in switchdev mode. Add pfcp module
>>> that allows to create a PFCP-type netdev. The netdev then can be
>>> passed to
>>> tc when creating a filter to indicate that PFCP filter should be created.
>>>
>>> To add a PFCP filter, a special netdev must be created and passed to tc
>>> command:
>>>
>>>     ip link add pfcp0 type pfcp
>>>     tc filter add dev eth0 ingress prio 1 flower pfcp_opts \
>>>       1:12ab/ff:fffffffffffffff0 skip_hw action mirred egress redirect \
>>>       dev pfcp0
>>>
>>> Changes in iproute2 [1] are required to use pfcp_opts in tc.
>>>
>>> ICE COMMS package is required as it contains PFCP profiles.
>>>
>>> Part of this patchset modifies IP_TUNNEL_*_OPTs, which were previously
>>> stored in a __be16. All possible values have already been used, making it
>>> impossible to add new ones.
>>>
>>> [1]
>>> https://lore.kernel.org/netdev/20230614091758.11180-1-marcin.szycik@linux.intel.com
>>> ---
>>> This patchset should be applied on top of the "boys" tree [2], as it
>>> depends on recent bitmap changes.
>>
>> Is this for comment only (RFC)? This doesn't seem to apply to iwl-next
>> and if this based on, and has dependencies from, another tree, I can't
>> apply them here.
> 
> It's not an RFC.
> The series contains generic code changes and must go directly through
> net-next. 

Should this be marked for 'net-next' rather than 'iwl-next' then?

Thanks,
Tony

> The dependency on the bitmap tree was discussed with Jakub and
> Yury and we agreed that the netdev guys will pull it before applying
> this one.
> 
> Thanks,
> Olek
Marcin Szycik Dec. 12, 2023, 10:45 a.m. UTC | #4
On 11.12.2023 22:23, Tony Nguyen wrote:
> 
> 
> On 12/11/2023 4:38 AM, Alexander Lobakin wrote:
>> From: Tony Nguyen <anthony.l.nguyen@intel.com>
>> Date: Fri, 8 Dec 2023 13:34:10 -0800
>>
>>>
>>>
>>> On 12/7/2023 8:49 AM, Marcin Szycik wrote:
>>>> Add support for creating PFCP filters in switchdev mode. Add pfcp module
>>>> that allows to create a PFCP-type netdev. The netdev then can be
>>>> passed to
>>>> tc when creating a filter to indicate that PFCP filter should be created.
>>>>
>>>> To add a PFCP filter, a special netdev must be created and passed to tc
>>>> command:
>>>>
>>>>     ip link add pfcp0 type pfcp
>>>>     tc filter add dev eth0 ingress prio 1 flower pfcp_opts \
>>>>       1:12ab/ff:fffffffffffffff0 skip_hw action mirred egress redirect \
>>>>       dev pfcp0
>>>>
>>>> Changes in iproute2 [1] are required to use pfcp_opts in tc.
>>>>
>>>> ICE COMMS package is required as it contains PFCP profiles.
>>>>
>>>> Part of this patchset modifies IP_TUNNEL_*_OPTs, which were previously
>>>> stored in a __be16. All possible values have already been used, making it
>>>> impossible to add new ones.
>>>>
>>>> [1]
>>>> https://lore.kernel.org/netdev/20230614091758.11180-1-marcin.szycik@linux.intel.com
>>>> ---
>>>> This patchset should be applied on top of the "boys" tree [2], as it
>>>> depends on recent bitmap changes.
>>>
>>> Is this for comment only (RFC)? This doesn't seem to apply to iwl-next
>>> and if this based on, and has dependencies from, another tree, I can't
>>> apply them here.
>>
>> It's not an RFC.
>> The series contains generic code changes and must go directly through
>> net-next. 
> 
> Should this be marked for 'net-next' rather than 'iwl-next' then?

My bad, sorry.
This series should go directly to net-next.

Thanks,
Marcin

> 
> Thanks,
> Tony
> 
>> The dependency on the bitmap tree was discussed with Jakub and
>> Yury and we agreed that the netdev guys will pull it before applying
>> this one.
>>
>> Thanks,
>> Olek
> _______________________________________________
> Intel-wired-lan mailing list
> Intel-wired-lan@osuosl.org
> https://lists.osuosl.org/mailman/listinfo/intel-wired-lan
Alexander Lobakin Dec. 15, 2023, 10:11 a.m. UTC | #5
From: Marcin Szycik <marcin.szycik@linux.intel.com>
Date: Tue, 12 Dec 2023 11:45:24 +0100

> 
> 
> On 11.12.2023 22:23, Tony Nguyen wrote:
>>
>>
>> On 12/11/2023 4:38 AM, Alexander Lobakin wrote:
>>> From: Tony Nguyen <anthony.l.nguyen@intel.com>
>>> Date: Fri, 8 Dec 2023 13:34:10 -0800
>>>
>>>>
>>>>
>>>> On 12/7/2023 8:49 AM, Marcin Szycik wrote:
>>>>> Add support for creating PFCP filters in switchdev mode. Add pfcp module
>>>>> that allows to create a PFCP-type netdev. The netdev then can be
>>>>> passed to
>>>>> tc when creating a filter to indicate that PFCP filter should be created.
>>>>>
>>>>> To add a PFCP filter, a special netdev must be created and passed to tc
>>>>> command:
>>>>>
>>>>>     ip link add pfcp0 type pfcp
>>>>>     tc filter add dev eth0 ingress prio 1 flower pfcp_opts \
>>>>>       1:12ab/ff:fffffffffffffff0 skip_hw action mirred egress redirect \
>>>>>       dev pfcp0
>>>>>
>>>>> Changes in iproute2 [1] are required to use pfcp_opts in tc.
>>>>>
>>>>> ICE COMMS package is required as it contains PFCP profiles.
>>>>>
>>>>> Part of this patchset modifies IP_TUNNEL_*_OPTs, which were previously
>>>>> stored in a __be16. All possible values have already been used, making it
>>>>> impossible to add new ones.
>>>>>
>>>>> [1]
>>>>> https://lore.kernel.org/netdev/20230614091758.11180-1-marcin.szycik@linux.intel.com
>>>>> ---
>>>>> This patchset should be applied on top of the "boys" tree [2], as it
>>>>> depends on recent bitmap changes.
>>>>
>>>> Is this for comment only (RFC)? This doesn't seem to apply to iwl-next
>>>> and if this based on, and has dependencies from, another tree, I can't
>>>> apply them here.
>>>
>>> It's not an RFC.
>>> The series contains generic code changes and must go directly through
>>> net-next. 
>>
>> Should this be marked for 'net-next' rather than 'iwl-next' then?
> 
> My bad, sorry.
> This series should go directly to net-next.
> 
> Thanks,
> Marcin
> 
>>
>> Thanks,
>> Tony
>>
>>> The dependency on the bitmap tree was discussed with Jakub and
>>> Yury and we agreed that the netdev guys will pull it before applying
>>> this one.

Ping? :s
Or should we resubmit?

Thanks,
Olek
Jakub Kicinski Dec. 15, 2023, 4:49 p.m. UTC | #6
On Fri, 15 Dec 2023 11:11:23 +0100 Alexander Lobakin wrote:
> Ping? :s
> Or should we resubmit?

Can you wait for next merge window instead?
We're getting flooded with patches as everyone seemingly tries to get
their own (i.e. the most important!) work merged before the end of 
the year. The set of PRs from the bitmap tree which Linus decided
not to pull is not empty. So we'd have to go figure out what's exactly
is in that branch we're supposed to pull, and whether it's fine.
It probably is, but you see, this is a problem which can be solved by
waiting, and letting Linus pull it himself. While the 150 patches we're
getting a day now have to be looked at.
Marcin Szycik Dec. 18, 2023, 10:04 a.m. UTC | #7
On 15.12.2023 17:49, Jakub Kicinski wrote:
> On Fri, 15 Dec 2023 11:11:23 +0100 Alexander Lobakin wrote:
>> Ping? :s
>> Or should we resubmit?
> 
> Can you wait for next merge window instead?
> We're getting flooded with patches as everyone seemingly tries to get
> their own (i.e. the most important!) work merged before the end of 
> the year. The set of PRs from the bitmap tree which Linus decided
> not to pull is not empty. So we'd have to go figure out what's exactly
> is in that branch we're supposed to pull, and whether it's fine.
> It probably is, but you see, this is a problem which can be solved by
> waiting, and letting Linus pull it himself. While the 150 patches we're
> getting a day now have to be looked at.

Let's wait to the next window then.

Thanks,
Marcin
Alexander Lobakin Dec. 18, 2023, 12:47 p.m. UTC | #8
From: Marcin Szycik <marcin.szycik@linux.intel.com>
Date: Mon, 18 Dec 2023 11:04:01 +0100

> 
> 
> On 15.12.2023 17:49, Jakub Kicinski wrote:
>> On Fri, 15 Dec 2023 11:11:23 +0100 Alexander Lobakin wrote:
>>> Ping? :s
>>> Or should we resubmit?
>>
>> Can you wait for next merge window instead?
>> We're getting flooded with patches as everyone seemingly tries to get
>> their own (i.e. the most important!) work merged before the end of 
>> the year. The set of PRs from the bitmap tree which Linus decided
>> not to pull is not empty. So we'd have to go figure out what's exactly
>> is in that branch we're supposed to pull, and whether it's fine.
>> It probably is, but you see, this is a problem which can be solved by
>> waiting, and letting Linus pull it himself. While the 150 patches we're
>> getting a day now have to be looked at.
> 
> Let's wait to the next window then.

Hey Yury,

Given that PFCP will be resent in the next window...

Your "boys" tree is in fact self-contained -- those are mostly
optimizations and cleanups, and for the new API -- bitmap_{read,write}()
-- it has internal users (after "bitmap: make bitmap_{get,set}_value8()
use bitmap_{read,write}()"). IOW, I don't see a reason for not merging
it into your main for-next tree (this week :p).
What do you think?

> 
> Thanks,
> Marcin

Thanks,
Olek
Yury Norov Dec. 18, 2023, 3:57 p.m. UTC | #9
+ Alexander Potapenko

On Mon, Dec 18, 2023 at 01:47:01PM +0100, Alexander Lobakin wrote:
> From: Marcin Szycik <marcin.szycik@linux.intel.com>
> Date: Mon, 18 Dec 2023 11:04:01 +0100
> 
> > 
> > 
> > On 15.12.2023 17:49, Jakub Kicinski wrote:
> >> On Fri, 15 Dec 2023 11:11:23 +0100 Alexander Lobakin wrote:
> >>> Ping? :s
> >>> Or should we resubmit?
> >>
> >> Can you wait for next merge window instead?
> >> We're getting flooded with patches as everyone seemingly tries to get
> >> their own (i.e. the most important!) work merged before the end of 
> >> the year. The set of PRs from the bitmap tree which Linus decided
> >> not to pull is not empty. So we'd have to go figure out what's exactly
> >> is in that branch we're supposed to pull, and whether it's fine.
> >> It probably is, but you see, this is a problem which can be solved by
> >> waiting, and letting Linus pull it himself. While the 150 patches we're
> >> getting a day now have to be looked at.
> > 
> > Let's wait to the next window then.
> 
> Hey Yury,
> 
> Given that PFCP will be resent in the next window...
> 
> Your "boys" tree is in fact self-contained -- those are mostly
> optimizations and cleanups, and for the new API -- bitmap_{read,write}()
> -- it has internal users (after "bitmap: make bitmap_{get,set}_value8()
> use bitmap_{read,write}()"). IOW, I don't see a reason for not merging
> it into your main for-next tree (this week :p).
> What do you think?

I think that there's already enough mess with this patch. Alexander
submitted new version of his MTE series together with the patch.

https://lore.kernel.org/lkml/ZXtciaxTKFBiui%2FX@yury-ThinkPad/T/

Now you're asking me to merge it separately. I don't want to undercut
arm64 folks.

Can you guys decide what you want? If you want to move
bitmap_read/write() with my branch, I need to send it in -next for
testing ASAP. And for that, as I already said, I need at least one
active user in current kernel tree. (Yes, bitmap_get_value8() counts.)

If you want to move it this way, please resend all the patches
together.

Thanks,
Yury
Alexander Potapenko Dec. 18, 2023, 4:16 p.m. UTC | #10
On Mon, Dec 18, 2023 at 4:57 PM Yury Norov <yury.norov@gmail.com> wrote:
>
> + Alexander Potapenko
>
> On Mon, Dec 18, 2023 at 01:47:01PM +0100, Alexander Lobakin wrote:
> > From: Marcin Szycik <marcin.szycik@linux.intel.com>
> > Date: Mon, 18 Dec 2023 11:04:01 +0100
> >
> > >
> > >
> > > On 15.12.2023 17:49, Jakub Kicinski wrote:
> > >> On Fri, 15 Dec 2023 11:11:23 +0100 Alexander Lobakin wrote:
> > >>> Ping? :s
> > >>> Or should we resubmit?
> > >>
> > >> Can you wait for next merge window instead?
> > >> We're getting flooded with patches as everyone seemingly tries to get
> > >> their own (i.e. the most important!) work merged before the end of
> > >> the year. The set of PRs from the bitmap tree which Linus decided
> > >> not to pull is not empty. So we'd have to go figure out what's exactly
> > >> is in that branch we're supposed to pull, and whether it's fine.
> > >> It probably is, but you see, this is a problem which can be solved by
> > >> waiting, and letting Linus pull it himself. While the 150 patches we're
> > >> getting a day now have to be looked at.
> > >
> > > Let's wait to the next window then.
> >
> > Hey Yury,
> >
> > Given that PFCP will be resent in the next window...
> >
> > Your "boys" tree is in fact self-contained -- those are mostly
> > optimizations and cleanups, and for the new API -- bitmap_{read,write}()
> > -- it has internal users (after "bitmap: make bitmap_{get,set}_value8()
> > use bitmap_{read,write}()"). IOW, I don't see a reason for not merging
> > it into your main for-next tree (this week :p).
> > What do you think?
>
> I think that there's already enough mess with this patch. Alexander
> submitted new version of his MTE series together with the patch.

Yeah, sorry about that. Because the MTE part of the patches was still
awaiting review, I thought it would be better to land the bitmap API
separately, but as you pointed out there should be at least one user
for it, which it wouldn't have in that case.

I don't have a strong preference about whether to submit the patches
before or after the end of year - in fact I don't think they are
urgent enough, and we'd better postpone them till January.

So unless Alexander has urgent fixes depending on my bitmap patches,
I'd suggest waiting till they are taken via the arm64 tree.

> https://lore.kernel.org/lkml/ZXtciaxTKFBiui%2FX@yury-ThinkPad/T/
>
> Now you're asking me to merge it separately. I don't want to undercut
> arm64 folks.
>
> Can you guys decide what you want? If you want to move
> bitmap_read/write() with my branch, I need to send it in -next for
> testing ASAP. And for that, as I already said, I need at least one
> active user in current kernel tree. (Yes, bitmap_get_value8() counts.)
>
> If you want to move it this way, please resend all the patches
> together.
>
> Thanks,
> Yury



--
Alexander Potapenko
Software Engineer

Google Germany GmbH
Erika-Mann-Straße, 33
80636 München

Geschäftsführer: Paul Manicle, Liana Sebastian
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Alexander Lobakin Dec. 19, 2023, 9:04 a.m. UTC | #11
From: Alexander Potapenko <glider@google.com>
Date: Mon, 18 Dec 2023 17:16:09 +0100

> On Mon, Dec 18, 2023 at 4:57 PM Yury Norov <yury.norov@gmail.com> wrote:
>>
>> + Alexander Potapenko
>>
>> On Mon, Dec 18, 2023 at 01:47:01PM +0100, Alexander Lobakin wrote:

[...]

>>> Hey Yury,
>>>
>>> Given that PFCP will be resent in the next window...
>>>
>>> Your "boys" tree is in fact self-contained -- those are mostly
>>> optimizations and cleanups, and for the new API -- bitmap_{read,write}()
>>> -- it has internal users (after "bitmap: make bitmap_{get,set}_value8()
>>> use bitmap_{read,write}()"). IOW, I don't see a reason for not merging
>>> it into your main for-next tree (this week :p).
>>> What do you think?
>>
>> I think that there's already enough mess with this patch. Alexander
>> submitted new version of his MTE series together with the patch.
> 
> Yeah, sorry about that. Because the MTE part of the patches was still
> awaiting review, I thought it would be better to land the bitmap API
> separately, but as you pointed out there should be at least one user
> for it, which it wouldn't have in that case.
> 
> I don't have a strong preference about whether to submit the patches
> before or after the end of year - in fact I don't think they are
> urgent enough, and we'd better postpone them till January.
> 
> So unless Alexander has urgent fixes depending on my bitmap patches,
> I'd suggest waiting till they are taken via the arm64 tree.

No, nothing urgent. Sounds good, no need to rush at the end of the dev
cycle.

> 
>> https://lore.kernel.org/lkml/ZXtciaxTKFBiui%2FX@yury-ThinkPad/T/
>>
>> Now you're asking me to merge it separately. I don't want to undercut
>> arm64 folks.
>>
>> Can you guys decide what you want? If you want to move
>> bitmap_read/write() with my branch, I need to send it in -next for
>> testing ASAP. And for that, as I already said, I need at least one
>> active user in current kernel tree. (Yes, bitmap_get_value8() counts.)
>>
>> If you want to move it this way, please resend all the patches
>> together.

[...]

Thanks,
Olek