[Weekly meetings] MoM - 28th of April 2022

[Weekly meetings] MoM - 28th of April 2022

[Mat Martineau]

Hello everyone -

Today was our 193rd meeting, with Paolo, Davide, Florian (Red Hat), 
Geliang (SuSE), Giray (Parkyeri), Pol (Tessares), Ossama, and me (Intel).


Meeting minutes:

Accepted patches:
     - The list of accepted patches can be seen on PatchWork:
       https://patchwork.kernel.org/project/mptcp/list/?state=3


     netdev (if mptcp ML is in cc) (by: Mat Martineau):

12827998  [net-next,7/7] selftests: mptcp: print extra msg in chk_csum_nr
12827997  [net-next,6/7] selftests: mptcp: check MP_FAIL response mibs
12827995  [net-next,5/7] mptcp: reset subflow when MP_FAIL doesn't respond
12827996  [net-next,4/7] mptcp: add MP_FAIL response support
12827994  [net-next,3/7] mptcp: add data lock for sk timers
12827993  [net-next,2/7] mptcp: use mptcp_stop_timer
12827992  [net-next,1/7] selftests: mptcp: add infinite map testcase
12824090  [net-next,8/8] selftests: mptcp: add infinite map mibs check
12824091  [net-next,7/8] mptcp: dump infinite_map field in 
mptcp_dump_mpext
12824089  [net-next,6/8] mptcp: add mib for infinite map sending
12824087  [net-next,5/8] mptcp: infinite mapping receiving
12824088  [net-next,4/8] mptcp: infinite mapping sending
12824086  [net-next,3/8] mptcp: track and update contiguous data status
12824085  [net-next,2/8] mptcp: add the fallback check
12824084  [net-next,1/8] mptcp: don't send RST for single subflow
   - Both of these series have been merged.


     our repo (by: Geliang Tang, kernel test robot, Paolo Abeni):

12826789  [mptcp-next] selftests/bpf: Drop duplicate max/min definitions
12826453  [mptcp-next] Squash to "selftests: bpf: test 
bpf_skc_to_mptcp_sock"
12826452  [mptcp-next] Squash to "bpf: add bpf_skc_to_mptcp_sock_proto"

12824418  mptcp: fix semicolon.cocci warnings

12821720  [v2,5/5] mptcp: add more offered MIBs counter.
12821719  [v2,4/5] mptcp: never shrink offered window
12821718  [v2,3/5] tcp: allow MPTCP to update the announced window.
12821717  [v2,2/5] mptcp: add mib for xmit window sharing
12821716  [v2,1/5] mptcp: really share subflow snd_wnd


Pending patches:
     - The list of pending patches can be seen on PatchWork:
       https://patchwork.kernel.org/project/mptcp/list/?state=*

     netdev (if mptcp ML is in cc) (by: Mat Martineau, Paolo Abeni):

12829644  [net-next,6/6] selftests: mptcp: Add tests for userspace PM type
12829643  [net-next,5/6] mptcp: Add a per-namespace sysctl to set the 
default p...
12829642  [net-next,4/6] mptcp: Make kernel path manager check for 
userspace-ma...
12829641  [net-next,3/6] mptcp: Bypass kernel PM when userspace PM is 
enabled
12829640  [net-next,2/6] mptcp: Add a member to mptcp_pm_data to track 
kernel v...
12829639  [net-next,1/6] mptcp: Remove redundant assignments in path 
manager in...

12819235  [RFC,2/2] mptcp: never shrink offered window
12819237  [RFC,1/2] tcp: allow MPTCP to update the announced window.
   - No comment, must be ok :)

     our repo (by: Dmytro SHYTYI, Geliang Tang, Jiapeng Chong, Mat 
Martineau, Matthieu Baerts, Paolo Abeni):

RFC:

12282219: RFC: [RESEND,RFC,2/4] tcp: move selected mptcp helpers to 
tcp.h/mptcp.h
12282221: RFC: [RESEND,RFC,4/4] tcp: parse tcp options contained in reset 
packets
12282223: RFC: [RESEND,RFC,mptpcp-next] mptcp: add ooo prune support
12282225: RFC: [RESEND,1/5] tcp: make two mptcp helpers available to tcp 
stack
12282227: RFC: [RESEND,5/5] mptcp: send fastclose if userspace closes 
socket with unread data
12321111: RFC: mptcp: Remove redundant assignment to remaining
12714439: RFC: [net-next,v2] net: mptcp, Fast Open Mechanism

New:
12829923: New: [mptcp-next,v2] selftests/bpf: Enable CONFIG_IKCONFIG_PROC 
in config
   - To be part of the bpf v2 series

12829996: New: [mptcp-next,v17,1/8] mptcp: add struct mptcp_sched_ops
12829997: New: [mptcp-next,v17,2/8] mptcp: add a new sysctl scheduler
12829998: New: [mptcp-next,v17,3/8] mptcp: add sched in mptcp_sock
12829999: New: [mptcp-next,v17,4/8] mptcp: add get_subflow wrappers
12830000: New: [mptcp-next,v17,5/8] mptcp: add bpf_mptcp_sched_ops
12830001: New: [mptcp-next,v17,6/8] mptcp: add last_snd write access
12830002: New: [mptcp-next,v17,7/8] selftests: bpf: add bpf_first 
scheduler
12830003: New: [mptcp-next,v17,8/8] selftests: bpf: add bpf_first test
     - Addressed Paolo's comments on v16
     - Geliang asks: what about using a separate BPF call for the 
retransmit scheduler?
       - v17 handling of retransmit seems ok to reviewers so far
     - Doesn't fully handle sending on multiple subflows yet - we still 
need to work out how the BPF api (with 'call again' option) ties in to the 
transmit loop.


Issues on Github:
     https://github.com/multipath-tcp/mptcp_net-next/issues/

  Upcoming:
      - Paolo has observed an issue where the MPTCP socket is left dangling 
up to a 60-second timeout, even after the owner process exits. Still need 
to investigate.


      Recently opened (latest from last week: 268)

   271  TCP_DEFER_ACCEPT socket option is not supported
     - No MPTCP support yet, running
   270  KASAN: use-after-free in dst_destroy (net/core/dst.c:118)  [bug] 
[selftests]
     - Appears to be upstream, not caused by MPTCP code
   269  Allow having a mix of v4/v6 subflows for the same socket 
[enhancement]
     - Enhancement request


     Bugs (opened, flagged as "bug" and assigned)

   264  selftests: diag: failing on the public CI with the new debug.config 
[bug] [selftests] @matttbe
   181  implement data_fin ack retransmission for subflow in  TIME_WAIT 
state [bug] @mjmartineau
     - Nothing to report


     Bugs (opened and flagged as "bug" and not assigned)

   265  self-tests #94 && #95 failing [bug] [selftests]
   248  packetdrill: more tests failing due to packets arriving later than 
expected [bug] [packetdrill]


     In Progress (opened, new feature and assigned)

   234  Packetdrill: Support MPC+DATA+checksum error [enhancement] 
[packetdrill] @spoorva
   216  The infinite mapping support: drop data [enhancement] @geliangtang
   167  packetdrill: add coverage for RM_ADDR [enhancement] [packetdrill] 
@dcaratti
    - Trying to check on IPv6 failures first

    75  BPF: packet scheduler [enhancement] @geliangtang
    - Actively revising - see v17 patch series above
    74  BPF: path manager [enhancement] @geliangtang


     For later (opened and not assigned assigned)

   271  TCP_DEFER_ACCEPT socket option is not supported
   269  Allow having a mix of v4/v6 subflows for the same socket 
[enhancement]
   266  Packetdrill: add MP_FAIL coverage [packetdrill]
   236  Review supported sockopts list [enhancement]
   222  Netlink event API: add SUBFLOW_CREATED event [enhancement]
   215  TCP Urgent pointer and MPTCP [enhancement]
   213  add MPTCP man page [enhancement]
   208  better handing of ssk memory pressure in the TX path [enhancement]
   202  Add sendmsg support for ancillary data [enhancement]
   197  more mibs needed [enhancement]
   180  Get an update when MPTCP fall back to TCP [enhancement]
   177  improve retransmit subflow selection [enhancement]
   169  packetdrill: add coverage for ADD_ADDR and MP_JOIN on a different 
port [enhancement] [packetdrill]
   150  remove completely workqueue usage [enhancement]
   141  avoid acquiring mptcp_data_lock() twice in the receive path 
[enhancement]
   133  PM: Closing the MPTCP connection when last subflow is not the 
initial one and its IP address is removed [enhancement]
   128  When the last subflow is closed without DATA_FIN and msk 
Established, close msk (after a timeout) [enhancement]
    79  allow 'force to MPTCP' mode: BPF [enhancement]
    78  notify the application (userspace) when a subflow is added/removed 
[enhancement]
    77  [gs]etsockopt: forward to new/existing SF [enhancement]
    76  [gs]etsockopt per subflow: BPF [enhancement]
    61  move msk clone after ctx creation [enhancement]
    59  (MP)TFO support [enhancement]
    57  After a few attempts of failed MPTCP, directly fallback to TCP for 
new connections [enhancement]
    48  MP_FASTCLOSE support (send part remaining) [enhancement]
    43  [syzkaller] Change syzkaller to exercise MPTCP inet_diag interface 
[enhancement] [syzkaller]
    41  reduce indirect call usage [enhancement]
    24  Revisit layout of struct mptcp_subflow_context [enhancement]


     Recently closed (since last week)

   268  kmemleak: suspected mem leak in msr_build_context() [bug] 
[selftests] @matttbe
     - Alignment issue in x86 arch code.
   267  Kmemleak report with mptcp_connect [bug] [selftests] @mjmartineau
     - Reference counting issue in userspace PM - fixes squashed
   186  Add netlink command support [enhancement] @kmaloor


FYI: Current Roadmap:
     - Bugs: https://github.com/multipath-tcp/mptcp_net-next/projects/2
     - Current/Coming merge window (5.18): 
https://github.com/multipath-tcp/mptcp_net-next/projects/14
     - For later: 
https://github.com/multipath-tcp/mptcp_net-next/projects/4



Patches to send to netdev:

            - [c492f39068e2] selftests/bpf: Drop duplicate max/min 
definitions (Geliang Tang)

     - Fixes for other trees:

         - [070aefe5b490] x86/pm: Fix false positive kmemleak report in 
msr_build_context() (Matthieu Baerts)
            - Merged: 
https://lore.kernel.org/all/165106098627.4207.9339334897773264387.tip-bot2@tip-bot2/

     - Fixes for -net:
        - None pending

     - Features for net-next:

         - [a03b05f10487] mptcp: really share subflow snd_wnd (Paolo Abeni)
         - [8296c4050142] mptcp: add mib for xmit window sharing (Paolo 
Abeni)
         - [64fa181719da] tcp: allow MPTCP to update the announced window 
(Paolo Abeni)
         - [1efb25c8f2f9] mptcp: never shrink offered window (Paolo Abeni)
         - [156b6c6a2b62] mptcp: add more offered MIBs counter (Paolo 
Abeni)
            - Ready to send

     - Features for net-next (next):

         - [8053fb8d49ed] selftests: mptcp: add MP_FAIL reset testcase 
(Geliang Tang)
           - Two pending issues: open issue regarding multiple MP_FAIL in 
self test, and dropping of bad data.
           - Prioritize MP_FAIL fixes
           - Follow up on ML to make sure Geliang is not blocked [Mat]

         - [89e7a9a44cd0] mptcp: bypass in-kernel PM restrictions for 
non-kernel PMs (Kishen Maloor)
         - [9c644de3762e] selftests: mptcp: ADD_ADDR echo test with missing 
userspace daemon (Mat Martineau)
         - [a7fcc034b220] mptcp: store remote id from MP_JOIN SYN/ACK in 
local ctx (Kishen Maloor)
         - [cab49bddc766] mptcp: reflect remote port (not 0) in ANNOUNCED 
events (Kishen Maloor)
         - [cd2b6a1ea78b] mptcp: establish subflows from either end of 
connection (Kishen Maloor)
         - [8ddb114e8dc8] mptcp: expose server_side attribute in MPTCP 
netlink events (Kishen Maloor)
            - Second batch, next to be sent after series already posted to 
netdev is merged.

         - [05854bd213ba] mptcp: allow ADD_ADDR reissuance by userspace PMs 
(Kishen Maloor)
         - [a7c6b9cb5ac9] mptcp: handle local addrs announced by userspace 
PMs (Kishen Maloor)
         - [d549cb7f5cbd] mptcp: read attributes of addr entries managed by 
userspace PMs (Kishen Maloor)
         - [9be55418a31f] mptcp: netlink: split mptcp_pm_parse_addr into 
two functions (Florian Westphal)
         - [08f35a937818] mptcp: netlink: Add MPTCP_PM_CMD_ANNOUNCE (Kishen 
Maloor)
         - [55262999192a] selftests: mptcp: support MPTCP_PM_CMD_ANNOUNCE 
(Kishen Maloor)
         - [141bd3b6ad7f] mptcp: netlink: Add MPTCP_PM_CMD_REMOVE (Kishen 
Maloor)
         - [6e543b5ba6f9] selftests: mptcp: support MPTCP_PM_CMD_REMOVE 
(Kishen Maloor)
         - [1d835bd5681f] mptcp: netlink: allow userspace-driven subflow 
establishment (Florian Westphal)
         - [3b08c6d29698] selftests: mptcp: support 
MPTCP_PM_CMD_SUBFLOW_CREATE (Kishen Maloor)
         - [c28e52832aeb] selftests: mptcp: support 
MPTCP_PM_CMD_SUBFLOW_DESTROY (Kishen Maloor)
         - [15ba7ea78c85] selftests: mptcp: capture netlink events (Kishen 
Maloor)
         - [05f43c2e5606] selftests: mptcp: create listeners to receive 
MPJs (Kishen Maloor)
         - [3402c68b325e] selftests: mptcp: functional tests for the 
userspace PM type (Kishen Maloor)
            - Third userspace PM batch

     - Features for other trees:

         - [ba947123d3d7] bpf: expose is_mptcp flag to bpf_tcp_sock 
(Nicolas Rybowski)
         - [2819631b4882] bpf: add bpf_skc_to_mptcp_sock_proto (Geliang 
Tang)
         - [e9620c459676] selftests: bpf: add MPTCP test base (Nicolas 
Rybowski)
         - [0ecdd13945c8] selftests: bpf: test bpf_skc_to_mptcp_sock 
(Geliang Tang)
         - [bbcf53e23497] selftests: bpf: verify token of struct mptcp_sock 
(Geliang Tang)
         - [c8c1887cd0fe] selftests: bpf: verify ca_name of struct 
mptcp_sock (Geliang Tang)
         - [702b1056a80c] selftests: bpf: verify first of struct mptcp_sock 
(Geliang Tang)
           - Might be ready for v2 - Mat will upstream.


Extra tests:
     - news about Syzkaller? (Christoph / Mat):
         - Nothing new related to MPTCP.

     - news about interop with mptcp.org/other stacks? (Christoph):
         - /

     - news about Intel's kbuild? (Mat):
         - No builds since 24 April, will check with kbuild team

     - packetdrill (Davide):
         - Nothing new

     - Patchew (Davide):
         - Network outage a few days ago, did not affect MPTCP. Dashboard 
looks ok now.

     - CI (Matth):
         - No changes

     - Other test issue
     - Paolo has observed some issues with subflow accounting, where the PM 
thinks there are more active subflows than there really are. Only affects 
in-kernel PM.


LPC:
     - https://lpc.events/event/16/abstracts/
     - Considering a talk proposal for MPTCP userspace path manager. (Mat)
     - Netdev conf also expected in Nov '22 (Lisbon?), likely not 
conflicting with IETF in London.


Next meeting:
     - Thursday, the 5th of May.
     - 15:00 UTC (8am PST, 4pm CET, 11am CST)
     - Still open to everyone!
     - https://annuel2.framapad.org/p/mptcp_upstreaming_20220505


--
Mat Martineau
Intel

Re: [Weekly meetings] MoM - 28th of April 2022 (MP_FAIL followup)

[Mat Martineau]

On Thu, 28 Apr 2022, Mat Martineau wrote:

>    - Features for net-next (next):
>
>        - [8053fb8d49ed] selftests: mptcp: add MP_FAIL reset testcase 
> (Geliang Tang)
>          - Two pending issues: open issue regarding multiple MP_FAIL in self 
> test, and dropping of bad data.
>          - Prioritize MP_FAIL fixes
>          - Follow up on ML to make sure Geliang is not blocked [Mat]
>

Geliang -

Were the two issues to resolve before upstreaming "selftests: mptcp: add 
MP_FAIL reset testcase" these:

1. https://github.com/multipath-tcp/mptcp_net-next/issues/265

2. Waiting for feedback on how to drop data on checksum failure.


For #1, Paolo has been making a lot of progress there. Since the bug is 
tracked, it will be clear when it is closed and no longer blocks 
upstreaming


For #2, do you have questions that I did not address in the "Infinite 
mapping and dropping data" message on the list (30-Mar-2022):

https://lore.kernel.org/mptcp/5e9cc0c5-1514-a198-5191-a765a67b9de4@linux.intel.com/#r

My conclusion there was that we don't need to drop data on checksum 
failure:

"""
I don't think it's necessary to discard any incoming data to
handle this single-subflow fallback scenario. As Matthieu mentioned in the
meeting, the checksum failure tells us that there was either MPTCP header
corruption or a mismatch between the data payload and the MPTCP mapping -
it is not an indication that the data was "bad". If the incoming data that
is read from the MPTCP socket matches the subflow's reassembled TCP
stream, then fallback was successful.
"""

Was that your concern, or is there a different issue?


Thanks!

--
Mat Martineau
Intel

Re: [Weekly meetings] MoM - 28th of April 2022 (MP_FAIL followup)

[Geliang Tang]

Hi Mat,

Mat Martineau <mathew.j.martineau@linux.intel.com> 于2022年5月4日周三 04:41写道：

>
>
> On Thu, 28 Apr 2022, Mat Martineau wrote:
>
> >    - Features for net-next (next):
> >
> >        - [8053fb8d49ed] selftests: mptcp: add MP_FAIL reset testcase
> > (Geliang Tang)
> >          - Two pending issues: open issue regarding multiple MP_FAIL in self
> > test, and dropping of bad data.
> >          - Prioritize MP_FAIL fixes
> >          - Follow up on ML to make sure Geliang is not blocked [Mat]
> >

With Paolo's commit "net/sched: act_pedit: really ensure the skb is
writable", all pending issues about MP_FAIL have been solved.

>
> Geliang -
>
> Were the two issues to resolve before upstreaming "selftests: mptcp: add
> MP_FAIL reset testcase" these:
>
> 1. https://github.com/multipath-tcp/mptcp_net-next/issues/265
>
> 2. Waiting for feedback on how to drop data on checksum failure.
>
>
> For #1, Paolo has been making a lot of progress there. Since the bug is
> tracked, it will be clear when it is closed and no longer blocks
> upstreaming
>

Now all these '2 MP_FAIL[s]' issues has been fixed:

095 MP_FAIL MP_RST                       syn[ ok ] - synack[ ok ] - ack[ ok ]
                                         sum[ ok ] - csum  [ ok ]
                                         ftx[fail] got 2 MP_FAIL[s] TX
expected 1
 - failrx[fail] got 2 MP_FAIL[s] RX expected 1
                                         rtx[fail] got 2 MP_RST[s] TX expected 1
 - rstrx [fail] got 2 MP_RST[s] RX expected 1
                                         itx[ ok ] - infirx[ ok ]

No '2 MP_FAIL[s]' issues in my test anymore.


But I still got another type of failures (0 MP_FAIL[s]) in my test:

002 MP_FAIL MP_RST: 5 corrupted pkts     syn[ ok ] - synack[ ok ] - ack[ ok ]
                                         sum[ ok ] - csum  [ ok ]
                                         ftx[ ok ] - failrx[fail] got
0 MP_FAIL[s] RX expected 1
                                         rtx[ ok ] - rstrx [fail] got
0 MP_RST[s] RX expected 1
                                         itx[ ok ] - infirx[ ok ]

It seems MP_FAIL is lost in some cases. MP_FAIL is sent, but not received.

I don't know whether Paolo is also looking at this 0 MP_FAIL[s] issue,
or whether it is necessary to open a new issue on Githut to track this
issue.

>
> For #2, do you have questions that I did not address in the "Infinite
> mapping and dropping data" message on the list (30-Mar-2022):
>
> https://lore.kernel.org/mptcp/5e9cc0c5-1514-a198-5191-a765a67b9de4@linux.intel.com/#r
>
> My conclusion there was that we don't need to drop data on checksum
> failure:
>
> """
> I don't think it's necessary to discard any incoming data to
> handle this single-subflow fallback scenario. As Matthieu mentioned in the
> meeting, the checksum failure tells us that there was either MPTCP header
> corruption or a mismatch between the data payload and the MPTCP mapping -
> it is not an indication that the data was "bad". If the incoming data that
> is read from the MPTCP socket matches the subflow's reassembled TCP
> stream, then fallback was successful.
> """
>
> Was that your concern, or is there a different issue?

Yes, all MP_FAIL dropping data issues are solved now. This 'singe
subflow and dropping data' issue you mentioned is solved.

Also, the 'multiple subflows and dropping data' issue is fixed by
Paolo's commit too:

Created /tmp/tmp.iTuFZCucnH (size 1024 KB) containing data sent by client
Created /tmp/tmp.YTN6JELVhq (size 1024 KB) containing data sent by server
002 MP_FAIL MP_RST: 1 corrupted pkts     syn[ ok ] - synack[ ok ] - ack[ ok ]
                                         sum[ ok ] - csum  [ ok ]
                                         ftx[ ok ] - failrx[ ok ]
                                         rtx[ ok ] - rstrx [ ok ]
                                         itx[ ok ] - infirx[ ok ]

Now no inverted byte is received in the MP_FAIL MP_RST test anymore.

Thanks,
-Geliang

>
>
> Thanks!
>
> --
> Mat Martineau
> Intel
>

Re: [Weekly meetings] MoM - 28th of April 2022 (MP_FAIL followup)

[Paolo Abeni]

On Wed, 2022-05-04 at 10:44 +0800, Geliang Tang wrote:
> Mat Martineau <mathew.j.martineau@linux.intel.com> 于2022年5月4日周三 04:41写道：
> 
> > 
> > 
> > On Thu, 28 Apr 2022, Mat Martineau wrote:
> > 
> > >    - Features for net-next (next):
> > > 
> > >        - [8053fb8d49ed] selftests: mptcp: add MP_FAIL reset testcase
> > > (Geliang Tang)
> > >          - Two pending issues: open issue regarding multiple MP_FAIL in self
> > > test, and dropping of bad data.
> > >          - Prioritize MP_FAIL fixes
> > >          - Follow up on ML to make sure Geliang is not blocked [Mat]
> > > 
> 
> With Paolo's commit "net/sched: act_pedit: really ensure the skb is
> writable", all pending issues about MP_FAIL have been solved.
> 
> > 
> > Geliang -
> > 
> > Were the two issues to resolve before upstreaming "selftests: mptcp: add
> > MP_FAIL reset testcase" these:
> > 
> > 1. https://github.com/multipath-tcp/mptcp_net-next/issues/265
> > 
> > 2. Waiting for feedback on how to drop data on checksum failure.
> > 
> > 
> > For #1, Paolo has been making a lot of progress there. Since the bug is
> > tracked, it will be clear when it is closed and no longer blocks
> > upstreaming
> > 
> 
> Now all these '2 MP_FAIL[s]' issues has been fixed:
> 
> 095 MP_FAIL MP_RST                       syn[ ok ] - synack[ ok ] - ack[ ok ]
>                                          sum[ ok ] - csum  [ ok ]
>                                          ftx[fail] got 2 MP_FAIL[s] TX
> expected 1
>  - failrx[fail] got 2 MP_FAIL[s] RX expected 1
>                                          rtx[fail] got 2 MP_RST[s] TX expected 1
>  - rstrx [fail] got 2 MP_RST[s] RX expected 1
>                                          itx[ ok ] - infirx[ ok ]
> 
> No '2 MP_FAIL[s]' issues in my test anymore.
> 
> 
> But I still got another type of failures (0 MP_FAIL[s]) in my test:
> 
> 002 MP_FAIL MP_RST: 5 corrupted pkts     syn[ ok ] - synack[ ok ] - ack[ ok ]
>                                          sum[ ok ] - csum  [ ok ]
>                                          ftx[ ok ] - failrx[fail] got
> 0 MP_FAIL[s] RX expected 1
>                                          rtx[ ok ] - rstrx [fail] got
> 0 MP_RST[s] RX expected 1
>                                          itx[ ok ] - infirx[ ok ]
> 
> It seems MP_FAIL is lost in some cases. MP_FAIL is sent, but not received.
> 
> I don't know whether Paolo is also looking at this 0 MP_FAIL[s] issue,
> or whether it is necessary to open a new issue on Githut to track this
> issue.

[browsers!] I'm not working on it right now. I can have a look at the
end of the week or the next one, if the issue is still pending.

Meanwhile, I suggest to track it with an individual/specific gitlab
issue - to collect all the relevant info in a single place.

Cheers,

Paolo

Re: [Weekly meetings] MoM - 28th of April 2022 (MP_FAIL followup)

[Geliang Tang]

Paolo Abeni <pabeni@redhat.com> 于2022年5月4日周三 22:05写道：
>
> On Wed, 2022-05-04 at 10:44 +0800, Geliang Tang wrote:
> > Mat Martineau <mathew.j.martineau@linux.intel.com> 于2022年5月4日周三 04:41写道：
> >
> > >
> > >
> > > On Thu, 28 Apr 2022, Mat Martineau wrote:
> > >
> > > >    - Features for net-next (next):
> > > >
> > > >        - [8053fb8d49ed] selftests: mptcp: add MP_FAIL reset testcase
> > > > (Geliang Tang)
> > > >          - Two pending issues: open issue regarding multiple MP_FAIL in self
> > > > test, and dropping of bad data.
> > > >          - Prioritize MP_FAIL fixes
> > > >          - Follow up on ML to make sure Geliang is not blocked [Mat]
> > > >
> >
> > With Paolo's commit "net/sched: act_pedit: really ensure the skb is
> > writable", all pending issues about MP_FAIL have been solved.
> >
> > >
> > > Geliang -
> > >
> > > Were the two issues to resolve before upstreaming "selftests: mptcp: add
> > > MP_FAIL reset testcase" these:
> > >
> > > 1. https://github.com/multipath-tcp/mptcp_net-next/issues/265
> > >
> > > 2. Waiting for feedback on how to drop data on checksum failure.
> > >
> > >
> > > For #1, Paolo has been making a lot of progress there. Since the bug is
> > > tracked, it will be clear when it is closed and no longer blocks
> > > upstreaming
> > >
> >
> > Now all these '2 MP_FAIL[s]' issues has been fixed:
> >
> > 095 MP_FAIL MP_RST                       syn[ ok ] - synack[ ok ] - ack[ ok ]
> >                                          sum[ ok ] - csum  [ ok ]
> >                                          ftx[fail] got 2 MP_FAIL[s] TX
> > expected 1
> >  - failrx[fail] got 2 MP_FAIL[s] RX expected 1
> >                                          rtx[fail] got 2 MP_RST[s] TX expected 1
> >  - rstrx [fail] got 2 MP_RST[s] RX expected 1
> >                                          itx[ ok ] - infirx[ ok ]
> >
> > No '2 MP_FAIL[s]' issues in my test anymore.
> >
> >
> > But I still got another type of failures (0 MP_FAIL[s]) in my test:
> >
> > 002 MP_FAIL MP_RST: 5 corrupted pkts     syn[ ok ] - synack[ ok ] - ack[ ok ]
> >                                          sum[ ok ] - csum  [ ok ]
> >                                          ftx[ ok ] - failrx[fail] got
> > 0 MP_FAIL[s] RX expected 1
> >                                          rtx[ ok ] - rstrx [fail] got
> > 0 MP_RST[s] RX expected 1
> >                                          itx[ ok ] - infirx[ ok ]
> >
> > It seems MP_FAIL is lost in some cases. MP_FAIL is sent, but not received.
> >
> > I don't know whether Paolo is also looking at this 0 MP_FAIL[s] issue,
> > or whether it is necessary to open a new issue on Githut to track this
> > issue.
>
> [browsers!] I'm not working on it right now. I can have a look at the
> end of the week or the next one, if the issue is still pending.

Hi Paolo, with your new commit, selftests: mptcp: fix MP_FAIL
test-case, v2,  I didn't find this issue again in my test.

>
> Meanwhile, I suggest to track it with an individual/specific gitlab
> issue - to collect all the relevant info in a single place.
>
> Cheers,
>
> Paolo
>