[MPTCP] Re: Additional part 3 squash suggestion

[MPTCP] Re: Additional part 3 squash suggestion

[Florian Westphal]

[-- Attachment #1: Type: text/plain, Size: 667 bytes --]

Mat Martineau <mathew.j.martineau(a)linux.intel.com> wrote:
> 
> Florian and Matthieu -
> 
> From the meeting notes, it looks like "subflow: place further subflows on
> new 'join_list'" is the commit that you're planning to squash.
> 
> What do you think about also squashing "mptcp: increment MIB counters in a
> few places" in to the previous MIB commit? Seems like our patch count for
> part 3 is still kind of high.

SGTM.  We could change the commit title as follows:

- mptcp: add MIB counter infrastructure
+ mptcp: add and use MIB counter infrastructure

and toss the commit description of the 'increment MIB' patch,
it doesn't add any value.

[MPTCP] Re: Additional part 3 squash suggestion

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 4514 bytes --]

On 23/03/2020 18:07, Matthieu Baerts wrote:
> Hi Paolo,
> 
> On 20/03/2020 02:09, Paolo Abeni wrote:
>> On Thu, 2020-03-19 at 16:24 -0700, Mat Martineau wrote:
>>> What do you think about also squashing "mptcp: increment MIB counters 
>>> in a
>>> few places" in to the previous MIB commit? Seems like our patch count 
>>> for
>>> part 3 is still kind of high.
>>
>> Additionally I think we could submit:
>>
>> "mptcp: re-check dsn before reading from subflow"
>> "sendmsg: truncate source buffer if mptcp sndbuf size was set from 
>> userspace"
>> "subflow: check parent mptcp socket on subflow state change"
>>
>> independently from part 3 (either before or later, as it fit better).
>>
>> Togethar with the planned squashing, that makes 17 patches for part 3,
>> which I hope should be fine.
>>
>> WDYT?
> 
> As agreed on IRC, I just moved these 3 topic/commits to the end of the 
> series:
> 
> - fb09a7a55b86: tg: revert 
> t/subflow-check-parent-mptcp-socket-on-subflow-state-change
> - e8bd34ac0d03: tg: revert 
> t/sendmsg-truncate-source-buffer-if-mptcp-sndbuf-size-was-set-from-userspace 
> 
> - dde8f988d769: tg: now depending on t/mptcp-add-MIB-counter-infrastructure
> - a9dba7031fe1: tg: now on top of 
> t/selftests-add-test-cases-for-MPTCP-MP_JOIN
> - 10ef670b2bbd: Revert "tg: revert 
> t/subflow-check-parent-mptcp-socket-on-subflow-state-change"
> - a44dbe63118d: Revert "tg: revert 
> t/sendmsg-truncate-source-buffer-if-mptcp-sndbuf-size-was-set-from-userspace" 
> 
> - 2574e7b58fa9: tg: now depending on 
> t/sendmsg-truncate-source-buffer-if-mptcp-sndbuf-size-was-set-from-userspace 
> 
> - ec499ad27205..267d3c188c55: result (empty as expected)
> 
> Tests + export is going to be launched soon.

Just added a small fix from Paolo:

- a8a525975bb3: Squash-to: "mptcp: queue data for mptcp level 
retransmission"
- 267d3c188c55..7deebe3e526b: result

And Mat' Signed-off-by:

[t/mptcp-Add-ADD_ADDR-handling 170851a2d026] tg:msg: add sender' 
Signed-off-by
[t/mptcp-Add-handling-of-incoming-MP_JOIN-requests 42c696413619] tg:msg: 
add sender' Signed-off-by
[t/mptcp-Add-handling-of-outgoing-MP_JOIN-requests f56086980bce] tg:msg: 
add sender' Signed-off-by
[t/mptcp-Implement-path-manager-interface-commands 6aae10d4559b] tg:msg: 
add sender' Signed-off-by
[t/mptcp-update-per-unacked-sequence-on-pkt-reception e17ffe398290] 
tg:msg: add sender' Signed-off-by
[t/mptcp-queue-data-for-mptcp-level-retransmission b0062356dfbf] tg:msg: 
add sender' Signed-off-by
[t/mptcp-introduce-MPTCP-retransmission-timer 582861f31a25] tg:msg: add 
sender' Signed-off-by
[t/mptcp-implement-memory-accounting-for-mptcp-rtx-queue db9c40354d23] 
tg:msg: add sender' Signed-off-by
[t/mptcp-allow-partial-cleaning-of-rtx-head-dfrag 1cba9c047d0b] tg:msg: 
add sender' Signed-off-by
[t/mptcp-rework-mptcp_sendmsg_frag-to-accept-optional-dfrag 
e686543c2dde] tg:msg: add sender' Signed-off-by
[t/mptcp-implement-and-use-MPTCP-level-retransmission 65b9b3193cb7] 
tg:msg: add sender' Signed-off-by
[t/mptcp-allow-dumping-subflow-context-to-userspace 39a49d56c932] 
tg:msg: add sender' Signed-off-by
[t/mptcp-add-MIB-counter-infrastructure dc5578e17935] tg:msg: add 
sender' Signed-off-by
[t/mptcp-add-netlink-based-PM 7fc3e2f6d899] tg:msg: add sender' 
Signed-off-by
[t/selftests-add-PM-netlink-functional-tests 17a238a85b83] tg:msg: add 
sender' Signed-off-by
[t/selftests-add-test-cases-for-MPTCP-MP_JOIN 2d607ee0ff03] tg:msg: add 
sender' Signed-off-by
- 7deebe3e526b..515b70745e1a: result (empty as expected)

And a few fixes for checkpatch:

- b425e5f01502: Squash-to: "mptcp: Implement path manager interface 
commands"
- 515b70745e1a..a447c9c5ef68: result

- b4ae646b5fc9: Squash-to: "mptcp: allow partial cleaning of rtx head dfrag"

- 751501f4cf39: Squash-to: "mptcp: implement and use MPTCP-level 
retransmission"
- a447c9c5ef68..36d843bc5ff4: result

- 85b1104827d1: Squash-to: "mptcp: add and use MIB counter infrastructure"
- 36d843bc5ff4..81343107d0f4: result

- 4ccbbc868719: Squash-to: "mptcp: add netlink-based PM"
- 81343107d0f4..67488f21ed90: result

- 31a00e0ee272: Squash-to: "selftests: add PM netlink functional tests"
- 67488f21ed90..c72cbd89741a: result

Tests + export are in progress.

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium

[MPTCP] Re: Additional part 3 squash suggestion

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 1826 bytes --]

Hi Paolo,

On 20/03/2020 02:09, Paolo Abeni wrote:
> On Thu, 2020-03-19 at 16:24 -0700, Mat Martineau wrote:
>> What do you think about also squashing "mptcp: increment MIB counters in a
>> few places" in to the previous MIB commit? Seems like our patch count for
>> part 3 is still kind of high.
> 
> Additionally I think we could submit:
> 
> "mptcp: re-check dsn before reading from subflow"
> "sendmsg: truncate source buffer if mptcp sndbuf size was set from userspace"
> "subflow: check parent mptcp socket on subflow state change"
> 
> independently from part 3 (either before or later, as it fit better).
> 
> Togethar with the planned squashing, that makes 17 patches for part 3,
> which I hope should be fine.
> 
> WDYT?

As agreed on IRC, I just moved these 3 topic/commits to the end of the 
series:

- fb09a7a55b86: tg: revert 
t/subflow-check-parent-mptcp-socket-on-subflow-state-change
- e8bd34ac0d03: tg: revert 
t/sendmsg-truncate-source-buffer-if-mptcp-sndbuf-size-was-set-from-userspace
- dde8f988d769: tg: now depending on t/mptcp-add-MIB-counter-infrastructure
- a9dba7031fe1: tg: now on top of 
t/selftests-add-test-cases-for-MPTCP-MP_JOIN
- 10ef670b2bbd: Revert "tg: revert 
t/subflow-check-parent-mptcp-socket-on-subflow-state-change"
- a44dbe63118d: Revert "tg: revert 
t/sendmsg-truncate-source-buffer-if-mptcp-sndbuf-size-was-set-from-userspace"
- 2574e7b58fa9: tg: now depending on 
t/sendmsg-truncate-source-buffer-if-mptcp-sndbuf-size-was-set-from-userspace
- ec499ad27205..267d3c188c55: result (empty as expected)

Tests + export is going to be launched soon.

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium

[MPTCP] Re: Additional part 3 squash suggestion

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 1286 bytes --]

Hi Mat, Florian,

On 20/03/2020 00:34, Florian Westphal wrote:
> Mat Martineau <mathew.j.martineau(a)linux.intel.com> wrote:
>>
>> Florian and Matthieu -
>>
>>  From the meeting notes, it looks like "subflow: place further subflows on
>> new 'join_list'" is the commit that you're planning to squash.
>>
>> What do you think about also squashing "mptcp: increment MIB counters in a
>> few places" in to the previous MIB commit? Seems like our patch count for
>> part 3 is still kind of high.
> 
> SGTM.  We could change the commit title as follows:
> 
> - mptcp: add MIB counter infrastructure
> + mptcp: add and use MIB counter infrastructure
> 
> and toss the commit description of the 'increment MIB' patch,
> it doesn't add any value.

Thank you for the idea and the instructions!

- 3eccc998b50b: mptcp: increment MIB counters in a few places
- 74ddb090f4a9: tg:msg: MIB counters are now used
- 1347fe98687a: tg: now depend on t/mptcp-add-MIB-counter-infrastructure
- 1f804516b871..c011ac35dcba: result (empty as expected)

Tests + export are in progress.

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium

[MPTCP] Re: Additional part 3 squash suggestion

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 678 bytes --]

On Thu, 2020-03-19 at 16:24 -0700, Mat Martineau wrote:
> What do you think about also squashing "mptcp: increment MIB counters in a 
> few places" in to the previous MIB commit? Seems like our patch count for 
> part 3 is still kind of high.

Additionally I think we could submit:

"mptcp: re-check dsn before reading from subflow" 
"sendmsg: truncate source buffer if mptcp sndbuf size was set from userspace"
"subflow: check parent mptcp socket on subflow state change"

independently from part 3 (either before or later, as it fit better).

Togethar with the planned squashing, that makes 17 patches for part 3,
which I hope should be fine.

WDYT?

Paolo