Re: [MPTCP] Restructuring of TCP code for implementing MPTCP

Re: [MPTCP] Restructuring of TCP code for implementing MPTCP

[Christoph Paasch]

[-- Attachment #1: Type: text/plain, Size: 1502 bytes --]

On 09/03/18 - 14:19:54, Mat Martineau wrote:
> 
> Hi Rao -
> 
> On Thu, 22 Feb 2018, Rao Shoaib wrote:
> 
> > Hi Folks,
> > 
> > When we met with David Miller and Eric Dumazet about a year ago to
> > discuss how to include MPTCP in the main Linux distribution, they wanted
> > to have some code that could be evaluated and used at the basis for
> > discussion.
> > 
> > The patches that I have submitted are for starting that discussion only
> > and nothing else. If the approach is acceptable we can do more cleanup,
> > If not we will have a discussion and hopefully agree on something.
> > 
> > I would like to send the patches, perhaps a week from now to the netdev
> > mailing list to start a discussion there and come up with a plan to
> > implement MPTCP.
> > 
> > Comments/Suggestions ?
> > 
> 
> I saw that you posted some questions on netdev this week and got some
> maintainer feedback: https://marc.info/?t=152039384300002&r=1&w=2
> 
> David mentions:
> 
> """
> Indirect calls are terrible and we are now looking at ways in which
> we can remove them from as many parts of the networking as possible.
> """
> 
> It's not immediately clear which techniques are being looked at (and both
> Eric and David did not elaborate when you followed up). I'm trying to find
> out which alternatives he's referring to - anyone else seen information on
> this?

I haven't seen/heard about anything, but also would love to know more about
it.



Christoph

Re: [MPTCP] Restructuring of TCP code for implementing MPTCP

[Rao Shoaib]

[-- Attachment #1: Type: text/plain, Size: 2780 bytes --]



On 03/09/2018 03:01 PM, Rao Shoaib wrote:
>
>
> On 03/09/2018 02:19 PM, Mat Martineau wrote:
>>
>> Hi Rao -
>>
>> On Thu, 22 Feb 2018, Rao Shoaib wrote:
>>
>>> Hi Folks,
>>>
>>> When we met with David Miller and Eric Dumazet about a year ago to 
>>> discuss how to include MPTCP in the main Linux distribution, they 
>>> wanted to have some code that could be evaluated and used at the 
>>> basis for discussion.
>>>
>>> The patches that I have submitted are for starting that discussion 
>>> only and nothing else. If the approach is acceptable we can do more 
>>> cleanup, If not we will have a discussion and hopefully agree on 
>>> something.
>>>
>>> I would like to send the patches, perhaps a week from now to the 
>>> netdev mailing list to start a discussion there and come up with a 
>>> plan to implement MPTCP.
>>>
>>> Comments/Suggestions ?
>>>
>>
>> I saw that you posted some questions on netdev this week and got some 
>> maintainer feedback: https://marc.info/?t=152039384300002&r=1&w=2
>>
>> David mentions:
>>
>> """
>> Indirect calls are terrible and we are now looking at ways in which
>> we can remove them from as many parts of the networking as possible.
>> """
>>
> Yes against all objections I did post to net-dev and got some very 
> informative answers. David is not allowing indirect function calls 
> because of the security issues. You should see the comments that Linus 
> has made about the fixes and folks asking for an alternative as they 
> are heavily used.
>
> https://www.theregister.co.uk/2018/01/22/intel_spectre_fix_linux/
Sorry wrong url -- See

https://news.ycombinator.com/item?id=16066968

Edit: To give a bit more background, predicting indirect calls and 
speculating into them is absolutely critical to getting good performance 
from OOP-ish code which tends to use them a lot (virtual functions in 
C++, function pointers in C). *The Linux kernel is (thankfully?) not 
very OOP-ish, but it does rely on indirect calls (function pointers) 
extensively.

*Rao*.
*
>
>> It's not immediately clear which techniques are being looked at (and 
>> both Eric and David did not elaborate when you followed up). I'm 
>> trying to find out which alternatives he's referring to - anyone else 
>> seen information on this?
> Nope. It's not clear what the solution will be, but that should not 
> have any impact on the patch that we submitted. As a simple fix we can 
> add if/else as there are only two ways to go.
>
> Please see my email and comment.
>
> Rao.
>
>> Regards,
>>
>> -- 
>> Mat Martineau
>> Intel OTC
>
> _______________________________________________
> mptcp mailing list
> mptcp(a)lists.01.org
> https://lists.01.org/mailman/listinfo/mptcp


[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 4701 bytes --]

Re: [MPTCP] Restructuring of TCP code for implementing MPTCP

[Rao Shoaib]

[-- Attachment #1: Type: text/plain, Size: 2052 bytes --]



On 03/09/2018 02:19 PM, Mat Martineau wrote:
>
> Hi Rao -
>
> On Thu, 22 Feb 2018, Rao Shoaib wrote:
>
>> Hi Folks,
>>
>> When we met with David Miller and Eric Dumazet about a year ago to 
>> discuss how to include MPTCP in the main Linux distribution, they 
>> wanted to have some code that could be evaluated and used at the 
>> basis for discussion.
>>
>> The patches that I have submitted are for starting that discussion 
>> only and nothing else. If the approach is acceptable we can do more 
>> cleanup, If not we will have a discussion and hopefully agree on 
>> something.
>>
>> I would like to send the patches, perhaps a week from now to the 
>> netdev mailing list to start a discussion there and come up with a 
>> plan to implement MPTCP.
>>
>> Comments/Suggestions ?
>>
>
> I saw that you posted some questions on netdev this week and got some 
> maintainer feedback: https://marc.info/?t=152039384300002&r=1&w=2
>
> David mentions:
>
> """
> Indirect calls are terrible and we are now looking at ways in which
> we can remove them from as many parts of the networking as possible.
> """
>
Yes against all objections I did post to net-dev and got some very 
informative answers. David is not allowing indirect function calls 
because of the security issues. You should see the comments that Linus 
has made about the fixes and folks asking for an alternative as they are 
heavily used.

https://www.theregister.co.uk/2018/01/22/intel_spectre_fix_linux/

> It's not immediately clear which techniques are being looked at (and 
> both Eric and David did not elaborate when you followed up). I'm 
> trying to find out which alternatives he's referring to - anyone else 
> seen information on this?
Nope. It's not clear what the solution will be, but that should not have 
any impact on the patch that we submitted. As a simple fix we can add 
if/else as there are only two ways to go.

Please see my email and comment.

Rao.

> Regards,
>
> -- 
> Mat Martineau
> Intel OTC

Re: [MPTCP] Restructuring of TCP code for implementing MPTCP

[Mat Martineau]

[-- Attachment #1: Type: text/plain, Size: 1325 bytes --]


Hi Rao -

On Thu, 22 Feb 2018, Rao Shoaib wrote:

> Hi Folks,
>
> When we met with David Miller and Eric Dumazet about a year ago to discuss 
> how to include MPTCP in the main Linux distribution, they wanted to have some 
> code that could be evaluated and used at the basis for discussion.
>
> The patches that I have submitted are for starting that discussion only and 
> nothing else. If the approach is acceptable we can do more cleanup, If not we 
> will have a discussion and hopefully agree on something.
>
> I would like to send the patches, perhaps a week from now to the netdev 
> mailing list to start a discussion there and come up with a plan to implement 
> MPTCP.
>
> Comments/Suggestions ?
>

I saw that you posted some questions on netdev this week and got some 
maintainer feedback: https://marc.info/?t=152039384300002&r=1&w=2

David mentions:

"""
Indirect calls are terrible and we are now looking at ways in which
we can remove them from as many parts of the networking as possible.
"""

It's not immediately clear which techniques are being looked at (and both 
Eric and David did not elaborate when you followed up). I'm trying to find 
out which alternatives he's referring to - anyone else seen information on 
this?

Regards,

--
Mat Martineau
Intel OTC

[MPTCP] Restructuring of TCP code for implementing MPTCP

[Rao Shoaib]

[-- Attachment #1: Type: text/plain, Size: 671 bytes --]

Hi Folks,

When we met with David Miller and Eric Dumazet about a year ago to 
discuss how to include MPTCP in the main Linux distribution, they wanted 
to have some code that could be evaluated and used at the basis for 
discussion.

The patches that I have submitted are for starting that discussion only 
and nothing else. If the approach is acceptable we can do more cleanup, 
If not we will have a discussion and hopefully agree on something.

I would like to send the patches, perhaps a week from now to the netdev 
mailing list to start a discussion there and come up with a plan to 
implement MPTCP.

Comments/Suggestions ?

Thanks,

Shoaib