Re: [lwip-users] upgrade from 1.4.1 to 2.0.3

[Sylvain Rochet]

Hi Mattia,

On Thu, Mar 08, 2018 at 04:39:03PM +0100, Mattia Settin wrote:
> Hi
> Thanks
> Buh how is possible that: with an operative system sys_timeouts_mbox_fetch 
> call
> sys_check_timeouts() which is the handle timeouts for NO_SYS==1.
> Is the comment of sys_check_timeouts()  wrong ?

Well, I agree the comment could be improved for clarity. When we fixed 
timers for NO_SYS==0 systems sys_check_timeouts was actually set 
static[1] so it wasn't possible to mistakenly call this function. Looks 
like it was changed when mbox handling moved to another file, therefore 
clearing the static modifier.

Sylvain


[1] 
http://git.savannah.nongnu.org/cgit/lwip.git/tree/src/core/timers.c?id=6adeb706a699da10f2841eea3671546407ec413a#n293


signature.asc
Description: Digital signature
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] upgrade from 1.4.1 to 2.0.3

[Mattia Settin]

Hi
Thanks
Buh how is possible that: with an operative system sys_timeouts_mbox_fetch call
sys_check_timeouts() which is the handle timeouts for NO_SYS==1.
Is the comment of sys_check_timeouts()  wrong ?



On Thu, Mar 8, 2018 at 4:24 PM, Sylvain Rochet 
wrote:

> Hi Mattia,
>
> On Thu, Mar 08, 2018 at 09:14:33AM +0100, Mattia Settin wrote:
> > Yes, but sys_now is required even for timeouts for NO_SYS==0
> (LWIP_TIMERS =
> > 1). It this correct ?
>
> This is correct, sys_now() is now necessary for all systems, furthermore
> it should be bound, at best, to a monotonic clock source.
>
> Sylvain
>
> ___
> lwip-users mailing list
> lwip-users@nongnu.org
> https://lists.nongnu.org/mailman/listinfo/lwip-users
>



-- 
Mattia Settin
Software and System Engineer
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] upgrade from 1.4.1 to 2.0.3

[Sylvain Rochet]

Hi Mattia,

On Thu, Mar 08, 2018 at 09:14:33AM +0100, Mattia Settin wrote:
> Yes, but sys_now is required even for timeouts for NO_SYS==0 (LWIP_TIMERS =
> 1). It this correct ?

This is correct, sys_now() is now necessary for all systems, furthermore 
it should be bound, at best, to a monotonic clock source.

Sylvain


signature.asc
Description: Digital signature
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] upgrade from 1.4.1 to 2.0.3

[Mattia Settin]

In addition I don't really get why
with NO_SYS = 0 (with FreeRTOS) sys_timeouts_mbox_fetch call
sys_check_timeouts() which is the handle timeouts for NO_SYS==1 (without OS)
Probably I have an too old port file for FreeRTOS.
Regards
m

On Thu, Mar 8, 2018 at 9:14 AM, Mattia Settin 
wrote:

> Yes, but sys_now is required even for timeouts for NO_SYS==0 (LWIP_TIMERS
> = 1). It this correct ?
>
> On Wed, Mar 7, 2018 at 8:02 PM, goldsi...@gmx.de  wrote:
>
>> On 07.03.2018 17:40, Mattia Settin wrote:
>>
>>> [..]
>>> The question is:
>>> It is now mandatory define/use sys_now() ?
>>>
>>
>> You can do without, but a number of features new to 2.0.x (or improved
>> there) require it. Right now, the list is:
>> - timeouts for NO_SYS==1
>> - LWIP_SO_SNDTIMEO
>> - LWIP_SO_LINGER
>> - lwiperf app
>> - LWIP_TCP_TIMESTAMPS
>> - PPPoS
>>
>> This list can grow longer in the future. As such, we can't ifdef out the
>> prototype for these configs and it's always there. You can try without and
>> you'll get a linker error once you enable code which actually uses it.
>>
>> Simon
>>
>> ___
>> lwip-users mailing list
>> lwip-users@nongnu.org
>> https://lists.nongnu.org/mailman/listinfo/lwip-users
>>
>
>
>
> --
> Mattia Settin
> Software and System Engineer
>
>
>


-- 
Mattia Settin
Software and System Engineer
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] upgrade from 1.4.1 to 2.0.3

[Mattia Settin]

Yes, but sys_now is required even for timeouts for NO_SYS==0 (LWIP_TIMERS =
1). It this correct ?

On Wed, Mar 7, 2018 at 8:02 PM, goldsi...@gmx.de  wrote:

> On 07.03.2018 17:40, Mattia Settin wrote:
>
>> [..]
>> The question is:
>> It is now mandatory define/use sys_now() ?
>>
>
> You can do without, but a number of features new to 2.0.x (or improved
> there) require it. Right now, the list is:
> - timeouts for NO_SYS==1
> - LWIP_SO_SNDTIMEO
> - LWIP_SO_LINGER
> - lwiperf app
> - LWIP_TCP_TIMESTAMPS
> - PPPoS
>
> This list can grow longer in the future. As such, we can't ifdef out the
> prototype for these configs and it's always there. You can try without and
> you'll get a linker error once you enable code which actually uses it.
>
> Simon
>
> ___
> lwip-users mailing list
> lwip-users@nongnu.org
> https://lists.nongnu.org/mailman/listinfo/lwip-users
>



-- 
Mattia Settin
Software and System Engineer
___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users

Re: [lwip-users] upgrade from 1.4.1 to 2.0.3

[goldsi...@gmx.de]

On 07.03.2018 17:40, Mattia Settin wrote:

[..]
The question is:
It is now mandatory define/use sys_now() ?


You can do without, but a number of features new to 2.0.x (or improved 
there) require it. Right now, the list is:

- timeouts for NO_SYS==1
- LWIP_SO_SNDTIMEO
- LWIP_SO_LINGER
- lwiperf app
- LWIP_TCP_TIMESTAMPS
- PPPoS

This list can grow longer in the future. As such, we can't ifdef out the 
prototype for these configs and it's always there. You can try without 
and you'll get a linker error once you enable code which actually uses it.


Simon

___
lwip-users mailing list
lwip-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lwip-users