Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [802.3_4PPOE] Detection cleared by SISM state diagrams



Hi Yair,

Yes, that also needs to happen.
However, we also need to prevent the ENTRY_* states from writing into
pd_4pair_cand continuously. They only need to do that when sism=TRUE.

Kind regards,

Lennart

On Tue, 2017-10-17 at 00:23 +0000, Yair Darshan wrote:
> Hi Lennart,
>  
> Isn't it more reasonable to set the sig_pri and sig_sec to FALSE in
> the top state machine at idle state?
> This will allow us to delete the if conditions from ENTRY_PRI and
> ENTRI_SEC and will also reset the signals for the single signature
> state machine, something that is not happening currently.
> (See below)
> Yair
>  
> 
>  
>  
>  
> -----Original Message-----
> From: Chris Bullock (bullock) [mailto:bullock@xxxxxxxxx] ;
> Sent: Saturday, October 14, 2017 2:19 AM
> To: STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx
> Subject: Re: [802.3_4PPOE] Detection cleared by SISM state diagrams
>  
> EXTERNAL EMAIL
>  
>  
> Hi Lennart,
>  
> Introducing a state before ENTRY_PRI and ENTRY_SEC will do the same
> thing.  I will support either option.  I will defer to you as to
> which option you think looks better.
>  
> Thanks,
> Chris
>  
> -----Original Message-----
> From: Lennart Yseboodt [mailto:lennartyseboodt@xxxxxxxxx]
> Sent: Friday, October 13, 2017 2:45 AM
> To: Chris Bullock (bullock) <bullock@xxxxxxxxx>; STDS-802-3-4PPOE@LIS
> TSERV.IEEE.ORG
> Subject: Re: [802.3_4PPOE] Detection cleared by SISM state diagrams
>  
> Hi Chris,
>  
> I like the concept to avoid continuously executing this block of code
> for a part of the state diagram that should be sleeping.
>  
> We can move it to SISM_START, but editorially that moves another
> chunk of text to a diagram that is already very full.
>  
> What if we introduce a state before ENTRY_PRI with no actions, and
> that exists to ENTRY_PRI when 'sism' ?
>  
> Does the same thing I think...
>  
> Lennart
>  
> On Wed, 2017-10-11 at 03:15 +0000, Chris Bullock (bullock) wrote:
> > We could take all of the assignments that are currently in the
> > ENTRY_PRI and ENTRY_SEC states, and move them to the SISM_START
> state.
> >
> > Thanks,
> > Chris
> >
> > From: Heath Stewart [mailto:00000855853231d4-dmarc-request@xxxxxxxx
> ]
> > Sent: Monday, October 09, 2017 11:23 AM
> > To: STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx
> > Subject: Re: [802.3_4PPOE] Detection cleared by SISM state diagrams
> >
> > Change to
> > if (sism & (CC_DET_SEQ != 2)) then
> >
> > -Heath
> >
> > On Mon, Oct 9, 2017 at 2:59 AM, Lennart Yseboodt <lennartyseboodt@g
> ma
> > il.com> wrote:
> > Hi,
> >
> > As of this cycle we have included the SISM state diagrams in our
> > simulation.
> >
> > When a single-signature is connected, ENTRY_PRI (and SEC) are
> executed
> > continuously because "!sism" is TRUE.
> >
> >
> > This has the effect to force sig_pri (and sig_sec) to 'invalid'
> > continuously, thereby breaking the main state diagram.
> > I don't know why that statement is there.
> >
> > Can we remove the IF statement and sig_pri <= invalid ?
> >
> > Kind regards,
> >
> > Lennart
> >
> >
> >
> > --
> > Heath Stewart
> > Design Center Manager, Mixed Signal
> >
> > Office   (805) 560-7658
> > Mobile  (805) 895-0499
> > Websites      analog.com, linear.com
> >
> > Linear Technology is now part of Analog Devices.  Learn more.
> >
>