Re: [myhdl-list] Constant ints in interfaces

[Keerthan JC <jck...@gm...>]

This is a bug, not a limitation. Please post additional testcases to the
issue I just created. https://github.com/jandecaluwe/myhdl/issues/33

I will look into this later this week.

On Mon, Mar 16, 2015 at 7:18 AM, SHEN Chen <she...@co...> wrote:

> Hi all,
>
> I encountered similar problem recently. This is a generic problem in
> the current implementation of interface.
>
> In your code, the following line is critical:
>          tiWIDTH_FPN = self.WIDTH_FPN
>
> With this line, Python creates a local variable named "tiWIDTH_FPN".
>
> In the previous versions of MyHDL, during conversion/elaboration, we
> put all local variables to "symdict".
> I found two places where symdict is constructed, in _extractHier and in
> _analyzeGens:
>    -
>
> https://github.com/jandecaluwe/myhdl/blob/master/myhdl/_extractHierarchy.py#L317
>    -
>
> https://github.com/jandecaluwe/myhdl/blob/master/myhdl/conversion/_analyze.py#L149
>
> Unfortunately, in the current implementation of interface, I couldn't
> find a place where self.WIDTH_FPN being added to symdict.
>
> I recently found that, if a signal isn't included in symdict, it will
> appear as constant in converted code.
> This is the origin of the remaining caveat I mentioned in pull request
> 21 (https://github.com/jandecaluwe/myhdl/pull/21).
>
> One workaround is simply to create a local copy of all signals in the
> interface.
> In my project, I even created an Interface base class, and defined an
> signals() member method, so in the logic generators I can write:
>     tiWIDTH_FPN, tiWIDTH_PRNU = self.signals()
>
> Unfortunately, Python does not allow the following:
>     createLocalVars( self.signals() )
> so you will have to hand-write the left-hand-side of the assignment, in
> order to create local variables.
>
>
>
> regards,
> shenchne
>
>
>
>
> On 2015-03-16 18:03, Josy Boelen wrote:
>
> >> class Interface(object): def __init__(self): self.a =
> >> Signal(intbv(0,
> >> min=-1000, max=1000)) self.c = 10 where c gets converted as
> >> interface_name.c. My understanding from the docs is that this should
> >> be
> >> supported, but I might be missing something. I noticed this
> >> previously
> >> with IntEnum,
> >
> > but
> >
> > Henry,
> >
> > I noticed similar behaviour when using a class (as I also have local
> > functions in the class), here is my code:
> > class CorrectionCoefficients(object):
> > ''' a VHDL-record like object '''
> > def __init__(self, WIDTH_PRNU, WIDTH_FPN, p = 0, f = 0):
> > self.WIDTH_FPN = WIDTH_FPN
> > self.WIDTH_PRNU = WIDTH_PRNU
> > self.fpn = Signal(intbv(f)[WIDTH_FPN:])
> > self.prnu = Signal(intbv(p)[WIDTH_PRNU:])
> >
> > def toCorrectionCoefficients(self, v):
> > # need an unique name as 'alias' otherwise the 'self.WIDTH_FPN'
> > shows up as a constant in the VHDL
> > tccWIDTH_FPN = self.WIDTH_FPN
> > @always_comb
> > def tcc():
> > lv = v
> > self.fpn.next = lv[tccWIDTH_FPN:]
> > self.prnu.next = lv[:tccWIDTH_FPN]
> >
> > return tcc
> >
> > def tointbv(self, y):
> > ''' build a flattened vector from the record's elements in
> > ascending order'''
> > # need an unique name as 'alias' otherwise the 'self.WIDTH_FPN'
> > shows up as a constant in the VHDL :)
> > tiWIDTH_FPN = self.WIDTH_FPN
> >
> > @always_comb
> > def ti():
> > y.next[tiWIDTH_FPN:] = self.fpn
> > y.next[:tiWIDTH_FPN] = self.prnu
> >
> > return ti
> >
> > Regards,
> >
> > Josy
> >
> >
>
> ------------------------------------------------------------------------------
> > Dive into the World of Parallel Programming The Go Parallel Website,
> > sponsored
> > by Intel and developed in partnership with Slashdot Media, is your
> > hub
> > for all
> > things parallel software development, from weekly thought leadership
> > blogs to
> > news, videos, case studies, tutorials and more. Take a look and join
> > the
> > conversation now. http://goparallel.sourceforge.net/
> > _______________________________________________
> > myhdl-list mailing list
> > myh...@li...
> > https://lists.sourceforge.net/lists/listinfo/myhdl-list
>
>
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website,
> sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for
> all
> things parallel software development, from weekly thought leadership blogs
> to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> myhdl-list mailing list
> myh...@li...
> https://lists.sourceforge.net/lists/listinfo/myhdl-list
>



-- 
have a nice day
-jck