This is the mail archive of the ecos-maintainers@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: New CPM/DPRAM allocator


On Wed, 2003-03-19 at 04:07, Nick Garnett wrote:
> John Dallaway <jld at ecoscentric dot com> writes:
> 
> > A few of questions:
> > 
> > 1) Does this incompatibilty affect _all_ targets which run both RedBoot and 
> > eCos?
> 
> Just PowerPC QUICC targets, like the MBX, Adder or Viper, for example.
> 
> > 2) Could the new allocator implementation be modified in any way to preserve 
> > compatibility?
> 
> Possibly. But I'll defer to Gary on this one.
> 

The more I look at this, the less I understand why it fails now.  The 
new allocation scheme should not clash - i.e. mixing new eCos with old
RedBoot should be OK.

Perhaps if I can figure out why this is failing, I can propose a change
which will not cause problems.

> > 3) To what extent might the new allocator perturb eCos functionality other 
> > than by a stepwise "it will work or it will be completely broken".
> 
> It shouldn't affect anything else. Failure modes are hard to specify
> since they depend on the contents of what is currently an
> uninitialized location in RAM. Different instances of RedBoot might
> have different values there.
> 
> We might even be able to add a test to the DPRAM code to detect an old
> RedBoot and generate an error message before proceeding.
> 
> > 4) Is a further break in compatibility possible as potential issues with the 
> > new allocator are ironed out?
> 
> Unlikely. The code is only a handful of lines. 
> 
> > 
> > We need to determine how to manage the switch over to the new CPM/DPRAM 
> > allocator smoothly for everyone. From my perspective, it is certainly 
> > unfortunate that this innovation has arrived between 2.0 beta and 2.0 
> > final. Were it not for the break in compatibility, I would vote against 
> > incorporating this change for 2.0 final since the change will (to a certain 
> > extent) invalidate beta testing by the net community. However, it would 
> > also be undesirable for 2.0 final RedBoot to be incompatible with anonCVS 
> > eCos from day 1. Users are likely to want to stick with a release version 
> > of the firmware while experimenting with eCos from anonCVS.
> > 
> 
> It only affects a handful of boards: MBX, VIPER, ADDER, TS1000. QUICC2
> targets like the tS6 and VADS are not affected, and non-QUICC MPC8xx
> targets like the CMA28x and FADS boards are not affected.
> 
> -- 
> Nick Garnett                    eCos Kernel Architect
> http://www.ecoscentric.com/     The eCos and RedBoot experts

-- 
------------------------------------------------------------
Gary Thomas                 |
MLB Associates              |  Consulting for the
+1 (970) 229-1963           |    Embedded world
http://www.mlbassoc.com/    |
email: <gary at mlbassoc dot com>  |
gpg: http://www.chez-thomas.org/gary/gpg_key.asc
------------------------------------------------------------


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]