[U-Boot] [ARM] Status for 2012.03

[Simon Glass]

Hi Marek,

On Fri, Mar 30, 2012 at 7:29 AM, Marek Vasut <marex at denx.de> wrote:
> Dear Simon Glass,
>
>> Hi Marek,
>>
>> On Fri, Mar 30, 2012 at 6:52 AM, Marek Vasut <marex at denx.de> wrote:
>> > Dear Simon Glass,
>> >
>> >> +Marek
>> >>
>> >> Hi Albert,
>> >>
>> >> On Fri, Mar 30, 2012 at 12:08 AM, Albert ARIBAUD
>> >>
>> >> <albert.u.boot at aribaud.net> wrote:
>> >> > Le 27/03/2012 22:18, Albert ARIBAUD a écrit :
>> >> >> Hi all,
>> >> >>
>> >> >>
>> >> >> I have been away from the list quite a few days, and will have a very
>> >> >> tricky time catching up, so with my apologies, I ask people who need
>> >> >> me to apply patches or pull requests for the upcoming release to
>> >> >> please send me a private e-mail and point me to the list message
>> >> >> that I should be acting upon.
>> >> >
>> >> > FWIW, ./MAKEALL arm on current u-boot-arm/master with CS toolchain
>> >> > 2010q1 builds 277 boards, of which none has errors and 167 have
>> >> > warnings.
>> >> >
>> >> > The total count of warnings is 248, distributed as:
>> >> >
>> >> > 4       post.c:503: warning: #warning "Not implemented yet"
>> >> >
>> >> > 79      mmc.c:132: warning: no return statement in function returning
>> >> > non-void
>> >> >
>> >> > 165     ohci-hcd.c:{1307,1310,1313}: warning: dereferencing pointer
>> >> > 'pretmp.NNN' does break strict-aliasing rules
>> >>
>> >> I did create a patch for this - it is marked delegated to Marek.
>> >>
>> >> http://patchwork.ozlabs.org/patch/134600/
>> >
>> > And stupid maintainer didn't look into it ... bad, bad maintainer. Accept
>> > my apology. But wasn't it decided with Mike that this was indeed a
>> > toolchain issue and not all toolchains choke on it?
>> :
>> :-)
>>
>> Yes. It does feel like a broken toolchain (I felt that the original
>> code should be warning free, even when I wrote my patch) but it seems
>> that Albert gets this warning as well. Maybe we should fix it?
>
> I don't see this with ELDK 4.2 or 5.0 or 5.1 ;-)

Nope. It only happens with some toolchains. That said, it blights my
MAKEALL at present.

>
> Maybe fundamental rewrite of that code might help instead of adding bandaids ;-)

Yes I'm sure, or even getting rid of the 3 separate versions of this
file in U-Boot. Hope springs eternal.

Regards,
Simon

>
>> >> > (the last one is actually 3 * 51 warnings)
>> >> >
>> >> > The first one is obvious and should be dealt with, but I consider it
>> >> > not to be a blocker, for only 4 boards.
>> >> >
>> >> > The second one seems straightforward to fix today, if not already
>> >> > submitted.
>> >> >
>> >> > The third one may be an issue, but it could also be a zealous warning
>> >> > from the specific CS toolchain version I am using. Anyway, it does not
>> >> > seem ARM related as such.
>> >>
>> >> Would be nice to get rid of it, but I think it is specific to some
>> >> toolchains.
>> >>
>> >> > Therefore I will issue the ARM pull request right away.
>> >> >
>> >> >
>> >> > Amicalement,
>> >> > --
>> >> > Albert.
>> >> > _______________________________________________
>> >> > U-Boot mailing list
>> >> > U-Boot at lists.denx.de
>> >> > http://lists.denx.de/mailman/listinfo/u-boot
>> >>
>> >> Regards,
>> >> Simon
>> >
>> > Best regards,
>> > Marek Vasut
>>
>> Regards,
>> Simon