[PATCH 12/15] configs: add qemu_arm64_lwip_defconfig

[Tom Rini]

On Fri, May 24, 2024 at 05:30:45PM +0200, Jerome Forissier wrote:
> 
> 
> On 5/24/24 17:08, Tom Rini wrote:
> > On Fri, May 24, 2024 at 12:09:33PM +0200, Jerome Forissier wrote:
> >>
> >>
> >> On 5/23/24 17:02, Tom Rini wrote:
> >>> On Wed, May 22, 2024 at 06:00:12PM +0200, Jerome Forissier wrote:
> >>>
> >>>> Add qemu_arm64_lwip_defconfig which was created from
> >>>> qemu_arm64_defconfig with CONFIG_NET_LWIP enabled.
> >>>>
> >>>> Signed-off-by: Jerome Forissier <jerome.forissier at linaro.org>
> >>>> ---
> >>>>  configs/qemu_arm64_lwip_defconfig | 70 +++++++++++++++++++++++++++++++
> >>>>  1 file changed, 70 insertions(+)
> >>>>  create mode 100644 configs/qemu_arm64_lwip_defconfig
> >>>>
> >>>> diff --git a/configs/qemu_arm64_lwip_defconfig b/configs/qemu_arm64_lwip_defconfig
> >>>> new file mode 100644
> >>>> index 0000000000..c8dadbce37
> >>>> --- /dev/null
> >>>> +++ b/configs/qemu_arm64_lwip_defconfig
> >>>> @@ -0,0 +1,70 @@
> >>>> +CONFIG_ARM=y
> >>>> +CONFIG_POSITION_INDEPENDENT=y
> >>>> +CONFIG_ARCH_QEMU=y
> >>>
> >>> Setting aside that I would like lwip to default to y, at least in the
> >>> first few iterations (so that testing is easier),
> >>
> >> I could certainly do this. By creating a new defconfig, my goal was to
> >> minimize the impacts on existing use cases.
> > 
> > Yes but then it makes it harder to out of the box see what the deltas
> > are.
> > 
> >>> this should be using
> >>> the #include method rather than duplicating the whole config.
> >>
> >> I've done that for v2. Let me know if you still prefer I change
> >> qemu_arm64_defconfig instead.
> > 
> > This is fine for v2, yeah. Once that's posted I'll grab everything and:
> > - Throw it at the small HW lab I have
> > - Enable it by default for everyone and throw it at CI
> 
> At this point you can expect quite a few failures (unresolved symbols to
> begin with) because so many parts of the network-related code peek & poke
> various parts of the net implementation without formally depending on NET
> (i.e., !NET_LWIP). Some more work is clearly needed to port to lwIP, how
> much is probably something the CI run will help understand.

OK, thanks for setting expectations. And FYI if you hadn't seen:
https://docs.u-boot.org/en/latest/develop/ci_testing.html you can
trigger them yourself, but Azure sometimes just has issues if we get a
slow runner sometimes.

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 659 bytes
Desc: not available
URL: <https://lists.denx.de/pipermail/u-boot/attachments/20240524/588407b3/attachment.sig>