linux kernel compiler warnings ignored

[Oskar Senft]

Hi Patrick

I'm not sure I follow ... Is that documented somewhere or can you elaborate?

I was hoping that I could build the kernel separately (as Andrew described)
and then combine that with "something" to end up with an image similar to
"obmc-phosphor-image-${MACHINE}.static.mtd".

Or is there an even easier way without having to build and flash a new
.static.mtd?

Thanks
Oskar.

On Fri, Apr 26, 2019 at 10:45 AM Patrick Venture <venture at google.com> wrote:

> On Fri, Apr 26, 2019 at 8:25 AM Oskar Senft <osk at google.com> wrote:
> >
> > Thank you!
> >
> > Is there an "easy" way to bake the resulting kernel image into an image
> for the SPI flash chip?
> >
> > I often make kernel-only changes and usually run through the whole
> process of committing to a local repo, then running bit bake and waiting
> for all the other stuff it also does. If there was a way to just build the
> kernel and replace the one in the SPI flash image with the new one, that
> would be awesome.
>
> Oskar, if you're using static-images still, there's an image-kernel
> file output from the build process.  this is supported if you drop it
> into /run/initramfs/ and reboot similarly to image-bmc.
>
> >
> > Oskar.
> >
> > On Fri, Apr 26, 2019 at 10:15 AM Andrew Jeffery <andrew at aj.id.au> wrote:
> >>
> >>
> >>
> >> On Fri, 26 Apr 2019, at 08:01, Patrick Venture wrote:
> >> > I'm getting feedback from a patchset I sent upstream that there are
> >> > compiler warnings.  Which surprised me because my OpenBMC builds with
> >> > those patches don't complain.  I imagine therefore we have them
> >> > silenced and was curious where we silence them.
> >>
> >> How are you building the kernel?
> >>
> >> I strongly recommend you build kernels directly from the kernel source
> tree rather than using bitbake, if that's what you're doing.
> >>
> >> In the kernel source tree, do:
> >>
> >> $ export ARCH=arm
> >> $ export CROSS_COMPILE=arm-linux-gnueabi-
> >> $ make aspeed_g5_defconfig
> >> $ make W=1 -j$(nproc)
> >>
> >> This exports are from memory, ymmv.
> >>
> >> In addition to giving you more output than Greg is after, it has a much
> lower write/compile/test latency.
> >>
> >> Hope that helps,
> >>
> >> Andrew
> >>
> >> >
> >> > Patrick
> >> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ozlabs.org/pipermail/openbmc/attachments/20190426/b6718ff6/attachment-0001.htm>