svn commit: r323371 - head/sys/conf

[Warner Losh]

On Sat, Sep 9, 2017 at 1:57 PM, Sean Bruno <sbruno at freebsd.org> wrote:

>
>
> On 09/09/17 13:51, Warner Losh wrote:
> >
> >
> > On Sat, Sep 9, 2017 at 1:19 PM, Sean Bruno <sbruno at freebsd.org
> > <mailto:sbruno at freebsd.org>> wrote:
> >
> >     Author: sbruno
> >     Date: Sat Sep  9 19:19:13 2017
> >     New Revision: 323371
> >     URL: https://svnweb.freebsd.org/changeset/base/323371
> >     <https://svnweb.freebsd.org/changeset/base/323371>
> >
> >     Log:
> >       r323359 instroduced an ARMv8 only uart(4) device to the tree but
> >     placed
> >       the driver in a place where it will be built for all targets.  x86
> >     doesn't
> >       have all the required build bits for this device.
> >
> >       Move the uart(4) device mvebu to arm64 only.
> >
> >     Modified:
> >       head/sys/conf/files
> >       head/sys/conf/files.arm64
> >
> >     Modified: head/sys/conf/files
> >     ============================================================
> ==================
> >     --- head/sys/conf/files Sat Sep  9 18:39:55 2017        (r323370)
> >     +++ head/sys/conf/files Sat Sep  9 19:19:13 2017        (r323371)
> >     @@ -3053,7 +3053,6 @@ dev/uart/uart_bus_puc.c           optional
> >     uart puc
> >      dev/uart/uart_bus_scc.c                optional uart scc
> >      dev/uart/uart_core.c           optional uart
> >      dev/uart/uart_dbg.c            optional uart gdb
> >     -dev/uart/uart_dev_mvebu.c      optional uart uart_mvebu
> >      dev/uart/uart_dev_ns8250.c     optional uart uart_ns8250 | uart
> >     uart_snps
> >      dev/uart/uart_dev_pl011.c      optional uart pl011
> >      dev/uart/uart_dev_quicc.c      optional uart quicc
> >
> >     Modified: head/sys/conf/files.arm64
> >     ============================================================
> ==================
> >     --- head/sys/conf/files.arm64   Sat Sep  9 18:39:55 2017
> >     (r323370)
> >     +++ head/sys/conf/files.arm64   Sat Sep  9 19:19:13 2017
> >     (r323371)
> >     @@ -176,6 +176,7 @@ dev/pci/pci_host_generic_fdt.c      optional
> >         pci fdt
> >      dev/psci/psci.c                        optional        psci
> >      dev/psci/psci_arm64.S          optional        psci
> >      dev/uart/uart_cpu_arm64.c      optional        uart
> >     +dev/uart/uart_dev_mvebu.c      optional        uart uart_mvebu
> >      dev/uart/uart_dev_pl011.c      optional        uart pl011
> >      dev/usb/controller/dwc_otg_hisi.c optional     dwcotg fdt
> >     soc_hisi_hi6220
> >      dev/usb/controller/ehci_mv.c   optional        ehci_mv fdt
> >
> >
> > This is a bad change since we may support running the mvebu in 32-bit
> > mode, in which case we'd need it duplicated.
> >
> > There's no harm in it being in the global files file, as it won't show
> > up automatically.
> >
> > Warner
> >
>
>
> I moved it around like this to unbreak the x86 build.
>
> I see explicit restrictions on the header inclusion in sys/modules/uart
> that would keep this from compiling on x86 and caused the breakage.
>
> I don't see an obvious way to get this device to build on 32bit and
> 64bit arm (unless files.arm is included automagically).
>
> Does this make sense?
>

No.

 -dev/uart/uart_dev_mvebu.c      optional uart uart_mvebu

This should only be enabled in kernel configs that have both 'uart' AND
'uart_mvebu' in them. What enabled uart_mvebu? I don't see it in GENERIC.

Warner