[PATCH] staging: gpib: Workaround for ppc build failure

[PATCH] staging: gpib: Workaround for ppc build failure

[Dave Penkler]

Make GPIB_FMH depend on !PPC

Reported_by: Stephen Rothwell <sfr@canb.auug.org.au>
Link: https://lore.kernel.org/all/20241015165538.634707e5@canb.auug.org.au/

Signed-off-by: Dave Penkler <dpenkler@gmail.com>
---
 drivers/staging/gpib/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/staging/gpib/Kconfig b/drivers/staging/gpib/Kconfig
index 95308d15a555..a9b811165f6b 100644
--- a/drivers/staging/gpib/Kconfig
+++ b/drivers/staging/gpib/Kconfig
@@ -128,7 +128,7 @@ config GPIB_FMH
        tristate "FMH FPGA based devices"
        select GPIB_COMMON
        select GPIB_NEC7210
-       depends on BROKEN
+       depends on !PPC
        depends on OF && PCI
        help
          GPIB driver for fmhess FPGA based devices
-- 
2.47.1

Re: [PATCH] staging: gpib: Workaround for ppc build failure

[Guenter Roeck]

On Wed, Dec 04, 2024 at 02:47:36PM +0100, Dave Penkler wrote:
> Make GPIB_FMH depend on !PPC
> 
> Reported_by: Stephen Rothwell <sfr@canb.auug.org.au>
> Link: https://lore.kernel.org/all/20241015165538.634707e5@canb.auug.org.au/
> 
> Signed-off-by: Dave Penkler <dpenkler@gmail.com>

This patch just exposes other build failures, for example
i386:allyesconfig or x86_64:allyesconfig.

Error log:
x86_64-linux-ld: vmlinux.o: in function `fmh_gpib_t1_delay':
fmh_gpib.c:(.text+0xfd49dbd): undefined reference to `nec7210_t1_delay'

There are many more missing functions. FWIW, I don't know how this is
supposed to work in the first place, since pretty much all gpib Makefiles
say

obj-m += <object file>

meaning they won't be built with allyesconfig. fmh_gpib is the only
exception with

obj-$(CONFIG_GPIB_FMH) += fmh_gpib.o

which means it _will_ be built with allyesconfig. So it is not surprising
that it fails to link allyesconfig builds. Actually, it will fail all builds
with CONFIG_GPIB_FMH=y.

Guenter

Re: [PATCH] staging: gpib: Workaround for ppc build failure

[Greg KH]

On Sun, Dec 15, 2024 at 03:23:55PM -0800, Guenter Roeck wrote:
> On Wed, Dec 04, 2024 at 02:47:36PM +0100, Dave Penkler wrote:
> > Make GPIB_FMH depend on !PPC
> > 
> > Reported_by: Stephen Rothwell <sfr@canb.auug.org.au>
> > Link: https://lore.kernel.org/all/20241015165538.634707e5@canb.auug.org.au/
> > 
> > Signed-off-by: Dave Penkler <dpenkler@gmail.com>
> 
> This patch just exposes other build failures, for example
> i386:allyesconfig or x86_64:allyesconfig.
> 
> Error log:
> x86_64-linux-ld: vmlinux.o: in function `fmh_gpib_t1_delay':
> fmh_gpib.c:(.text+0xfd49dbd): undefined reference to `nec7210_t1_delay'
> 
> There are many more missing functions. FWIW, I don't know how this is
> supposed to work in the first place, since pretty much all gpib Makefiles
> say
> 
> obj-m += <object file>
> 
> meaning they won't be built with allyesconfig. fmh_gpib is the only
> exception with
> 
> obj-$(CONFIG_GPIB_FMH) += fmh_gpib.o
> 
> which means it _will_ be built with allyesconfig. So it is not surprising
> that it fails to link allyesconfig builds. Actually, it will fail all builds
> with CONFIG_GPIB_FMH=y.

Arnd has a patch for this that I'll be queueing up this week and getting
to Linus also this week.

thanks,

greg k-h