Retronetworking

Since July 14, the master-dahdi-linux jenkins jobs are failing for torvalds-master and stable-linux-6.1.y but continue to pass for stable-linux-5.10.y and stable-linux-4.19.y

The error in question is triggered by the use of -Werror=address

build/drivers/dahdi/xpp/xbus-core.c: In function 'xbus_check_unique':
/build/drivers/dahdi/xpp/xbus-core.c:116:13: error: the comparison will always evaluate as 'true' for the address of 'label' will never be NULL [-Werror=address]
  116 |         if (xbus->label && *(xbus->label)) {
      |             ^~~~
In file included from /build/drivers/dahdi/xpp/xbus-core.c:39:
/build/drivers/dahdi/xpp/xbus-core.h:187:14: note: 'label' declared here
  187 |         char label[LABEL_SIZE];
      |              ^~~~~
  CC [M]  /build/drivers/dahdi/voicebus/GpakCust.o
  CC [M]  /build/drivers/dahdi/voicebus/GpakApi.o
cc1: all warnings being treated as errors
make[3]: *** [scripts/Makefile.build:252: /build/drivers/dahdi/xpp/xbus-core.o] Error 1
make[2]: *** [scripts/Makefile.build:494: /build/drivers/dahdi/xpp] Error 2
make[2]: *** Waiting for unfinished jobs....

I don't understand how this might have changed overnight. I can continue to build it locally on my debian unstable with gcc-12.3.0, even against the debian 6.1.27-1 kernel.

The big qustion is probably where does that -Werror=address suddenly come from?

We know that DAHDI is full of things that have triggered compiler warnings for years, but we don't want to diverge even more from master than needed, particularly for hardware that we don't have to test any code changes on.