Re: [tip:sched/core] BUILD SUCCESS 3387ce4d8a5f2956fab827edf499fe6780e83faa

[Nathan Chancellor]

Hello,

On Tue, Mar 22, 2022 at 06:36:16AM +0800, kernel test robot wrote:
> tree/branch: https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git sched/core
> branch HEAD: 3387ce4d8a5f2956fab827edf499fe6780e83faa  headers/prep: Fix header to build standalone: <linux/psi.h>
> 
> elapsed time: 730m
> 
> configs tested: 114
> configs skipped: 3
> 
> The following configs have been built successfully.
> More configs may be tested in the coming days.
> 
> gcc tested configs:
...
> arm                              allyesconfig
> arm                              allmodconfig

This caught my eye due to my earlier report:

https://lore.kernel.org/r/YjiddAnoCCz7Tbt3@dev-arch.thelio-3990X/

Are you sure these configurations built sucessfully?


$ git show -s
commit 3387ce4d8a5f2956fab827edf499fe6780e83faa
Author: Ingo Molnar <mingo@xxxxxxxxxx>
Date:   Mon Mar 21 11:05:50 2022 +0100

    headers/prep: Fix header to build standalone: <linux/psi.h>
    
    Add the <linux/cgroup-defs.h> dependency to <linux/psi.h>, because
    cgroup_move_task() will dereference 'struct css_set'.
    
    ( Only older toolchains are affected, due to variations in
      the implementation of rcu_assign_pointer() et al. )
    
    Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
    Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
    Reported-by: Sachin Sant <sachinp@xxxxxxxxxxxxx>
    Reported-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
    Reported-by: Borislav Petkov <bp@xxxxxxxxx>
    Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>

$ make -skj"$(nproc)" ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- allmodconfig kernel/sched/
In file included from kernel/sched/fair.c:52:
kernel/sched/sched.h:87:11: fatal error: asm/paravirt_api_clock.h: No such file or directory
   87 | # include <asm/paravirt_api_clock.h>
      |           ^~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make[3]: *** [scripts/Makefile.build:288: kernel/sched/fair.o] Error 1
In file included from kernel/sched/build_utility.c:52:
kernel/sched/sched.h:87:11: fatal error: asm/paravirt_api_clock.h: No such file or directory
   87 | # include <asm/paravirt_api_clock.h>
      |           ^~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make[3]: *** [scripts/Makefile.build:288: kernel/sched/build_utility.o] Error 1
In file included from kernel/sched/build_policy.c:33:
kernel/sched/sched.h:87:11: fatal error: asm/paravirt_api_clock.h: No such file or directory
   87 | # include <asm/paravirt_api_clock.h>
      |           ^~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make[3]: *** [scripts/Makefile.build:288: kernel/sched/build_policy.o] Error 1
In file included from kernel/sched/core.c:81:
kernel/sched/sched.h:87:11: fatal error: asm/paravirt_api_clock.h: No such file or directory
   87 | # include <asm/paravirt_api_clock.h>
      |           ^~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
...

I am curious to know how this error was not caught by these builds?
Perhaps an old .config from a previous allmodconfig is being reused,
rather than just running "make allmodconfig"? If that is the case, it
seems like that methodology will miss out on new configuration options
over time.

Cheers,
Nathan