Re: [PATCH 2/2] ipc/sem.c: move wake_up_process out of the spinlocksection

[Manfred Spraul]

On 05/12/2010 07:34 PM, Manfred Spraul wrote:
> The effect would be that e.g. a semctl(SETALL) operation might change 
> sem_otime.
> semctl(SETALL) must only change sem_ctime (and sem_otime only if it 
> causes a wakeup
> and the woken up thread modifies the array)
Just for your information: I'm not sure why semctl(SETALL) changes 
sem_ctime at all.
According to the opengroup, it shouldn't do that.

On FreeBSD, it doesn't.

Any proposals how to fix it?
Should I remove the update to sem_ctime?


> [manfred@cores linux-2.6]$ touch ipc/sem.o
It should have been "touch ipc/sem.c":

> [manfred@cores linux-2.6]$ touch ipc/sem.c
> [manfred@cores linux-2.6]$ make ipc
>   CHK     include/linux/version.h
>   CHK     include/generated/utsrelease.h
>  CALL    scripts/checksyscalls.sh
>  CC      ipc/sem.o
>  LD      ipc/built-in.o

--
    Manfred
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/