Re: [Qemu-devel] [PATCH] Fix signal handling of SIG_IPI when io-thread is enabled

[Jan Kiszka <jan.kiszka@siemens.com>]

On 2011-06-15 07:20, Alexandre Raymond wrote:
> Both the signal thread (via sigwait()) and the cpu thread (via
> a normal signal handler) were attempting to catch SIG_IPI.

Why? Ahh, because of qemu_cpu_kick_self: raise(SIG_IPI)! That should
generate a per-process SIG_IPI. And that may not only affect Darwin.
Looks good.

Acked-by: Jan Kiszka <jan.kiszka@siemens.com>

> 
> This resulted in random freezes under Darwin.
> 
> This patch separates SIG_IPI from the rest of the signals handled
> by the signal thread, because it is independently caught by the cpu
> thread.
> 
> Signed-off-by: Alexandre Raymond <cerbere@gmail.com>
> ---
>  cpus.c |   10 +++++++++-
>  1 files changed, 9 insertions(+), 1 deletions(-)
> 
> diff --git a/cpus.c b/cpus.c
> index 18a1522..84ffd1c 100644
> --- a/cpus.c
> +++ b/cpus.c
> @@ -394,10 +394,18 @@ static int qemu_signal_init(void)
>      sigaddset(&set, SIGUSR2);
>      pthread_sigmask(SIG_UNBLOCK, &set, NULL);
>  
> +    /*
> +     * SIG_IPI must be blocked in the main thread and must not be caught
> +     * by sigwait() in the signal thread. Otherwise, the cpu thread will
> +     * not catch it reliably.
> +     */
> +    sigemptyset(&set);
> +    sigaddset(&set, SIG_IPI);
> +    pthread_sigmask(SIG_BLOCK, &set, NULL);
> +
>      sigemptyset(&set);
>      sigaddset(&set, SIGIO);
>      sigaddset(&set, SIGALRM);
> -    sigaddset(&set, SIG_IPI);
>      sigaddset(&set, SIGBUS);
>  #else
>      sigemptyset(&set);

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux