| Message ID | 1308173586-5344-1-git-send-email-levinsasha928@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 06/16/2011 05:33 AM, Sasha Levin wrote: > Send signals only to the thread which handles console injection > instead of the entire process. > > This prevents every single thread from having to ignore SIGALRM > and will help scaling when running large amount of VCPU threads. > > Suggested-by: Ingo Molnar <mingo@elte.hu> > Signed-off-by: Sasha Levin <levinsasha928@gmail.com> This one is much simpler than the sig_block patches and also resolves the getaddrinfo() problem for me in virtio net thread. Tested-by: Asias He <asias.hejun@gmail.com>
* Asias He <asias.hejun@gmail.com> wrote: > On 06/16/2011 05:33 AM, Sasha Levin wrote: > > Send signals only to the thread which handles console injection > > instead of the entire process. > > > > This prevents every single thread from having to ignore SIGALRM > > and will help scaling when running large amount of VCPU threads. > > > > Suggested-by: Ingo Molnar <mingo@elte.hu> > > Signed-off-by: Sasha Levin <levinsasha928@gmail.com> > > > This one is much simpler than the sig_block patches and also resolves > the getaddrinfo() problem for me in virtio net thread. > > Tested-by: Asias He <asias.hejun@gmail.com> yep, very nice! Acked-by: Ingo Molnar <mingo@elte.hu> Thanks, Ingo -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 06/16/2011 03:29 PM, Ingo Molnar wrote: > * Asias He <asias.hejun@gmail.com> wrote: > >> On 06/16/2011 05:33 AM, Sasha Levin wrote: >>> Send signals only to the thread which handles console injection >>> instead of the entire process. >>> >>> This prevents every single thread from having to ignore SIGALRM >>> and will help scaling when running large amount of VCPU threads. >>> >>> Suggested-by: Ingo Molnar <mingo@elte.hu> >>> Signed-off-by: Sasha Levin <levinsasha928@gmail.com> >> >> >> This one is much simpler than the sig_block patches and also resolves >> the getaddrinfo() problem for me in virtio net thread. >> >> Tested-by: Asias He <asias.hejun@gmail.com> > > yep, very nice! > > Acked-by: Ingo Molnar <mingo@elte.hu> > > Thanks, > > Ingo > Hi, Pekka Ping. Mind to merge this patch?
diff --git a/tools/kvm/kvm.c b/tools/kvm/kvm.c index 2c56a79..22871db 100644 --- a/tools/kvm/kvm.c +++ b/tools/kvm/kvm.c @@ -28,6 +28,7 @@ #include <fcntl.h> #include <time.h> #include <sys/eventfd.h> +#include <asm/unistd.h> #define DEFINE_KVM_EXIT_REASON(reason) [reason] = #reason @@ -529,8 +530,9 @@ void kvm__start_timer(struct kvm *kvm) memset(&sev, 0, sizeof(struct sigevent)); sev.sigev_value.sival_int = 0; - sev.sigev_notify = SIGEV_SIGNAL; + sev.sigev_notify = SIGEV_THREAD_ID; sev.sigev_signo = SIGALRM; + sev._sigev_un._tid = syscall(__NR_gettid); if (timer_create(CLOCK_REALTIME, &sev, &kvm->timerid) < 0) die("timer_create()");
Send signals only to the thread which handles console injection instead of the entire process. This prevents every single thread from having to ignore SIGALRM and will help scaling when running large amount of VCPU threads. Suggested-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Sasha Levin <levinsasha928@gmail.com> --- tools/kvm/kvm.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-)