From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [patch 3/4] QEMU/KVM: ignore SIG_IPI signals in userspace Date: Tue, 08 Apr 2008 04:17:13 +0300 Message-ID: <47FAC799.4010601@qumranet.com> References: <20080405182246.761430758@localhost.localdomain> <20080405182320.098125884@localhost.localdomain> <47F8949B.8000409@qumranet.com> <20080407202624.GA4254@dmt> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Cc: kvm-devel@lists.sourceforge.net To: Marcelo Tosatti Return-path: In-Reply-To: <20080407202624.GA4254@dmt> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: kvm-devel-bounces@lists.sourceforge.net Errors-To: kvm-devel-bounces@lists.sourceforge.net List-Id: kvm.vger.kernel.org Marcelo Tosatti wrote: > On Sun, Apr 06, 2008 at 12:15:07PM +0300, Avi Kivity wrote: > >> Marcelo Tosatti wrote: >> >>> Otherwise a signal can be received in userspace and a vcpu goes back >>> to the kernel while it should stay still. >>> >>> Signed-off-by: Marcelo Tosatti >>> >>> Index: kvm-userspace.io/qemu/qemu-kvm.c >>> =================================================================== >>> --- kvm-userspace.io.orig/qemu/qemu-kvm.c >>> +++ kvm-userspace.io/qemu/qemu-kvm.c >>> @@ -350,7 +350,6 @@ static void *ap_main_loop(void *_env) >>> vcpu->env = env; >>> vcpu->env->thread_id = kvm_get_thread_id(); >>> sigfillset(&signals); >>> - sigdelset(&signals, SIG_IPI); >>> sigprocmask(SIG_BLOCK, &signals, NULL); >>> kvm_create_vcpu(kvm_context, env->cpu_index); >>> kvm_qemu_init_env(env); >>> >>> >>> >> Does this work with -no-kvm-irqchip? >> > > Yes. SIG_IPI was blocked before the IO thread. > > Okay (of course; it is blocked, but still dequeued by sigtimedwait). >> I think we need to fix the kernel to handle random signals. Otherwise >> even attaching a debugger can change guest behavior (I think). >> > > Well ptrace forces signals so SIGSTOP is delivered even though the child > has blocked them. > > Attaching a debugger does change behaviour since SIGSTOP will send a > vcpu back to userspace. > > Can you be more specific? > > I misunderstood. I thought something about a spurious signal being received in the kernel, dropping it out of hlt state, and confusing the guest; but the real issue is the signal getting lost completely if it is delivered to userspace instead of sigtimedwait(), which this patch fixes. -- Any sufficiently difficult bug is indistinguishable from a feature. ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Register now and save $200. Hurry, offer ends at 11:59 p.m., Monday, April 7! Use priority code J8TLD2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone