From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:49423) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1SiCli-00039C-B6 for qemu-devel@nongnu.org; Fri, 22 Jun 2012 18:55:59 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1SiClg-0004nz-7I for qemu-devel@nongnu.org; Fri, 22 Jun 2012 18:55:57 -0400 Received: from mout.web.de ([212.227.17.12]:58150) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1SiClf-0004nj-Sy for qemu-devel@nongnu.org; Fri, 22 Jun 2012 18:55:56 -0400 Message-ID: <4FE4F7F5.7030400@web.de> Date: Sat, 23 Jun 2012 00:55:49 +0200 From: Jan Kiszka MIME-Version: 1.0 References: <4FE4F56D.1020201@web.de> In-Reply-To: <4FE4F56D.1020201@web.de> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="------------enig7560E6D252B45BBA09841F5F" Subject: Re: [Qemu-devel] [PATCH] kvm: First step to push iothread lock out of inner run loop List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel Cc: Liu Ping Fan , kvm , Marcelo Tosatti , Alexander Graf , Avi Kivity , Anthony Liguori This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --------------enig7560E6D252B45BBA09841F5F Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Should have declared this [RFC] in the subject and CC'ed kvm... On 2012-06-23 00:45, Jan Kiszka wrote: > This sketches a possible path to get rid of the iothread lock on vmexit= s > in KVM mode. On x86, the the in-kernel irqchips has to be used because > we otherwise need to synchronize APIC and other per-cpu state accesses > that could be changed concurrently. Not yet fully analyzed is the NMI > injection path in the absence of an APIC. >=20 > s390x should be fine without specific locking as their pre/post-run > callbacks are empty. Power requires locking for the pre-run callback. >=20 > This patch is untested, but a similar version was successfully used in > a x86 setup with a network I/O path that needed no central iothread > locking anymore (required special MMIO exit handling). > --- > kvm-all.c | 18 ++++++++++++++++-- > target-i386/kvm.c | 7 +++++++ > target-ppc/kvm.c | 4 ++++ > 3 files changed, 27 insertions(+), 2 deletions(-) >=20 > diff --git a/kvm-all.c b/kvm-all.c > index f8e4328..9c3e26f 100644 > --- a/kvm-all.c > +++ b/kvm-all.c > @@ -1460,6 +1460,8 @@ int kvm_cpu_exec(CPUArchState *env) > return EXCP_HLT; > } > =20 > + qemu_mutex_unlock_iothread(); > + > do { > if (env->kvm_vcpu_dirty) { > kvm_arch_put_registers(env, KVM_PUT_RUNTIME_STATE); > @@ -1476,14 +1478,16 @@ int kvm_cpu_exec(CPUArchState *env) > */ > qemu_cpu_kick_self(); > } > - qemu_mutex_unlock_iothread(); > =20 > run_ret =3D kvm_vcpu_ioctl(env, KVM_RUN, 0); > =20 > - qemu_mutex_lock_iothread(); > kvm_arch_post_run(env, run); > =20 > + /* TODO: push coalesced mmio flushing to the point where we ac= cess > + * devices that are using it (currently VGA and E1000). */ > + qemu_mutex_lock_iothread(); > kvm_flush_coalesced_mmio_buffer(); > + qemu_mutex_unlock_iothread(); > =20 > if (run_ret < 0) { > if (run_ret =3D=3D -EINTR || run_ret =3D=3D -EAGAIN) { > @@ -1499,19 +1503,23 @@ int kvm_cpu_exec(CPUArchState *env) > switch (run->exit_reason) { > case KVM_EXIT_IO: > DPRINTF("handle_io\n"); > + qemu_mutex_lock_iothread(); > kvm_handle_io(run->io.port, > (uint8_t *)run + run->io.data_offset, > run->io.direction, > run->io.size, > run->io.count); > + qemu_mutex_unlock_iothread(); > ret =3D 0; > break; > case KVM_EXIT_MMIO: > DPRINTF("handle_mmio\n"); > + qemu_mutex_lock_iothread(); > cpu_physical_memory_rw(run->mmio.phys_addr, > run->mmio.data, > run->mmio.len, > run->mmio.is_write); > + qemu_mutex_unlock_iothread(); > ret =3D 0; > break; > case KVM_EXIT_IRQ_WINDOW_OPEN: > @@ -1520,7 +1528,9 @@ int kvm_cpu_exec(CPUArchState *env) > break; > case KVM_EXIT_SHUTDOWN: > DPRINTF("shutdown\n"); > + qemu_mutex_lock_iothread(); > qemu_system_reset_request(); > + qemu_mutex_unlock_iothread(); > ret =3D EXCP_INTERRUPT; > break; > case KVM_EXIT_UNKNOWN: > @@ -1533,11 +1543,15 @@ int kvm_cpu_exec(CPUArchState *env) > break; > default: > DPRINTF("kvm_arch_handle_exit\n"); > + qemu_mutex_lock_iothread(); > ret =3D kvm_arch_handle_exit(env, run); > + qemu_mutex_unlock_iothread(); > break; > } > } while (ret =3D=3D 0); > =20 > + qemu_mutex_lock_iothread(); > + > if (ret < 0) { > cpu_dump_state(env, stderr, fprintf, CPU_DUMP_CODE); > vm_stop(RUN_STATE_INTERNAL_ERROR); > diff --git a/target-i386/kvm.c b/target-i386/kvm.c > index 0d0d8f6..0ad64d1 100644 > --- a/target-i386/kvm.c > +++ b/target-i386/kvm.c > @@ -1631,7 +1631,10 @@ void kvm_arch_pre_run(CPUX86State *env, struct k= vm_run *run) > =20 > /* Inject NMI */ > if (env->interrupt_request & CPU_INTERRUPT_NMI) { > + qemu_mutex_lock_iothread(); > env->interrupt_request &=3D ~CPU_INTERRUPT_NMI; > + qemu_mutex_unlock_iothread(); > + > DPRINTF("injected NMI\n"); > ret =3D kvm_vcpu_ioctl(env, KVM_NMI); > if (ret < 0) { > @@ -1641,6 +1644,8 @@ void kvm_arch_pre_run(CPUX86State *env, struct kv= m_run *run) > } > =20 > if (!kvm_irqchip_in_kernel()) { > + qemu_mutex_lock_iothread(); > + > /* Force the VCPU out of its inner loop to process any INIT re= quests > * or pending TPR access reports. */ > if (env->interrupt_request & > @@ -1682,6 +1687,8 @@ void kvm_arch_pre_run(CPUX86State *env, struct kv= m_run *run) > =20 > DPRINTF("setting tpr\n"); > run->cr8 =3D cpu_get_apic_tpr(env->apic_state); > + > + qemu_mutex_unlock_iothread(); > } > } > =20 > diff --git a/target-ppc/kvm.c b/target-ppc/kvm.c > index c09cc39..60d91a5 100644 > --- a/target-ppc/kvm.c > +++ b/target-ppc/kvm.c > @@ -471,6 +471,8 @@ void kvm_arch_pre_run(CPUPPCState *env, struct kvm_= run *run) > int r; > unsigned irq; > =20 > + qemu_mutex_lock_iothread(); > + > /* PowerPC QEMU tracks the various core input pins (interrupt, cri= tical > * interrupt, reset, etc) in PPC-specific env->irq_input_state. */= > if (!cap_interrupt_level && > @@ -497,6 +499,8 @@ void kvm_arch_pre_run(CPUPPCState *env, struct kvm_= run *run) > /* We don't know if there are more interrupts pending after this. = However, > * the guest will return to userspace in the course of handling th= is one > * anyways, so we will get a chance to deliver the rest. */ > + > + qemu_mutex_unlock_iothread(); > } > =20 > void kvm_arch_post_run(CPUPPCState *env, struct kvm_run *run) >=20 --------------enig7560E6D252B45BBA09841F5F Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk/k9/UACgkQitSsb3rl5xQplgCgnoQk+UNru7YV8szplacrI+jA v6cAoJ45mpoq3Saaj+8AOSLcPP+P2lgr =r/8t -----END PGP SIGNATURE----- --------------enig7560E6D252B45BBA09841F5F--