[Qemu-devel] [PULL 0/2] ppc-for-2.11 queue 20171114

[Qemu-devel] [PULL 0/2] ppc-for-2.11 queue 20171114

[David Gibson]

The following changes since commit 4e8a737c092d25fe5ff99ac7bff01c222c1c017d:

  Merge remote-tracking branch 'remotes/kraxel/tags/vga-20171110-pull-request' into staging (2017-11-13 14:33:29 +0000)

are available in the git repository at:

  git://github.com/dgibson/qemu.git tags/ppc-for-2.11-20171114

for you to fetch changes up to dcb556fc6ad879f7d29fb233a273a1ea3aa0549c:

  xics/kvm: synchonize state before 'info pic' (2017-11-14 11:12:42 +1100)

----------------------------------------------------------------
ppc patch queue for 2017-11-14

Another couple of fixes for qemu-2.11.

----------------------------------------------------------------
Greg Kurz (1):
      xics/kvm: synchonize state before 'info pic'

Sam Bobroff (1):
      target/ppc: correct htab shift for hash on radix

 hw/intc/xics.c        | 11 +++++++++++
 hw/intc/xics_kvm.c    | 19 +++++++++++++++++++
 hw/ppc/spapr_hcall.c  | 13 ++++++-------
 include/hw/ppc/xics.h |  2 ++
 4 files changed, 38 insertions(+), 7 deletions(-)

[Qemu-devel] [PULL 1/2] target/ppc: correct htab shift for hash on radix

[David Gibson]

From: Sam Bobroff <sam.bobroff@au1.ibm.com>

KVM HV will soon support running a guest in hash mode on a POWER9 host
running in radix mode (see [1]), however the guest currently fails to
boot.

This is because the "htab_shift" value (the size of the MMU's hash
table) is added to the device tree before KVM has had a chance to
change it. If the host is in hash mode, KVM does not need to change it
and so the problem is not seen, but when the host is in radix mode a
change is required and we see a problem.

To fix this, move the call spapr_setup_hpt_and_vrma() (where
htab_shift could be changed) up a little so that it's called before
spapr_h_cas_compose_response() (where htab_shift is added to the
device tree).

Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>

[1] See http://www.spinics.net/lists/kvm-ppc/msg13057.html
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/ppc/spapr_hcall.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/hw/ppc/spapr_hcall.c b/hw/ppc/spapr_hcall.c
index 0d59d1534d..be22a6b289 100644
--- a/hw/ppc/spapr_hcall.c
+++ b/hw/ppc/spapr_hcall.c
@@ -1636,6 +1636,12 @@ static target_ulong h_client_architecture_support(PowerPCCPU *cpu,
     spapr->cas_legacy_guest_workaround = !spapr_ovec_test(ov1_guest,
                                                           OV1_PPC_3_00);
     if (!spapr->cas_reboot) {
+        /* If ppc_spapr_reset() did not set up a HPT but one is necessary
+         * (because the guest isn't going to use radix) then set it up here. */
+        if ((spapr->patb_entry & PATBE1_GR) && !guest_radix) {
+            /* legacy hash or new hash: */
+            spapr_setup_hpt_and_vrma(spapr);
+        }
         spapr->cas_reboot =
             (spapr_h_cas_compose_response(spapr, args[1], args[2],
                                           ov5_updates) != 0);
@@ -1644,13 +1650,6 @@ static target_ulong h_client_architecture_support(PowerPCCPU *cpu,
 
     if (spapr->cas_reboot) {
         qemu_system_reset_request(SHUTDOWN_CAUSE_GUEST_RESET);
-    } else {
-        /* If ppc_spapr_reset() did not set up a HPT but one is necessary
-         * (because the guest isn't going to use radix) then set it up here. */
-        if ((spapr->patb_entry & PATBE1_GR) && !guest_radix) {
-            /* legacy hash or new hash: */
-            spapr_setup_hpt_and_vrma(spapr);
-        }
     }
 
     return H_SUCCESS;
-- 
2.13.6

[Qemu-devel] [PULL 2/2] xics/kvm: synchonize state before 'info pic'

[David Gibson]

From: Greg Kurz <groug@kaod.org>

When using the emulated XICS, the 'info pic' monitor command shows:

CPU 0 XIRR=ff000000 ((nil)) PP=ff MFRR=ff
ICS 1000..13ff 0x10040060340
  1000 MSI 05 00
  1001 MSI 05 00
  1002 MSI 05 00
  1003 MSI ff 00
  1004 LSI ff 00
  1005 LSI ff 00
  1006 LSI ff 00
  1007 LSI ff 00
  1008 MSI 05 00
  1009 MSI 05 00
  100a MSI 05 00
  100b MSI 05 00
  100c MSI 05 00

but when using the in-kernel XICS with the very same guest, we get:

CPU 0 XIRR=00000000 ((nil)) PP=ff MFRR=ff
ICS 1000..13ff 0x10032e00340
  1000 MSI ff 00
  1001 MSI ff 00
  1002 MSI ff 00
  1003 MSI ff 00
  1004 LSI ff 00
  1005 LSI ff 00
  1006 LSI ff 00
  1007 LSI ff 00
  1008 MSI ff 00
  1009 MSI ff 00
  100a MSI ff 00
  100b MSI ff 00
  100c MSI ff 00

ie, all irqs are masked and XIRR is null, while we should get the
same output as with the emulated XICS.

If the guest is then migrated, 'info pic' shows the expected values
on both source and destination.

The problem is that QEMU doesn't synchronize with KVM before printing
the XICS state. Migration happens to fix the output because it enforces
synchronization with KVM.

To fix the invalid output of 'info pic', this patch introduces a new
synchronize_state operation for both ICPStateClass and ICSStateClass.
The ICP operation relies on run_on_cpu() in order to kick the vCPU
and avoid sleeping on KVM_GET_ONE_REG.

Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/intc/xics.c        | 11 +++++++++++
 hw/intc/xics_kvm.c    | 19 +++++++++++++++++++
 include/hw/ppc/xics.h |  2 ++
 3 files changed, 32 insertions(+)

diff --git a/hw/intc/xics.c b/hw/intc/xics.c
index cc9816e7f2..a1cc0e420c 100644
--- a/hw/intc/xics.c
+++ b/hw/intc/xics.c
@@ -40,11 +40,17 @@
 
 void icp_pic_print_info(ICPState *icp, Monitor *mon)
 {
+    ICPStateClass *icpc = ICP_GET_CLASS(icp);
     int cpu_index = icp->cs ? icp->cs->cpu_index : -1;
 
     if (!icp->output) {
         return;
     }
+
+    if (icpc->synchronize_state) {
+        icpc->synchronize_state(icp);
+    }
+
     monitor_printf(mon, "CPU %d XIRR=%08x (%p) PP=%02x MFRR=%02x\n",
                    cpu_index, icp->xirr, icp->xirr_owner,
                    icp->pending_priority, icp->mfrr);
@@ -52,6 +58,7 @@ void icp_pic_print_info(ICPState *icp, Monitor *mon)
 
 void ics_pic_print_info(ICSState *ics, Monitor *mon)
 {
+    ICSStateClass *icsc = ICS_BASE_GET_CLASS(ics);
     uint32_t i;
 
     monitor_printf(mon, "ICS %4x..%4x %p\n",
@@ -61,6 +68,10 @@ void ics_pic_print_info(ICSState *ics, Monitor *mon)
         return;
     }
 
+    if (icsc->synchronize_state) {
+        icsc->synchronize_state(ics);
+    }
+
     for (i = 0; i < ics->nr_irqs; i++) {
         ICSIRQState *irq = ics->irqs + i;
 
diff --git a/hw/intc/xics_kvm.c b/hw/intc/xics_kvm.c
index 3091ad3ac2..89fb20e2c5 100644
--- a/hw/intc/xics_kvm.c
+++ b/hw/intc/xics_kvm.c
@@ -81,6 +81,18 @@ static void icp_get_kvm_state(ICPState *icp)
         & KVM_REG_PPC_ICP_PPRI_MASK;
 }
 
+static void do_icp_synchronize_state(CPUState *cpu, run_on_cpu_data arg)
+{
+    icp_get_kvm_state(arg.host_ptr);
+}
+
+static void icp_synchronize_state(ICPState *icp)
+{
+    if (icp->cs) {
+        run_on_cpu(icp->cs, do_icp_synchronize_state, RUN_ON_CPU_HOST_PTR(icp));
+    }
+}
+
 static int icp_set_kvm_state(ICPState *icp, int version_id)
 {
     uint64_t state;
@@ -156,6 +168,7 @@ static void icp_kvm_class_init(ObjectClass *klass, void *data)
     icpc->post_load = icp_set_kvm_state;
     icpc->realize = icp_kvm_realize;
     icpc->reset = icp_kvm_reset;
+    icpc->synchronize_state = icp_synchronize_state;
 }
 
 static const TypeInfo icp_kvm_info = {
@@ -234,6 +247,11 @@ static void ics_get_kvm_state(ICSState *ics)
     }
 }
 
+static void ics_synchronize_state(ICSState *ics)
+{
+    ics_get_kvm_state(ics);
+}
+
 static int ics_set_kvm_state(ICSState *ics, int version_id)
 {
     uint64_t state;
@@ -347,6 +365,7 @@ static void ics_kvm_class_init(ObjectClass *klass, void *data)
     icsc->realize = ics_kvm_realize;
     icsc->pre_save = ics_get_kvm_state;
     icsc->post_load = ics_set_kvm_state;
+    icsc->synchronize_state = ics_synchronize_state;
 }
 
 static const TypeInfo ics_kvm_info = {
diff --git a/include/hw/ppc/xics.h b/include/hw/ppc/xics.h
index 28d248abad..2df99be111 100644
--- a/include/hw/ppc/xics.h
+++ b/include/hw/ppc/xics.h
@@ -69,6 +69,7 @@ struct ICPStateClass {
     void (*pre_save)(ICPState *icp);
     int (*post_load)(ICPState *icp, int version_id);
     void (*reset)(ICPState *icp);
+    void (*synchronize_state)(ICPState *icp);
 };
 
 struct ICPState {
@@ -119,6 +120,7 @@ struct ICSStateClass {
     void (*reject)(ICSState *s, uint32_t irq);
     void (*resend)(ICSState *s);
     void (*eoi)(ICSState *s, uint32_t irq);
+    void (*synchronize_state)(ICSState *s);
 };
 
 struct ICSState {
-- 
2.13.6

Re: [Qemu-devel] [PULL 0/2] ppc-for-2.11 queue 20171114

[Peter Maydell]

On 14 November 2017 at 04:37, David Gibson <david@gibson.dropbear.id.au> wrote:
> The following changes since commit 4e8a737c092d25fe5ff99ac7bff01c222c1c017d:
>
>   Merge remote-tracking branch 'remotes/kraxel/tags/vga-20171110-pull-request' into staging (2017-11-13 14:33:29 +0000)
>
> are available in the git repository at:
>
>   git://github.com/dgibson/qemu.git tags/ppc-for-2.11-20171114
>
> for you to fetch changes up to dcb556fc6ad879f7d29fb233a273a1ea3aa0549c:
>
>   xics/kvm: synchonize state before 'info pic' (2017-11-14 11:12:42 +1100)
>
> ----------------------------------------------------------------
> ppc patch queue for 2017-11-14
>
> Another couple of fixes for qemu-2.11.
>
> ----------------------------------------------------------------
> Greg Kurz (1):
>       xics/kvm: synchonize state before 'info pic'
>
> Sam Bobroff (1):
>       target/ppc: correct htab shift for hash on radix

Applied, thanks.

-- PMM