diff mbox series

[v5,06/13] hw/core/qdev: handle parent bus change regarding resettable

Message ID 20191018150630.31099-7-damien.hedde@greensocs.com (mailing list archive)
State New, archived
Headers show
Series Multi-phase reset mechanism | expand

Commit Message

Damien Hedde Oct. 18, 2019, 3:06 p.m. UTC
In qdev_set_parent_bus(), when changing the parent bus of a
realized device, if the source and destination buses are not in the
same reset state, some adaptation are required. This patch adds
needed call to resettable_change_parent() to make sure a device reset
state stays coherent with its parent bus.

The addition is a no-op if:
1. the device being parented is not realized.
2. the device is realized, but both buses are not under reset.

Case 2 means that as long as qdev_set_parent_bus() is called
during the machine realization procedure (which is before the
machine reset so nothing is in reset), it is a no op.

There are 49 call sites of qdev_set_parent_bus(). All but one fall
into the no-op case:
+ 28 calls related to virtio (in hw/{s390x,display,virtio}/
  {vhost,virtio}-xxx.c) to set a _vdev_/_vgpu_ composing device
  parent bus just before realizing the _vdev_/_vgpu_.
+ hw/qdev.c: when creating a device in qdev_try_create()
+ hw/sysbus.c: when initializing a device in the sysbus
+ hw/display/virtio-gpu-pci.c: before realizing VirtIOGPUPCIBase/vgpu
+ hw/display/virtio-vga.c: before realizing VirtIOVGABase/vgpu
+ hw/i386/amd_iommu.c: before realizing AMDVIState/pci
+ hw/misc/auxbus.c: when creating an AUXBus
+ hw/misc/auxbus.c: when creating an AUXBus child
+ hw/misc/macio/macio.c: when initializing a MACIOState child
+ hw/misc/macio/macio.c: before realizing NewWorldMacIOState/pmu
+ hw/misc/macio/macio.c: before realizing NewWorldMacIOState/cuda
+ hw/pci-host/designware.c: before realizing DesignwarePCIEHost/root
+ hw/pci-host/gpex.c: before realizing GPEXHost/root
+ hw/pci-host/prep.c: when initializaing PREPPCIState/pci_dev
+ hw/pci-host/q35.c: before realizing Q35PCIHost/mch
+ hw/pci-host/versatile.c: when initializing PCIVPBState/pci_dev
+ hw/pci-host/xilinx-pcie.c: before realizing XilinxPCIEHost/root
+ hw/s390x/event-facility.c: when creating SCLPEventFacility/
                             TYPE_SCLP_QUIESCE
+ hw/s390x/event-facility.c: ditto with SCLPEventFacility/
                             TYPE_SCLP_CPU_HOTPLUG
+ hw/s390x/sclp.c: Not trivial because it is called on a SLCPDevice
  just after realizing it. Ok because at this point the destination
  bus (sysbus) is not in reset; the realize step is before the
  machine reset.
+ hw/sd/core.c: Not OK. Used in sdbus_reparent_card(). See below.
+ hw/ssi/ssi.c: Used to put spi slave on spi bus and connect the cs
  line in ssi_auto_connect_slave(). Ok because this function is only
  used in realize step in hw/ssi/aspeed_smc.ci, hw/ssi/imx_spi.c,
  hw/ssi/mss-spi.c, hw/ssi/xilinx_spi.c and hw/ssi/xilinx_spips.c.
+ hw/xen/xen-legacy-backend.c: when creating a XenLegacyDevice device
+ qdev-monitor.c: in device hotplug creation procedure before realize

Note that this commit alone will have no effect, right now there is no
use of resettable API to reset anything. So a bus will never be tagged
as in-reset by this same API.

The one place where side-effect will occurs is in hw/sd/core.c in
sdbus_reparent_card(). This function is only used in the raspi machines,
including during the sysbus reset procedure. This case will be fixed by
a following commit before globally enabling resettable API for sysbus
reset.

Signed-off-by: Damien Hedde <damien.hedde@greensocs.com>
---

Exhaustive list of the 28 virtio caller to qdev_set_parent_bus():
+ hw/display/virtio-gpu-pci.c     -> VirtIOGPUPCIBase/vgpu realize
+ hw/display_virtio-vga.c         -> VirtIOVGABase/vgpu realize
+ hw/s390x/vhost-vsock-ccw.c      -> VHostVSockCCWState/vdev realize
+ hw/s390x/virtio-ccw-9p.c        -> V9fsCCWState/vdev realize
+ hw/s390x/virtio-ccw-balloon.c   -> VirtIOBalloonCcw/vdev realize
+ hw/s390x/virtio-ccw-blk.c       -> VirtIOBlkCcw/vdev realize
+ hw/s390x/virtio-ccw-crypto.c    -> VirtIOCryptoCcw/vdev realize
+ hw/s390x/virtio-ccw-gpu.c       -> VirtIOGPUCcw/vdev realize
+ hw/s390x/virtio-ccw-input.c     -> VirtIOInputCcw/vdev realize
+ hw/s390x/virtio-ccw-net.c       -> VirtIONetCcw/vdev realize
+ hw/s390x/virtio-ccw-rng.c       -> VirtIORNGCcw/vdev realize
+ hw/s390x/virtio-ccw-scsi.c      -> VirtIOSCSICcw/vdev realize
+ hw/s390x/virtio-ccw-scsi.c      -> VHostSCSICcw/vdev realize
+ hw/s390x/virtio-ccw-serial.c    -> VirtioSerialCcw/vdev realize
+ hw/virtio/vhost-scsi-pci.c      -> VHostSCSIPCI/vdev realize
+ hw/virtio/vhost-user-blk-pci.c  -> VHostUserBlkPCI/vdev realize
+ hw/virtio/vhost-user-scsi-pci.c -> VHostUserSCSIPCI/vdev realize
+ hw/virtio/vhost-vsock-pci.c     -> VHostVSockPCI/vdev realize
+ hw/virtio/virtio-9p-pci.c       -> V9fsPCIState/vdev realize
+ hw/virtio/virtio-balloon-pci.c  -> VirtIOBalloonPCI/vdev realize
+ hw/virtio/virtio-blk-pci.c      -> VirtIOBlkPCI/vdev realize
+ hw/virtio/virtio-crypto-pci.c   -> VirtIOCryptoPCI/vdev realize
+ hw/virtio/virtio-input-pci.c    -> VirtIOInputPCI/vdev realize
+ hw/virtio/virtio-net-pci.c      -> VirtIONetPCI/vdev realize
+ hw/virtio/virtio-pmem-pci.c     -> VirtIOPMEMPCI/vdev realize
+ hw/virtio/virtio-rng-pci.c      -> VirtIORngPCI/vdev realize
+ hw/virtio/virtio-scsi-pci.c     -> VirtIOSCSIPCI/vdev realize
+ hw/virtio/virtio-serial-pci.c   -> VirtIOSerialPCI/vdev realize
---
 hw/core/qdev.c | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

Comments

Peter Maydell Nov. 29, 2019, 6:41 p.m. UTC | #1
On Fri, 18 Oct 2019 at 16:07, Damien Hedde <damien.hedde@greensocs.com> wrote:
>
> In qdev_set_parent_bus(), when changing the parent bus of a
> realized device, if the source and destination buses are not in the
> same reset state, some adaptation are required. This patch adds

"adaptations"

> needed call to resettable_change_parent() to make sure a device reset
> state stays coherent with its parent bus.
>
> The addition is a no-op if:
> 1. the device being parented is not realized.
> 2. the device is realized, but both buses are not under reset.
>
> Case 2 means that as long as qdev_set_parent_bus() is called
> during the machine realization procedure (which is before the
> machine reset so nothing is in reset), it is a no op.
>
> There are 49 call sites of qdev_set_parent_bus(). All but one fall
> into the no-op case:
> + 28 calls related to virtio (in hw/{s390x,display,virtio}/
>   {vhost,virtio}-xxx.c) to set a _vdev_/_vgpu_ composing device
>   parent bus just before realizing the _vdev_/_vgpu_.
> + hw/qdev.c: when creating a device in qdev_try_create()
> + hw/sysbus.c: when initializing a device in the sysbus
> + hw/display/virtio-gpu-pci.c: before realizing VirtIOGPUPCIBase/vgpu
> + hw/display/virtio-vga.c: before realizing VirtIOVGABase/vgpu
> + hw/i386/amd_iommu.c: before realizing AMDVIState/pci
> + hw/misc/auxbus.c: when creating an AUXBus
> + hw/misc/auxbus.c: when creating an AUXBus child
> + hw/misc/macio/macio.c: when initializing a MACIOState child
> + hw/misc/macio/macio.c: before realizing NewWorldMacIOState/pmu
> + hw/misc/macio/macio.c: before realizing NewWorldMacIOState/cuda
> + hw/pci-host/designware.c: before realizing DesignwarePCIEHost/root
> + hw/pci-host/gpex.c: before realizing GPEXHost/root
> + hw/pci-host/prep.c: when initializaing PREPPCIState/pci_dev
> + hw/pci-host/q35.c: before realizing Q35PCIHost/mch
> + hw/pci-host/versatile.c: when initializing PCIVPBState/pci_dev
> + hw/pci-host/xilinx-pcie.c: before realizing XilinxPCIEHost/root
> + hw/s390x/event-facility.c: when creating SCLPEventFacility/
>                              TYPE_SCLP_QUIESCE
> + hw/s390x/event-facility.c: ditto with SCLPEventFacility/
>                              TYPE_SCLP_CPU_HOTPLUG
> + hw/s390x/sclp.c: Not trivial because it is called on a SLCPDevice
>   just after realizing it. Ok because at this point the destination
>   bus (sysbus) is not in reset; the realize step is before the
>   machine reset.
> + hw/sd/core.c: Not OK. Used in sdbus_reparent_card(). See below.
> + hw/ssi/ssi.c: Used to put spi slave on spi bus and connect the cs
>   line in ssi_auto_connect_slave(). Ok because this function is only
>   used in realize step in hw/ssi/aspeed_smc.ci, hw/ssi/imx_spi.c,
>   hw/ssi/mss-spi.c, hw/ssi/xilinx_spi.c and hw/ssi/xilinx_spips.c.
> + hw/xen/xen-legacy-backend.c: when creating a XenLegacyDevice device
> + qdev-monitor.c: in device hotplug creation procedure before realize

This is a really useful analysis to have in the commit message;
thanks!

(Side note: I wonder if the sclp.c case could be reordered so
it realizes the device after parenting it? Anyway, not something
to worry about now.)

> Note that this commit alone will have no effect, right now there is no
> use of resettable API to reset anything. So a bus will never be tagged
> as in-reset by this same API.
>
> The one place where side-effect will occurs is in hw/sd/core.c in
> sdbus_reparent_card(). This function is only used in the raspi machines,
> including during the sysbus reset procedure. This case will be fixed by
> a following commit before globally enabling resettable API for sysbus
> reset.
>
> Signed-off-by: Damien Hedde <damien.hedde@greensocs.com>

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM
Cornelia Huck Dec. 3, 2019, 11:37 a.m. UTC | #2
On Fri, 29 Nov 2019 18:41:26 +0000
Peter Maydell <peter.maydell@linaro.org> wrote:

> On Fri, 18 Oct 2019 at 16:07, Damien Hedde <damien.hedde@greensocs.com> wrote:
> >
> > In qdev_set_parent_bus(), when changing the parent bus of a
> > realized device, if the source and destination buses are not in the
> > same reset state, some adaptation are required. This patch adds  
> 
> "adaptations"
> 
> > needed call to resettable_change_parent() to make sure a device reset
> > state stays coherent with its parent bus.
> >
> > The addition is a no-op if:
> > 1. the device being parented is not realized.
> > 2. the device is realized, but both buses are not under reset.
> >
> > Case 2 means that as long as qdev_set_parent_bus() is called
> > during the machine realization procedure (which is before the
> > machine reset so nothing is in reset), it is a no op.
> >
> > There are 49 call sites of qdev_set_parent_bus(). All but one fall
> > into the no-op case:
> > + 28 calls related to virtio (in hw/{s390x,display,virtio}/
> >   {vhost,virtio}-xxx.c) to set a _vdev_/_vgpu_ composing device
> >   parent bus just before realizing the _vdev_/_vgpu_.
> > + hw/qdev.c: when creating a device in qdev_try_create()
> > + hw/sysbus.c: when initializing a device in the sysbus
> > + hw/display/virtio-gpu-pci.c: before realizing VirtIOGPUPCIBase/vgpu
> > + hw/display/virtio-vga.c: before realizing VirtIOVGABase/vgpu
> > + hw/i386/amd_iommu.c: before realizing AMDVIState/pci
> > + hw/misc/auxbus.c: when creating an AUXBus
> > + hw/misc/auxbus.c: when creating an AUXBus child
> > + hw/misc/macio/macio.c: when initializing a MACIOState child
> > + hw/misc/macio/macio.c: before realizing NewWorldMacIOState/pmu
> > + hw/misc/macio/macio.c: before realizing NewWorldMacIOState/cuda
> > + hw/pci-host/designware.c: before realizing DesignwarePCIEHost/root
> > + hw/pci-host/gpex.c: before realizing GPEXHost/root
> > + hw/pci-host/prep.c: when initializaing PREPPCIState/pci_dev
> > + hw/pci-host/q35.c: before realizing Q35PCIHost/mch
> > + hw/pci-host/versatile.c: when initializing PCIVPBState/pci_dev
> > + hw/pci-host/xilinx-pcie.c: before realizing XilinxPCIEHost/root
> > + hw/s390x/event-facility.c: when creating SCLPEventFacility/
> >                              TYPE_SCLP_QUIESCE
> > + hw/s390x/event-facility.c: ditto with SCLPEventFacility/
> >                              TYPE_SCLP_CPU_HOTPLUG
> > + hw/s390x/sclp.c: Not trivial because it is called on a SLCPDevice
> >   just after realizing it. Ok because at this point the destination
> >   bus (sysbus) is not in reset; the realize step is before the
> >   machine reset.
> > + hw/sd/core.c: Not OK. Used in sdbus_reparent_card(). See below.
> > + hw/ssi/ssi.c: Used to put spi slave on spi bus and connect the cs
> >   line in ssi_auto_connect_slave(). Ok because this function is only
> >   used in realize step in hw/ssi/aspeed_smc.ci, hw/ssi/imx_spi.c,
> >   hw/ssi/mss-spi.c, hw/ssi/xilinx_spi.c and hw/ssi/xilinx_spips.c.
> > + hw/xen/xen-legacy-backend.c: when creating a XenLegacyDevice device
> > + qdev-monitor.c: in device hotplug creation procedure before realize  
> 
> This is a really useful analysis to have in the commit message;
> thanks!
> 
> (Side note: I wonder if the sclp.c case could be reordered so
> it realizes the device after parenting it? Anyway, not something
> to worry about now.)

As far as I can see, that should work. This code is a bit weird anyway;
the problem is that we need the sysbus somewhere in there... I'm
wondering if that can be handled in a different way. But agreed, that
is something we can revisit later.

> 
> > Note that this commit alone will have no effect, right now there is no
> > use of resettable API to reset anything. So a bus will never be tagged
> > as in-reset by this same API.
> >
> > The one place where side-effect will occurs is in hw/sd/core.c in
> > sdbus_reparent_card(). This function is only used in the raspi machines,
> > including during the sysbus reset procedure. This case will be fixed by
> > a following commit before globally enabling resettable API for sysbus
> > reset.
> >
> > Signed-off-by: Damien Hedde <damien.hedde@greensocs.com>  
> 
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> 
> thanks
> -- PMM
>
diff mbox series

Patch

diff --git a/hw/core/qdev.c b/hw/core/qdev.c
index 81784c73fb..3933f62d0c 100644
--- a/hw/core/qdev.c
+++ b/hw/core/qdev.c
@@ -96,23 +96,29 @@  static void bus_add_child(BusState *bus, DeviceState *child)
 
 void qdev_set_parent_bus(DeviceState *dev, BusState *bus)
 {
-    bool replugging = dev->parent_bus != NULL;
+    BusState *old_parent_bus = dev->parent_bus;
 
-    if (replugging) {
+    if (old_parent_bus) {
         trace_qdev_update_parent_bus(dev, dev->parent_bus, bus);
         /*
          * Keep a reference to the device while it's not plugged into
          * any bus, to avoid it potentially evaporating when it is
          * dereffed in bus_remove_child().
+         * Also keep the ref of the parent bus until the end, so that
+         * we can safely call resettable_change_parent() below.
          */
         object_ref(OBJECT(dev));
         bus_remove_child(dev->parent_bus, dev);
-        object_unref(OBJECT(dev->parent_bus));
     }
     dev->parent_bus = bus;
     object_ref(OBJECT(bus));
     bus_add_child(bus, dev);
-    if (replugging) {
+    if (dev->realized) {
+        resettable_change_parent(OBJECT(dev), OBJECT(bus),
+                                 OBJECT(old_parent_bus));
+    }
+    if (old_parent_bus) {
+        object_unref(OBJECT(old_parent_bus));
         object_unref(OBJECT(dev));
     }
 }