| Message ID | 154774530571.1208625.5318753891932747707.stgit@bahia.lan |
|---|---|
| State | New |
| Headers | show |
| Series | spapr: Add support for PHB hotplug | expand |
On 1/17/19 6:15 PM, Greg Kurz wrote: > Every PHB needs to claim 4 LSIs to support legacy PCI devices. This is > currently done at PHB realize. When using in-kernel XICS (or upcoming > in-kernel XIVE), QEMU synchronizes the state of all irqs, including > these LSIs, later on at machine reset. > > In order to support PHB hotplug, we need a way to tell KVM about the > LSIs that doesn't require a machine reset. Since these irq numbers are > fixed values derived from the PHB index, let's identify them all at > machine init. Older machines that don't have fixed irq numbers cannot > support PHB hotplug and keep the existing behavior. > > Signed-off-by: Greg Kurz <groug@kaod.org> > --- > hw/ppc/spapr.c | 6 ++++++ > hw/ppc/spapr_pci.c | 20 ++++++++++++++++++-- > include/hw/pci-host/spapr.h | 2 ++ > 3 files changed, 26 insertions(+), 2 deletions(-) > > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c > index 26f8e55cc25e..9189b4d3a9d6 100644 > --- a/hw/ppc/spapr.c > +++ b/hw/ppc/spapr.c > @@ -2802,6 +2802,12 @@ static void spapr_machine_init(MachineState *machine) > > phb = spapr_create_default_phb(); > > + if (!smc->legacy_irq_allocation) { > + for (i = 0; i < SPAPR_MAX_PHBS; i++) { > + spapr_phb_set_lsis(i, spapr); > + } > + } > + Can't we be more brutal and do the LSI setting in spapr_irq_init() for the whole [ SPAPR_IRQ_PCI_LSI ... SPAPR_IRQ_MSI -1 ] range ? C. > for (i = 0; i < nb_nics; i++) { > NICInfo *nd = &nd_table[i]; > > diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c > index ec9d4d28004c..f5f13a4d4816 100644 > --- a/hw/ppc/spapr_pci.c > +++ b/hw/ppc/spapr_pci.c > @@ -1559,6 +1559,22 @@ static void spapr_pci_unplug_request(HotplugHandler *plug_handler, > } > } > > +static uint32_t spapr_phb_index_to_lsi(int phb_index, int irq_index) > +{ > + return SPAPR_IRQ_PCI_LSI + phb_index * PCI_NUM_PINS + irq_index; > +} > + > +void spapr_phb_set_lsis(int index, sPAPRMachineState *spapr) > +{ > + int i; > + > + for (i = 0; i < PCI_NUM_PINS; i++) { > + uint32_t irq = spapr_phb_index_to_lsi(index, i); > + > + spapr_irq_set_type(spapr, irq, true); > + } > +} > + > static void spapr_phb_realize(DeviceState *dev, Error **errp) > { > /* We don't use SPAPR_MACHINE() in order to exit gracefully if the user > @@ -1726,7 +1742,7 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp) > > /* Initialize the LSI table */ > for (i = 0; i < PCI_NUM_PINS; i++) { > - uint32_t irq = SPAPR_IRQ_PCI_LSI + sphb->index * PCI_NUM_PINS + i; > + uint32_t irq = spapr_phb_index_to_lsi(sphb->index, i); > Error *local_err = NULL; > > if (smc->legacy_irq_allocation) { > @@ -1736,6 +1752,7 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp) > "can't allocate LSIs: "); > return; > } > + spapr_irq_set_type(spapr, irq, true); > } > > spapr_irq_claim(spapr, irq, &local_err); > @@ -1743,7 +1760,6 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp) > error_propagate_prepend(errp, local_err, "can't allocate LSIs: "); > return; > } > - spapr_irq_set_type(spapr, irq, true); > > sphb->lsi_table[i].irq = irq; > } > diff --git a/include/hw/pci-host/spapr.h b/include/hw/pci-host/spapr.h > index e0e683c32469..bb0ae7fdd41d 100644 > --- a/include/hw/pci-host/spapr.h > +++ b/include/hw/pci-host/spapr.h > @@ -164,4 +164,6 @@ static inline void spapr_phb_vfio_reset(DeviceState *qdev) > > void spapr_phb_dma_reset(sPAPRPHBState *sphb); > > +void spapr_phb_set_lsis(int index, sPAPRMachineState *spapr); > + > #endif /* PCI_HOST_SPAPR_H */ >
On Fri, 18 Jan 2019 13:38:02 +0100 Cédric Le Goater <clg@kaod.org> wrote: > On 1/17/19 6:15 PM, Greg Kurz wrote: > > Every PHB needs to claim 4 LSIs to support legacy PCI devices. This is > > currently done at PHB realize. When using in-kernel XICS (or upcoming > > in-kernel XIVE), QEMU synchronizes the state of all irqs, including > > these LSIs, later on at machine reset. > > > > In order to support PHB hotplug, we need a way to tell KVM about the > > LSIs that doesn't require a machine reset. Since these irq numbers are > > fixed values derived from the PHB index, let's identify them all at > > machine init. Older machines that don't have fixed irq numbers cannot > > support PHB hotplug and keep the existing behavior. > > > > Signed-off-by: Greg Kurz <groug@kaod.org> > > --- > > hw/ppc/spapr.c | 6 ++++++ > > hw/ppc/spapr_pci.c | 20 ++++++++++++++++++-- > > include/hw/pci-host/spapr.h | 2 ++ > > 3 files changed, 26 insertions(+), 2 deletions(-) > > > > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c > > index 26f8e55cc25e..9189b4d3a9d6 100644 > > --- a/hw/ppc/spapr.c > > +++ b/hw/ppc/spapr.c > > @@ -2802,6 +2802,12 @@ static void spapr_machine_init(MachineState *machine) > > > > phb = spapr_create_default_phb(); > > > > + if (!smc->legacy_irq_allocation) { > > + for (i = 0; i < SPAPR_MAX_PHBS; i++) { > > + spapr_phb_set_lsis(i, spapr); > > + } > > + } > > + > > Can't we be more brutal and do the LSI setting in spapr_irq_init() > for the whole [ SPAPR_IRQ_PCI_LSI ... SPAPR_IRQ_MSI -1 ] range ? > I like this. And we should even set the type of all IRQs for the machine lifetime since it is an invariant. I'll give it a try for v4. > C. > > > for (i = 0; i < nb_nics; i++) { > > NICInfo *nd = &nd_table[i]; > > > > diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c > > index ec9d4d28004c..f5f13a4d4816 100644 > > --- a/hw/ppc/spapr_pci.c > > +++ b/hw/ppc/spapr_pci.c > > @@ -1559,6 +1559,22 @@ static void spapr_pci_unplug_request(HotplugHandler *plug_handler, > > } > > } > > > > +static uint32_t spapr_phb_index_to_lsi(int phb_index, int irq_index) > > +{ > > + return SPAPR_IRQ_PCI_LSI + phb_index * PCI_NUM_PINS + irq_index; > > +} > > + > > +void spapr_phb_set_lsis(int index, sPAPRMachineState *spapr) > > +{ > > + int i; > > + > > + for (i = 0; i < PCI_NUM_PINS; i++) { > > + uint32_t irq = spapr_phb_index_to_lsi(index, i); > > + > > + spapr_irq_set_type(spapr, irq, true); > > + } > > +} > > + > > static void spapr_phb_realize(DeviceState *dev, Error **errp) > > { > > /* We don't use SPAPR_MACHINE() in order to exit gracefully if the user > > @@ -1726,7 +1742,7 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp) > > > > /* Initialize the LSI table */ > > for (i = 0; i < PCI_NUM_PINS; i++) { > > - uint32_t irq = SPAPR_IRQ_PCI_LSI + sphb->index * PCI_NUM_PINS + i; > > + uint32_t irq = spapr_phb_index_to_lsi(sphb->index, i); > > Error *local_err = NULL; > > > > if (smc->legacy_irq_allocation) { > > @@ -1736,6 +1752,7 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp) > > "can't allocate LSIs: "); > > return; > > } > > + spapr_irq_set_type(spapr, irq, true); > > } > > > > spapr_irq_claim(spapr, irq, &local_err); > > @@ -1743,7 +1760,6 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp) > > error_propagate_prepend(errp, local_err, "can't allocate LSIs: "); > > return; > > } > > - spapr_irq_set_type(spapr, irq, true); > > > > sphb->lsi_table[i].irq = irq; > > } > > diff --git a/include/hw/pci-host/spapr.h b/include/hw/pci-host/spapr.h > > index e0e683c32469..bb0ae7fdd41d 100644 > > --- a/include/hw/pci-host/spapr.h > > +++ b/include/hw/pci-host/spapr.h > > @@ -164,4 +164,6 @@ static inline void spapr_phb_vfio_reset(DeviceState *qdev) > > > > void spapr_phb_dma_reset(sPAPRPHBState *sphb); > > > > +void spapr_phb_set_lsis(int index, sPAPRMachineState *spapr); > > + > > #endif /* PCI_HOST_SPAPR_H */ > > >
diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c index 26f8e55cc25e..9189b4d3a9d6 100644 --- a/hw/ppc/spapr.c +++ b/hw/ppc/spapr.c @@ -2802,6 +2802,12 @@ static void spapr_machine_init(MachineState *machine) phb = spapr_create_default_phb(); + if (!smc->legacy_irq_allocation) { + for (i = 0; i < SPAPR_MAX_PHBS; i++) { + spapr_phb_set_lsis(i, spapr); + } + } + for (i = 0; i < nb_nics; i++) { NICInfo *nd = &nd_table[i]; diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c index ec9d4d28004c..f5f13a4d4816 100644 --- a/hw/ppc/spapr_pci.c +++ b/hw/ppc/spapr_pci.c @@ -1559,6 +1559,22 @@ static void spapr_pci_unplug_request(HotplugHandler *plug_handler, } } +static uint32_t spapr_phb_index_to_lsi(int phb_index, int irq_index) +{ + return SPAPR_IRQ_PCI_LSI + phb_index * PCI_NUM_PINS + irq_index; +} + +void spapr_phb_set_lsis(int index, sPAPRMachineState *spapr) +{ + int i; + + for (i = 0; i < PCI_NUM_PINS; i++) { + uint32_t irq = spapr_phb_index_to_lsi(index, i); + + spapr_irq_set_type(spapr, irq, true); + } +} + static void spapr_phb_realize(DeviceState *dev, Error **errp) { /* We don't use SPAPR_MACHINE() in order to exit gracefully if the user @@ -1726,7 +1742,7 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp) /* Initialize the LSI table */ for (i = 0; i < PCI_NUM_PINS; i++) { - uint32_t irq = SPAPR_IRQ_PCI_LSI + sphb->index * PCI_NUM_PINS + i; + uint32_t irq = spapr_phb_index_to_lsi(sphb->index, i); Error *local_err = NULL; if (smc->legacy_irq_allocation) { @@ -1736,6 +1752,7 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp) "can't allocate LSIs: "); return; } + spapr_irq_set_type(spapr, irq, true); } spapr_irq_claim(spapr, irq, &local_err); @@ -1743,7 +1760,6 @@ static void spapr_phb_realize(DeviceState *dev, Error **errp) error_propagate_prepend(errp, local_err, "can't allocate LSIs: "); return; } - spapr_irq_set_type(spapr, irq, true); sphb->lsi_table[i].irq = irq; } diff --git a/include/hw/pci-host/spapr.h b/include/hw/pci-host/spapr.h index e0e683c32469..bb0ae7fdd41d 100644 --- a/include/hw/pci-host/spapr.h +++ b/include/hw/pci-host/spapr.h @@ -164,4 +164,6 @@ static inline void spapr_phb_vfio_reset(DeviceState *qdev) void spapr_phb_dma_reset(sPAPRPHBState *sphb); +void spapr_phb_set_lsis(int index, sPAPRMachineState *spapr); + #endif /* PCI_HOST_SPAPR_H */
Every PHB needs to claim 4 LSIs to support legacy PCI devices. This is currently done at PHB realize. When using in-kernel XICS (or upcoming in-kernel XIVE), QEMU synchronizes the state of all irqs, including these LSIs, later on at machine reset. In order to support PHB hotplug, we need a way to tell KVM about the LSIs that doesn't require a machine reset. Since these irq numbers are fixed values derived from the PHB index, let's identify them all at machine init. Older machines that don't have fixed irq numbers cannot support PHB hotplug and keep the existing behavior. Signed-off-by: Greg Kurz <groug@kaod.org> --- hw/ppc/spapr.c | 6 ++++++ hw/ppc/spapr_pci.c | 20 ++++++++++++++++++-- include/hw/pci-host/spapr.h | 2 ++ 3 files changed, 26 insertions(+), 2 deletions(-)