[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [PATCH v5 24/30] ARM: vITS: handle MOVI command
On Thu, 6 Apr 2017, Andre Przywara wrote: > On 06/04/17 01:47, Stefano Stabellini wrote: > > On Thu, 6 Apr 2017, Andre Przywara wrote: > >> The MOVI command moves the interrupt affinity from one redistributor > >> (read: VCPU) to another. > >> For now migration of "live" LPIs is not yet implemented, but we store > >> the changed affinity in the host LPI structure and in our virtual ITTE. > >> > >> Signed-off-by: Andre Przywara <andre.przywara@xxxxxxx> > >> --- > >> xen/arch/arm/gic-v3-its.c | 24 ++++++++++++++++++++ > >> xen/arch/arm/gic-v3-lpi.c | 15 +++++++++++++ > >> xen/arch/arm/vgic-v3-its.c | 47 > >> ++++++++++++++++++++++++++++++++++++++++ > >> xen/include/asm-arm/gic_v3_its.h | 4 ++++ > >> 4 files changed, 90 insertions(+) > >> > >> diff --git a/xen/arch/arm/gic-v3-its.c b/xen/arch/arm/gic-v3-its.c > >> index d970119..a57e63a 100644 > >> --- a/xen/arch/arm/gic-v3-its.c > >> +++ b/xen/arch/arm/gic-v3-its.c > >> @@ -851,6 +851,30 @@ struct pending_irq *gicv3_assign_guest_event(struct > >> domain *d, > >> return pirq; > >> } > >> > >> +/* Changes the target VCPU for a given host LPI assigned to a domain. */ > >> +int gicv3_lpi_change_vcpu(struct domain *d, paddr_t vdoorbell, > >> + uint32_t vdevid, uint32_t veventid, > >> + unsigned int vcpu_id) > >> +{ > >> + uint32_t host_lpi; > >> + struct its_devices *dev; > >> + > >> + spin_lock(&d->arch.vgic.its_devices_lock); > >> + dev = get_its_device(d, vdoorbell, vdevid); > >> + if ( dev ) > >> + host_lpi = get_host_lpi(dev, veventid); > >> + else > >> + host_lpi = 0; > >> + spin_unlock(&d->arch.vgic.its_devices_lock); > >> + > >> + if ( !host_lpi ) > >> + return -ENOENT; > >> + > >> + gicv3_lpi_update_host_vcpuid(host_lpi, vcpu_id); > > > > we need to call vgic_migrate_irq > > Mmmh, are you sure? The very first statement there reads: > > /* nothing to do for virtual interrupts */ > if ( p->desc == NULL ) > return; > > Also my understanding of this command is that it only affects future > MSIs, so ITS translations. If an MSI has already been translated to a > certain redistributor and the ITS signaled it already, this command has > no effect on this particular one. > Also we don't care about benign races, so if this command comes in a tad > to late for a just happening MSI, this is the same problem on real hardware. It is more complicated than I thought. I would add a in-code comment here like this: TODO: we do not change physical irq affinity, in response to a virtual movi command. In other words, the physical LPI will still be delivered to the same pcpu. In addition to the comment, I wouldn't mind if also a warning was printed (with gdprintk). > >> + return 0; > >> +} > >> + > >> /* Scan the DT for any ITS nodes and create a list of host ITSes out of > >> it. */ > >> void gicv3_its_dt_init(const struct dt_device_node *node) > >> { > >> diff --git a/xen/arch/arm/gic-v3-lpi.c b/xen/arch/arm/gic-v3-lpi.c > >> index c997ed5..b9960aa 100644 > >> --- a/xen/arch/arm/gic-v3-lpi.c > >> +++ b/xen/arch/arm/gic-v3-lpi.c > >> @@ -234,6 +234,21 @@ void gicv3_lpi_update_host_entry(uint32_t host_lpi, > >> int domain_id, > >> write_u64_atomic(&hlpip->data, hlpi.data); > >> } > >> > >> +int gicv3_lpi_update_host_vcpuid(uint32_t host_lpi, unsigned int vcpu_id) > >> +{ > >> + union host_lpi *hlpip; > >> + > >> + ASSERT(host_lpi >= LPI_OFFSET); > >> + > >> + host_lpi -= LPI_OFFSET; > >> + > >> + hlpip = &lpi_data.host_lpis[host_lpi / HOST_LPIS_PER_PAGE][host_lpi % > >> HOST_LPIS_PER_PAGE]; > >> + > >> + write_u16_atomic(&hlpip->vcpu_id, vcpu_id); > >> + > >> + return 0; > >> +} > >> + > >> static int gicv3_lpi_allocate_pendtable(uint64_t *reg) > >> { > >> uint64_t val; > >> diff --git a/xen/arch/arm/vgic-v3-its.c b/xen/arch/arm/vgic-v3-its.c > >> index 079dd44..6afb915 100644 > >> --- a/xen/arch/arm/vgic-v3-its.c > >> +++ b/xen/arch/arm/vgic-v3-its.c > >> @@ -508,6 +508,47 @@ static int its_handle_mapti(struct virt_its *its, > >> uint64_t *cmdptr) > >> return 0; > >> } > >> > >> +static int its_handle_movi(struct virt_its *its, uint64_t *cmdptr) > >> +{ > >> + uint32_t devid = its_cmd_get_deviceid(cmdptr); > >> + uint32_t eventid = its_cmd_get_id(cmdptr); > >> + int collid = its_cmd_get_collection(cmdptr); > >> + struct pending_irq *p; > >> + struct vcpu *vcpu; > >> + uint32_t vlpi; > >> + int ret = -1; > >> + > >> + spin_lock(&its->its_lock); > >> + /* Check for a mapped LPI and get the LPI number. */ > >> + if ( !read_itte_locked(its, devid, eventid, &vcpu, &vlpi) ) > >> + goto out_unlock; > >> + > >> + /* Check the new collection ID and get the new VCPU pointer */ > >> + vcpu = get_vcpu_from_collection(its, collid); > >> + if ( !vcpu ) > >> + goto out_unlock; > >> + > >> + /* Update our cached vcpu_id in the pending_irq. */ > >> + p = its->d->arch.vgic.handler->lpi_to_pending(its->d, vlpi); > >> + p->vcpu_id = vcpu->vcpu_id; > >> + > >> + /* Now store the new collection in the translation table. */ > >> + if ( !write_itte_locked(its, devid, eventid, collid, vlpi, &vcpu) ) > >> + goto out_unlock; > >> + > >> + spin_unlock(&its->its_lock); > >> + > >> + /* TODO: lookup currently-in-guest virtual IRQs and migrate them? */ > >> + > >> + return gicv3_lpi_change_vcpu(its->d, its->doorbell_address, > >> + devid, eventid, vcpu->vcpu_id); > >> + > >> +out_unlock: > >> + spin_unlock(&its->its_lock); > >> + > >> + return ret; > >> +} > >> + > >> #define ITS_CMD_BUFFER_SIZE(baser) ((((baser) & 0xff) + 1) << 12) > >> > >> /* > >> @@ -552,6 +593,12 @@ static int vgic_its_handle_cmds(struct domain *d, > >> struct virt_its *its) > >> case GITS_CMD_MAPTI: > >> ret = its_handle_mapti(its, command); > >> break; > >> + case GITS_CMD_MOVALL: > >> + gdprintk(XENLOG_G_INFO, "ITS: ignoring MOVALL command\n"); > >> + break; > >> + case GITS_CMD_MOVI: > >> + ret = its_handle_movi(its, command); > >> + break; > >> case GITS_CMD_SYNC: > >> /* We handle ITS commands synchronously, so we ignore SYNC. */ > >> break; > >> diff --git a/xen/include/asm-arm/gic_v3_its.h > >> b/xen/include/asm-arm/gic_v3_its.h > >> index 30aa1ef..daae143 100644 > >> --- a/xen/include/asm-arm/gic_v3_its.h > >> +++ b/xen/include/asm-arm/gic_v3_its.h > >> @@ -177,8 +177,12 @@ void gicv3_free_host_lpi_block(uint32_t first_lpi); > >> struct pending_irq *gicv3_assign_guest_event(struct domain *d, paddr_t > >> doorbell, > >> uint32_t devid, uint32_t > >> eventid, > >> struct vcpu *v, uint32_t > >> virt_lpi); > >> +int gicv3_lpi_change_vcpu(struct domain *d, paddr_t doorbell, > >> + uint32_t devid, uint32_t eventid, > >> + unsigned int vcpu_id); > >> void gicv3_lpi_update_host_entry(uint32_t host_lpi, int domain_id, > >> unsigned int vcpu_id, uint32_t virt_lpi); > >> +int gicv3_lpi_update_host_vcpuid(uint32_t host_lpi, unsigned int vcpu_id); > >> > >> #else > >> > >> -- > >> 2.8.2 > >> > _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxx https://lists.xen.org/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |