[xen master] ARM: new VGIC: vgic-init: implement vgic_init

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[xen master] ARM: new VGIC: vgic-init: implement vgic_init

patchbot
commit a373635990073ee3cc8df2dbb9696c868fa07114
Author:     Andre Przywara <[hidden email]>
AuthorDate: Thu Feb 8 18:40:45 2018 +0000
Commit:     Stefano Stabellini <[hidden email]>
CommitDate: Thu Mar 29 10:37:10 2018 -0700

    ARM: new VGIC: vgic-init: implement vgic_init
   
    This patch allocates and initializes the data structures used to model
    the vgic distributor and virtual cpu interfaces. At that stage the
    number of IRQs and number of virtual CPUs is frozen.
    Implement the various functions that the Xen arch code is expecting to
    call during domain and VCPU setup to initialize the VGIC.
    Their prototypes are already in existing header files.
   
    This is based on Linux commit ad275b8bb1e6, written by Eric Auger.
   
    Signed-off-by: Andre Przywara <[hidden email]>
    Reviewed-by: Julien Grall <[hidden email]>
    Acked-by: Stefano Stabellini <[hidden email]>
---
 xen/arch/arm/vgic/vgic-init.c | 201 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 201 insertions(+)

diff --git a/xen/arch/arm/vgic/vgic-init.c b/xen/arch/arm/vgic/vgic-init.c
index d091c92ed0..bfd3d09edb 100644
--- a/xen/arch/arm/vgic/vgic-init.c
+++ b/xen/arch/arm/vgic/vgic-init.c
@@ -15,11 +15,83 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <xen/lib.h>
 #include <xen/sched.h>
 #include <asm/new_vgic.h>
 
 #include "vgic.h"
 
+/*
+ * Initialization rules: there are multiple stages to the vgic
+ * initialization, both for the distributor and the CPU interfaces.  The basic
+ * idea is that even though the VGIC is not functional or not requested from
+ * user space, the critical path of the run loop can still call VGIC functions
+ * that just won't do anything, without them having to check additional
+ * initialization flags to ensure they don't look at uninitialized data
+ * structures.
+ *
+ * Distributor:
+ *
+ * - vgic_early_init(): initialization of static data that doesn't
+ *   depend on any sizing information or emulation type. No allocation
+ *   is allowed there.
+ *
+ * - vgic_init(): allocation and initialization of the generic data
+ *   structures that depend on sizing information (number of CPUs,
+ *   number of interrupts). Also initializes the vcpu specific data
+ *   structures. Can be executed lazily for GICv2.
+ *
+ * CPU Interface:
+ *
+ * - vgic_vcpu_early_init(): initialization of static data that
+ *   doesn't depend on any sizing information or emulation type. No
+ *   allocation is allowed there.
+ */
+
+/**
+ * vgic_vcpu_early_init() - Initialize static VGIC VCPU data structures
+ * @vcpu: The VCPU whose VGIC data structures whould be initialized
+ *
+ * Only do initialization, but do not actually enable the VGIC CPU interface
+ * yet.
+ */
+static void vgic_vcpu_early_init(struct vcpu *vcpu)
+{
+    struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic;
+    unsigned int i;
+
+    INIT_LIST_HEAD(&vgic_cpu->ap_list_head);
+    spin_lock_init(&vgic_cpu->ap_list_lock);
+
+    /*
+     * Enable and configure all SGIs to be edge-triggered and
+     * configure all PPIs as level-triggered.
+     */
+    for ( i = 0; i < VGIC_NR_PRIVATE_IRQS; i++ )
+    {
+        struct vgic_irq *irq = &vgic_cpu->private_irqs[i];
+
+        INIT_LIST_HEAD(&irq->ap_list);
+        spin_lock_init(&irq->irq_lock);
+        irq->intid = i;
+        irq->vcpu = NULL;
+        irq->target_vcpu = vcpu;
+        irq->targets = 1U << vcpu->vcpu_id;
+        atomic_set(&irq->refcount, 0);
+        if ( vgic_irq_is_sgi(i) )
+        {
+            /* SGIs */
+            irq->enabled = 1;
+            irq->config = VGIC_CONFIG_EDGE;
+        }
+        else
+        {
+            /* PPIs */
+            irq->config = VGIC_CONFIG_LEVEL;
+        }
+    }
+}
+
 /* CREATION */
 
 /**
@@ -50,6 +122,135 @@ int domain_vgic_register(struct domain *d, int *mmio_count)
     return 0;
 }
 
+/* INIT/DESTROY */
+
+/**
+ * domain_vgic_init: initialize the dist data structures
+ * @d: domain pointer
+ * @nr_spis: number of SPIs
+ */
+int domain_vgic_init(struct domain *d, unsigned int nr_spis)
+{
+    struct vgic_dist *dist = &d->arch.vgic;
+    unsigned int i;
+    int ret;
+
+    /* The number of SPIs must be a multiple of 32 per the GIC spec. */
+    nr_spis = ROUNDUP(nr_spis, 32);
+
+    /* Limit the number of virtual SPIs supported to (1020 - 32) = 988  */
+    if ( nr_spis > (1020 - NR_LOCAL_IRQS) )
+        return -EINVAL;
+
+    dist->nr_spis = nr_spis;
+    dist->spis = xzalloc_array(struct vgic_irq, nr_spis);
+    if ( !dist->spis )
+        return  -ENOMEM;
+
+    /*
+     * In the following code we do not take the irq struct lock since
+     * no other action on irq structs can happen while the VGIC is
+     * not initialized yet:
+     * If someone wants to inject an interrupt or does a MMIO access, we
+     * require prior initialization in case of a virtual GICv3 or trigger
+     * initialization when using a virtual GICv2.
+     */
+    for ( i = 0; i < nr_spis; i++ )
+    {
+        struct vgic_irq *irq = &dist->spis[i];
+
+        irq->intid = i + VGIC_NR_PRIVATE_IRQS;
+        INIT_LIST_HEAD(&irq->ap_list);
+        spin_lock_init(&irq->irq_lock);
+        irq->vcpu = NULL;
+        irq->target_vcpu = NULL;
+        atomic_set(&irq->refcount, 0);
+        if ( dist->version == GIC_V2 )
+            irq->targets = 0;
+        else
+            irq->mpidr = 0;
+    }
+
+    INIT_LIST_HEAD(&dist->lpi_list_head);
+    spin_lock_init(&dist->lpi_list_lock);
+
+    if ( dist->version == GIC_V2 )
+        ret = vgic_v2_map_resources(d);
+    else
+        ret = -ENXIO;
+
+    if ( ret )
+        return ret;
+
+    /* allocated_irqs() is used by Xen to find available vIRQs */
+    d->arch.vgic.allocated_irqs =
+        xzalloc_array(unsigned long, BITS_TO_LONGS(vgic_num_irqs(d)));
+    if ( !d->arch.vgic.allocated_irqs )
+        return -ENOMEM;
+
+    /* vIRQ0-15 (SGIs) are reserved */
+    for ( i = 0; i < NR_GIC_SGI; i++ )
+        set_bit(i, d->arch.vgic.allocated_irqs);
+
+    return 0;
+}
+
+/**
+ * vcpu_vgic_init() - Register VCPU-specific KVM iodevs
+ * was: kvm_vgic_vcpu_init()
+ * Xen: adding vgic_vx_enable() call
+ * @vcpu: pointer to the VCPU being created and initialized
+ */
+int vcpu_vgic_init(struct vcpu *vcpu)
+{
+    int ret = 0;
+
+    vgic_vcpu_early_init(vcpu);
+
+    if ( gic_hw_version() == GIC_V2 )
+        vgic_v2_enable(vcpu);
+    else
+        ret = -ENXIO;
+
+    return ret;
+}
+
+void domain_vgic_free(struct domain *d)
+{
+    struct vgic_dist *dist = &d->arch.vgic;
+        int i, ret;
+
+    for ( i = 0; i < dist->nr_spis; i++ )
+    {
+        struct vgic_irq *irq = vgic_get_irq(d, NULL, 32 + i);
+
+        if ( !irq->hw )
+            continue;
+
+        ret = release_guest_irq(d, irq->hwintid);
+        if ( ret )
+            dprintk(XENLOG_G_WARNING,
+                    "d%u: Failed to release virq %u ret = %d\n",
+                    d->domain_id, 32 + i, ret);
+    }
+
+    dist->ready = false;
+    dist->initialized = false;
+
+    xfree(dist->spis);
+    xfree(dist->allocated_irqs);
+    dist->nr_spis = 0;
+}
+
+int vcpu_vgic_free(struct vcpu *vcpu)
+{
+    struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic;
+
+    INIT_LIST_HEAD(&vgic_cpu->ap_list_head);
+
+    return 0;
+}
+
 /*
  * Local variables:
  * mode: C
--
generated by git-patchbot for /home/xen/git/xen.git#master

_______________________________________________
Xen-changelog mailing list
[hidden email]
https://lists.xenproject.org/xen-changelog