051089a2ee
-----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAABAgAGBQJaDdh4AAoJELDendYovxMvPFAH/2QjTys2ydIAdmwke4odpJ7U xuy7HOQCzOeZ5YsZthzCBsN90VmnDM7X7CcB8weSdjcKlXMSAWD+J1RgkL2iAJhI 8tzIEXECrlNuz4V5mX9TmMgtPCr4qzU3fsts0pZy4fYDq1PVWDefqOwEtbpbWabb wRSMq/nTb9iASTMgheSC0WfhJneqtJ+J20zrzkGPCBPRFcwfppeP8/7vpkmJslBi eH/pfchICM4w093T/BfavnsPvhLdjgRuwVzn6+e46s4tLnZAxnLRVQ7SXZXzBORq /dL/qC0XH3YXdU+XfIs//giZsmLns6SxZaMr4vs6TxFtuzZBKpLtkOKo9zndvxk= =sZY5 -----END PGP SIGNATURE----- Merge tag 'for-linus-4.15-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip Pull xen updates from Juergen Gross: "Xen features and fixes for v4.15-rc1 Apart from several small fixes it contains the following features: - a series by Joao Martins to add vdso support of the pv clock interface - a series by Juergen Gross to add support for Xen pv guests to be able to run on 5 level paging hosts - a series by Stefano Stabellini adding the Xen pvcalls frontend driver using a paravirtualized socket interface" * tag 'for-linus-4.15-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip: (34 commits) xen/pvcalls: fix potential endless loop in pvcalls-front.c xen/pvcalls: Add MODULE_LICENSE() MAINTAINERS: xen, kvm: track pvclock-abi.h changes x86/xen/time: setup vcpu 0 time info page x86/xen/time: set pvclock flags on xen_time_init() x86/pvclock: add setter for pvclock_pvti_cpu0_va ptp_kvm: probe for kvm guest availability xen/privcmd: remove unused variable pageidx xen: select grant interface version xen: update arch/x86/include/asm/xen/cpuid.h xen: add grant interface version dependent constants to gnttab_ops xen: limit grant v2 interface to the v1 functionality xen: re-introduce support for grant v2 interface xen: support priv-mapping in an HVM tools domain xen/pvcalls: remove redundant check for irq >= 0 xen/pvcalls: fix unsigned less than zero error check xen/time: Return -ENODEV from xen_get_wallclock() xen/pvcalls-front: mark expected switch fall-through xen: xenbus_probe_frontend: mark expected switch fall-throughs xen/time: do not decrease steal time after live migration on xen ...
184 lines
4.1 KiB
C
184 lines
4.1 KiB
C
// SPDX-License-Identifier: GPL-2.0
|
|
/*
|
|
* Xen stolen ticks accounting.
|
|
*/
|
|
#include <linux/kernel.h>
|
|
#include <linux/kernel_stat.h>
|
|
#include <linux/math64.h>
|
|
#include <linux/gfp.h>
|
|
#include <linux/slab.h>
|
|
|
|
#include <asm/paravirt.h>
|
|
#include <asm/xen/hypervisor.h>
|
|
#include <asm/xen/hypercall.h>
|
|
|
|
#include <xen/events.h>
|
|
#include <xen/features.h>
|
|
#include <xen/interface/xen.h>
|
|
#include <xen/interface/vcpu.h>
|
|
#include <xen/xen-ops.h>
|
|
|
|
/* runstate info updated by Xen */
|
|
static DEFINE_PER_CPU(struct vcpu_runstate_info, xen_runstate);
|
|
|
|
static DEFINE_PER_CPU(u64[4], old_runstate_time);
|
|
|
|
/* return an consistent snapshot of 64-bit time/counter value */
|
|
static u64 get64(const u64 *p)
|
|
{
|
|
u64 ret;
|
|
|
|
if (BITS_PER_LONG < 64) {
|
|
u32 *p32 = (u32 *)p;
|
|
u32 h, l, h2;
|
|
|
|
/*
|
|
* Read high then low, and then make sure high is
|
|
* still the same; this will only loop if low wraps
|
|
* and carries into high.
|
|
* XXX some clean way to make this endian-proof?
|
|
*/
|
|
do {
|
|
h = READ_ONCE(p32[1]);
|
|
l = READ_ONCE(p32[0]);
|
|
h2 = READ_ONCE(p32[1]);
|
|
} while(h2 != h);
|
|
|
|
ret = (((u64)h) << 32) | l;
|
|
} else
|
|
ret = READ_ONCE(*p);
|
|
|
|
return ret;
|
|
}
|
|
|
|
static void xen_get_runstate_snapshot_cpu_delta(
|
|
struct vcpu_runstate_info *res, unsigned int cpu)
|
|
{
|
|
u64 state_time;
|
|
struct vcpu_runstate_info *state;
|
|
|
|
BUG_ON(preemptible());
|
|
|
|
state = per_cpu_ptr(&xen_runstate, cpu);
|
|
|
|
do {
|
|
state_time = get64(&state->state_entry_time);
|
|
rmb(); /* Hypervisor might update data. */
|
|
*res = READ_ONCE(*state);
|
|
rmb(); /* Hypervisor might update data. */
|
|
} while (get64(&state->state_entry_time) != state_time ||
|
|
(state_time & XEN_RUNSTATE_UPDATE));
|
|
}
|
|
|
|
static void xen_get_runstate_snapshot_cpu(struct vcpu_runstate_info *res,
|
|
unsigned int cpu)
|
|
{
|
|
int i;
|
|
|
|
xen_get_runstate_snapshot_cpu_delta(res, cpu);
|
|
|
|
for (i = 0; i < 4; i++)
|
|
res->time[i] += per_cpu(old_runstate_time, cpu)[i];
|
|
}
|
|
|
|
void xen_manage_runstate_time(int action)
|
|
{
|
|
static struct vcpu_runstate_info *runstate_delta;
|
|
struct vcpu_runstate_info state;
|
|
int cpu, i;
|
|
|
|
switch (action) {
|
|
case -1: /* backup runstate time before suspend */
|
|
if (unlikely(runstate_delta))
|
|
pr_warn_once("%s: memory leak as runstate_delta is not NULL\n",
|
|
__func__);
|
|
|
|
runstate_delta = kmalloc_array(num_possible_cpus(),
|
|
sizeof(*runstate_delta),
|
|
GFP_ATOMIC);
|
|
if (unlikely(!runstate_delta)) {
|
|
pr_warn("%s: failed to allocate runstate_delta\n",
|
|
__func__);
|
|
return;
|
|
}
|
|
|
|
for_each_possible_cpu(cpu) {
|
|
xen_get_runstate_snapshot_cpu_delta(&state, cpu);
|
|
memcpy(runstate_delta[cpu].time, state.time,
|
|
sizeof(runstate_delta[cpu].time));
|
|
}
|
|
|
|
break;
|
|
|
|
case 0: /* backup runstate time after resume */
|
|
if (unlikely(!runstate_delta)) {
|
|
pr_warn("%s: cannot accumulate runstate time as runstate_delta is NULL\n",
|
|
__func__);
|
|
return;
|
|
}
|
|
|
|
for_each_possible_cpu(cpu) {
|
|
for (i = 0; i < 4; i++)
|
|
per_cpu(old_runstate_time, cpu)[i] +=
|
|
runstate_delta[cpu].time[i];
|
|
}
|
|
|
|
break;
|
|
|
|
default: /* do not accumulate runstate time for checkpointing */
|
|
break;
|
|
}
|
|
|
|
if (action != -1 && runstate_delta) {
|
|
kfree(runstate_delta);
|
|
runstate_delta = NULL;
|
|
}
|
|
}
|
|
|
|
/*
|
|
* Runstate accounting
|
|
*/
|
|
void xen_get_runstate_snapshot(struct vcpu_runstate_info *res)
|
|
{
|
|
xen_get_runstate_snapshot_cpu(res, smp_processor_id());
|
|
}
|
|
|
|
/* return true when a vcpu could run but has no real cpu to run on */
|
|
bool xen_vcpu_stolen(int vcpu)
|
|
{
|
|
return per_cpu(xen_runstate, vcpu).state == RUNSTATE_runnable;
|
|
}
|
|
|
|
u64 xen_steal_clock(int cpu)
|
|
{
|
|
struct vcpu_runstate_info state;
|
|
|
|
xen_get_runstate_snapshot_cpu(&state, cpu);
|
|
return state.time[RUNSTATE_runnable] + state.time[RUNSTATE_offline];
|
|
}
|
|
|
|
void xen_setup_runstate_info(int cpu)
|
|
{
|
|
struct vcpu_register_runstate_memory_area area;
|
|
|
|
area.addr.v = &per_cpu(xen_runstate, cpu);
|
|
|
|
if (HYPERVISOR_vcpu_op(VCPUOP_register_runstate_memory_area,
|
|
xen_vcpu_nr(cpu), &area))
|
|
BUG();
|
|
}
|
|
|
|
void __init xen_time_setup_guest(void)
|
|
{
|
|
bool xen_runstate_remote;
|
|
|
|
xen_runstate_remote = !HYPERVISOR_vm_assist(VMASST_CMD_enable,
|
|
VMASST_TYPE_runstate_update_flag);
|
|
|
|
pv_time_ops.steal_clock = xen_steal_clock;
|
|
|
|
static_key_slow_inc(¶virt_steal_enabled);
|
|
if (xen_runstate_remote)
|
|
static_key_slow_inc(¶virt_steal_rq_enabled);
|
|
}
|