From: Marcelo Tosatti <mtosatti@redhat.com>
To: qemu-devel <qemu-devel@nongnu.org>, kvm-devel <kvm@vger.kernel.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>,
Vitaly Kuznetsov <vkuznets@redhat.com>,
Eduardo Habkost <ehabkost@redhat.com>
Subject: [PATCH] kvm: i386: allow TSC to differ by NTP correction bounds without TSC scaling
Date: Mon, 15 Jun 2020 09:01:27 -0300 [thread overview]
Message-ID: <20200615120127.GB224592@fuller.cnet> (raw)
The Linux TSC calibration procedure is subject to small variations
(its common to see +-1 kHz difference between reboots on a given CPU, for example).
So migrating a guest between two hosts with identical processor can fail, in case
of a small variation in calibrated TSC between them.
Allow a conservative 250ppm error between host TSC and VM TSC frequencies,
rather than requiring an exact match. NTP daemon in the guest can
correct this difference.
Also change migration to accept this bound.
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
diff --git a/target/i386/kvm.c b/target/i386/kvm.c
index 34f8387..257fee4 100644
--- a/target/i386/kvm.c
+++ b/target/i386/kvm.c
@@ -739,23 +739,44 @@ static int kvm_arch_set_tsc_khz(CPUState *cs)
{
X86CPU *cpu = X86_CPU(cs);
CPUX86State *env = &cpu->env;
- int r;
+ int r, cur_freq;
+ bool set_ioctl = false;
if (!env->tsc_khz) {
return 0;
}
- r = kvm_check_extension(cs->kvm_state, KVM_CAP_TSC_CONTROL) ?
+ cur_freq = kvm_check_extension(cs->kvm_state, KVM_CAP_GET_TSC_KHZ) ?
+ kvm_vcpu_ioctl(cs, KVM_GET_TSC_KHZ) : -ENOTSUP;
+
+ /*
+ * If TSC scaling is supported, attempt to set TSC frequency.
+ */
+ if (kvm_check_extension(cs->kvm_state, KVM_CAP_TSC_CONTROL)) {
+ set_ioctl = true;
+ }
+
+ /*
+ * If desired TSC frequency is within bounds of NTP correction,
+ * attempt to set TSC frequency.
+ */
+ if (cur_freq != -ENOTSUP && freq_within_bounds(cur_freq, env->tsc_khz)) {
+ set_ioctl = true;
+ }
+
+ r = set_ioctl ?
kvm_vcpu_ioctl(cs, KVM_SET_TSC_KHZ, env->tsc_khz) :
-ENOTSUP;
+
if (r < 0) {
/* When KVM_SET_TSC_KHZ fails, it's an error only if the current
* TSC frequency doesn't match the one we want.
*/
- int cur_freq = kvm_check_extension(cs->kvm_state, KVM_CAP_GET_TSC_KHZ) ?
- kvm_vcpu_ioctl(cs, KVM_GET_TSC_KHZ) :
- -ENOTSUP;
- if (cur_freq <= 0 || cur_freq != env->tsc_khz) {
+ cur_freq = kvm_check_extension(cs->kvm_state, KVM_CAP_GET_TSC_KHZ) ?
+ kvm_vcpu_ioctl(cs, KVM_GET_TSC_KHZ) :
+ -ENOTSUP;
+ if (cur_freq <= 0 || (cur_freq != env->tsc_khz &&
+ !freq_within_bounds(cur_freq, env->tsc_khz))) {
warn_report("TSC frequency mismatch between "
"VM (%" PRId64 " kHz) and host (%d kHz), "
"and TSC scaling unavailable",
diff --git a/target/i386/kvm_i386.h b/target/i386/kvm_i386.h
index 00bde7a..ebf9a64 100644
--- a/target/i386/kvm_i386.h
+++ b/target/i386/kvm_i386.h
@@ -47,4 +47,20 @@ bool kvm_has_x2apic_api(void);
bool kvm_hv_vpindex_settable(void);
+/*
+ * Check whether target_freq is within conservative
+ * ntp correctable bounds (250ppm) of freq
+ */
+static inline bool freq_within_bounds(int freq, int target_freq)
+{
+ int max_freq = freq + (freq * 250 / 1000000);
+ int min_freq = freq - (freq * 250 / 1000000);
+
+ if (target_freq >= min_freq && target_freq <= max_freq) {
+ return true;
+ }
+
+ return false;
+}
+
#endif
diff --git a/target/i386/machine.c b/target/i386/machine.c
index 0c96531..b052654 100644
--- a/target/i386/machine.c
+++ b/target/i386/machine.c
@@ -300,7 +300,7 @@ static int cpu_post_load(void *opaque, int version_id)
int i;
if (env->tsc_khz && env->user_tsc_khz &&
- env->tsc_khz != env->user_tsc_khz) {
+ !freq_within_bounds(env->tsc_khz, env->user_tsc_khz)) {
error_report("Mismatch between user-specified TSC frequency and "
"migrated TSC frequency");
return -EINVAL;
next reply other threads:[~2020-06-15 12:12 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-06-15 12:01 Marcelo Tosatti [this message]
2020-06-16 16:58 ` [PATCH v2] kvm: i386: allow TSC to differ by NTP correction bounds without TSC scaling Marcelo Tosatti
2020-06-16 21:35 ` no-reply
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200615120127.GB224592@fuller.cnet \
--to=mtosatti@redhat.com \
--cc=ehabkost@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=vkuznets@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).