[PATCHv2] x86/hyperv: Hold cpus_read_lock() on assigning reenlightenment vector

From: Dmitry Safonov
Date: Mon Jun 17 2019 - 12:44:57 EST

KVM support may be compiled as dynamic module, which triggers the
following splat on modprobe (under CONFIG_DEBUG_PREEMPT):

KVM: vmx: using Hyper-V Enlightened VMCS
BUG: using smp_processor_id() in preemptible [00000000] code: modprobe/466
caller is debug_smp_processor_id+0x17/0x19
CPU: 0 PID: 466 Comm: modprobe Kdump: loaded Not tainted 4.19.43 #1
Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS 090007 06/02/2017
Call Trace:
kvm_arch_init+0x14a/0x163 [kvm]
kvm_init+0x30/0x259 [kvm]
vmx_init+0xed/0x3db [kvm_intel]

Hold cpus_read_lock() so that MSR will be written for an online CPU,
even if set_hv_tscchange_cb() gets being preempted.
While at it, cleanup smp_processor_id()'s in hv_cpu_init() and add a
lockdep assert into hv_cpu_die().

Fixes: 93286261de1b4 ("x86/hyperv: Reenlightenment notifications

Cc: Andy Lutomirski <luto@xxxxxxxxxx>
Cc: Borislav Petkov <bp@xxxxxxxxx>
Cc: Cathy Avery <cavery@xxxxxxxxxx>
Cc: Haiyang Zhang <haiyangz@xxxxxxxxxxxxx>
Cc: "H. Peter Anvin" <hpa@xxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: "K. Y. Srinivasan" <kys@xxxxxxxxxxxxx>
Cc: "Michael Kelley (EOSG)" <Michael.H.Kelley@xxxxxxxxxxxxx>
Cc: Mohammed Gamal <mmorsy@xxxxxxxxxx>
Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Radim KrÄmÃÅ <rkrcmar@xxxxxxxxxx>
Cc: Roman Kagan <rkagan@xxxxxxxxxxxxx>
Cc: Sasha Levin <sashal@xxxxxxxxxx>
Cc: Stephen Hemminger <sthemmin@xxxxxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>

Cc: devel@xxxxxxxxxxxxxxxxxxxxxx
Cc: kvm@xxxxxxxxxxxxxxx
Cc: linux-hyperv@xxxxxxxxxxxxxxx
Cc: x86@xxxxxxxxxx
Reported-by: Prasanna Panchamukhi <panchamukhi@xxxxxxxxxx>
Signed-off-by: Dmitry Safonov <dima@xxxxxxxxxx>
v1 link: lkml.kernel.org/r/20190611212003.26382-1-dima@xxxxxxxxxx

NOTE that I hadn't a chance to test v2 on hyperv machine so far,
ONLY BUILD TESTED. (In hope that the patch still makes sense and Kbuild
bot will report any issue).

arch/x86/hyperv/hv_init.c | 16 +++++++++++++---
1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/arch/x86/hyperv/hv_init.c b/arch/x86/hyperv/hv_init.c
index 1608050e9df9..ec7fd7d6c125 100644
--- a/arch/x86/hyperv/hv_init.c
+++ b/arch/x86/hyperv/hv_init.c
@@ -20,6 +20,7 @@
#include <linux/clockchips.h>
#include <linux/hyperv.h>
#include <linux/slab.h>
+#include <linux/cpu.h>
#include <linux/cpuhotplug.h>

@@ -91,7 +92,7 @@ EXPORT_SYMBOL_GPL(hv_max_vp_index);
static int hv_cpu_init(unsigned int cpu)
u64 msr_vp_index;
- struct hv_vp_assist_page **hvp = &hv_vp_assist_page[smp_processor_id()];
+ struct hv_vp_assist_page **hvp = &hv_vp_assist_page[cpu];
void **input_arg;
struct page *pg;

@@ -103,7 +104,7 @@ static int hv_cpu_init(unsigned int cpu)


- hv_vp_index[smp_processor_id()] = msr_vp_index;
+ hv_vp_index[cpu] = msr_vp_index;

if (msr_vp_index > hv_max_vp_index)
hv_max_vp_index = msr_vp_index;
@@ -182,7 +183,6 @@ void set_hv_tscchange_cb(void (*cb)(void))
struct hv_reenlightenment_control re_ctrl = {
.enabled = 1,
- .target_vp = hv_vp_index[smp_processor_id()]
struct hv_tsc_emulation_control emu_ctrl = {.enabled = 1};

@@ -196,7 +196,16 @@ void set_hv_tscchange_cb(void (*cb)(void))
/* Make sure callback is registered before we write to MSRs */

+ /*
+ * As reenlightenment vector is global, there is no difference which
+ * CPU will register MSR, though it should be an online CPU.
+ * hv_cpu_die() callback guarantees that on CPU teardown
+ * another CPU will re-register MSR back.
+ */
+ cpus_read_lock();
+ re_ctrl.target_vp = hv_vp_index[raw_smp_processor_id()];
wrmsrl(HV_X64_MSR_REENLIGHTENMENT_CONTROL, *((u64 *)&re_ctrl));
+ cpus_read_unlock();
wrmsrl(HV_X64_MSR_TSC_EMULATION_CONTROL, *((u64 *)&emu_ctrl));
@@ -239,6 +248,7 @@ static int hv_cpu_die(unsigned int cpu)

rdmsrl(HV_X64_MSR_REENLIGHTENMENT_CONTROL, *((u64 *)&re_ctrl));
if (re_ctrl.target_vp == hv_vp_index[cpu]) {
+ lockdep_assert_cpus_held();
/* Reassign to some other online CPU */
new_cpu = cpumask_any_but(cpu_online_mask, cpu);