From: james.morse@arm.com (James Morse)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH v4 15/13] firmware: arm_sdei: be more robust against cpu-hotplug
Date: Wed, 8 Nov 2017 16:06:24 +0000 [thread overview]
Message-ID: <20171108160624.10355-1-james.morse@arm.com> (raw)
In-Reply-To: <5A031E9E.2090809@arm.com>
dpm_suspend() calls the freeze/thaw callbacks for hibernate before
disable_non_bootcpus() takes down secondaries.
This leads to a fun race where the freeze/thaw callbacks reset the
SDEI interface (as we may be restoring a kernel with a different
layout due to KASLR), then the cpu-hotplug callbacks come in to
save the current state, which has already been reset.
I tried to solve this with a 'frozen' flag that stops the hotplug
callback from overwriting the saved values. Instead this just
moves the race around and makes it even harder to think about.
Instead, make it look like the secondaries have gone offline.
Call cpuhp_remove_state() in the freeze callback, this will call the
teardown hook on all online CPUs, then remove the state. This saves
all private events and makes future CPU up/down events invisible.
Change sdei_event_unregister_all()/sdei_reregister_events() to
only save/restore shared events, which are all that is left. With
this we can remove the frozen flag. We can remove the device
suspend/resume calls too as cpuhotplug's teardown call has masked
the CPUs.
All that is left is the reboot notifier, (which was abusing the
frozen flag). Call cpuhp_remove_state() to make it look like
secondary CPUs have gone offline.
Suggested-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: James Morse <james.morse@arm.com>
---
drivers/firmware/arm_sdei.c | 60 +++++++++++++++++++++++----------------------
1 file changed, 31 insertions(+), 29 deletions(-)
diff --git a/drivers/firmware/arm_sdei.c b/drivers/firmware/arm_sdei.c
index 65a8f122f545..d50634a25954 100644
--- a/drivers/firmware/arm_sdei.c
+++ b/drivers/firmware/arm_sdei.c
@@ -86,9 +86,6 @@ struct sdei_event {
static LIST_HEAD(sdei_events);
static DEFINE_SPINLOCK(sdei_events_lock);
-/* When frozen, cpu-hotplug notifiers shouldn't unregister/re-register events */
-static bool frozen;
-
static DEFINE_PER_CPU(u64, sdei_running_event) = -1;
/* Private events are registered/enabled via IPI passing one of these */
@@ -625,15 +622,18 @@ EXPORT_SYMBOL(sdei_event_unregister);
/*
* unregister events, but don't destroy them as they are re-registered by
- * sdei_reregister_events().
+ * sdei_reregister_shared().
*/
-static int sdei_event_unregister_all(void)
+static int sdei_unregister_shared(void)
{
int err = 0;
struct sdei_event *event;
spin_lock(&sdei_events_lock);
list_for_each_entry(event, &sdei_events, list) {
+ if (event->type != SDEI_EVENT_TYPE_SHARED)
+ continue;
+
err = _sdei_event_unregister(event);
if (err)
break;
@@ -841,13 +841,16 @@ static int sdei_reregister_event(struct sdei_event *event)
return err;
}
-static int sdei_reregister_events(void)
+static int sdei_reregister_shared(void)
{
int err = 0;
struct sdei_event *event;
spin_lock(&sdei_events_lock);
list_for_each_entry(event, &sdei_events, list) {
+ if (event->type != SDEI_EVENT_TYPE_SHARED)
+ continue;
+
err = sdei_reregister_event(event);
if (err)
break;
@@ -862,11 +865,6 @@ static int sdei_cpuhp_down(unsigned int cpu)
struct sdei_event *event;
struct sdei_crosscall_args arg;
- if (frozen) {
- /* All events unregistered */
- return sdei_mask_local_cpu();
- }
-
/* un-register private events */
spin_lock(&sdei_events_lock);
list_for_each_entry(event, &sdei_events, list) {
@@ -890,11 +888,6 @@ static int sdei_cpuhp_up(unsigned int cpu)
struct sdei_event *event;
struct sdei_crosscall_args arg;
- if (frozen) {
- /* Events will be re-registered when we thaw. */
- return sdei_unmask_local_cpu();
- }
-
/* re-register/enable private events */
spin_lock(&sdei_events_lock);
list_for_each_entry(event, &sdei_events, list) {
@@ -1004,22 +997,33 @@ static int sdei_device_freeze(struct device *dev)
{
int err;
- frozen = true;
- err = sdei_event_unregister_all();
+ /* save and unregister private events */
+ cpuhp_remove_state(CPUHP_AP_ARM_SDEI_STARTING);
+
+ err = sdei_unregister_shared();
if (err)
return err;
- return sdei_device_suspend(dev);
+ return 0;
}
static int sdei_device_thaw(struct device *dev)
{
int err;
- sdei_device_resume(dev);
+ /* re-register shared events */
+ err = sdei_reregister_shared();
+ if (err) {
+ pr_warn("Failed to re-register shared events...\n");
+ sdei_mark_interface_broken();
+ return err;
+ }
+
+ err = cpuhp_setup_state(CPUHP_AP_ARM_SDEI_STARTING, "SDEI",
+ &sdei_cpuhp_up, &sdei_cpuhp_down);
+ if (err)
+ pr_warn("Failed to re-register CPU hotplug notifier...\n");
- err = sdei_reregister_events();
- frozen = false;
return err;
}
@@ -1048,15 +1052,13 @@ static const struct dev_pm_ops sdei_pm_ops = {
static int sdei_reboot_notifier(struct notifier_block *nb, unsigned long action,
void *data)
{
- on_each_cpu(&_ipi_mask_cpu, NULL, true);
-
- sdei_platform_reset();
-
/*
- * There is now no point trying to unregister private events if we go on
- * to take CPUs offline.
+ * We are going to reset the interface, after this there is no point
+ * doing work when we take CPUs offline.
*/
- frozen = true;
+ cpuhp_remove_state(CPUHP_AP_ARM_SDEI_STARTING);
+
+ sdei_platform_reset();
return NOTIFY_OK;
}
--
2.15.0
next prev parent reply other threads:[~2017-11-08 16:06 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-17 17:44 [PATCH v4 00/13] arm64/firmware: Software Delegated Exception Interface James Morse
2017-10-17 17:44 ` [PATCH v4 01/13] KVM: arm64: Store vcpu on the stack during __guest_enter() James Morse
2017-10-17 17:44 ` [PATCH v4 02/13] KVM: arm/arm64: Convert kvm_host_cpu_state to a static per-cpu allocation James Morse
2017-10-17 17:44 ` [PATCH v4 03/13] KVM: arm64: Change hyp_panic()s dependency on tpidr_el2 James Morse
2017-10-17 17:44 ` [PATCH v4 04/13] arm64: alternatives: use tpidr_el2 on VHE hosts James Morse
2017-10-18 11:00 ` Catalin Marinas
2017-10-17 17:44 ` [PATCH v4 05/13] KVM: arm64: Stop save/restoring host tpidr_el1 on VHE James Morse
2017-10-17 17:44 ` [PATCH v4 06/13] Docs: dt: add devicetree binding for describing arm64 SDEI firmware James Morse
2017-10-17 17:44 ` [PATCH v4 07/13] firmware: arm_sdei: Add driver for Software Delegated Exceptions James Morse
2017-10-18 11:09 ` Catalin Marinas
2017-10-17 17:44 ` [PATCH v4 08/13] arm64: Add vmap_stack header file James Morse
2017-10-18 11:10 ` Catalin Marinas
2017-10-17 17:44 ` [PATCH v4 09/13] arm64: kernel: Add arch-specific SDEI entry code and CPU masking James Morse
2017-10-18 11:18 ` Catalin Marinas
2017-10-17 17:44 ` [PATCH v4 10/13] firmware: arm_sdei: Add support for CPU and system power states James Morse
2017-10-18 11:11 ` Catalin Marinas
2017-10-18 17:17 ` Will Deacon
2017-10-24 17:34 ` James Morse
2017-10-25 14:43 ` Will Deacon
2017-10-17 17:44 ` [PATCH v4 11/13] firmware: arm_sdei: add support for CPU private events James Morse
2017-10-18 11:11 ` Catalin Marinas
2017-10-18 17:19 ` Will Deacon
2017-10-24 17:34 ` James Morse
2017-11-01 15:59 ` James Morse
2017-10-17 17:44 ` [PATCH v4 12/13] arm64: acpi: Remove __init from acpi_psci_use_hvc() for use by SDEI James Morse
2017-10-18 11:11 ` Catalin Marinas
2017-10-17 17:44 ` [PATCH v4 13/13] firmware: arm_sdei: Discover SDEI support via ACPI James Morse
2017-10-18 11:12 ` Catalin Marinas
2017-10-18 17:16 ` [PATCH v4 00/13] arm64/firmware: Software Delegated Exception Interface Will Deacon
2017-10-30 15:58 ` [PATCH v4 14/13] firmware: arm_sdei: Move cpuhotplug registration later James Morse
2017-11-01 15:59 ` [PATCH 15/13] firmware: arm_sdei: move the frozen flag under the spinlock James Morse
2017-11-08 15:11 ` James Morse
2017-11-08 16:06 ` James Morse [this message]
2017-11-13 11:01 ` [PATCH v4 15/13] firmware: arm_sdei: be more robust against cpu-hotplug Will Deacon
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=20171108160624.10355-1-james.morse@arm.com \
--to=james.morse@arm.com \
--cc=linux-arm-kernel@lists.infradead.org \
/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).