This is a note to let you know that I've just added the patch titled
futex: Fix OWNER_DEAD fixup
to the 4.14-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum…
The filename of the patch is:
futex-fix-owner_dead-fixup.patch
and it can be found in the queue-4.14 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable(a)vger.kernel.org> know about it.
>From a97cb0e7b3f4c6297fd857055ae8e895f402f501 Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz(a)infradead.org>
Date: Mon, 22 Jan 2018 11:39:47 +0100
Subject: futex: Fix OWNER_DEAD fixup
From: Peter Zijlstra <peterz(a)infradead.org>
commit a97cb0e7b3f4c6297fd857055ae8e895f402f501 upstream.
Both Geert and DaveJ reported that the recent futex commit:
c1e2f0eaf015 ("futex: Avoid violating the 10th rule of futex")
introduced a problem with setting OWNER_DEAD. We set the bit on an
uninitialized variable and then entirely optimize it away as a
dead-store.
Move the setting of the bit to where it is more useful.
Reported-by: Geert Uytterhoeven <geert(a)linux-m68k.org>
Reported-by: Dave Jones <davej(a)codemonkey.org.uk>
Signed-off-by: Peter Zijlstra (Intel) <peterz(a)infradead.org>
Cc: Andrew Morton <akpm(a)linux-foundation.org>
Cc: Linus Torvalds <torvalds(a)linux-foundation.org>
Cc: Paul E. McKenney <paulmck(a)us.ibm.com>
Cc: Peter Zijlstra <peterz(a)infradead.org>
Cc: Thomas Gleixner <tglx(a)linutronix.de>
Fixes: c1e2f0eaf015 ("futex: Avoid violating the 10th rule of futex")
Link: http://lkml.kernel.org/r/20180122103947.GD2228@hirez.programming.kicks-ass.…
Signed-off-by: Ingo Molnar <mingo(a)kernel.org>
Cc: Ozkan Sezer <sezeroz(a)gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org>
---
kernel/futex.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -2311,9 +2311,6 @@ static int fixup_pi_state_owner(u32 __us
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock);
oldowner = pi_state->owner;
- /* Owner died? */
- if (!pi_state->owner)
- newtid |= FUTEX_OWNER_DIED;
/*
* We are here because either:
@@ -2374,6 +2371,9 @@ retry:
}
newtid = task_pid_vnr(newowner) | FUTEX_WAITERS;
+ /* Owner died? */
+ if (!pi_state->owner)
+ newtid |= FUTEX_OWNER_DIED;
if (get_futex_value_locked(&uval, uaddr))
goto handle_fault;
Patches currently in stable-queue which might be from peterz(a)infradead.org are
queue-4.14/futex-fix-owner_dead-fixup.patch
When introducing support for irqchip in userspace we needed a way to
mask the timer signal to prevent the guest continuously exiting due to a
screaming timer.
We did this by disabling the corresponding percpu interrupt on the
host interrupt controller, because we cannot rely on the host system
having a GIC, and therefore cannot make any assumptions about having an
active state to hide the timer signal.
Unfortunately, when introducing this feature, it became entirely
possible that a VCPU which belongs to a VM that has a userspace irqchip
can disable the vtimer irq on the host on some physical CPU, and then go
away without ever enabling the vtimer irq on that physical CPU again.
This means that using irqchips in userspace on a system that also
supports running VMs with an in-kernel GIC can prevent forward progress
from in-kernel GIC VMs.
Later on, when we started taking virtual timer interrupts in the arch
timer code, we would also leave this timer state active for userspace
irqchip VMs, because we leave it up to a VGIC-enabled guest to
deactivate the hardware IRQ using the HW bit in the LR.
Both issues are solved by only using the enable/disable trick on systems
that do not have a host GIC which supports the active state, because all
VMs on such systems must use irqchips in userspace. Systems that have a
working GIC with support for an active state use the active state to
mask the timer signal for both userspace and in-kernel irqchips.
Cc: Alexander Graf <agraf(a)suse.de>
Cc: <stable(a)vger.kernel.org> # v4.12+
Fixes: d9e139778376 ("KVM: arm/arm64: Support arch timers with a userspace gic")
Signed-off-by: Christoffer Dall <christoffer.dall(a)linaro.org>
---
This conflicts horribly with everything when applied to either
kvmarm/queue or kvmarm/master. Therefore, this patch is written for
(and applies to) v4.15 with kvmarm/queue merged and should therefore
apply cleanly after v4.16-rc1. An example with this patch applied can
be found on kvmarm/temp-for-v4.16-rc2. I plan on sending this along
with any other potential fixes post v4.16-rc1.
Changes since v1:
- Added userspace_irqchip() wrapper to simplify logic
- Changed has_gic_active_state to a static key
- Fixed typos in commentary and commit message
- Clear the active state on sync for userspace irqchips on systems
with a working GIC.
- Get rid of __timer_snapshot_state() in unmaks_vtimer_irq_user()
because kvm_timer_should_fire() has already been reworked in other
patches to look at the timer state when it's loaded onto the
hardware. As a result, we don't need the
__timer_snapshot_state() indirection anymore and this logic is
now inlined in vtimer_save_state().
virt/kvm/arm/arch_timer.c | 116 +++++++++++++++++++++++++---------------------
1 file changed, 64 insertions(+), 52 deletions(-)
diff --git a/virt/kvm/arm/arch_timer.c b/virt/kvm/arm/arch_timer.c
index 70268c0bec79..70f4c30918eb 100644
--- a/virt/kvm/arm/arch_timer.c
+++ b/virt/kvm/arm/arch_timer.c
@@ -36,6 +36,8 @@ static struct timecounter *timecounter;
static unsigned int host_vtimer_irq;
static u32 host_vtimer_irq_flags;
+static DEFINE_STATIC_KEY_FALSE(has_gic_active_state);
+
static const struct kvm_irq_level default_ptimer_irq = {
.irq = 30,
.level = 1,
@@ -56,6 +58,12 @@ u64 kvm_phys_timer_read(void)
return timecounter->cc->read(timecounter->cc);
}
+static inline bool userspace_irqchip(struct kvm *kvm)
+{
+ return static_branch_unlikely(&userspace_irqchip_in_use) &&
+ unlikely(!irqchip_in_kernel(kvm));
+}
+
static void soft_timer_start(struct hrtimer *hrt, u64 ns)
{
hrtimer_start(hrt, ktime_add_ns(ktime_get(), ns),
@@ -69,25 +77,6 @@ static void soft_timer_cancel(struct hrtimer *hrt, struct work_struct *work)
cancel_work_sync(work);
}
-static void kvm_vtimer_update_mask_user(struct kvm_vcpu *vcpu)
-{
- struct arch_timer_context *vtimer = vcpu_vtimer(vcpu);
-
- /*
- * When using a userspace irqchip with the architected timers, we must
- * prevent continuously exiting from the guest, and therefore mask the
- * physical interrupt by disabling it on the host interrupt controller
- * when the virtual level is high, such that the guest can make
- * forward progress. Once we detect the output level being
- * de-asserted, we unmask the interrupt again so that we exit from the
- * guest when the timer fires.
- */
- if (vtimer->irq.level)
- disable_percpu_irq(host_vtimer_irq);
- else
- enable_percpu_irq(host_vtimer_irq, 0);
-}
-
static irqreturn_t kvm_arch_timer_handler(int irq, void *dev_id)
{
struct kvm_vcpu *vcpu = *(struct kvm_vcpu **)dev_id;
@@ -106,9 +95,9 @@ static irqreturn_t kvm_arch_timer_handler(int irq, void *dev_id)
if (kvm_timer_should_fire(vtimer))
kvm_timer_update_irq(vcpu, true, vtimer);
- if (static_branch_unlikely(&userspace_irqchip_in_use) &&
- unlikely(!irqchip_in_kernel(vcpu->kvm)))
- kvm_vtimer_update_mask_user(vcpu);
+ if (userspace_irqchip(vcpu->kvm) &&
+ !static_branch_unlikely(&has_gic_active_state))
+ disable_percpu_irq(host_vtimer_irq);
return IRQ_HANDLED;
}
@@ -290,8 +279,7 @@ static void kvm_timer_update_irq(struct kvm_vcpu *vcpu, bool new_level,
trace_kvm_timer_update_irq(vcpu->vcpu_id, timer_ctx->irq.irq,
timer_ctx->irq.level);
- if (!static_branch_unlikely(&userspace_irqchip_in_use) ||
- likely(irqchip_in_kernel(vcpu->kvm))) {
+ if (!userspace_irqchip(vcpu->kvm)) {
ret = kvm_vgic_inject_irq(vcpu->kvm, vcpu->vcpu_id,
timer_ctx->irq.irq,
timer_ctx->irq.level,
@@ -350,12 +338,6 @@ static void kvm_timer_update_state(struct kvm_vcpu *vcpu)
phys_timer_emulate(vcpu);
}
-static void __timer_snapshot_state(struct arch_timer_context *timer)
-{
- timer->cnt_ctl = read_sysreg_el0(cntv_ctl);
- timer->cnt_cval = read_sysreg_el0(cntv_cval);
-}
-
static void vtimer_save_state(struct kvm_vcpu *vcpu)
{
struct arch_timer_cpu *timer = &vcpu->arch.timer_cpu;
@@ -367,8 +349,10 @@ static void vtimer_save_state(struct kvm_vcpu *vcpu)
if (!vtimer->loaded)
goto out;
- if (timer->enabled)
- __timer_snapshot_state(vtimer);
+ if (timer->enabled) {
+ vtimer->cnt_ctl = read_sysreg_el0(cntv_ctl);
+ vtimer->cnt_cval = read_sysreg_el0(cntv_cval);
+ }
/* Disable the virtual timer */
write_sysreg_el0(0, cntv_ctl);
@@ -460,23 +444,43 @@ static void set_cntvoff(u64 cntvoff)
kvm_call_hyp(__kvm_timer_set_cntvoff, low, high);
}
-static void kvm_timer_vcpu_load_vgic(struct kvm_vcpu *vcpu)
+static inline void set_vtimer_irq_phys_active(struct kvm_vcpu *vcpu, bool active)
+{
+ int r;
+ r = irq_set_irqchip_state(host_vtimer_irq, IRQCHIP_STATE_ACTIVE, active);
+ WARN_ON(r);
+}
+
+static void kvm_timer_vcpu_load_gic(struct kvm_vcpu *vcpu)
{
struct arch_timer_context *vtimer = vcpu_vtimer(vcpu);
bool phys_active;
- int ret;
- phys_active = kvm_vgic_map_is_active(vcpu, vtimer->irq.irq);
-
- ret = irq_set_irqchip_state(host_vtimer_irq,
- IRQCHIP_STATE_ACTIVE,
- phys_active);
- WARN_ON(ret);
+ if (irqchip_in_kernel(vcpu->kvm))
+ phys_active = kvm_vgic_map_is_active(vcpu, vtimer->irq.irq);
+ else
+ phys_active = vtimer->irq.level;
+ set_vtimer_irq_phys_active(vcpu, phys_active);
}
-static void kvm_timer_vcpu_load_user(struct kvm_vcpu *vcpu)
+static void kvm_timer_vcpu_load_nogic(struct kvm_vcpu *vcpu)
{
- kvm_vtimer_update_mask_user(vcpu);
+ struct arch_timer_context *vtimer = vcpu_vtimer(vcpu);
+
+ /*
+ * When using a userspace irqchip with the architected timers and a
+ * host interrupt controller that doesn't support an active state, we
+ * must still prevent continuously exiting from the guest, and
+ * therefore mask the physical interrupt by disabling it on the host
+ * interrupt controller when the virtual level is high, such that the
+ * guest can make forward progress. Once we detect the output level
+ * being de-asserted, we unmask the interrupt again so that we exit
+ * from the guest when the timer fires.
+ */
+ if (vtimer->irq.level)
+ disable_percpu_irq(host_vtimer_irq);
+ else
+ enable_percpu_irq(host_vtimer_irq, host_vtimer_irq_flags);
}
void kvm_timer_vcpu_load(struct kvm_vcpu *vcpu)
@@ -487,10 +491,10 @@ void kvm_timer_vcpu_load(struct kvm_vcpu *vcpu)
if (unlikely(!timer->enabled))
return;
- if (unlikely(!irqchip_in_kernel(vcpu->kvm)))
- kvm_timer_vcpu_load_user(vcpu);
+ if (static_branch_likely(&has_gic_active_state))
+ kvm_timer_vcpu_load_gic(vcpu);
else
- kvm_timer_vcpu_load_vgic(vcpu);
+ kvm_timer_vcpu_load_nogic(vcpu);
set_cntvoff(vtimer->cntvoff);
@@ -555,18 +559,24 @@ static void unmask_vtimer_irq_user(struct kvm_vcpu *vcpu)
{
struct arch_timer_context *vtimer = vcpu_vtimer(vcpu);
- if (unlikely(!irqchip_in_kernel(vcpu->kvm))) {
- __timer_snapshot_state(vtimer);
- if (!kvm_timer_should_fire(vtimer)) {
- kvm_timer_update_irq(vcpu, false, vtimer);
- kvm_vtimer_update_mask_user(vcpu);
- }
+ if (!kvm_timer_should_fire(vtimer)) {
+ kvm_timer_update_irq(vcpu, false, vtimer);
+ if (static_branch_likely(&has_gic_active_state))
+ set_vtimer_irq_phys_active(vcpu, false);
+ else
+ enable_percpu_irq(host_vtimer_irq, host_vtimer_irq_flags);
}
}
void kvm_timer_sync_hwstate(struct kvm_vcpu *vcpu)
{
- unmask_vtimer_irq_user(vcpu);
+ struct arch_timer_cpu *timer = &vcpu->arch.timer_cpu;
+
+ if (unlikely(!timer->enabled))
+ return;
+
+ if (unlikely(!irqchip_in_kernel(vcpu->kvm)))
+ unmask_vtimer_irq_user(vcpu);
}
int kvm_timer_vcpu_reset(struct kvm_vcpu *vcpu)
@@ -753,6 +763,8 @@ int kvm_timer_hyp_init(bool has_gic)
kvm_err("kvm_arch_timer: error setting vcpu affinity\n");
goto out_free_irq;
}
+
+ static_branch_enable(&has_gic_active_state);
}
kvm_info("virtual timer IRQ%d\n", host_vtimer_irq);
--
2.14.2
This is a note to let you know that I've just added the patch titled
um: Stop abusing __KERNEL__
to the 3.18-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum…
The filename of the patch is:
um-stop-abusing-__kernel__.patch
and it can be found in the queue-3.18 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable(a)vger.kernel.org> know about it.
>From 298e20ba8c197e8d429a6c8671550c41c7919033 Mon Sep 17 00:00:00 2001
From: Richard Weinberger <richard(a)nod.at>
Date: Sun, 31 May 2015 19:50:57 +0200
Subject: um: Stop abusing __KERNEL__
From: Richard Weinberger <richard(a)nod.at>
commit 298e20ba8c197e8d429a6c8671550c41c7919033 upstream.
Currently UML is abusing __KERNEL__ to distinguish between
kernel and host code (os-Linux). It is better to use a custom
define such that existing users of __KERNEL__ don't get confused.
Signed-off-by: Richard Weinberger <richard(a)nod.at>
Cc: Greg Hackmann <ghackmann(a)google.com>
Cc: Bernie Innocenti <codewiz(a)google.com>
Cc: Lorenzo Colitti <lorenzo(a)google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org>
---
arch/um/Makefile | 7 ++++---
arch/um/drivers/mconsole.h | 2 +-
arch/um/include/shared/init.h | 4 ++--
arch/um/include/shared/user.h | 2 +-
arch/x86/um/shared/sysdep/tls.h | 6 +++---
5 files changed, 11 insertions(+), 10 deletions(-)
--- a/arch/um/Makefile
+++ b/arch/um/Makefile
@@ -68,9 +68,10 @@ KBUILD_CFLAGS += $(CFLAGS) $(CFLAGS-y) -
KBUILD_AFLAGS += $(ARCH_INCLUDE)
-USER_CFLAGS = $(patsubst $(KERNEL_DEFINES),,$(patsubst -D__KERNEL__,,\
- $(patsubst -I%,,$(KBUILD_CFLAGS)))) $(ARCH_INCLUDE) $(MODE_INCLUDE) \
- $(filter -I%,$(CFLAGS)) -D_FILE_OFFSET_BITS=64 -idirafter include
+USER_CFLAGS = $(patsubst $(KERNEL_DEFINES),,$(patsubst -I%,,$(KBUILD_CFLAGS))) \
+ $(ARCH_INCLUDE) $(MODE_INCLUDE) $(filter -I%,$(CFLAGS)) \
+ -D_FILE_OFFSET_BITS=64 -idirafter include \
+ -D__KERNEL__ -D__UM_HOST__
#This will adjust *FLAGS accordingly to the platform.
include $(srctree)/$(ARCH_DIR)/Makefile-os-$(OS)
--- a/arch/um/drivers/mconsole.h
+++ b/arch/um/drivers/mconsole.h
@@ -7,7 +7,7 @@
#ifndef __MCONSOLE_H__
#define __MCONSOLE_H__
-#ifndef __KERNEL__
+#ifdef __UM_HOST__
#include <stdint.h>
#define u32 uint32_t
#endif
--- a/arch/um/include/shared/init.h
+++ b/arch/um/include/shared/init.h
@@ -40,7 +40,7 @@
typedef int (*initcall_t)(void);
typedef void (*exitcall_t)(void);
-#ifndef __KERNEL__
+#ifdef __UM_HOST__
#ifndef __section
# define __section(S) __attribute__ ((__section__(#S)))
#endif
@@ -131,7 +131,7 @@ extern struct uml_param __uml_setup_star
#define __uml_postsetup_call __used __section(.uml.postsetup.init)
#define __uml_exit_call __used __section(.uml.exitcall.exit)
-#ifndef __KERNEL__
+#ifdef __UM_HOST__
#define __define_initcall(level,fn) \
static initcall_t __initcall_##fn __used \
--- a/arch/um/include/shared/user.h
+++ b/arch/um/include/shared/user.h
@@ -17,7 +17,7 @@
#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0]))
/* This is to get size_t */
-#ifdef __KERNEL__
+#ifndef __UM_HOST__
#include <linux/types.h>
#else
#include <stddef.h>
--- a/arch/x86/um/shared/sysdep/tls.h
+++ b/arch/x86/um/shared/sysdep/tls.h
@@ -1,7 +1,7 @@
#ifndef _SYSDEP_TLS_H
#define _SYSDEP_TLS_H
-# ifndef __KERNEL__
+#ifdef __UM_HOST__
/* Change name to avoid conflicts with the original one from <asm/ldt.h>, which
* may be named user_desc (but in 2.4 and in header matching its API was named
@@ -22,11 +22,11 @@ typedef struct um_dup_user_desc {
#endif
} user_desc_t;
-# else /* __KERNEL__ */
+#else /* __UM_HOST__ */
typedef struct user_desc user_desc_t;
-# endif /* __KERNEL__ */
+#endif /* __UM_HOST__ */
extern int os_set_thread_area(user_desc_t *info, int pid);
extern int os_get_thread_area(user_desc_t *info, int pid);
Patches currently in stable-queue which might be from richard(a)nod.at are
queue-3.18/um-stop-abusing-__kernel__.patch
queue-3.18/um-link-vmlinux-with-no-pie.patch
queue-3.18/um-remove-copy-paste-code-from-init.h.patch
This is a note to let you know that I've just added the patch titled
um: Remove copy&paste code from init.h
to the 3.18-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum…
The filename of the patch is:
um-remove-copy-paste-code-from-init.h.patch
and it can be found in the queue-3.18 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable(a)vger.kernel.org> know about it.
>From 30b11ee9ae23d78de66b9ae315880af17a64ba83 Mon Sep 17 00:00:00 2001
From: Richard Weinberger <richard(a)nod.at>
Date: Sun, 31 May 2015 22:15:58 +0200
Subject: um: Remove copy&paste code from init.h
From: Richard Weinberger <richard(a)nod.at>
commit 30b11ee9ae23d78de66b9ae315880af17a64ba83 upstream.
As we got rid of the __KERNEL__ abuse, we can directly
include linux/compiler.h now.
This also allows gcc 5 to build UML.
Reported-by: Hans-Werner Hilse <hwhilse(a)gmail.com>
Signed-off-by: Richard Weinberger <richard(a)nod.at>
Cc: Greg Hackmann <ghackmann(a)google.com>
Cc: Bernie Innocenti <codewiz(a)google.com>
Cc: Lorenzo Colitti <lorenzo(a)google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org>
---
arch/um/include/shared/init.h | 22 +---------------------
1 file changed, 1 insertion(+), 21 deletions(-)
--- a/arch/um/include/shared/init.h
+++ b/arch/um/include/shared/init.h
@@ -40,28 +40,8 @@
typedef int (*initcall_t)(void);
typedef void (*exitcall_t)(void);
-#ifdef __UM_HOST__
-#ifndef __section
-# define __section(S) __attribute__ ((__section__(#S)))
-#endif
-
-#if __GNUC__ == 3
-
-#if __GNUC_MINOR__ >= 3
-# define __used __attribute__((__used__))
-#else
-# define __used __attribute__((__unused__))
-#endif
-
-#else
-#if __GNUC__ == 4
-# define __used __attribute__((__used__))
-#endif
-#endif
-
-#else
#include <linux/compiler.h>
-#endif
+
/* These are for everybody (although not all archs will actually
discard it in modules) */
#define __init __section(.init.text)
Patches currently in stable-queue which might be from richard(a)nod.at are
queue-3.18/um-stop-abusing-__kernel__.patch
queue-3.18/um-link-vmlinux-with-no-pie.patch
queue-3.18/um-remove-copy-paste-code-from-init.h.patch
This is a note to let you know that I've just added the patch titled
um: link vmlinux with -no-pie
to the 3.18-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum…
The filename of the patch is:
um-link-vmlinux-with-no-pie.patch
and it can be found in the queue-3.18 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable(a)vger.kernel.org> know about it.
>From 883354afbc109c57f925ccc19840055193da0cc0 Mon Sep 17 00:00:00 2001
From: Thomas Meyer <thomas(a)m3y3r.de>
Date: Sun, 20 Aug 2017 13:26:04 +0200
Subject: um: link vmlinux with -no-pie
From: Thomas Meyer <thomas(a)m3y3r.de>
commit 883354afbc109c57f925ccc19840055193da0cc0 upstream.
Debian's gcc defaults to pie. The global Makefile already defines the -fno-pie option.
Link UML dynamic kernel image also with -no-pie to fix the build.
Signed-off-by: Thomas Meyer <thomas(a)m3y3r.de>
Signed-off-by: Richard Weinberger <richard(a)nod.at>
Cc: Bernie Innocenti <codewiz(a)google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org>
---
arch/um/Makefile | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/arch/um/Makefile
+++ b/arch/um/Makefile
@@ -116,7 +116,7 @@ archheaders:
archprepare: include/generated/user_constants.h
LINK-$(CONFIG_LD_SCRIPT_STATIC) += -static
-LINK-$(CONFIG_LD_SCRIPT_DYN) += -Wl,-rpath,/lib
+LINK-$(CONFIG_LD_SCRIPT_DYN) += -Wl,-rpath,/lib $(call cc-option, -no-pie)
CFLAGS_NO_HARDENING := $(call cc-option, -fno-PIC,) $(call cc-option, -fno-pic,) \
$(call cc-option, -fno-stack-protector,) \
Patches currently in stable-queue which might be from thomas(a)m3y3r.de are
queue-3.18/um-link-vmlinux-with-no-pie.patch
This is a note to let you know that I've just added the patch titled
Input: do not emit unneeded EV_SYN when suspending
to the 3.18-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum…
The filename of the patch is:
input-do-not-emit-unneeded-ev_syn-when-suspending.patch
and it can be found in the queue-3.18 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable(a)vger.kernel.org> know about it.
>From 00159f19a5057cb779146afce1cceede692af346 Mon Sep 17 00:00:00 2001
From: Dmitry Torokhov <dmitry.torokhov(a)gmail.com>
Date: Thu, 6 Aug 2015 19:15:30 -0700
Subject: Input: do not emit unneeded EV_SYN when suspending
From: Dmitry Torokhov <dmitry.torokhov(a)gmail.com>
commit 00159f19a5057cb779146afce1cceede692af346 upstream.
Do not emit EV_SYN/SYN_REPORT on suspend if there were no keys that are
still pressed as we are suspending the device (and in all other cases when
input core is forcibly releasing keys via input_dev_release_keys() call).
Reviewed-by: Benson Leung <bleung(a)chromium.org>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov(a)gmail.com>
Signed-off-by: Bo Hu <bohu(a)google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org>
diff --git a/drivers/input/input.c b/drivers/input/input.c
index 78d24990a816..5391abd28b27 100644
--- a/drivers/input/input.c
+++ b/drivers/input/input.c
@@ -674,13 +674,19 @@ EXPORT_SYMBOL(input_close_device);
*/
static void input_dev_release_keys(struct input_dev *dev)
{
+ bool need_sync = false;
int code;
if (is_event_supported(EV_KEY, dev->evbit, EV_MAX)) {
- for_each_set_bit(code, dev->key, KEY_CNT)
+ for_each_set_bit(code, dev->key, KEY_CNT) {
input_pass_event(dev, EV_KEY, code, 0);
+ need_sync = true;
+ }
+
+ if (need_sync)
+ input_pass_event(dev, EV_SYN, SYN_REPORT, 1);
+
memset(dev->key, 0, sizeof(dev->key));
- input_pass_event(dev, EV_SYN, SYN_REPORT, 1);
}
}
Patches currently in stable-queue which might be from dmitry.torokhov(a)gmail.com are
queue-3.18/input-do-not-emit-unneeded-ev_syn-when-suspending.patch
Hi Marcin,
Since it's been a week, could you confirm the patch is ok as-is or do
you think some comment(s) from James should be incorporated ?
On Tue, Jan 23, 2018 at 3:17 PM, James Hogan <jhogan(a)kernel.org> wrote:
> On Thu, Dec 21, 2017 at 10:00:59PM +0100, Mathieu Malaterre wrote:
>> From: Marcin Nowakowski <marcin.nowakowski(a)mips.com>
>>
>> Change 73fbc1eba7ff added a fix to ensure that the memory range between
>
> Please refer to commits with e.g. commit 73fbc1eba7ff ("MIPS: fix
> mem=X@Y commandline processing").
>
>> PHYS_OFFSET and low memory address specified by mem= cmdline argument is
>> not later processed by free_all_bootmem.
>> This change was incorrect for systems where the commandline specifies
>> more than 1 mem argument, as it will cause all memory between
>> PHYS_OFFSET and each of the memory offsets to be marked as reserved,
>> which results in parts of the RAM marked as reserved (Creator CI20's
>> u-boot has a default commandline argument 'mem=256M@0x0
>> mem=768M@0x30000000').
>>
>> Change the behaviour to ensure that only the range between PHYS_OFFSET
>> and the lowest start address of the memories is marked as protected.
>>
>> This change also ensures that the range is marked protected even if it's
>> only defined through the devicetree and not only via commandline
>> arguments.
>>
>> Reported-by: Mathieu Malaterre <mathieu.malaterre(a)gmail.com>
>> Signed-off-by: Marcin Nowakowski <marcin.nowakowski(a)mips.com>
>> Fixes: 73fbc1eba7ff ("MIPS: fix mem=X@Y commandline processing")
>> Cc: <stable(a)vger.kernel.org> # v4.11
>
> I'm guessing that should technically be v4.11+
My fault, if this is the only change, I can re-submit.
>> ---
>> v2: Use updated email adress, add tag for stable.
>> arch/mips/kernel/setup.c | 19 ++++++++++++++++---
>> 1 file changed, 16 insertions(+), 3 deletions(-)
>>
>> diff --git a/arch/mips/kernel/setup.c b/arch/mips/kernel/setup.c
>> index 702c678de116..f19d61224c71 100644
>> --- a/arch/mips/kernel/setup.c
>> +++ b/arch/mips/kernel/setup.c
>> @@ -375,6 +375,7 @@ static void __init bootmem_init(void)
>> unsigned long reserved_end;
>> unsigned long mapstart = ~0UL;
>> unsigned long bootmap_size;
>> + phys_addr_t ramstart = ~0UL;
>
> Although practically it might not matter, technically phys_addr_t may be
> 64-bits (CONFIG_PHYS_ADDR_T_64BIT) even on a 32-bit kernels, in which
> case ~0UL may not be sufficiently large.
>
> Maybe that should be ~(phys_addr_t)0, or perhaps (phys_addr_t)ULLONG_MAX
> to match add_memory_region().
>
>> bool bootmap_valid = false;
>> int i;
>>
>> @@ -395,6 +396,21 @@ static void __init bootmem_init(void)
>> max_low_pfn = 0;
>>
>> /*
>> + * Reserve any memory between the start of RAM and PHYS_OFFSET
>> + */
>> + for (i = 0; i < boot_mem_map.nr_map; i++) {
>> + if (boot_mem_map.map[i].type != BOOT_MEM_RAM)
>> + continue;
>> +
>> + ramstart = min(ramstart, boot_mem_map.map[i].addr);
>
> Is it worth incorporating this into the existing loop below ...
>
>> + }
>> +
>> + if (ramstart > PHYS_OFFSET)
>> + add_memory_region(PHYS_OFFSET, ramstart - PHYS_OFFSET,
>> + BOOT_MEM_RESERVED);
>
> ... and this then placed below that loop?
>
> Otherwise I can't find fault with this patch, though i'm not intimately
> familiar with bootmem.
>
> Cheers
> James
>
>> +
>> +
>> + /*
>> * Find the highest page frame number we have available.
>> */
>> for (i = 0; i < boot_mem_map.nr_map; i++) {
>> @@ -664,9 +680,6 @@ static int __init early_parse_mem(char *p)
>>
>> add_memory_region(start, size, BOOT_MEM_RAM);
>>
>> - if (start && start > PHYS_OFFSET)
>> - add_memory_region(PHYS_OFFSET, start - PHYS_OFFSET,
>> - BOOT_MEM_RESERVED);
>> return 0;
>> }
>> early_param("mem", early_parse_mem);
>> --
>> 2.11.0
>>
When introducing support for irqchip in userspace we needed a way to
mask the timer signal to prevent the guest continuously exiting due to a
screaming timer.
We did this by disabling the corresponding percpu interrupt on the
host interrupt controller, because we cannot rely on the host system
having a GIC, and therefore cannot make any assumptions about having an
active state to hide the timer signal.
Unfortunately, when introducing this feature, it became entirely
possible that a VCPU which belongs to a VM that has a userspace irqchip
can disable the vtimer irq on the host on some physical CPU, and then go
away without ever enabling the vimter irq on that physical CPU again.
This means that using irqchips in userspace on a system that also
supports running VMs with an in-kernel GIC can prevent forward progress
from in-kernel GIC VMs.
Later on, when we started taking virtual timer interrupts in the arch
timer code, we would also leave this timer state active for userspace
irqchip VMs, because we leave it up to a VGIC-enabled guest to
deactivate the hardware IRQ using the HW bit in the LR.
Both issues are solved by only using the enable/disable trick on systems
that do not have a host GIC which supports the active state, because all
VMs on such systems must use irqchips in userspace. Systems that have a
working GIC with support for an active state use the active state to
mask the timer signal for both userspace an in-kernel irqchips.
Cc: Alexander Graf <agraf(a)suse.de>
Cc: <stable(a)vger.kernel.org> # v4.12+
Fixes: d9e139778376 ("KVM: arm/arm64: Support arch timers with a userspace gic")
Signed-off-by: Christoffer Dall <christoffer.dall(a)linaro.org>
---
This conflicts horribly with everything when applied to either
kvmarm/queue or kvmarm/master. Therefore, this patch is written for
(and applies to) v4.15 with kvmarm/queue merged and should therefore
apply cleanly after v4.16-rc1. An example with this patch applied can
be found on kvmarm/temp-for-v4.16-rc2. I plan on sending this along
with any other potential fixes post v4.16-rc1.
virt/kvm/arm/arch_timer.c | 77 ++++++++++++++++++++++++++---------------------
1 file changed, 42 insertions(+), 35 deletions(-)
diff --git a/virt/kvm/arm/arch_timer.c b/virt/kvm/arm/arch_timer.c
index 70268c0bec79..228906ceb722 100644
--- a/virt/kvm/arm/arch_timer.c
+++ b/virt/kvm/arm/arch_timer.c
@@ -35,6 +35,7 @@
static struct timecounter *timecounter;
static unsigned int host_vtimer_irq;
static u32 host_vtimer_irq_flags;
+static bool has_gic_active_state;
static const struct kvm_irq_level default_ptimer_irq = {
.irq = 30,
@@ -69,25 +70,6 @@ static void soft_timer_cancel(struct hrtimer *hrt, struct work_struct *work)
cancel_work_sync(work);
}
-static void kvm_vtimer_update_mask_user(struct kvm_vcpu *vcpu)
-{
- struct arch_timer_context *vtimer = vcpu_vtimer(vcpu);
-
- /*
- * When using a userspace irqchip with the architected timers, we must
- * prevent continuously exiting from the guest, and therefore mask the
- * physical interrupt by disabling it on the host interrupt controller
- * when the virtual level is high, such that the guest can make
- * forward progress. Once we detect the output level being
- * de-asserted, we unmask the interrupt again so that we exit from the
- * guest when the timer fires.
- */
- if (vtimer->irq.level)
- disable_percpu_irq(host_vtimer_irq);
- else
- enable_percpu_irq(host_vtimer_irq, 0);
-}
-
static irqreturn_t kvm_arch_timer_handler(int irq, void *dev_id)
{
struct kvm_vcpu *vcpu = *(struct kvm_vcpu **)dev_id;
@@ -107,8 +89,8 @@ static irqreturn_t kvm_arch_timer_handler(int irq, void *dev_id)
kvm_timer_update_irq(vcpu, true, vtimer);
if (static_branch_unlikely(&userspace_irqchip_in_use) &&
- unlikely(!irqchip_in_kernel(vcpu->kvm)))
- kvm_vtimer_update_mask_user(vcpu);
+ unlikely(!irqchip_in_kernel(vcpu->kvm)) && !has_gic_active_state)
+ disable_percpu_irq(host_vtimer_irq);
return IRQ_HANDLED;
}
@@ -460,13 +442,16 @@ static void set_cntvoff(u64 cntvoff)
kvm_call_hyp(__kvm_timer_set_cntvoff, low, high);
}
-static void kvm_timer_vcpu_load_vgic(struct kvm_vcpu *vcpu)
+static void kvm_timer_vcpu_load_gic(struct kvm_vcpu *vcpu)
{
struct arch_timer_context *vtimer = vcpu_vtimer(vcpu);
bool phys_active;
int ret;
- phys_active = kvm_vgic_map_is_active(vcpu, vtimer->irq.irq);
+ if (irqchip_in_kernel(vcpu->kvm))
+ phys_active = kvm_vgic_map_is_active(vcpu, vtimer->irq.irq);
+ else
+ phys_active = vtimer->irq.level;
ret = irq_set_irqchip_state(host_vtimer_irq,
IRQCHIP_STATE_ACTIVE,
@@ -474,9 +459,24 @@ static void kvm_timer_vcpu_load_vgic(struct kvm_vcpu *vcpu)
WARN_ON(ret);
}
-static void kvm_timer_vcpu_load_user(struct kvm_vcpu *vcpu)
+static void kvm_timer_vcpu_load_nogic(struct kvm_vcpu *vcpu)
{
- kvm_vtimer_update_mask_user(vcpu);
+ struct arch_timer_context *vtimer = vcpu_vtimer(vcpu);
+
+ /*
+ * When using a userspace irqchip with the architected timers and a
+ * host interrupt controller that doesn't support an active state, we
+ * must still we must prevent continuously exiting from the guest, and
+ * therefore mask the physical interrupt by disabling it on the host
+ * interrupt controller when the virtual level is high, such that the
+ * guest can make forward progress. Once we detect the output level
+ * being de-asserted, we unmask the interrupt again so that we exit
+ * from the guest when the timer fires.
+ */
+ if (vtimer->irq.level)
+ disable_percpu_irq(host_vtimer_irq);
+ else
+ enable_percpu_irq(host_vtimer_irq, host_vtimer_irq_flags);
}
void kvm_timer_vcpu_load(struct kvm_vcpu *vcpu)
@@ -487,10 +487,10 @@ void kvm_timer_vcpu_load(struct kvm_vcpu *vcpu)
if (unlikely(!timer->enabled))
return;
- if (unlikely(!irqchip_in_kernel(vcpu->kvm)))
- kvm_timer_vcpu_load_user(vcpu);
+ if (has_gic_active_state)
+ kvm_timer_vcpu_load_gic(vcpu);
else
- kvm_timer_vcpu_load_vgic(vcpu);
+ kvm_timer_vcpu_load_nogic(vcpu);
set_cntvoff(vtimer->cntvoff);
@@ -555,18 +555,23 @@ static void unmask_vtimer_irq_user(struct kvm_vcpu *vcpu)
{
struct arch_timer_context *vtimer = vcpu_vtimer(vcpu);
- if (unlikely(!irqchip_in_kernel(vcpu->kvm))) {
- __timer_snapshot_state(vtimer);
- if (!kvm_timer_should_fire(vtimer)) {
- kvm_timer_update_irq(vcpu, false, vtimer);
- kvm_vtimer_update_mask_user(vcpu);
- }
+ __timer_snapshot_state(vtimer);
+ if (!kvm_timer_should_fire(vtimer)) {
+ kvm_timer_update_irq(vcpu, false, vtimer);
+ if (!has_gic_active_state)
+ enable_percpu_irq(host_vtimer_irq, host_vtimer_irq_flags);
}
}
void kvm_timer_sync_hwstate(struct kvm_vcpu *vcpu)
{
- unmask_vtimer_irq_user(vcpu);
+ struct arch_timer_cpu *timer = &vcpu->arch.timer_cpu;
+
+ if (unlikely(!timer->enabled))
+ return;
+
+ if (unlikely(!irqchip_in_kernel(vcpu->kvm)))
+ unmask_vtimer_irq_user(vcpu);
}
int kvm_timer_vcpu_reset(struct kvm_vcpu *vcpu)
@@ -753,6 +758,8 @@ int kvm_timer_hyp_init(bool has_gic)
kvm_err("kvm_arch_timer: error setting vcpu affinity\n");
goto out_free_irq;
}
+
+ has_gic_active_state = true;
}
kvm_info("virtual timer IRQ%d\n", host_vtimer_irq);
--
2.14.2