From: Valentin Schneider <vschneid(a)redhat.com>
commit 05c6257433b7212f07a7e53479a8ab038fc1666a upstream.
Attempting to get a crash dump out of a debug PREEMPT_RT kernel via an NMI
panic() doesn't work. The cause of that lies in the PREEMPT_RT definition
of mutex_trylock():
if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEXES) && WARN_ON_ONCE(!in_task()))
return 0;
This prevents an nmi_panic() from executing the main body of
__crash_kexec() which does the actual kexec into the kdump kernel. The
warning and return are explained by:
6ce47fd961fa ("rtmutex: Warn if trylock is called from hard/softirq context")
[...]
The reasons for this are:
1) There is a potential deadlock in the slowpath
2) Another cpu which blocks on the rtmutex will boost the task
which allegedly locked the rtmutex, but that cannot work
because the hard/softirq context borrows the task context.
Furthermore, grabbing the lock isn't NMI safe, so do away with kexec_mutex
and replace it with an atomic variable. This is somewhat overzealous as
*some* callsites could keep using a mutex (e.g. the sysfs-facing ones
like crash_shrink_memory()), but this has the benefit of involving a
single unified lock and preventing any future NMI-related surprises.
Tested by triggering NMI panics via:
$ echo 1 > /proc/sys/kernel/panic_on_unrecovered_nmi
$ echo 1 > /proc/sys/kernel/unknown_nmi_panic
$ echo 1 > /proc/sys/kernel/panic
$ ipmitool power diag
Link: https://lkml.kernel.org/r/20220630223258.4144112-3-vschneid@redhat.com
Fixes: 6ce47fd961fa ("rtmutex: Warn if trylock is called from hard/softirq context")
Signed-off-by: Valentin Schneider <vschneid(a)redhat.com>
Cc: Arnd Bergmann <arnd(a)arndb.de>
Cc: Baoquan He <bhe(a)redhat.com>
Cc: "Eric W . Biederman" <ebiederm(a)xmission.com>
Cc: Juri Lelli <jlelli(a)redhat.com>
Cc: Luis Claudio R. Goncalves <lgoncalv(a)redhat.com>
Cc: Miaohe Lin <linmiaohe(a)huawei.com>
Cc: Petr Mladek <pmladek(a)suse.com>
Cc: Sebastian Andrzej Siewior <bigeasy(a)linutronix.de>
Cc: Thomas Gleixner <tglx(a)linutronix.de>
Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org>
Cc: stable(a)vger.kernel.org # 5.10+
Signed-off-by: Wen Yang <wenyang.linux(a)foxmail.com>
---
kernel/kexec.c | 11 ++++-------
kernel/kexec_core.c | 20 ++++++++++----------
kernel/kexec_file.c | 4 ++--
kernel/kexec_internal.h | 15 ++++++++++++++-
4 files changed, 30 insertions(+), 20 deletions(-)
diff --git a/kernel/kexec.c b/kernel/kexec.c
index 9c7aef8f4bb6..f0f0c6555454 100644
--- a/kernel/kexec.c
+++ b/kernel/kexec.c
@@ -112,13 +112,10 @@ static int do_kexec_load(unsigned long entry, unsigned long nr_segments,
/*
* Because we write directly to the reserved memory region when loading
- * crash kernels we need a mutex here to prevent multiple crash kernels
- * from attempting to load simultaneously, and to prevent a crash kernel
- * from loading over the top of a in use crash kernel.
- *
- * KISS: always take the mutex.
+ * crash kernels we need a serialization here to prevent multiple crash
+ * kernels from attempting to load simultaneously.
*/
- if (!mutex_trylock(&kexec_mutex))
+ if (!kexec_trylock())
return -EBUSY;
if (flags & KEXEC_ON_CRASH) {
@@ -184,7 +181,7 @@ static int do_kexec_load(unsigned long entry, unsigned long nr_segments,
kimage_free(image);
out_unlock:
- mutex_unlock(&kexec_mutex);
+ kexec_unlock();
return ret;
}
diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
index e47870f30728..7a8104d48997 100644
--- a/kernel/kexec_core.c
+++ b/kernel/kexec_core.c
@@ -45,7 +45,7 @@
#include <crypto/sha.h>
#include "kexec_internal.h"
-DEFINE_MUTEX(kexec_mutex);
+atomic_t __kexec_lock = ATOMIC_INIT(0);
/* Per cpu memory for storing cpu states in case of system crash. */
note_buf_t __percpu *crash_notes;
@@ -943,7 +943,7 @@ int kexec_load_disabled;
*/
void __noclone __crash_kexec(struct pt_regs *regs)
{
- /* Take the kexec_mutex here to prevent sys_kexec_load
+ /* Take the kexec_lock here to prevent sys_kexec_load
* running on one cpu from replacing the crash kernel
* we are using after a panic on a different cpu.
*
@@ -951,7 +951,7 @@ void __noclone __crash_kexec(struct pt_regs *regs)
* of memory the xchg(&kexec_crash_image) would be
* sufficient. But since I reuse the memory...
*/
- if (mutex_trylock(&kexec_mutex)) {
+ if (kexec_trylock()) {
if (kexec_crash_image) {
struct pt_regs fixed_regs;
@@ -960,7 +960,7 @@ void __noclone __crash_kexec(struct pt_regs *regs)
machine_crash_shutdown(&fixed_regs);
machine_kexec(kexec_crash_image);
}
- mutex_unlock(&kexec_mutex);
+ kexec_unlock();
}
}
STACK_FRAME_NON_STANDARD(__crash_kexec);
@@ -993,13 +993,13 @@ ssize_t crash_get_memory_size(void)
{
ssize_t size = 0;
- if (!mutex_trylock(&kexec_mutex))
+ if (!kexec_trylock())
return -EBUSY;
if (crashk_res.end != crashk_res.start)
size = resource_size(&crashk_res);
- mutex_unlock(&kexec_mutex);
+ kexec_unlock();
return size;
}
@@ -1019,7 +1019,7 @@ int crash_shrink_memory(unsigned long new_size)
unsigned long old_size;
struct resource *ram_res;
- if (!mutex_trylock(&kexec_mutex))
+ if (!kexec_trylock())
return -EBUSY;
if (kexec_crash_image) {
@@ -1058,7 +1058,7 @@ int crash_shrink_memory(unsigned long new_size)
insert_resource(&iomem_resource, ram_res);
unlock:
- mutex_unlock(&kexec_mutex);
+ kexec_unlock();
return ret;
}
@@ -1130,7 +1130,7 @@ int kernel_kexec(void)
{
int error = 0;
- if (!mutex_trylock(&kexec_mutex))
+ if (!kexec_trylock())
return -EBUSY;
if (!kexec_image) {
error = -EINVAL;
@@ -1205,7 +1205,7 @@ int kernel_kexec(void)
#endif
Unlock:
- mutex_unlock(&kexec_mutex);
+ kexec_unlock();
return error;
}
diff --git a/kernel/kexec_file.c b/kernel/kexec_file.c
index fff11916aba3..b9c857782ada 100644
--- a/kernel/kexec_file.c
+++ b/kernel/kexec_file.c
@@ -343,7 +343,7 @@ SYSCALL_DEFINE5(kexec_file_load, int, kernel_fd, int, initrd_fd,
image = NULL;
- if (!mutex_trylock(&kexec_mutex))
+ if (!kexec_trylock())
return -EBUSY;
dest_image = &kexec_image;
@@ -415,7 +415,7 @@ SYSCALL_DEFINE5(kexec_file_load, int, kernel_fd, int, initrd_fd,
if ((flags & KEXEC_FILE_ON_CRASH) && kexec_crash_image)
arch_kexec_protect_crashkres();
- mutex_unlock(&kexec_mutex);
+ kexec_unlock();
kimage_free(image);
return ret;
}
diff --git a/kernel/kexec_internal.h b/kernel/kexec_internal.h
index 39d30ccf8d87..49d4e3ab9c96 100644
--- a/kernel/kexec_internal.h
+++ b/kernel/kexec_internal.h
@@ -15,7 +15,20 @@ int kimage_is_destination_range(struct kimage *image,
int machine_kexec_post_load(struct kimage *image);
-extern struct mutex kexec_mutex;
+/*
+ * Whatever is used to serialize accesses to the kexec_crash_image needs to be
+ * NMI safe, as __crash_kexec() can happen during nmi_panic(), so here we use a
+ * "simple" atomic variable that is acquired with a cmpxchg().
+ */
+extern atomic_t __kexec_lock;
+static inline bool kexec_trylock(void)
+{
+ return atomic_cmpxchg_acquire(&__kexec_lock, 0, 1) == 0;
+}
+static inline void kexec_unlock(void)
+{
+ atomic_set_release(&__kexec_lock, 0);
+}
#ifdef CONFIG_KEXEC_FILE
#include <linux/purgatory.h>
--
2.37.2
From: Valentin Schneider <vschneid(a)redhat.com>
commit 7bb5da0d490b2d836c5218f5186ee588d2145310 upstream.
Patch series "kexec, panic: Making crash_kexec() NMI safe", v4.
This patch (of 2):
Most acquistions of kexec_mutex are done via mutex_trylock() - those were
a direct "translation" from:
8c5a1cf0ad3a ("kexec: use a mutex for locking rather than xchg()")
there have however been two additions since then that use mutex_lock():
crash_get_memory_size() and crash_shrink_memory().
A later commit will replace said mutex with an atomic variable, and
locking operations will become atomic_cmpxchg(). Rather than having those
mutex_lock() become while (atomic_cmpxchg(&lock, 0, 1)), turn them into
trylocks that can return -EBUSY on acquisition failure.
This does halve the printable size of the crash kernel, but that's still
neighbouring 2G for 32bit kernels which should be ample enough.
Link: https://lkml.kernel.org/r/20220630223258.4144112-1-vschneid@redhat.com
Link: https://lkml.kernel.org/r/20220630223258.4144112-2-vschneid@redhat.com
Signed-off-by: Valentin Schneider <vschneid(a)redhat.com>
Cc: Arnd Bergmann <arnd(a)arndb.de>
Cc: "Eric W . Biederman" <ebiederm(a)xmission.com>
Cc: Juri Lelli <jlelli(a)redhat.com>
Cc: Luis Claudio R. Goncalves <lgoncalv(a)redhat.com>
Cc: Miaohe Lin <linmiaohe(a)huawei.com>
Cc: Petr Mladek <pmladek(a)suse.com>
Cc: Sebastian Andrzej Siewior <bigeasy(a)linutronix.de>
Cc: Thomas Gleixner <tglx(a)linutronix.de>
Cc: Baoquan He <bhe(a)redhat.com>
Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org>
Cc: stable(a)vger.kernel.org # 5.10+
Signed-off-by: Wen Yang <wenyang.linux(a)foxmail.com>
---
include/linux/kexec.h | 2 +-
kernel/kexec_core.c | 12 ++++++++----
kernel/ksysfs.c | 7 ++++++-
3 files changed, 15 insertions(+), 6 deletions(-)
diff --git a/include/linux/kexec.h b/include/linux/kexec.h
index a1f12e959bba..3c1deba496c9 100644
--- a/include/linux/kexec.h
+++ b/include/linux/kexec.h
@@ -380,8 +380,8 @@ extern note_buf_t __percpu *crash_notes;
extern bool kexec_in_progress;
int crash_shrink_memory(unsigned long new_size);
-size_t crash_get_memory_size(void);
void crash_free_reserved_phys_range(unsigned long begin, unsigned long end);
+ssize_t crash_get_memory_size(void);
void arch_kexec_protect_crashkres(void);
void arch_kexec_unprotect_crashkres(void);
diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
index c589c7a9562c..e47870f30728 100644
--- a/kernel/kexec_core.c
+++ b/kernel/kexec_core.c
@@ -989,13 +989,16 @@ void crash_kexec(struct pt_regs *regs)
}
}
-size_t crash_get_memory_size(void)
+ssize_t crash_get_memory_size(void)
{
- size_t size = 0;
+ ssize_t size = 0;
+
+ if (!mutex_trylock(&kexec_mutex))
+ return -EBUSY;
- mutex_lock(&kexec_mutex);
if (crashk_res.end != crashk_res.start)
size = resource_size(&crashk_res);
+
mutex_unlock(&kexec_mutex);
return size;
}
@@ -1016,7 +1019,8 @@ int crash_shrink_memory(unsigned long new_size)
unsigned long old_size;
struct resource *ram_res;
- mutex_lock(&kexec_mutex);
+ if (!mutex_trylock(&kexec_mutex))
+ return -EBUSY;
if (kexec_crash_image) {
ret = -ENOENT;
diff --git a/kernel/ksysfs.c b/kernel/ksysfs.c
index 35859da8bd4f..e20c19e3ba49 100644
--- a/kernel/ksysfs.c
+++ b/kernel/ksysfs.c
@@ -106,7 +106,12 @@ KERNEL_ATTR_RO(kexec_crash_loaded);
static ssize_t kexec_crash_size_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
- return sprintf(buf, "%zu\n", crash_get_memory_size());
+ ssize_t size = crash_get_memory_size();
+
+ if (size < 0)
+ return size;
+
+ return sprintf(buf, "%zd\n", size);
}
static ssize_t kexec_crash_size_store(struct kobject *kobj,
struct kobj_attribute *attr,
--
2.37.2
From: Arnd Bergmann <arnd(a)arndb.de>
commit 4b692e861619353ce069e547a67c8d0e32d9ef3d upstream.
Patch series "compat: remove compat_alloc_user_space", v5.
Going through compat_alloc_user_space() to convert indirect system call
arguments tends to add complexity compared to handling the native and
compat logic in the same code.
This patch (of 6):
The locking is the same between the native and compat version of
sys_kexec_load(), so it can be done in the common implementation to reduce
duplication.
Link: https://lkml.kernel.org/r/20210727144859.4150043-1-arnd@kernel.org
Link: https://lkml.kernel.org/r/20210727144859.4150043-2-arnd@kernel.org
Signed-off-by: Arnd Bergmann <arnd(a)arndb.de>
Co-developed-by: Eric Biederman <ebiederm(a)xmission.com>
Co-developed-by: Christoph Hellwig <hch(a)infradead.org>
Acked-by: "Eric W. Biederman" <ebiederm(a)xmission.com>
Cc: Catalin Marinas <catalin.marinas(a)arm.com>
Cc: Will Deacon <will(a)kernel.org>
Cc: Thomas Bogendoerfer <tsbogend(a)alpha.franken.de>
Cc: "James E.J. Bottomley" <James.Bottomley(a)HansenPartnership.com>
Cc: Helge Deller <deller(a)gmx.de>
Cc: Michael Ellerman <mpe(a)ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh(a)kernel.crashing.org>
Cc: Paul Mackerras <paulus(a)samba.org>
Cc: Heiko Carstens <hca(a)linux.ibm.com>
Cc: Vasily Gorbik <gor(a)linux.ibm.com>
Cc: Christian Borntraeger <borntraeger(a)de.ibm.com>
Cc: "David S. Miller" <davem(a)davemloft.net>
Cc: Thomas Gleixner <tglx(a)linutronix.de>
Cc: Ingo Molnar <mingo(a)redhat.com>
Cc: Borislav Petkov <bp(a)alien8.de>
Cc: "H. Peter Anvin" <hpa(a)zytor.com>
Cc: Al Viro <viro(a)zeniv.linux.org.uk>
Cc: Feng Tang <feng.tang(a)intel.com>
Cc: Christoph Hellwig <hch(a)lst.de>
Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds(a)linux-foundation.org>
Cc: stable(a)vger.kernel.org # 5.10+
Signed-off-by: Wen Yang <wenyang.linux(a)foxmail.com>
---
kernel/kexec.c | 44 ++++++++++++++++----------------------------
1 file changed, 16 insertions(+), 28 deletions(-)
diff --git a/kernel/kexec.c b/kernel/kexec.c
index c82c6c06f051..9c7aef8f4bb6 100644
--- a/kernel/kexec.c
+++ b/kernel/kexec.c
@@ -110,6 +110,17 @@ static int do_kexec_load(unsigned long entry, unsigned long nr_segments,
unsigned long i;
int ret;
+ /*
+ * Because we write directly to the reserved memory region when loading
+ * crash kernels we need a mutex here to prevent multiple crash kernels
+ * from attempting to load simultaneously, and to prevent a crash kernel
+ * from loading over the top of a in use crash kernel.
+ *
+ * KISS: always take the mutex.
+ */
+ if (!mutex_trylock(&kexec_mutex))
+ return -EBUSY;
+
if (flags & KEXEC_ON_CRASH) {
dest_image = &kexec_crash_image;
if (kexec_crash_image)
@@ -121,7 +132,8 @@ static int do_kexec_load(unsigned long entry, unsigned long nr_segments,
if (nr_segments == 0) {
/* Uninstall image */
kimage_free(xchg(dest_image, NULL));
- return 0;
+ ret = 0;
+ goto out_unlock;
}
if (flags & KEXEC_ON_CRASH) {
/*
@@ -134,7 +146,7 @@ static int do_kexec_load(unsigned long entry, unsigned long nr_segments,
ret = kimage_alloc_init(&image, entry, nr_segments, segments, flags);
if (ret)
- return ret;
+ goto out_unlock;
if (flags & KEXEC_PRESERVE_CONTEXT)
image->preserve_context = 1;
@@ -171,6 +183,8 @@ static int do_kexec_load(unsigned long entry, unsigned long nr_segments,
arch_kexec_protect_crashkres();
kimage_free(image);
+out_unlock:
+ mutex_unlock(&kexec_mutex);
return ret;
}
@@ -247,21 +261,8 @@ SYSCALL_DEFINE4(kexec_load, unsigned long, entry, unsigned long, nr_segments,
((flags & KEXEC_ARCH_MASK) != KEXEC_ARCH_DEFAULT))
return -EINVAL;
- /* Because we write directly to the reserved memory
- * region when loading crash kernels we need a mutex here to
- * prevent multiple crash kernels from attempting to load
- * simultaneously, and to prevent a crash kernel from loading
- * over the top of a in use crash kernel.
- *
- * KISS: always take the mutex.
- */
- if (!mutex_trylock(&kexec_mutex))
- return -EBUSY;
-
result = do_kexec_load(entry, nr_segments, segments, flags);
- mutex_unlock(&kexec_mutex);
-
return result;
}
@@ -301,21 +302,8 @@ COMPAT_SYSCALL_DEFINE4(kexec_load, compat_ulong_t, entry,
return -EFAULT;
}
- /* Because we write directly to the reserved memory
- * region when loading crash kernels we need a mutex here to
- * prevent multiple crash kernels from attempting to load
- * simultaneously, and to prevent a crash kernel from loading
- * over the top of a in use crash kernel.
- *
- * KISS: always take the mutex.
- */
- if (!mutex_trylock(&kexec_mutex))
- return -EBUSY;
-
result = do_kexec_load(entry, nr_segments, ksegments, flags);
- mutex_unlock(&kexec_mutex);
-
return result;
}
#endif
--
2.37.2