panic, kexec: make __crash_kexec() NMI safe

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:

  6ce47fd961 ("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: 6ce47fd961 ("rtmutex: Warn if trylock is called from hard/softirq context")
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Baoquan He <bhe@redhat.com>
Cc: "Eric W . Biederman" <ebiederm@xmission.com>
Cc: Juri Lelli <jlelli@redhat.com>
Cc: Luis Claudio R. Goncalves <lgoncalv@redhat.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

kernel/kexec.c

@@ -93,13 +93,10 @@ 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) {
@@ -165,7 +162,7 @@ out:
 
 	kimage_free(image);
 out_unlock:
-	mutex_unlock(&kexec_mutex);
+	kexec_unlock();
 	return ret;
 }
 

kernel/kexec_core.c

@@ -46,7 +46,7 @@
 #include <crypto/hash.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;
@@ -959,7 +959,7 @@ late_initcall(kexec_core_sysctl_init);
  */
 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.
 	 *
@@ -967,7 +967,7 @@ 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;
 
@@ -976,7 +976,7 @@ 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);
@@ -1008,13 +1008,13 @@ 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;
 }
 
@@ -1025,7 +1025,7 @@ 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) {
@@ -1064,7 +1064,7 @@ int crash_shrink_memory(unsigned long new_size)
 	insert_resource(&iomem_resource, ram_res);
 
 unlock:
-	mutex_unlock(&kexec_mutex);
+	kexec_unlock();
 	return ret;
 }
 
@@ -1136,7 +1136,7 @@ int kernel_kexec(void)
 {
 	int error = 0;
 
-	if (!mutex_trylock(&kexec_mutex))
+	if (!kexec_trylock())
 		return -EBUSY;
 	if (!kexec_image) {
 		error = -EINVAL;
@@ -1212,6 +1212,6 @@ int kernel_kexec(void)
 #endif
 
  Unlock:
-	mutex_unlock(&kexec_mutex);
+	kexec_unlock();
 	return error;
 }

kernel/kexec_file.c

@@ -339,7 +339,7 @@ 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;
@@ -411,7 +411,7 @@ out:
 	if ((flags & KEXEC_FILE_ON_CRASH) && kexec_crash_image)
 		arch_kexec_protect_crashkres();
 
-	mutex_unlock(&kexec_mutex);
+	kexec_unlock();
 	kimage_free(image);
 	return ret;
 }

kernel/kexec_internal.h

@@ -13,7 +13,20 @@ void kimage_terminate(struct kimage *image);
 int kimage_is_destination_range(struct kimage *image,
 				unsigned long start, unsigned long end);
 
-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>