Merge tag 'for-linus-5.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml

Pull UML updates from Richard Weinberger:

- A new timer mode, time travel, for testing with UML

- Many bugixes/improvements for the serial line driver

- Various bugfixes

* tag 'for-linus-5.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml:
um: fix build without CONFIG_UML_TIME_TRAVEL_SUPPORT
um: Fix kcov crash during startup
um: configs: Remove useless UEVENT_HELPER_PATH
um: Support time travel mode
um: Pass nsecs to os timer functions
um: Remove drivers/ssl.h
um: Don't garbage collect in deactivate_all_fds()
um: Silence lockdep complaint about mmap_sem
um: Remove locking in deactivate_all_fds()
um: Timer code cleanup
um: fix os_timer_one_shot()
um: Fix IRQ controller regression on console read

arch/um/Kconfig

···
 
 	  If unsure, say Y.
 
+config UML_TIME_TRAVEL_SUPPORT
+	bool
+	prompt "Support time-travel mode (e.g. for test execution)"
+	help
+	  Enable this option to support time travel inside the UML instance.
+
+	  After enabling this option, two modes are accessible at runtime
+	  (selected by the kernel command line), see the kernel's command-
+	  line help for more details.
+
+	  It is safe to say Y, but you probably don't need this.
+
 endmenu
 
 source "arch/um/drivers/Kconfig"

arch/um/configs/i386_defconfig

···
 CONFIG_CON_CHAN="pts"
 CONFIG_SSL_CHAN="pts"
 CONFIG_UML_SOUND=m
-CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug"
 CONFIG_DEVTMPFS=y
 CONFIG_DEVTMPFS_MOUNT=y
 CONFIG_BLK_DEV_UBD=y

arch/um/configs/x86_64_defconfig

···
 CONFIG_CON_CHAN="pts"
 CONFIG_SSL_CHAN="pts"
 CONFIG_UML_SOUND=m
-CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug"
 CONFIG_DEVTMPFS=y
 CONFIG_DEVTMPFS_MOUNT=y
 CONFIG_BLK_DEV_UBD=y

arch/um/drivers/chan_kern.c

···
 	return err;
 }
 
+/* Items are added in IRQ context, when free_irq can't be called, and
+ * removed in process context, when it can.
+ * This handles interrupt sources which disappear, and which need to
+ * be permanently disabled.  This is discovered in IRQ context, but
+ * the freeing of the IRQ must be done later.
+ */
+static DEFINE_SPINLOCK(irqs_to_free_lock);
+static LIST_HEAD(irqs_to_free);
+
+void free_irqs(void)
+{
+	struct chan *chan;
+	LIST_HEAD(list);
+	struct list_head *ele;
+	unsigned long flags;
+
+	spin_lock_irqsave(&irqs_to_free_lock, flags);
+	list_splice_init(&irqs_to_free, &list);
+	spin_unlock_irqrestore(&irqs_to_free_lock, flags);
+
+	list_for_each(ele, &list) {
+		chan = list_entry(ele, struct chan, free_list);
+
+		if (chan->input && chan->enabled)
+			um_free_irq(chan->line->driver->read_irq, chan);
+		if (chan->output && chan->enabled)
+			um_free_irq(chan->line->driver->write_irq, chan);
+		chan->enabled = 0;
+	}
+}
+
 static void close_one_chan(struct chan *chan, int delay_free_irq)
 {
+	unsigned long flags;
+
 	if (!chan->opened)
 		return;
 
-    /* we can safely call free now - it will be marked
-     *  as free and freed once the IRQ stopped processing
-     */
-	if (chan->input && chan->enabled)
-		um_free_irq(chan->line->driver->read_irq, chan);
-	if (chan->output && chan->enabled)
-		um_free_irq(chan->line->driver->write_irq, chan);
-	chan->enabled = 0;
+	if (delay_free_irq) {
+		spin_lock_irqsave(&irqs_to_free_lock, flags);
+		list_add(&chan->free_list, &irqs_to_free);
+		spin_unlock_irqrestore(&irqs_to_free_lock, flags);
+	} else {
+		if (chan->input && chan->enabled)
+			um_free_irq(chan->line->driver->read_irq, chan);
+		if (chan->output && chan->enabled)
+			um_free_irq(chan->line->driver->write_irq, chan);
+		chan->enabled = 0;
+	}
 	if (chan->ops->close != NULL)
 		(*chan->ops->close)(chan->fd, chan->data);
 

arch/um/drivers/ssl.c

···
 #include <linux/console.h>
 #include <asm/termbits.h>
 #include <asm/irq.h>
-#include "ssl.h"
 #include "chan.h"
 #include <init.h>
 #include <irq_user.h>

arch/um/drivers/ssl.h

···
-/* 
- * Copyright (C) 2000 Jeff Dike ([e-mail])
- * Licensed under the GPL
- */
-
-#ifndef __SSL_H__
-#define __SSL_H__
-
-extern int ssl_read(int fd, int line);
-extern void ssl_receive_char(int line, char ch);
-
-#endif
-

arch/um/include/asm/mmu_context.h

···
 	 * when the new ->mm is used for the first time.
 	 */
 	__switch_mm(&new->context.id);
-	down_write(&new->mmap_sem);
+	down_write_nested(&new->mmap_sem, 1);
 	uml_setup_stubs(new);
 	up_write(&new->mmap_sem);
 }

arch/um/include/shared/os.h

···
 
 /* time.c */
 extern void os_idle_sleep(unsigned long long nsecs);
-extern int os_timer_create(void* timer);
-extern int os_timer_set_interval(void* timer, void* its);
-extern int os_timer_one_shot(int ticks);
-extern long long os_timer_disable(void);
-extern long os_timer_remain(void* timer);
+extern int os_timer_create(void);
+extern int os_timer_set_interval(unsigned long long nsecs);
+extern int os_timer_one_shot(unsigned long long nsecs);
+extern void os_timer_disable(void);
 extern void uml_idle_timer(void);
 extern long long os_persistent_clock_emulation(void);
 extern long long os_nsecs(void);
-extern long long os_vnsecs(void);
 
 /* skas/mem.c */
 extern long run_syscall_stub(struct mm_id * mm_idp,

arch/um/include/shared/timer-internal.h

···
 #define TIMER_MULTIPLIER 256
 #define TIMER_MIN_DELTA  500
 
+enum time_travel_mode {
+	TT_MODE_OFF,
+	TT_MODE_BASIC,
+	TT_MODE_INFCPU,
+};
+
+enum time_travel_timer_mode {
+	TT_TMR_DISABLED,
+	TT_TMR_ONESHOT,
+	TT_TMR_PERIODIC,
+};
+
+#ifdef CONFIG_UML_TIME_TRAVEL_SUPPORT
+extern enum time_travel_mode time_travel_mode;
+extern unsigned long long time_travel_time;
+extern enum time_travel_timer_mode time_travel_timer_mode;
+extern unsigned long long time_travel_timer_expiry;
+extern unsigned long long time_travel_timer_interval;
+
+static inline void time_travel_set_time(unsigned long long ns)
+{
+	time_travel_time = ns;
+}
+
+static inline void time_travel_set_timer(enum time_travel_timer_mode mode,
+					 unsigned long long expiry)
+{
+	time_travel_timer_mode = mode;
+	time_travel_timer_expiry = expiry;
+}
+#else
+#define time_travel_mode TT_MODE_OFF
+#define time_travel_time 0
+#define time_travel_timer_expiry 0
+#define time_travel_timer_interval 0
+
+static inline void time_travel_set_time(unsigned long long ns)
+{
+}
+
+static inline void time_travel_set_timer(enum time_travel_timer_mode mode,
+					 unsigned long long expiry)
+{
+}
+
+#define time_travel_timer_mode TT_TMR_DISABLED
+#endif
+
 #endif

arch/um/kernel/irq.c

···
 #include <irq_user.h>
 
 
+extern void free_irqs(void);
+
 /* When epoll triggers we do not know why it did so
  * we can also have different IRQs for read and write.
  * This is why we keep a small irq_fd array for each fd -
···
 			}
 		}
 	}
+
+	free_irqs();
 }
 
 static int assign_epoll_events_to_irq(struct irq_entry *irq_entry)
···
  */
 int deactivate_all_fds(void)
 {
-	unsigned long flags;
 	struct irq_entry *to_free;
 
-	spin_lock_irqsave(&irq_lock, flags);
 	/* Stop IO. The IRQ loop has no lock so this is our
 	 * only way of making sure we are safe to dispose
 	 * of all IRQ handlers
···
 		);
 		to_free = to_free->next;
 	}
-	garbage_collect_irq_entries();
-	spin_unlock_irqrestore(&irq_lock, flags);
+	/* don't garbage collect - we can no longer call kfree() here */
 	os_close_epoll_fd();
 	return 0;
 }

arch/um/kernel/process.c

···
 	kmalloc_ok = save_kmalloc_ok;
 }
 
+static void time_travel_sleep(unsigned long long duration)
+{
+	unsigned long long next = time_travel_time + duration;
+
+	if (time_travel_mode != TT_MODE_INFCPU)
+		os_timer_disable();
+
+	if (time_travel_timer_mode != TT_TMR_DISABLED ||
+	    time_travel_timer_expiry < next) {
+		if (time_travel_timer_mode == TT_TMR_ONESHOT)
+			time_travel_set_timer(TT_TMR_DISABLED, 0);
+		/*
+		 * time_travel_time will be adjusted in the timer
+		 * IRQ handler so it works even when the signal
+		 * comes from the OS timer
+		 */
+		deliver_alarm();
+	} else {
+		time_travel_set_time(next);
+	}
+
+	if (time_travel_mode != TT_MODE_INFCPU) {
+		if (time_travel_timer_mode == TT_TMR_PERIODIC)
+			os_timer_set_interval(time_travel_timer_interval);
+		else if (time_travel_timer_mode == TT_TMR_ONESHOT)
+			os_timer_one_shot(time_travel_timer_expiry - next);
+	}
+}
+
+static void um_idle_sleep(void)
+{
+	unsigned long long duration = UM_NSEC_PER_SEC;
+
+	if (time_travel_mode != TT_MODE_OFF) {
+		time_travel_sleep(duration);
+	} else {
+		os_idle_sleep(duration);
+	}
+}
+
 void arch_cpu_idle(void)
 {
 	cpu_tasks[current_thread_info()->cpu].pid = os_getpid();
-	os_idle_sleep(UM_NSEC_PER_SEC);
+	um_idle_sleep();
 	local_irq_enable();
 }
 

arch/um/kernel/skas/Makefile

···
 CFLAGS_clone.o := $(CFLAGS_NO_HARDENING)
 UNPROFILE_OBJS := clone.o
 
+KCOV_INSTRUMENT := n
+
 include arch/um/scripts/Makefile.rules

arch/um/kernel/skas/syscall.c

···
 #include <sysdep/ptrace.h>
 #include <sysdep/ptrace_user.h>
 #include <sysdep/syscalls.h>
+#include <shared/timer-internal.h>
 
 void handle_syscall(struct uml_pt_regs *r)
 {
 	struct pt_regs *regs = container_of(r, struct pt_regs, regs);
 	int syscall;
+
+	/*
+	 * If we have infinite CPU resources, then make every syscall also a
+	 * preemption point, since we don't have any other preemption in this
+	 * case, and kernel threads would basically never run until userspace
+	 * went to sleep, even if said userspace interacts with the kernel in
+	 * various ways.
+	 */
+	if (time_travel_mode == TT_MODE_INFCPU)
+		schedule();
 
 	/* Initialize the syscall number and default return value. */
 	UPT_SYSCALL_NR(r) = PT_SYSCALL_NR(r->gp);

arch/um/kernel/time.c

···
 #include <kern_util.h>
 #include <os.h>
 #include <timer-internal.h>
+#include <shared/init.h>
+
+#ifdef CONFIG_UML_TIME_TRAVEL_SUPPORT
+enum time_travel_mode time_travel_mode;
+unsigned long long time_travel_time;
+enum time_travel_timer_mode time_travel_timer_mode;
+unsigned long long time_travel_timer_expiry;
+unsigned long long time_travel_timer_interval;
+
+static bool time_travel_start_set;
+static unsigned long long time_travel_start;
+#else
+#define time_travel_start_set 0
+#define time_travel_start 0
+#endif
 
 void timer_handler(int sig, struct siginfo *unused_si, struct uml_pt_regs *regs)
 {
 	unsigned long flags;
+
+	if (time_travel_mode != TT_MODE_OFF)
+		time_travel_set_time(time_travel_timer_expiry);
 
 	local_irq_save(flags);
 	do_IRQ(TIMER_IRQ, regs);
···
 
 static int itimer_shutdown(struct clock_event_device *evt)
 {
-	os_timer_disable();
+	if (time_travel_mode != TT_MODE_OFF)
+		time_travel_set_timer(TT_TMR_DISABLED, 0);
+
+	if (time_travel_mode != TT_MODE_INFCPU)
+		os_timer_disable();
+
 	return 0;
 }
 
 static int itimer_set_periodic(struct clock_event_device *evt)
 {
-	os_timer_set_interval(NULL, NULL);
+	unsigned long long interval = NSEC_PER_SEC / HZ;
+
+	if (time_travel_mode != TT_MODE_OFF)
+		time_travel_set_timer(TT_TMR_PERIODIC,
+				      time_travel_time + interval);
+
+	if (time_travel_mode != TT_MODE_INFCPU)
+		os_timer_set_interval(interval);
+
 	return 0;
 }
 
 static int itimer_next_event(unsigned long delta,
 			     struct clock_event_device *evt)
 {
-	return os_timer_one_shot(delta);
+	delta += 1;
+
+	if (time_travel_mode != TT_MODE_OFF)
+		time_travel_set_timer(TT_TMR_ONESHOT,
+				      time_travel_time + delta);
+
+	if (time_travel_mode != TT_MODE_INFCPU)
+		return os_timer_one_shot(delta);
+
+	return 0;
 }
 
 static int itimer_one_shot(struct clock_event_device *evt)
 {
-	os_timer_one_shot(1);
-	return 0;
+	return itimer_next_event(0, evt);
 }
 
 static struct clock_event_device timer_clockevent = {
···
 
 static u64 timer_read(struct clocksource *cs)
 {
+	if (time_travel_mode != TT_MODE_OFF) {
+		/*
+		 * We make reading the timer cost a bit so that we don't get
+		 * stuck in loops that expect time to move more than the
+		 * exact requested sleep amount, e.g. python's socket server,
+		 * see https://bugs.python.org/issue37026.
+		 */
+		time_travel_set_time(time_travel_time + TIMER_MULTIPLIER);
+		return time_travel_time / TIMER_MULTIPLIER;
+	}
+
 	return os_nsecs() / TIMER_MULTIPLIER;
 }
 
···
 		printk(KERN_ERR "register_timer : request_irq failed - "
 		       "errno = %d\n", -err);
 
-	err = os_timer_create(NULL);
+	err = os_timer_create();
 	if (err != 0) {
 		printk(KERN_ERR "creation of timer failed - errno = %d\n", -err);
 		return;
···
 
 void read_persistent_clock64(struct timespec64 *ts)
 {
-	long long nsecs = os_persistent_clock_emulation();
+	long long nsecs;
+
+	if (time_travel_start_set)
+		nsecs = time_travel_start + time_travel_time;
+	else
+		nsecs = os_persistent_clock_emulation();
 
 	set_normalized_timespec64(ts, nsecs / NSEC_PER_SEC,
 				  nsecs % NSEC_PER_SEC);
···
 	timer_set_signal_handler();
 	late_time_init = um_timer_setup;
 }
+
+#ifdef CONFIG_UML_TIME_TRAVEL_SUPPORT
+unsigned long calibrate_delay_is_known(void)
+{
+	if (time_travel_mode == TT_MODE_INFCPU)
+		return 1;
+	return 0;
+}
+
+int setup_time_travel(char *str)
+{
+	if (strcmp(str, "=inf-cpu") == 0) {
+		time_travel_mode = TT_MODE_INFCPU;
+		timer_clockevent.name = "time-travel-timer-infcpu";
+		timer_clocksource.name = "time-travel-clock";
+		return 1;
+	}
+
+	if (!*str) {
+		time_travel_mode = TT_MODE_BASIC;
+		timer_clockevent.name = "time-travel-timer";
+		timer_clocksource.name = "time-travel-clock";
+		return 1;
+	}
+
+	return -EINVAL;
+}
+
+__setup("time-travel", setup_time_travel);
+__uml_help(setup_time_travel,
+"time-travel\n"
+"This option just enables basic time travel mode, in which the clock/timers\n"
+"inside the UML instance skip forward when there's nothing to do, rather than\n"
+"waiting for real time to elapse. However, instance CPU speed is limited by\n"
+"the real CPU speed, so e.g. a 10ms timer will always fire after ~10ms wall\n"
+"clock (but quicker when there's nothing to do).\n"
+"\n"
+"time-travel=inf-cpu\n"
+"This enables time travel mode with infinite processing power, in which there\n"
+"are no wall clock timers, and any CPU processing happens - as seen from the\n"
+"guest - instantly. This can be useful for accurate simulation regardless of\n"
+"debug overhead, physical CPU speed, etc. but is somewhat dangerous as it can\n"
+"easily lead to getting stuck (e.g. if anything in the system busy loops).\n");
+
+int setup_time_travel_start(char *str)
+{
+	int err;
+
+	err = kstrtoull(str, 0, &time_travel_start);
+	if (err)
+		return err;
+
+	time_travel_start_set = 1;
+	return 1;
+}
+
+__setup("time-travel-start", setup_time_travel_start);
+__uml_help(setup_time_travel_start,
+"time-travel-start=<seconds>\n"
+"Configure the UML instance's wall clock to start at this value rather than\n"
+"the host's wall clock at the time of UML boot.\n");
+#endif

arch/um/os-Linux/time.c

···
 
 static inline long long timespec_to_ns(const struct timespec *ts)
 {
-	return ((long long) ts->tv_sec * UM_NSEC_PER_SEC) +
-		ts->tv_nsec;
+	return ((long long) ts->tv_sec * UM_NSEC_PER_SEC) + ts->tv_nsec;
 }
 
-long long os_persistent_clock_emulation (void) {
+long long os_persistent_clock_emulation(void)
+{
 	struct timespec realtime_tp;
 
 	clock_gettime(CLOCK_REALTIME, &realtime_tp);
···
 /**
  * os_timer_create() - create an new posix (interval) timer
  */
-int os_timer_create(void* timer) {
+int os_timer_create(void)
+{
+	timer_t *t = &event_high_res_timer;
 
-	timer_t* t = timer;
-
-	if(t == NULL) {
-		t = &event_high_res_timer;
-	}
-
-	if (timer_create(
-		CLOCK_MONOTONIC,
-		NULL,
-		t) == -1) {
+	if (timer_create(CLOCK_MONOTONIC, NULL, t) == -1)
 		return -1;
-	}
-	return 0;
-}
-
-int os_timer_set_interval(void* timer, void* i)
-{
-	struct itimerspec its;
-	unsigned long long nsec;
-	timer_t* t = timer;
-	struct itimerspec* its_in = i;
-
-	if(t == NULL) {
-		t = &event_high_res_timer;
-	}
-
-	nsec = UM_NSEC_PER_SEC / UM_HZ;
-
-	if(its_in != NULL) {
-		its.it_value.tv_sec = its_in->it_value.tv_sec;
-		its.it_value.tv_nsec = its_in->it_value.tv_nsec;
-	} else {
-		its.it_value.tv_sec = 0;
-		its.it_value.tv_nsec = nsec;
-	}
-
-	its.it_interval.tv_sec = 0;
-	its.it_interval.tv_nsec = nsec;
-
-	if(timer_settime(*t, 0, &its, NULL) == -1) {
-		return -errno;
-	}
 
 	return 0;
 }
 
-/**
- * os_timer_remain() - returns the remaining nano seconds of the given interval
- *                     timer
- * Because this is the remaining time of an interval timer, which correspondends
- * to HZ, this value can never be bigger than one second. Just
- * the nanosecond part of the timer is returned.
- * The returned time is relative to the start time of the interval timer.
- * Return an negative value in an error case.
- */
-long os_timer_remain(void* timer)
+int os_timer_set_interval(unsigned long long nsecs)
 {
 	struct itimerspec its;
-	timer_t* t = timer;
 
-	if(t == NULL) {
-		t = &event_high_res_timer;
-	}
+	its.it_value.tv_sec = nsecs / UM_NSEC_PER_SEC;
+	its.it_value.tv_nsec = nsecs % UM_NSEC_PER_SEC;
 
-	if(timer_gettime(t, &its) == -1) {
+	its.it_interval.tv_sec = nsecs / UM_NSEC_PER_SEC;
+	its.it_interval.tv_nsec = nsecs % UM_NSEC_PER_SEC;
+
+	if (timer_settime(event_high_res_timer, 0, &its, NULL) == -1)
 		return -errno;
-	}
 
-	return its.it_value.tv_nsec;
+	return 0;
 }
 
-int os_timer_one_shot(int ticks)
+int os_timer_one_shot(unsigned long long nsecs)
 {
-	struct itimerspec its;
-	unsigned long long nsec;
-	unsigned long sec;
+	struct itimerspec its = {
+		.it_value.tv_sec = nsecs / UM_NSEC_PER_SEC,
+		.it_value.tv_nsec = nsecs % UM_NSEC_PER_SEC,
 
-    nsec = (ticks + 1);
-    sec = nsec / UM_NSEC_PER_SEC;
-	nsec = nsec % UM_NSEC_PER_SEC;
-
-	its.it_value.tv_sec = nsec / UM_NSEC_PER_SEC;
-	its.it_value.tv_nsec = nsec;
-
-	its.it_interval.tv_sec = 0;
-	its.it_interval.tv_nsec = 0; // we cheat here
+		.it_interval.tv_sec = 0,
+		.it_interval.tv_nsec = 0, // we cheat here
+	};
 
 	timer_settime(event_high_res_timer, 0, &its, NULL);
 	return 0;
···
 
 /**
  * os_timer_disable() - disable the posix (interval) timer
- * Returns the remaining interval timer time in nanoseconds
  */
-long long os_timer_disable(void)
+void os_timer_disable(void)
 {
 	struct itimerspec its;
 
 	memset(&its, 0, sizeof(struct itimerspec));
-	timer_settime(event_high_res_timer, 0, &its, &its);
-
-	return its.it_value.tv_sec * UM_NSEC_PER_SEC + its.it_value.tv_nsec;
-}
-
-long long os_vnsecs(void)
-{
-	struct timespec ts;
-
-	clock_gettime(CLOCK_PROCESS_CPUTIME_ID,&ts);
-	return timespec_to_ns(&ts);
+	timer_settime(event_high_res_timer, 0, &its, NULL);
 }
 
 long long os_nsecs(void)
···
  */
 void os_idle_sleep(unsigned long long nsecs)
 {
-	struct timespec ts;
-
-	if (nsecs <= 0) {
-		return;
-	}
-
-	ts = ((struct timespec) {
-			.tv_sec  = nsecs / UM_NSEC_PER_SEC,
-			.tv_nsec = nsecs % UM_NSEC_PER_SEC
-	});
+	struct timespec ts = {
+		.tv_sec  = nsecs / UM_NSEC_PER_SEC,
+		.tv_nsec = nsecs % UM_NSEC_PER_SEC
+	};
 
 	/*
 	 * Relay the signal if clock_nanosleep is interrupted.
 	 */
-	if (clock_nanosleep(CLOCK_MONOTONIC, 0, &ts, NULL)) {
+	if (clock_nanosleep(CLOCK_MONOTONIC, 0, &ts, NULL))
 		deliver_alarm();
-	}
 }