c/r: support for pending signals (v2)

c/r: support for pending signals (v2)

[Oren Laadan]

This is the 2nd round of c/r support for pending signals. This
version adds process itimers (real/virtual/prof). Changes:

  - Fix compilation warnings
  - [Louis Rilling] Remove SIGQUEUE_PREALLOC flag from queued signals
  - [Louis Rilling] Fail if task has posix-timers or SI_TIMER signal

This applies on top of ckpt-v17-dev (44430daae5d6...)

Oren.

[PATCH 1/4] c/r: blocked and template for shared signals (v2)

[Oren Laadan]

This patch adds checkpoint/restart of blocked signals mask
(t->blocked) and a template for shared signals (t->signal).

Because t->signal sharing is tied to threads, we ensure proper sharing
of t->signal (struct signal_struct) for threads only.

Access to t->signal is protected by locking t->sighand->lock.
Therefore, the usual checkpoint_obj() invoking the callback
checkpoint_signal(ctx, signal) is insufficient because the task
pointer is unavailable.

Instead, handling of t->signal sharing is explicit using helpers
like ckpt_obj_lookup_add(), ckpt_obj_fetch() and ckpt_obj_insert().
The actual state is saved (if needed) _after_ the task_objs data.

To prevent tasks from handling restored signals during restart,
set their mask to block all signals and only restore the original
mask at the very end (before the last sync point).

Introduce per-task pointer 'ckpt_data' to temporary store data
for restore actions that are deferred to the end (like restoring
the signal block mask).

Signed-off-by: Oren Laadan <orenl-eQaUEPhvms7ENvBUuze7eA@public.gmane.org>
---
 checkpoint/objhash.c           |    9 +++
 checkpoint/process.c           |   65 ++++++++++++++++++++++-
 checkpoint/signal.c            |  114 ++++++++++++++++++++++++++++++++++++---
 include/linux/checkpoint.h     |    6 ++
 include/linux/checkpoint_hdr.h |   14 +++++-
 5 files changed, 197 insertions(+), 11 deletions(-)

diff --git a/checkpoint/objhash.c b/checkpoint/objhash.c
index 3452e83..18bea71 100644
--- a/checkpoint/objhash.c
+++ b/checkpoint/objhash.c
@@ -290,6 +290,15 @@ static struct ckpt_obj_ops ckpt_obj_ops[] = {
 		.checkpoint = checkpoint_sighand,
 		.restore = restore_sighand,
 	},
+	/* signal object */
+	{
+		.obj_name = "SIGNAL",
+		.obj_type = CKPT_OBJ_SIGNAL,
+		.ref_drop = obj_no_drop,
+		.ref_grab = obj_no_grab,
+		.checkpoint = checkpoint_bad,
+		.restore = restore_bad,
+	},
 	/* ns object */
 	{
 		.obj_name = "NSPROXY",
diff --git a/checkpoint/process.c b/checkpoint/process.c
index d76ab2c..40b2580 100644
--- a/checkpoint/process.c
+++ b/checkpoint/process.c
@@ -181,7 +181,8 @@ static int checkpoint_task_objs(struct ckpt_ctx *ctx, struct task_struct *t)
 	int files_objref;
 	int mm_objref;
 	int sighand_objref;
-	int ret;
+	int signal_objref;
+	int first, ret;
 
 	/*
 	 * Shared objects may have dependencies among them: task->mm
@@ -224,14 +225,37 @@ static int checkpoint_task_objs(struct ckpt_ctx *ctx, struct task_struct *t)
 		return sighand_objref;
 	}
 
+	/*
+	 * Handle t->signal differently because the checkpoint method
+	 * for t->signal needs access to owning task_struct to access
+	 * t->sighand (to lock/unlock). First explicitly determine if
+	 * need to save, and only below invoke checkpoint_obj_signal()
+	 * if needed.
+	 */
+	signal_objref = ckpt_obj_lookup_add(ctx, t->signal,
+					    CKPT_OBJ_SIGNAL, &first);
+	ckpt_debug("signal: objref %d\n", signal_objref);
+	if (signal_objref < 0)
+		return signal_objref;
+
 	h = ckpt_hdr_get_type(ctx, sizeof(*h), CKPT_HDR_TASK_OBJS);
 	if (!h)
 		return -ENOMEM;
 	h->files_objref = files_objref;
 	h->mm_objref = mm_objref;
 	h->sighand_objref = sighand_objref;
+	h->signal_objref = signal_objref;
 	ret = ckpt_write_obj(ctx, &h->h);
 	ckpt_hdr_put(ctx, h);
+	if (ret < 0)
+		return ret;
+
+	/* actually save t->signal, if need to */
+	if (first)
+		ret = checkpoint_obj_signal(ctx, t);
+	if (ret < 0)
+		ckpt_write_err(ctx, "task %d (%s), signal_struct: %d",
+			       task_pid_vnr(t), t->comm, ret);
 
 	return ret;
 }
@@ -374,6 +398,10 @@ int checkpoint_task(struct ckpt_ctx *ctx, struct task_struct *t)
 		goto out;
 	ret = checkpoint_cpu(ctx, t);
 	ckpt_debug("cpu %d\n", ret);
+	if (ret < 0)
+		goto out;
+	ret = checkpoint_task_signal(ctx, t);
+	ckpt_debug("task-signal %d\n", ret);
  out:
 	return ret;
 }
@@ -547,6 +575,11 @@ static int restore_task_objs(struct ckpt_ctx *ctx)
 
 	ret = restore_obj_sighand(ctx, h->sighand_objref);
 	ckpt_debug("sighand: ret %d (%p)\n", ret, current->sighand);
+	if (ret < 0)
+		goto out;
+
+	ret = restore_obj_signal(ctx, h->signal_objref);
+	ckpt_debug("signal: ret %d (%p)\n", ret, current->signal);
  out:
 	ckpt_hdr_put(ctx, h);
 	return ret;
@@ -684,11 +717,37 @@ int restore_restart_block(struct ckpt_ctx *ctx)
 	return ret;
 }
 
+/* pre_restore_task - prepare the task for restore */
+static int pre_restore_task(struct ckpt_ctx *ctx)
+{
+	sigset_t sigset;
+
+	/*
+	 * Block task's signals to avoid interruptions due to signals,
+	 * say, from restored timers, file descriptors etc. Signals
+	 * will be unblocked when restore completes.
+	 *
+	 * NOTE: tasks with file descriptors set to send a SIGKILL as
+	 * i/o notification may fail the restart if a signal occurs
+	 * before that task completed its restore. FIX ?
+	 */
+	sigfillset(&sigset);
+	sigdelset(&sigset, SIGKILL);
+	sigdelset(&sigset, SIGSTOP);
+	sigprocmask(SIG_SETMASK, &sigset, NULL);
+
+	return 0;
+}
+
 /* read the entire state of the current task */
 int restore_task(struct ckpt_ctx *ctx)
 {
 	int ret;
 
+	ret = pre_restore_task(ctx);
+	if (ret < 0)
+		goto out;
+
 	ret = restore_task_struct(ctx);
 	ckpt_debug("task %d\n", ret);
 	if (ret < 0)
@@ -716,6 +775,10 @@ int restore_task(struct ckpt_ctx *ctx)
 		goto out;
 	ret = restore_creds(ctx);
 	ckpt_debug("creds: ret %d\n", ret);
+	if (ret < 0)
+		goto out;
+
+	ret = restore_task_signal(ctx);
  out:
 	return ret;
 }
diff --git a/checkpoint/signal.c b/checkpoint/signal.c
index 506476b..70df7c4 100644
--- a/checkpoint/signal.c
+++ b/checkpoint/signal.c
@@ -28,7 +28,7 @@ static inline void load_sigset(sigset_t *sigset, struct ckpt_hdr_sigset *h)
 }
 
 /***********************************************************************
- * Checkpoint
+ * sighand checkpoint/collect/restart
  */
 
 int do_checkpoint_sighand(struct ckpt_ctx *ctx, struct sighand_struct *sighand)
@@ -81,10 +81,6 @@ int checkpoint_obj_sighand(struct ckpt_ctx *ctx, struct task_struct *t)
 	return objref;
 }
 
-/***********************************************************************
- * Collect
- */
-
 int ckpt_collect_sighand(struct ckpt_ctx *ctx, struct task_struct *t)
 {
 	struct sighand_struct *sighand;
@@ -101,10 +97,6 @@ int ckpt_collect_sighand(struct ckpt_ctx *ctx, struct task_struct *t)
 	return ret;
 }
 
-/***********************************************************************
- * Restart
- */
-
 struct sighand_struct *do_restore_sighand(struct ckpt_ctx *ctx)
 {
 	struct ckpt_hdr_sighand *h;
@@ -168,3 +160,107 @@ int restore_obj_sighand(struct ckpt_ctx *ctx, int sighand_objref)
 
 	return 0;
 }
+
+/***********************************************************************
+ * signal checkpoint/restart
+ */
+
+static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
+{
+	struct ckpt_hdr_signal *h;
+	int ret;
+
+	h = ckpt_hdr_get_type(ctx, sizeof(*h), CKPT_HDR_SIGNAL);
+	if (!h)
+		return -ENOMEM;
+
+	/* fill in later */
+
+	ret = ckpt_write_obj(ctx, &h->h);
+	ckpt_hdr_put(ctx, h);
+	return ret;
+}
+
+int checkpoint_obj_signal(struct ckpt_ctx *ctx, struct task_struct *t)
+{
+	BUG_ON(t->flags & PF_EXITING);
+	return checkpoint_signal(ctx, t);
+}
+
+static int restore_signal(struct ckpt_ctx *ctx)
+{
+	struct ckpt_hdr_signal *h;
+
+	h = ckpt_read_obj_type(ctx, sizeof(*h), CKPT_HDR_SIGNAL);
+	if (IS_ERR(h))
+		return PTR_ERR(h);
+
+	/* fill in later */
+
+	ckpt_hdr_put(ctx, h);
+	return 0;
+}
+
+int restore_obj_signal(struct ckpt_ctx *ctx, int signal_objref)
+{
+	struct signal_struct *signal;
+	int ret = 0;
+
+	signal = ckpt_obj_fetch(ctx, signal_objref, CKPT_OBJ_SIGNAL);
+	if (!IS_ERR(signal)) {
+		/*
+		 * signal_struct is already shared properly as it is
+		 * tied to thread groups. Since thread relationships
+		 * are already restore now, t->signal must match.
+		 */
+		if (signal != current->signal)
+			ret = -EINVAL;
+	} else if (PTR_ERR(signal) == -EINVAL) {
+		/* first timer: add to hash and restore our t->signal */
+		ret = ckpt_obj_insert(ctx, current->signal,
+				      signal_objref, CKPT_OBJ_SIGNAL);
+		if (ret >= 0)
+			ret = restore_signal(ctx);
+	} else {
+		ret = PTR_ERR(signal);
+	}
+
+	return ret;
+}
+
+int checkpoint_task_signal(struct ckpt_ctx *ctx, struct task_struct *t)
+{
+	struct ckpt_hdr_signal_task *h;
+	int ret;
+
+	h = ckpt_hdr_get_type(ctx, sizeof(*h), CKPT_HDR_SIGNAL_TASK);
+	if (!h)
+		return -ENOMEM;
+
+	fill_sigset(&h->blocked, &t->blocked);
+
+	ret = ckpt_write_obj(ctx, &h->h);
+	ckpt_hdr_put(ctx, h);
+	return ret;
+}
+
+int restore_task_signal(struct ckpt_ctx *ctx)
+{
+	struct ckpt_hdr_signal_task *h;
+	sigset_t blocked;
+
+	h = ckpt_read_obj_type(ctx, sizeof(*h), CKPT_HDR_SIGNAL_TASK);
+	if (IS_ERR(h))
+		return PTR_ERR(h);
+
+	load_sigset(&blocked, &h->blocked);
+	/* silently remove SIGKILL, SIGSTOP */
+	sigdelset(&blocked, SIGKILL);
+	sigdelset(&blocked, SIGSTOP);
+
+	sigprocmask(SIG_SETMASK, &blocked, NULL);
+	recalc_sigpending();
+
+	ckpt_hdr_put(ctx, h);
+	return 0;
+}
diff --git a/include/linux/checkpoint.h b/include/linux/checkpoint.h
index 45cf0fe..87759c6 100644
--- a/include/linux/checkpoint.h
+++ b/include/linux/checkpoint.h
@@ -240,6 +240,12 @@ extern int ckpt_collect_sighand(struct ckpt_ctx *ctx, struct task_struct *t);
 extern int checkpoint_sighand(struct ckpt_ctx *ctx, void *ptr);
 extern void *restore_sighand(struct ckpt_ctx *ctx);
 
+extern int checkpoint_obj_signal(struct ckpt_ctx *ctx, struct task_struct *t);
+extern int restore_obj_signal(struct ckpt_ctx *ctx, int signal_objref);
+
+extern int checkpoint_task_signal(struct ckpt_ctx *ctx, struct task_struct *t);
+extern int restore_task_signal(struct ckpt_ctx *ctx);
+
 /* useful macros to copy fields and buffers to/from ckpt_hdr_xxx structures */
 #define CKPT_CPT 1
 #define CKPT_RST 2
diff --git a/include/linux/checkpoint_hdr.h b/include/linux/checkpoint_hdr.h
index 4b85956..63afdc7 100644
--- a/include/linux/checkpoint_hdr.h
+++ b/include/linux/checkpoint_hdr.h
@@ -87,6 +87,8 @@ enum {
 	CKPT_HDR_IPC_SEM,
 
 	CKPT_HDR_SIGHAND = 601,
+	CKPT_HDR_SIGNAL,
+	CKPT_HDR_SIGNAL_TASK,
 
 	CKPT_HDR_TAIL = 9001,
 
@@ -115,6 +117,7 @@ enum obj_type {
 	CKPT_OBJ_FILE,
 	CKPT_OBJ_MM,
 	CKPT_OBJ_SIGHAND,
+	CKPT_OBJ_SIGNAL,
 	CKPT_OBJ_NS,
 	CKPT_OBJ_UTS_NS,
 	CKPT_OBJ_IPC_NS,
@@ -209,7 +212,6 @@ struct ckpt_hdr_task {
 	__u32 compat_robust_futex_list; /* a compat __user ptr */
 	__u32 robust_futex_head_len;
 	__u64 robust_futex_list; /* a __user ptr */
-
 } __attribute__((aligned(8)));
 
 /* Posix capabilities */
@@ -285,6 +287,7 @@ struct ckpt_hdr_task_objs {
 	__s32 files_objref;
 	__s32 mm_objref;
 	__s32 sighand_objref;
+	__s32 signal_objref;
 } __attribute__((aligned(8)));
 
 /* restart blocks */
@@ -427,6 +430,15 @@ struct ckpt_hdr_sighand {
 	struct ckpt_hdr_sigaction action[0];
 } __attribute__((aligned(8)));
 
+struct ckpt_hdr_signal {
+	struct ckpt_hdr h;
+} __attribute__((aligned(8)));
+
+struct ckpt_hdr_signal_task {
+	struct ckpt_hdr h;
+	struct ckpt_hdr_sigset blocked;
+} __attribute__((aligned(8)));
+
 /* ipc commons */
 struct ckpt_hdr_ipcns {
 	struct ckpt_hdr h;
-- 
1.6.0.4

[PATCH 2/4] c/r: checkpoint/restart of rlimit (v2)

[Oren Laadan]

This patch adds checkpoint and restart of rlimit information
that is part of shared signal_struct.

Signed-off-by: Oren Laadan <orenl-eQaUEPhvms7ENvBUuze7eA@public.gmane.org>
---
 checkpoint/checkpoint.c        |    2 ++
 checkpoint/restart.c           |    3 +++
 checkpoint/signal.c            |   27 +++++++++++++++++++++++----
 include/linux/checkpoint_hdr.h |   17 +++++++++++++++++
 include/linux/resource.h       |    4 ++++
 kernel/sys.c                   |   36 +++++++++++++++++++++++-------------
 6 files changed, 72 insertions(+), 17 deletions(-)

diff --git a/checkpoint/checkpoint.c b/checkpoint/checkpoint.c
index c368270..870b16d 100644
--- a/checkpoint/checkpoint.c
+++ b/checkpoint/checkpoint.c
@@ -199,6 +199,8 @@ static void fill_kernel_const(struct ckpt_hdr_const *h)
 	h->uts_version_len = sizeof(uts->version);
 	h->uts_machine_len = sizeof(uts->machine);
 	h->uts_domainname_len = sizeof(uts->domainname);
+	/* rlimit */
+	h->rlimit_nlimits = RLIM_NLIMITS;
 }
 
 /* write the checkpoint header */
diff --git a/checkpoint/restart.c b/checkpoint/restart.c
index 65cafd9..e80df05 100644
--- a/checkpoint/restart.c
+++ b/checkpoint/restart.c
@@ -317,6 +317,9 @@ static int check_kernel_const(struct ckpt_hdr_const *h)
 		return -EINVAL;
 	if (h->uts_domainname_len != sizeof(uts->domainname))
 		return -EINVAL;
+	/* rlimit */
+	if (h->rlimit_nlimits != RLIM_NLIMITS)
+		return -EINVAL;
 
 	return 0;
 }
diff --git a/checkpoint/signal.c b/checkpoint/signal.c
index 70df7c4..b3f1d3e 100644
--- a/checkpoint/signal.c
+++ b/checkpoint/signal.c
@@ -14,6 +14,7 @@
 #include <linux/sched.h>
 #include <linux/signal.h>
 #include <linux/errno.h>
+#include <linux/resource.h>
 #include <linux/checkpoint.h>
 #include <linux/checkpoint_hdr.h>
 
@@ -168,13 +169,22 @@ int restore_obj_sighand(struct ckpt_ctx *ctx, int sighand_objref)
 static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
 {
 	struct ckpt_hdr_signal *h;
+	struct signal_struct *signal;
+	struct rlimit *rlim;
 	int ret;
 
 	h = ckpt_hdr_get_type(ctx, sizeof(*h), CKPT_HDR_SIGNAL);
 	if (!h)
 		return -ENOMEM;
 
-	/* fill in later */
+	signal = t->signal;
+	rlim = signal->rlim;
+
+	/* rlimit */
+	for (i = 0; i < RLIM_NLIMITS; i++) {
+		h->rlim[i].rlim_cur = rlim[i].rlim_cur;
+		h->rlim[i].rlim_max = rlim[i].rlim_max;
+	}
 
 	ret = ckpt_write_obj(ctx, &h->h);
 	ckpt_hdr_put(ctx, h);
@@ -190,15 +200,24 @@ int checkpoint_obj_signal(struct ckpt_ctx *ctx, struct task_struct *t)
 static int restore_signal(struct ckpt_ctx *ctx)
 {
 	struct ckpt_hdr_signal *h;
+	struct rlimit rlim;
+	int i, ret;
 
 	h = ckpt_read_obj_type(ctx, sizeof(*h), CKPT_HDR_SIGNAL);
 	if (IS_ERR(h))
 		return PTR_ERR(h);
 
-	/* fill in later */
-
+	/* rlimit */
+	for (i = 0; i < RLIM_NLIMITS; i++) {
+		rlim.rlim_cur = h->rlim[i].rlim_cur;
+		rlim.rlim_max = h->rlim[i].rlim_max;
+		ret = do_setrlimit(i, &rlim);
+		if (ret < 0)
+			break;
+	}
+ out:
 	ckpt_hdr_put(ctx, h);
-	return 0;
+	return ret;
 }
 
 int restore_obj_signal(struct ckpt_ctx *ctx, int signal_objref)
diff --git a/include/linux/checkpoint_hdr.h b/include/linux/checkpoint_hdr.h
index 63afdc7..0fd83b0 100644
--- a/include/linux/checkpoint_hdr.h
+++ b/include/linux/checkpoint_hdr.h
@@ -143,6 +143,8 @@ struct ckpt_hdr_const {
 	__u16 uts_version_len;
 	__u16 uts_machine_len;
 	__u16 uts_domainname_len;
+	/* rlimit */
+	__u16 rlimit_nlimits;
 } __attribute__((aligned(8)));
 
 /* checkpoint image header */
@@ -430,8 +432,23 @@ struct ckpt_hdr_sighand {
 	struct ckpt_hdr_sigaction action[0];
 } __attribute__((aligned(8)));
 
+struct ckpt_hdr_rlimit {
+	__u64 rlim_cur;
+	__u64 rlim_max;
+} __attribute__((aligned(8)));
+
+/* cannot include <linux/resource.h> from userspace, so define: */
+#define CKPT_RLIM_NLIMITS  16
+#ifdef __KERNEL__
+#include <linux/resource.h>
+#if CKPT_RLIM_NLIMITS != RLIM_NLIMITS
+#error CKPT_RLIM_NLIMIT size is wrong per asm-generic/resource.h
+#endif
+#endif
+
 struct ckpt_hdr_signal {
 	struct ckpt_hdr h;
+	struct ckpt_hdr_rlimit rlim[CKPT_RLIM_NLIMITS];
 } __attribute__((aligned(8)));
 
 struct ckpt_hdr_signal_task {
diff --git a/include/linux/resource.h b/include/linux/resource.h
index 40fc7e6..87e1bf3 100644
--- a/include/linux/resource.h
+++ b/include/linux/resource.h
@@ -72,4 +72,8 @@ struct rlimit {
 
 int getrusage(struct task_struct *p, int who, struct rusage __user *ru);
 
+#ifdef __KERNEL__
+extern int do_setrlimit(unsigned int resource, struct rlimit *rlim);
+#endif
+
 #endif
diff --git a/kernel/sys.c b/kernel/sys.c
index da4f9e0..0979a3f 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -1144,40 +1144,39 @@ SYSCALL_DEFINE2(old_getrlimit, unsigned int, resource,
 
 #endif
 
-SYSCALL_DEFINE2(setrlimit, unsigned int, resource, struct rlimit __user *, rlim)
+int do_setrlimit(unsigned int resource, struct rlimit *new_rlim)
 {
-	struct rlimit new_rlim, *old_rlim;
+	struct rlimit *old_rlim;
 	int retval;
 
 	if (resource >= RLIM_NLIMITS)
 		return -EINVAL;
-	if (copy_from_user(&new_rlim, rlim, sizeof(*rlim)))
-		return -EFAULT;
-	if (new_rlim.rlim_cur > new_rlim.rlim_max)
+	if (new_rlim->rlim_cur > new_rlim->rlim_max)
 		return -EINVAL;
+
 	old_rlim = current->signal->rlim + resource;
-	if ((new_rlim.rlim_max > old_rlim->rlim_max) &&
+	if ((new_rlim->rlim_max > old_rlim->rlim_max) &&
 	    !capable(CAP_SYS_RESOURCE))
 		return -EPERM;
-	if (resource == RLIMIT_NOFILE && new_rlim.rlim_max > sysctl_nr_open)
+	if (resource == RLIMIT_NOFILE && new_rlim->rlim_max > sysctl_nr_open)
 		return -EPERM;
 
-	retval = security_task_setrlimit(resource, &new_rlim);
+	retval = security_task_setrlimit(resource, new_rlim);
 	if (retval)
 		return retval;
 
-	if (resource == RLIMIT_CPU && new_rlim.rlim_cur == 0) {
+	if (resource == RLIMIT_CPU && new_rlim->rlim_cur == 0) {
 		/*
 		 * The caller is asking for an immediate RLIMIT_CPU
 		 * expiry.  But we use the zero value to mean "it was
 		 * never set".  So let's cheat and make it one second
 		 * instead
 		 */
-		new_rlim.rlim_cur = 1;
+		new_rlim->rlim_cur = 1;
 	}
 
 	task_lock(current->group_leader);
-	*old_rlim = new_rlim;
+	*old_rlim = *new_rlim;
 	task_unlock(current->group_leader);
 
 	if (resource != RLIMIT_CPU)
@@ -1189,14 +1188,25 @@ SYSCALL_DEFINE2(setrlimit, unsigned int, resource, struct rlimit __user *, rlim)
 	 * very long-standing error, and fixing it now risks breakage of
 	 * applications, so we live with it
 	 */
-	if (new_rlim.rlim_cur == RLIM_INFINITY)
+	if (new_rlim->rlim_cur == RLIM_INFINITY)
 		goto out;
 
-	update_rlimit_cpu(new_rlim.rlim_cur);
+	update_rlimit_cpu(new_rlim->rlim_cur);
 out:
 	return 0;
 }
 
+SYSCALL_DEFINE2(setrlimit, unsigned int, resource, struct rlimit __user *, rlim)
+{
+	struct rlimit new_rlim;
+
+	if (resource >= RLIM_NLIMITS)
+		return -EINVAL;
+	if (copy_from_user(&new_rlim, rlim, sizeof(*rlim)))
+		return -EFAULT;
+	return do_setrlimit(resource, &new_rlim);
+}
+
 /*
  * It would make sense to put struct rusage in the task_struct,
  * except that would make the task_struct be *really big*.  After
-- 
1.6.0.4

[PATCH 3/4] c/r: pending signals (private, shared) (v2)

[Oren Laadan]

This patch adds checkpoint and restart of pending signals queues:
struct sigpending, both per-task t->sigpending and shared (per-
thread-group) t->signal->shared_sigpending.

To checkpoint pending signals (private/shared) we first detach the
signal queue (and copy the mask) to a separate struct sigpending.
This separate structure can be iterated through without locking.

Once the state is saved, we re-attaches (prepends) the original signal
queue back to the original struct sigpending.

Signals that arrive(d) in the meantime will be suitably queued after
these (for real-time signals). Repeated non-realtime signals will not
be queued because they will already be marked in the pending mask,
that remains as is. This is the expected behavior of non-realtime
signals.

Changelog [v2]:
  - Fix compilation warnings
  - [Louis Rilling] Remove SIGQUEUE_PREALLOC flag from queued signals
  - [Louis Rilling] Fail if task has posix-timers or SI_TIMER signal

Signed-off-by: Oren Laadan <orenl-eQaUEPhvms7ENvBUuze7eA@public.gmane.org>
---
 checkpoint/signal.c            |  269 +++++++++++++++++++++++++++++++++++++++-
 include/linux/checkpoint_hdr.h |   23 ++++
 2 files changed, 291 insertions(+), 1 deletions(-)

diff --git a/checkpoint/signal.c b/checkpoint/signal.c
index b3f1d3e..c9da06b 100644
--- a/checkpoint/signal.c
+++ b/checkpoint/signal.c
@@ -166,12 +166,106 @@ int restore_obj_sighand(struct ckpt_ctx *ctx, int sighand_objref)
  * signal checkpoint/restart
  */
 
+static void fill_siginfo(struct ckpt_hdr_siginfo *si, siginfo_t *info)
+{
+	si->signo = info->si_signo;
+	si->_errno = info->si_errno;
+	si->code = info->si_code;
+
+	/* TODO: convert info->si_uid to uid_objref */
+
+	switch(info->si_code & __SI_MASK) {
+	case __SI_TIMER:
+		si->pid = info->si_tid;
+		si->uid = info->si_overrun;
+		si->sigval_int = info->si_int;
+		si->utime = info->si_sys_private;
+		break;
+	case __SI_POLL:
+		si->pid = info->si_band;
+		si->sigval_int = info->si_fd;
+		break;
+	case __SI_FAULT:
+		si->sigval_ptr = (unsigned long) info->si_addr;
+#ifdef __ARCH_SI_TRAPNO
+		si->sigval_int = info->si_trapno;
+#endif
+		break;
+	case __SI_CHLD:
+		si->pid = info->si_pid;
+		si->uid = info->si_uid;
+		si->sigval_int = info->si_status;
+		si->stime = info->si_stime;
+		si->utime = info->si_utime;
+		break;
+	/*
+	 * case __SI_KILL:
+	 * case __SI_RT:
+	 * case __SI_MESGQ:
+	 */
+	default:
+		si->pid = info->si_pid;
+		si->uid = info->si_uid;
+		si->sigval_ptr = (unsigned long) info->si_ptr;
+		break;
+	}
+}
+
+/*
+ * To checkpoint pending signals (private/shared) the caller moves the
+ * signal queue (and copies the mask) to a separate struct sigpending,
+ * therefore we can iterate through it without locking.
+ * After we return, the caller re-attaches (prepends) the original
+ * signal queue to the original struct sigpending. Thus, signals that
+ * arrive(d) in the meantime will be suitably queued after these.
+ * Finally, repeated non-realtime signals will not be queued because
+ * they will already be marked in the pending mask, that remains as is.
+ * This is the expected behavior of non-realtime signals.
+ */
+static int checkpoint_sigpending(struct ckpt_ctx *ctx,
+				 struct sigpending *pending)
+{
+	struct ckpt_hdr_sigpending *h;
+	struct ckpt_hdr_siginfo *si;
+	struct sigqueue *q;
+	int nr_pending = 0;
+	int ret;
+
+	list_for_each_entry(q, &pending->list, list) {
+		/* TODO: remove after adding support for posix-timers */
+		if ((q->info.si_code & __SI_MASK) == __SI_TIMER) {
+			ckpt_write_err(ctx, "bad signal type SI_TIMER\n");
+			return -ENOTSUPP;
+		}
+		nr_pending++;
+	}
+
+	h = ckpt_hdr_get_type(ctx, nr_pending * sizeof(*si) + sizeof(*h),
+			      CKPT_HDR_SIGPENDING);
+	if (!h)
+		return -ENOMEM;
+
+	h->nr_pending = nr_pending;
+	fill_sigset(&h->signal, &pending->signal);
+
+	si = h->siginfo;
+	list_for_each_entry(q, &pending->list, list)
+		fill_siginfo(si++, &q->info);
+
+	ret = ckpt_write_obj(ctx, &h->h);
+	ckpt_hdr_put(ctx, h);
+
+	return ret;
+}
+
 static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
 {
 	struct ckpt_hdr_signal *h;
 	struct signal_struct *signal;
+	struct sigpending shared_pending;
 	struct rlimit *rlim;
-	int ret;
+	unsigned long flags;
+	int i, ret;
 
 	h = ckpt_hdr_get_type(ctx, sizeof(*h), CKPT_HDR_SIGNAL);
 	if (!h)
@@ -180,13 +274,45 @@ static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
 	signal = t->signal;
 	rlim = signal->rlim;
 
+	INIT_LIST_HEAD(&shared_pending.list);
+
+	/* temporarily borrow signal queue - see chekcpoint_sigpending() */
+	if (!lock_task_sighand(t, &flags)) {
+		pr_warning("c/r: [%d] without sighand\n", task_pid_vnr(t));
+		goto out;
+	}
+
+	/* TODO: remove after adding support for posix-timers */
+	if (!list_empty(&signal->posix_timers)) {
+		ckpt_write_err(ctx, "task %d (%s) uses posix-timers\n",
+			       task_pid_vnr(t), t->comm);
+		unlock_task_sighand(t, &flags);
+		ret = -ENOTSUPP;
+		goto out;
+	}
+
+	list_splice_init(&signal->shared_pending.list, &shared_pending.list);
+	shared_pending.signal = signal->shared_pending.signal;
+
 	/* rlimit */
 	for (i = 0; i < RLIM_NLIMITS; i++) {
 		h->rlim[i].rlim_cur = rlim[i].rlim_cur;
 		h->rlim[i].rlim_max = rlim[i].rlim_max;
 	}
+	unlock_task_sighand(t, &flags);
 
 	ret = ckpt_write_obj(ctx, &h->h);
+	if (!ret)
+		ret = checkpoint_sigpending(ctx, &shared_pending);
+
+	/* return the borrowed queue */
+	if (!lock_task_sighand(t, &flags)) {
+		pr_warning("c/r: [%d] sighand disappeared\n", task_pid_vnr(t));
+		goto out;
+	}
+	list_splice(&shared_pending.list, &signal->shared_pending.list);
+	unlock_task_sighand(t, &flags);
+ out:
 	ckpt_hdr_put(ctx, h);
 	return ret;
 }
@@ -197,9 +323,99 @@ int checkpoint_obj_signal(struct ckpt_ctx *ctx, struct task_struct *t)
 	return checkpoint_signal(ctx, t);
 }
 
+static int load_siginfo(siginfo_t *info, struct ckpt_hdr_siginfo *si)
+{
+	if (!valid_signal(si->signo))
+		return -EINVAL;
+
+	info->si_signo = si->signo;
+	info->si_errno = si->_errno;
+	info->si_code = si->code;
+
+	/* TODO: validate remaining signal fields */
+
+	switch(info->si_code & __SI_MASK) {
+	case __SI_TIMER:
+		info->si_tid = si->pid;
+		info->si_overrun = si->uid;
+		info->si_int = si->sigval_int;
+		info->si_sys_private = si->utime;
+		break;
+	case __SI_POLL:
+		info->si_band = si->pid;
+		info->si_fd = si->sigval_int;
+		break;
+	case __SI_FAULT:
+		info->si_addr = (void __user *) (unsigned long) si->sigval_ptr;
+#ifdef __ARCH_SI_TRAPNO
+		info->si_trapno = si->sigval_int;
+#endif
+		break;
+	case __SI_CHLD:
+		info->si_pid = si->pid;
+		info->si_uid = si->uid;
+		info->si_status = si->sigval_int;
+		info->si_stime = si->stime;
+		info->si_utime = si->utime;
+		break;
+	case __SI_KILL:
+	case __SI_RT:
+	case __SI_MESGQ:
+		info->si_pid = si->pid;
+		info->si_uid = si->uid;
+		info->si_ptr = (void __user *) (unsigned long) si->sigval_ptr;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int restore_sigpending(struct ckpt_ctx *ctx, struct sigpending *pending)
+{
+	struct ckpt_hdr_sigpending *h;
+	struct ckpt_hdr_siginfo *si;
+	struct sigqueue *q;
+	int ret = 0;
+
+	h = ckpt_read_buf_type(ctx, 0, CKPT_HDR_SIGPENDING);
+	if (IS_ERR(h))
+		return PTR_ERR(h);
+
+	INIT_LIST_HEAD(&pending->list);
+	load_sigset(&pending->signal, &h->signal);
+
+	si = h->siginfo;
+	while (h->nr_pending--) {
+		q = sigqueue_alloc();
+		if (!q) {
+			ret = -ENOMEM;
+			break;
+		}
+
+		ret = load_siginfo(&q->info, si++);
+		if (ret < 0) {
+			sigqueue_free(q);
+			break;
+		}
+
+		q->flags &= ~SIGQUEUE_PREALLOC;
+		list_add_tail(&pending->list, &q->list);
+	}
+
+	if (ret < 0)
+		flush_sigqueue(pending);
+
+	ckpt_hdr_put(ctx, h);
+	return ret;
+}
+
 static int restore_signal(struct ckpt_ctx *ctx)
 {
 	struct ckpt_hdr_signal *h;
+	struct sigpending new_pending;
+	struct sigpending *pending;
 	struct rlimit rlim;
 	int i, ret;
 
@@ -215,6 +431,17 @@ static int restore_signal(struct ckpt_ctx *ctx)
 		if (ret < 0)
 			break;
 	}
+
+	ret = restore_sigpending(ctx, &new_pending);
+	if (ret < 0)
+		goto out;
+
+	spin_lock_irq(&current->sighand->siglock);
+	pending = &current->signal->shared_pending;
+	flush_sigqueue(pending);
+	pending->signal = new_pending.signal;
+	list_splice_init(&new_pending.list, &pending->list);
+	spin_unlock_irq(&current->sighand->siglock);
  out:
 	ckpt_hdr_put(ctx, h);
 	return ret;
@@ -250,8 +477,34 @@ int restore_obj_signal(struct ckpt_ctx *ctx, int signal_objref)
 int checkpoint_task_signal(struct ckpt_ctx *ctx, struct task_struct *t)
 {
 	struct ckpt_hdr_signal_task *h;
+	struct sigpending pending;
+	unsigned long flags;
 	int ret;
 
+	INIT_LIST_HEAD(&pending.list);
+
+	/* temporarily borrow signal queue - see chekcpoint_sigpending() */
+	if (!lock_task_sighand(t, &flags)) {
+		pr_warning("c/r: [%d] without sighand\n", task_pid_vnr(t));
+		return -EAGAIN;
+	}
+	list_splice_init(&t->pending.list, &pending.list);
+	pending.signal = t->pending.signal;
+	unlock_task_sighand(t, &flags);
+
+	ret = checkpoint_sigpending(ctx, &pending);
+
+	/* re-attach the borrowed queue */
+	if (!lock_task_sighand(t, &flags)) {
+		pr_warning("c/r: [%d] sighand disappeared\n", task_pid_vnr(t));
+		return -EAGAIN;
+	}
+	list_splice(&pending.list, &t->pending.list);
+	unlock_task_sighand(t, &flags);
+
+	if (ret < 0)
+		return ret;
+
 	h = ckpt_hdr_get_type(ctx, sizeof(*h), CKPT_HDR_SIGNAL_TASK);
 	if (!h)
 		return -ENOMEM;
@@ -266,7 +519,21 @@ int checkpoint_task_signal(struct ckpt_ctx *ctx, struct task_struct *t)
 int restore_task_signal(struct ckpt_ctx *ctx)
 {
 	struct ckpt_hdr_signal_task *h;
+	struct sigpending new_pending;
+	struct sigpending *pending;
 	sigset_t blocked;
+	int ret;
+
+	ret = restore_sigpending(ctx, &new_pending);
+	if (ret < 0)
+		return ret;
+
+	spin_lock_irq(&current->sighand->siglock);
+	pending = &current->pending;
+	flush_sigqueue(pending);
+	pending->signal = new_pending.signal;
+	list_splice_init(&new_pending.list, &pending->list);
+	spin_unlock_irq(&current->sighand->siglock);
 
 	h = ckpt_read_obj_type(ctx, sizeof(*h), CKPT_HDR_SIGNAL_TASK);
 	if (IS_ERR(h))
diff --git a/include/linux/checkpoint_hdr.h b/include/linux/checkpoint_hdr.h
index 0fd83b0..9dcb160 100644
--- a/include/linux/checkpoint_hdr.h
+++ b/include/linux/checkpoint_hdr.h
@@ -89,6 +89,7 @@ enum {
 	CKPT_HDR_SIGHAND = 601,
 	CKPT_HDR_SIGNAL,
 	CKPT_HDR_SIGNAL_TASK,
+	CKPT_HDR_SIGPENDING,
 
 	CKPT_HDR_TAIL = 9001,
 
@@ -432,6 +433,28 @@ struct ckpt_hdr_sighand {
 	struct ckpt_hdr_sigaction action[0];
 } __attribute__((aligned(8)));
 
+#ifndef HAVE_ARCH_SIGINFO_T
+struct ckpt_hdr_siginfo {
+	__u32 signo;
+	__u32 _errno;
+	__u32 code;
+
+	__u32 pid;
+	__s32 uid;
+	__u32 sigval_int;
+	__u64 sigval_ptr;
+	__u64 utime;
+	__u64 stime;
+} __attribute__((aligned(8)));
+#endif
+
+struct ckpt_hdr_sigpending {
+	struct ckpt_hdr h;
+	__u32 nr_pending;
+	struct ckpt_hdr_sigset signal;
+	struct ckpt_hdr_siginfo siginfo[0];
+} __attribute__((aligned(8)));
+
 struct ckpt_hdr_rlimit {
 	__u64 rlim_cur;
 	__u64 rlim_max;
-- 
1.6.0.4

[PATCH 4/4] c/r: support for real/virt/prof itimers (v2)

[Oren Laadan]

This patch adds support for real/virt/prof itimers.
Expiry and the interval values are both saved in nanoseconds.

Signed-off-by: Oren Laadan <orenl-eQaUEPhvms7ENvBUuze7eA@public.gmane.org>
---
 checkpoint/signal.c            |   72 ++++++++++++++++++++++++++++++++++++++++
 include/linux/checkpoint_hdr.h |    6 +++
 2 files changed, 78 insertions(+), 0 deletions(-)

diff --git a/checkpoint/signal.c b/checkpoint/signal.c
index c9da06b..18b3e2d 100644
--- a/checkpoint/signal.c
+++ b/checkpoint/signal.c
@@ -15,6 +15,7 @@
 #include <linux/signal.h>
 #include <linux/errno.h>
 #include <linux/resource.h>
+#include <linux/timer.h>
 #include <linux/checkpoint.h>
 #include <linux/checkpoint_hdr.h>
 
@@ -264,6 +265,7 @@ static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
 	struct signal_struct *signal;
 	struct sigpending shared_pending;
 	struct rlimit *rlim;
+	struct timeval tval;
 	unsigned long flags;
 	int i, ret;
 
@@ -299,6 +301,49 @@ static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
 		h->rlim[i].rlim_cur = rlim[i].rlim_cur;
 		h->rlim[i].rlim_max = rlim[i].rlim_max;
 	}
+
+	/* real/virt/prof itimers */
+	h->it_real_incr = ktime_to_ns(signal->it_real_incr);
+	/*
+	 * (a) If the timer is now active (did not expire), compute
+	 * the time delta.
+	 * (b) If the timer expired, and delivered the signal already,
+	 * then set the expire (value) to the interval
+	 * (c) If the timer expired but did not yet deliver the
+	 * signal, then set the expire to the minimum possible, so it
+	 * will expire immediately after restart succeeds.
+	 */
+	if (hrtimer_active(&signal->real_timer)) {
+		ktime_t delta = hrtimer_get_remaining(&signal->real_timer);
+		/*
+		 * If the timer expired after the the test above, then
+		 * set the expire to the minimum possible (because by
+		 * now the pending signals have been saved already, but
+		 * without the signal from this very expiry).
+		 */
+		ckpt_debug("active ! %lld\n", delta.tv64);
+		if (delta.tv64 <= 0)
+			delta.tv64 = NSEC_PER_USEC;
+		h->it_real_value = ktime_to_ns(delta);
+	} else if (sigismember(&signal->shared_pending.signal, SIGALRM)) {
+		/* case (b) */
+		h->it_real_value = h->it_real_incr;
+	} else {
+		/* case (c) */
+		h->it_real_value =
+			ktime_to_ns((ktime_t) { .tv64 = NSEC_PER_USEC });
+	}
+
+	cputime_to_timeval(signal->it_virt_expires, &tval);
+	h->it_virt_value = timeval_to_ns(&tval);
+	cputime_to_timeval(signal->it_virt_incr, &tval);
+	h->it_virt_incr = timeval_to_ns(&tval);
+
+	cputime_to_timeval(signal->it_prof_expires, &tval);
+	h->it_prof_value = timeval_to_ns(&tval);
+	cputime_to_timeval(signal->it_prof_incr, &tval);
+	h->it_prof_incr = timeval_to_ns(&tval);
+
 	unlock_task_sighand(t, &flags);
 
 	ret = ckpt_write_obj(ctx, &h->h);
@@ -416,6 +461,7 @@ static int restore_signal(struct ckpt_ctx *ctx)
 	struct ckpt_hdr_signal *h;
 	struct sigpending new_pending;
 	struct sigpending *pending;
+	struct itimerval itimer;
 	struct rlimit rlim;
 	int i, ret;
 
@@ -436,12 +482,38 @@ static int restore_signal(struct ckpt_ctx *ctx)
 	if (ret < 0)
 		goto out;
 
+	/*
+	 * Reset real/virt/prof itimer (in case they were set), to
+	 * prevent unwanted signals after flushing current signals
+	 * and before restoring original real/virt/prof itimer.
+	 */
+	itimer.it_value = (struct timeval) { .tv_sec = 0, .tv_usec = 0 };
+	itimer.it_interval =  (struct timeval) { .tv_sec = 0, .tv_usec = 0 };
+	do_setitimer(ITIMER_REAL, &itimer, NULL);
+	do_setitimer(ITIMER_VIRTUAL, &itimer, NULL);
+	do_setitimer(ITIMER_PROF, &itimer, NULL);
+
 	spin_lock_irq(&current->sighand->siglock);
 	pending = &current->signal->shared_pending;
 	flush_sigqueue(pending);
 	pending->signal = new_pending.signal;
 	list_splice_init(&new_pending.list, &pending->list);
 	spin_unlock_irq(&current->sighand->siglock);
+
+	/* real/virt/prof itimers */
+	itimer.it_value = ns_to_timeval(h->it_real_value);
+	itimer.it_interval = ns_to_timeval(h->it_real_incr);
+	ret = do_setitimer(ITIMER_REAL, &itimer, NULL);
+	if (ret < 0)
+		goto out;
+	itimer.it_value = ns_to_timeval(h->it_virt_value);
+	itimer.it_interval = ns_to_timeval(h->it_virt_incr);
+	ret = do_setitimer(ITIMER_VIRTUAL, &itimer, NULL);
+	if (ret < 0)
+		goto out;
+	itimer.it_value = ns_to_timeval(h->it_prof_value);
+	itimer.it_interval = ns_to_timeval(h->it_prof_incr);
+	ret = do_setitimer(ITIMER_PROF, &itimer, NULL);
  out:
 	ckpt_hdr_put(ctx, h);
 	return ret;
diff --git a/include/linux/checkpoint_hdr.h b/include/linux/checkpoint_hdr.h
index 9dcb160..a0ec25f 100644
--- a/include/linux/checkpoint_hdr.h
+++ b/include/linux/checkpoint_hdr.h
@@ -472,6 +472,12 @@ struct ckpt_hdr_rlimit {
 struct ckpt_hdr_signal {
 	struct ckpt_hdr h;
 	struct ckpt_hdr_rlimit rlim[CKPT_RLIM_NLIMITS];
+	__u64 it_real_value;
+	__u64 it_real_incr;
+	__u64 it_virt_value;
+	__u64 it_virt_incr;
+	__u64 it_prof_value;
+	__u64 it_prof_incr;
 } __attribute__((aligned(8)));
 
 struct ckpt_hdr_signal_task {
-- 
1.6.0.4

Re: [PATCH 4/4] c/r: support for real/virt/prof itimers (v2)

[Louis Rilling]

[-- Attachment #1.1: Type: text/plain, Size: 7367 bytes --]

Hi Oren,

On 04/08/09  5:09 -0400, Oren Laadan wrote:
> This patch adds support for real/virt/prof itimers.
> Expiry and the interval values are both saved in nanoseconds.
> 
> Signed-off-by: Oren Laadan <orenl-eQaUEPhvms7ENvBUuze7eA@public.gmane.org>
> ---
>  checkpoint/signal.c            |   72 ++++++++++++++++++++++++++++++++++++++++
>  include/linux/checkpoint_hdr.h |    6 +++
>  2 files changed, 78 insertions(+), 0 deletions(-)
> 
> diff --git a/checkpoint/signal.c b/checkpoint/signal.c
> index c9da06b..18b3e2d 100644
> --- a/checkpoint/signal.c
> +++ b/checkpoint/signal.c
> @@ -15,6 +15,7 @@
>  #include <linux/signal.h>
>  #include <linux/errno.h>
>  #include <linux/resource.h>
> +#include <linux/timer.h>
>  #include <linux/checkpoint.h>
>  #include <linux/checkpoint_hdr.h>
>  
> @@ -264,6 +265,7 @@ static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
>  	struct signal_struct *signal;
>  	struct sigpending shared_pending;
>  	struct rlimit *rlim;
> +	struct timeval tval;
>  	unsigned long flags;
>  	int i, ret;
>  
> @@ -299,6 +301,49 @@ static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
>  		h->rlim[i].rlim_cur = rlim[i].rlim_cur;
>  		h->rlim[i].rlim_max = rlim[i].rlim_max;
>  	}
> +
> +	/* real/virt/prof itimers */
> +	h->it_real_incr = ktime_to_ns(signal->it_real_incr);
> +	/*
> +	 * (a) If the timer is now active (did not expire), compute
> +	 * the time delta.
> +	 * (b) If the timer expired, and delivered the signal already,
> +	 * then set the expire (value) to the interval
> +	 * (c) If the timer expired but did not yet deliver the
> +	 * signal, then set the expire to the minimum possible, so it
> +	 * will expire immediately after restart succeeds.
> +	 */
> +	if (hrtimer_active(&signal->real_timer)) {
> +		ktime_t delta = hrtimer_get_remaining(&signal->real_timer);
> +		/*
> +		 * If the timer expired after the the test above, then
> +		 * set the expire to the minimum possible (because by
> +		 * now the pending signals have been saved already, but
> +		 * without the signal from this very expiry).
> +		 */
> +		ckpt_debug("active ! %lld\n", delta.tv64);
> +		if (delta.tv64 <= 0)
> +			delta.tv64 = NSEC_PER_USEC;
> +		h->it_real_value = ktime_to_ns(delta);
> +	} else if (sigismember(&signal->shared_pending.signal, SIGALRM)) {
> +		/* case (b) */
> +		h->it_real_value = h->it_real_incr;
> +	} else {
> +		/* case (c) */
> +		h->it_real_value =
> +			ktime_to_ns((ktime_t) { .tv64 = NSEC_PER_USEC });
> +	}

AFAICS the third case catches expired timers having not delivered the signal yet
as well as inactive timers. If the latter, at restart the timer will be
rearmed while it should not.

However, reading hrtimer code I understand that case (c) cannot happen. IOW if
the timer has expired but has not sent SIGALRM yet, then hrtimer_active()
returns true, and the first block already handles that case correctly.

Moreover case (b) only happens if the timer was not automatically rearmed,
that is it_real_incr == 0. So in this case the signal was already sent (and
picked for checkpointing if not handled), and h->it_real_value should be set to
0.

So the following could be actually better:

	h->it_real_incr = ktime_to_ns(signal->it_real_incr);
	/*
	 * (a) If the timer is now active, compute the time delta.
	 * (b) Otherwise, siglock ensures that 	 */
	if (hrtimer_active(&signal->real_timer)) {
		ktime_t delta = hrtimer_get_remaining(&signal->real_timer);
		/*
		 * If the timer expired after the the test above, then
		 * set the expire to the minimum possible (because by
		 * now the pending signals have been saved already, but
		 * the signal from this very expiry won't be sent before we
		 * release t->sighand->siglock.
		 */
		ckpt_debug("active ! %lld\n", delta.tv64);
		if (delta.tv64 <= 0)
			delta.tv64 = NSEC_PER_USEC;
		h->it_real_value = ktime_to_ns(delta);
	} else {
		BUG_ON(h->it_real_incr);
		h->it_real_value = 0;
	}

Louis

> +
> +	cputime_to_timeval(signal->it_virt_expires, &tval);
> +	h->it_virt_value = timeval_to_ns(&tval);
> +	cputime_to_timeval(signal->it_virt_incr, &tval);
> +	h->it_virt_incr = timeval_to_ns(&tval);
> +
> +	cputime_to_timeval(signal->it_prof_expires, &tval);
> +	h->it_prof_value = timeval_to_ns(&tval);
> +	cputime_to_timeval(signal->it_prof_incr, &tval);
> +	h->it_prof_incr = timeval_to_ns(&tval);
> +
>  	unlock_task_sighand(t, &flags);
>  
>  	ret = ckpt_write_obj(ctx, &h->h);
> @@ -416,6 +461,7 @@ static int restore_signal(struct ckpt_ctx *ctx)
>  	struct ckpt_hdr_signal *h;
>  	struct sigpending new_pending;
>  	struct sigpending *pending;
> +	struct itimerval itimer;
>  	struct rlimit rlim;
>  	int i, ret;
>  
> @@ -436,12 +482,38 @@ static int restore_signal(struct ckpt_ctx *ctx)
>  	if (ret < 0)
>  		goto out;
>  
> +	/*
> +	 * Reset real/virt/prof itimer (in case they were set), to
> +	 * prevent unwanted signals after flushing current signals
> +	 * and before restoring original real/virt/prof itimer.
> +	 */
> +	itimer.it_value = (struct timeval) { .tv_sec = 0, .tv_usec = 0 };
> +	itimer.it_interval =  (struct timeval) { .tv_sec = 0, .tv_usec = 0 };
> +	do_setitimer(ITIMER_REAL, &itimer, NULL);
> +	do_setitimer(ITIMER_VIRTUAL, &itimer, NULL);
> +	do_setitimer(ITIMER_PROF, &itimer, NULL);
> +
>  	spin_lock_irq(&current->sighand->siglock);
>  	pending = &current->signal->shared_pending;
>  	flush_sigqueue(pending);
>  	pending->signal = new_pending.signal;
>  	list_splice_init(&new_pending.list, &pending->list);
>  	spin_unlock_irq(&current->sighand->siglock);
> +
> +	/* real/virt/prof itimers */
> +	itimer.it_value = ns_to_timeval(h->it_real_value);
> +	itimer.it_interval = ns_to_timeval(h->it_real_incr);
> +	ret = do_setitimer(ITIMER_REAL, &itimer, NULL);
> +	if (ret < 0)
> +		goto out;
> +	itimer.it_value = ns_to_timeval(h->it_virt_value);
> +	itimer.it_interval = ns_to_timeval(h->it_virt_incr);
> +	ret = do_setitimer(ITIMER_VIRTUAL, &itimer, NULL);
> +	if (ret < 0)
> +		goto out;
> +	itimer.it_value = ns_to_timeval(h->it_prof_value);
> +	itimer.it_interval = ns_to_timeval(h->it_prof_incr);
> +	ret = do_setitimer(ITIMER_PROF, &itimer, NULL);
>   out:
>  	ckpt_hdr_put(ctx, h);
>  	return ret;
> diff --git a/include/linux/checkpoint_hdr.h b/include/linux/checkpoint_hdr.h
> index 9dcb160..a0ec25f 100644
> --- a/include/linux/checkpoint_hdr.h
> +++ b/include/linux/checkpoint_hdr.h
> @@ -472,6 +472,12 @@ struct ckpt_hdr_rlimit {
>  struct ckpt_hdr_signal {
>  	struct ckpt_hdr h;
>  	struct ckpt_hdr_rlimit rlim[CKPT_RLIM_NLIMITS];
> +	__u64 it_real_value;
> +	__u64 it_real_incr;
> +	__u64 it_virt_value;
> +	__u64 it_virt_incr;
> +	__u64 it_prof_value;
> +	__u64 it_prof_incr;
>  } __attribute__((aligned(8)));
>  
>  struct ckpt_hdr_signal_task {
> -- 
> 1.6.0.4
> 
> _______________________________________________
> Containers mailing list
> Containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org
> https://lists.linux-foundation.org/mailman/listinfo/containers

-- 
Dr Louis Rilling			Kerlabs
Skype: louis.rilling			Batiment Germanium
Phone: (+33|0) 6 80 89 08 23		80 avenue des Buttes de Coesmes
http://www.kerlabs.com/			35700 Rennes

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

[-- Attachment #2: Type: text/plain, Size: 206 bytes --]

_______________________________________________
Containers mailing list
Containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org
https://lists.linux-foundation.org/mailman/listinfo/containers

Re: [PATCH 4/4] c/r: support for real/virt/prof itimers (v2)

[Oren Laadan]

Louis Rilling wrote:
> Hi Oren,
> 
> On 04/08/09  5:09 -0400, Oren Laadan wrote:
>> This patch adds support for real/virt/prof itimers.
>> Expiry and the interval values are both saved in nanoseconds.
>>
>> Signed-off-by: Oren Laadan <orenl-eQaUEPhvms7ENvBUuze7eA@public.gmane.org>
>> ---
>>  checkpoint/signal.c            |   72 ++++++++++++++++++++++++++++++++++++++++
>>  include/linux/checkpoint_hdr.h |    6 +++
>>  2 files changed, 78 insertions(+), 0 deletions(-)
>>
>> diff --git a/checkpoint/signal.c b/checkpoint/signal.c
>> index c9da06b..18b3e2d 100644
>> --- a/checkpoint/signal.c
>> +++ b/checkpoint/signal.c
>> @@ -15,6 +15,7 @@
>>  #include <linux/signal.h>
>>  #include <linux/errno.h>
>>  #include <linux/resource.h>
>> +#include <linux/timer.h>
>>  #include <linux/checkpoint.h>
>>  #include <linux/checkpoint_hdr.h>
>>  
>> @@ -264,6 +265,7 @@ static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
>>  	struct signal_struct *signal;
>>  	struct sigpending shared_pending;
>>  	struct rlimit *rlim;
>> +	struct timeval tval;
>>  	unsigned long flags;
>>  	int i, ret;
>>  
>> @@ -299,6 +301,49 @@ static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
>>  		h->rlim[i].rlim_cur = rlim[i].rlim_cur;
>>  		h->rlim[i].rlim_max = rlim[i].rlim_max;
>>  	}
>> +
>> +	/* real/virt/prof itimers */
>> +	h->it_real_incr = ktime_to_ns(signal->it_real_incr);
>> +	/*
>> +	 * (a) If the timer is now active (did not expire), compute
>> +	 * the time delta.
>> +	 * (b) If the timer expired, and delivered the signal already,
>> +	 * then set the expire (value) to the interval
>> +	 * (c) If the timer expired but did not yet deliver the
>> +	 * signal, then set the expire to the minimum possible, so it
>> +	 * will expire immediately after restart succeeds.
>> +	 */
>> +	if (hrtimer_active(&signal->real_timer)) {
>> +		ktime_t delta = hrtimer_get_remaining(&signal->real_timer);
>> +		/*
>> +		 * If the timer expired after the the test above, then
>> +		 * set the expire to the minimum possible (because by
>> +		 * now the pending signals have been saved already, but
>> +		 * without the signal from this very expiry).
>> +		 */
>> +		ckpt_debug("active ! %lld\n", delta.tv64);
>> +		if (delta.tv64 <= 0)
>> +			delta.tv64 = NSEC_PER_USEC;
>> +		h->it_real_value = ktime_to_ns(delta);
>> +	} else if (sigismember(&signal->shared_pending.signal, SIGALRM)) {
>> +		/* case (b) */
>> +		h->it_real_value = h->it_real_incr;
>> +	} else {
>> +		/* case (c) */
>> +		h->it_real_value =
>> +			ktime_to_ns((ktime_t) { .tv64 = NSEC_PER_USEC });
>> +	}
> 
> AFAICS the third case catches expired timers having not delivered the signal yet
> as well as inactive timers. If the latter, at restart the timer will be
> rearmed while it should not.
> 
> However, reading hrtimer code I understand that case (c) cannot happen. IOW if
> the timer has expired but has not sent SIGALRM yet, then hrtimer_active()
> returns true, and the first block already handles that case correctly.

Yes, I re-read the code and (c) should not happen, because the timer
cannot become inactive without having sent SIGALRM as long as we hold
the siglock.

> Moreover case (b) only happens if the timer was not automatically rearmed,
> that is it_real_incr == 0. So in this case the signal was already sent (and
> picked for checkpointing if not handled), and h->it_real_value should be set to
> 0.

Here is the scenario I had in mind:

1) kernel/hrtimer.c:__run_hrtimer() is called with the timer fires,
sets the timer state to HRTIMER_STATE_CALLBACK and then invokes
the callback.

2) for itimers, the callback is kernel/itimer.c:it_real_fn(): it
sends the signal and returns HRTIMER_NORESTART.

3) back to __run_hrtimer() which does _not_ call enqueue_hrtimer(),
and then _clears_ the HRTIMER_STATE_CALLBACK. This results in the
state being zero -- that is, hrtimer_active() will be false.

So in #2 SIGALRM was sent, and then in #3 the timer became inactive
without holding siglock, and not re-armed. (if it_real_incr > 0, it
should and will be re-armed when the signal is dequeued).

Therefore when we restart, we need to set it_real_incr to its saved
value, and it_real_value to 0. Then, if SIGALRM was sent, it will
be dequeued after restart, and depending on it_real_incr the timer
will, or will not, be re-armed. If SIGALRM was not sent, then it
must be that it_real_incr is also 0.

IOW, in both cases (SIGALRM sent/not sent) of the inactive timer,
the value of it_real_incr should be 0, therefore the code you
suggest below is correct, except for the BUG_ON statement.

(In restart, it's not a big deal if these values are "incorrect",
because the worst harm it can do is send a superfluous SIGALRM).

> 
> So the following could be actually better:
> 
> 	h->it_real_incr = ktime_to_ns(signal->it_real_incr);
> 	/*
> 	 * (a) If the timer is now active, compute the time delta.
> 	 * (b) Otherwise, siglock ensures that 	 */
> 	if (hrtimer_active(&signal->real_timer)) {
> 		ktime_t delta = hrtimer_get_remaining(&signal->real_timer);
> 		/*
> 		 * If the timer expired after the the test above, then
> 		 * set the expire to the minimum possible (because by
> 		 * now the pending signals have been saved already, but
> 		 * the signal from this very expiry won't be sent before we
> 		 * release t->sighand->siglock.
> 		 */
> 		ckpt_debug("active ! %lld\n", delta.tv64);
> 		if (delta.tv64 <= 0)
> 			delta.tv64 = NSEC_PER_USEC;
> 		h->it_real_value = ktime_to_ns(delta);
> 	} else {
> 		BUG_ON(h->it_real_incr);
> 		h->it_real_value = 0;
> 	}
> 
> Louis

Thanks,

Oren.

Re: [PATCH 4/4] c/r: support for real/virt/prof itimers (v2)

[Louis Rilling]

[-- Attachment #1.1: Type: text/plain, Size: 6933 bytes --]

On 04/08/09 14:43 -0400, Oren Laadan wrote:
> 
> 
> Louis Rilling wrote:
> > Hi Oren,
> > 
> > On 04/08/09  5:09 -0400, Oren Laadan wrote:
> >> This patch adds support for real/virt/prof itimers.
> >> Expiry and the interval values are both saved in nanoseconds.
> >>
> >> Signed-off-by: Oren Laadan <orenl-eQaUEPhvms7ENvBUuze7eA@public.gmane.org>
> >> ---
> >>  checkpoint/signal.c            |   72 ++++++++++++++++++++++++++++++++++++++++
> >>  include/linux/checkpoint_hdr.h |    6 +++
> >>  2 files changed, 78 insertions(+), 0 deletions(-)
> >>
> >> diff --git a/checkpoint/signal.c b/checkpoint/signal.c
> >> index c9da06b..18b3e2d 100644
> >> --- a/checkpoint/signal.c
> >> +++ b/checkpoint/signal.c
> >> @@ -15,6 +15,7 @@
> >>  #include <linux/signal.h>
> >>  #include <linux/errno.h>
> >>  #include <linux/resource.h>
> >> +#include <linux/timer.h>
> >>  #include <linux/checkpoint.h>
> >>  #include <linux/checkpoint_hdr.h>
> >>  
> >> @@ -264,6 +265,7 @@ static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
> >>  	struct signal_struct *signal;
> >>  	struct sigpending shared_pending;
> >>  	struct rlimit *rlim;
> >> +	struct timeval tval;
> >>  	unsigned long flags;
> >>  	int i, ret;
> >>  
> >> @@ -299,6 +301,49 @@ static int checkpoint_signal(struct ckpt_ctx *ctx, struct task_struct *t)
> >>  		h->rlim[i].rlim_cur = rlim[i].rlim_cur;
> >>  		h->rlim[i].rlim_max = rlim[i].rlim_max;
> >>  	}
> >> +
> >> +	/* real/virt/prof itimers */
> >> +	h->it_real_incr = ktime_to_ns(signal->it_real_incr);
> >> +	/*
> >> +	 * (a) If the timer is now active (did not expire), compute
> >> +	 * the time delta.
> >> +	 * (b) If the timer expired, and delivered the signal already,
> >> +	 * then set the expire (value) to the interval
> >> +	 * (c) If the timer expired but did not yet deliver the
> >> +	 * signal, then set the expire to the minimum possible, so it
> >> +	 * will expire immediately after restart succeeds.
> >> +	 */
> >> +	if (hrtimer_active(&signal->real_timer)) {
> >> +		ktime_t delta = hrtimer_get_remaining(&signal->real_timer);
> >> +		/*
> >> +		 * If the timer expired after the the test above, then
> >> +		 * set the expire to the minimum possible (because by
> >> +		 * now the pending signals have been saved already, but
> >> +		 * without the signal from this very expiry).
> >> +		 */
> >> +		ckpt_debug("active ! %lld\n", delta.tv64);
> >> +		if (delta.tv64 <= 0)
> >> +			delta.tv64 = NSEC_PER_USEC;
> >> +		h->it_real_value = ktime_to_ns(delta);
> >> +	} else if (sigismember(&signal->shared_pending.signal, SIGALRM)) {
> >> +		/* case (b) */
> >> +		h->it_real_value = h->it_real_incr;
> >> +	} else {
> >> +		/* case (c) */
> >> +		h->it_real_value =
> >> +			ktime_to_ns((ktime_t) { .tv64 = NSEC_PER_USEC });
> >> +	}
> > 
> > AFAICS the third case catches expired timers having not delivered the signal yet
> > as well as inactive timers. If the latter, at restart the timer will be
> > rearmed while it should not.
> > 
> > However, reading hrtimer code I understand that case (c) cannot happen. IOW if
> > the timer has expired but has not sent SIGALRM yet, then hrtimer_active()
> > returns true, and the first block already handles that case correctly.
> 
> Yes, I re-read the code and (c) should not happen, because the timer
> cannot become inactive without having sent SIGALRM as long as we hold
> the siglock.
> 
> > Moreover case (b) only happens if the timer was not automatically rearmed,
> > that is it_real_incr == 0. So in this case the signal was already sent (and
> > picked for checkpointing if not handled), and h->it_real_value should be set to
> > 0.
> 
> Here is the scenario I had in mind:
> 
> 1) kernel/hrtimer.c:__run_hrtimer() is called with the timer fires,
> sets the timer state to HRTIMER_STATE_CALLBACK and then invokes
> the callback.
> 
> 2) for itimers, the callback is kernel/itimer.c:it_real_fn(): it
> sends the signal and returns HRTIMER_NORESTART.
> 
> 3) back to __run_hrtimer() which does _not_ call enqueue_hrtimer(),
> and then _clears_ the HRTIMER_STATE_CALLBACK. This results in the
> state being zero -- that is, hrtimer_active() will be false.
> 
> So in #2 SIGALRM was sent, and then in #3 the timer became inactive
> without holding siglock, and not re-armed. (if it_real_incr > 0, it
> should and will be re-armed when the signal is dequeued).
> 
> Therefore when we restart, we need to set it_real_incr to its saved
> value, and it_real_value to 0. Then, if SIGALRM was sent, it will
> be dequeued after restart, and depending on it_real_incr the timer
> will, or will not, be re-armed. If SIGALRM was not sent, then it
> must be that it_real_incr is also 0.
> 
> IOW, in both cases (SIGALRM sent/not sent) of the inactive timer,
> the value of it_real_incr should be 0, therefore the code you

I guess that you meant it_real_value, not it_real_incr.

> suggest below is correct, except for the BUG_ON statement.

You're right. I missed the fact that the timer was only rearmed when dequeuing
SIGALRM, which is a bit counter-intuitive since blocking SIGALRM then introduces
a drift in timer events.

> 
> (In restart, it's not a big deal if these values are "incorrect",
> because the worst harm it can do is send a superfluous SIGALRM).

I'm not sure that superfuous SIGALRM are harmless, especially if the process
just deactivated all timers. However, IIUC there will be none so it's ok.

Thanks,

Louis

> 
> > 
> > So the following could be actually better:
> > 
> > 	h->it_real_incr = ktime_to_ns(signal->it_real_incr);
> > 	/*
> > 	 * (a) If the timer is now active, compute the time delta.
> > 	 * (b) Otherwise, siglock ensures that 	 */
> > 	if (hrtimer_active(&signal->real_timer)) {
> > 		ktime_t delta = hrtimer_get_remaining(&signal->real_timer);
> > 		/*
> > 		 * If the timer expired after the the test above, then
> > 		 * set the expire to the minimum possible (because by
> > 		 * now the pending signals have been saved already, but
> > 		 * the signal from this very expiry won't be sent before we
> > 		 * release t->sighand->siglock.
> > 		 */
> > 		ckpt_debug("active ! %lld\n", delta.tv64);
> > 		if (delta.tv64 <= 0)
> > 			delta.tv64 = NSEC_PER_USEC;
> > 		h->it_real_value = ktime_to_ns(delta);
> > 	} else {
> > 		BUG_ON(h->it_real_incr);
> > 		h->it_real_value = 0;
> > 	}
> > 
> > Louis
> 
> Thanks,
> 
> Oren.
> 
> _______________________________________________
> Containers mailing list
> Containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org
> https://lists.linux-foundation.org/mailman/listinfo/containers

-- 
Dr Louis Rilling			Kerlabs
Skype: louis.rilling			Batiment Germanium
Phone: (+33|0) 6 80 89 08 23		80 avenue des Buttes de Coesmes
http://www.kerlabs.com/			35700 Rennes

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

[-- Attachment #2: Type: text/plain, Size: 206 bytes --]

_______________________________________________
Containers mailing list
Containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org
https://lists.linux-foundation.org/mailman/listinfo/containers