From: Kei Tokunaga <ktokunag@redhat.com>
Date: Mon, 28 Apr 2008 16:32:09 -0400
Subject: [misc] add CPU hotplug support for relay functions
Message-id: 48163449.8070009@redhat.com
O-Subject: [RHEL5.3][PATCH] add CPU hotplug support for relay functions
Bugzilla: 441523

bz441523
https://bugzilla.redhat.com/show_bug.cgi?id=441523

Description
===========
  The message buffers for relay function are only allocated for online CPUs
  at the time when relay_open() is called.  So, if any offline CPUs become
  online using hotplug after relay_open() gets called, relay function starts
  trying to access to the message buffers of newly onlined CPUs and the kernel
  panics.

kABI status
===========
  There are some additions to structure rchan, which cause kABI changes.  So
  they are wrapped with #ifndef __GENKSYMS__.

Brew status
===========
  Built on all platforms.
  http://brewweb.devel.redhat.com/brew/taskinfo?taskID=1295273

Upstream status
===============
  Upstream has a different way of this as it has added one more argument to
  relay_open().  Since that breaks kABI, I removed the change from the backported
  patch for RHEL.

  http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=23c887522e912ca494950796a95df8dd210f4b01
  http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=c9b3febc5b9c55a76b838c977b078195ec8bb95e

Test status
===========
  Confirmed this issue doesn't occur on 2.6.18-90.el5 with the patch applied on
  Fujitsu PRIMEQUEST.  What I did is to offline all CPUs, except one, to start
  running relay function using blktrace/blkparse commands, and to online the
  offlined CPUs.

Thanks,
Kei

Acked-by: Prarit Bhargava <prarit@redhat.com>

diff --git a/include/linux/relay.h b/include/linux/relay.h
index 24accb4..05bf63b 100644
--- a/include/linux/relay.h
+++ b/include/linux/relay.h
@@ -64,6 +64,12 @@ struct rchan
 	void *private_data;		/* for user-defined data */
 	size_t last_toobig;		/* tried to log event > subbuf size */
 	struct rchan_buf *buf[NR_CPUS]; /* per-cpu channel buffers */
+#ifndef __GENKSYMS__
+	int is_global;			/* One global buffer ? */
+	struct list_head list;		/* for channel list */
+	struct dentry *parent;		/* parent dentry passed to open */
+	char base_filename[NAME_MAX];	/* saved base filename */
+#endif
 };
 
 /*
diff --git a/kernel/relay.c b/kernel/relay.c
index 0ad4ec3..d82d6da 100644
--- a/kernel/relay.c
+++ b/kernel/relay.c
@@ -1,12 +1,14 @@
 /*
  * Public API and common code for kernel->userspace relay file support.
  *
- * See Documentation/filesystems/relayfs.txt for an overview of relayfs.
+ * See Documentation/filesystems/relay.txt for an overview.
  *
  * Copyright (C) 2002-2005 - Tom Zanussi (zanussi@us.ibm.com), IBM Corp
  * Copyright (C) 1999-2005 - Karim Yaghmour (karim@opersys.com)
  *
  * Moved to kernel/relay.c by Paul Mundt, 2006.
+ * November 2006 - CPU hotplug support by Mathieu Desnoyers
+ * 	(mathieu.desnoyers@polymtl.ca)
  *
  * This file is released under the GPL.
  */
@@ -18,6 +20,11 @@
 #include <linux/relay.h>
 #include <linux/vmalloc.h>
 #include <linux/mm.h>
+#include <linux/cpu.h>
+
+/* list of open channels, for cpu hotplug */
+static DEFINE_MUTEX(relay_channels_mutex);
+static LIST_HEAD(relay_channels);
 
 /*
  * close() vm_op implementation for relay file mapping.
@@ -188,6 +195,7 @@ void relay_destroy_buf(struct rchan_buf *buf)
 			__free_page(buf->page_array[i]);
 		kfree(buf->page_array);
 	}
+	chan->buf[buf->cpu] = NULL;
 	kfree(buf->padding);
 	kfree(buf);
 	kref_put(&chan->kref, relay_destroy_channel);
@@ -361,51 +369,70 @@ static inline void __relay_reset(struct rchan_buf *buf, unsigned int init)
 void relay_reset(struct rchan *chan)
 {
 	unsigned int i;
-	struct rchan_buf *prev = NULL;
 
 	if (!chan)
 		return;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i] || chan->buf[i] == prev)
-			break;
-		__relay_reset(chan->buf[i], 0);
-		prev = chan->buf[i];
+	if (chan->is_global && chan->buf[0]) {
+		__relay_reset(chan->buf[0], 0);
+		return;
 	}
+
+	mutex_lock(&relay_channels_mutex);
+	for_each_online_cpu(i)
+		if (chan->buf[i])
+			__relay_reset(chan->buf[i], 0);
+	mutex_unlock(&relay_channels_mutex);
 }
 EXPORT_SYMBOL_GPL(relay_reset);
 
-/**
+/*
  *	relay_open_buf - create a new relay channel buffer
  *
- *	Internal - used by relay_open().
+ *	used by relay_open() and CPU hotplug.
  */
-static struct rchan_buf *relay_open_buf(struct rchan *chan,
-					const char *filename,
-					struct dentry *parent,
-					int *is_global)
+static struct rchan_buf *relay_open_buf(struct rchan *chan, unsigned int cpu)
 {
-	struct rchan_buf *buf;
+ 	struct rchan_buf *buf = NULL;
 	struct dentry *dentry;
+	char *tmpname;
 
-	if (*is_global)
+	if (chan->is_global)
 		return chan->buf[0];
 
+	tmpname = kzalloc(NAME_MAX + 1, GFP_KERNEL);
+	if (!tmpname)
+		goto end;
+	snprintf(tmpname, NAME_MAX, "%s%d", chan->base_filename, cpu);
+
 	buf = relay_create_buf(chan);
 	if (!buf)
-		return NULL;
+		goto free_name;
+
+	buf->cpu = cpu;
+	__relay_reset(buf, 1);
 
 	/* Create file in fs */
-	dentry = chan->cb->create_buf_file(filename, parent, S_IRUSR,
-					   buf, is_global);
-	if (!dentry) {
-		relay_destroy_buf(buf);
-		return NULL;
-	}
+	dentry = chan->cb->create_buf_file(tmpname, chan->parent, S_IRUSR,
+					   buf, &chan->is_global);
+	if (!dentry)
+		goto free_buf;
 
 	buf->dentry = dentry;
-	__relay_reset(buf, 1);
 
+	if(chan->is_global) {
+		chan->buf[0] = buf;
+		buf->cpu = 0;
+	}
+
+	goto free_name;
+
+free_buf:
+	relay_destroy_buf(buf);
+	buf = NULL;
+free_name:
+	kfree(tmpname);
+end:
 	return buf;
 }
 
@@ -447,6 +474,47 @@ static inline void setup_callbacks(struct rchan *chan,
 }
 
 /**
+ *
+ * 	relay_hotcpu_callback - CPU hotplug callback
+ * 	@nb: notifier block
+ * 	@action: hotplug action to take
+ * 	@hcpu: CPU number
+ *
+ * 	Returns the success/failure of the operation. (NOTIFY_OK, NOTIFY_BAD)
+ */
+static int __cpuinit relay_hotcpu_callback(struct notifier_block *nb,
+				unsigned long action,
+				void *hcpu)
+{
+	unsigned int hotcpu = (unsigned long)hcpu;
+	struct rchan *chan;
+
+	switch(action) {
+	case CPU_UP_PREPARE:
+		mutex_lock(&relay_channels_mutex);
+		list_for_each_entry(chan, &relay_channels, list) {
+			if (chan->buf[hotcpu])
+				continue;
+			chan->buf[hotcpu] = relay_open_buf(chan, hotcpu);
+			if(!chan->buf[hotcpu]) {
+				printk(KERN_ERR
+					"relay_hotcpu_callback: cpu %d buffer "
+					"creation failed\n", hotcpu);
+				mutex_unlock(&relay_channels_mutex);
+				return NOTIFY_BAD;
+			}
+		}
+		mutex_unlock(&relay_channels_mutex);
+		break;
+	case CPU_DEAD:
+		/* No need to flush the cpu : will be flushed upon
+		 * final relay_flush() call. */
+		break;
+	}
+	return NOTIFY_OK;
+}
+
+/**
  *	relay_open - create a new relay channel
  *	@base_filename: base name of files to create
  *	@parent: dentry of parent directory, NULL for root directory
@@ -469,9 +537,6 @@ struct rchan *relay_open(const char *base_filename,
 {
 	unsigned int i;
 	struct rchan *chan;
-	char *tmpname;
-	int is_global = 0;
-
 	if (!base_filename)
 		return NULL;
 
@@ -486,38 +551,31 @@ struct rchan *relay_open(const char *base_filename,
 	chan->n_subbufs = n_subbufs;
 	chan->subbuf_size = subbuf_size;
 	chan->alloc_size = FIX_SIZE(subbuf_size * n_subbufs);
+	chan->parent = parent;
+	strlcpy(chan->base_filename, base_filename, NAME_MAX);
 	setup_callbacks(chan, cb);
 	kref_init(&chan->kref);
 
-	tmpname = kmalloc(NAME_MAX + 1, GFP_KERNEL);
-	if (!tmpname)
-		goto free_chan;
-
+	mutex_lock(&relay_channels_mutex);
 	for_each_online_cpu(i) {
-		sprintf(tmpname, "%s%d", base_filename, i);
-		chan->buf[i] = relay_open_buf(chan, tmpname, parent,
-					      &is_global);
+		chan->buf[i] = relay_open_buf(chan, i);
 		if (!chan->buf[i])
 			goto free_bufs;
-
-		chan->buf[i]->cpu = i;
 	}
+	list_add(&chan->list, &relay_channels);
+	mutex_unlock(&relay_channels_mutex);
 
-	kfree(tmpname);
 	return chan;
 
 free_bufs:
-	for (i = 0; i < NR_CPUS; i++) {
+	for_each_online_cpu(i) {
 		if (!chan->buf[i])
 			break;
 		relay_close_buf(chan->buf[i]);
-		if (is_global)
-			break;
 	}
-	kfree(tmpname);
 
-free_chan:
 	kref_put(&chan->kref, relay_destroy_channel);
+	mutex_unlock(&relay_channels_mutex);
 	return NULL;
 }
 EXPORT_SYMBOL_GPL(relay_open);
@@ -617,24 +675,26 @@ EXPORT_SYMBOL_GPL(relay_subbufs_consumed);
 void relay_close(struct rchan *chan)
 {
 	unsigned int i;
-	struct rchan_buf *prev = NULL;
 
 	if (!chan)
 		return;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i] || chan->buf[i] == prev)
-			break;
-		relay_close_buf(chan->buf[i]);
-		prev = chan->buf[i];
-	}
+	mutex_lock(&relay_channels_mutex);
+	if (chan->is_global && chan->buf[0])
+		relay_close_buf(chan->buf[0]);
+	else
+		for_each_possible_cpu(i)
+			if (chan->buf[i])
+				relay_close_buf(chan->buf[i]);
 
 	if (chan->last_toobig)
 		printk(KERN_WARNING "relay: one or more items not logged "
 		       "[item size (%Zd) > sub-buffer size (%Zd)]\n",
 		       chan->last_toobig, chan->subbuf_size);
 
+	list_del(&chan->list);
 	kref_put(&chan->kref, relay_destroy_channel);
+	mutex_unlock(&relay_channels_mutex);
 }
 EXPORT_SYMBOL_GPL(relay_close);
 
@@ -647,17 +707,20 @@ EXPORT_SYMBOL_GPL(relay_close);
 void relay_flush(struct rchan *chan)
 {
 	unsigned int i;
-	struct rchan_buf *prev = NULL;
 
 	if (!chan)
 		return;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i] || chan->buf[i] == prev)
-			break;
-		relay_switch_subbuf(chan->buf[i], 0);
-		prev = chan->buf[i];
+	if (chan->is_global && chan->buf[0]) {
+		relay_switch_subbuf(chan->buf[0], 0);
+		return;
 	}
+
+	mutex_lock(&relay_channels_mutex);
+	for_each_possible_cpu(i)
+		if (chan->buf[i])
+			relay_switch_subbuf(chan->buf[i], 0);
+	mutex_unlock(&relay_channels_mutex);
 }
 EXPORT_SYMBOL_GPL(relay_flush);
 
@@ -1018,3 +1081,12 @@ struct file_operations relay_file_operations = {
 	.sendfile       = relay_file_sendfile,
 };
 EXPORT_SYMBOL_GPL(relay_file_operations);
+
+static __init int relay_init(void)
+{
+
+	hotcpu_notifier(relay_hotcpu_callback, 0);
+	return 0;
+}
+
+module_init(relay_init);