[PULL] module, param and stop_machine patches

[PULL] module, param and stop_machine patches

[Rusty Russell]

The following changes since commit 2515ddc6db8eb49a79f0fe5e67ff09ac7c81eab4:
  Paul Mundt (1):
        binfmt_elf_fdpic: Update for cputime changes.

are available in the git repository at:

  ssh://master.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-for-linus.git master

Andi Kleen (1):
      Remove stop_machine during module load v2

Heiko Carstens (4):
      Call init_workqueues before pre smp initcalls.
      workqueue: introduce create_rt_workqueue
      stop_machine: use workqueues instead of kernel threads
      stop_machine: fix error code handling on multiple cpus

Rusty Russell (6):
      module: simplify load_module.
      module: check kernel param length at compile time, not runtime
      param: Fix duplicate module prefixes
      core_param() for genuinely core kernel parameters
      Make initcall_debug a core_param
      Make panic= and panic_on_oops into core_params

 include/linux/module.h      |    6 +-
 include/linux/moduleparam.h |   25 ++++-
 include/linux/workqueue.h   |   18 ++--
 init/main.c                 |   12 +--
 kernel/module.c             |  284 ++++++++++++++++++------------------------
 kernel/panic.c              |   17 +--
 kernel/params.c             |  274 ++++++++++++++++++++++-------------------
 kernel/stop_machine.c       |  120 +++++++-----------
 kernel/workqueue.c          |    7 +-
 9 files changed, 364 insertions(+), 399 deletions(-)

diff --git a/include/linux/module.h b/include/linux/module.h
index 5d2970c..3bfed01 100644
--- a/include/linux/module.h
+++ b/include/linux/module.h
@@ -29,7 +29,7 @@
 #define MODULE_SYMBOL_PREFIX ""
 #endif
 
-#define MODULE_NAME_LEN (64 - sizeof(unsigned long))
+#define MODULE_NAME_LEN MAX_PARAM_PREFIX_LEN
 
 struct kernel_symbol
 {
@@ -60,6 +60,7 @@ struct module_kobject
 	struct kobject kobj;
 	struct module *mod;
 	struct kobject *drivers_dir;
+	struct module_param_attrs *mp;
 };
 
 /* These are either module local, or the kernel's dummy ones. */
@@ -242,7 +243,6 @@ struct module
 
 	/* Sysfs stuff. */
 	struct module_kobject mkobj;
-	struct module_param_attrs *param_attrs;
 	struct module_attribute *modinfo_attrs;
 	const char *version;
 	const char *srcversion;
@@ -277,7 +277,7 @@ struct module
 
 	/* Exception table */
 	unsigned int num_exentries;
-	const struct exception_table_entry *extable;
+	struct exception_table_entry *extable;
 
 	/* Startup function. */
 	int (*init)(void);
diff --git a/include/linux/moduleparam.h b/include/linux/moduleparam.h
index ec62438..e4af339 100644
--- a/include/linux/moduleparam.h
+++ b/include/linux/moduleparam.h
@@ -13,6 +13,9 @@
 #define MODULE_PARAM_PREFIX KBUILD_MODNAME "."
 #endif
 
+/* Chosen so that structs with an unsigned long line up. */
+#define MAX_PARAM_PREFIX_LEN (64 - sizeof(unsigned long))
+
 #ifdef MODULE
 #define ___module_cat(a,b) __mod_ ## a ## b
 #define __module_cat(a,b) ___module_cat(a,b)
@@ -79,7 +82,8 @@ struct kparam_array
 #define __module_param_call(prefix, name, set, get, arg, perm)		\
 	/* Default value instead of permissions? */			\
 	static int __param_perm_check_##name __attribute__((unused)) =	\
-	BUILD_BUG_ON_ZERO((perm) < 0 || (perm) > 0777 || ((perm) & 2));	\
+	BUILD_BUG_ON_ZERO((perm) < 0 || (perm) > 0777 || ((perm) & 2))	\
+	+ BUILD_BUG_ON_ZERO(sizeof(""prefix) > MAX_PARAM_PREFIX_LEN);	\
 	static const char __param_str_##name[] = prefix #name;		\
 	static struct kernel_param __moduleparam_const __param_##name	\
 	__used								\
@@ -100,6 +104,25 @@ struct kparam_array
 #define module_param(name, type, perm)				\
 	module_param_named(name, name, type, perm)
 
+#ifndef MODULE
+/**
+ * core_param - define a historical core kernel parameter.
+ * @name: the name of the cmdline and sysfs parameter (often the same as var)
+ * @var: the variable
+ * @type: the type (for param_set_##type and param_get_##type)
+ * @perm: visibility in sysfs
+ *
+ * core_param is just like module_param(), but cannot be modular and
+ * doesn't add a prefix (such as "printk.").  This is for compatibility
+ * with __setup(), and it makes sense as truly core parameters aren't
+ * tied to the particular file they're in.
+ */
+#define core_param(name, var, type, perm)				\
+	param_check_##type(name, &(var));				\
+	__module_param_call("", name, param_set_##type, param_get_##type, \
+			    &var, perm)
+#endif /* !MODULE */
+
 /* Actually copy string: maxlen param is usually sizeof(string). */
 #define module_param_string(name, string, len, perm)			\
 	static const struct kparam_string __param_string_##name		\
diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h
index 5c158c4..89a5a12 100644
--- a/include/linux/workqueue.h
+++ b/include/linux/workqueue.h
@@ -149,11 +149,11 @@ struct execute_work {
 
 extern struct workqueue_struct *
 __create_workqueue_key(const char *name, int singlethread,
-		       int freezeable, struct lock_class_key *key,
+		       int freezeable, int rt, struct lock_class_key *key,
 		       const char *lock_name);
 
 #ifdef CONFIG_LOCKDEP
-#define __create_workqueue(name, singlethread, freezeable)	\
+#define __create_workqueue(name, singlethread, freezeable, rt)	\
 ({								\
 	static struct lock_class_key __key;			\
 	const char *__lock_name;				\
@@ -164,17 +164,19 @@ __create_workqueue_key(const char *name, int singlethread,
 		__lock_name = #name;				\
 								\
 	__create_workqueue_key((name), (singlethread),		\
-			       (freezeable), &__key,		\
+			       (freezeable), (rt), &__key,	\
 			       __lock_name);			\
 })
 #else
-#define __create_workqueue(name, singlethread, freezeable)	\
-	__create_workqueue_key((name), (singlethread), (freezeable), NULL, NULL)
+#define __create_workqueue(name, singlethread, freezeable, rt)	\
+	__create_workqueue_key((name), (singlethread), (freezeable), (rt), \
+			       NULL, NULL)
 #endif
 
-#define create_workqueue(name) __create_workqueue((name), 0, 0)
-#define create_freezeable_workqueue(name) __create_workqueue((name), 1, 1)
-#define create_singlethread_workqueue(name) __create_workqueue((name), 1, 0)
+#define create_workqueue(name) __create_workqueue((name), 0, 0, 0)
+#define create_rt_workqueue(name) __create_workqueue((name), 0, 0, 1)
+#define create_freezeable_workqueue(name) __create_workqueue((name), 1, 1, 0)
+#define create_singlethread_workqueue(name) __create_workqueue((name), 1, 0, 0)
 
 extern void destroy_workqueue(struct workqueue_struct *wq);
 
diff --git a/init/main.c b/init/main.c
index 3e17a3b..6c7fd13 100644
--- a/init/main.c
+++ b/init/main.c
@@ -697,13 +697,7 @@ asmlinkage void __init start_kernel(void)
 }
 
 static int initcall_debug;
-
-static int __init initcall_debug_setup(char *str)
-{
-	initcall_debug = 1;
-	return 1;
-}
-__setup("initcall_debug", initcall_debug_setup);
+core_param(initcall_debug, initcall_debug, bool, 0644);
 
 int do_one_initcall(initcall_t fn)
 {
@@ -773,8 +767,6 @@ static void __init do_initcalls(void)
 static void __init do_basic_setup(void)
 {
 	rcu_init_sched(); /* needed by module_init stage. */
-	/* drivers will send hotplug events */
-	init_workqueues();
 	usermodehelper_init();
 	driver_init();
 	init_irq_proc();
@@ -858,6 +850,8 @@ static int __init kernel_init(void * unused)
 
 	cad_pid = task_pid(current);
 
+	init_workqueues();
+
 	smp_prepare_cpus(setup_max_cpus);
 
 	do_pre_smp_initcalls();
diff --git a/kernel/module.c b/kernel/module.c
index 0d8d21e..c0f1826 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -42,6 +42,7 @@
 #include <linux/string.h>
 #include <linux/mutex.h>
 #include <linux/unwind.h>
+#include <linux/rculist.h>
 #include <asm/uaccess.h>
 #include <asm/cacheflush.h>
 #include <linux/license.h>
@@ -63,7 +64,7 @@
 #define INIT_OFFSET_MASK (1UL << (BITS_PER_LONG-1))
 
 /* List of modules, protected by module_mutex or preempt_disable
- * (add/delete uses stop_machine). */
+ * (delete uses stop_machine/add uses RCU list operations). */
 static DEFINE_MUTEX(module_mutex);
 static LIST_HEAD(modules);
 
@@ -132,6 +133,29 @@ static unsigned int find_sec(Elf_Ehdr *hdr,
 	return 0;
 }
 
+/* Find a module section, or NULL. */
+static void *section_addr(Elf_Ehdr *hdr, Elf_Shdr *shdrs,
+			  const char *secstrings, const char *name)
+{
+	/* Section 0 has sh_addr 0. */
+	return (void *)shdrs[find_sec(hdr, shdrs, secstrings, name)].sh_addr;
+}
+
+/* Find a module section, or NULL.  Fill in number of "objects" in section. */
+static void *section_objs(Elf_Ehdr *hdr,
+			  Elf_Shdr *sechdrs,
+			  const char *secstrings,
+			  const char *name,
+			  size_t object_size,
+			  unsigned int *num)
+{
+	unsigned int sec = find_sec(hdr, sechdrs, secstrings, name);
+
+	/* Section 0 has sh_addr 0 and sh_size 0. */
+	*num = sechdrs[sec].sh_size / object_size;
+	return (void *)sechdrs[sec].sh_addr;
+}
+
 /* Provided by the linker */
 extern const struct kernel_symbol __start___ksymtab[];
 extern const struct kernel_symbol __stop___ksymtab[];
@@ -218,7 +242,7 @@ static bool each_symbol(bool (*fn)(const struct symsearch *arr,
 	if (each_symbol_in_section(arr, ARRAY_SIZE(arr), NULL, fn, data))
 		return true;
 
-	list_for_each_entry(mod, &modules, list) {
+	list_for_each_entry_rcu(mod, &modules, list) {
 		struct symsearch arr[] = {
 			{ mod->syms, mod->syms + mod->num_syms, mod->crcs,
 			  NOT_GPL_ONLY, false },
@@ -1394,17 +1418,6 @@ static void mod_kobject_remove(struct module *mod)
 }
 
 /*
- * link the module with the whole machine is stopped with interrupts off
- * - this defends against kallsyms not taking locks
- */
-static int __link_module(void *_mod)
-{
-	struct module *mod = _mod;
-	list_add(&mod->list, &modules);
-	return 0;
-}
-
-/*
  * unlink the module with the whole machine is stopped with interrupts off
  * - this defends against kallsyms not taking locks
  */
@@ -1789,32 +1802,20 @@ static inline void add_kallsyms(struct module *mod,
 }
 #endif /* CONFIG_KALLSYMS */
 
-#ifdef CONFIG_DYNAMIC_PRINTK_DEBUG
-static void dynamic_printk_setup(Elf_Shdr *sechdrs, unsigned int verboseindex)
+static void dynamic_printk_setup(struct mod_debug *debug, unsigned int num)
 {
-	struct mod_debug *debug_info;
-	unsigned long pos, end;
-	unsigned int num_verbose;
-
-	pos = sechdrs[verboseindex].sh_addr;
-	num_verbose = sechdrs[verboseindex].sh_size /
-				sizeof(struct mod_debug);
-	end = pos + (num_verbose * sizeof(struct mod_debug));
+#ifdef CONFIG_DYNAMIC_PRINTK_DEBUG
+	unsigned int i;
 
-	for (; pos < end; pos += sizeof(struct mod_debug)) {
-		debug_info = (struct mod_debug *)pos;
-		register_dynamic_debug_module(debug_info->modname,
-			debug_info->type, debug_info->logical_modname,
-			debug_info->flag_names, debug_info->hash,
-			debug_info->hash2);
+	for (i = 0; i < num; i++) {
+		register_dynamic_debug_module(debug[i].modname,
+					      debug[i].type,
+					      debug[i].logical_modname,
+					      debug[i].flag_names,
+					      debug[i].hash, debug[i].hash2);
 	}
-}
-#else
-static inline void dynamic_printk_setup(Elf_Shdr *sechdrs,
-					unsigned int verboseindex)
-{
-}
 #endif /* CONFIG_DYNAMIC_PRINTK_DEBUG */
+}
 
 static void *module_alloc_update_bounds(unsigned long size)
 {
@@ -1843,37 +1844,14 @@ static noinline struct module *load_module(void __user *umod,
 	unsigned int i;
 	unsigned int symindex = 0;
 	unsigned int strindex = 0;
-	unsigned int setupindex;
-	unsigned int exindex;
-	unsigned int exportindex;
-	unsigned int modindex;
-	unsigned int obsparmindex;
-	unsigned int infoindex;
-	unsigned int gplindex;
-	unsigned int crcindex;
-	unsigned int gplcrcindex;
-	unsigned int versindex;
-	unsigned int pcpuindex;
-	unsigned int gplfutureindex;
-	unsigned int gplfuturecrcindex;
+	unsigned int modindex, versindex, infoindex, pcpuindex;
 	unsigned int unwindex = 0;
-#ifdef CONFIG_UNUSED_SYMBOLS
-	unsigned int unusedindex;
-	unsigned int unusedcrcindex;
-	unsigned int unusedgplindex;
-	unsigned int unusedgplcrcindex;
-#endif
-	unsigned int markersindex;
-	unsigned int markersstringsindex;
-	unsigned int verboseindex;
-	unsigned int tracepointsindex;
-	unsigned int tracepointsstringsindex;
-	unsigned int mcountindex;
+	unsigned int num_kp, num_mcount;
+	struct kernel_param *kp;
 	struct module *mod;
 	long err = 0;
 	void *percpu = NULL, *ptr = NULL; /* Stops spurious gcc warning */
-	void *mseg;
-	struct exception_table_entry *extable;
+	unsigned long *mseg;
 	mm_segment_t old_fs;
 
 	DEBUGP("load_module: umod=%p, len=%lu, uargs=%p\n",
@@ -1937,6 +1915,7 @@ static noinline struct module *load_module(void __user *umod,
 		err = -ENOEXEC;
 		goto free_hdr;
 	}
+	/* This is temporary: point mod into copy of data. */
 	mod = (void *)sechdrs[modindex].sh_addr;
 
 	if (symindex == 0) {
@@ -1946,22 +1925,6 @@ static noinline struct module *load_module(void __user *umod,
 		goto free_hdr;
 	}
 
-	/* Optional sections */
-	exportindex = find_sec(hdr, sechdrs, secstrings, "__ksymtab");
-	gplindex = find_sec(hdr, sechdrs, secstrings, "__ksymtab_gpl");
-	gplfutureindex = find_sec(hdr, sechdrs, secstrings, "__ksymtab_gpl_future");
-	crcindex = find_sec(hdr, sechdrs, secstrings, "__kcrctab");
-	gplcrcindex = find_sec(hdr, sechdrs, secstrings, "__kcrctab_gpl");
-	gplfuturecrcindex = find_sec(hdr, sechdrs, secstrings, "__kcrctab_gpl_future");
-#ifdef CONFIG_UNUSED_SYMBOLS
-	unusedindex = find_sec(hdr, sechdrs, secstrings, "__ksymtab_unused");
-	unusedgplindex = find_sec(hdr, sechdrs, secstrings, "__ksymtab_unused_gpl");
-	unusedcrcindex = find_sec(hdr, sechdrs, secstrings, "__kcrctab_unused");
-	unusedgplcrcindex = find_sec(hdr, sechdrs, secstrings, "__kcrctab_unused_gpl");
-#endif
-	setupindex = find_sec(hdr, sechdrs, secstrings, "__param");
-	exindex = find_sec(hdr, sechdrs, secstrings, "__ex_table");
-	obsparmindex = find_sec(hdr, sechdrs, secstrings, "__obsparm");
 	versindex = find_sec(hdr, sechdrs, secstrings, "__versions");
 	infoindex = find_sec(hdr, sechdrs, secstrings, ".modinfo");
 	pcpuindex = find_pcpusec(hdr, sechdrs, secstrings);
@@ -2117,42 +2080,57 @@ static noinline struct module *load_module(void __user *umod,
 	if (err < 0)
 		goto cleanup;
 
-	/* Set up EXPORTed & EXPORT_GPLed symbols (section 0 is 0 length) */
-	mod->num_syms = sechdrs[exportindex].sh_size / sizeof(*mod->syms);
-	mod->syms = (void *)sechdrs[exportindex].sh_addr;
-	if (crcindex)
-		mod->crcs = (void *)sechdrs[crcindex].sh_addr;
-	mod->num_gpl_syms = sechdrs[gplindex].sh_size / sizeof(*mod->gpl_syms);
-	mod->gpl_syms = (void *)sechdrs[gplindex].sh_addr;
-	if (gplcrcindex)
-		mod->gpl_crcs = (void *)sechdrs[gplcrcindex].sh_addr;
-	mod->num_gpl_future_syms = sechdrs[gplfutureindex].sh_size /
-					sizeof(*mod->gpl_future_syms);
-	mod->gpl_future_syms = (void *)sechdrs[gplfutureindex].sh_addr;
-	if (gplfuturecrcindex)
-		mod->gpl_future_crcs = (void *)sechdrs[gplfuturecrcindex].sh_addr;
+	/* Now we've got everything in the final locations, we can
+	 * find optional sections. */
+	kp = section_objs(hdr, sechdrs, secstrings, "__param", sizeof(*kp),
+			  &num_kp);
+	mod->syms = section_objs(hdr, sechdrs, secstrings, "__ksymtab",
+				 sizeof(*mod->syms), &mod->num_syms);
+	mod->crcs = section_addr(hdr, sechdrs, secstrings, "__kcrctab");
+	mod->gpl_syms = section_objs(hdr, sechdrs, secstrings, "__ksymtab_gpl",
+				     sizeof(*mod->gpl_syms),
+				     &mod->num_gpl_syms);
+	mod->gpl_crcs = section_addr(hdr, sechdrs, secstrings, "__kcrctab_gpl");
+	mod->gpl_future_syms = section_objs(hdr, sechdrs, secstrings,
+					    "__ksymtab_gpl_future",
+					    sizeof(*mod->gpl_future_syms),
+					    &mod->num_gpl_future_syms);
+	mod->gpl_future_crcs = section_addr(hdr, sechdrs, secstrings,
+					    "__kcrctab_gpl_future");
 
 #ifdef CONFIG_UNUSED_SYMBOLS
-	mod->num_unused_syms = sechdrs[unusedindex].sh_size /
-					sizeof(*mod->unused_syms);
-	mod->num_unused_gpl_syms = sechdrs[unusedgplindex].sh_size /
-					sizeof(*mod->unused_gpl_syms);
-	mod->unused_syms = (void *)sechdrs[unusedindex].sh_addr;
-	if (unusedcrcindex)
-		mod->unused_crcs = (void *)sechdrs[unusedcrcindex].sh_addr;
-	mod->unused_gpl_syms = (void *)sechdrs[unusedgplindex].sh_addr;
-	if (unusedgplcrcindex)
-		mod->unused_gpl_crcs
-			= (void *)sechdrs[unusedgplcrcindex].sh_addr;
+	mod->unused_syms = section_objs(hdr, sechdrs, secstrings,
+					"__ksymtab_unused",
+					sizeof(*mod->unused_syms),
+					&mod->num_unused_syms);
+	mod->unused_crcs = section_addr(hdr, sechdrs, secstrings,
+					"__kcrctab_unused");
+	mod->unused_gpl_syms = section_objs(hdr, sechdrs, secstrings,
+					    "__ksymtab_unused_gpl",
+					    sizeof(*mod->unused_gpl_syms),
+					    &mod->num_unused_gpl_syms);
+	mod->unused_gpl_crcs = section_addr(hdr, sechdrs, secstrings,
+					    "__kcrctab_unused_gpl");
+#endif
+
+#ifdef CONFIG_MARKERS
+	mod->markers = section_objs(hdr, sechdrs, secstrings, "__markers",
+				    sizeof(*mod->markers), &mod->num_markers);
+#endif
+#ifdef CONFIG_TRACEPOINTS
+	mod->tracepoints = section_objs(hdr, sechdrs, secstrings,
+					"__tracepoints",
+					sizeof(*mod->tracepoints),
+					&mod->num_tracepoints);
 #endif
 
 #ifdef CONFIG_MODVERSIONS
-	if ((mod->num_syms && !crcindex)
-	    || (mod->num_gpl_syms && !gplcrcindex)
-	    || (mod->num_gpl_future_syms && !gplfuturecrcindex)
+	if ((mod->num_syms && !mod->crcs)
+	    || (mod->num_gpl_syms && !mod->gpl_crcs)
+	    || (mod->num_gpl_future_syms && !mod->gpl_future_crcs)
 #ifdef CONFIG_UNUSED_SYMBOLS
-	    || (mod->num_unused_syms && !unusedcrcindex)
-	    || (mod->num_unused_gpl_syms && !unusedgplcrcindex)
+	    || (mod->num_unused_syms && !mod->unused_crcs)
+	    || (mod->num_unused_gpl_syms && !mod->unused_gpl_crcs)
 #endif
 		) {
 		printk(KERN_WARNING "%s: No versions for exported symbols.\n", mod->name);
@@ -2161,16 +2139,6 @@ static noinline struct module *load_module(void __user *umod,
 			goto cleanup;
 	}
 #endif
-	markersindex = find_sec(hdr, sechdrs, secstrings, "__markers");
- 	markersstringsindex = find_sec(hdr, sechdrs, secstrings,
-					"__markers_strings");
-	verboseindex = find_sec(hdr, sechdrs, secstrings, "__verbose");
-	tracepointsindex = find_sec(hdr, sechdrs, secstrings, "__tracepoints");
-	tracepointsstringsindex = find_sec(hdr, sechdrs, secstrings,
-					"__tracepoints_strings");
-
-	mcountindex = find_sec(hdr, sechdrs, secstrings,
-			       "__mcount_loc");
 
 	/* Now do relocations. */
 	for (i = 1; i < hdr->e_shnum; i++) {
@@ -2193,28 +2161,16 @@ static noinline struct module *load_module(void __user *umod,
 		if (err < 0)
 			goto cleanup;
 	}
-#ifdef CONFIG_MARKERS
-	mod->markers = (void *)sechdrs[markersindex].sh_addr;
-	mod->num_markers =
-		sechdrs[markersindex].sh_size / sizeof(*mod->markers);
-#endif
-#ifdef CONFIG_TRACEPOINTS
-	mod->tracepoints = (void *)sechdrs[tracepointsindex].sh_addr;
-	mod->num_tracepoints =
-		sechdrs[tracepointsindex].sh_size / sizeof(*mod->tracepoints);
-#endif
-
 
         /* Find duplicate symbols */
 	err = verify_export_symbols(mod);
-
 	if (err < 0)
 		goto cleanup;
 
   	/* Set up and sort exception table */
-	mod->num_exentries = sechdrs[exindex].sh_size / sizeof(*mod->extable);
-	mod->extable = extable = (void *)sechdrs[exindex].sh_addr;
-	sort_extable(extable, extable + mod->num_exentries);
+	mod->extable = section_objs(hdr, sechdrs, secstrings, "__ex_table",
+				    sizeof(*mod->extable), &mod->num_exentries);
+	sort_extable(mod->extable, mod->extable + mod->num_exentries);
 
 	/* Finally, copy percpu area over. */
 	percpu_modcopy(mod->percpu, (void *)sechdrs[pcpuindex].sh_addr,
@@ -2223,11 +2179,17 @@ static noinline struct module *load_module(void __user *umod,
 	add_kallsyms(mod, sechdrs, symindex, strindex, secstrings);
 
 	if (!mod->taints) {
+		struct mod_debug *debug;
+		unsigned int num_debug;
+
 #ifdef CONFIG_MARKERS
 		marker_update_probe_range(mod->markers,
 			mod->markers + mod->num_markers);
 #endif
-	dynamic_printk_setup(sechdrs, verboseindex);
+		debug = section_objs(hdr, sechdrs, secstrings, "__verbose",
+				     sizeof(*debug), &num_debug);
+		dynamic_printk_setup(debug, num_debug);
+
 #ifdef CONFIG_TRACEPOINTS
 		tracepoint_update_probe_range(mod->tracepoints,
 			mod->tracepoints + mod->num_tracepoints);
@@ -2235,8 +2197,9 @@ static noinline struct module *load_module(void __user *umod,
 	}
 
 	/* sechdrs[0].sh_size is always zero */
-	mseg = (void *)sechdrs[mcountindex].sh_addr;
-	ftrace_init_module(mseg, mseg + sechdrs[mcountindex].sh_size);
+	mseg = section_objs(hdr, sechdrs, secstrings, "__mcount_loc",
+			    sizeof(*mseg), &num_mcount);
+	ftrace_init_module(mseg, mseg + num_mcount);
 
 	err = module_finalize(hdr, sechdrs, mod);
 	if (err < 0)
@@ -2261,30 +2224,24 @@ static noinline struct module *load_module(void __user *umod,
 	set_fs(old_fs);
 
 	mod->args = args;
-	if (obsparmindex)
+	if (section_addr(hdr, sechdrs, secstrings, "__obsparm"))
 		printk(KERN_WARNING "%s: Ignoring obsolete parameters\n",
 		       mod->name);
 
 	/* Now sew it into the lists so we can get lockdep and oops
-         * info during argument parsing.  Noone should access us, since
-         * strong_try_module_get() will fail. */
-	stop_machine(__link_module, mod, NULL);
-
-	/* Size of section 0 is 0, so this works well if no params */
-	err = parse_args(mod->name, mod->args,
-			 (struct kernel_param *)
-			 sechdrs[setupindex].sh_addr,
-			 sechdrs[setupindex].sh_size
-			 / sizeof(struct kernel_param),
-			 NULL);
+	 * info during argument parsing.  Noone should access us, since
+	 * strong_try_module_get() will fail.
+	 * lockdep/oops can run asynchronous, so use the RCU list insertion
+	 * function to insert in a way safe to concurrent readers.
+	 * The mutex protects against concurrent writers.
+	 */
+	list_add_rcu(&mod->list, &modules);
+
+	err = parse_args(mod->name, mod->args, kp, num_kp, NULL);
 	if (err < 0)
 		goto unlink;
 
-	err = mod_sysfs_setup(mod,
-			      (struct kernel_param *)
-			      sechdrs[setupindex].sh_addr,
-			      sechdrs[setupindex].sh_size
-			      / sizeof(struct kernel_param));
+	err = mod_sysfs_setup(mod, kp, num_kp);
 	if (err < 0)
 		goto unlink;
 	add_sect_attrs(mod, hdr->e_shnum, secstrings, sechdrs);
@@ -2473,7 +2430,7 @@ const char *module_address_lookup(unsigned long addr,
 	const char *ret = NULL;
 
 	preempt_disable();
-	list_for_each_entry(mod, &modules, list) {
+	list_for_each_entry_rcu(mod, &modules, list) {
 		if (within(addr, mod->module_init, mod->init_size)
 		    || within(addr, mod->module_core, mod->core_size)) {
 			if (modname)
@@ -2496,7 +2453,7 @@ int lookup_module_symbol_name(unsigned long addr, char *symname)
 	struct module *mod;
 
 	preempt_disable();
-	list_for_each_entry(mod, &modules, list) {
+	list_for_each_entry_rcu(mod, &modules, list) {
 		if (within(addr, mod->module_init, mod->init_size) ||
 		    within(addr, mod->module_core, mod->core_size)) {
 			const char *sym;
@@ -2520,7 +2477,7 @@ int lookup_module_symbol_attrs(unsigned long addr, unsigned long *size,
 	struct module *mod;
 
 	preempt_disable();
-	list_for_each_entry(mod, &modules, list) {
+	list_for_each_entry_rcu(mod, &modules, list) {
 		if (within(addr, mod->module_init, mod->init_size) ||
 		    within(addr, mod->module_core, mod->core_size)) {
 			const char *sym;
@@ -2547,7 +2504,7 @@ int module_get_kallsym(unsigned int symnum, unsigned long *value, char *type,
 	struct module *mod;
 
 	preempt_disable();
-	list_for_each_entry(mod, &modules, list) {
+	list_for_each_entry_rcu(mod, &modules, list) {
 		if (symnum < mod->num_symtab) {
 			*value = mod->symtab[symnum].st_value;
 			*type = mod->symtab[symnum].st_info;
@@ -2590,7 +2547,7 @@ unsigned long module_kallsyms_lookup_name(const char *name)
 			ret = mod_find_symname(mod, colon+1);
 		*colon = ':';
 	} else {
-		list_for_each_entry(mod, &modules, list)
+		list_for_each_entry_rcu(mod, &modules, list)
 			if ((ret = mod_find_symname(mod, name)) != 0)
 				break;
 	}
@@ -2693,7 +2650,7 @@ const struct exception_table_entry *search_module_extables(unsigned long addr)
 	struct module *mod;
 
 	preempt_disable();
-	list_for_each_entry(mod, &modules, list) {
+	list_for_each_entry_rcu(mod, &modules, list) {
 		if (mod->num_exentries == 0)
 			continue;
 
@@ -2719,7 +2676,7 @@ int is_module_address(unsigned long addr)
 
 	preempt_disable();
 
-	list_for_each_entry(mod, &modules, list) {
+	list_for_each_entry_rcu(mod, &modules, list) {
 		if (within(addr, mod->module_core, mod->core_size)) {
 			preempt_enable();
 			return 1;
@@ -2740,7 +2697,7 @@ struct module *__module_text_address(unsigned long addr)
 	if (addr < module_addr_min || addr > module_addr_max)
 		return NULL;
 
-	list_for_each_entry(mod, &modules, list)
+	list_for_each_entry_rcu(mod, &modules, list)
 		if (within(addr, mod->module_init, mod->init_text_size)
 		    || within(addr, mod->module_core, mod->core_text_size))
 			return mod;
@@ -2765,8 +2722,11 @@ void print_modules(void)
 	char buf[8];
 
 	printk("Modules linked in:");
-	list_for_each_entry(mod, &modules, list)
+	/* Most callers should already have preempt disabled, but make sure */
+	preempt_disable();
+	list_for_each_entry_rcu(mod, &modules, list)
 		printk(" %s%s", mod->name, module_flags(mod, buf));
+	preempt_enable();
 	if (last_unloaded_module[0])
 		printk(" [last unloaded: %s]", last_unloaded_module);
 	printk("\n");
diff --git a/kernel/panic.c b/kernel/panic.c
index bda561e..6513aac 100644
--- a/kernel/panic.c
+++ b/kernel/panic.c
@@ -34,13 +34,6 @@ ATOMIC_NOTIFIER_HEAD(panic_notifier_list);
 
 EXPORT_SYMBOL(panic_notifier_list);
 
-static int __init panic_setup(char *str)
-{
-	panic_timeout = simple_strtoul(str, NULL, 0);
-	return 1;
-}
-__setup("panic=", panic_setup);
-
 static long no_blink(long time)
 {
 	return 0;
@@ -218,13 +211,6 @@ void add_taint(unsigned flag)
 }
 EXPORT_SYMBOL(add_taint);
 
-static int __init pause_on_oops_setup(char *str)
-{
-	pause_on_oops = simple_strtoul(str, NULL, 0);
-	return 1;
-}
-__setup("pause_on_oops=", pause_on_oops_setup);
-
 static void spin_msec(int msecs)
 {
 	int i;
@@ -384,3 +370,6 @@ void __stack_chk_fail(void)
 }
 EXPORT_SYMBOL(__stack_chk_fail);
 #endif
+
+core_param(panic, panic_timeout, int, 0644);
+core_param(pause_on_oops, pause_on_oops, int, 0644);
diff --git a/kernel/params.c b/kernel/params.c
index afc46a2..b077f1b 100644
--- a/kernel/params.c
+++ b/kernel/params.c
@@ -373,6 +373,8 @@ int param_get_string(char *buffer, struct kernel_param *kp)
 }
 
 /* sysfs output in /sys/modules/XYZ/parameters/ */
+#define to_module_attr(n) container_of(n, struct module_attribute, attr);
+#define to_module_kobject(n) container_of(n, struct module_kobject, kobj);
 
 extern struct kernel_param __start___param[], __stop___param[];
 
@@ -384,6 +386,7 @@ struct param_attribute
 
 struct module_param_attrs
 {
+	unsigned int num;
 	struct attribute_group grp;
 	struct param_attribute attrs[0];
 };
@@ -434,69 +437,84 @@ static ssize_t param_attr_store(struct module_attribute *mattr,
 
 #ifdef CONFIG_SYSFS
 /*
- * param_sysfs_setup - setup sysfs support for one module or KBUILD_MODNAME
- * @mk: struct module_kobject (contains parent kobject)
- * @kparam: array of struct kernel_param, the actual parameter definitions
- * @num_params: number of entries in array
- * @name_skip: offset where the parameter name start in kparam[].name. Needed for built-in "modules"
+ * add_sysfs_param - add a parameter to sysfs
+ * @mk: struct module_kobject
+ * @kparam: the actual parameter definition to add to sysfs
+ * @name: name of parameter
  *
- * Create a kobject for a (per-module) group of parameters, and create files
- * in sysfs. A pointer to the param_kobject is returned on success,
- * NULL if there's no parameter to export, or other ERR_PTR(err).
+ * Create a kobject if for a (per-module) parameter if mp NULL, and
+ * create file in sysfs.  Returns an error on out of memory.  Always cleans up
+ * if there's an error.
  */
-static __modinit struct module_param_attrs *
-param_sysfs_setup(struct module_kobject *mk,
-		  struct kernel_param *kparam,
-		  unsigned int num_params,
-		  unsigned int name_skip)
+static __modinit int add_sysfs_param(struct module_kobject *mk,
+				     struct kernel_param *kp,
+				     const char *name)
 {
-	struct module_param_attrs *mp;
-	unsigned int valid_attrs = 0;
-	unsigned int i, size[2];
-	struct param_attribute *pattr;
-	struct attribute **gattr;
-	int err;
-
-	for (i=0; i<num_params; i++) {
-		if (kparam[i].perm)
-			valid_attrs++;
+	struct module_param_attrs *new;
+	struct attribute **attrs;
+	int err, num;
+
+	/* We don't bother calling this with invisible parameters. */
+	BUG_ON(!kp->perm);
+
+	if (!mk->mp) {
+		num = 0;
+		attrs = NULL;
+	} else {
+		num = mk->mp->num;
+		attrs = mk->mp->grp.attrs;
 	}
 
-	if (!valid_attrs)
-		return NULL;
-
-	size[0] = ALIGN(sizeof(*mp) +
-			valid_attrs * sizeof(mp->attrs[0]),
-			sizeof(mp->grp.attrs[0]));
-	size[1] = (valid_attrs + 1) * sizeof(mp->grp.attrs[0]);
-
-	mp = kzalloc(size[0] + size[1], GFP_KERNEL);
-	if (!mp)
-		return ERR_PTR(-ENOMEM);
+	/* Enlarge. */
+	new = krealloc(mk->mp,
+		       sizeof(*mk->mp) + sizeof(mk->mp->attrs[0]) * (num+1),
+		       GFP_KERNEL);
+	if (!new) {
+		kfree(mk->mp);
+		err = -ENOMEM;
+		goto fail;
+	}
+	attrs = krealloc(attrs, sizeof(new->grp.attrs[0])*(num+2), GFP_KERNEL);
+	if (!attrs) {
+		err = -ENOMEM;
+		goto fail_free_new;
+	}
 
-	mp->grp.name = "parameters";
-	mp->grp.attrs = (void *)mp + size[0];
+	/* Sysfs wants everything zeroed. */
+	memset(new, 0, sizeof(*new));
+	memset(&new->attrs[num], 0, sizeof(new->attrs[num]));
+	memset(&attrs[num], 0, sizeof(attrs[num]));
+	new->grp.name = "parameters";
+	new->grp.attrs = attrs;
+
+	/* Tack new one on the end. */
+	new->attrs[num].param = kp;
+	new->attrs[num].mattr.show = param_attr_show;
+	new->attrs[num].mattr.store = param_attr_store;
+	new->attrs[num].mattr.attr.name = (char *)name;
+	new->attrs[num].mattr.attr.mode = kp->perm;
+	new->num = num+1;
+
+	/* Fix up all the pointers, since krealloc can move us */
+	for (num = 0; num < new->num; num++)
+		new->grp.attrs[num] = &new->attrs[num].mattr.attr;
+	new->grp.attrs[num] = NULL;
+
+	mk->mp = new;
+	return 0;
 
-	pattr = &mp->attrs[0];
-	gattr = &mp->grp.attrs[0];
-	for (i = 0; i < num_params; i++) {
-		struct kernel_param *kp = &kparam[i];
-		if (kp->perm) {
-			pattr->param = kp;
-			pattr->mattr.show = param_attr_show;
-			pattr->mattr.store = param_attr_store;
-			pattr->mattr.attr.name = (char *)&kp->name[name_skip];
-			pattr->mattr.attr.mode = kp->perm;
-			*(gattr++) = &(pattr++)->mattr.attr;
-		}
-	}
-	*gattr = NULL;
+fail_free_new:
+	kfree(new);
+fail:
+	mk->mp = NULL;
+	return err;
+}
 
-	if ((err = sysfs_create_group(&mk->kobj, &mp->grp))) {
-		kfree(mp);
-		return ERR_PTR(err);
-	}
-	return mp;
+static void free_module_param_attrs(struct module_kobject *mk)
+{
+	kfree(mk->mp->grp.attrs);
+	kfree(mk->mp);
+	mk->mp = NULL;
 }
 
 #ifdef CONFIG_MODULES
@@ -506,21 +524,33 @@ param_sysfs_setup(struct module_kobject *mk,
  * @kparam: module parameters (array)
  * @num_params: number of module parameters
  *
- * Adds sysfs entries for module parameters, and creates a link from
- * /sys/module/[mod->name]/parameters to /sys/parameters/[mod->name]/
+ * Adds sysfs entries for module parameters under
+ * /sys/module/[mod->name]/parameters/
  */
 int module_param_sysfs_setup(struct module *mod,
 			     struct kernel_param *kparam,
 			     unsigned int num_params)
 {
-	struct module_param_attrs *mp;
+	int i, err;
+	bool params = false;
+
+	for (i = 0; i < num_params; i++) {
+		if (kparam[i].perm == 0)
+			continue;
+		err = add_sysfs_param(&mod->mkobj, &kparam[i], kparam[i].name);
+		if (err)
+			return err;
+		params = true;
+	}
 
-	mp = param_sysfs_setup(&mod->mkobj, kparam, num_params, 0);
-	if (IS_ERR(mp))
-		return PTR_ERR(mp);
+	if (!params)
+		return 0;
 
-	mod->param_attrs = mp;
-	return 0;
+	/* Create the param group. */
+	err = sysfs_create_group(&mod->mkobj.kobj, &mod->mkobj.mp->grp);
+	if (err)
+		free_module_param_attrs(&mod->mkobj);
+	return err;
 }
 
 /*
@@ -532,43 +562,55 @@ int module_param_sysfs_setup(struct module *mod,
  */
 void module_param_sysfs_remove(struct module *mod)
 {
-	if (mod->param_attrs) {
-		sysfs_remove_group(&mod->mkobj.kobj,
-				   &mod->param_attrs->grp);
+	if (mod->mkobj.mp) {
+		sysfs_remove_group(&mod->mkobj.kobj, &mod->mkobj.mp->grp);
 		/* We are positive that no one is using any param
 		 * attrs at this point.  Deallocate immediately. */
-		kfree(mod->param_attrs);
-		mod->param_attrs = NULL;
+		free_module_param_attrs(&mod->mkobj);
 	}
 }
 #endif
 
-/*
- * kernel_param_sysfs_setup - wrapper for built-in params support
- */
-static void __init kernel_param_sysfs_setup(const char *name,
-					    struct kernel_param *kparam,
-					    unsigned int num_params,
-					    unsigned int name_skip)
+static void __init kernel_add_sysfs_param(const char *name,
+					  struct kernel_param *kparam,
+					  unsigned int name_skip)
 {
 	struct module_kobject *mk;
-	int ret;
+	struct kobject *kobj;
+	int err;
 
-	mk = kzalloc(sizeof(struct module_kobject), GFP_KERNEL);
-	BUG_ON(!mk);
-
-	mk->mod = THIS_MODULE;
-	mk->kobj.kset = module_kset;
-	ret = kobject_init_and_add(&mk->kobj, &module_ktype, NULL, "%s", name);
-	if (ret) {
-		kobject_put(&mk->kobj);
-		printk(KERN_ERR "Module '%s' failed to be added to sysfs, "
-		      "error number %d\n", name, ret);
-		printk(KERN_ERR	"The system will be unstable now.\n");
-		return;
+	kobj = kset_find_obj(module_kset, name);
+	if (kobj) {
+		/* We already have one.  Remove params so we can add more. */
+		mk = to_module_kobject(kobj);
+		/* We need to remove it before adding parameters. */
+		sysfs_remove_group(&mk->kobj, &mk->mp->grp);
+	} else {
+		mk = kzalloc(sizeof(struct module_kobject), GFP_KERNEL);
+		BUG_ON(!mk);
+
+		mk->mod = THIS_MODULE;
+		mk->kobj.kset = module_kset;
+		err = kobject_init_and_add(&mk->kobj, &module_ktype, NULL,
+					   "%s", name);
+		if (err) {
+			kobject_put(&mk->kobj);
+			printk(KERN_ERR "Module '%s' failed add to sysfs, "
+			       "error number %d\n", name, err);
+			printk(KERN_ERR	"The system will be unstable now.\n");
+			return;
+		}
+		/* So that exit path is even. */
+		kobject_get(&mk->kobj);
 	}
-	param_sysfs_setup(mk, kparam, num_params, name_skip);
+
+	/* These should not fail at boot. */
+	err = add_sysfs_param(mk, kparam, kparam->name + name_skip);
+	BUG_ON(err);
+	err = sysfs_create_group(&mk->kobj, &mk->mp->grp);
+	BUG_ON(err);
 	kobject_uevent(&mk->kobj, KOBJ_ADD);
+	kobject_put(&mk->kobj);
 }
 
 /*
@@ -579,60 +621,36 @@ static void __init kernel_param_sysfs_setup(const char *name,
  * The "module" name (KBUILD_MODNAME) is stored before a dot, the
  * "parameter" name is stored behind a dot in kernel_param->name. So,
  * extract the "module" name for all built-in kernel_param-eters,
- * and for all who have the same, call kernel_param_sysfs_setup.
+ * and for all who have the same, call kernel_add_sysfs_param.
  */
 static void __init param_sysfs_builtin(void)
 {
-	struct kernel_param *kp, *kp_begin = NULL;
-	unsigned int i, name_len, count = 0;
-	char modname[MODULE_NAME_LEN + 1] = "";
+	struct kernel_param *kp;
+	unsigned int name_len;
+	char modname[MODULE_NAME_LEN];
 
-	for (i=0; i < __stop___param - __start___param; i++) {
+	for (kp = __start___param; kp < __stop___param; kp++) {
 		char *dot;
-		size_t max_name_len;
 
-		kp = &__start___param[i];
-		max_name_len =
-			min_t(size_t, MODULE_NAME_LEN, strlen(kp->name));
+		if (kp->perm == 0)
+			continue;
 
-		dot = memchr(kp->name, '.', max_name_len);
+		dot = strchr(kp->name, '.');
 		if (!dot) {
-			DEBUGP("couldn't find period in first %d characters "
-			       "of %s\n", MODULE_NAME_LEN, kp->name);
-			continue;
-		}
-		name_len = dot - kp->name;
-
- 		/* new kbuild_modname? */
-		if (strlen(modname) != name_len
-		    || strncmp(modname, kp->name, name_len) != 0) {
-			/* add a new kobject for previous kernel_params. */
-			if (count)
-				kernel_param_sysfs_setup(modname,
-							 kp_begin,
-							 count,
-							 strlen(modname)+1);
-
-			strncpy(modname, kp->name, name_len);
-			modname[name_len] = '\0';
-			count = 0;
-			kp_begin = kp;
+			/* This happens for core_param() */
+			strcpy(modname, "kernel");
+			name_len = 0;
+		} else {
+			name_len = dot - kp->name + 1;
+			strlcpy(modname, kp->name, name_len);
 		}
-		count++;
+		kernel_add_sysfs_param(modname, kp, name_len);
 	}
-
-	/* last kernel_params need to be registered as well */
-	if (count)
-		kernel_param_sysfs_setup(modname, kp_begin, count,
-					 strlen(modname)+1);
 }
 
 
 /* module-related sysfs stuff */
 
-#define to_module_attr(n) container_of(n, struct module_attribute, attr);
-#define to_module_kobject(n) container_of(n, struct module_kobject, kobj);
-
 static ssize_t module_attr_show(struct kobject *kobj,
 				struct attribute *attr,
 				char *buf)
diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c
index af3c7ce..8aff79d 100644
--- a/kernel/stop_machine.c
+++ b/kernel/stop_machine.c
@@ -37,9 +37,13 @@ struct stop_machine_data {
 /* Like num_online_cpus(), but hotplug cpu uses us, so we need this. */
 static unsigned int num_threads;
 static atomic_t thread_ack;
-static struct completion finished;
 static DEFINE_MUTEX(lock);
 
+static struct workqueue_struct *stop_machine_wq;
+static struct stop_machine_data active, idle;
+static const cpumask_t *active_cpus;
+static void *stop_machine_work;
+
 static void set_state(enum stopmachine_state newstate)
 {
 	/* Reset ack counter. */
@@ -51,21 +55,26 @@ static void set_state(enum stopmachine_state newstate)
 /* Last one to ack a state moves to the next state. */
 static void ack_state(void)
 {
-	if (atomic_dec_and_test(&thread_ack)) {
-		/* If we're the last one to ack the EXIT, we're finished. */
-		if (state == STOPMACHINE_EXIT)
-			complete(&finished);
-		else
-			set_state(state + 1);
-	}
+	if (atomic_dec_and_test(&thread_ack))
+		set_state(state + 1);
 }
 
-/* This is the actual thread which stops the CPU.  It exits by itself rather
- * than waiting for kthread_stop(), because it's easier for hotplug CPU. */
-static int stop_cpu(struct stop_machine_data *smdata)
+/* This is the actual function which stops the CPU. It runs
+ * in the context of a dedicated stopmachine workqueue. */
+static void stop_cpu(struct work_struct *unused)
 {
 	enum stopmachine_state curstate = STOPMACHINE_NONE;
-
+	struct stop_machine_data *smdata = &idle;
+	int cpu = smp_processor_id();
+	int err;
+
+	if (!active_cpus) {
+		if (cpu == first_cpu(cpu_online_map))
+			smdata = &active;
+	} else {
+		if (cpu_isset(cpu, *active_cpus))
+			smdata = &active;
+	}
 	/* Simple state machine */
 	do {
 		/* Chill out and ensure we re-read stopmachine_state. */
@@ -78,9 +87,11 @@ static int stop_cpu(struct stop_machine_data *smdata)
 				hard_irq_disable();
 				break;
 			case STOPMACHINE_RUN:
-				/* |= allows error detection if functions on
-				 * multiple CPUs. */
-				smdata->fnret |= smdata->fn(smdata->data);
+				/* On multiple CPUs only a single error code
+				 * is needed to tell that something failed. */
+				err = smdata->fn(smdata->data);
+				if (err)
+					smdata->fnret = err;
 				break;
 			default:
 				break;
@@ -90,7 +101,6 @@ static int stop_cpu(struct stop_machine_data *smdata)
 	} while (curstate != STOPMACHINE_EXIT);
 
 	local_irq_enable();
-	do_exit(0);
 }
 
 /* Callback for CPUs which aren't supposed to do anything. */
@@ -101,78 +111,34 @@ static int chill(void *unused)
 
 int __stop_machine(int (*fn)(void *), void *data, const cpumask_t *cpus)
 {
-	int i, err;
-	struct stop_machine_data active, idle;
-	struct task_struct **threads;
+	struct work_struct *sm_work;
+	int i;
 
+	/* Set up initial state. */
+	mutex_lock(&lock);
+	num_threads = num_online_cpus();
+	active_cpus = cpus;
 	active.fn = fn;
 	active.data = data;
 	active.fnret = 0;
 	idle.fn = chill;
 	idle.data = NULL;
 
-	/* This could be too big for stack on large machines. */
-	threads = kcalloc(NR_CPUS, sizeof(threads[0]), GFP_KERNEL);
-	if (!threads)
-		return -ENOMEM;
-
-	/* Set up initial state. */
-	mutex_lock(&lock);
-	init_completion(&finished);
-	num_threads = num_online_cpus();
 	set_state(STOPMACHINE_PREPARE);
 
-	for_each_online_cpu(i) {
-		struct stop_machine_data *smdata = &idle;
-		struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 };
-
-		if (!cpus) {
-			if (i == first_cpu(cpu_online_map))
-				smdata = &active;
-		} else {
-			if (cpu_isset(i, *cpus))
-				smdata = &active;
-		}
-
-		threads[i] = kthread_create((void *)stop_cpu, smdata, "kstop%u",
-					    i);
-		if (IS_ERR(threads[i])) {
-			err = PTR_ERR(threads[i]);
-			threads[i] = NULL;
-			goto kill_threads;
-		}
-
-		/* Place it onto correct cpu. */
-		kthread_bind(threads[i], i);
-
-		/* Make it highest prio. */
-		if (sched_setscheduler_nocheck(threads[i], SCHED_FIFO, &param))
-			BUG();
-	}
-
-	/* We've created all the threads.  Wake them all: hold this CPU so one
+	/* Schedule the stop_cpu work on all cpus: hold this CPU so one
 	 * doesn't hit this CPU until we're ready. */
 	get_cpu();
-	for_each_online_cpu(i)
-		wake_up_process(threads[i]);
-
+	for_each_online_cpu(i) {
+		sm_work = percpu_ptr(stop_machine_work, i);
+		INIT_WORK(sm_work, stop_cpu);
+		queue_work_on(i, stop_machine_wq, sm_work);
+	}
 	/* This will release the thread on our CPU. */
 	put_cpu();
-	wait_for_completion(&finished);
+	flush_workqueue(stop_machine_wq);
 	mutex_unlock(&lock);
-
-	kfree(threads);
-
 	return active.fnret;
-
-kill_threads:
-	for_each_online_cpu(i)
-		if (threads[i])
-			kthread_stop(threads[i]);
-	mutex_unlock(&lock);
-
-	kfree(threads);
-	return err;
 }
 
 int stop_machine(int (*fn)(void *), void *data, const cpumask_t *cpus)
@@ -187,3 +153,11 @@ int stop_machine(int (*fn)(void *), void *data, const cpumask_t *cpus)
 	return ret;
 }
 EXPORT_SYMBOL_GPL(stop_machine);
+
+static int __init stop_machine_init(void)
+{
+	stop_machine_wq = create_rt_workqueue("kstop");
+	stop_machine_work = alloc_percpu(struct work_struct);
+	return 0;
+}
+early_initcall(stop_machine_init);
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 714afad..f928f2a 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -62,6 +62,7 @@ struct workqueue_struct {
 	const char *name;
 	int singlethread;
 	int freezeable;		/* Freeze threads during suspend */
+	int rt;
 #ifdef CONFIG_LOCKDEP
 	struct lockdep_map lockdep_map;
 #endif
@@ -766,6 +767,7 @@ init_cpu_workqueue(struct workqueue_struct *wq, int cpu)
 
 static int create_workqueue_thread(struct cpu_workqueue_struct *cwq, int cpu)
 {
+	struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 };
 	struct workqueue_struct *wq = cwq->wq;
 	const char *fmt = is_single_threaded(wq) ? "%s" : "%s/%d";
 	struct task_struct *p;
@@ -781,7 +783,8 @@ static int create_workqueue_thread(struct cpu_workqueue_struct *cwq, int cpu)
 	 */
 	if (IS_ERR(p))
 		return PTR_ERR(p);
-
+	if (cwq->wq->rt)
+		sched_setscheduler_nocheck(p, SCHED_FIFO, &param);
 	cwq->thread = p;
 
 	return 0;
@@ -801,6 +804,7 @@ static void start_workqueue_thread(struct cpu_workqueue_struct *cwq, int cpu)
 struct workqueue_struct *__create_workqueue_key(const char *name,
 						int singlethread,
 						int freezeable,
+						int rt,
 						struct lock_class_key *key,
 						const char *lock_name)
 {
@@ -822,6 +826,7 @@ struct workqueue_struct *__create_workqueue_key(const char *name,
 	lockdep_init_map(&wq->lockdep_map, lock_name, key, 0);
 	wq->singlethread = singlethread;
 	wq->freezeable = freezeable;
+	wq->rt = rt;
 	INIT_LIST_HEAD(&wq->list);
 
 	if (singlethread) {

Re: [PULL] module, param and stop_machine patches

[Linus Torvalds]

On Wed, 22 Oct 2008, Rusty Russell wrote:
> 
> Heiko Carstens (4):
>       Call init_workqueues before pre smp initcalls.

This one actually causes a regression for me on one of my machines.

Don't ask me why, but it's very repeatable, and it goes away when that 
commit is reverted. See the appended commit that I have queued up (but not 
yet pushed out) in my tree for details..

		Linus
---
From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Sat, 25 Oct 2008 14:04:40 -0700
Subject: [PATCH] Revert "Call init_workqueues before pre smp initcalls."

This reverts commit a802dd0eb5fc97a50cf1abb1f788a8f6cc5db635, which
causes an odd bootup failure on one of my machines.  The boot hangs
after printing out

	Found volume group "VolGroup00" using metadata type lvm2

where the next message (that never prints out) should be

	2 logical volume(s) in Volume Group"VolGroup00" now active.

Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/init/main.c b/init/main.c
index 130d1a0..70e4eed 100644
--- a/init/main.c
+++ b/init/main.c
@@ -768,6 +768,8 @@ static void __init do_initcalls(void)
 static void __init do_basic_setup(void)
 {
 	rcu_init_sched(); /* needed by module_init stage. */
+	/* drivers will send hotplug events */
+	init_workqueues();
 	usermodehelper_init();
 	driver_init();
 	init_irq_proc();
@@ -851,8 +853,6 @@ static int __init kernel_init(void * unused)
 
 	cad_pid = task_pid(current);
 
-	init_workqueues();
-
 	smp_prepare_cpus(setup_max_cpus);
 
 	do_pre_smp_initcalls();

Re: [PULL] module, param and stop_machine patches

[Rusty Russell]

On Sunday 26 October 2008 08:17:38 Linus Torvalds wrote:
> On Wed, 22 Oct 2008, Rusty Russell wrote:
> > Heiko Carstens (4):
> >       Call init_workqueues before pre smp initcalls.
>
> This one actually causes a regression for me on one of my machines.

Thanks, Heiko tracked this down; he's probably sleeping now but Hugh and Walt
reported this fixes it for them and it makes sense.

---
Subject: commit a802dd0e breaks console keyboard input
Date: Sat, 25 Oct 2008 16:16:22 +0200
From: Heiko Carstens <heiko.carstens@de.ibm.com>

The patch below moves the init call to its old place but makes the
stop_machine initialization happen later. I added a BUG_ON to catch possible
early calls, but there shouldn't be any. I think the patch should solve
the regression you are seeing. Could you give it a try please?
The patch applies on top of latest Linus' git tree.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
---
 include/linux/stop_machine.h |    6 ++++++
 init/main.c                  |    5 +++--
 kernel/stop_machine.c        |    9 ++++++---
 3 files changed, 15 insertions(+), 5 deletions(-)

Index: linux-2.6/include/linux/stop_machine.h
===================================================================
--- linux-2.6.orig/include/linux/stop_machine.h
+++ linux-2.6/include/linux/stop_machine.h
@@ -35,6 +35,9 @@ int stop_machine(int (*fn)(void *), void
  * won't come or go while it's being called.  Used by hotplug cpu.
  */
 int __stop_machine(int (*fn)(void *), void *data, const cpumask_t *cpus);
+
+void stop_machine_init(void);
+
 #else
 
 static inline int stop_machine(int (*fn)(void *), void *data,
@@ -46,5 +49,8 @@ static inline int stop_machine(int (*fn)
 	local_irq_enable();
 	return ret;
 }
+
+static inline void stop_machine_init(void) { }
+
 #endif /* CONFIG_SMP */
 #endif /* _LINUX_STOP_MACHINE */
Index: linux-2.6/init/main.c
===================================================================
--- linux-2.6.orig/init/main.c
+++ linux-2.6/init/main.c
@@ -38,6 +38,7 @@
 #include <linux/moduleparam.h>
 #include <linux/kallsyms.h>
 #include <linux/writeback.h>
+#include <linux/stop_machine.h>
 #include <linux/cpu.h>
 #include <linux/cpuset.h>
 #include <linux/cgroup.h>
@@ -768,6 +769,8 @@ static void __init do_initcalls(void)
 static void __init do_basic_setup(void)
 {
 	rcu_init_sched(); /* needed by module_init stage. */
+	init_workqueues();
+	stop_machine_init();
 	usermodehelper_init();
 	driver_init();
 	init_irq_proc();
@@ -851,8 +854,6 @@ static int __init kernel_init(void * unu
 
 	cad_pid = task_pid(current);
 
-	init_workqueues();
-
 	smp_prepare_cpus(setup_max_cpus);
 
 	do_pre_smp_initcalls();
Index: linux-2.6/kernel/stop_machine.c
===================================================================
--- linux-2.6.orig/kernel/stop_machine.c
+++ linux-2.6/kernel/stop_machine.c
@@ -43,6 +43,7 @@ static struct workqueue_struct *stop_mac
 static struct stop_machine_data active, idle;
 static const cpumask_t *active_cpus;
 static void *stop_machine_work;
+static int stop_machine_initialized;
 
 static void set_state(enum stopmachine_state newstate)
 {
@@ -114,6 +115,7 @@ int __stop_machine(int (*fn)(void *), vo
 	struct work_struct *sm_work;
 	int i;
 
+	BUG_ON(!stop_machine_initialized);
 	/* Set up initial state. */
 	mutex_lock(&lock);
 	num_threads = num_online_cpus();
@@ -154,10 +156,11 @@ int stop_machine(int (*fn)(void *), void
 }
 EXPORT_SYMBOL_GPL(stop_machine);
 
-static int __init stop_machine_init(void)
+void __init stop_machine_init(void)
 {
 	stop_machine_wq = create_rt_workqueue("kstop");
+	BUG_ON(!stop_machine_wq);
 	stop_machine_work = alloc_percpu(struct work_struct);
-	return 0;
+	BUG_ON(!stop_machine_work);
+	stop_machine_initialized = 1;
 }
-early_initcall(stop_machine_init);

Re: [PULL] module, param and stop_machine patches

[Linus Torvalds]

On Sun, 26 Oct 2008, Rusty Russell wrote:
> 
> Thanks, Heiko tracked this down; he's probably sleeping now but Hugh and Walt
> reported this fixes it for them and it makes sense.

I'm not seeing any "tracked it down".

Yes, it moves it back to its old place, but doesn't say why the new one 
didn't work.

And it then mixes things up with 'stop_machine_init()' mess. Why does that 
need to run so early?

Is there any reason why the real patch isn't just to make 
'stop_machine_init' a 'core_initcall()' instead of 'early_initcall()'?

IOW, I don't think that patch is anything but a "hey, test if it works 
with this". None of the changes or the problems are explained. Nor do I 
see a sign-off from Heiko on it.

I also don't want to see more BUG_ON()'s there. Make them warnings or 
something.

		Linus

Re: [PULL] module, param and stop_machine patches

[Linus Torvalds]

On Sat, 25 Oct 2008, Linus Torvalds wrote:
> 
> Is there any reason why the real patch isn't just to make 
> 'stop_machine_init' a 'core_initcall()' instead of 'early_initcall()'?

IOW, something like this (untested)

		Linus

---
 init/main.c           |    3 +--
 kernel/stop_machine.c |    2 +-
 2 files changed, 2 insertions(+), 3 deletions(-)

diff --git a/init/main.c b/init/main.c
index 130d1a0..7e117a2 100644
--- a/init/main.c
+++ b/init/main.c
@@ -768,6 +768,7 @@ static void __init do_initcalls(void)
 static void __init do_basic_setup(void)
 {
 	rcu_init_sched(); /* needed by module_init stage. */
+	init_workqueues();
 	usermodehelper_init();
 	driver_init();
 	init_irq_proc();
@@ -851,8 +852,6 @@ static int __init kernel_init(void * unused)
 
 	cad_pid = task_pid(current);
 
-	init_workqueues();
-
 	smp_prepare_cpus(setup_max_cpus);
 
 	do_pre_smp_initcalls();
diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c
index 8aff79d..9bc4c00 100644
--- a/kernel/stop_machine.c
+++ b/kernel/stop_machine.c
@@ -160,4 +160,4 @@ static int __init stop_machine_init(void)
 	stop_machine_work = alloc_percpu(struct work_struct);
 	return 0;
 }
-early_initcall(stop_machine_init);
+core_initcall(stop_machine_init);

Re: [PULL] module, param and stop_machine patches

[Linus Torvalds]

On Sat, 25 Oct 2008, Linus Torvalds wrote:
>
> -early_initcall(stop_machine_init);
> +core_initcall(stop_machine_init);

Btw, this obviously implies that all stop_machine() users had better be 
later in the boot sequence.

And no, I didn't check that exhaustively. There's the tracing 
infrastructure, and it's also core_initcall(), but it should be linked 
later. And module loading will happen later. mm/page_alloc.c has special 
code for bootup, and intel-rng.c should be a normal driver.

So it _should_ be ok. But I really didn't test. And my default kernel 
config doesn't even have CONFIG_STOP_MACHINE enabled, so I'm not the best 
person to test. 

		Linus

Re: [PULL] module, param and stop_machine patches

[walt]

Linus Torvalds wrote:
>
> On Sat, 25 Oct 2008, Linus Torvalds wrote:
>> Is there any reason why the real patch isn't just to make
>> 'stop_machine_init' a 'core_initcall()' instead of 'early_initcall()'?
>
> IOW, something like this (untested)
>
> 		Linus

<patch snipped for brevity>

This patch also fixes the keyboard problem I reported originally.

Re: [PULL] module, param and stop_machine patches

[Rusty Russell]

On Sunday 26 October 2008 09:33:43 Linus Torvalds wrote:
> On Sun, 26 Oct 2008, Rusty Russell wrote:
> > Thanks, Heiko tracked this down; he's probably sleeping now but Hugh and
> > Walt reported this fixes it for them and it makes sense.
>
> I'm not seeing any "tracked it down".

He tracked it down to moving init_workqueues() too early, so he moved that
back.

> And it then mixes things up with 'stop_machine_init()' mess. Why does that
> need to run so early?

The S/390 guys want to run it stop_machine v. early, so when Heiko introduced
stop_machine_init() he made it an early_initcall().

> IOW, I don't think that patch is anything but a "hey, test if it works
> with this". None of the changes or the problems are explained.

Indeed.

Turns out it's the cpu_online_map difference.  If init_workqueues() is called
too early, only the boot cpu is set.  We then only create_workqueue_thread()
for the boot cpu.

If CONFIG_HOTPLUG_CPU=y, it's fine since the hotplug callback will create the
workqueue threads for the other cpus as they come up.  Without it, the kevent
workqueues on non-boot cpus don't get processed.

Still boots for me, but was a bit sick (varying, but no keyboard was one
symptom).

> Nor do I see a sign-off from Heiko on it.

I thought mine would be sufficient since we both work for IBM?

> I also don't want to see more BUG_ON()'s there. Make them warnings or
> something.

OK, I just killed them.  It'll crash anyway.

Rusty.

Subject: Fix boot problems caused by moving init_workqueues()
Date: Sat, 25 Oct 2008 16:16:22 +0200
From: Heiko Carstens <heiko.carstens@de.ibm.com>

The patch below moves the init_workqueues() call to its old place but
makes the stop_machine initialization happen later (but still before
core_initcalls).

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
---
 include/linux/stop_machine.h |    6 ++++++
 init/main.c                  |    5 +++--
 kernel/stop_machine.c        |    9 ++++++---
 3 files changed, 15 insertions(+), 5 deletions(-)

diff -r 2b5f764088ee include/linux/stop_machine.h
--- a/include/linux/stop_machine.h	Sun Oct 26 18:50:23 2008 +1100
+++ b/include/linux/stop_machine.h	Sun Oct 26 19:13:04 2008 +1100
@@ -35,6 +35,9 @@ int stop_machine(int (*fn)(void *), void
  * won't come or go while it's being called.  Used by hotplug cpu.
  */
 int __stop_machine(int (*fn)(void *), void *data, const cpumask_t *cpus);
+
+void stop_machine_init(void);
+
 #else
 
 static inline int stop_machine(int (*fn)(void *), void *data,
@@ -46,5 +49,8 @@ static inline int stop_machine(int (*fn)
 	local_irq_enable();
 	return ret;
 }
+
+static inline void stop_machine_init(void) { }
+
 #endif /* CONFIG_SMP */
 #endif /* _LINUX_STOP_MACHINE */
diff -r 2b5f764088ee init/main.c
--- a/init/main.c	Sun Oct 26 18:50:23 2008 +1100
+++ b/init/main.c	Sun Oct 26 19:13:04 2008 +1100
@@ -38,6 +38,7 @@
 #include <linux/moduleparam.h>
 #include <linux/kallsyms.h>
 #include <linux/writeback.h>
+#include <linux/stop_machine.h>
 #include <linux/cpu.h>
 #include <linux/cpuset.h>
 #include <linux/cgroup.h>
@@ -767,6 +768,8 @@ static void __init do_basic_setup(void)
 static void __init do_basic_setup(void)
 {
 	rcu_init_sched(); /* needed by module_init stage. */
+	init_workqueues();
+	stop_machine_init();
 	usermodehelper_init();
 	driver_init();
 	init_irq_proc();
@@ -850,8 +853,6 @@ static int __init kernel_init(void * unu
 
 	cad_pid = task_pid(current);
 
-	init_workqueues();
-
 	smp_prepare_cpus(setup_max_cpus);
 
 	do_pre_smp_initcalls();
diff -r 2b5f764088ee kernel/stop_machine.c
--- a/kernel/stop_machine.c	Sun Oct 26 18:50:23 2008 +1100
+++ b/kernel/stop_machine.c	Sun Oct 26 19:13:04 2008 +1100
@@ -154,10 +154,8 @@ int stop_machine(int (*fn)(void *), void
 }
 EXPORT_SYMBOL_GPL(stop_machine);
 
-static int __init stop_machine_init(void)
+void __init stop_machine_init(void)
 {
 	stop_machine_wq = create_rt_workqueue("kstop");
 	stop_machine_work = alloc_percpu(struct work_struct);
-	return 0;
 }
-early_initcall(stop_machine_init);

Re: [PULL] module, param and stop_machine patches

[Heiko Carstens]

On Sun, Oct 26, 2008 at 07:16:16PM +1100, Rusty Russell wrote:
> On Sunday 26 October 2008 09:33:43 Linus Torvalds wrote:
> > On Sun, 26 Oct 2008, Rusty Russell wrote:
> > > Thanks, Heiko tracked this down; he's probably sleeping now but Hugh and
> > > Walt reported this fixes it for them and it makes sense.
> > I'm not seeing any "tracked it down".
> He tracked it down to moving init_workqueues() too early, so he moved that
> back.
> 
> > And it then mixes things up with 'stop_machine_init()' mess. Why does that
> > need to run so early?
> 
> The S/390 guys want to run it stop_machine v. early, so when Heiko introduced
> stop_machine_init() he made it an early_initcall().
> 
> > IOW, I don't think that patch is anything but a "hey, test if it works
> > with this". None of the changes or the problems are explained.
> 
> Indeed.
> 
> Turns out it's the cpu_online_map difference.  If init_workqueues() is called
> too early, only the boot cpu is set.  We then only create_workqueue_thread()
> for the boot cpu.
> 
> If CONFIG_HOTPLUG_CPU=y, it's fine since the hotplug callback will create the
> workqueue threads for the other cpus as they come up.  Without it, the kevent
> workqueues on non-boot cpus don't get processed.
> 
> Still boots for me, but was a bit sick (varying, but no keyboard was one
> symptom).

Yes, it's all my fault. I always think in terms of CONFIG_HOTPLUG_CPU=y, so
I couldn't make any sense of the bug reports and just reverted the
init_workqueues() call move and added an explicit stop_machine_init() call,
so that we don't depend on linkage order.
Thanks for tracking it down, Rusty!

> > Nor do I see a sign-off from Heiko on it.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>

But I guess you don't need that anymore since you already committed a
fix for this.

Thanks,
Heiko

Re: [PULL] module, param and stop_machine patches

[Linus Torvalds]

On Sun, 26 Oct 2008, Rusty Russell wrote:
> On Sunday 26 October 2008 09:33:43 Linus Torvalds wrote:
> >
> > I'm not seeing any "tracked it down".
> 
> He tracked it down to moving init_workqueues() too early, so he moved that
> back.

Yeah. I had done as well by my bisection. One of the reasons I didn't push 
out my initial fix (just the plain revert) was exactly the fact that I 
don't think that is sufficient. I want to _understand_ what was going on, 
and get a feel from the commit message that makes me go "yeah, ok, makes 
sense".

> Turns out it's the cpu_online_map difference.  If init_workqueues() is called
> too early, only the boot cpu is set.  We then only create_workqueue_thread()
> for the boot cpu.

Indeed. I figured it out yesterday, since I didn't get an explanation soon 
enough for my impatient self. So I ended up committing the fix as the 
minimal patch I sent out earlier, with a fairly minimal explanation. I'll 
happily take more fixes, but I think it should be ok.

If S390 wants to do stop_machine() even before some of the basic 
core_initcalls, I suspect it will have to change, but nothing I saw 
indicated that it really needs to.

> If CONFIG_HOTPLUG_CPU=y, it's fine since the hotplug callback will create the
> workqueue threads for the other cpus as they come up.  Without it, the kevent
> workqueues on non-boot cpus don't get processed.

.. and my default workstation config indeed hit this, since I normally 
have a fairly minimal setup.

> > Nor do I see a sign-off from Heiko on it.
> 
> I thought mine would be sufficient since we both work for IBM?

Technically, yes, since you all sign your life away and do this for work, 
but he normally signs off his patches, so the lack of sign-off didn't make 
me go "hmm, this looks odd from a legal copyright standpoint", but rather 
"hmm, maybe Heiko doesn't feel comfy with the patch since he didn't sign 
off".

To me, the legal side of the sign-off is actually secondary. The whole 
thing was designed to be useful even in the absense of it. I've grown very 
fond of seeing who has seen the patch (including who _should_ have seen it 
but didn't react - the "cc" part ;)

			Linus