qsnet-suse-2.6.patch

非常经典的一个分布式系统

+/* IOPROC SYNC PAGE
+ *
+ * Called when a memory map is synchronised with its disk image i.e. when the 
+ * msync() syscall is invoked. Any future read or write to the associated page
+ * by the IOPROC should cause the page to be marked as referenced or modified.
+ *
+ * Not currently called as msync() calls ioproc_sync_range() instead
+ *
+ * Called holding the mm->page_table_lock
+ */
+static inline void 
+ioproc_sync_page(struct vm_area_struct *vma, unsigned long addr)
+{
+	struct ioproc_ops *cp;
+
+	for (cp = vma->vm_mm->ioproc_ops; cp; cp = cp->next)
+		if (cp->sync_page)
+			cp->sync_page(cp->arg, vma, addr);
+}
+
+/* IOPROC INVALIDATE PAGE
+ *
+ * Called whenever a valid PTE is unloaded e.g. when a page is unmapped by the
+ * user or paged out by the kernel. 
+ *
+ * After this call the IOPROC must not access the physical memory again unless
+ * a new translation is loaded.
+ *
+ * Called holding the mm->page_table_lock
+ */
+static inline void 
+ioproc_invalidate_page(struct vm_area_struct *vma, unsigned long addr)
+{
+	struct ioproc_ops *cp;
+
+	for (cp = vma->vm_mm->ioproc_ops; cp; cp = cp->next)
+		if (cp->invalidate_page)
+			cp->invalidate_page(cp->arg, vma, addr);
+}
+
+/* IOPROC UPDATE PAGE
+ *
+ * Called whenever a valid PTE is loaded e.g. mmaping memory, moving the brk 
+ * up, when breaking COW or faulting in an anoymous page of memory.
+ *
+ * These give the IOPROC device the opportunity to load translations 
+ * speculatively, which can improve performance by avoiding device translation
+ * faults.
+ *
+ * Called holding the mm->page_table_lock
+ */
+static inline void 
+ioproc_update_page(struct vm_area_struct *vma, unsigned long addr)
+{
+	struct ioproc_ops *cp;
+
+	for (cp = vma->vm_mm->ioproc_ops; cp; cp = cp->next)
+		if (cp->update_page)
+			cp->update_page(cp->arg, vma, addr);
+}
+
+#else
+
+/* ! CONFIG_IOPROC so make all hooks empty */
+
+#define ioproc_release(mm)			do { } while (0)
+
+#define ioproc_sync_range(vma,start,end)		do { } while (0)
+
+#define ioproc_invalidate_range(vma, start,end)	do { } while (0)
+
+#define ioproc_update_range(vma, start, end)	do { } while (0)
+
+#define ioproc_change_protection(vma, start, end, prot)	do { } while (0)
+
+#define ioproc_sync_page(vma, addr)		do { } while (0)
+
+#define ioproc_invalidate_page(vma, addr)	do { } while (0)
+
+#define ioproc_update_page(vma, addr)		do { } while (0)
+
+#endif /* CONFIG_IOPROC */
+
+#endif /* __LINUX_IOPROC_H__ */
Index: LINUX-SRC-TREE/include/linux/ptrack.h
===================================================================
--- /dev/null
+++ LINUX-SRC-TREE/include/linux/ptrack.h
@@ -0,0 +1,65 @@
+/*
+ *    Copyright (C) 2000  Regents of the University of California
+ *
+ *    This program is free software; you can redistribute it and/or modify
+ *    it under the terms of the GNU General Public License as published by
+ *    the Free Software Foundation; either version 2 of the License, or
+ *    (at your option) any later version.
+ *
+ *    This program is distributed in the hope that it will be useful,
+ *    but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *    GNU General Public License for more details.
+ *
+ *    You should have received a copy of the GNU General Public License
+ *    along with this program; if not, write to the Free Software
+ *    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ * Derived from exit_actn.c by
+ *    Copyright (C) 2003 Quadrics Ltd.
+ *
+ */
+#ifndef __LINUX_PTRACK_H
+#define __LINUX_PTRACK_H
+
+/* 
+ * Process tracking - this allows a module to keep track of processes
+ * in order that it can manage all tasks derived from a single process.
+ */
+
+#define PTRACK_PHASE_CLONE	1
+#define PTRACK_PHASE_CLONE_FAIL	2
+#define PTRACK_PHASE_EXEC	3
+#define PTRACK_PHASE_EXIT      	4
+
+#define PTRACK_FINISHED		0
+#define PTRACK_INNHERIT		1
+#define PTRACK_DENIED		2
+
+#ifdef CONFIG_PTRACK
+
+typedef int (*ptrack_callback_t)(void *arg, int phase, struct task_struct *child);
+
+struct ptrack_desc {
+       struct list_head        link;
+       ptrack_callback_t       callback;
+       void                   *arg;
+};
+
+extern int     ptrack_register (ptrack_callback_t callback, void *arg);
+extern void    ptrack_deregister (ptrack_callback_t callback, void *arg);
+extern int     ptrack_registered (ptrack_callback_t callback, void *arg);
+
+extern int     ptrack_call_callbacks (int phase, struct task_struct *child);
+
+#define INIT_TASK_PTRACK(tsk) \
+	.ptrack_list = LIST_HEAD_INIT(tsk.ptrack_list)
+
+#else
+#define ptrack_call_callbacks (phase, child) (0)
+
+#define INIT_TASK_PTRACK(tsk)
+
+#endif
+
+#endif /* __LINUX_PTRACK_H */
Index: LINUX-SRC-TREE/include/linux/sched.h
===================================================================
--- LINUX-SRC-TREE.orig/include/linux/sched.h
+++ LINUX-SRC-TREE/include/linux/sched.h
@@ -188,6 +188,9 @@ asmlinkage void schedule(void);
 extern int max_timeslice, min_timeslice;
 
 struct namespace;
+#ifdef CONFIG_IOPROC
+struct ioproc_ops;
+#endif
 
 /* Maximum number of active map areas.. This is a random (large) number */
 #define DEFAULT_MAX_MAP_COUNT	65536
@@ -241,6 +244,11 @@ struct mm_struct {
 	struct kioctx		default_kioctx;
 
 	unsigned long hiwater_rss, hiwater_vm;
+
+#ifdef CONFIG_IOPROC
+	/* hooks for io devices with advanced RDMA capabilities */
+	struct ioproc_ops	*ioproc_ops;
+#endif
 };
 
 extern int mmlist_nr;
@@ -603,6 +611,10 @@ struct task_struct {
 	struct rw_semaphore pagg_sem;
 #endif
 
+#ifdef CONFIG_PTRACK
+/* process tracking callback */
+	struct list_head ptrack_list;
+#endif
 };
 
 static inline pid_t process_group(struct task_struct *tsk)
Index: LINUX-SRC-TREE/ipc/shm.c
===================================================================
--- LINUX-SRC-TREE.orig/ipc/shm.c
+++ LINUX-SRC-TREE/ipc/shm.c
@@ -27,6 +27,7 @@
 #include <linux/shmem_fs.h>
 #include <linux/security.h>
 #include <linux/audit.h>
+#include <linux/module.h>
 #include <linux/trigevent_hooks.h>
 #include <asm/uaccess.h>
 
@@ -879,6 +880,44 @@ asmlinkage long sys_shmdt(char __user *s
 	return audit_result(retval);
 }
 
+/*
+ * Mark all segments created by this process for destruction
+ */
+int shm_cleanup (void)
+{
+	int i;
+
+	down(&shm_ids.sem);
+
+	for (i = 0; i <= shm_ids.max_id; i++) {
+		struct shmid_kernel *shp;
+
+		shp = shm_lock(i);
+		if (shp != NULL) {
+			/* mark this segment for destruction if we created it */
+			if (current->pid == shp->shm_cprid)
+			{
+				/* copy of IPC_RMID code */
+				if (shp->shm_nattch) {
+					shp->shm_flags |= SHM_DEST;
+					/* do not find it any more */
+					shp->shm_perm.key = IPC_PRIVATE;
+				} else {
+					shm_destroy(shp);
+					continue;
+				}
+			}
+
+			shm_unlock(shp);
+		}
+	}
+
+	up(&shm_ids.sem);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(shm_cleanup);
+
 #ifdef CONFIG_PROC_FS
 static int sysvipc_shm_read_proc(char *buffer, char **start, off_t offset, int length, int *eof, void *data)
 {
Index: LINUX-SRC-TREE/kernel/exit.c
===================================================================
--- LINUX-SRC-TREE.orig/kernel/exit.c
+++ LINUX-SRC-TREE/kernel/exit.c
@@ -40,6 +40,8 @@
 /* tng related changes */
 int (*tng_exitfunc)(int) = NULL;
 
+#include <linux/ptrack.h>
+
 extern void sem_exit (void);
 extern struct task_struct *child_reaper;
 void (*do_eop_acct) (int, struct task_struct *);
@@ -848,6 +850,8 @@ asmlinkage NORET_TYPE void do_exit(long 
 		audit_exit(tsk, code);
 	audit_free(tsk->audit);
 #endif
+ 	/* Notify any ptrack callbacks of the process exit */
+ 	ptrack_call_callbacks (PTRACK_PHASE_EXIT, NULL);
 	__exit_mm(tsk);
 
 	if (unlikely(tng_exitfunc))
Index: LINUX-SRC-TREE/kernel/fork.c
===================================================================
--- LINUX-SRC-TREE.orig/kernel/fork.c
+++ LINUX-SRC-TREE/kernel/fork.c
@@ -14,6 +14,7 @@
 #include <linux/config.h>
 #include <linux/slab.h>
 #include <linux/init.h>
+#include <linux/ptrack.h>
 #include <linux/unistd.h>
 #include <linux/smp_lock.h>
 #include <linux/module.h>
@@ -432,6 +433,9 @@ static struct mm_struct * mm_init(struct
 	mm->page_table_lock = SPIN_LOCK_UNLOCKED;
 	mm->ioctx_list_lock = RW_LOCK_UNLOCKED;
 	mm->ioctx_list = NULL;
+#ifdef CONFIG_IOPROC
+	mm->ioproc_ops = NULL;
+#endif
 	mm->default_kioctx = (struct kioctx)INIT_KIOCTX(mm->default_kioctx, *mm);
 	mm->free_area_cache = TASK_UNMAPPED_BASE;
 
@@ -1276,6 +1280,11 @@ long do_fork(unsigned long clone_flags,
          	      audit_fork(current, p);
 #endif
 
+		if (ptrack_call_callbacks(PTRACK_PHASE_CLONE, p)) {
+			sigaddset(&p->pending.signal, SIGKILL);
+			set_tsk_thread_flag(p, TIF_SIGPENDING);
+		}
+
 		/* Trace the event  */
 		TRIG_EVENT(fork_hook, clone_flags, p, pid);
 		if (!(clone_flags & CLONE_STOPPED)) {
Index: LINUX-SRC-TREE/kernel/Kconfig
===================================================================
--- /dev/null
+++ LINUX-SRC-TREE/kernel/Kconfig
@@ -0,0 +1,14 @@
+#
+# Kernel subsystem specific config
+# 
+
+# Support for Process Tracking callbacks
+#
+config PTRACK
+	bool "Enable PTRACK process tracking hooks"
+	default y
+	help
+	This option enables hooks to be called when processes are
+	created and destoryed in order for a resource management 
+	system to know which processes are a member of a "job" and 
+	to be able to clean up when the job is terminated.
Index: LINUX-SRC-TREE/kernel/Makefile
===================================================================
--- LINUX-SRC-TREE.orig/kernel/Makefile
+++ LINUX-SRC-TREE/kernel/Makefile
@@ -29,6 +29,7 @@ obj-$(CONFIG_LTT) += ltt/
 obj-$(CONFIG_KPROBES) += kprobes.o
 obj-$(CONFIG_CPUSETS) += cpuset.o
 obj-$(CONFIG_CKRM_CPU_SCHEDULE) += ckrm_classqueue.o ckrm_sched.o
+obj-$(CONFIG_PTRACK) += ptrack.o
 
 ifneq ($(CONFIG_IA64),y)
 # According to Alan Modra <[e-mail]>, the -fno-omit-frame-pointer is
Index: LINUX-SRC-TREE/kernel/ptrack.c
===================================================================
--- /dev/null
+++ LINUX-SRC-TREE/kernel/ptrack.c
@@ -0,0 +1,145 @@
+/*
+ *    Copyright (C) 2000  Regents of the University of California
+ *
+ *    This program is free software; you can redistribute it and/or modify
+ *    it under the terms of the GNU General Public License as published by
+ *    the Free Software Foundation; either version 2 of the License, or
+ *    (at your option) any later version.
+ *
+ *    This program is distributed in the hope that it will be useful,
+ *    but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *    GNU General Public License for more details.
+ *
+ *    You should have received a copy of the GNU General Public License
+ *    along with this program; if not, write to the Free Software
+ *    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ * Derived from exit_actn.c by
+ *    Copyright (C) 2003 Quadrics Ltd.
+ */
+
+
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/sched.h>
+#include <linux/ptrack.h>
+#include <linux/slab.h>
+#include <linux/list.h>
+
+#include <asm/errno.h>
+
+int
+ptrack_register (ptrack_callback_t callback, void *arg)
+{
+       struct ptrack_desc *desc = kmalloc (sizeof (struct ptrack_desc), GFP_KERNEL);
+       
+       if (desc == NULL)
+               return -ENOMEM;
+
+	desc->callback = callback;
+	desc->arg      = arg;
+       
+       list_add_tail (&desc->link, &current->ptrack_list);
+       
+       return 0;
+}
+
+void
+ptrack_deregister (ptrack_callback_t callback, void *arg)
+{      
+       struct list_head *el, *nel;
+       
+       list_for_each_safe (el, nel, &current->ptrack_list) {
+               struct ptrack_desc *desc = list_entry (el, struct ptrack_desc, link);
+               
+               if (desc->callback == callback && desc->arg == arg) {
+                       list_del (&desc->link);
+                       kfree (desc);
+               }
+       }
+}
+
+int
+ptrack_registered (ptrack_callback_t callback, void *arg)
+{
+       struct list_head *el;
+       
+       list_for_each (el, &current->ptrack_list) {
+               struct ptrack_desc *desc = list_entry (el, struct ptrack_desc, link);
+               
+               if (desc->callback == callback && desc->arg == arg)
+                       return 1;
+       }
+       return 0;
+}      
+        
+int
+ptrack_call_callbacks (int phase, struct task_struct *child)
+{
+       struct list_head *el, *nel;
+       struct ptrack_desc *new;
+	int res;
+