qsnet-rhel4-2.6.patch

非常经典的一个分布式系统

+-- Last update Daniel J Blueman - 24 Mar 2006
Index: linux-269-5502/mm/ioproc.c
===================================================================
--- /dev/null
+++ linux-269-5502/mm/ioproc.c
@@ -0,0 +1,52 @@
+/*
+ *    Copyright (C) 2006 Quadrics Ltd
+ *
+ *    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
+ */
+
+/*
+ * Registration for IO processor page table updates.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+
+#include <linux/mm.h>
+#include <linux/ioproc.h>
+
+int ioproc_register_ops(struct mm_struct *mm, struct ioproc_ops *ip)
+{
+	ip->next = mm->ioproc_ops;
+	mm->ioproc_ops = ip;
+
+	return 0;
+}
+
+EXPORT_SYMBOL_GPL(ioproc_register_ops);
+
+int ioproc_unregister_ops(struct mm_struct *mm, struct ioproc_ops *ip)
+{
+	struct ioproc_ops **tmp;
+
+	for (tmp = &mm->ioproc_ops; *tmp && *tmp != ip; tmp = &(*tmp)->next) ;
+	if (*tmp) {
+		*tmp = ip->next;
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+EXPORT_SYMBOL_GPL(ioproc_unregister_ops);
Index: linux-269-5502/mm/hugetlb.c
===================================================================
--- linux-269-5502.orig/mm/hugetlb.c
+++ linux-269-5502/mm/hugetlb.c
@@ -10,6 +10,7 @@
 #include <linux/hugetlb.h>
 #include <linux/sysctl.h>
 #include <linux/highmem.h>
+#include <linux/ioproc.h>
 
 const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL;
 static unsigned long nr_huge_pages, free_huge_pages;
@@ -260,6 +261,7 @@ void zap_hugepage_range(struct vm_area_s
 	struct mm_struct *mm = vma->vm_mm;
 
 	spin_lock(&mm->page_table_lock);
+	ioproc_invalidate_range(vma, start, start + length);
 	unmap_hugepage_range(vma, start, start + length);
 	spin_unlock(&mm->page_table_lock);
 }
Index: linux-269-5502/mm/Kconfig
===================================================================
--- /dev/null
+++ linux-269-5502/mm/Kconfig
@@ -0,0 +1,15 @@
+#
+# VM subsystem specific config
+# 
+
+# Support for IO processors which have advanced RDMA capabilities
+#
+config IOPROC
+	bool "Enable IOPROC VM hooks"
+	depends on MMU
+	default y
+	help
+	This option enables hooks in the VM subsystem so that IO devices which
+	incorporate advanced RDMA capabilities can be kept in sync with CPU 
+	page table changes.
+	See Documentation/vm/ioproc.txt for more details.
Index: linux-269-5502/mm/Makefile
===================================================================
--- linux-269-5502.orig/mm/Makefile
+++ linux-269-5502/mm/Makefile
@@ -16,6 +16,7 @@ obj-$(CONFIG_SWAP)	+= page_io.o swap_sta
 obj-$(CONFIG_X86_4G)	+= usercopy.o
 obj-$(CONFIG_HUGETLBFS)	+= hugetlb.o
 obj-$(CONFIG_NUMA) 	+= mempolicy.o
+obj-$(CONFIG_IOPROC)    += ioproc.o
 obj-$(CONFIG_SHMEM) += shmem.o
 obj-$(CONFIG_TINY_SHMEM) += tiny-shmem.o
 
Index: linux-269-5502/mm/mprotect.c
===================================================================
--- linux-269-5502.orig/mm/mprotect.c
+++ linux-269-5502/mm/mprotect.c
@@ -10,6 +10,7 @@
 
 #include <linux/mm.h>
 #include <linux/hugetlb.h>
+#include <linux/ioproc.h>
 #include <linux/slab.h>
 #include <linux/shm.h>
 #include <linux/mman.h>
@@ -100,6 +101,7 @@ change_protection(struct vm_area_struct 
 	if (start >= end)
 		BUG();
 	spin_lock(&current->mm->page_table_lock);
+	ioproc_change_protection(vma, start, end, newprot);
 	do {
 		change_pmd_range(dir, start, end - start, newprot);
 		start = (start + PGDIR_SIZE) & PGDIR_MASK;
Index: linux-269-5502/mm/msync.c
===================================================================
--- linux-269-5502.orig/mm/msync.c
+++ linux-269-5502/mm/msync.c
@@ -12,6 +12,7 @@
 #include <linux/mm.h>
 #include <linux/mman.h>
 #include <linux/hugetlb.h>
+#include <linux/ioproc.h>
 
 #include <asm/pgtable.h>
 #include <asm/tlbflush.h>
@@ -115,6 +116,7 @@ static int filemap_sync(struct vm_area_s
 
 	if (address >= end)
 		BUG();
+	ioproc_sync_range(vma, address, end);
 	do {
 		error |= filemap_sync_pmd_range(dir, address, end, vma, flags);
 		address = (address + PGDIR_SIZE) & PGDIR_MASK;
Index: linux-269-5502/mm/mremap.c
===================================================================
--- linux-269-5502.orig/mm/mremap.c
+++ linux-269-5502/mm/mremap.c
@@ -9,6 +9,7 @@
 
 #include <linux/mm.h>
 #include <linux/hugetlb.h>
+#include <linux/ioproc.h>
 #include <linux/slab.h>
 #include <linux/shm.h>
 #include <linux/mman.h>
@@ -148,6 +149,8 @@ static unsigned long move_page_tables(st
 {
 	unsigned long offset;
 
+	ioproc_invalidate_range(vma, old_addr, old_addr + len);
+	ioproc_invalidate_range(vma, new_addr, new_addr + len);
 	flush_cache_range(vma, old_addr, old_addr + len);
 
 	/*
Index: linux-269-5502/mm/fremap.c
===================================================================
--- linux-269-5502.orig/mm/fremap.c
+++ linux-269-5502/mm/fremap.c
@@ -12,6 +12,7 @@
 #include <linux/mman.h>
 #include <linux/pagemap.h>
 #include <linux/swapops.h>
+#include <linux/ioproc.h>
 #include <linux/rmap.h>
 #include <linux/module.h>
 
@@ -29,6 +30,7 @@ static inline void zap_pte(struct mm_str
 	if (pte_present(pte)) {
 		unsigned long pfn = pte_pfn(pte);
 
+		ioproc_invalidate_page(vma, addr);
 		flush_cache_page(vma, addr);
 		pte = ptep_clear_flush(vma, addr, ptep);
 		if (pfn_valid(pfn)) {
@@ -93,6 +95,7 @@ int install_page(struct mm_struct *mm, s
 	pte_val = *pte;
 	pte_unmap(pte);
 	update_mmu_cache(vma, addr, pte_val);
+	ioproc_update_page(vma, addr);
 
 	err = 0;
 err_unlock:
@@ -132,6 +135,7 @@ int install_file_pte(struct mm_struct *m
 	pte_val = *pte;
 	pte_unmap(pte);
 	update_mmu_cache(vma, addr, pte_val);
+	ioproc_update_page(vma, addr);
 	spin_unlock(&mm->page_table_lock);
 	return 0;
 
Index: linux-269-5502/mm/rmap.c
===================================================================
--- linux-269-5502.orig/mm/rmap.c
+++ linux-269-5502/mm/rmap.c
@@ -51,6 +51,7 @@
 #include <linux/slab.h>
 #include <linux/init.h>
 #include <linux/rmap.h>
+#include <linux/ioproc.h>
 #include <linux/rcupdate.h>
 
 #include <asm/tlbflush.h>
@@ -566,6 +567,7 @@ static int try_to_unmap_one(struct page 
 	}
 
 	/* Nuke the page table entry. */
+	ioproc_invalidate_page(vma, address);
 	flush_cache_page(vma, address);
 	pteval = ptep_clear_flush(vma, address, pte);
 
@@ -673,6 +675,7 @@ static void try_to_unmap_cluster(unsigne
 			continue;
 
 		/* Nuke the page table entry. */
+		ioproc_invalidate_page(vma, address);
 		flush_cache_page(vma, address);
 		pteval = ptep_clear_flush(vma, address, pte);
 
Index: linux-269-5502/mm/memory.c
===================================================================
--- linux-269-5502.orig/mm/memory.c
+++ linux-269-5502/mm/memory.c
@@ -43,6 +43,7 @@
 #include <linux/swap.h>
 #include <linux/highmem.h>
 #include <linux/pagemap.h>
+#include <linux/ioproc.h>
 #include <linux/rmap.h>
 #include <linux/module.h>
 #include <linux/init.h>
@@ -630,6 +631,7 @@ void zap_page_range(struct vm_area_struc
 
 	lru_add_drain();
 	spin_lock(&mm->page_table_lock);
+ 	ioproc_invalidate_range(vma, address, end);
 	tlb = tlb_gather_mmu(mm, 0);
 	unmap_vmas(&tlb, mm, vma, address, end, &nr_accounted, details);
 	tlb_finish_mmu(tlb, address, end);
@@ -998,6 +1000,7 @@ int zeromap_page_range(struct vm_area_st
 		BUG();
 
 	spin_lock(&mm->page_table_lock);
+	ioproc_invalidate_range(vma, beg, end);
 	do {
 		pmd_t *pmd = pmd_alloc(mm, dir, address);
 		error = -ENOMEM;
@@ -1012,6 +1015,7 @@ int zeromap_page_range(struct vm_area_st
 	/*
 	 * Why flush? zeromap_pte_range has a BUG_ON for !pte_none()
 	 */
+	ioproc_update_range(vma, beg, end);
 	flush_tlb_range(vma, beg, end);
 	spin_unlock(&mm->page_table_lock);
 	return error;
@@ -1092,6 +1096,7 @@ int remap_page_range(struct vm_area_stru
 	vma->vm_flags |= VM_IO | VM_RESERVED;
 
 	spin_lock(&mm->page_table_lock);
+	ioproc_invalidate_range(vma, beg, end);
 	do {
 		pmd_t *pmd = pmd_alloc(mm, dir, from);
 		error = -ENOMEM;
@@ -1106,6 +1111,7 @@ int remap_page_range(struct vm_area_stru
 	/*
 	 * Why flush? remap_pte_range has a BUG_ON for !pte_none()
 	 */
+	ioproc_update_range(vma, beg, end);
 	flush_tlb_range(vma, beg, end);
 	spin_unlock(&mm->page_table_lock);
 	return error;
@@ -1194,6 +1200,7 @@ static int do_wp_page(struct mm_struct *
 			update_mmu_cache(vma, address, entry);
 			lazy_mmu_prot_update(entry);
 			pte_unmap(page_table);
+			ioproc_update_page(vma, address);
 			spin_unlock(&mm->page_table_lock);
 			return VM_FAULT_MINOR;
 		}
@@ -1226,6 +1233,7 @@ static int do_wp_page(struct mm_struct *
 			++mm->rss;
 		else
 			page_remove_rmap(old_page);
+		ioproc_invalidate_page(vma, address);
 		break_cow(vma, new_page, address, page_table);
 		lru_cache_add_active(new_page);
 		page_add_anon_rmap(new_page, vma, address);
@@ -1234,6 +1242,7 @@ static int do_wp_page(struct mm_struct *
 		new_page = old_page;
 	}
 	pte_unmap(page_table);
+	ioproc_update_page(vma, address);
 	page_cache_release(new_page);
 	page_cache_release(old_page);
 	spin_unlock(&mm->page_table_lock);
@@ -1630,6 +1639,7 @@ static int do_swap_page(struct mm_struct
 	update_mmu_cache(vma, address, pte);
 	lazy_mmu_prot_update(pte);
 	pte_unmap(page_table);
+	ioproc_update_page(vma, address);
 	spin_unlock(&mm->page_table_lock);
 out:
 	return ret;
@@ -1695,6 +1705,7 @@ do_anonymous_page(struct mm_struct *mm, 
 	/* No need to invalidate - it was non-present before */
 	update_mmu_cache(vma, addr, entry);
 	lazy_mmu_prot_update(entry);
+	ioproc_update_page(vma, addr);
 	spin_unlock(&mm->page_table_lock);
 out:
 	return VM_FAULT_MINOR;
@@ -1813,6 +1824,7 @@ retry:
 	/* no need to invalidate: a not-present page shouldn't be cached */
 	update_mmu_cache(vma, address, entry);
 	lazy_mmu_prot_update(entry);
+	ioproc_update_page(vma, address);
 	spin_unlock(&mm->page_table_lock);
 out:
 	return ret;
@@ -1998,6 +2010,7 @@ int make_pages_present(unsigned long add
 		return ret;
 	return ret == len ? 0 : -1;
 }
+EXPORT_SYMBOL(make_pages_present);
 
 /* 
  * Map a vmalloc()-space virtual address to the physical page.
Index: linux-269-5502/mm/mmap.c
===================================================================
--- linux-269-5502.orig/mm/mmap.c
+++ linux-269-5502/mm/mmap.c
@@ -15,6 +15,7 @@
 #include <linux/init.h>
 #include <linux/file.h>
 #include <linux/fs.h>
+#include <linux/ioproc.h>
 #include <linux/personality.h>
 #include <linux/security.h>
 #include <linux/hugetlb.h>
@@ -1703,6 +1704,7 @@ static void unmap_region(struct mm_struc
 	unsigned long nr_accounted = 0;
 
 	lru_add_drain();
+	ioproc_invalidate_range(vma, start, end);
 	tlb = tlb_gather_mmu(mm, 0);
 	unmap_vmas(&tlb, mm, vma, start, end, &nr_accounted, NULL);
 	vm_unacct_memory(nr_accounted);
@@ -1995,6 +1997,7 @@ void exit_mmap(struct mm_struct *mm)
 
 	spin_lock(&mm->page_table_lock);
 
+	ioproc_release(mm);
 	tlb = tlb_gather_mmu(mm, 1);
 	flush_cache_mm(mm);
 	/* Use ~0UL here to ensure all VMAs in the mm are unmapped */
Index: linux-269-5502/ipc/shm.c
===================================================================
--- linux-269-5502.orig/ipc/shm.c
+++ linux-269-5502/ipc/shm.c
@@ -26,6 +26,7 @@
 #include <linux/proc_fs.h>
 #include <linux/shmem_fs.h>
 #include <linux/security.h>
+#include <linux/module.h>
 #include <linux/audit.h>
 #include <asm/uaccess.h>
 
@@ -856,6 +857,44 @@ asmlinkage long sys_shmdt(char __user *s
 	return 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-269-5502/include/linux/init_task.h
===================================================================
--- linux-269-5502.orig/include/linux/init_task.h
+++ linux-269-5502/include/linux/init_task.h
@@ -2,6 +2,7 @@
 #define _LINUX__INIT_TASK_H
 
 #include <linux/file.h>