[Linux-ia64] HugeTLB Page patch for IA-64 2.5.60 kernel

[Linux-ia64] HugeTLB Page patch for IA-64 2.5.60 kernel

[Seth, Rohit]

[-- Attachment #1.1: Type: text/plain, Size: 261 bytes --]

Please find attached a hugetlb page patch for IA-64 2.5.60 kernel .
Changes in generic files are mostly backported from 2.5.62 (to get
hugetlb support functioning properly).  

Let me know if anyone sees any issues.

thanks,
rohit


 <<patch.2560>> 

[-- Attachment #1.2: Type: text/html, Size: 856 bytes --]

[-- Attachment #2: patch.2560 --]
[-- Type: application/octet-stream, Size: 5841 bytes --]

--- include/linux/hugetlb.h.org	Wed Feb 19 12:11:16 2003
+++ include/linux/hugetlb.h	Fri Feb 21 10:46:55 2003
@@ -26,6 +26,7 @@
 					unsigned long address);
 struct page *follow_huge_pmd(struct mm_struct *mm, unsigned long address,
 				pmd_t *pmd, int write);
+int is_aligned_hugepage_range(unsigned long addr, unsigned long len);
 int pmd_huge(pmd_t pmd);
 
 extern int htlbpage_max;
@@ -56,6 +57,7 @@
 #define hugepage_vma(mm, addr)			0
 #define mark_mm_hugetlb(mm, vma)		do { } while (0)
 #define follow_huge_pmd(mm, addr, pmd, write)	0
+#define is_aligned_hugepage_range(addr, len)	0
 #define pmd_huge(x)	0
 
 #ifndef HPAGE_MASK
--- include/asm-ia64/page.h.org	Fri Feb 21 10:41:55 2003
+++ include/asm-ia64/page.h	Fri Feb 21 11:35:00 2003
@@ -128,6 +128,9 @@
 # define htlbpage_to_page(x)	((REGION_NUMBER(x) << 61)				\
 				 | (REGION_OFFSET(x) >> (HPAGE_SHIFT-PAGE_SHIFT)))
 # define HUGETLB_PAGE_ORDER	(HPAGE_SHIFT - PAGE_SHIFT)
+extern int  is_invalid_hugepage_range(unsigned long addr, unsigned long len);
+#else
+#define is_invalid_hugepage_range(addr, len) 0
 #endif
 
 static __inline__ int
--- mm/mmap.c.org	Wed Feb 19 11:45:51 2003
+++ mm/mmap.c	Wed Feb 19 12:56:07 2003
@@ -797,10 +797,26 @@
 		unsigned long pgoff, unsigned long flags)
 {
 	if (flags & MAP_FIXED) {
+		unsigned long ret;
+
 		if (addr > TASK_SIZE - len)
 			return -ENOMEM;
 		if (addr & ~PAGE_MASK)
 			return -EINVAL;
+		if (file && is_file_hugepages(file)) 
+		/* If the request is for hugepages, then make sure that addr
+		 * and length is properly aligned.
+		 */
+			ret = is_aligned_hugepage_range(addr, len);
+		else 
+		/* 
+		 * Make sure that a normal request is not falling
+		 * in reserved hugepage range.  For some archs like IA-64, 
+		 * there is seperate region for hugepages.
+		 */
+			ret = is_invalid_hugepage_range(addr, len);
+		if (ret)
+			return ret;
 		return addr;
 	}
 
--- fs/hugetlbfs/inode.c.org	Wed Feb 19 11:01:43 2003
+++ fs/hugetlbfs/inode.c	Wed Feb 19 11:03:27 2003
@@ -45,6 +45,7 @@
 {
 	struct inode *inode =file->f_dentry->d_inode;
 	struct address_space *mapping = inode->i_mapping;
+	loff_t len;
 	int ret;
 
 	if (!capable(CAP_IPC_LOCK))
@@ -65,6 +66,10 @@
 	vma->vm_flags |= VM_HUGETLB | VM_RESERVED;
 	vma->vm_ops = &hugetlb_vm_ops;
 	ret = hugetlb_prefault(mapping, vma);
+	len = (loff_t)(vma->vm_end - vma->vm_start) +
+		((loff_t)vma->vm_pgoff << PAGE_SHIFT);
+	if (ret == 0 && inode->i_size < len)
+		inode->i_size = len;
 	up(&inode->i_sem);
 	return ret;
 }
@@ -211,7 +216,7 @@
 		list_add(&inode->i_list, &inode_unused);
 	}
 	inodes_stat.nr_unused++;
-	if (!super_block | (super_block->s_flags & MS_ACTIVE)) {
+	if (!super_block || (super_block->s_flags & MS_ACTIVE)) {
 		spin_unlock(&inode_lock);
 		return;
 	}
--- arch/ia64/mm/hugetlbpage.c.org	Thu Feb 13 11:24:44 2003
+++ arch/ia64/mm/hugetlbpage.c	Fri Feb 21 18:17:58 2003
@@ -95,6 +95,31 @@
 	set_pte(page_table, entry);
 	return;
 }
+/*
+ * This function checks for proper alignment of input addr and len parameters.
+ */
+int is_aligned_hugepage_range(unsigned long addr, unsigned long len)
+{
+	if (len & ~HPAGE_MASK)
+		return -EINVAL;
+	if (addr & ~HPAGE_MASK)
+		return -EINVAL;
+	if (REGION_NUMBER(addr) != REGION_HPAGE)
+		return -EINVAL;
+
+	return 0;
+}
+/* This function checks if the address and address+len falls out of HugeTLB region.  It
+ * return -EINVAL if any part of address range falls in HugeTLB region.
+ */
+int  is_invalid_hugepage_range(unsigned long addr, unsigned long len)
+{
+	if (REGION_NUMBER(addr) == REGION_HPAGE)
+		return -EINVAL;
+	if (REGION_NUMBER(addr+len) == REGION_HPAGE)
+		return -EINVAL;
+	return 0;
+}
 
 int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src,
 			struct vm_area_struct *vma)
@@ -158,6 +183,39 @@
 	return i;
 }
 
+struct vm_area_struct *hugepage_vma(struct mm_struct *mm, unsigned long addr)
+{
+	if (mm->used_hugetlb) {
+		if (REGION_NUMBER(addr) == REGION_HPAGE) {
+			struct vm_area_struct *vma = find_vma(mm, addr);
+			if (vma && is_vm_hugetlb_page(vma))
+				return vma;
+		}
+	}
+	return NULL;
+}
+
+struct page *follow_huge_addr(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, int write)
+{
+	struct page *page;
+	pte_t *ptep;
+
+	ptep = huge_pte_offset(mm, addr);
+	page = pte_page(*ptep);
+	page += ((addr & ~HPAGE_MASK) >> PAGE_SHIFT);
+	get_page(page);
+	return page;
+}
+int pmd_huge(pmd_t pmd)
+{
+	return 0;
+}
+struct page *
+follow_huge_pmd(struct mm_struct *mm, unsigned long address, pmd_t *pmd, int write)
+{
+	return NULL;
+}
+
 void free_huge_page(struct page *page)
 {
 	BUG_ON(page_count(page));
@@ -189,8 +247,6 @@
 	BUG_ON(start & (HPAGE_SIZE - 1));
 	BUG_ON(end & (HPAGE_SIZE - 1));
 
-	spin_lock(&htlbpage_lock);
-	spin_unlock(&htlbpage_lock);
 	for (address = start; address < end; address += HPAGE_SIZE) {
 		pte = huge_pte_offset(mm, address);
 		if (pte_none(*pte))
@@ -242,8 +298,12 @@
 				ret = -ENOMEM;
 				goto out;
 			}
-			add_to_page_cache(page, mapping, idx, GFP_ATOMIC);
+			ret = add_to_page_cache(page, mapping, idx, GFP_ATOMIC);
 			unlock_page(page);
+			if (ret) {
+				free_huge_page(page);
+				goto out;
+			}
 		}
 		set_huge_pte(mm, vma, page, pte, vma->vm_flags & VM_WRITE);
 	}
@@ -287,8 +347,8 @@
 				break;
 		}
 		page = list_entry(p, struct page, list);
-		if ((page_zone(page))->name[0] != 'H') // Look for non-Highmem
-				map = page;
+		if (!PageHighMem(page))
+			map = page;
 	}
 	if (map) {
 		list_del(&map->list);
@@ -302,8 +362,8 @@
 
 int set_hugetlb_mem_size(int count)
 {
-	int j, lcount;
-	struct page *page, *map;
+	int  lcount;
+	struct page *page ;
 	extern long htlbzone_pages;
 	extern struct list_head htlbpage_freelist;
 
@@ -402,5 +462,4 @@
 
 struct vm_operations_struct hugetlb_vm_ops = {
 	.nopage =	hugetlb_nopage,
-	.close =	zap_hugetlb_resources,
 };