hi Linus-
please accept this patch that implements the mincore() system call for
ia32 Linux. this is a precursor patch for an inital ia32 madvise()
implementation. the patch is against Linux 2.3.49-pre2. thanks.
Binary files Linux-2.3.49-pre2/ID and linux/ID differ
diff -ruN Linux-2.3.49-pre2/arch/i386/kernel/entry.S linux/arch/i386/kernel/entry.S
--- Linux-2.3.49-pre2/arch/i386/kernel/entry.S Thu Mar 2 15:59:14 2000
+++ linux/arch/i386/kernel/entry.S Thu Mar 2 16:17:30 2000
@@ -638,6 +638,7 @@
.long SYMBOL_NAME(sys_setfsuid) /* 215 */
.long SYMBOL_NAME(sys_setfsgid)
.long SYMBOL_NAME(sys_pivot_root)
+ .long SYMBOL_NAME(sys_mincore)
/*
@@ -646,6 +647,6 @@
* entries. Don't panic if you notice that this hasn't
* been shrunk every time we add a new system call.
*/
- .rept NR_syscalls-217
+ .rept NR_syscalls-218
.long SYMBOL_NAME(sys_ni_syscall)
.endr
diff -ruN Linux-2.3.49-pre2/include/asm-i386/unistd.h linux/include/asm-i386/unistd.h
--- Linux-2.3.49-pre2/include/asm-i386/unistd.h Thu Mar 2 15:58:11 2000
+++ linux/include/asm-i386/unistd.h Thu Mar 2 16:18:08 2000
@@ -222,6 +222,7 @@
#define __NR_setfsuid32 215
#define __NR_setfsgid32 216
#define __NR_pivot_root 217
+#define __NR_mincore 218
/* user-visible error numbers are in the range -1 - -124: see <asm-i386/errno.h> */
diff -ruN Linux-2.3.49-pre2/include/linux/mm.h linux/include/linux/mm.h
--- Linux-2.3.49-pre2/include/linux/mm.h Thu Mar 2 15:59:49 2000
+++ linux/include/linux/mm.h Thu Mar 2 16:19:29 2000
@@ -106,6 +106,7 @@
void (*protect)(struct vm_area_struct *area, unsigned long, size_t, unsigned int newprot);
int (*sync)(struct vm_area_struct *area, unsigned long, size_t, unsigned int flags);
void (*advise)(struct vm_area_struct *area, unsigned long, size_t, unsigned int advise);
+ unsigned char (*incore)(struct vm_area_struct *area, unsigned long);
struct page * (*nopage)(struct vm_area_struct * area, unsigned long address, int write_access);
struct page * (*wppage)(struct vm_area_struct * area, unsigned long address, struct page * page);
int (*swapout)(struct page *, struct file *);
@@ -446,6 +447,8 @@
size_t size, unsigned int flags);
extern struct page *filemap_nopage(struct vm_area_struct * area,
unsigned long address, int no_share);
+extern unsigned char filemap_incore(struct vm_area_struct * vma,
+ unsigned long pgoff);
/*
* GFP bitmasks..
diff -ruN Linux-2.3.49-pre2/ipc/shm.c linux/ipc/shm.c
--- Linux-2.3.49-pre2/ipc/shm.c Thu Mar 2 15:59:28 2000
+++ linux/ipc/shm.c Thu Mar 2 16:25:24 2000
@@ -65,6 +65,7 @@
static void killseg (int shmid);
static void shm_open (struct vm_area_struct *shmd);
static void shm_close (struct vm_area_struct *shmd);
+static unsigned char shm_incore (struct vm_area_struct *shmd, unsigned long pgoff);
static struct page * shm_nopage(struct vm_area_struct *, unsigned long, int);
static int shm_swapout(struct page *, struct file *);
#ifdef CONFIG_PROC_FS
@@ -623,6 +624,7 @@
static struct vm_operations_struct shm_vm_ops = {
open: shm_open, /* open - callback for a new vm-area open */
close: shm_close, /* close - callback for when the vm-area is released */
+ incore: shm_incore,
nopage: shm_nopage,
swapout: shm_swapout,
};
@@ -864,6 +866,40 @@
{
return 0;
}
+
+/*
+ * is page present?
+ *
+ * shm has a special incore method because we need to synchronize
+ * with the shm swapper (shm_swap) while finding the page.
+ */
+static unsigned char shm_incore(struct vm_area_struct * shmd,
+ unsigned long pgoff)
+{
+ unsigned char present = 0;
+ pte_t pte;
+ struct shmid_kernel * shp =
+ (struct shmid_kernel *) shmd->vm_private_data;
+
+ down(&shp->sem);
+ if(shp != shm_lock(shp->id))
+ BUG();
+
+ pte = SHM_ENTRY(shp, pgoff);
+
+ /*
+ * the pte isn't present if the page is swapped, or if it hasn't
+ * been touched yet.
+ */
+ if (pte_present(pte))
+ present = 1;
+
+ shm_unlock(shp->id);
+ up(&shp->sem);
+
+ return present;
+}
+
/*
* page not present ... go through shm_dir
diff -ruN Linux-2.3.49-pre2/mm/filemap.c linux/mm/filemap.c
--- Linux-2.3.49-pre2/mm/filemap.c Thu Mar 2 15:59:28 2000
+++ linux/mm/filemap.c Thu Mar 2 16:39:10 2000
@@ -707,6 +707,28 @@
return page;
}
+/*
+ * Later we can get more picky about what "in core" means precisely
+ * for a filemapped page. For now, simply check to see if the page
+ * is in the page cache, and is up to date; i.e. that no page-in
+ * operation would be required at this time if an application were
+ * to map and access this page.
+ */
+unsigned char filemap_incore(struct vm_area_struct * vma, unsigned long pgoff)
+{
+ unsigned char present = 0;
+ struct address_space * as = &vma->vm_file->f_dentry->d_inode->i_data;
+ struct page * page, ** hash = page_hash(as, pgoff);
+
+ spin_lock(&pagecache_lock);
+ page = __find_page_nolock(as, pgoff, *hash);
+ if ((page) && (Page_Uptodate(page)))
+ present = 1;
+ spin_unlock(&pagecache_lock);
+
+ return present;
+}
+
#if 0
#define PROFILE_READAHEAD
#define DEBUG_READAHEAD
@@ -1612,6 +1634,7 @@
static struct vm_operations_struct file_shared_mmap = {
unmap: filemap_unmap, /* unmap - we need to sync the pages */
sync: filemap_sync,
+ incore: filemap_incore,
nopage: filemap_nopage,
swapout: filemap_swapout,
};
@@ -1623,6 +1646,7 @@
* know they can't ever get write permissions..)
*/
static struct vm_operations_struct file_private_mmap = {
+ incore: filemap_incore,
nopage: filemap_nopage,
};
diff -ruN Linux-2.3.49-pre2/mm/mmap.c linux/mm/mmap.c
--- Linux-2.3.49-pre2/mm/mmap.c Thu Mar 2 15:59:00 2000
+++ linux/mm/mmap.c Thu Mar 2 16:41:08 2000
@@ -729,6 +729,140 @@
return ret;
}
+static long mincore_area(struct vm_area_struct * vma,
+ unsigned long start, unsigned long end, unsigned char * vec)
+{
+ long error, i, remaining;
+ unsigned char * tmp;
+ unsigned char (*incore)(struct vm_area_struct * , unsigned long);
+
+ error = -ENOMEM;
+ if (!vma->vm_ops || !vma->vm_ops->incore)
+ return error;
+ incore = vma->vm_ops->incore;
+
+ start = ((start - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
+ if (end > vma->vm_end)
+ end = vma->vm_end;
+ end = ((end - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
+
+ error = -EAGAIN;
+ tmp = (unsigned char *) __get_free_page(GFP_KERNEL);
+ if (!tmp)
+ return error;
+
+ /* (end - start) is # of pages, and also # of bytes in "vec */
+ remaining = (end - start),
+
+ error = 0;
+ for (i = 0; remaining > 0; remaining -= PAGE_SIZE, i++) {
+ int j = 0;
+ long thispiece = (remaining < PAGE_SIZE) ?
+ remaining : PAGE_SIZE;
+
+ while (j < thispiece)
+ tmp[j++] = incore(vma, start++);
+
+ if (copy_to_user(vec + PAGE_SIZE * i, tmp, thispiece)) {
+ error = -EFAULT;
+ break;
+ }
+ }
+
+ free_page((unsigned long) tmp);
+ return error;
+}
+
+/*
+ * The mincore(2) system call.
+ *
+ * mincore() returns the memory residency status of the pages in the
+ * current process's address space specified by [addr, addr + len).
+ * The status is returned in a vector of bytes. The least significant
+ * bit of each byte is 1 if the referenced page is in memory, otherwise
+ * it is zero.
+ *
+ * Because the status of a page can change after mincore() checks it
+ * but before it returns to the application, the returned vector may
+ * contain stale information. Only locked pages are guaranteed to
+ * remain in memory.
+ *
+ * return values:
+ * zero - success
+ * -EFAULT - vec points to an illegal address
+ * -EINVAL - addr is not a multiple of PAGE_CACHE_SIZE,
+ * or len has a nonpositive value
+ * -ENOMEM - Addresses in the range [addr, addr + len] are
+ * invalid for the address space of this process, or
+ * specify one or more pages which are not currently
+ * mapped
+ * -EAGAIN - A kernel resource was temporarily unavailable.
+ */
+asmlinkage long sys_mincore(unsigned long start, size_t len,
+ unsigned char * vec)
+{
+ int index = 0;
+ unsigned long end;
+ struct vm_area_struct * vma;
+ int unmapped_error = 0;
+ long error = -EINVAL;
+
+ down(¤t->mm->mmap_sem);
+
+ if (start & ~PAGE_MASK)
+ goto out;
+ len = (len + ~PAGE_MASK) & PAGE_MASK;
+ end = start + len;
+ if (end < start)
+ goto out;
+
+ error = 0;
+ if (end == start)
+ goto out;
+
+ /*
+ * If the interval [start,end) covers some unmapped address
+ * ranges, just ignore them, but return -ENOMEM at the end.
+ */
+ vma = find_vma(current->mm, start);
+ for (;;) {
+ /* Still start < end. */
+ error = -ENOMEM;
+ if (!vma)
+ goto out;
+
+ /* Here start < vma->vm_end. */
+ if (start < vma->vm_start) {
+ unmapped_error = -ENOMEM;
+ start = vma->vm_start;
+ }
+
+ /* Here vma->vm_start <= start < vma->vm_end. */
+ if (end <= vma->vm_end) {
+ if (start < end) {
+ error = mincore_area(vma, start, end,
+ &vec[index]);
+ if (error)
+ goto out;
+ }
+ error = unmapped_error;
+ goto out;
+ }
+
+ /* Here vma->vm_start <= start < vma->vm_end < end. */
+ error = mincore_area(vma, start, vma->vm_end, &vec[index]);
+ if (error)
+ goto out;
+ index += (vma->vm_end - start) >> PAGE_CACHE_SHIFT;
+ start = vma->vm_end;
+ vma = vma->vm_next;
+ }
+
+out:
+ up(¤t->mm->mmap_sem);
+ return error;
+}
+
/*
* this is really a simplified "do_mmap". it only handles
* anonymous maps. eventually we may be able to do some
- Chuck Lever
-- corporate: <chuckl@netscape.com> personal: <chucklever@netscape.net> or <cel@monkey.org>The Linux Scalability project: http://www.citi.umich.edu/projects/linux-scalability/
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Mar 07 2000 - 21:00:13 EST