Merge bk://linuxusb.bkbits.net/linus-2.5

into home.transmeta.com:/home/torvalds/v2.5/linux

arch/cris/drivers/examples/kiobuftest.c

-/*
- * Example showing how to pin down a range of virtual pages from user-space
- * to be able to do for example DMA directly into them.
- *
- * It is necessary because the pages the virtual pointers reference, might
- * not exist in memory (could be mapped to the zero-page, filemapped etc)
- * and DMA cannot trigger the MMU to force them in (and would have time 
- * contraints making it impossible to wait for it anyway).
- *
- * Author:  Bjorn Wesen
- *
- * $Log: kiobuftest.c,v $
- * Revision 1.1.1.1  2001/12/17 13:59:27  bjornw
- * Import of Linux 2.5.1
- *
- * Revision 1.2  2001/02/27 13:52:50  bjornw
- * malloc.h -> slab.h
- *
- * Revision 1.1  2001/01/19 15:57:49  bjornw
- * Example of how to do direct HW -> user-mode DMA
- *
- *
- */
-
-#include <linux/module.h>
-#include <linux/sched.h>
-#include <linux/slab.h>
-#include <linux/errno.h>
-#include <linux/kernel.h>
-#include <linux/fs.h>
-#include <linux/string.h>
-#include <linux/init.h>
-#include <linux/mm.h>
-#include <linux/iobuf.h>
-
-#define KIOBUFTEST_MAJOR 124  /* in the local range, experimental */
-
-
-static ssize_t
-kiobuf_read(struct file *filp, char *buf, size_t len, loff_t *ppos)
-{
-
-     struct kiobuf *iobuf;
-     int res, i;
- 
-     /* Make a kiobuf that maps the entire length the reader has given
-      * us
-      */
-
-     res = alloc_kiovec(1, &iobuf);
-     if (res)
-	     return res;
-     
-     if((res = map_user_kiobuf(READ, iobuf, (unsigned long)buf, len))) {
-	     printk("map_user_kiobuf failed, return %d\n", res);
-	     return res;
-     }
-
-     /* At this point, the virtual area buf[0] -> buf[len-1] will
-      * have corresponding pages mapped in physical memory and locked
-      * until we unmap the kiobuf. They cannot be swapped out or moved
-      * around.
-      */
-
-     printk("nr_pages == %d\noffset == %d\nlength == %d\n",
-	    iobuf->nr_pages, iobuf->offset, iobuf->length);
-
-     for(i = 0; i < iobuf->nr_pages; i++) {
-	     printk("page_add(maplist[%d]) == 0x%x\n", i,
-		    page_address(iobuf->maplist[i]));
-     }
-
-     /* This is the place to create the necessary scatter-gather vector
-      * for the DMA using the iobuf->maplist array and page_address
-      * (don't forget __pa if the DMA needs the actual physical DRAM address)
-      * and run it.
-      */
-
-
-
-
-     /* Release the mapping and exit */
-
-     unmap_kiobuf(iobuf); /* The unlock_kiobuf is implicit here */
-
-     return len;
-}
-
-
-static struct file_operations kiobuf_fops = {
-	owner:    THIS_MODULE,
-	read:     kiobuf_read
-};
-
-static int __init
-kiobuftest_init(void)
-{
-	int res;
-
-	/* register char device */
-
-	res = register_chrdev(KIOBUFTEST_MAJOR, "kiobuftest", &kiobuf_fops);
-	if(res < 0) {
-		printk(KERN_ERR "kiobuftest: couldn't get a major number.\n");
-		return res;
-	}
-
-	printk("Initializing kiobuf-test device\n");
-}
-
-module_init(kiobuftest_init);

arch/i386/mm/discontig.c

@@ -70,6 +70,7 @@ static void __init allocate_pgdat(int nid)
 	node_datasz = PFN_UP(sizeof(struct pglist_data));
 	NODE_DATA(nid) = (pg_data_t *)(__va(min_low_pfn << PAGE_SHIFT));
 	min_low_pfn += node_datasz;
+	memset(NODE_DATA(nid), 0, sizeof(struct pglist_data));
 }
 
 /*

drivers/block/Config.in

@@ -49,7 +49,6 @@ fi
 dep_bool '  Initial RAM disk (initrd) support' CONFIG_BLK_DEV_INITRD $CONFIG_BLK_DEV_RAM
 
 if [ "$CONFIG_X86" = "y" -o "$CONFIG_PPC32" = "y" ]; then
-#   bool 'Support for Large Block Devices' CONFIG_LBD
-   define_bool CONFIG_LBD y
+   bool 'Support for Large Block Devices' CONFIG_LBD
 fi
 endmenu

drivers/block/ll_rw_blk.c

@@ -1878,7 +1878,7 @@ void generic_make_request(struct bio *bio)
  */
 int submit_bio(int rw, struct bio *bio)
 {
-	int count = bio_sectors(bio) >> 1;
+	int count = bio_sectors(bio);
 
 	BUG_ON(!bio->bi_end_io);
 	BIO_BUG_ON(!bio->bi_size);

drivers/media/video/bttv-risc.c

@@ -29,7 +29,6 @@
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/pci.h>
-#include <linux/iobuf.h>
 #include <linux/vmalloc.h>
 #include <linux/interrupt.h>
 #include <asm/page.h>

drivers/media/video/bttvp.h

@@ -31,7 +31,6 @@
 #include <linux/i2c.h>
 #include <linux/i2c-algo-bit.h>
 #include <linux/videodev.h>
-#include <linux/iobuf.h>
 #include <linux/pci.h>
 #include <asm/scatterlist.h>
 

drivers/media/video/video-buf.c

@@ -18,8 +18,8 @@
 
 #include <linux/init.h>
 #include <linux/module.h>
-#include <linux/iobuf.h>
 #include <linux/vmalloc.h>
+#include <linux/pagemap.h>
 #include <linux/slab.h>
 #include <linux/pci.h>
 #include <linux/interrupt.h>
@@ -65,32 +65,31 @@ videobuf_vmalloc_to_sg(unsigned char *virt, int nr_pages)
 	return NULL;
 }
 
-struct scatterlist*
-videobuf_iobuf_to_sg(struct kiobuf *iobuf)
+struct scatterlist *
+videobuf_pages_to_sg(struct page **pages, int nr_pages, int offset)
 {
 	struct scatterlist *sglist;
 	int i = 0;
-
-	sglist = kmalloc(sizeof(struct scatterlist) * iobuf->nr_pages,
-			 GFP_KERNEL);
+	
+	if (NULL == pages[0])
+		return NULL;
+	sglist = kmalloc(sizeof(*sglist) * nr_pages, GFP_KERNEL);
 	if (NULL == sglist)
 		return NULL;
-	memset(sglist,0,sizeof(struct scatterlist) * iobuf->nr_pages);
+	memset(sglist, 0, sizeof(*sglist) * nr_pages);
 
-	if (NULL == iobuf->maplist[0])
-		goto err;
-	if (PageHighMem(iobuf->maplist[0]))
+	if (PageHighMem(pages[0]))
 		/* DMA to highmem pages might not work */
 		goto err;
-	sglist[0].page   = iobuf->maplist[0];
-	sglist[0].offset = iobuf->offset;
-	sglist[0].length = PAGE_SIZE - iobuf->offset;
-	for (i = 1; i < iobuf->nr_pages; i++) {
-		if (NULL == iobuf->maplist[i])
+	sglist[0].page   = pages[0];
+	sglist[0].offset = offset;
+	sglist[0].length = PAGE_SIZE - offset;
+	for (i = 1; i < nr_pages; i++) {
+		if (NULL == pages[i])
 			goto err;
-		if (PageHighMem(iobuf->maplist[i]))
+		if (PageHighMem(pages[i]))
 			goto err;
-		sglist[i].page   = iobuf->maplist[i];
+		sglist[i].page   = pages[i];
 		sglist[i].length = PAGE_SIZE;
 	}
 	return sglist;
@@ -100,6 +99,30 @@ videobuf_iobuf_to_sg(struct kiobuf *iobuf)
 	return NULL;
 }
 
+int videobuf_lock(struct page **pages, int nr_pages)
+{
+	int i;
+
+	for (i = 0; i < nr_pages; i++)
+		if (TestSetPageLocked(pages[i]))
+			goto err;
+	return 0;
+
+ err:
+	while (i > 0)
+		unlock_page(pages[--i]);
+	return -EINVAL;
+}
+
+int videobuf_unlock(struct page **pages, int nr_pages)
+{
+	int i;
+
+	for (i = 0; i < nr_pages; i++)
+		unlock_page(pages[i]);
+	return 0;
+}
+
 /* --------------------------------------------------------------------- */
 
 int videobuf_dma_init_user(struct videobuf_dmabuf *dma, int direction,
@@ -113,14 +136,21 @@ int videobuf_dma_init_user(struct videobuf_dmabuf *dma, int direction,
 	case PCI_DMA_TODEVICE:   rw = WRITE; break;
 	default:                 BUG();
 	}
-	if (0 != (err = alloc_kiovec(1,&dma->iobuf)))
-		return err;
-	if (0 != (err = map_user_kiobuf(rw, dma->iobuf, data, size))) {
-		dprintk(1,"map_user_kiobuf: %d\n",err);
-		return err;
-	}
-	dma->nr_pages = dma->iobuf->nr_pages;
-	return 0;
+
+	dma->offset   = data & PAGE_MASK;
+	dma->nr_pages = ((((data+size) & ~PAGE_MASK) -
+			  (data & ~PAGE_MASK)) >> PAGE_SHIFT) +1;
+	dma->pages    = kmalloc(dma->nr_pages * sizeof(struct page*),
+				GFP_KERNEL);
+	if (NULL == dma->pages)
+		return -ENOMEM;
+	down_read(&current->mm->mmap_sem);
+        err = get_user_pages(current,current->mm,
+			     data, dma->nr_pages,
+			     rw == READ, 0, /* don't force */
+			     dma->pages, NULL);
+	up_read(&current->mm->mmap_sem);
+	return err;
 }
 
 int videobuf_dma_init_kernel(struct videobuf_dmabuf *dma, int direction,
@@ -144,13 +174,15 @@ int videobuf_dma_pci_map(struct pci_dev *dev, struct videobuf_dmabuf *dma)
 	if (0 == dma->nr_pages)
 		BUG();
 	
-	if (dma->iobuf) {
-		if (0 != (err = lock_kiovec(1,&dma->iobuf,1))) {
-			dprintk(1,"lock_kiovec: %d\n",err);
+	if (dma->pages) {
+		if (0 != (err = videobuf_lock(dma->pages, dma->nr_pages))) {
+			dprintk(1,"videobuf_lock_pages: %d\n",err);
 			return err;
 		}
-		dma->sglist = videobuf_iobuf_to_sg(dma->iobuf);
+		dma->sglist = videobuf_pages_to_sg(dma->pages, dma->nr_pages,
+						   dma->offset);
 	}
+
 	if (dma->vmalloc) {
 		dma->sglist = videobuf_vmalloc_to_sg
 			(dma->vmalloc,dma->nr_pages);
@@ -160,7 +192,7 @@ int videobuf_dma_pci_map(struct pci_dev *dev, struct videobuf_dmabuf *dma)
 		return -ENOMEM;
 	}
 	dma->sglen = pci_map_sg(dev,dma->sglist,dma->nr_pages,
-				 dma->direction);
+				dma->direction);
 	return 0;
 }
 
@@ -182,8 +214,8 @@ int videobuf_dma_pci_unmap(struct pci_dev *dev, struct videobuf_dmabuf *dma)
 	kfree(dma->sglist);
 	dma->sglist = NULL;
 	dma->sglen = 0;
-	if (dma->iobuf)
-		unlock_kiovec(1,&dma->iobuf);
+	if (dma->pages)
+		videobuf_lock(dma->pages, dma->nr_pages);
 	return 0;
 }
 
@@ -192,11 +224,14 @@ int videobuf_dma_free(struct videobuf_dmabuf *dma)
 	if (dma->sglen)
 		BUG();
 
-	if (dma->iobuf) {
-		unmap_kiobuf(dma->iobuf);
-		free_kiovec(1,&dma->iobuf);
-		dma->iobuf = NULL;
+	if (dma->pages) {
+		int i;
+		for (i=0; i < dma->nr_pages; i++)
+			page_cache_release(dma->pages[i]);
+		kfree(dma->pages);
+		dma->pages = NULL;
 	}
+
 	if (dma->vmalloc) {
 		vfree(dma->vmalloc);
 		dma->vmalloc = NULL;
@@ -959,6 +994,7 @@ int videobuf_mmap_mapper(struct vm_area_struct *vma,
 	map->q       = q;
 	vma->vm_ops  = &videobuf_vm_ops;
 	vma->vm_flags |= VM_DONTEXPAND;
+	vma->vm_flags &= ~VM_IO; /* using shared anonymous pages */
 	vma->vm_private_data = map;
 	dprintk(1,"mmap %p: %08lx-%08lx pgoff %08lx bufs %d-%d\n",
 		map,vma->vm_start,vma->vm_end,vma->vm_pgoff,first,last);
@@ -972,7 +1008,6 @@ int videobuf_mmap_mapper(struct vm_area_struct *vma,
 /* --------------------------------------------------------------------- */
 
 EXPORT_SYMBOL_GPL(videobuf_vmalloc_to_sg);
-EXPORT_SYMBOL_GPL(videobuf_iobuf_to_sg);
 
 EXPORT_SYMBOL_GPL(videobuf_dma_init_user);
 EXPORT_SYMBOL_GPL(videobuf_dma_init_kernel);

drivers/media/video/video-buf.h

@@ -28,11 +28,12 @@
 struct scatterlist* videobuf_vmalloc_to_sg(unsigned char *virt, int nr_pages);
 
 /*
- * Return a scatterlist for a locked iobuf (NULL on errors).  Memory
+ * Return a scatterlist for a an array of userpages (NULL on errors).  Memory
  * for the scatterlist is allocated using kmalloc.  The caller must
  * free the memory.
  */
-struct scatterlist* videobuf_iobuf_to_sg(struct kiobuf *iobuf);
+struct scatterlist *videobuf_pages_to_sg(struct page **pages, int nr_pages,
+					 int offset);
 
 /* --------------------------------------------------------------------- */
 
@@ -57,7 +58,8 @@ struct scatterlist* videobuf_iobuf_to_sg(struct kiobuf *iobuf);
 
 struct videobuf_dmabuf {
 	/* for userland buffer */
-	struct kiobuf       *iobuf;
+	struct page         **pages;
+	int                 offset;
 
 	/* for kernel buffers */
 	void                *vmalloc;

fs/Makefile

@@ -11,7 +11,7 @@ export-objs :=	open.o dcache.o buffer.o bio.o inode.o dquot.o mpage.o aio.o \
 obj-y :=	open.o read_write.o devices.o file_table.o buffer.o \
 		bio.o super.o block_dev.o char_dev.o stat.o exec.o pipe.o \
 		namei.o fcntl.o ioctl.o readdir.o select.o fifo.o locks.o \
-		dcache.o inode.o attr.o bad_inode.o file.o iobuf.o dnotify.o \
+		dcache.o inode.o attr.o bad_inode.o file.o dnotify.o \
 		filesystems.o namespace.o seq_file.o xattr.o libfs.o \
 		fs-writeback.o mpage.o direct-io.o aio.o
 

fs/aio.c

@@ -9,6 +9,7 @@
  *	See ../COPYING for licensing terms.
  */
 #include <linux/kernel.h>
+#include <linux/init.h>
 #include <linux/errno.h>
 #include <linux/time.h>
 #include <linux/aio_abi.h>
@@ -21,15 +22,9 @@
 #include <linux/file.h>
 #include <linux/mm.h>
 #include <linux/mman.h>
-#include <linux/vmalloc.h>
-#include <linux/iobuf.h>
 #include <linux/slab.h>
 #include <linux/timer.h>
-#include <linux/brlock.h>
 #include <linux/aio.h>
-#include <linux/smp_lock.h>
-#include <linux/compiler.h>
-#include <linux/brlock.h>
 #include <linux/module.h>
 #include <linux/highmem.h>
 #include <linux/workqueue.h>

fs/bio.c

@@ -20,7 +20,7 @@
 #include <linux/bio.h>
 #include <linux/blk.h>
 #include <linux/slab.h>
-#include <linux/iobuf.h>
+#include <linux/init.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/mempool.h>
@@ -438,128 +438,6 @@ int bio_add_page(struct bio *bio, struct page *page, unsigned int len,
 	return 0;
 }
 
-static int bio_end_io_kio(struct bio *bio, unsigned int bytes_done, int error)
-{
-	struct kiobuf *kio = (struct kiobuf *) bio->bi_private;
-
-	if (bio->bi_size)
-		return 1;
-
-	end_kio_request(kio, error);
-	bio_put(bio);
-	return 0;
-}
-
-/**
- * ll_rw_kio - submit a &struct kiobuf for I/O
- * @rw:   %READ or %WRITE
- * @kio:   the kiobuf to do I/O on
- * @bdev:   target device
- * @sector:   start location on disk
- *
- * Description:
- *   ll_rw_kio will map the page list inside the &struct kiobuf to
- *   &struct bio and queue them for I/O. The kiobuf given must describe
- *   a continous range of data, and must be fully prepared for I/O.
- **/
-void ll_rw_kio(int rw, struct kiobuf *kio, struct block_device *bdev, sector_t sector)
-{
-	int i, offset, size, err, map_i, total_nr_pages, nr_pages;
-	struct bio *bio;
-
-	err = 0;
-	if ((rw & WRITE) && bdev_read_only(bdev)) {
-		printk("ll_rw_bio: WRITE to ro device %s\n", bdevname(bdev));
-		err = -EPERM;
-		goto out;
-	}
-
-	if (!kio->nr_pages) {
-		err = -EINVAL;
-		goto out;
-	}
-
-	/*
-	 * maybe kio is bigger than the max we can easily map into a bio.
-	 * if so, split it up in appropriately sized chunks.
-	 */
-	total_nr_pages = kio->nr_pages;
-	offset = kio->offset & ~PAGE_MASK;
-	size = kio->length;
-
-	atomic_set(&kio->io_count, 1);
-
-	map_i = 0;
-
-next_chunk:
-	nr_pages = BIO_MAX_PAGES;
-	if (nr_pages > total_nr_pages)
-		nr_pages = total_nr_pages;
-
-	atomic_inc(&kio->io_count);
-
-	/*
-	 * allocate bio and do initial setup
-	 */
-	if ((bio = bio_alloc(GFP_NOIO, nr_pages)) == NULL) {
-		err = -ENOMEM;
-		goto out;
-	}
-
-	bio->bi_sector = sector;
-	bio->bi_bdev = bdev;
-	bio->bi_idx = 0;
-	bio->bi_end_io = bio_end_io_kio;
-	bio->bi_private = kio;
-
-	for (i = 0; i < nr_pages; i++, map_i++) {
-		int nbytes = PAGE_SIZE - offset;
-
-		if (nbytes > size)
-			nbytes = size;
-
-		BUG_ON(kio->maplist[map_i] == NULL);
-
-		/*
-		 * if we can't add this page to the bio, submit for i/o
-		 * and alloc a new one if needed
-		 */
-		if (bio_add_page(bio, kio->maplist[map_i], nbytes, offset))
-			break;
-
-		/*
-		 * kiobuf only has an offset into the first page
-		 */
-		offset = 0;
-
-		sector += nbytes >> 9;
-		size -= nbytes;
-		total_nr_pages--;
-		kio->offset += nbytes;
-	}
-
-	submit_bio(rw, bio);
-
-	if (total_nr_pages)
-		goto next_chunk;
-
-	if (size) {
-		printk("ll_rw_kio: size %d left (kio %d)\n", size, kio->length);
-		BUG();
-	}
-
-out:
-	if (err)
-		kio->errno = err;
-
-	/*
-	 * final atomic_dec of io_count to match our initial setting of 1.
-	 * I/O may or may not have completed at this point, final completion
-	 * handler is only run on last decrement.
-	 */
-	end_kio_request(kio, !err);
-}
-
 /**
  * bio_endio - end I/O on a bio
  * @bio:	bio
@@ -662,7 +540,6 @@ module_init(init_bio);
 
 EXPORT_SYMBOL(bio_alloc);
 EXPORT_SYMBOL(bio_put);
-EXPORT_SYMBOL(ll_rw_kio);
 EXPORT_SYMBOL(bio_endio);
 EXPORT_SYMBOL(bio_init);
 EXPORT_SYMBOL(bio_copy);

fs/block_dev.c

@@ -14,7 +14,6 @@
 #include <linux/major.h>
 #include <linux/devfs_fs_kernel.h>
 #include <linux/smp_lock.h>
-#include <linux/iobuf.h>
 #include <linux/highmem.h>
 #include <linux/blkdev.h>
 #include <linux/module.h>

fs/buffer.c

@@ -28,7 +28,6 @@
 #include <linux/blkdev.h>
 #include <linux/file.h>
 #include <linux/quotaops.h>
-#include <linux/iobuf.h>
 #include <linux/highmem.h>
 #include <linux/module.h>
 #include <linux/writeback.h>
@@ -2300,65 +2299,6 @@ sector_t generic_block_bmap(struct address_space *mapping, sector_t block,
 	return tmp.b_blocknr;
 }
 
-/*
- * Start I/O on a physical range of kernel memory, defined by a vector
- * of kiobuf structs (much like a user-space iovec list).
- *
- * The kiobuf must already be locked for IO.  IO is submitted
- * asynchronously: you need to check page->locked and page->uptodate.
- *
- * It is up to the caller to make sure that there are enough blocks
- * passed in to completely map the iobufs to disk.
- */
-int brw_kiovec(int rw, int nr, struct kiobuf *iovec[],
-	       struct block_device *bdev, sector_t b[], int size)
-{
-	int		transferred;
-	int		i;
-	int		err;
-	struct kiobuf *	iobuf;
-
-	if (!nr)
-		return 0;
-	
-	/* 
-	 * First, do some alignment and validity checks 
-	 */
-	for (i = 0; i < nr; i++) {
-		iobuf = iovec[i];
-		if ((iobuf->offset & (size-1)) || (iobuf->length & (size-1)))
-			return -EINVAL;
-		if (!iobuf->nr_pages)
-			panic("brw_kiovec: iobuf not initialised");
-	}
-
-	/* 
-	 * OK to walk down the iovec doing page IO on each page we find. 
-	 */
-	for (i = 0; i < nr; i++) {
-		iobuf = iovec[i];
-		iobuf->errno = 0;
-
-		ll_rw_kio(rw, iobuf, bdev, b[i] * (size >> 9));
-	}
-
-	/*
-	 * now they are all submitted, wait for completion
-	 */
-	transferred = 0;
-	err = 0;
-	for (i = 0; i < nr; i++) {
-		iobuf = iovec[i];
-		kiobuf_wait_for_io(iobuf);
-		if (iobuf->errno && !err)
-			err = iobuf->errno;
-		if (!err)
-			transferred += iobuf->length;
-	}
-
-	return err ? err : transferred;
-}
-
 static int end_bio_bh_io_sync(struct bio *bio, unsigned int bytes_done, int err)
 {
 	struct buffer_head *bh = bio->bi_private;

fs/dcache.c

@@ -328,7 +328,7 @@ static inline void prune_one_dentry(struct dentry * dentry)
  * all the dentries are in use.
  */
  
-void prune_dcache(int count)
+static void prune_dcache(int count)
 {
 	spin_lock(&dcache_lock);
 	for (; count ; count--) {
@@ -572,25 +572,24 @@ void shrink_dcache_anon(struct list_head *head)
  * This is called from kswapd when we think we need some
  * more memory. 
  */
-int shrink_dcache_memory(int ratio, unsigned int gfp_mask)
+static int shrink_dcache_memory(int nr, unsigned int gfp_mask)
 {
-	int entries = dentry_stat.nr_dentry / ratio + 1;
-	/*
-	 * Nasty deadlock avoidance.
-	 *
-	 * ext2_new_block->getblk->GFP->shrink_dcache_memory->prune_dcache->
-	 * prune_one_dentry->dput->dentry_iput->iput->inode->i_sb->s_op->
-	 * put_inode->ext2_discard_prealloc->ext2_free_blocks->lock_super->
-	 * DEADLOCK.
-	 *
-	 * We should make sure we don't hold the superblock lock over
-	 * block allocations, but for now:
-	 */
-	if (!(gfp_mask & __GFP_FS))
-		return 0;
-
-	prune_dcache(entries);
-	return entries;
+	if (nr) {
+		/*
+		 * Nasty deadlock avoidance.
+		 *
+	 	 * ext2_new_block->getblk->GFP->shrink_dcache_memory->
+		 * prune_dcache->prune_one_dentry->dput->dentry_iput->iput->
+		 * inode->i_sb->s_op->put_inode->ext2_discard_prealloc->
+		 * ext2_free_blocks->lock_super->DEADLOCK.
+	 	 *
+	 	 * We should make sure we don't hold the superblock lock over
+	 	 * block allocations, but for now:
+		 */
+		if (gfp_mask & __GFP_FS)
+			prune_dcache(nr);
+	}
+	return dentry_stat.nr_dentry;
 }
 
 #define NAME_ALLOC_LEN(len)	((len+16) & ~15)
@@ -1330,6 +1329,8 @@ static void __init dcache_init(unsigned long mempages)
 					 NULL, NULL);
 	if (!dentry_cache)
 		panic("Cannot create dentry cache");
+	
+	set_shrinker(DEFAULT_SEEKS, shrink_dcache_memory);
 
 #if PAGE_SHIFT < 13
 	mempages >>= (13 - PAGE_SHIFT);
@@ -1375,9 +1376,6 @@ kmem_cache_t *names_cachep;
 /* SLAB cache for file structures */
 kmem_cache_t *filp_cachep;
 
-/* SLAB cache for dquot structures */
-kmem_cache_t *dquot_cachep;
-
 EXPORT_SYMBOL(d_genocide);
 
 extern void bdev_cache_init(void);
@@ -1397,14 +1395,6 @@ void __init vfs_caches_init(unsigned long mempages)
 	if(!filp_cachep)
 		panic("Cannot create filp SLAB cache");
 
-#if defined (CONFIG_QUOTA)
-	dquot_cachep = kmem_cache_create("dquot", 
-			sizeof(struct dquot), sizeof(unsigned long) * 4,
-			SLAB_HWCACHE_ALIGN, NULL, NULL);
-	if (!dquot_cachep)
-		panic("Cannot create dquot SLAB cache");
-#endif
-
 	dcache_init(mempages);
 	inode_init(mempages);
 	files_init(mempages); 

fs/dquot.c

@@ -55,6 +55,7 @@
 #include <linux/errno.h>
 #include <linux/kernel.h>
 #include <linux/fs.h>
+#include <linux/mm.h>
 #include <linux/time.h>
 #include <linux/types.h>
 #include <linux/string.h>
@@ -481,14 +482,14 @@ static void prune_dqcache(int count)
  * more memory
  */
 
-int shrink_dqcache_memory(int ratio, unsigned int gfp_mask)
+static int shrink_dqcache_memory(int nr, unsigned int gfp_mask)
 {
-	int entries = dqstats.allocated_dquots / ratio + 1;
-
-	lock_kernel();
-	prune_dqcache(entries);
-	unlock_kernel();
-	return entries;
+	if (nr) {
+		lock_kernel();
+		prune_dqcache(nr);
+		unlock_kernel();
+	}
+	return dqstats.allocated_dquots;
 }
 
 /*
@@ -1490,6 +1491,9 @@ static ctl_table sys_table[] = {
 	{},
 };
 
+/* SLAB cache for dquot structures */
+kmem_cache_t *dquot_cachep;
+
 static int __init dquot_init(void)
 {
 	int i;
@@ -1499,9 +1503,17 @@ static int __init dquot_init(void)
 		INIT_LIST_HEAD(dquot_hash + i);
 	printk(KERN_NOTICE "VFS: Disk quotas v%s\n", __DQUOT_VERSION__);
 
+	dquot_cachep = kmem_cache_create("dquot", 
+			sizeof(struct dquot), sizeof(unsigned long) * 4,
+			SLAB_HWCACHE_ALIGN, NULL, NULL);
+	if (!dquot_cachep)
+		panic("Cannot create dquot SLAB cache");
+
+	set_shrinker(DEFAULT_SEEKS, shrink_dqcache_memory);
+
 	return 0;
 }
-__initcall(dquot_init);
+module_init(dquot_init);
 
 EXPORT_SYMBOL(register_quota_format);
 EXPORT_SYMBOL(unregister_quota_format);

fs/fcntl.c

@@ -10,7 +10,6 @@
 #include <linux/dnotify.h>
 #include <linux/smp_lock.h>
 #include <linux/slab.h>
-#include <linux/iobuf.h>
 #include <linux/module.h>
 #include <linux/security.h>
 

fs/file_table.c

@@ -11,7 +11,6 @@
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/smp_lock.h>
-#include <linux/iobuf.h>
 #include <linux/fs.h>
 #include <linux/security.h>
 

fs/inode.c

@@ -243,22 +243,25 @@ void clear_inode(struct inode *inode)
  * Dispose-list gets a local list with local inodes in it, so it doesn't
  * need to worry about list corruption and SMP locks.
  */
-static void dispose_list(struct list_head * head)
+static void dispose_list(struct list_head *head)
 {
-	struct list_head * inode_entry;
-	struct inode * inode;
+	int nr_disposed = 0;
+
+	while (!list_empty(head)) {
+		struct inode *inode;
 
-	while ((inode_entry = head->next) != head)
-	{
-		list_del(inode_entry);
+		inode = list_entry(head->next, struct inode, i_list);
+		list_del(&inode->i_list);
 
-		inode = list_entry(inode_entry, struct inode, i_list);
 		if (inode->i_data.nrpages)
 			truncate_inode_pages(&inode->i_data, 0);
 		clear_inode(inode);
 		destroy_inode(inode);
-		inodes_stat.nr_inodes--;
+		nr_disposed++;
 	}
+	spin_lock(&inode_lock);
+	inodes_stat.nr_inodes -= nr_disposed;
+	spin_unlock(&inode_lock);
 }
 
 /*
@@ -377,7 +380,7 @@ int invalidate_device(kdev_t dev, int do_sync)
 	 !inode_has_buffers(inode))
 #define INODE(entry)	(list_entry(entry, struct inode, i_list))
 
-void prune_icache(int goal)
+static inline void prune_icache(int goal)
 {
 	LIST_HEAD(list);
 	struct list_head *entry, *freeable = &list;
@@ -417,23 +420,19 @@ void prune_icache(int goal)
  * This is called from kswapd when we think we need some
  * more memory. 
  */
-int shrink_icache_memory(int ratio, unsigned int gfp_mask)
+static int shrink_icache_memory(int nr, unsigned int gfp_mask)
 {
-	int entries = inodes_stat.nr_inodes / ratio + 1;
-	/*
-	 * Nasty deadlock avoidance..
-	 *
-	 * We may hold various FS locks, and we don't
-	 * want to recurse into the FS that called us
-	 * in clear_inode() and friends..
-	 */
-	if (!(gfp_mask & __GFP_FS))
-		return 0;
-
-	prune_icache(entries);
-	return entries;
+	if (nr) {
+		/*
+		 * Nasty deadlock avoidance.  We may hold various FS locks,
+		 * and we don't want to recurse into the FS that called us
+		 * in clear_inode() and friends..
+	 	 */
+		if (gfp_mask & __GFP_FS)
+			prune_icache(nr);
+	}
+	return inodes_stat.nr_inodes;
 }
-EXPORT_SYMBOL(shrink_icache_memory);
 
 /*
  * Called with the inode lock held.
@@ -1226,4 +1225,6 @@ void __init inode_init(unsigned long mempages)
 					 NULL);
 	if (!inode_cachep)
 		panic("cannot create inode slab cache");
+
+	set_shrinker(DEFAULT_SEEKS, shrink_icache_memory);
 }

fs/iobuf.c

-/*
- * iobuf.c
- *
- * Keep track of the general-purpose IO-buffer structures used to track
- * abstract kernel-space io buffers.
- * 
- */
-
-#include <linux/iobuf.h>
-#include <linux/slab.h>
-
-int end_kio_request(struct kiobuf *kiobuf, int uptodate)
-{
-	int ret = 1;
-
-	if ((!uptodate) && !kiobuf->errno)
-		kiobuf->errno = -EIO;
-
-	if (atomic_dec_and_test(&kiobuf->io_count)) {
-		ret = 0;
-		if (kiobuf->end_io)
-			kiobuf->end_io(kiobuf);
-		wake_up(&kiobuf->wait_queue);
-	}
-
-	return ret;
-}
-
-static void kiobuf_init(struct kiobuf *iobuf)
-{
-	init_waitqueue_head(&iobuf->wait_queue);
-	atomic_set(&iobuf->io_count, 0);
-	iobuf->array_len = KIO_STATIC_PAGES;
-	iobuf->maplist   = iobuf->map_array;
-	iobuf->nr_pages   = 0;
-	iobuf->locked   = 0;
-	iobuf->io_count.counter   = 0;
-	iobuf->end_io = NULL;
-}
-
-int alloc_kiovec(int nr, struct kiobuf **bufp)
-{
-	int i;
-	struct kiobuf *iobuf;
-	
-	for (i = 0; i < nr; i++) {
-		iobuf = kmalloc(sizeof(struct kiobuf), GFP_KERNEL);
-		if (!iobuf) {
-			free_kiovec(i, bufp);
-			return -ENOMEM;
-		}
-		kiobuf_init(iobuf);
-		bufp[i] = iobuf;
-	}
-	
-	return 0;
-}
-
-void free_kiovec(int nr, struct kiobuf **bufp) 
-{
-	int i;
-	struct kiobuf *iobuf;
-	
-	for (i = 0; i < nr; i++) {
-		iobuf = bufp[i];
-		if (iobuf->locked)
-			unlock_kiovec(1, &iobuf);
-		if (iobuf->array_len > KIO_STATIC_PAGES)
-			kfree (iobuf->maplist);
-		kfree(bufp[i]);
-	}
-}
-
-int expand_kiobuf(struct kiobuf *iobuf, int wanted)
-{
-	struct page ** maplist;
-	
-	if (iobuf->array_len >= wanted)
-		return 0;
-	
-	maplist = (struct page **) 
-		kmalloc(wanted * sizeof(struct page **), GFP_KERNEL);
-	if (!maplist)
-		return -ENOMEM;
-
-	/* Did it grow while we waited? */
-	if (iobuf->array_len >= wanted) {
-		kfree(maplist);
-		return 0;
-	}
-	
-	memcpy (maplist, iobuf->maplist, iobuf->array_len * sizeof(struct page **));
-
-	if (iobuf->array_len > KIO_STATIC_PAGES)
-		kfree (iobuf->maplist);
-	
-	iobuf->maplist   = maplist;
-	iobuf->array_len = wanted;
-	return 0;
-}
-
-
-void kiobuf_wait_for_io(struct kiobuf *kiobuf)
-{
-	struct task_struct *tsk = current;
-	DECLARE_WAITQUEUE(wait, tsk);
-
-	if (atomic_read(&kiobuf->io_count) == 0)
-		return;
-
-	add_wait_queue(&kiobuf->wait_queue, &wait);
-repeat:
-	set_task_state(tsk, TASK_UNINTERRUPTIBLE);
-	if (atomic_read(&kiobuf->io_count) != 0) {
-		blk_run_queues();
-		schedule();
-		if (atomic_read(&kiobuf->io_count) != 0)
-			goto repeat;
-	}
-	tsk->state = TASK_RUNNING;
-	remove_wait_queue(&kiobuf->wait_queue, &wait);
-}
-
-
-

fs/open.c

@@ -14,7 +14,6 @@
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/tty.h>
-#include <linux/iobuf.h>
 #include <linux/namei.h>
 #include <linux/backing-dev.h>
 #include <linux/security.h>

fs/xfs/linux/xfs_aops.c

@@ -34,7 +34,6 @@
 #include <linux/mm.h>
 #include <linux/pagemap.h>
 #include <linux/mpage.h>
-#include <linux/iobuf.h>
 
 
 STATIC int delalloc_convert(struct inode *, struct page *, int, int);

fs/xfs/linux/xfs_ioctl.c

@@ -35,7 +35,6 @@
 #include <xfs_dfrag.h>
 #include <linux/dcache.h>
 #include <linux/namei.h>
-#include <linux/iobuf.h>
 
 
 extern int xfs_change_file_space(bhv_desc_t *, int,
@@ -605,6 +604,7 @@ xfs_ioctl(
 		 * it is set to the file system block size to
 		 * avoid having to do block zeroing on short writes.
 		 */
+#define KIO_MAX_ATOMIC_IO 512	/* FIXME: what do we really want here? */
 		da.d_maxiosz = XFS_FSB_TO_B(mp,
 				XFS_B_TO_FSBT(mp, KIO_MAX_ATOMIC_IO << 10));
 

include/linux/buffer_head.h

@@ -32,7 +32,6 @@ enum bh_state_bits {
 #define MAX_BUF_PER_PAGE (PAGE_CACHE_SIZE / 512)
 
 struct page;
-struct kiobuf;
 struct buffer_head;
 struct address_space;
 typedef void (bh_end_io_t)(struct buffer_head *bh, int uptodate);

include/linux/dcache.h

@@ -180,17 +180,6 @@ extern void shrink_dcache_parent(struct dentry *);
 extern void shrink_dcache_anon(struct list_head *);
 extern int d_invalidate(struct dentry *);
 
-/* dcache memory management */
-extern int shrink_dcache_memory(int, unsigned int);
-extern void prune_dcache(int);
-
-/* icache memory management (defined in linux/fs/inode.c) */
-extern int shrink_icache_memory(int, unsigned int);
-extern void prune_icache(int);
-
-/* quota cache memory management (defined in linux/fs/dquot.c) */
-extern int shrink_dqcache_memory(int, unsigned int);
-
 /* only used at mount-time */
 extern struct dentry * d_alloc_root(struct inode *);
 

include/linux/iobuf.h

-/*
- * iobuf.h
- *
- * Defines the structures used to track abstract kernel-space io buffers.
- *
- */
-
-#ifndef __LINUX_IOBUF_H
-#define __LINUX_IOBUF_H
-
-#include <linux/mm.h>
-#include <linux/init.h>
-#include <linux/wait.h>
-#include <asm/atomic.h>
-
-/*
- * The kiobuf structure describes a physical set of pages reserved
- * locked for IO.  The reference counts on each page will have been
- * incremented, and the flags field will indicate whether or not we have
- * pre-locked all of the pages for IO.
- *
- * kiobufs may be passed in arrays to form a kiovec, but we must
- * preserve the property that no page is present more than once over the
- * entire iovec.
- */
-
-#define KIO_MAX_ATOMIC_IO	512 /* in kb */
-#define KIO_STATIC_PAGES	(KIO_MAX_ATOMIC_IO / (PAGE_SIZE >> 10) + 1)
-#define KIO_MAX_SECTORS		(KIO_MAX_ATOMIC_IO * 2)
-
-/* The main kiobuf struct */
-
-struct kiobuf 
-{
-	int		nr_pages;	/* Pages actually referenced */
-	int		array_len;	/* Space in the allocated lists */
-	int		offset;		/* Offset to start of valid data */
-	int		length;		/* Number of valid bytes of data */
-
-	/* Keep separate track of the physical addresses and page
-	 * structs involved.  If we do IO to a memory-mapped device
-	 * region, there won't necessarily be page structs defined for
-	 * every address. */
-
-	struct page **	maplist;
-
-	unsigned int	locked : 1;	/* If set, pages has been locked */
-	
-	/* Always embed enough struct pages for atomic IO */
-	struct page *	map_array[KIO_STATIC_PAGES];
-	sector_t	blocks[KIO_MAX_SECTORS];
-
-	/* Dynamic state for IO completion: */
-	atomic_t	io_count;	/* IOs still in progress */
-	int		errno;		/* Status of completed IO */
-	void		(*end_io) (struct kiobuf *); /* Completion callback */
-	wait_queue_head_t wait_queue;
-};
-
-
-/* mm/memory.c */
-
-int	map_user_kiobuf(int rw, struct kiobuf *, unsigned long va, size_t len);
-void	unmap_kiobuf(struct kiobuf *iobuf);
-int	lock_kiovec(int nr, struct kiobuf *iovec[], int wait);
-int	unlock_kiovec(int nr, struct kiobuf *iovec[]);
-void	mark_dirty_kiobuf(struct kiobuf *iobuf, int bytes);
-
-/* fs/iobuf.c */
-
-int	end_kio_request(struct kiobuf *, int);
-void	simple_wakeup_kiobuf(struct kiobuf *);
-int	alloc_kiovec(int nr, struct kiobuf **);
-void	free_kiovec(int nr, struct kiobuf **);
-int	expand_kiobuf(struct kiobuf *, int);
-void	kiobuf_wait_for_io(struct kiobuf *);
-extern int alloc_kiobuf_bhs(struct kiobuf *);
-extern void free_kiobuf_bhs(struct kiobuf *);
-
-/* fs/buffer.c */
-
-int	brw_kiovec(int rw, int nr, struct kiobuf *iovec[], 
-		   struct block_device *bdev, sector_t [], int size);
-
-/* fs/bio.c */
-void	ll_rw_kio(int rw, struct kiobuf *kio, struct block_device *bdev, sector_t block);
-
-#endif /* __LINUX_IOBUF_H */

include/linux/mm.h

@@ -391,6 +391,29 @@ extern	int free_hugepages(struct vm_area_struct *);
 #endif
 
 
+/*
+ * Prototype to add a shrinker callback for ageable caches.
+ * 
+ * These functions are passed a count `nr_to_scan' and a gfpmask.  They should
+ * scan `nr_to_scan' objects, attempting to free them.
+ *
+ * The callback must the number of objects which remain in the cache.
+ *
+ * The callback will be passes nr_to_scan == 0 when the VM is querying the
+ * cache size, so a fastpath for that case is appropriate.
+ */
+typedef int (*shrinker_t)(int nr_to_scan, unsigned int gfp_mask);
+
+/*
+ * Add an aging callback.  The int is the number of 'seeks' it takes
+ * to recreate one of the objects that these functions age.
+ */
+
+#define DEFAULT_SEEKS 2
+struct shrinker;
+extern struct shrinker *set_shrinker(int, shrinker_t);
+extern void remove_shrinker(struct shrinker *shrinker);
+
 /*
  * If the mapping doesn't provide a set_page_dirty a_op, then
  * just fall through and assume that it wants buffer_heads.

init/main.c

@@ -24,7 +24,6 @@
 #include <linux/smp_lock.h>
 #include <linux/blk.h>
 #include <linux/hdreg.h>
-#include <linux/iobuf.h>
 #include <linux/bootmem.h>
 #include <linux/tty.h>
 #include <linux/percpu.h>

kernel/ksyms.c

@@ -33,7 +33,6 @@
 #include <linux/swap.h>
 #include <linux/ctype.h>
 #include <linux/file.h>
-#include <linux/iobuf.h>
 #include <linux/console.h>
 #include <linux/poll.h>
 #include <linux/mmzone.h>
@@ -103,6 +102,8 @@ EXPORT_SYMBOL(kmem_cache_shrink);
 EXPORT_SYMBOL(kmem_cache_alloc);
 EXPORT_SYMBOL(kmem_cache_free);
 EXPORT_SYMBOL(kmem_cache_size);
+EXPORT_SYMBOL(set_shrinker);
+EXPORT_SYMBOL(remove_shrinker);
 EXPORT_SYMBOL(kmalloc);
 EXPORT_SYMBOL(kfree);
 EXPORT_SYMBOL(vfree);
@@ -246,7 +247,6 @@ EXPORT_SYMBOL(dput);
 EXPORT_SYMBOL(have_submounts);
 EXPORT_SYMBOL(d_find_alias);
 EXPORT_SYMBOL(d_prune_aliases);
-EXPORT_SYMBOL(prune_dcache);
 EXPORT_SYMBOL(shrink_dcache_sb);
 EXPORT_SYMBOL(shrink_dcache_parent);
 EXPORT_SYMBOL(shrink_dcache_anon);
@@ -438,18 +438,6 @@ EXPORT_SYMBOL(__br_write_lock);
 EXPORT_SYMBOL(__br_write_unlock);
 #endif
 
-/* Kiobufs */
-EXPORT_SYMBOL(alloc_kiovec);
-EXPORT_SYMBOL(free_kiovec);
-EXPORT_SYMBOL(expand_kiobuf);
-
-EXPORT_SYMBOL(map_user_kiobuf);
-EXPORT_SYMBOL(unmap_kiobuf);
-EXPORT_SYMBOL(lock_kiovec);
-EXPORT_SYMBOL(unlock_kiovec);
-EXPORT_SYMBOL(brw_kiovec);
-EXPORT_SYMBOL(kiobuf_wait_for_io);
-
 #ifdef HAVE_DISABLE_HLT
 EXPORT_SYMBOL(disable_hlt);
 EXPORT_SYMBOL(enable_hlt);

mm/filemap.c

@@ -20,7 +20,6 @@
 #include <linux/pagemap.h>
 #include <linux/file.h>
 #include <linux/uio.h>
-#include <linux/iobuf.h>
 #include <linux/hash.h>
 #include <linux/writeback.h>
 #include <linux/pagevec.h>

mm/memory.c

@@ -40,7 +40,6 @@
 #include <linux/mm.h>
 #include <linux/mman.h>
 #include <linux/swap.h>
-#include <linux/iobuf.h>
 #include <linux/highmem.h>
 #include <linux/pagemap.h>
 #include <linux/vcache.h>
@@ -504,7 +503,7 @@ follow_page(struct mm_struct *mm, unsigned long address, int write)
 /* 
  * Given a physical address, is there a useful struct page pointing to
  * it?  This may become more complex in the future if we start dealing
- * with IO-aperture pages in kiobufs.
+ * with IO-aperture pages for direct-IO.
  */
 
 static inline struct page *get_page_map(struct page *page)
@@ -589,224 +588,6 @@ int get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
 	return i;
 }
 
-/*
- * Force in an entire range of pages from the current process's user VA,
- * and pin them in physical memory.  
- */
-#define dprintk(x...)
-
-int map_user_kiobuf(int rw, struct kiobuf *iobuf, unsigned long va, size_t len)
-{
-	int pgcount, err;
-	struct mm_struct *	mm;
-	
-	/* Make sure the iobuf is not already mapped somewhere. */
-	if (iobuf->nr_pages)
-		return -EINVAL;
-
-	mm = current->mm;
-	dprintk ("map_user_kiobuf: begin\n");
-	
-	pgcount = (va + len + PAGE_SIZE - 1)/PAGE_SIZE - va/PAGE_SIZE;
-	/* mapping 0 bytes is not permitted */
-	if (!pgcount) BUG();
-	err = expand_kiobuf(iobuf, pgcount);
-	if (err)
-		return err;
-
-	iobuf->locked = 0;
-	iobuf->offset = va & (PAGE_SIZE-1);
-	iobuf->length = len;
-	
-	/* Try to fault in all of the necessary pages */
-	down_read(&mm->mmap_sem);
-	/* rw==READ means read from disk, write into memory area */
-	err = get_user_pages(current, mm, va, pgcount,
-			(rw==READ), 0, iobuf->maplist, NULL);
-	up_read(&mm->mmap_sem);
-	if (err < 0) {
-		unmap_kiobuf(iobuf);
-		dprintk ("map_user_kiobuf: end %d\n", err);
-		return err;
-	}
-	iobuf->nr_pages = err;
-	while (pgcount--) {
-		/* FIXME: flush superflous for rw==READ,
-		 * probably wrong function for rw==WRITE
-		 */
-		flush_dcache_page(iobuf->maplist[pgcount]);
-	}
-	dprintk ("map_user_kiobuf: end OK\n");
-	return 0;
-}
-
-/*
- * Mark all of the pages in a kiobuf as dirty 
- *
- * We need to be able to deal with short reads from disk: if an IO error
- * occurs, the number of bytes read into memory may be less than the
- * size of the kiobuf, so we have to stop marking pages dirty once the
- * requested byte count has been reached.
- */
-
-void mark_dirty_kiobuf(struct kiobuf *iobuf, int bytes)
-{
-	int index, offset, remaining;
-	struct page *page;
-	
-	index = iobuf->offset >> PAGE_SHIFT;
-	offset = iobuf->offset & ~PAGE_MASK;
-	remaining = bytes;
-	if (remaining > iobuf->length)
-		remaining = iobuf->length;
-	
-	while (remaining > 0 && index < iobuf->nr_pages) {
-		page = iobuf->maplist[index];
-		
-		if (!PageReserved(page))
-			set_page_dirty(page);
-
-		remaining -= (PAGE_SIZE - offset);
-		offset = 0;
-		index++;
-	}
-}
-
-/*
- * Unmap all of the pages referenced by a kiobuf.  We release the pages,
- * and unlock them if they were locked. 
- */
-
-void unmap_kiobuf (struct kiobuf *iobuf) 
-{
-	int i;
-	struct page *map;
-	
-	for (i = 0; i < iobuf->nr_pages; i++) {
-		map = iobuf->maplist[i];
-		if (map) {
-			if (iobuf->locked)
-				unlock_page(map);
-			/* FIXME: cache flush missing for rw==READ
-			 * FIXME: call the correct reference counting function
-			 */
-			page_cache_release(map);
-		}
-	}
-	
-	iobuf->nr_pages = 0;
-	iobuf->locked = 0;
-}
-
-
-/*
- * Lock down all of the pages of a kiovec for IO.
- *
- * If any page is mapped twice in the kiovec, we return the error -EINVAL.
- *
- * The optional wait parameter causes the lock call to block until all
- * pages can be locked if set.  If wait==0, the lock operation is
- * aborted if any locked pages are found and -EAGAIN is returned.
- */
-
-int lock_kiovec(int nr, struct kiobuf *iovec[], int wait)
-{
-	struct kiobuf *iobuf;
-	int i, j;
-	struct page *page, **ppage;
-	int doublepage = 0;
-	int repeat = 0;
-	
- repeat:
-	
-	for (i = 0; i < nr; i++) {
-		iobuf = iovec[i];
-
-		if (iobuf->locked)
-			continue;
-
-		ppage = iobuf->maplist;
-		for (j = 0; j < iobuf->nr_pages; ppage++, j++) {
-			page = *ppage;
-			if (!page)
-				continue;
-			
-			if (TestSetPageLocked(page)) {
-				while (j--) {
-					struct page *tmp = *--ppage;
-					if (tmp)
-						unlock_page(tmp);
-				}
-				goto retry;
-			}
-		}
-		iobuf->locked = 1;
-	}
-
-	return 0;
-	
- retry:
-	
-	/* 
-	 * We couldn't lock one of the pages.  Undo the locking so far,
-	 * wait on the page we got to, and try again.  
-	 */
-	
-	unlock_kiovec(nr, iovec);
-	if (!wait)
-		return -EAGAIN;
-	
-	/* 
-	 * Did the release also unlock the page we got stuck on?
-	 */
-	if (!PageLocked(page)) {
-		/* 
-		 * If so, we may well have the page mapped twice
-		 * in the IO address range.  Bad news.  Of
-		 * course, it _might_ just be a coincidence,
-		 * but if it happens more than once, chances
-		 * are we have a double-mapped page. 
-		 */
-		if (++doublepage >= 3) 
-			return -EINVAL;
-		
-		/* Try again...  */
-		wait_on_page_locked(page);
-	}
-	
-	if (++repeat < 16)
-		goto repeat;
-	return -EAGAIN;
-}
-
-/*
- * Unlock all of the pages of a kiovec after IO.
- */
-
-int unlock_kiovec(int nr, struct kiobuf *iovec[])
-{
-	struct kiobuf *iobuf;
-	int i, j;
-	struct page *page, **ppage;
-	
-	for (i = 0; i < nr; i++) {
-		iobuf = iovec[i];
-
-		if (!iobuf->locked)
-			continue;
-		iobuf->locked = 0;
-		
-		ppage = iobuf->maplist;
-		for (j = 0; j < iobuf->nr_pages; ppage++, j++) {
-			page = *ppage;
-			if (!page)
-				continue;
-			unlock_page(page);
-		}
-	}
-	return 0;
-}
-
 static inline void zeromap_pte_range(pte_t * pte, unsigned long address,
                                      unsigned long size, pgprot_t prot)
 {

mm/msync.c

@@ -137,6 +137,9 @@ static int msync_interval(struct vm_area_struct * vma,
 	int ret = 0;
 	struct file * file = vma->vm_file;
 
+	if ((flags & MS_INVALIDATE) && (vma->vm_flags & VM_LOCKED))
+		return -EBUSY;
+
 	if (file && (vma->vm_flags & VM_SHARED)) {
 		ret = filemap_sync(vma, start, end-start, flags);
 
@@ -173,6 +176,8 @@ asmlinkage long sys_msync(unsigned long start, size_t len, int flags)
 		goto out;
 	if (start & ~PAGE_MASK)
 		goto out;
+	if ((flags & MS_ASYNC) && (flags & MS_SYNC))
+		goto out;
 	error = -ENOMEM;
 	len = (len + ~PAGE_MASK) & PAGE_MASK;
 	end = start + len;

mm/page_alloc.c

@@ -1147,6 +1147,8 @@ static void *vmstat_start(struct seq_file *m, loff_t *pos)
 	if (!ps)
 		return ERR_PTR(-ENOMEM);
 	get_full_page_state(ps);
+	ps->pgpgin /= 2;		/* sectors -> kbytes */
+	ps->pgpgout /= 2;
 	return (unsigned long *)ps + *pos;
 }
 

mm/vmscan.c

@@ -77,9 +77,94 @@ static long total_memory;
 #define prefetchw_prev_lru_page(_page, _base, _field) do { } while (0)
 #endif
 
-#ifndef CONFIG_QUOTA
-#define shrink_dqcache_memory(ratio, gfp_mask) do { } while (0)
-#endif
+/*
+ * The list of shrinker callbacks used by to apply pressure to
+ * ageable caches.
+ */
+struct shrinker {
+	shrinker_t		shrinker;
+	struct list_head	list;
+	int			seeks;	/* seeks to recreate an obj */
+	int			nr;	/* objs pending delete */
+};
+
+static LIST_HEAD(shrinker_list);
+static DECLARE_MUTEX(shrinker_sem);
+
+/*
+ * Add a shrinker callback to be called from the vm
+ */
+struct shrinker *set_shrinker(int seeks, shrinker_t theshrinker)
+{
+        struct shrinker *shrinker;
+
+        shrinker = kmalloc(sizeof(*shrinker), GFP_KERNEL);
+        if (shrinker) {
+	        shrinker->shrinker = theshrinker;
+	        shrinker->seeks = seeks;
+	        shrinker->nr = 0;
+	        down(&shrinker_sem);
+	        list_add(&shrinker->list, &shrinker_list);
+	        up(&shrinker_sem);
+	}
+	return shrinker;
+}
+
+/*
+ * Remove one
+ */
+void remove_shrinker(struct shrinker *shrinker)
+{
+	down(&shrinker_sem);
+	list_del(&shrinker->list);
+	up(&shrinker_sem);
+	kfree(shrinker);
+}
+ 
+#define SHRINK_BATCH 32
+/*
+ * Call the shrink functions to age shrinkable caches
+ *
+ * Here we assume it costs one seek to replace a lru page and that it also
+ * takes a seek to recreate a cache object.  With this in mind we age equal
+ * percentages of the lru and ageable caches.  This should balance the seeks
+ * generated by these structures.
+ *
+ * If the vm encounted mapped pages on the LRU it increase the pressure on
+ * slab to avoid swapping.
+ *
+ * FIXME: do not do for zone highmem
+ */
+static int shrink_slab(int scanned,  unsigned int gfp_mask)
+{
+	struct list_head *lh;
+	int pages;
+
+	if (down_trylock(&shrinker_sem))
+		return 0;
+
+	pages = nr_used_zone_pages();
+	list_for_each(lh, &shrinker_list) {
+		struct shrinker *shrinker;
+		int entries;
+		unsigned long delta;
+
+		shrinker = list_entry(lh, struct shrinker, list);
+		entries = (*shrinker->shrinker)(0, gfp_mask);
+		if (!entries)
+			continue;
+		delta = scanned * shrinker->seeks * entries;
+		shrinker->nr += delta / (pages + 1);
+		if (shrinker->nr > SHRINK_BATCH) {
+			int nr = shrinker->nr;
+
+			shrinker->nr = 0;
+			(*shrinker->shrinker)(nr, gfp_mask);
+		}
+	}
+	up(&shrinker_sem);
+	return 0;
+}
 
 /* Must be called with page's pte_chain_lock held. */
 static inline int page_mapping_inuse(struct page * page)
@@ -626,32 +711,6 @@ shrink_zone(struct zone *zone, int max_scan, unsigned int gfp_mask,
 				max_scan, nr_mapped);
 }
 
-/*
- * FIXME: don't do this for ZONE_HIGHMEM
- */
-/*
- * Here we assume it costs one seek to replace a lru page and that it also
- * takes a seek to recreate a cache object.  With this in mind we age equal
- * percentages of the lru and ageable caches.  This should balance the seeks
- * generated by these structures.
- *
- * NOTE: for now I do this for all zones.  If we find this is too aggressive
- * on large boxes we may want to exclude ZONE_HIGHMEM.
- *
- * If we're encountering mapped pages on the LRU then increase the pressure on
- * slab to avoid swapping.
- */
-static void shrink_slab(int total_scanned, int gfp_mask)
-{
-	int shrink_ratio;
-	int pages = nr_used_zone_pages();
-
-	shrink_ratio = (pages / (total_scanned + 1)) + 1;
-	shrink_dcache_memory(shrink_ratio, gfp_mask);
-	shrink_icache_memory(shrink_ratio, gfp_mask);
-	shrink_dqcache_memory(shrink_ratio, gfp_mask);
-}
-
 /*
  * This is the direct reclaim path, for page-allocating processes.  We only
  * try to reclaim pages from zones which will satisfy the caller's allocation
@@ -695,7 +754,7 @@ shrink_caches(struct zone *classzone, int priority, int *total_scanned,
 	}
 	return ret;
 }
-
+ 
 /*
  * This is the main entry point to direct page reclaim.
  *