From: Nick Piggin Date: Fri, 8 Feb 2008 12:19:49 +0000 (-0800) Subject: rewrite rd X-Git-Tag: v2.6.25-rc1~189 X-Git-Url: http://pilppa.com/gitweb/?a=commitdiff_plain;h=9db5579be4bb5320c3248f6acf807aedf05ae143;p=linux-2.6-omap-h63xx.git rewrite rd This is a rewrite of the ramdisk block device driver. The old one is really difficult because it effectively implements a block device which serves data out of its own buffer cache. It relies on the dirty bit being set, to pin its backing store in cache, however there are non trivial paths which can clear the dirty bit (eg. try_to_free_buffers()), which had recently lead to data corruption. And in general it is completely wrong for a block device driver to do this. The new one is more like a regular block device driver. It has no idea about vm/vfs stuff. It's backing store is similar to the buffer cache (a simple radix-tree of pages), but it doesn't know anything about page cache (the pages in the radix tree are not pagecache pages). There is one slight downside -- direct block device access and filesystem metadata access goes through an extra copy and gets stored in RAM twice. However, this downside is only slight, because the real buffercache of the device is now reclaimable (because we're not playing crazy games with it), so under memory intensive situations, footprint should effectively be the same -- maybe even a slight advantage to the new driver because it can also reclaim buffer heads. The fact that it now goes through all the regular vm/fs paths makes it much more useful for testing, too. text data bss dec hex filename 2837 849 384 4070 fe6 drivers/block/rd.o 3528 371 12 3911 f47 drivers/block/brd.o Text is larger, but data and bss are smaller, making total size smaller. A few other nice things about it: - Similar structure and layout to the new loop device handlinag. - Dynamic ramdisk creation. - Runtime flexible buffer head size (because it is no longer part of the ramdisk code). - Boot / load time flexible ramdisk size, which could easily be extended to a per-ramdisk runtime changeable size (eg. with an ioctl). - Can use highmem for the backing store. [akpm@linux-foundation.org: fix build] [byron.bbradley@gmail.com: make rd_size non-static] Signed-off-by: Nick Piggin Signed-off-by: Byron Bradley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- diff --git a/MAINTAINERS b/MAINTAINERS index a7e6ef3dae1..29f4f4dc145 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3254,6 +3254,11 @@ W: http://rt2x00.serialmonkey.com/ S: Maintained F: drivers/net/wireless/rt2x00/ +RAMDISK RAM BLOCK DEVICE DRIVER +P: Nick Piggin +M: npiggin@suse.de +S: Maintained + RANDOM NUMBER DRIVER P: Matt Mackall M: mpm@selenic.com diff --git a/drivers/block/Kconfig b/drivers/block/Kconfig index 64e5148d82b..8be67cd3fe0 100644 --- a/drivers/block/Kconfig +++ b/drivers/block/Kconfig @@ -322,7 +322,7 @@ config BLK_DEV_UB If unsure, say N. config BLK_DEV_RAM - tristate "RAM disk support" + tristate "RAM block device support" ---help--- Saying Y here will allow you to use a portion of your RAM memory as a block device, so that you can make file systems on it, read and @@ -357,16 +357,6 @@ config BLK_DEV_RAM_SIZE The default value is 4096 kilobytes. Only change this if you know what you are doing. -config BLK_DEV_RAM_BLOCKSIZE - int "Default RAM disk block size (bytes)" - depends on BLK_DEV_RAM - default "1024" - help - The default value is 1024 bytes. PAGE_SIZE is a much more - efficient choice however. The default is kept to ensure initrd - setups function - apparently needed by the rd_load_image routine - that supposes the filesystem in the image uses a 1024 blocksize. - config CDROM_PKTCDVD tristate "Packet writing on CD/DVD media" depends on !UML diff --git a/drivers/block/Makefile b/drivers/block/Makefile index 7691505a2e1..01c972415cb 100644 --- a/drivers/block/Makefile +++ b/drivers/block/Makefile @@ -11,7 +11,7 @@ obj-$(CONFIG_AMIGA_FLOPPY) += amiflop.o obj-$(CONFIG_PS3_DISK) += ps3disk.o obj-$(CONFIG_ATARI_FLOPPY) += ataflop.o obj-$(CONFIG_AMIGA_Z2RAM) += z2ram.o -obj-$(CONFIG_BLK_DEV_RAM) += rd.o +obj-$(CONFIG_BLK_DEV_RAM) += brd.o obj-$(CONFIG_BLK_DEV_LOOP) += loop.o obj-$(CONFIG_BLK_DEV_PS2) += ps2esdi.o obj-$(CONFIG_BLK_DEV_XD) += xd.o diff --git a/drivers/block/brd.c b/drivers/block/brd.c new file mode 100644 index 00000000000..50b659bedc8 --- /dev/null +++ b/drivers/block/brd.c @@ -0,0 +1,548 @@ +/* + * Ram backed block device driver. + * + * Copyright (C) 2007 Nick Piggin + * Copyright (C) 2007 Novell Inc. + * + * Parts derived from drivers/block/rd.c, and drivers/block/loop.c, copyright + * of their respective owners. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include /* invalidate_bh_lrus() */ + +#include + +#define SECTOR_SHIFT 9 +#define PAGE_SECTORS_SHIFT (PAGE_SHIFT - SECTOR_SHIFT) +#define PAGE_SECTORS (1 << PAGE_SECTORS_SHIFT) + +/* + * Each block ramdisk device has a radix_tree brd_pages of pages that stores + * the pages containing the block device's contents. A brd page's ->index is + * its offset in PAGE_SIZE units. This is similar to, but in no way connected + * with, the kernel's pagecache or buffer cache (which sit above our block + * device). + */ +struct brd_device { + int brd_number; + int brd_refcnt; + loff_t brd_offset; + loff_t brd_sizelimit; + unsigned brd_blocksize; + + struct request_queue *brd_queue; + struct gendisk *brd_disk; + struct list_head brd_list; + + /* + * Backing store of pages and lock to protect it. This is the contents + * of the block device. + */ + spinlock_t brd_lock; + struct radix_tree_root brd_pages; +}; + +/* + * Look up and return a brd's page for a given sector. + */ +static struct page *brd_lookup_page(struct brd_device *brd, sector_t sector) +{ + pgoff_t idx; + struct page *page; + + /* + * The page lifetime is protected by the fact that we have opened the + * device node -- brd pages will never be deleted under us, so we + * don't need any further locking or refcounting. + * + * This is strictly true for the radix-tree nodes as well (ie. we + * don't actually need the rcu_read_lock()), however that is not a + * documented feature of the radix-tree API so it is better to be + * safe here (we don't have total exclusion from radix tree updates + * here, only deletes). + */ + rcu_read_lock(); + idx = sector >> PAGE_SECTORS_SHIFT; /* sector to page index */ + page = radix_tree_lookup(&brd->brd_pages, idx); + rcu_read_unlock(); + + BUG_ON(page && page->index != idx); + + return page; +} + +/* + * Look up and return a brd's page for a given sector. + * If one does not exist, allocate an empty page, and insert that. Then + * return it. + */ +static struct page *brd_insert_page(struct brd_device *brd, sector_t sector) +{ + pgoff_t idx; + struct page *page; + + page = brd_lookup_page(brd, sector); + if (page) + return page; + + /* + * Must use NOIO because we don't want to recurse back into the + * block or filesystem layers from page reclaim. + */ + page = alloc_page(GFP_NOIO | __GFP_HIGHMEM | __GFP_ZERO); + if (!page) + return NULL; + + if (radix_tree_preload(GFP_NOIO)) { + __free_page(page); + return NULL; + } + + spin_lock(&brd->brd_lock); + idx = sector >> PAGE_SECTORS_SHIFT; + if (radix_tree_insert(&brd->brd_pages, idx, page)) { + __free_page(page); + page = radix_tree_lookup(&brd->brd_pages, idx); + BUG_ON(!page); + BUG_ON(page->index != idx); + } else + page->index = idx; + spin_unlock(&brd->brd_lock); + + radix_tree_preload_end(); + + return page; +} + +/* + * Free all backing store pages and radix tree. This must only be called when + * there are no other users of the device. + */ +#define FREE_BATCH 16 +static void brd_free_pages(struct brd_device *brd) +{ + unsigned long pos = 0; + struct page *pages[FREE_BATCH]; + int nr_pages; + + do { + int i; + + nr_pages = radix_tree_gang_lookup(&brd->brd_pages, + (void **)pages, pos, FREE_BATCH); + + for (i = 0; i < nr_pages; i++) { + void *ret; + + BUG_ON(pages[i]->index < pos); + pos = pages[i]->index; + ret = radix_tree_delete(&brd->brd_pages, pos); + BUG_ON(!ret || ret != pages[i]); + __free_page(pages[i]); + } + + pos++; + + /* + * This assumes radix_tree_gang_lookup always returns as + * many pages as possible. If the radix-tree code changes, + * so will this have to. + */ + } while (nr_pages == FREE_BATCH); +} + +/* + * copy_to_brd_setup must be called before copy_to_brd. It may sleep. + */ +static int copy_to_brd_setup(struct brd_device *brd, sector_t sector, size_t n) +{ + unsigned int offset = (sector & (PAGE_SECTORS-1)) << SECTOR_SHIFT; + size_t copy; + + copy = min_t(size_t, n, PAGE_SIZE - offset); + if (!brd_insert_page(brd, sector)) + return -ENOMEM; + if (copy < n) { + sector += copy >> SECTOR_SHIFT; + if (!brd_insert_page(brd, sector)) + return -ENOMEM; + } + return 0; +} + +/* + * Copy n bytes from src to the brd starting at sector. Does not sleep. + */ +static void copy_to_brd(struct brd_device *brd, const void *src, + sector_t sector, size_t n) +{ + struct page *page; + void *dst; + unsigned int offset = (sector & (PAGE_SECTORS-1)) << SECTOR_SHIFT; + size_t copy; + + copy = min_t(size_t, n, PAGE_SIZE - offset); + page = brd_lookup_page(brd, sector); + BUG_ON(!page); + + dst = kmap_atomic(page, KM_USER1); + memcpy(dst + offset, src, copy); + kunmap_atomic(dst, KM_USER1); + + if (copy < n) { + src += copy; + sector += copy >> SECTOR_SHIFT; + copy = n - copy; + page = brd_lookup_page(brd, sector); + BUG_ON(!page); + + dst = kmap_atomic(page, KM_USER1); + memcpy(dst, src, copy); + kunmap_atomic(dst, KM_USER1); + } +} + +/* + * Copy n bytes to dst from the brd starting at sector. Does not sleep. + */ +static void copy_from_brd(void *dst, struct brd_device *brd, + sector_t sector, size_t n) +{ + struct page *page; + void *src; + unsigned int offset = (sector & (PAGE_SECTORS-1)) << SECTOR_SHIFT; + size_t copy; + + copy = min_t(size_t, n, PAGE_SIZE - offset); + page = brd_lookup_page(brd, sector); + if (page) { + src = kmap_atomic(page, KM_USER1); + memcpy(dst, src + offset, copy); + kunmap_atomic(src, KM_USER1); + } else + memset(dst, 0, copy); + + if (copy < n) { + dst += copy; + sector += copy >> SECTOR_SHIFT; + copy = n - copy; + page = brd_lookup_page(brd, sector); + if (page) { + src = kmap_atomic(page, KM_USER1); + memcpy(dst, src, copy); + kunmap_atomic(src, KM_USER1); + } else + memset(dst, 0, copy); + } +} + +/* + * Process a single bvec of a bio. + */ +static int brd_do_bvec(struct brd_device *brd, struct page *page, + unsigned int len, unsigned int off, int rw, + sector_t sector) +{ + void *mem; + int err = 0; + + if (rw != READ) { + err = copy_to_brd_setup(brd, sector, len); + if (err) + goto out; + } + + mem = kmap_atomic(page, KM_USER0); + if (rw == READ) { + copy_from_brd(mem + off, brd, sector, len); + flush_dcache_page(page); + } else + copy_to_brd(brd, mem + off, sector, len); + kunmap_atomic(mem, KM_USER0); + +out: + return err; +} + +static int brd_make_request(struct request_queue *q, struct bio *bio) +{ + struct block_device *bdev = bio->bi_bdev; + struct brd_device *brd = bdev->bd_disk->private_data; + int rw; + struct bio_vec *bvec; + sector_t sector; + int i; + int err = -EIO; + + sector = bio->bi_sector; + if (sector + (bio->bi_size >> SECTOR_SHIFT) > + get_capacity(bdev->bd_disk)) + goto out; + + rw = bio_rw(bio); + if (rw == READA) + rw = READ; + + bio_for_each_segment(bvec, bio, i) { + unsigned int len = bvec->bv_len; + err = brd_do_bvec(brd, bvec->bv_page, len, + bvec->bv_offset, rw, sector); + if (err) + break; + sector += len >> SECTOR_SHIFT; + } + +out: + bio_endio(bio, err); + + return 0; +} + +static int brd_ioctl(struct inode *inode, struct file *file, + unsigned int cmd, unsigned long arg) +{ + int error; + struct block_device *bdev = inode->i_bdev; + struct brd_device *brd = bdev->bd_disk->private_data; + + if (cmd != BLKFLSBUF) + return -ENOTTY; + + /* + * ram device BLKFLSBUF has special semantics, we want to actually + * release and destroy the ramdisk data. + */ + mutex_lock(&bdev->bd_mutex); + error = -EBUSY; + if (bdev->bd_openers <= 1) { + /* + * Invalidate the cache first, so it isn't written + * back to the device. + * + * Another thread might instantiate more buffercache here, + * but there is not much we can do to close that race. + */ + invalidate_bh_lrus(); + truncate_inode_pages(bdev->bd_inode->i_mapping, 0); + brd_free_pages(brd); + error = 0; + } + mutex_unlock(&bdev->bd_mutex); + + return error; +} + +static struct block_device_operations brd_fops = { + .owner = THIS_MODULE, + .ioctl = brd_ioctl, +}; + +/* + * And now the modules code and kernel interface. + */ +static int rd_nr; +int rd_size = CONFIG_BLK_DEV_RAM_SIZE; +module_param(rd_nr, int, 0); +MODULE_PARM_DESC(rd_nr, "Maximum number of brd devices"); +module_param(rd_size, int, 0); +MODULE_PARM_DESC(rd_size, "Size of each RAM disk in kbytes."); +MODULE_LICENSE("GPL"); +MODULE_ALIAS_BLOCKDEV_MAJOR(RAMDISK_MAJOR); + +#ifndef MODULE +/* Legacy boot options - nonmodular */ +static int __init ramdisk_size(char *str) +{ + rd_size = simple_strtol(str, NULL, 0); + return 1; +} +static int __init ramdisk_size2(char *str) +{ + return ramdisk_size(str); +} +__setup("ramdisk=", ramdisk_size); +__setup("ramdisk_size=", ramdisk_size2); +#endif + +/* + * The device scheme is derived from loop.c. Keep them in synch where possible + * (should share code eventually). + */ +static LIST_HEAD(brd_devices); +static DEFINE_MUTEX(brd_devices_mutex); + +static struct brd_device *brd_alloc(int i) +{ + struct brd_device *brd; + struct gendisk *disk; + + brd = kzalloc(sizeof(*brd), GFP_KERNEL); + if (!brd) + goto out; + brd->brd_number = i; + spin_lock_init(&brd->brd_lock); + INIT_RADIX_TREE(&brd->brd_pages, GFP_ATOMIC); + + brd->brd_queue = blk_alloc_queue(GFP_KERNEL); + if (!brd->brd_queue) + goto out_free_dev; + blk_queue_make_request(brd->brd_queue, brd_make_request); + blk_queue_max_sectors(brd->brd_queue, 1024); + blk_queue_bounce_limit(brd->brd_queue, BLK_BOUNCE_ANY); + + disk = brd->brd_disk = alloc_disk(1); + if (!disk) + goto out_free_queue; + disk->major = RAMDISK_MAJOR; + disk->first_minor = i; + disk->fops = &brd_fops; + disk->private_data = brd; + disk->queue = brd->brd_queue; + sprintf(disk->disk_name, "ram%d", i); + set_capacity(disk, rd_size * 2); + + return brd; + +out_free_queue: + blk_cleanup_queue(brd->brd_queue); +out_free_dev: + kfree(brd); +out: + return NULL; +} + +static void brd_free(struct brd_device *brd) +{ + put_disk(brd->brd_disk); + blk_cleanup_queue(brd->brd_queue); + brd_free_pages(brd); + kfree(brd); +} + +static struct brd_device *brd_init_one(int i) +{ + struct brd_device *brd; + + list_for_each_entry(brd, &brd_devices, brd_list) { + if (brd->brd_number == i) + goto out; + } + + brd = brd_alloc(i); + if (brd) { + add_disk(brd->brd_disk); + list_add_tail(&brd->brd_list, &brd_devices); + } +out: + return brd; +} + +static void brd_del_one(struct brd_device *brd) +{ + list_del(&brd->brd_list); + del_gendisk(brd->brd_disk); + brd_free(brd); +} + +static struct kobject *brd_probe(dev_t dev, int *part, void *data) +{ + struct brd_device *brd; + struct kobject *kobj; + + mutex_lock(&brd_devices_mutex); + brd = brd_init_one(dev & MINORMASK); + kobj = brd ? get_disk(brd->brd_disk) : ERR_PTR(-ENOMEM); + mutex_unlock(&brd_devices_mutex); + + *part = 0; + return kobj; +} + +static int __init brd_init(void) +{ + int i, nr; + unsigned long range; + struct brd_device *brd, *next; + + /* + * brd module now has a feature to instantiate underlying device + * structure on-demand, provided that there is an access dev node. + * However, this will not work well with user space tool that doesn't + * know about such "feature". In order to not break any existing + * tool, we do the following: + * + * (1) if rd_nr is specified, create that many upfront, and this + * also becomes a hard limit. + * (2) if rd_nr is not specified, create 1 rd device on module + * load, user can further extend brd device by create dev node + * themselves and have kernel automatically instantiate actual + * device on-demand. + */ + if (rd_nr > 1UL << MINORBITS) + return -EINVAL; + + if (rd_nr) { + nr = rd_nr; + range = rd_nr; + } else { + nr = CONFIG_BLK_DEV_RAM_COUNT; + range = 1UL << MINORBITS; + } + + if (register_blkdev(RAMDISK_MAJOR, "ramdisk")) + return -EIO; + + for (i = 0; i < nr; i++) { + brd = brd_alloc(i); + if (!brd) + goto out_free; + list_add_tail(&brd->brd_list, &brd_devices); + } + + /* point of no return */ + + list_for_each_entry(brd, &brd_devices, brd_list) + add_disk(brd->brd_disk); + + blk_register_region(MKDEV(RAMDISK_MAJOR, 0), range, + THIS_MODULE, brd_probe, NULL, NULL); + + printk(KERN_INFO "brd: module loaded\n"); + return 0; + +out_free: + list_for_each_entry_safe(brd, next, &brd_devices, brd_list) { + list_del(&brd->brd_list); + brd_free(brd); + } + + unregister_blkdev(RAMDISK_MAJOR, "brd"); + return -ENOMEM; +} + +static void __exit brd_exit(void) +{ + unsigned long range; + struct brd_device *brd, *next; + + range = rd_nr ? rd_nr : 1UL << MINORBITS; + + list_for_each_entry_safe(brd, next, &brd_devices, brd_list) + brd_del_one(brd); + + blk_unregister_region(MKDEV(RAMDISK_MAJOR, 0), range); + unregister_blkdev(RAMDISK_MAJOR, "ramdisk"); +} + +module_init(brd_init); +module_exit(brd_exit); + diff --git a/drivers/block/rd.c b/drivers/block/rd.c deleted file mode 100644 index 06e23be7090..00000000000 --- a/drivers/block/rd.c +++ /dev/null @@ -1,537 +0,0 @@ -/* - * ramdisk.c - Multiple RAM disk driver - gzip-loading version - v. 0.8 beta. - * - * (C) Chad Page, Theodore Ts'o, et. al, 1995. - * - * This RAM disk is designed to have filesystems created on it and mounted - * just like a regular floppy disk. - * - * It also does something suggested by Linus: use the buffer cache as the - * RAM disk data. This makes it possible to dynamically allocate the RAM disk - * buffer - with some consequences I have to deal with as I write this. - * - * This code is based on the original ramdisk.c, written mostly by - * Theodore Ts'o (TYT) in 1991. The code was largely rewritten by - * Chad Page to use the buffer cache to store the RAM disk data in - * 1995; Theodore then took over the driver again, and cleaned it up - * for inclusion in the mainline kernel. - * - * The original CRAMDISK code was written by Richard Lyons, and - * adapted by Chad Page to use the new RAM disk interface. Theodore - * Ts'o rewrote it so that both the compressed RAM disk loader and the - * kernel decompressor uses the same inflate.c codebase. The RAM disk - * loader now also loads into a dynamic (buffer cache based) RAM disk, - * not the old static RAM disk. Support for the old static RAM disk has - * been completely removed. - * - * Loadable module support added by Tom Dyas. - * - * Further cleanups by Chad Page (page0588@sundance.sjsu.edu): - * Cosmetic changes in #ifdef MODULE, code movement, etc. - * When the RAM disk module is removed, free the protected buffers - * Default RAM disk size changed to 2.88 MB - * - * Added initrd: Werner Almesberger & Hans Lermen, Feb '96 - * - * 4/25/96 : Made RAM disk size a parameter (default is now 4 MB) - * - Chad Page - * - * Add support for fs images split across >1 disk, Paul Gortmaker, Mar '98 - * - * Make block size and block size shift for RAM disks a global macro - * and set blk_size for -ENOSPC, Werner Fink , Apr '99 - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include /* for invalidate_bdev() */ -#include -#include -#include -#include - -#include - -/* Various static variables go here. Most are used only in the RAM disk code. - */ - -static struct gendisk *rd_disks[CONFIG_BLK_DEV_RAM_COUNT]; -static struct block_device *rd_bdev[CONFIG_BLK_DEV_RAM_COUNT];/* Protected device data */ -static struct request_queue *rd_queue[CONFIG_BLK_DEV_RAM_COUNT]; - -/* - * Parameters for the boot-loading of the RAM disk. These are set by - * init/main.c (from arguments to the kernel command line) or from the - * architecture-specific setup routine (from the stored boot sector - * information). - */ -int rd_size = CONFIG_BLK_DEV_RAM_SIZE; /* Size of the RAM disks */ -/* - * It would be very desirable to have a soft-blocksize (that in the case - * of the ramdisk driver is also the hardblocksize ;) of PAGE_SIZE because - * doing that we'll achieve a far better MM footprint. Using a rd_blocksize of - * BLOCK_SIZE in the worst case we'll make PAGE_SIZE/BLOCK_SIZE buffer-pages - * unfreeable. With a rd_blocksize of PAGE_SIZE instead we are sure that only - * 1 page will be protected. Depending on the size of the ramdisk you - * may want to change the ramdisk blocksize to achieve a better or worse MM - * behaviour. The default is still BLOCK_SIZE (needed by rd_load_image that - * supposes the filesystem in the image uses a BLOCK_SIZE blocksize). - */ -static int rd_blocksize = CONFIG_BLK_DEV_RAM_BLOCKSIZE; - -/* - * Copyright (C) 2000 Linus Torvalds. - * 2000 Transmeta Corp. - * aops copied from ramfs. - */ - -/* - * If a ramdisk page has buffers, some may be uptodate and some may be not. - * To bring the page uptodate we zero out the non-uptodate buffers. The - * page must be locked. - */ -static void make_page_uptodate(struct page *page) -{ - if (page_has_buffers(page)) { - struct buffer_head *bh = page_buffers(page); - struct buffer_head *head = bh; - - do { - if (!buffer_uptodate(bh)) { - memset(bh->b_data, 0, bh->b_size); - /* - * akpm: I'm totally undecided about this. The - * buffer has just been magically brought "up to - * date", but nobody should want to be reading - * it anyway, because it hasn't been used for - * anything yet. It is still in a "not read - * from disk yet" state. - * - * But non-uptodate buffers against an uptodate - * page are against the rules. So do it anyway. - */ - set_buffer_uptodate(bh); - } - } while ((bh = bh->b_this_page) != head); - } else { - memset(page_address(page), 0, PAGE_CACHE_SIZE); - } - flush_dcache_page(page); - SetPageUptodate(page); -} - -static int ramdisk_readpage(struct file *file, struct page *page) -{ - if (!PageUptodate(page)) - make_page_uptodate(page); - unlock_page(page); - return 0; -} - -static int ramdisk_prepare_write(struct file *file, struct page *page, - unsigned offset, unsigned to) -{ - if (!PageUptodate(page)) - make_page_uptodate(page); - return 0; -} - -static int ramdisk_commit_write(struct file *file, struct page *page, - unsigned offset, unsigned to) -{ - set_page_dirty(page); - return 0; -} - -/* - * ->writepage to the blockdev's mapping has to redirty the page so that the - * VM doesn't go and steal it. We return AOP_WRITEPAGE_ACTIVATE so that the VM - * won't try to (pointlessly) write the page again for a while. - * - * Really, these pages should not be on the LRU at all. - */ -static int ramdisk_writepage(struct page *page, struct writeback_control *wbc) -{ - if (!PageUptodate(page)) - make_page_uptodate(page); - SetPageDirty(page); - if (wbc->for_reclaim) - return AOP_WRITEPAGE_ACTIVATE; - unlock_page(page); - return 0; -} - -/* - * This is a little speedup thing: short-circuit attempts to write back the - * ramdisk blockdev inode to its non-existent backing store. - */ -static int ramdisk_writepages(struct address_space *mapping, - struct writeback_control *wbc) -{ - return 0; -} - -/* - * ramdisk blockdev pages have their own ->set_page_dirty() because we don't - * want them to contribute to dirty memory accounting. - */ -static int ramdisk_set_page_dirty(struct page *page) -{ - if (!TestSetPageDirty(page)) - return 1; - return 0; -} - -/* - * releasepage is called by pagevec_strip/try_to_release_page if - * buffers_heads_over_limit is true. Without a releasepage function - * try_to_free_buffers is called instead. That can unset the dirty - * bit of our ram disk pages, which will be eventually freed, even - * if the page is still in use. - */ -static int ramdisk_releasepage(struct page *page, gfp_t dummy) -{ - return 0; -} - -static const struct address_space_operations ramdisk_aops = { - .readpage = ramdisk_readpage, - .prepare_write = ramdisk_prepare_write, - .commit_write = ramdisk_commit_write, - .writepage = ramdisk_writepage, - .set_page_dirty = ramdisk_set_page_dirty, - .writepages = ramdisk_writepages, - .releasepage = ramdisk_releasepage, -}; - -static int rd_blkdev_pagecache_IO(int rw, struct bio_vec *vec, sector_t sector, - struct address_space *mapping) -{ - pgoff_t index = sector >> (PAGE_CACHE_SHIFT - 9); - unsigned int vec_offset = vec->bv_offset; - int offset = (sector << 9) & ~PAGE_CACHE_MASK; - int size = vec->bv_len; - int err = 0; - - do { - int count; - struct page *page; - char *src; - char *dst; - - count = PAGE_CACHE_SIZE - offset; - if (count > size) - count = size; - size -= count; - - page = grab_cache_page(mapping, index); - if (!page) { - err = -ENOMEM; - goto out; - } - - if (!PageUptodate(page)) - make_page_uptodate(page); - - index++; - - if (rw == READ) { - src = kmap_atomic(page, KM_USER0) + offset; - dst = kmap_atomic(vec->bv_page, KM_USER1) + vec_offset; - } else { - src = kmap_atomic(vec->bv_page, KM_USER0) + vec_offset; - dst = kmap_atomic(page, KM_USER1) + offset; - } - offset = 0; - vec_offset += count; - - memcpy(dst, src, count); - - kunmap_atomic(src, KM_USER0); - kunmap_atomic(dst, KM_USER1); - - if (rw == READ) - flush_dcache_page(vec->bv_page); - else - set_page_dirty(page); - unlock_page(page); - put_page(page); - } while (size); - - out: - return err; -} - -/* - * Basically, my strategy here is to set up a buffer-head which can't be - * deleted, and make that my Ramdisk. If the request is outside of the - * allocated size, we must get rid of it... - * - * 19-JAN-1998 Richard Gooch Added devfs support - * - */ -static int rd_make_request(struct request_queue *q, struct bio *bio) -{ - struct block_device *bdev = bio->bi_bdev; - struct address_space * mapping = bdev->bd_inode->i_mapping; - sector_t sector = bio->bi_sector; - unsigned long len = bio->bi_size >> 9; - int rw = bio_data_dir(bio); - struct bio_vec *bvec; - int ret = 0, i; - - if (sector + len > get_capacity(bdev->bd_disk)) - goto fail; - - if (rw==READA) - rw=READ; - - bio_for_each_segment(bvec, bio, i) { - ret |= rd_blkdev_pagecache_IO(rw, bvec, sector, mapping); - sector += bvec->bv_len >> 9; - } - if (ret) - goto fail; - - bio_endio(bio, 0); - return 0; -fail: - bio_io_error(bio); - return 0; -} - -static int rd_ioctl(struct inode *inode, struct file *file, - unsigned int cmd, unsigned long arg) -{ - int error; - struct block_device *bdev = inode->i_bdev; - - if (cmd != BLKFLSBUF) - return -ENOTTY; - - /* - * special: we want to release the ramdisk memory, it's not like with - * the other blockdevices where this ioctl only flushes away the buffer - * cache - */ - error = -EBUSY; - mutex_lock(&bdev->bd_mutex); - if (bdev->bd_openers <= 2) { - truncate_inode_pages(bdev->bd_inode->i_mapping, 0); - error = 0; - } - mutex_unlock(&bdev->bd_mutex); - return error; -} - -/* - * This is the backing_dev_info for the blockdev inode itself. It doesn't need - * writeback and it does not contribute to dirty memory accounting. - */ -static struct backing_dev_info rd_backing_dev_info = { - .ra_pages = 0, /* No readahead */ - .capabilities = BDI_CAP_NO_ACCT_DIRTY | BDI_CAP_NO_WRITEBACK | BDI_CAP_MAP_COPY, - .unplug_io_fn = default_unplug_io_fn, -}; - -/* - * This is the backing_dev_info for the files which live atop the ramdisk - * "device". These files do need writeback and they do contribute to dirty - * memory accounting. - */ -static struct backing_dev_info rd_file_backing_dev_info = { - .ra_pages = 0, /* No readahead */ - .capabilities = BDI_CAP_MAP_COPY, /* Does contribute to dirty memory */ - .unplug_io_fn = default_unplug_io_fn, -}; - -static int rd_open(struct inode *inode, struct file *filp) -{ - unsigned unit = iminor(inode); - - if (rd_bdev[unit] == NULL) { - struct block_device *bdev = inode->i_bdev; - struct address_space *mapping; - unsigned bsize; - gfp_t gfp_mask; - - inode = igrab(bdev->bd_inode); - rd_bdev[unit] = bdev; - bdev->bd_openers++; - bsize = bdev_hardsect_size(bdev); - bdev->bd_block_size = bsize; - inode->i_blkbits = blksize_bits(bsize); - inode->i_size = get_capacity(bdev->bd_disk)<<9; - - mapping = inode->i_mapping; - mapping->a_ops = &ramdisk_aops; - mapping->backing_dev_info = &rd_backing_dev_info; - bdev->bd_inode_backing_dev_info = &rd_file_backing_dev_info; - - /* - * Deep badness. rd_blkdev_pagecache_IO() needs to allocate - * pagecache pages within a request_fn. We cannot recur back - * into the filesystem which is mounted atop the ramdisk, because - * that would deadlock on fs locks. And we really don't want - * to reenter rd_blkdev_pagecache_IO when we're already within - * that function. - * - * So we turn off __GFP_FS and __GFP_IO. - * - * And to give this thing a hope of working, turn on __GFP_HIGH. - * Hopefully, there's enough regular memory allocation going on - * for the page allocator emergency pools to keep the ramdisk - * driver happy. - */ - gfp_mask = mapping_gfp_mask(mapping); - gfp_mask &= ~(__GFP_FS|__GFP_IO); - gfp_mask |= __GFP_HIGH; - mapping_set_gfp_mask(mapping, gfp_mask); - } - - return 0; -} - -static struct block_device_operations rd_bd_op = { - .owner = THIS_MODULE, - .open = rd_open, - .ioctl = rd_ioctl, -}; - -/* - * Before freeing the module, invalidate all of the protected buffers! - */ -static void __exit rd_cleanup(void) -{ - int i; - - for (i = 0; i < CONFIG_BLK_DEV_RAM_COUNT; i++) { - struct block_device *bdev = rd_bdev[i]; - rd_bdev[i] = NULL; - if (bdev) { - invalidate_bdev(bdev); - blkdev_put(bdev); - } - del_gendisk(rd_disks[i]); - put_disk(rd_disks[i]); - blk_cleanup_queue(rd_queue[i]); - } - unregister_blkdev(RAMDISK_MAJOR, "ramdisk"); - - bdi_destroy(&rd_file_backing_dev_info); - bdi_destroy(&rd_backing_dev_info); -} - -/* - * This is the registration and initialization section of the RAM disk driver - */ -static int __init rd_init(void) -{ - int i; - int err; - - err = bdi_init(&rd_backing_dev_info); - if (err) - goto out2; - - err = bdi_init(&rd_file_backing_dev_info); - if (err) { - bdi_destroy(&rd_backing_dev_info); - goto out2; - } - - err = -ENOMEM; - - if (rd_blocksize > PAGE_SIZE || rd_blocksize < 512 || - !is_power_of_2(rd_blocksize)) { - printk("RAMDISK: wrong blocksize %d, reverting to defaults\n", - rd_blocksize); - rd_blocksize = BLOCK_SIZE; - } - - for (i = 0; i < CONFIG_BLK_DEV_RAM_COUNT; i++) { - rd_disks[i] = alloc_disk(1); - if (!rd_disks[i]) - goto out; - - rd_queue[i] = blk_alloc_queue(GFP_KERNEL); - if (!rd_queue[i]) { - put_disk(rd_disks[i]); - goto out; - } - } - - if (register_blkdev(RAMDISK_MAJOR, "ramdisk")) { - err = -EIO; - goto out; - } - - for (i = 0; i < CONFIG_BLK_DEV_RAM_COUNT; i++) { - struct gendisk *disk = rd_disks[i]; - - blk_queue_make_request(rd_queue[i], &rd_make_request); - blk_queue_hardsect_size(rd_queue[i], rd_blocksize); - - /* rd_size is given in kB */ - disk->major = RAMDISK_MAJOR; - disk->first_minor = i; - disk->fops = &rd_bd_op; - disk->queue = rd_queue[i]; - disk->flags |= GENHD_FL_SUPPRESS_PARTITION_INFO; - sprintf(disk->disk_name, "ram%d", i); - set_capacity(disk, rd_size * 2); - add_disk(rd_disks[i]); - } - - /* rd_size is given in kB */ - printk("RAMDISK driver initialized: " - "%d RAM disks of %dK size %d blocksize\n", - CONFIG_BLK_DEV_RAM_COUNT, rd_size, rd_blocksize); - - return 0; -out: - while (i--) { - put_disk(rd_disks[i]); - blk_cleanup_queue(rd_queue[i]); - } - bdi_destroy(&rd_backing_dev_info); - bdi_destroy(&rd_file_backing_dev_info); -out2: - return err; -} - -module_init(rd_init); -module_exit(rd_cleanup); - -/* options - nonmodular */ -#ifndef MODULE -static int __init ramdisk_size(char *str) -{ - rd_size = simple_strtol(str,NULL,0); - return 1; -} -static int __init ramdisk_blocksize(char *str) -{ - rd_blocksize = simple_strtol(str,NULL,0); - return 1; -} -__setup("ramdisk_size=", ramdisk_size); -__setup("ramdisk_blocksize=", ramdisk_blocksize); -#endif - -/* options - modular */ -module_param(rd_size, int, 0); -MODULE_PARM_DESC(rd_size, "Size of each RAM disk in kbytes."); -module_param(rd_blocksize, int, 0); -MODULE_PARM_DESC(rd_blocksize, "Blocksize of each RAM disk in bytes."); -MODULE_ALIAS_BLOCKDEV_MAJOR(RAMDISK_MAJOR); - -MODULE_LICENSE("GPL"); diff --git a/fs/buffer.c b/fs/buffer.c index 826baf4f04b..11b002e01d6 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -1436,6 +1436,7 @@ void invalidate_bh_lrus(void) { on_each_cpu(invalidate_bh_lru, NULL, 1, 1); } +EXPORT_SYMBOL_GPL(invalidate_bh_lrus); void set_bh_page(struct buffer_head *bh, struct page *page, unsigned long offset)