diff -urN lvm-ref/CREDITS lvm-aa/CREDITS --- lvm-ref/CREDITS Thu Nov 23 03:28:42 2000 +++ lvm-aa/CREDITS Sun Nov 26 05:00:08 2000 @@ -1440,6 +1440,13 @@ S: 2300 Copenhagen S S: Denmark +N: Heinz Mauelshagen +E: mge@EZ-Darmstadt.Telekom.de +D: Logical Volume Manager +S: Bartningstr. 12 +S: 64289 Darmstadt +S: Germany + N: Mike McLagan E: mike.mclagan@linux.org W: http://www.invlogic.com/~mmclagan diff -urN lvm-ref/Documentation/Configure.help lvm-aa/Documentation/Configure.help --- lvm-ref/Documentation/Configure.help Sun Nov 26 04:53:30 2000 +++ lvm-aa/Documentation/Configure.help Sun Nov 26 05:00:08 2000 @@ -1002,6 +1002,30 @@ called on26.o. You must also have a high-level driver for the type of device that you want to support. +Logical Volume Manager (LVM) support +CONFIG_BLK_DEV_LVM + This driver lets you combine several hard disks, hard disk partitions, + multiple devices or even loop devices (for evaluation purposes) into + a volume group. Imagine a volume group as a kind of virtual disk. + Logical volumes, which can be thought of as virtual partitions, + can be created in the volume group. You can resize volume groups and + logical volumes after creation time, corresponding to new capacity needs. + Logical volumes are accessed as block devices named + /dev/VolumeGroupName/LogicalVolumeName. + + For details see /usr/src/linux/Documentaion/LVM-HOWTO. + + To get the newest software see . + +Logical Volume Manager proc filesystem information +CONFIG_LVM_PROC_FS + If you say Y here, you are able to access overall Logical Volume Manager, + Volume Group, Logical and Physical Volume information in /proc/lvm. + + To use this option, you have to check, that the "proc filesystem support" + (CONFIG_PROC_FS) is enabled too. + + Multiple devices driver support CONFIG_BLK_DEV_MD This driver lets you combine several hard disk partitions into one diff -urN lvm-ref/Documentation/LVM-HOWTO lvm-aa/Documentation/LVM-HOWTO --- lvm-ref/Documentation/LVM-HOWTO Thu Jan 1 01:00:00 1970 +++ lvm-aa/Documentation/LVM-HOWTO Sun Nov 26 05:00:08 2000 @@ -0,0 +1,118 @@ +Heinz Mauelshagen's LVM (Logical Volume Manager) howto. 01/28/1999 + + +Abstract: +--------- +The LVM adds a kind of virtual disks and virtual partitions functionality +to the Linux operating system + +It achieves this by adding an additional layer between the physical peripherals +and the i/o interface in the kernel. + +This allows the concatenation of several disk partitions or total disks +(so-called physical volumes or PVs) or even multiple devices +to form a storage pool (so-called Volume Group or VG) with +allocation units called physical extents (called PE). +You can think of the volume group as a virtual disk. +Please see scenario below. + +Some or all PEs of this VG then can be allocated to so-called Logical Volumes +or LVs in units called logical extents or LEs. +Each LE is mapped to a corresponding PE. +LEs and PEs are equal in size. +Logical volumes are a kind of virtual partitions. + + +The LVs can be used through device special files similar to the known +/dev/sd[a-z]* or /dev/hd[a-z]* named /dev/VolumeGroupName/LogicalVolumeName. + +But going beyond this, you are able to extend or reduce +VGs _AND_ LVs at runtime! + +So... +If for example the capacity of a LV gets too small and your VG containing +this LV is full, you could add another PV to that VG and simply extend +the LV afterwards. +If you reduce or delete a LV you can use the freed capacity for different +LVs in the same VG. + + +The above scenario looks like this: + + /------------------------------------------\ + | /--PV2---\ VG 1 /--PVn---\ | + | |-VGDA---| |-VGDA-- | | + | |PE1PE2..| |PE1PE2..| | + | | | ...... | | | + | | | | | | + | | /-----------------------\ | | + | | \-------LV 1------------/ | | + | | ..PEn| | ..PEn| | + | \--------/ \--------/ | + \------------------------------------------/ + +PV 1 could be /dev/sdc1 sized 3GB +PV n could be /dev/sde1 sized 4GB +VG 1 could be test_vg +LV 1 could be /dev/test_vg/test_lv +VGDA is the volume group descriptor area holding the LVM metadata +PE1 up to PEn is the number of physical extents on each disk(partition) + + + +Installation steps see INSTALL and insmod(1)/modprobe(1), kmod/kerneld(8) +to load the logical volume manager module if you did not bind it +into the kernel. + + +Configuration steps for getting the above scenario: + +1. Set the partition system id to 0xFE on /dev/sdc1 and /dev/sde1. + +2. do a "pvcreate /dev/sd[ce]1" + For testing purposes you can use more than one partition on a disk. + You should not use more than one partition because in the case of + a striped LV you'll have a performance breakdown. + +3. do a "vgcreate test_vg /dev/sd[ce]1" to create the new VG named "test_vg" + which has the total capacity of both partitions. + vgcreate activates (transfers the metadata into the LVM driver in the kernel) + the new volume group too to be able to create LVs in the next step. + +4. do a "lvcreate -L1500 -ntest_lv test_vg" to get a 1500MB linear LV named + "test_lv" and it's block device special "/dev/test_vg/test_lv". + + Or do a "lvcreate -i2 -I4 -l1500 -nanother_test_lv test_vg" to get a 100 LE + large logical volume with 2 stripes and stripesize 4 KB. + +5. For example generate a filesystem in one LV with + "mke2fs /dev/test_vg/test_lv" and mount it. + +6. extend /dev/test_vg/test_lv to 1600MB with relative size by + "lvextend -L+100 /dev/test_vg/test_lv" + or with absolute size by + "lvextend -L1600 /dev/test_vg/test_lv" + +7. reduce /dev/test_vg/test_lv to 900 logical extents with relative extents by + "lvreduce -l-700 /dev/test_vg/test_lv" + or with absolute extents by + "lvreduce -l900 /dev/test_vg/test_lv" + +9. rename a VG by deactivating it with + "vgchange -an test_vg" # only VGs with _no_ open LVs can be deactivated! + "vgrename test_vg whatever" + and reactivate it again by + "vgchange -ay whatever" + +9. rename a LV after closing it by + "lvchange -an /dev/whatever/test_lv" # only closed LVs can be deactivated + "lvrename /dev/whatever/test_lv /dev/whatever/whatvolume" + or by + "lvrename whatever test_lv whatvolume" + and reactivate it again by + "lvchange -ay /dev/whatever/whatvolume" + +10. if you have the resize2fs program from e2fsprogs 1.19 or later and/or the + GNU ext2resize tools, you are able to resize the ext2 type filesystems + contained in logical volumes without destroyiing the data by + "e2fsadm -L+100 /dev/test_vg/another_test_lv" diff -urN lvm-ref/MAINTAINERS lvm-aa/MAINTAINERS --- lvm-ref/MAINTAINERS Thu Nov 23 03:28:43 2000 +++ lvm-aa/MAINTAINERS Sun Nov 26 05:00:08 2000 @@ -602,6 +602,13 @@ W: http://people.redhat.com/zab/maestro/ S: Supported +LOGICAL VOLUME MANAGER +P: Heinz Mauelshagen +M: linux-LVM@EZ-Darmstadt.Telekom.de +L: linux-LVM@sistina.com +W: http://www.sistina.com/lvm +S: Maintained + M68K P: Jes Sorensen M: Jes.Sorensen@cern.ch diff -urN lvm-ref/drivers/block/Config.in lvm-aa/drivers/block/Config.in --- lvm-ref/drivers/block/Config.in Thu Nov 23 03:28:43 2000 +++ lvm-aa/drivers/block/Config.in Sun Nov 26 05:00:08 2000 @@ -97,6 +97,10 @@ comment 'Additional Block Devices' +tristate 'Logical volume manager (LVM) support' CONFIG_BLK_DEV_LVM N +if [ "$CONFIG_BLK_DEV_LVM" != "n" ]; then + bool ' LVM information in proc filesystem' CONFIG_LVM_PROC_FS Y +fi tristate 'Loopback device support' CONFIG_BLK_DEV_LOOP if [ "$CONFIG_NET" = "y" ]; then tristate 'Network block device support' CONFIG_BLK_DEV_NBD diff -urN lvm-ref/drivers/block/Makefile lvm-aa/drivers/block/Makefile --- lvm-ref/drivers/block/Makefile Thu Nov 23 03:28:43 2000 +++ lvm-aa/drivers/block/Makefile Sun Nov 26 05:00:08 2000 @@ -266,6 +266,14 @@ endif endif +ifeq ($(CONFIG_BLK_DEV_LVM),y) +L_OBJS += lvm.o lvm-snap.o +else + ifeq ($(CONFIG_BLK_DEV_LVM),m) + M_OBJS += lvm-mod.o + endif +endif + ifeq ($(CONFIG_BLK_DEV_MD),y) LX_OBJS += md.o @@ -321,6 +329,9 @@ endif include $(TOPDIR)/Rules.make + +lvm-mod.o: lvm.o lvm-snap.o + $(LD) -r -o $@ lvm.o lvm-snap.o ide-mod.o: ide.o $(IDE_OBJS) $(LD) $(LD_RFLAG) -r -o $@ ide.o $(IDE_OBJS) diff -urN lvm-ref/drivers/block/README.lvm lvm-aa/drivers/block/README.lvm --- lvm-ref/drivers/block/README.lvm Thu Jan 1 01:00:00 1970 +++ lvm-aa/drivers/block/README.lvm Sun Nov 26 05:00:08 2000 @@ -0,0 +1,8 @@ + +This is the Logical Volume Manager driver for Linux, + +Tools, library that manage logical volumes can be found +at . + +There you can obtain actual driver versions too. + diff -urN lvm-ref/drivers/block/genhd.c lvm-aa/drivers/block/genhd.c --- lvm-ref/drivers/block/genhd.c Thu Nov 23 03:28:43 2000 +++ lvm-aa/drivers/block/genhd.c Sun Nov 26 05:00:08 2000 @@ -54,6 +54,11 @@ le32_to_cpu(__a); \ }) +#if defined CONFIG_BLK_DEV_LVM || defined CONFIG_BLK_DEV_LVM_MODULE +#include +void ( *lvm_hd_name_ptr) ( char *, int) = NULL; +#endif + struct gendisk *gendisk_head = NULL; static int current_minor = 0; @@ -135,6 +140,14 @@ break; case MD_MAJOR: unit -= 'a'-'0'; + break; +#if defined CONFIG_BLK_DEV_LVM || defined CONFIG_BLK_DEV_LVM_MODULE + case LVM_BLK_MAJOR: + *buf = 0; + if ( lvm_hd_name_ptr != NULL) + ( lvm_hd_name_ptr) ( buf, minor); + return buf; +#endif } part = minor & ((1 << hd->minor_shift) - 1); if (hd->major >= SCSI_DISK1_MAJOR && hd->major <= SCSI_DISK7_MAJOR) { diff -urN lvm-ref/drivers/block/ll_rw_blk.c lvm-aa/drivers/block/ll_rw_blk.c --- lvm-ref/drivers/block/ll_rw_blk.c Sun Nov 26 04:53:17 2000 +++ lvm-aa/drivers/block/ll_rw_blk.c Sun Nov 26 05:00:08 2000 @@ -26,6 +26,14 @@ #include +#if defined CONFIG_BLK_DEV_LVM || defined CONFIG_BLK_DEV_LVM_MODULE +#include + /* function pointer to the LVM driver remapping function + which will be setup during driver/module init; neccessary + to be able to load LVM as a module */ +int (*lvm_map_ptr) (struct buffer_head *, int) = NULL; +#endif + /* * The request-struct contains all necessary data * to load a nr of sectors into memory @@ -660,6 +668,9 @@ if (!req) { /* MD and loop can't handle plugging without deadlocking */ if (major != MD_MAJOR && major != LOOP_MAJOR && +#if defined CONFIG_BLK_DEV_LVM || defined CONFIG_BLK_DEV_LVM_MODULE + major != LVM_BLK_MAJOR && +#endif major != DDV_MAJOR && major != NBD_MAJOR) plug_device(blk_dev + major); /* is atomic */ } else switch (major) { @@ -829,13 +840,34 @@ correct_size, bh[i]->b_size); goto sorry; } - - /* Md remaps blocks now */ + /* LVM and MD remap blocks now */ +#if defined CONFIG_BLK_DEV_LVM || defined CONFIG_BLK_DEV_LVM_MODULE + major = MAJOR(bh[i]->b_dev); + if (major == LVM_BLK_MAJOR) { + if (lvm_map_ptr == NULL) { + printk(KERN_ERR + "Bad lvm_map_ptr in ll_rw_block\n"); + goto sorry; + } + if ((lvm_map_ptr) (bh[i], rw) != 0) { + printk(KERN_ERR + "Bad lvm_map in ll_rw_block\n"); + goto sorry; + } + /* remap major too ... */ + major = MAJOR(bh[i]->b_rdev); + } else { + bh[i]->b_rdev = bh[i]->b_dev; + bh[i]->b_rsector = bh[i]->b_blocknr * (bh[i]->b_size >> 9); + } +#else bh[i]->b_rdev = bh[i]->b_dev; bh[i]->b_rsector=bh[i]->b_blocknr*(bh[i]->b_size >> 9); +#endif #ifdef CONFIG_BLK_DEV_MD if (major==MD_MAJOR && - md_map (MINOR(bh[i]->b_dev), &bh[i]->b_rdev, + /* changed v to allow LVM to remap */ + md_map (MINOR(bh[i]->b_rdev), &bh[i]->b_rdev, &bh[i]->b_rsector, bh[i]->b_size >> 9)) { printk (KERN_ERR "Bad md_map in ll_rw_block\n"); @@ -854,8 +886,10 @@ if (bh[i]) { set_bit(BH_Req, &bh[i]->b_state); #ifdef CONFIG_BLK_DEV_MD - if (MAJOR(bh[i]->b_dev) == MD_MAJOR) { - md_make_request(MINOR (bh[i]->b_dev), rw, bh[i]); + /* changed v to allow LVM to remap */ + if (MAJOR(bh[i]->b_rdev) == MD_MAJOR) { + /* remap device for MD too v */ + md_make_request(MINOR (bh[i]->b_rdev), rw, bh[i]); continue; } #endif @@ -1040,6 +1074,9 @@ #ifdef CONFIG_SJCD sjcd_init(); #endif CONFIG_SJCD +#ifdef CONFIG_BLK_DEV_LVM + lvm_init(); +#endif #ifdef CONFIG_BLK_DEV_MD md_init(); #endif CONFIG_BLK_DEV_MD diff -urN lvm-ref/drivers/block/lvm-snap.c lvm-aa/drivers/block/lvm-snap.c --- lvm-ref/drivers/block/lvm-snap.c Thu Jan 1 01:00:00 1970 +++ lvm-aa/drivers/block/lvm-snap.c Sun Nov 26 05:00:08 2000 @@ -0,0 +1,431 @@ +/* + * kernel/lvm-snap.c + * + * Copyright (C) 2000 Andrea Arcangeli SuSE + * + * LVM snapshot driver is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * LVM driver is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with GNU CC; see the file COPYING. If not, write to + * the Free Software Foundation, 59 Temple Place - Suite 330, + * Boston, MA 02111-1307, USA. + * + */ + +#include +#include +#include +#include +#include +#include +#include + + +static char *lvm_snap_version __attribute ((unused)) = "LVM 0.8final (15/02/2000)\n"; + +extern const char *const lvm_name; +extern int lvm_blocksizes[]; + +void lvm_snapshot_release(lv_t *); + +#define hashfn(dev,block,mask,chunk_size) \ + ((HASHDEV(dev)^((block)/(chunk_size))) & (mask)) + +static inline lv_block_exception_t * +lvm_find_exception_table(kdev_t org_dev, unsigned long org_start, lv_t * lv) +{ + struct list_head * hash_table = lv->lv_snapshot_hash_table, * next; + unsigned long mask = lv->lv_snapshot_hash_mask; + int chunk_size = lv->lv_chunk_size; + lv_block_exception_t * ret; + int i = 0; + + hash_table = &hash_table[hashfn(org_dev, org_start, mask, chunk_size)]; + ret = NULL; + for (next = hash_table->next; next != hash_table; next = next->next) + { + lv_block_exception_t * exception; + + exception = list_entry(next, lv_block_exception_t, hash); + if (exception->rsector_org == org_start && + exception->rdev_org == org_dev) + { + if (i) + { + /* fun, isn't it? :) */ + list_del(next); + list_add(next, hash_table); + } + ret = exception; + break; + } + i++; + } + return ret; +} + +static inline void lvm_hash_link(lv_block_exception_t * exception, + kdev_t org_dev, unsigned long org_start, + lv_t * lv) +{ + struct list_head * hash_table = lv->lv_snapshot_hash_table; + unsigned long mask = lv->lv_snapshot_hash_mask; + int chunk_size = lv->lv_chunk_size; + + hash_table = &hash_table[hashfn(org_dev, org_start, mask, chunk_size)]; + list_add(&exception->hash, hash_table); +} + +int lvm_snapshot_remap_block(kdev_t * org_dev, unsigned long * org_sector, + unsigned long pe_start, lv_t * lv) +{ + int ret; + unsigned long pe_off, pe_adjustment, __org_start; + kdev_t __org_dev; + int chunk_size = lv->lv_chunk_size; + lv_block_exception_t * exception; + + pe_off = pe_start % chunk_size; + pe_adjustment = (*org_sector-pe_off) % chunk_size; + __org_start = *org_sector - pe_adjustment; + __org_dev = *org_dev; + + ret = 0; + exception = lvm_find_exception_table(__org_dev, __org_start, lv); + if (exception) + { + *org_dev = exception->rdev_new; + *org_sector = exception->rsector_new + pe_adjustment; + ret = 1; + } + return ret; +} + +static void lvm_drop_snapshot(lv_t * lv_snap, const char * reason) +{ + kdev_t last_dev; + int i; + + /* no exception storage space available for this snapshot + or error on this snapshot --> release it */ + invalidate_buffers(lv_snap->lv_dev); + + last_dev = 0; + for (i = 0; i < lv_snap->lv_remap_ptr; i++) { + if ( lv_snap->lv_block_exception[i].rdev_new != last_dev) { + last_dev = lv_snap->lv_block_exception[i].rdev_new; + invalidate_buffers(last_dev); + } + } + + lvm_snapshot_release(lv_snap); + + printk(KERN_INFO + "%s -- giving up to snapshot %s on %s due %s\n", + lvm_name, lv_snap->lv_snapshot_org->lv_name, lv_snap->lv_name, + reason); +} + +static inline void lvm_snapshot_prepare_blocks(unsigned long * blocks, + unsigned long start, + int nr_sectors, + int blocksize) +{ + int i, sectors_per_block, nr_blocks; + + sectors_per_block = blocksize >> 9; + nr_blocks = nr_sectors / sectors_per_block; + start /= sectors_per_block; + + for (i = 0; i < nr_blocks; i++) + blocks[i] = start++; +} + +static inline int get_blksize(kdev_t dev) +{ + int correct_size = BLOCK_SIZE, i, major; + + major = MAJOR(dev); + if (blksize_size[major]) + { + i = blksize_size[major][MINOR(dev)]; + if (i) + correct_size = i; + } + return correct_size; +} + +#ifdef DEBUG_SNAPSHOT +static inline void invalidate_snap_cache(unsigned long start, unsigned long nr, + kdev_t dev) +{ + struct buffer_head * bh; + int sectors_per_block, i, blksize, minor; + + minor = MINOR(dev); + blksize = lvm_blocksizes[minor]; + sectors_per_block = blksize >> 9; + nr /= sectors_per_block; + start /= sectors_per_block; + + for (i = 0; i < nr; i++) + { + bh = get_hash_table(dev, start++, blksize); + if (bh) + bforget(bh); + } +} +#endif + +/* + * copy on write handler for one snapshot logical volume + * + * read the original blocks and store it/them on the new one(s). + * if there is no exception storage space free any longer --> release snapshot. + * + * this routine gets called for each _first_ write to a physical chunk. + */ +int lvm_snapshot_COW(kdev_t org_phys_dev, + unsigned long org_phys_sector, + unsigned long org_pe_start, + unsigned long org_virt_sector, + lv_t * lv_snap) +{ + const char * reason; + unsigned long org_start, snap_start, virt_start, pe_off; + int idx = lv_snap->lv_remap_ptr, chunk_size = lv_snap->lv_chunk_size; + kdev_t snap_phys_dev; + struct kiobuf * iobuf; + unsigned long blocks[KIO_MAX_SECTORS]; + int blksize_snap, blksize_org, min_blksize, max_blksize; + int max_sectors, nr_sectors; + + /* check if we are out of snapshot space */ + if (idx >= lv_snap->lv_remap_end) + goto fail_out_of_space; + + /* calculate physical boundaries of source chunk */ + pe_off = org_pe_start % chunk_size; + org_start = org_phys_sector - ((org_phys_sector-pe_off) % chunk_size); + virt_start = org_virt_sector - (org_phys_sector - org_start); + + /* calculate physical boundaries of destination chunk */ + snap_phys_dev = lv_snap->lv_block_exception[idx].rdev_new; + snap_start = lv_snap->lv_block_exception[idx].rsector_new; + +#ifdef DEBUG_SNAPSHOT + printk(KERN_INFO + "%s -- COW: " + "org %02d:%02d faulting %lu start %lu, " + "snap %02d:%02d start %lu, " + "size %d, pe_start %lu pe_off %lu, virt_sec %lu\n", + lvm_name, + MAJOR(org_phys_dev), MINOR(org_phys_dev), org_phys_sector, + org_start, + MAJOR(snap_phys_dev), MINOR(snap_phys_dev), snap_start, + chunk_size, + org_pe_start, pe_off, + org_virt_sector); +#endif + + iobuf = lv_snap->lv_iobuf; + + blksize_org = get_blksize(org_phys_dev); + blksize_snap = get_blksize(snap_phys_dev); + max_blksize = max(blksize_org, blksize_snap); + min_blksize = min(blksize_org, blksize_snap); + max_sectors = KIO_MAX_SECTORS * (min_blksize>>9); + + if (chunk_size % (max_blksize>>9)) + goto fail_blksize; + + while (chunk_size) + { + nr_sectors = min(chunk_size, max_sectors); + chunk_size -= nr_sectors; + + iobuf->length = nr_sectors << 9; + + lvm_snapshot_prepare_blocks(blocks, org_start, + nr_sectors, blksize_org); + if (brw_kiovec(READ, 1, &iobuf, org_phys_dev, + blocks, blksize_org) != (nr_sectors<<9)) + goto fail_raw_read; + + lvm_snapshot_prepare_blocks(blocks, snap_start, + nr_sectors, blksize_snap); + if (brw_kiovec(WRITE, 1, &iobuf, snap_phys_dev, + blocks, blksize_snap) != (nr_sectors<<9)) + goto fail_raw_write; + } + +#ifdef DEBUG_SNAPSHOT + /* invalidate the logcial snapshot buffer cache */ + invalidate_snap_cache(virt_start, lv_snap->lv_chunk_size, + lv_snap->lv_dev); +#endif + + /* the original chunk is now stored on the snapshot volume + so update the execption table */ + lv_snap->lv_block_exception[idx].rdev_org = org_phys_dev; + lv_snap->lv_block_exception[idx].rsector_org = org_start; + lvm_hash_link(lv_snap->lv_block_exception + idx, + org_phys_dev, org_start, lv_snap); + lv_snap->lv_remap_ptr = idx + 1; + return 0; + + /* slow path */ + out: + lvm_drop_snapshot(lv_snap, reason); + return -1; + + fail_out_of_space: + reason = "out of space"; + goto out; + fail_raw_read: + reason = "read error"; + goto out; + fail_raw_write: + reason = "write error"; + goto out; + fail_blksize: + reason = "blocksize error"; + goto out; +} + +static int lvm_snapshot_alloc_iobuf_pages(struct kiobuf * iobuf, int sectors) +{ + int bytes, nr_pages, err, i; + + bytes = sectors << 9; + nr_pages = (bytes + ~PAGE_MASK) >> PAGE_SHIFT; + err = expand_kiobuf(iobuf, nr_pages); + if (err) + goto out; + + err = -ENOMEM; + iobuf->locked = 0; + iobuf->nr_pages = 0; + for (i = 0; i < nr_pages; i++) + { + struct page * page; + +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,3,27) + page = alloc_page(GFP_KERNEL); + if (!page) + goto out; +#else + { + unsigned long addr = __get_free_page(GFP_USER); + if (!addr) + goto out; + iobuf->pagelist[i] = addr; + page = mem_map + MAP_NR(addr); + } +#endif + + iobuf->maplist[i] = page; + iobuf->nr_pages++; + } + iobuf->offset = 0; + + err = 0; + out: + return err; +} + +static int calc_max_buckets(void) +{ + unsigned long mem; + + mem = num_physpages << PAGE_SHIFT; + mem /= 100; + mem *= 2; + mem /= sizeof(struct list_head); + + return mem; +} + +static int lvm_snapshot_alloc_hash_table(lv_t * lv) +{ + int err; + unsigned long buckets, max_buckets, size; + struct list_head * hash; + + buckets = lv->lv_remap_end; + max_buckets = calc_max_buckets(); + buckets = min(buckets, max_buckets); + while (buckets & (buckets-1)) + buckets &= (buckets-1); + + size = buckets * sizeof(struct list_head); + + err = -ENOMEM; + hash = vmalloc(size); + lv->lv_snapshot_hash_table = hash; + + if (!hash) + goto out; + + lv->lv_snapshot_hash_mask = buckets-1; + while (buckets--) + INIT_LIST_HEAD(hash+buckets); + err = 0; + out: + return err; +} + +int lvm_snapshot_alloc(lv_t * lv_snap) +{ + int err, blocksize, max_sectors; + + err = alloc_kiovec(1, &lv_snap->lv_iobuf); + if (err) + goto out; + + blocksize = lvm_blocksizes[MINOR(lv_snap->lv_dev)]; + max_sectors = KIO_MAX_SECTORS << (PAGE_SHIFT-9); + + err = lvm_snapshot_alloc_iobuf_pages(lv_snap->lv_iobuf, max_sectors); + if (err) + goto out_free_kiovec; + + err = lvm_snapshot_alloc_hash_table(lv_snap); + if (err) + goto out_free_kiovec; + out: + return err; + + out_free_kiovec: + unmap_kiobuf(lv_snap->lv_iobuf); + free_kiovec(1, &lv_snap->lv_iobuf); + goto out; +} + +void lvm_snapshot_release(lv_t * lv) +{ + if (lv->lv_block_exception) + { + vfree(lv->lv_block_exception); + lv->lv_block_exception = NULL; + } + if (lv->lv_snapshot_hash_table) + { + vfree(lv->lv_snapshot_hash_table); + lv->lv_snapshot_hash_table = NULL; + } + if (lv->lv_iobuf) + { + unmap_kiobuf(lv->lv_iobuf); + free_kiovec(1, &lv->lv_iobuf); + lv->lv_iobuf = NULL; + } +} diff -urN lvm-ref/drivers/block/lvm.c lvm-aa/drivers/block/lvm.c --- lvm-ref/drivers/block/lvm.c Thu Jan 1 01:00:00 1970 +++ lvm-aa/drivers/block/lvm.c Sun Nov 26 05:00:19 2000 @@ -0,0 +1,2571 @@ +/* + * kernel/lvm.c + * + * Copyright (C) 1997 - 2000 Heinz Mauelshagen, Germany + * + * February-November 1997 + * April-May,July-August,November 1998 + * January-March,May,July,September,October 1999 + * + * + * LVM driver is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * LVM driver is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with GNU CC; see the file COPYING. If not, write to + * the Free Software Foundation, 59 Temple Place - Suite 330, + * Boston, MA 02111-1307, USA. + * + */ + +/* + * Changelog + * + * 09/11/1997 - added chr ioctls VG_STATUS_GET_COUNT + * and VG_STATUS_GET_NAMELIST + * 18/01/1998 - change lvm_chr_open/close lock handling + * 30/04/1998 - changed LV_STATUS ioctl to LV_STATUS_BYNAME and + * - added LV_STATUS_BYINDEX ioctl + * - used lvm_status_byname_req_t and + * lvm_status_byindex_req_t vars + * 04/05/1998 - added multiple device support + * 08/05/1998 - added support to set/clear extendable flag in volume group + * 09/05/1998 - changed output of lvm_proc_get_info() because of + * support for free (eg. longer) logical volume names + * 12/05/1998 - added spin_locks (thanks to Pascal van Dam + * ) + * 25/05/1998 - fixed handling of locked PEs in lvm_map() and lvm_chr_ioctl() + * 26/05/1998 - reactivated verify_area by access_ok + * 07/06/1998 - used vmalloc/vfree instead of kmalloc/kfree to go + * beyond 128/256 KB max allocation limit per call + * - #ifdef blocked spin_lock calls to avoid compile errors + * with 2.0.x + * 11/06/1998 - another enhancement to spinlock code in lvm_chr_open() + * and use of LVM_VERSION_CODE instead of my own macros + * (thanks to Michael Marxmeier ) + * 07/07/1998 - added statistics in lvm_map() + * 08/07/1998 - saved statistics in do_lv_extend_reduce() + * 25/07/1998 - used __initfunc macro + * 02/08/1998 - changes for official char/block major numbers + * 07/08/1998 - avoided init_module() and cleanup_module() to be static + * 30/08/1998 - changed VG lv_open counter from sum of LV lv_open counters + * to sum of LVs open (no matter how often each is) + * 01/09/1998 - fixed lvm_gendisk.part[] index error + * 07/09/1998 - added copying of lv_current_pe-array + * in LV_STATUS_BYINDEX ioctl + * 17/11/1998 - added KERN_* levels to printk + * 13/01/1999 - fixed LV index bug in do_lv_create() which hit lvrename + * 07/02/1999 - fixed spinlock handling bug in case of LVM_RESET + * by moving spinlock code from lvm_chr_open() + * to lvm_chr_ioctl() + * - added LVM_LOCK_LVM ioctl to lvm_chr_ioctl() + * - allowed LVM_RESET and retrieval commands to go ahead; + * only other update ioctls are blocked now + * - fixed pv->pe to NULL for pv_status + * - using lv_req structure in lvm_chr_ioctl() now + * - fixed NULL ptr reference bug in do_lv_extend_reduce() + * caused by uncontiguous PV array in lvm_chr_ioctl(VG_REDUCE) + * 09/02/1999 - changed BLKRASET and BLKRAGET in lvm_chr_ioctl() to + * handle lgoical volume private read ahead sector + * - implemented LV read_ahead handling with lvm_blk_read() + * and lvm_blk_write() + * 10/02/1999 - implemented 2.[12].* support function lvm_hd_name() + * to be used in drivers/block/genhd.c by disk_name() + * 12/02/1999 - fixed index bug in lvm_blk_ioctl(), HDIO_GETGEO + * - enhanced gendisk insert/remove handling + * 16/02/1999 - changed to dynamic block minor number allocation to + * have as much as 99 volume groups with 256 logical volumes + * as the grand total; this allows having 1 volume group with + * up to 256 logical volumes in it + * 21/02/1999 - added LV open count information to proc filesystem + * - substituted redundant LVM_RESET code by calls + * to do_vg_remove() + * 22/02/1999 - used schedule_timeout() to be more responsive + * in case of do_vg_remove() with lots of logical volumes + * 19/03/1999 - fixed NULL pointer bug in module_init/lvm_init + * 17/05/1999 - used DECLARE_WAIT_QUEUE_HEAD macro (>2.3.0) + * - enhanced lvm_hd_name support + * 03/07/1999 - avoided use of KERNEL_VERSION macro based ifdefs and + * memcpy_tofs/memcpy_fromfs macro redefinitions + * 06/07/1999 - corrected reads/writes statistic counter copy in case + * of striped logical volume + * 28/07/1999 - implemented snapshot logical volumes + * - lvm_chr_ioctl + * - LV_STATUS_BYINDEX + * - LV_STATUS_BYNAME + * - do_lv_create + * - do_lv_remove + * - lvm_map + * - new lvm_snapshot_remap_block + * - new lvm_snapshot_remap_new_block + * 08/10/1999 - implemented support for multiple snapshots per + * original logical volume + * 12/10/1999 - support for 2.3.19 + * 11/11/1999 - support for 2.3.28 + * 21/11/1999 - changed lvm_map() interface to buffer_head based + * 19/12/1999 - support for 2.3.33 + * 01/01/2000 - changed locking concept in lvm_map(), + * do_vg_create() and do_lv_remove() + * 07/07/2000 - support for 2.2.16 (backport from 2.4) by Andreas Dilger + * + * + */ + + +/* + * TODO + * + * - implement special handling of unavailable physical volumes + * + */ + +char *lvm_version = "LVM version 0.8final by Heinz Mauelshagen (22/2/2000)\n"; +char *lvm_short_version = "version 0.8final (22/02/2000)"; + +#define MAJOR_NR LVM_BLK_MAJOR +#define DEVICE_OFF(device) +#define DEVICE_REQUEST lvm_dummy_device_request +#define LOCAL_END_REQUEST + +#include +#include + +#ifdef MODVERSIONS +# undef MODULE +# define MODULE +# include +#endif + +#ifdef MODULE +# include +#endif + +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#ifdef CONFIG_KERNELD +#include +#endif + +#include +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 0) +#include +#endif + +#include +#include + +#define LVM_CORRECT_READ_AHEAD(a) \ +do { \ + if ((a) < LVM_MIN_READ_AHEAD) \ + (a) = LVM_MIN_READ_AHEAD; \ + if ((a) > LVM_MAX_READ_AHEAD) \ + (a) = LVM_MAX_READ_AHEAD; \ +} while(0) + +#ifndef WRITEA +#define WRITEA WRITE +#endif + + +/* + * External function prototypes + */ +#ifdef MODULE +int init_module ( void); +void cleanup_module ( void); +#else +extern int lvm_init ( void); +#endif + +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 30) +static void lvm_dummy_device_request ( request_queue_t*); +#else +static void lvm_dummy_device_request ( void); +#endif +static int lvm_blk_ioctl ( struct inode *, struct file *, uint, ulong); +static int lvm_blk_open ( struct inode *, struct file *); + +static int lvm_chr_open ( struct inode *, struct file *); + +static int lvm_chr_release ( struct inode *, struct file *); +static int lvm_blk_release ( struct inode *, struct file *); + +static int lvm_chr_ioctl ( struct inode *, struct file *, uint, ulong); + +#if defined CONFIG_LVM_PROC_FS && defined CONFIG_PROC_FS +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 30) +static int lvm_proc_get_info ( char *, char **, off_t, int); +static int (*lvm_proc_get_info_ptr)(char *, char **, off_t, int) = + &lvm_proc_get_info; +#else +static int lvm_proc_get_info ( char *, char **, off_t, int, int); +#endif +#endif + +#ifdef LVM_HD_NAME +void lvm_hd_name ( char*, int); +#endif + +/* external snapshot calls */ +int lvm_snapshot_remap_block ( kdev_t*, ulong*, unsigned long, lv_t*); +int lvm_snapshot_COW(kdev_t, unsigned long, unsigned long, + unsigned long, lv_t *); +int lvm_snapshot_alloc(lv_t *); +void lvm_snapshot_release(lv_t *); + +/* End external function prototypes */ + + +/* + * Internal function prototypes + */ +static void lvm_init_vars ( void); +extern int (*lvm_map_ptr) ( struct buffer_head*, int); + + +#ifdef LVM_HD_NAME +extern void (*lvm_hd_name_ptr) ( char*, int); +#endif +static int lvm_map ( struct buffer_head*, int); +static int do_vg_create ( int, void *); +static int do_vg_remove ( int); +static int do_lv_create ( int, char *, lv_t *); +static int do_lv_remove ( int, char *, int); +static int do_lv_extend_reduce ( int, char *, lv_t *); +static void lvm_geninit ( struct gendisk *); +#ifdef LVM_GET_INODE + static struct inode *lvm_get_inode ( kdev_t); + void lvm_clear_inode ( struct inode *); +#endif +inline int lvm_strlen ( char *); +inline void lvm_memcpy ( char *, char *, int); +inline int lvm_strcmp ( char *, char *); +inline char *lvm_strrchr ( char *, char c); +/* END Internal function prototypes */ + + +/* volume group descriptor area pointers */ +static vg_t *vg[ABS_MAX_VG + 1]; +static pv_t *pvp = NULL; +static lv_t *lvp = NULL; +static pe_t *pep = NULL; +static pe_t *pep1 = NULL; + + +/* map from block minor number to VG and LV numbers */ +typedef struct { + int vg_number; + int lv_number; +} vg_lv_map_t; +static vg_lv_map_t vg_lv_map[ABS_MAX_LV]; + + +/* Request structures (lvm_chr_ioctl()) */ +static pv_change_req_t pv_change_req; +static pv_flush_req_t pv_flush_req; +static pv_status_req_t pv_status_req; +static pe_lock_req_t pe_lock_req; +static le_remap_req_t le_remap_req; +static lv_req_t lv_req; + +#ifdef LVM_TOTAL_RESET +static int lvm_reset_spindown = 0; +#endif + +static char pv_name[NAME_LEN]; +/* static char rootvg[NAME_LEN] = { 0, }; */ +static uint lv_open = 0; +const char *const lvm_name = LVM_NAME; +static int lock = 0; +static int loadtime = 0; +static uint vg_count = 0; +static long lvm_chr_open_count = 0; +static ushort lvm_iop_version = LVM_DRIVER_IOP_VERSION; +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 0) +static DECLARE_WAIT_QUEUE_HEAD ( lvm_wait); +static DECLARE_WAIT_QUEUE_HEAD ( lvm_map_wait); +#else +struct wait_queue *lvm_wait = NULL; +struct wait_queue *lvm_map_wait = NULL; +#endif + +static spinlock_t lvm_lock = SPIN_LOCK_UNLOCKED; + +#if defined CONFIG_LVM_PROC_FS && defined CONFIG_PROC_FS +#if LINUX_VERSION_CODE < KERNEL_VERSION ( 2, 3, 31) +static struct proc_dir_entry lvm_proc_entry = { + 0, 3, LVM_NAME, S_IFREG | S_IRUGO, + 1, 0, 0, 0, + NULL, + lvm_proc_get_info, + NULL, NULL, NULL, NULL, NULL, +}; +#endif +#endif + +static struct file_operations lvm_chr_fops = { + ioctl: lvm_chr_ioctl, + open: lvm_chr_open, + release: lvm_chr_release, +}; + +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 3, 38) +static struct file_operations lvm_blk_fops = { + read: block_read, + write: block_write, + ioctl: lvm_blk_ioctl, + open: lvm_blk_open, + release: lvm_blk_release, + fsync: block_fsync, +}; +#else +static struct block_device_operations lvm_blk_fops = +{ + open: lvm_blk_open, + release: lvm_blk_release, + ioctl: lvm_blk_ioctl, +}; +#endif + +/* gendisk structures */ +static struct hd_struct lvm_hd_struct[MAX_LV]; +int lvm_blocksizes[MAX_LV] = { 0, }; +static int lvm_size[MAX_LV] = { 0, }; +static struct gendisk lvm_gendisk = { + MAJOR_NR, /* major # */ + LVM_NAME, /* name of major */ + 0, /* number of times minor is shifted + to get real minor */ + 1, /* maximum partitions per device */ + MAX_LV, /* maximum number of real devices */ + lvm_geninit, /* initialization called before we + do other things */ + lvm_hd_struct, /* partition table */ + lvm_size, /* device size in blocks, copied + to block_size[] */ + MAX_LV, /* number or real devices */ + NULL, /* internal */ + NULL, /* pointer to next gendisk struct (internal) */ +}; + + +#ifdef MODULE +/* + * Module initialization... + */ +int init_module ( void) +#else +/* + * Driver initialization... + */ +#ifdef __initfunc +__initfunc ( int lvm_init ( void)) +#else +int __init lvm_init ( void) +#endif +#endif /* #ifdef MODULE */ +{ + struct gendisk *gendisk_ptr = NULL; + + lvm_init_vars (); + + /* insert our gendisk at the corresponding major */ + lvm_geninit ( &lvm_gendisk); + if ( gendisk_head != NULL) { + gendisk_ptr = gendisk_head; + while ( gendisk_ptr->next != NULL && + gendisk_ptr->major > lvm_gendisk.major) { + gendisk_ptr = gendisk_ptr->next; + } + lvm_gendisk.next = gendisk_ptr->next; + gendisk_ptr->next = &lvm_gendisk; + } else { + gendisk_head = &lvm_gendisk; + lvm_gendisk.next = NULL; + } + + /* reference from drivers/block/ll_rw_blk.c */ + lvm_map_ptr = lvm_map; + +#ifdef LVM_HD_NAME + /* reference from drivers/block/genhd.c */ + lvm_hd_name_ptr = lvm_hd_name; +#endif + +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 30) + blk_init_queue ( BLK_DEFAULT_QUEUE ( MAJOR_NR), lvm_dummy_device_request); +#else + blk_dev[MAJOR_NR].request_fn = lvm_dummy_device_request; + blk_dev[MAJOR_NR].current_request = NULL; +#endif + + /* optional read root VGDA */ +/* + if ( *rootvg != 0) { + vg_read_with_pv_and_lv ( rootvg, &vg); + } +*/ + + if ( register_chrdev ( LVM_CHAR_MAJOR, lvm_name, &lvm_chr_fops) < 0) { + printk ( KERN_ERR "%s -- register_chrdev failed\n", lvm_name); + return -EIO; + } + if ( register_blkdev ( MAJOR_NR, lvm_name, &lvm_blk_fops) < 0) { + printk ( "%s -- register_blkdev failed\n", lvm_name); + if ( unregister_chrdev ( LVM_CHAR_MAJOR, lvm_name) < 0) + printk ( KERN_ERR "%s -- unregister_chrdev failed\n", lvm_name); + return -EIO; + } + +#if defined CONFIG_LVM_PROC_FS && defined CONFIG_PROC_FS +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 25) + create_proc_info_entry ( LVM_NAME, S_IFREG | S_IRUGO, + &proc_root, lvm_proc_get_info_ptr); +# else + proc_register ( &proc_root, &lvm_proc_entry); +# endif +#endif + + printk ( KERN_INFO + "%s%s -- " +#ifdef MODULE + "Module" +#else + "Driver" +#endif + " successfully initialized\n", + lvm_version, lvm_name); + + return 0; +} /* init_module () / lvm_init () */ + + +#ifdef MODULE +/* + * Module cleanup... + */ +void cleanup_module ( void) { + struct gendisk *gendisk_ptr = NULL, *gendisk_ptr_prev = NULL; + + if ( unregister_chrdev ( LVM_CHAR_MAJOR, lvm_name) < 0) { + printk ( KERN_ERR "%s -- unregister_chrdev failed\n", lvm_name); + } + if ( unregister_blkdev ( MAJOR_NR, lvm_name) < 0) { + printk ( KERN_ERR "%s -- unregister_blkdev failed\n", lvm_name); + } + +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 30) + blk_cleanup_queue ( BLK_DEFAULT_QUEUE ( MAJOR_NR)); +#else + blk_dev[MAJOR_NR].request_fn = NULL; + blk_dev[MAJOR_NR].current_request = NULL; +#endif + + gendisk_ptr = gendisk_ptr_prev = gendisk_head; + while ( gendisk_ptr != NULL) { + if ( gendisk_ptr == &lvm_gendisk) break; + gendisk_ptr_prev = gendisk_ptr; + gendisk_ptr = gendisk_ptr->next; + } + /* delete our gendisk from chain */ + if ( gendisk_ptr == &lvm_gendisk) gendisk_ptr_prev->next = gendisk_ptr->next; + + blk_size[MAJOR_NR] = NULL; + blksize_size[MAJOR_NR] = NULL; + +#if defined CONFIG_LVM_PROC_FS && defined CONFIG_PROC_FS +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 30) + remove_proc_entry ( LVM_NAME, &proc_root); +# else + proc_unregister ( &proc_root, lvm_proc_entry.low_ino); +# endif +#endif + + /* reference from linux/drivers/block/ll_rw_blk.c */ + lvm_map_ptr = NULL; + +#ifdef LVM_HD_NAME + /* reference from linux/drivers/block/genhd.c */ + lvm_hd_name_ptr = NULL; +#endif + + printk ( KERN_INFO "%s -- Module successfully deactivated\n", lvm_name); + + return; +} /* void cleanup_module () */ +#endif /* #ifdef MODULE */ + + +/* + * support function to initialize lvm variables + */ +#ifdef __initfunc +__initfunc ( void lvm_init_vars ( void)) +#else +void __init lvm_init_vars ( void) +#endif +{ + int v; + + loadtime = CURRENT_TIME; + + lvm_lock = SPIN_LOCK_UNLOCKED; + + pe_lock_req.lock = UNLOCK_PE; + pe_lock_req.data.lv_dev = pe_lock_req.data.pv_dev = 0; + pe_lock_req.data.pv_offset = 0; + + /* Initialize VG pointers */ + for ( v = 0; v <= ABS_MAX_VG; v++) vg[v] = NULL; + + /* Initialize LV -> VG association */ + for ( v = 0; v < ABS_MAX_LV; v++) { + /* index ABS_MAX_VG never used for real VG */ + vg_lv_map[v].vg_number = ABS_MAX_VG; + vg_lv_map[v].lv_number = -1; + } + + return; +} /* lvm_init_vars () */ + + +/******************************************************************** + * + * Character device functions + * + ********************************************************************/ + +/* + * character device open routine + */ +static int lvm_chr_open ( struct inode *inode, + struct file *file) { + int minor = MINOR ( inode->i_rdev); + +#ifdef DEBUG + printk ( KERN_DEBUG + "%s -- lvm_chr_open MINOR: %d VG#: %d mode: 0x%X lock: %d\n", + lvm_name, minor, VG_CHR(minor), file->f_mode, lock); +#endif + + /* super user validation */ + if ( ! capable(CAP_SYS_ADMIN)) return -EACCES; + + /* Group special file open */ + if ( VG_CHR(minor) > MAX_VG) return -ENXIO; + +#ifdef MODULE + MOD_INC_USE_COUNT; +#endif + + lvm_chr_open_count++; + return 0; +} /* lvm_chr_open () */ + + +/* + * character device i/o-control routine + * + * Only one changing process can do ioctl at one time, others will block. + * + */ +static int lvm_chr_ioctl ( struct inode *inode, struct file *file, + uint command, ulong a) { + int minor = MINOR ( inode->i_rdev); + int extendable; + ulong l, le, p, v; + ulong size; + void *arg = ( void*) a; +#ifdef LVM_GET_INODE + struct inode *inode_sav; +#endif + lv_status_byname_req_t lv_status_byname_req; + lv_status_byindex_req_t lv_status_byindex_req; + lv_t lv; + + /* otherwise cc will complain about unused variables */ + ( void) lvm_lock; + + +#ifdef DEBUG_IOCTL + printk ( KERN_DEBUG + "%s -- lvm_chr_ioctl: command: 0x%X MINOR: %d " + "VG#: %d mode: 0x%X\n", + lvm_name, command, minor, VG_CHR(minor), file->f_mode); +#endif + +#ifdef LVM_TOTAL_RESET + if ( lvm_reset_spindown > 0) return -EACCES; +#endif + + + /* Main command switch */ + switch ( command) { + /* lock the LVM */ + case LVM_LOCK_LVM: +lock_try_again: + spin_lock ( &lvm_lock); + if( lock != 0 && lock != current->pid ) { +#ifdef DEBUG_IOCTL + printk ( KERN_INFO "lvm_chr_ioctl: %s is locked by pid %d ...\n", + lvm_name, lock); +#endif + spin_unlock ( &lvm_lock); + interruptible_sleep_on ( &lvm_wait); + if ( current->sigpending != 0) return -EINTR; +#ifdef LVM_TOTAL_RESET + if ( lvm_reset_spindown > 0) return -EACCES; +#endif + goto lock_try_again; + } + lock = current->pid; + spin_unlock ( &lvm_lock); + return 0; + + + /* check lvm version to ensure driver/tools+lib interoperability */ + case LVM_GET_IOP_VERSION: + if ( copy_to_user ( arg, &lvm_iop_version, sizeof ( ushort)) != 0) + return -EFAULT; + return 0; + + +#ifdef LVM_TOTAL_RESET + /* lock reset function */ + case LVM_RESET: + lvm_reset_spindown = 1; + for ( v = 0; v < ABS_MAX_VG; v++) { + if ( vg[v] != NULL) { + do_vg_remove ( v); + } + } + +#ifdef MODULE + while ( GET_USE_COUNT ( &__this_module) < 1) + MOD_INC_USE_COUNT; + while ( GET_USE_COUNT ( &__this_module) > 1) + MOD_DEC_USE_COUNT; +#endif /* MODULE */ + lock = 0; /* release lock */ + wake_up_interruptible ( &lvm_wait); + return 0; +#endif /* LVM_TOTAL_RESET */ + + + /* lock/unlock i/o to a physical extent to move it to another + physical volume (move's done in user space's pvmove) */ + case PE_LOCK_UNLOCK: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_from_user ( &pe_lock_req, arg, sizeof ( pe_lock_req_t)) != 0) + return -EFAULT; + + switch ( pe_lock_req.lock) { + case LOCK_PE: + for ( p = 0; p < vg[VG_CHR(minor)]->pv_max; p++) { + if ( vg[VG_CHR(minor)]->pv[p] != NULL && + pe_lock_req.data.pv_dev == + vg[VG_CHR(minor)]->pv[p]->pv_dev) + break; + } + + if ( p == vg[VG_CHR(minor)]->pv_max) return -ENXIO; + + pe_lock_req.lock = UNLOCK_PE; + fsync_dev ( pe_lock_req.data.lv_dev); + pe_lock_req.lock = LOCK_PE; + break; + + case UNLOCK_PE: + pe_lock_req.lock = UNLOCK_PE; + pe_lock_req.data.lv_dev = pe_lock_req.data.pv_dev = 0; + pe_lock_req.data.pv_offset = 0; + wake_up ( &lvm_map_wait); + break; + + default: + return -EINVAL; + } + + return 0; + + + /* remap a logical extent (after moving the physical extent) */ + case LE_REMAP: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_from_user ( &le_remap_req, arg, + sizeof ( le_remap_req_t)) != 0) + return -EFAULT; + + for ( l = 0; l < vg[VG_CHR(minor)]->lv_max; l++) { + if ( vg[VG_CHR(minor)]->lv[l] != NULL && + lvm_strcmp ( vg[VG_CHR(minor)]->lv[l]->lv_name, + le_remap_req.lv_name) == 0) { + for ( le = 0; le < vg[VG_CHR(minor)]->lv[l]->lv_allocated_le; + le++) { + if ( vg[VG_CHR(minor)]->lv[l]->lv_current_pe[le].dev == + le_remap_req.old_dev && + vg[VG_CHR(minor)]->lv[l]->lv_current_pe[le].pe == + le_remap_req.old_pe) { + vg[VG_CHR(minor)]->lv[l]->lv_current_pe[le].dev = + le_remap_req.new_dev; + vg[VG_CHR(minor)]->lv[l]->lv_current_pe[le].pe = + le_remap_req.new_pe; + return 0; + } + } + return -EINVAL; + } + } + + return -ENXIO; + + + /* create a VGDA */ + case VG_CREATE: + return do_vg_create ( minor, arg); + + + /* remove an inactive VGDA */ + case VG_REMOVE: + return do_vg_remove ( minor); + + + /* extend a volume group */ + case VG_EXTEND: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( vg[VG_CHR(minor)]->pv_cur < vg[VG_CHR(minor)]->pv_max) { + for ( p = 0; p < vg[VG_CHR(minor)]->pv_max; p++) { + if ( vg[VG_CHR(minor)]->pv[p] == NULL) { + if ( ( vg[VG_CHR(minor)]->pv[p] = + kmalloc ( sizeof ( pv_t), GFP_USER)) == NULL) { + printk ( KERN_CRIT + "%s -- VG_EXTEND: kmalloc error PV at line %d\n", + lvm_name, __LINE__); + return -ENOMEM; + } + if ( copy_from_user ( vg[VG_CHR(minor)]->pv[p], arg, + sizeof ( pv_t)) != 0) + return -EFAULT; + + vg[VG_CHR(minor)]->pv[p]->pv_status = PV_ACTIVE; + /* We don't need the PE list + in kernel space like LVs pe_t list */ + vg[VG_CHR(minor)]->pv[p]->pe = NULL; + vg[VG_CHR(minor)]->pv_cur++; + vg[VG_CHR(minor)]->pv_act++; + vg[VG_CHR(minor)]->pe_total += + vg[VG_CHR(minor)]->pv[p]->pe_total; +#ifdef LVM_GET_INODE + /* insert a dummy inode for fs_may_mount */ + vg[VG_CHR(minor)]->pv[p]->inode = + lvm_get_inode ( vg[VG_CHR(minor)]->pv[p]->pv_dev); +#endif + return 0; + } + } + } + return -EPERM; + + + /* reduce a volume group */ + case VG_REDUCE: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_from_user ( pv_name, arg, sizeof ( pv_name)) != 0) + return -EFAULT; + + for ( p = 0; p < vg[VG_CHR(minor)]->pv_max; p++) { + if ( vg[VG_CHR(minor)]->pv[p] != NULL && + lvm_strcmp ( vg[VG_CHR(minor)]->pv[p]->pv_name, + pv_name) == 0) { + if ( vg[VG_CHR(minor)]->pv[p]->lv_cur > 0) return -EPERM; + vg[VG_CHR(minor)]->pe_total -= + vg[VG_CHR(minor)]->pv[p]->pe_total; + vg[VG_CHR(minor)]->pv_cur--; + vg[VG_CHR(minor)]->pv_act--; +#ifdef DEBUG_VFREE + printk ( KERN_DEBUG + "%s -- kfree %d\n", lvm_name, __LINE__); +#endif +#ifdef LVM_GET_INODE + lvm_clear_inode ( vg[VG_CHR(minor)]->pv[p]->inode); +#endif + kfree ( vg[VG_CHR(minor)]->pv[p]); + /* Make PV pointer array contiguous */ + for ( ; p < vg[VG_CHR(minor)]->pv_max-1; p++) + vg[VG_CHR(minor)]->pv[p] = vg[VG_CHR(minor)]->pv[p + 1]; + vg[VG_CHR(minor)]->pv[p + 1] = NULL; + return 0; + } + } + return -ENXIO; + + + /* set/clear extendability flag of volume group */ + case VG_SET_EXTENDABLE: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_from_user ( &extendable, arg, sizeof ( extendable)) != 0) + return -EFAULT; + + if ( extendable == VG_EXTENDABLE || + extendable == ~VG_EXTENDABLE) { + if ( extendable == VG_EXTENDABLE) + vg[VG_CHR(minor)]->vg_status |= VG_EXTENDABLE; + else + vg[VG_CHR(minor)]->vg_status &= ~VG_EXTENDABLE; + } else return -EINVAL; + return 0; + + + /* get volume group data (only the vg_t struct) */ + case VG_STATUS: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_to_user ( arg, vg[VG_CHR(minor)], sizeof ( vg_t)) != 0) + return -EFAULT; + + return 0; + + + /* get volume group count */ + case VG_STATUS_GET_COUNT: + if ( copy_to_user ( arg, &vg_count, sizeof ( vg_count)) != 0) + return -EFAULT; + + return 0; + + + /* get volume group count */ + case VG_STATUS_GET_NAMELIST: + for ( l = v = 0; v < ABS_MAX_VG; v++) { + if ( vg[v] != NULL) { + if ( copy_to_user ( arg + l++ * NAME_LEN, + vg[v]->vg_name, + NAME_LEN) != 0) + return -EFAULT; + } + } + return 0; + + + /* create, remove, extend or reduce a logical volume */ + case LV_CREATE: + case LV_REMOVE: + case LV_EXTEND: + case LV_REDUCE: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_from_user ( &lv_req, arg, sizeof ( lv_req)) != 0) + return -EFAULT; + + if ( command != LV_REMOVE) { + if ( copy_from_user ( &lv, lv_req.lv, sizeof ( lv_t)) != 0) + return -EFAULT; + } + + switch ( command) { + case LV_CREATE: + return do_lv_create ( minor, lv_req.lv_name, &lv); + + case LV_REMOVE: + return do_lv_remove ( minor, lv_req.lv_name, -1); + + case LV_EXTEND: + case LV_REDUCE: + return do_lv_extend_reduce ( minor, lv_req.lv_name, &lv); + } + + + /* get status of a logical volume by name */ + case LV_STATUS_BYNAME: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_from_user ( &lv_status_byname_req, arg, + sizeof ( lv_status_byname_req_t)) != 0) + return -EFAULT; + + if ( lv_status_byname_req.lv == NULL) return -EINVAL; + if ( copy_from_user ( &lv, lv_status_byname_req.lv, + sizeof ( lv_t)) != 0) + return -EFAULT; + + for ( l = 0; l < vg[VG_CHR(minor)]->lv_max; l++) { + if ( vg[VG_CHR(minor)]->lv[l] != NULL && + lvm_strcmp ( vg[VG_CHR(minor)]->lv[l]->lv_name, + lv_status_byname_req.lv_name) == 0) { + if ( copy_to_user ( lv_status_byname_req.lv, + vg[VG_CHR(minor)]->lv[l], + sizeof ( lv_t)) != 0) + return -EFAULT; + + if ( lv.lv_current_pe != NULL) { + size = vg[VG_CHR(minor)]->lv[l]->lv_allocated_le * + sizeof ( pe_t); + if ( copy_to_user ( lv.lv_current_pe, + vg[VG_CHR(minor)]->lv[l]->lv_current_pe, + size) != 0) + return -EFAULT; + } + return 0; + } + } + return -ENXIO; + + + /* get status of a logical volume by index */ + case LV_STATUS_BYINDEX: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_from_user ( &lv_status_byindex_req, arg, + sizeof ( lv_status_byindex_req)) != 0) + return -EFAULT; + + if ( ( lvp = lv_status_byindex_req.lv) == NULL) return -EINVAL; + l = lv_status_byindex_req.lv_index; + if ( vg[VG_CHR(minor)]->lv[l] == NULL) return -ENXIO; + + if ( copy_from_user ( &lv, lvp, sizeof ( lv_t)) != 0) + return -EFAULT; + + if ( copy_to_user ( lvp, vg[VG_CHR(minor)]->lv[l], + sizeof ( lv_t)) != 0) + return -EFAULT; + + if ( lv.lv_current_pe != NULL) { + size = vg[VG_CHR(minor)]->lv[l]->lv_allocated_le * sizeof ( pe_t); + if ( copy_to_user ( lv.lv_current_pe, + vg[VG_CHR(minor)]->lv[l]->lv_current_pe, + size) != 0) + return -EFAULT; + } + return 0; + + + /* change a physical volume */ + case PV_CHANGE: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_from_user ( &pv_change_req, arg, + sizeof ( pv_change_req)) != 0) + return -EFAULT; + + for ( p = 0; p < vg[VG_CHR(minor)]->pv_max; p++) { + if ( vg[VG_CHR(minor)]->pv[p] != NULL && + lvm_strcmp ( vg[VG_CHR(minor)]->pv[p]->pv_name, + pv_change_req.pv_name) == 0) { +#ifdef LVM_GET_INODE + inode_sav = vg[VG_CHR(minor)]->pv[p]->inode; +#endif + if ( copy_from_user ( vg[VG_CHR(minor)]->pv[p], + pv_change_req.pv, + sizeof ( pv_t)) != 0) + return -EFAULT; + + /* We don't need the PE list + in kernel space as with LVs pe_t list */ + vg[VG_CHR(minor)]->pv[p]->pe = NULL; +#ifdef LVM_GET_INODE + vg[VG_CHR(minor)]->pv[p]->inode = inode_sav; +#endif + return 0; + } + } + return -ENXIO; + + + /* get physical volume data (pv_t structure only) */ + case PV_STATUS: + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + if ( copy_from_user ( &pv_status_req, arg, + sizeof ( pv_status_req)) != 0) + return -EFAULT; + + for ( p = 0; p < vg[VG_CHR(minor)]->pv_max; p++) { + if ( vg[VG_CHR(minor)]->pv[p] != NULL) { + if ( lvm_strcmp ( vg[VG_CHR(minor)]->pv[p]->pv_name, + pv_status_req.pv_name) == 0) { + if ( copy_to_user ( pv_status_req.pv, + vg[VG_CHR(minor)]->pv[p], + sizeof ( pv_t)) != 0) + return -EFAULT; + return 0; + } + } + } + return -ENXIO; + + + /* physical volume buffer flush/invalidate */ + case PV_FLUSH: + if ( copy_from_user ( &pv_flush_req, arg, sizeof ( pv_flush_req)) != 0) + return -EFAULT; + + for ( v = 0; v < ABS_MAX_VG; v++) { + if ( vg[v] == NULL) continue; + for ( p = 0; p < vg[v]->pv_max; p++) { + if ( vg[v]->pv[p] != NULL && + lvm_strcmp ( vg[v]->pv[p]->pv_name, + pv_flush_req.pv_name) == 0) { + fsync_dev ( vg[v]->pv[p]->pv_dev); + invalidate_buffers ( vg[v]->pv[p]->pv_dev); + return 0; + } + } + } + return 0; + + + default: + printk ( KERN_WARNING + "%s -- lvm_chr_ioctl: unknown command %x\n", + lvm_name, command); + return -EINVAL; + } + + return 0; +} /* lvm_chr_ioctl */ + + +/* + * character device close routine + */ +static int lvm_chr_release ( struct inode *inode, struct file *file) +{ +#ifdef DEBUG + int minor = MINOR ( inode->i_rdev); + printk ( KERN_DEBUG + "%s -- lvm_chr_release VG#: %d\n", lvm_name, VG_CHR(minor)); +#endif + +#ifdef MODULE + if ( GET_USE_COUNT ( &__this_module) > 0) MOD_DEC_USE_COUNT; +#endif + +#ifdef LVM_TOTAL_RESET + if ( lvm_reset_spindown > 0) { + lvm_reset_spindown = 0; + lvm_chr_open_count = 1; + } +#endif + + if ( lvm_chr_open_count > 0) lvm_chr_open_count--; + if ( lock == current->pid) { + lock = 0; /* release lock */ + wake_up_interruptible ( &lvm_wait); + } + + return 0; +} /* lvm_chr_release () */ + + + +/******************************************************************** + * + * Block device functions + * + ********************************************************************/ + +/* + * block device open routine + */ +static int lvm_blk_open ( struct inode *inode, struct file *file) { + int minor = MINOR ( inode->i_rdev); + +#ifdef DEBUG_LVM_BLK_OPEN + printk ( KERN_DEBUG + "%s -- lvm_blk_open MINOR: %d VG#: %d LV#: %d mode: 0x%X\n", + lvm_name, minor, VG_BLK(minor), LV_BLK(minor), file->f_mode); +#endif + +#ifdef LVM_TOTAL_RESET + if ( lvm_reset_spindown > 0) return -EPERM; +#endif + + if ( vg[VG_BLK(minor)] != NULL && + ( vg[VG_BLK(minor)]->vg_status & VG_ACTIVE) && + vg[VG_BLK(minor)]->lv[LV_BLK(minor)] != NULL && + LV_BLK(minor) >= 0 && + LV_BLK(minor) < vg[VG_BLK(minor)]->lv_max) { + + /* Check parallel LV spindown (LV remove) */ + if ( vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_status & LV_SPINDOWN) + return -EPERM; + + /* Check inactive LV and open for read/write */ + if ( file->f_mode & O_RDWR) { + if ( ! ( vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_status & LV_ACTIVE)) + return -EPERM; + if ( ! ( vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_access & LV_WRITE)) + return -EACCES; + } + + if ( vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_open == 0) + vg[VG_BLK(minor)]->lv_open++; + vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_open++; + +#ifdef MODULE + MOD_INC_USE_COUNT; +#endif + +#ifdef DEBUG_LVM_BLK_OPEN + printk ( KERN_DEBUG + "%s -- lvm_blk_open MINOR: %d VG#: %d LV#: %d size: %d\n", + lvm_name, minor, VG_BLK(minor), LV_BLK(minor), + vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_size); +#endif + + return 0; + } + + return -ENXIO; +} /* lvm_blk_open () */ + + +/* + * block device i/o-control routine + */ +static int lvm_blk_ioctl (struct inode *inode, struct file *file, + uint command, ulong a) { + int minor = MINOR ( inode->i_rdev); + void *arg = ( void*) a; + struct hd_geometry *hd = ( struct hd_geometry *) a; + +#ifdef DEBUG_IOCTL + printk ( KERN_DEBUG + "%s -- lvm_blk_ioctl MINOR: %d command: 0x%X arg: %X " + "VG#: %dl LV#: %d\n", + lvm_name, minor, command, ( ulong) arg, + VG_BLK(minor), LV_BLK(minor)); +#endif + + switch ( command) { + /* return device size */ + case BLKGETSIZE: +#ifdef DEBUG_IOCTL + printk ( KERN_DEBUG + "%s -- lvm_blk_ioctl -- BLKGETSIZE: %u\n", + lvm_name, vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_size); +#endif + copy_to_user ( ( long*) arg, &vg[VG_BLK(minor)]->\ + lv[LV_BLK(minor)]->lv_size, + sizeof ( vg[VG_BLK(minor)]->\ + lv[LV_BLK(minor)]->lv_size)); + break; + + + /* flush buffer cache */ + case BLKFLSBUF: + /* super user validation */ + if ( ! capable (CAP_SYS_ADMIN)) return -EACCES; + +#ifdef DEBUG_IOCTL + printk ( KERN_DEBUG + "%s -- lvm_blk_ioctl -- BLKFLSBUF\n", lvm_name); +#endif + fsync_dev ( inode->i_rdev); + invalidate_buffers(inode->i_rdev); + break; + + + /* set read ahead for block device */ + case BLKRASET: + /* super user validation */ + if ( ! capable (CAP_SYS_ADMIN)) return -EACCES; + +#ifdef DEBUG_IOCTL + printk ( KERN_DEBUG + "%s -- lvm_blk_ioctl -- BLKRASET: %d sectors for %02X:%02X\n", + lvm_name, ( long) arg, MAJOR( inode->i_rdev), minor); +#endif + if ( ( long) arg < LVM_MIN_READ_AHEAD || + ( long) arg > LVM_MAX_READ_AHEAD) return -EINVAL; + read_ahead[MAJOR_NR] = + vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_read_ahead = ( long) arg; + break; + + + /* get current read ahead setting */ + case BLKRAGET: +#ifdef DEBUG_IOCTL + printk ( KERN_DEBUG + "%s -- lvm_blk_ioctl -- BLKRAGET\n", lvm_name); +#endif + copy_to_user ( ( long*) arg, + &vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_read_ahead, + sizeof ( vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->\ + lv_read_ahead)); + break; + + + /* get disk geometry */ + case HDIO_GETGEO: +#ifdef DEBUG_IOCTL + printk ( KERN_DEBUG + "%s -- lvm_blk_ioctl -- HDIO_GETGEO\n", lvm_name); +#endif + if ( hd == NULL) return -EINVAL; + { + unsigned char heads = 64; + unsigned char sectors = 32; + long start = 0; + short cylinders = vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_size / + heads / sectors; + + if ( copy_to_user ( ( char*) &hd->heads, &heads, + sizeof ( heads)) != 0 || + copy_to_user ( ( char*) &hd->sectors, §ors, + sizeof ( sectors)) != 0 || + copy_to_user ( ( short*) &hd->cylinders, + &cylinders, sizeof ( cylinders)) != 0 || + copy_to_user ( ( long*) &hd->start, &start, + sizeof ( start)) != 0) + return -EFAULT; + } + +#ifdef DEBUG_IOCTL + printk ( KERN_DEBUG + "%s -- lvm_blk_ioctl -- cylinders: %d\n", + lvm_name, vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->\ + lv_size / heads / sectors); +#endif + break; + + + /* set access flags of a logical volume */ + case LV_SET_ACCESS: + /* super user validation */ + if ( ! capable (CAP_SYS_ADMIN)) return -EACCES; + vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_access = ( ulong) arg; + if ( vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_access & LV_WRITE) + set_device_ro(vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_dev, 0); + else + set_device_ro(vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_dev, 1); + break; + break; + + + /* set status flags of a logical volume */ + case LV_SET_STATUS: + /* super user validation */ + if ( ! capable (CAP_SYS_ADMIN)) return -EACCES; + if ( ! ( ( ulong) arg & LV_ACTIVE) && + vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_open > 1) return -EPERM; + vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_status = ( ulong) arg; + break; + + + /* set allocation flags of a logical volume */ + case LV_SET_ALLOCATION: + /* super user validation */ + if ( ! capable (CAP_SYS_ADMIN)) return -EACCES; + vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_allocation = ( ulong) arg; + break; + + + default: + printk ( KERN_WARNING + "%s -- lvm_blk_ioctl: unknown command %d\n", + lvm_name, command); + return -EINVAL; + } + + return 0; +} /* lvm_blk_ioctl () */ + + +/* + * block device close routine + */ +static int lvm_blk_release ( struct inode *inode, struct file *file) +{ + int minor = MINOR ( inode->i_rdev); + +#ifdef DEBUG + printk ( KERN_DEBUG + "%s -- lvm_blk_release MINOR: %d VG#: %d LV#: %d\n", + lvm_name, minor, VG_BLK(minor), LV_BLK(minor)); +#endif + + sync_dev ( inode->i_rdev); + if ( vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_open == 1) + vg[VG_BLK(minor)]->lv_open--; + vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_open--; + +#ifdef MODULE + MOD_DEC_USE_COUNT; +#endif + + return 0; +} /* lvm_blk_release () */ + + +#if defined CONFIG_LVM_PROC_FS && defined CONFIG_PROC_FS +/* + * Support function /proc-Filesystem + */ +#define LVM_PROC_BUF ( i == 0 ? dummy_buf : &buf[sz]) + +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 25) +static int lvm_proc_get_info ( char *page, char **start, off_t pos, int count) +#else +static int lvm_proc_get_info ( char *page, char **start, off_t pos, + int count, int whence) +#endif +{ + int c, i, l, p, v, vg_counter, pv_counter, lv_counter, lv_open_counter, + lv_open_total, pe_t_bytes, lv_block_exception_t_bytes, seconds; + static off_t sz; + off_t sz_last; + char allocation_flag, inactive_flag, rw_flag, stripes_flag; + char *lv_name = NULL; + static char *buf = NULL; + static char dummy_buf[160]; /* sized for 2 lines */ + +#ifdef DEBUG_LVM_PROC_GET_INFO + printk ( KERN_DEBUG + "%s - lvm_proc_get_info CALLED pos: %lu count: %d whence: %d\n", + lvm_name, pos, count, whence); +#endif + + if ( pos == 0 || buf == NULL) { + sz_last = vg_counter = pv_counter = lv_counter = lv_open_counter = \ + lv_open_total = pe_t_bytes = lv_block_exception_t_bytes = 0; + + /* search for activity */ + for ( v = 0; v < ABS_MAX_VG; v++) { + if ( vg[v] != NULL) { + vg_counter++; + pv_counter += vg[v]->pv_cur; + lv_counter += vg[v]->lv_cur; + if ( vg[v]->lv_cur > 0) { + for ( l = 0; l < vg[v]->lv_max; l++) { + if ( vg[v]->lv[l] != NULL) { + pe_t_bytes += vg[v]->lv[l]->lv_allocated_le; + if ( vg[v]->lv[l]->lv_block_exception != NULL) { + lv_block_exception_t_bytes += + vg[v]->lv[l]->lv_remap_end; + } + if ( vg[v]->lv[l]->lv_open > 0) { + lv_open_counter++; + lv_open_total += vg[v]->lv[l]->lv_open; + } + } + } + } + } + } + pe_t_bytes *= sizeof ( pe_t); + lv_block_exception_t_bytes *= sizeof ( lv_block_exception_t); + + if ( buf != NULL) { +#ifdef DEBUG_VFREE + printk ( KERN_DEBUG + "%s -- vfree %d\n", lvm_name, __LINE__); +#endif + vfree ( buf); + buf = NULL; + } + + /* 2 times: first to get size to allocate buffer, + 2nd to fill the vmalloced buffer */ + for ( i = 0; i < 2; i++) { + sz = 0; + sz += sprintf ( LVM_PROC_BUF, + "LVM " +#ifdef MODULE + "module" +#else + "driver" +#endif + " %s\n\n" + "Total: %d VG%s %d PV%s %d LV%s ", + lvm_short_version, + vg_counter, vg_counter == 1 ? "" : "s", + pv_counter, pv_counter == 1 ? "" : "s", + lv_counter, lv_counter == 1 ? "" : "s"); + sz += sprintf ( LVM_PROC_BUF, + "(%d LV%s open", + lv_open_counter, + lv_open_counter == 1 ? "" : "s"); + if ( lv_open_total > 0) sz += sprintf ( LVM_PROC_BUF, + " %d times)\n", + lv_open_total); + else sz += sprintf ( LVM_PROC_BUF, ")"); + sz += sprintf ( LVM_PROC_BUF, + "\nGlobal: %lu bytes vmalloced IOP version: %d ", + vg_counter * sizeof ( vg_t) + + pv_counter * sizeof ( pv_t) + + lv_counter * sizeof ( lv_t) + + pe_t_bytes + lv_block_exception_t_bytes + sz_last, + lvm_iop_version); + + seconds = CURRENT_TIME - loadtime; + if ( seconds < 0) loadtime = CURRENT_TIME + seconds; + if ( seconds / 86400 > 0) { + sz += sprintf ( LVM_PROC_BUF, "%d day%s ", + seconds / 86400, + seconds / 86400 == 0 || + seconds / 86400 > 1 ? "s": ""); + } + sz += sprintf ( LVM_PROC_BUF, "%d:%02d:%02d active\n", + ( seconds % 86400) / 3600, + ( seconds % 3600) / 60, + seconds % 60); + + if ( vg_counter > 0) { + for ( v = 0; v < ABS_MAX_VG; v++) { + /* volume group */ + if ( vg[v] != NULL) { + inactive_flag = ' '; + if ( ! ( vg[v]->vg_status & VG_ACTIVE)) + inactive_flag = 'I'; + sz += sprintf ( LVM_PROC_BUF, + "\nVG: %c%s [%d PV, %d LV/%d open] " + " PE Size: %d KB\n" + " Usage [KB/PE]: %d /%d total " + "%d /%d used %d /%d free", + inactive_flag, + vg[v]->vg_name, + vg[v]->pv_cur, + vg[v]->lv_cur, + vg[v]->lv_open, + vg[v]->pe_size >> 1, + vg[v]->pe_size * vg[v]->pe_total >> 1, + vg[v]->pe_total, + vg[v]->pe_allocated * vg[v]->pe_size >> 1, + vg[v]->pe_allocated, + ( vg[v]->pe_total - vg[v]->pe_allocated) * + vg[v]->pe_size >> 1, + vg[v]->pe_total - vg[v]->pe_allocated); + + /* physical volumes */ + sz += sprintf ( LVM_PROC_BUF, + "\n PV%s ", + vg[v]->pv_cur == 1 ? ": " : "s:"); + c = 0; + for ( p = 0; p < vg[v]->pv_max; p++) { + if ( vg[v]->pv[p] != NULL) { + inactive_flag = 'A'; + if ( ! ( vg[v]->pv[p]->pv_status & PV_ACTIVE)) + inactive_flag = 'I'; + allocation_flag = 'A'; + if ( ! ( vg[v]->pv[p]->pv_allocatable & PV_ALLOCATABLE)) + allocation_flag = 'N'; + sz += sprintf ( LVM_PROC_BUF, + "[%c%c] %-21s %8d /%-6d " + "%8d /%-6d %8d /%-6d", + inactive_flag, + allocation_flag, + vg[v]->pv[p]->pv_name, + vg[v]->pv[p]->pe_total * + vg[v]->pv[p]->pe_size >> 1, + vg[v]->pv[p]->pe_total, + vg[v]->pv[p]->pe_allocated * + vg[v]->pv[p]->pe_size >> 1, + vg[v]->pv[p]->pe_allocated, + ( vg[v]->pv[p]->pe_total - + vg[v]->pv[p]->pe_allocated) * + vg[v]->pv[p]->pe_size >> 1, + vg[v]->pv[p]->pe_total - + vg[v]->pv[p]->pe_allocated); + c++; + if ( c < vg[v]->pv_cur) sz += sprintf ( LVM_PROC_BUF, + "\n "); + } + } + + /* logical volumes */ + sz += sprintf ( LVM_PROC_BUF, + "\n LV%s ", + vg[v]->lv_cur == 1 ? ": " : "s:"); + c = 0; + for ( l = 0; l < vg[v]->lv_max; l++) { + if ( vg[v]->lv[l] != NULL) { + inactive_flag = 'A'; + if ( ! ( vg[v]->lv[l]->lv_status & LV_ACTIVE)) + inactive_flag = 'I'; + rw_flag = 'R'; + if ( vg[v]->lv[l]->lv_access & LV_WRITE) rw_flag = 'W'; + allocation_flag = 'D'; + if ( vg[v]->lv[l]->lv_allocation & LV_CONTIGUOUS) + allocation_flag = 'C'; + stripes_flag = 'L'; + if ( vg[v]->lv[l]->lv_stripes > 1) stripes_flag = 'S'; + sz += sprintf ( LVM_PROC_BUF, + "[%c%c%c%c", + inactive_flag, + rw_flag, + allocation_flag, + stripes_flag); + if ( vg[v]->lv[l]->lv_stripes > 1) + sz += sprintf ( LVM_PROC_BUF, "%-2d", + vg[v]->lv[l]->lv_stripes); + else + sz += sprintf ( LVM_PROC_BUF, " "); + lv_name = lvm_strrchr ( vg[v]->lv[l]->lv_name, '/'); + if ( lv_name != NULL) lv_name++; + else lv_name = vg[v]->lv[l]->lv_name; + sz += sprintf ( LVM_PROC_BUF, "] %-25s", lv_name); + if ( lvm_strlen ( lv_name) > 25) + sz += sprintf ( LVM_PROC_BUF, + "\n "); + sz += sprintf ( LVM_PROC_BUF, "%9d /%-6d ", + vg[v]->lv[l]->lv_size >> 1, + vg[v]->lv[l]->lv_size / vg[v]->pe_size); + + if ( vg[v]->lv[l]->lv_open == 0) + sz += sprintf ( LVM_PROC_BUF, "close"); + else + sz += sprintf ( LVM_PROC_BUF, "%dx open", + vg[v]->lv[l]->lv_open); + c++; + if ( c < vg[v]->lv_cur) sz += sprintf ( LVM_PROC_BUF, + "\n "); + } + } + if ( vg[v]->lv_cur == 0) + sz += sprintf ( LVM_PROC_BUF, "none"); + sz += sprintf ( LVM_PROC_BUF, "\n"); + } + } + } + + if ( buf == NULL) { + if ( ( buf = vmalloc ( sz)) == NULL) { + sz = 0; + return sprintf ( page, "%s - vmalloc error at line %d\n", + lvm_name, __LINE__); + } + } + sz_last = sz; + } + } + + if ( pos > sz - 1) { + vfree ( buf); + buf = NULL; + return 0; + } + + *start = &buf[pos]; + if ( sz - pos < count) return sz - pos; + else return count; +} /* lvm_proc_get_info () */ +#endif /* #if defined CONFIG_LVM_PROC_FS && defined CONFIG_PROC_FS */ + + +/* + * block device support function for /usr/src/linux/drivers/block/ll_rw_blk.c + * (see init_module/lvm_init) + */ +static int lvm_map ( struct buffer_head *bh, int rw) { + int minor = MINOR ( bh->b_dev); + int ret = 0; + ulong index; + ulong size = bh->b_size >> 9; + ulong rsector_tmp = bh->b_blocknr * size; + ulong rsector_sav; + kdev_t rdev_tmp = bh->b_dev; + kdev_t rdev_sav; + lv_t *lv = vg[VG_BLK(minor)]->lv[LV_BLK(minor)]; + unsigned long pe_start; + + + if ( ! ( lv->lv_status & LV_ACTIVE)) { + printk ( KERN_ALERT + "%s - lvm_map: ll_rw_blk for inactive LV %s\n", + lvm_name, lv->lv_name); + return -1; + } + +/* +if ( lv->lv_access & LV_SNAPSHOT) +printk ( "%s -- %02d:%02d block: %lu rw: %d\n", lvm_name, MAJOR ( bh->b_dev), MINOR ( bh->b_dev), bh->b_blocknr, rw); +*/ + + /* take care of snapshot chunk writes before + check for writable logical volume */ + if ( ( lv->lv_access & LV_SNAPSHOT) && + MAJOR ( bh->b_dev) != 0 && + MAJOR ( bh->b_dev) != MAJOR_NR && + ( rw == WRITEA || rw == WRITE)) + { +/* +printk ( "%s -- doing snapshot write for %02d:%02d[%02d:%02d] b_blocknr: %lu b_rsector: %lu\n", lvm_name, MAJOR ( bh->b_dev), MINOR ( bh->b_dev), MAJOR ( bh->b_dev), MINOR ( bh->b_dev), bh->b_blocknr, bh->b_rsector); +*/ + return 0; + } + + if ( ( rw == WRITE || rw == WRITEA) && + ! ( lv->lv_access & LV_WRITE)) { + printk ( KERN_CRIT + "%s - lvm_map: ll_rw_blk write for readonly LV %s\n", + lvm_name, lv->lv_name); + return -1; + } + + +#ifdef DEBUG_MAP + printk ( KERN_DEBUG + "%s - lvm_map minor:%d *rdev: %02d:%02d *rsector: %lu " + "size:%lu\n", + lvm_name, minor, + MAJOR ( rdev_tmp), + MINOR ( rdev_tmp), + rsector_tmp, size); +#endif + + if ( rsector_tmp + size > lv->lv_size) { + printk ( KERN_ALERT + "%s - lvm_map *rsector: %lu or size: %lu wrong for" + " minor: %2d\n", lvm_name, rsector_tmp, size, minor); + return -1; + } + + rsector_sav = rsector_tmp; + rdev_sav = rdev_tmp; + +lvm_second_remap: + /* linear mapping */ + if ( lv->lv_stripes < 2) { + index = rsector_tmp / vg[VG_BLK(minor)]->pe_size; /* get the index */ + pe_start = lv->lv_current_pe[index].pe; + rsector_tmp = lv->lv_current_pe[index].pe + + ( rsector_tmp % vg[VG_BLK(minor)]->pe_size); + rdev_tmp = lv->lv_current_pe[index].dev; + +#ifdef DEBUG_MAP + printk ( KERN_DEBUG + "lv_current_pe[%ld].pe: %d rdev: %02d:%02d rsector:%ld\n", + index, + lv->lv_current_pe[index].pe, + MAJOR ( rdev_tmp), + MINOR ( rdev_tmp), + rsector_tmp); +#endif + + /* striped mapping */ + } else { + ulong stripe_index; + ulong stripe_length; + + stripe_length = vg[VG_BLK(minor)]->pe_size * lv->lv_stripes; + stripe_index = ( rsector_tmp % stripe_length) / lv->lv_stripesize; + index = rsector_tmp / stripe_length + + ( stripe_index % lv->lv_stripes) * + ( lv->lv_allocated_le / lv->lv_stripes); + pe_start = lv->lv_current_pe[index].pe; + rsector_tmp = lv->lv_current_pe[index].pe + + ( rsector_tmp % stripe_length) - + ( stripe_index % lv->lv_stripes) * lv->lv_stripesize - + stripe_index / lv->lv_stripes * + ( lv->lv_stripes - 1) * lv->lv_stripesize; + rdev_tmp = lv->lv_current_pe[index].dev; + +#ifdef DEBUG_MAP + printk(KERN_DEBUG + "lv_current_pe[%ld].pe: %d rdev: %02d:%02d rsector:%ld\n" + "stripe_length: %ld stripe_index: %ld\n", + index, + lv->lv_current_pe[index].pe, + MAJOR ( rdev_tmp), + MINOR ( rdev_tmp), + rsector_tmp, + stripe_length, + stripe_index); +#endif + } + + /* handle physical extents on the move */ + if ( pe_lock_req.lock == LOCK_PE) { + if ( rdev_tmp == pe_lock_req.data.pv_dev && + rsector_tmp >= pe_lock_req.data.pv_offset && + rsector_tmp < ( pe_lock_req.data.pv_offset + + vg[VG_BLK(minor)]->pe_size)) { + sleep_on ( &lvm_map_wait); + rsector_tmp = rsector_sav; + rdev_tmp = rdev_sav; + goto lvm_second_remap; + } + } + + /* statistic */ + if ( rw == WRITE || rw == WRITEA) + lv->lv_current_pe[index].writes++; + else + lv->lv_current_pe[index].reads++; + + /* snapshot volume exception handling on physical device address base */ + if ( lv->lv_access & ( LV_SNAPSHOT | LV_SNAPSHOT_ORG)) { + /* original logical volume */ + if ( lv->lv_access & LV_SNAPSHOT_ORG) { + if ( rw == WRITE || rw == WRITEA) + { + lv_t *lv_ptr; + + /* start with first snapshot and loop thrugh all of them */ + for ( lv_ptr = lv->lv_snapshot_next; + lv_ptr != NULL; + lv_ptr = lv_ptr->lv_snapshot_next) { + down(&lv_ptr->lv_snapshot_sem); + /* do we still have exception storage for this snapshot free? */ + if ( lv_ptr->lv_block_exception != NULL) { + rdev_sav = rdev_tmp; + rsector_sav = rsector_tmp; + if (!lvm_snapshot_remap_block(&rdev_tmp, + &rsector_tmp, + pe_start, + lv_ptr)) + /* create a new mapping */ + ret = lvm_snapshot_COW(rdev_tmp, + rsector_tmp, + pe_start, + rsector_sav, + lv_ptr); + rdev_tmp = rdev_sav; + rsector_tmp = rsector_sav; + } + up(&lv_ptr->lv_snapshot_sem); + } + } + } else { + /* remap snapshot logical volume */ + down(&lv->lv_snapshot_sem); + if ( lv->lv_block_exception != NULL) + lvm_snapshot_remap_block ( &rdev_tmp, &rsector_tmp, pe_start, lv); + up(&lv->lv_snapshot_sem); + } + } + + bh->b_rdev = rdev_tmp; + bh->b_rsector = rsector_tmp; + + return ret; +} /* lvm_map () */ + + +/* + * lvm_map snapshot logical volume support functions + */ + +/* + * end lvm_map snapshot logical volume support functions + */ + + +/* + * internal support functions + */ + +#ifdef LVM_HD_NAME +/* + * generate "hard disk" name + */ +void lvm_hd_name ( char *buf, int minor) { + int len = 0; + + if ( vg[VG_BLK(minor)] == NULL || + vg[VG_BLK(minor)]->lv[LV_BLK(minor)] == NULL) return; + len = lvm_strlen ( vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_name) - 5; + lvm_memcpy ( buf, &vg[VG_BLK(minor)]->lv[LV_BLK(minor)]->lv_name[5], len); + buf[len] = 0; + return; +} +#endif + + +/* + * this one never should be called... + */ +#if LINUX_VERSION_CODE > KERNEL_VERSION ( 2, 3, 30) +static void lvm_dummy_device_request ( request_queue_t *t) +#else +static void lvm_dummy_device_request ( void) +#endif +{ + printk ( KERN_EMERG + "%s -- oops, got lvm request for %02d:%02d [sector: %lu]\n", + lvm_name, + MAJOR ( CURRENT->rq_dev), + MINOR ( CURRENT->rq_dev), + CURRENT->sector); + return; +} + + +/* + * character device support function VGDA create + */ +int do_vg_create ( int minor, void *arg) { + int snaporg_minor = 0; + ulong l, p; + lv_t lv; + vg_t *vg_ptr; + + if ( vg[VG_CHR(minor)] != NULL) return -EPERM; + + if ( ( vg_ptr = kmalloc ( sizeof ( vg_t), GFP_USER)) == NULL) { + printk ( KERN_CRIT + "%s -- VG_CREATE: kmalloc error VG at line %d\n", + lvm_name, __LINE__); + return -ENOMEM; + } + + /* get the volume group structure */ + if ( copy_from_user ( vg_ptr, arg, sizeof ( vg_t)) != 0) { + kfree ( vg_ptr); + return -EFAULT; + } + + /* we are not that active so far... */ + vg_ptr->vg_status &= ~VG_ACTIVE; + vg[VG_CHR(minor)] = vg_ptr; + + vg[VG_CHR(minor)]->pe_allocated = 0; + if ( vg[VG_CHR(minor)]->pv_max > ABS_MAX_PV) { + printk ( KERN_WARNING + "%s -- Can't activate VG: ABS_MAX_PV too small\n", + lvm_name); + kfree ( vg[VG_CHR(minor)]); + vg[VG_CHR(minor)] = NULL; + return -EPERM; + } + if ( vg[VG_CHR(minor)]->lv_max > ABS_MAX_LV) { + printk ( KERN_WARNING + "%s -- Can't activate VG: ABS_MAX_LV too small for %u\n", + lvm_name, vg[VG_CHR(minor)]->lv_max); + kfree ( vg[VG_CHR(minor)]); + vg[VG_CHR(minor)] = NULL; + return -EPERM; + } + + /* get the physical volume structures */ + vg[VG_CHR(minor)]->pv_act = vg[VG_CHR(minor)]->pv_cur = 0; + for ( p = 0; p < vg[VG_CHR(minor)]->pv_max; p++) { + /* user space address */ + if ( ( pvp = vg[VG_CHR(minor)]->pv[p]) != NULL) { + vg[VG_CHR(minor)]->pv[p] = kmalloc ( sizeof ( pv_t), GFP_USER); + if ( vg[VG_CHR(minor)]->pv[p] == NULL) { + printk ( KERN_CRIT + "%s -- VG_CREATE: kmalloc error PV at line %d\n", + lvm_name, __LINE__); + do_vg_remove ( minor); + return -ENOMEM; + } + if ( copy_from_user ( vg[VG_CHR(minor)]->pv[p], pvp, + sizeof ( pv_t)) != 0) { + do_vg_remove ( minor); + return -EFAULT; + } + + /* We don't need the PE list + in kernel space as with LVs pe_t list (see below) */ + vg[VG_CHR(minor)]->pv[p]->pe = NULL; + vg[VG_CHR(minor)]->pv[p]->pe_allocated = 0; + vg[VG_CHR(minor)]->pv[p]->pv_status = PV_ACTIVE; + vg[VG_CHR(minor)]->pv_act++; + vg[VG_CHR(minor)]->pv_cur++; + +#ifdef LVM_GET_INODE + /* insert a dummy inode for fs_may_mount */ + vg[VG_CHR(minor)]->pv[p]->inode = + lvm_get_inode ( vg[VG_CHR(minor)]->pv[p]->pv_dev); +#endif + } + } + + /* get the logical volume structures */ + vg[VG_CHR(minor)]->lv_cur = 0; + for ( l = 0; l < vg[VG_CHR(minor)]->lv_max; l++) { + /* user space address */ + if ( ( lvp = vg[VG_CHR(minor)]->lv[l]) != NULL) { + if ( copy_from_user ( &lv, lvp, sizeof ( lv_t)) != 0) { + do_vg_remove ( minor); + return -EFAULT; + } + vg[VG_CHR(minor)]->lv[l] = NULL; + { + int err; + + err = do_lv_create(minor, lv.lv_name, &lv); + if (err) + { + do_vg_remove(minor); + return err; + } + } + } + } + + /* Second path to correct snapshot logical volumes which are not + in place during first path above */ + for ( l = 0; l < vg[VG_CHR(minor)]->lv_max; l++) { + if ( vg[VG_CHR(minor)]->lv[l] != NULL && + vg[VG_CHR(minor)]->lv[l]->lv_access & LV_SNAPSHOT) { + snaporg_minor = vg[VG_CHR(minor)]->lv[l]->lv_snapshot_minor; + if ( vg[VG_CHR(minor)]->lv[LV_BLK(snaporg_minor)] != NULL) { + /* get pointer to original logical volume */ + lv_t *lv_ptr = vg[VG_CHR(minor)]->lv[l]->lv_snapshot_org = + vg[VG_CHR(minor)]->lv[LV_BLK(snaporg_minor)]; + + /* set necessary fields of original logical volume */ + lv_ptr->lv_access |= LV_SNAPSHOT_ORG; + lv_ptr->lv_snapshot_minor = 0; + lv_ptr->lv_snapshot_org = lv_ptr; + lv_ptr->lv_snapshot_prev = NULL; + + /* find last snapshot logical volume in the chain */ + while ( lv_ptr->lv_snapshot_next != NULL) + lv_ptr = lv_ptr->lv_snapshot_next; + + /* set back pointer to this last one in our new logical volume */ + vg[VG_CHR(minor)]->lv[l]->lv_snapshot_prev = lv_ptr; + + /* last logical volume now points to our new snapshot volume */ + lv_ptr->lv_snapshot_next = vg[VG_CHR(minor)]->lv[l]; + + /* now point to the new one */ + lv_ptr = lv_ptr->lv_snapshot_next; + + /* set necessary fields of new snapshot logical volume */ + lv_ptr->lv_snapshot_next = NULL; + lv_ptr->lv_current_pe = + vg[VG_CHR(minor)]->lv[LV_BLK(snaporg_minor)]->lv_current_pe; + lv_ptr->lv_allocated_le = + vg[VG_CHR(minor)]->lv[LV_BLK(snaporg_minor)]->lv_allocated_le; + lv_ptr->lv_current_le = + vg[VG_CHR(minor)]->lv[LV_BLK(snaporg_minor)]->lv_current_le; + lv_ptr->lv_size = + vg[VG_CHR(minor)]->lv[LV_BLK(snaporg_minor)]->lv_size; + } + } + } + + vg_count++; + + /* let's go active */ + vg[VG_CHR(minor)]->vg_status |= VG_ACTIVE; + +#ifdef MODULE + MOD_INC_USE_COUNT; +#endif + return 0; +} /* do_vg_create () */ + + +/* + * character device support function VGDA remove + */ +static int do_vg_remove ( int minor) { + int i; + + if ( vg[VG_CHR(minor)] == NULL) return -ENXIO; + +#ifdef LVM_TOTAL_RESET + if ( vg[VG_CHR(minor)]->lv_open > 0 && lvm_reset_spindown == 0) +#else + if ( vg[VG_CHR(minor)]->lv_open > 0) +#endif + return -EPERM; + + /* let's go inactive */ + vg[VG_CHR(minor)]->vg_status &= ~VG_ACTIVE; + + /* free LVs */ + /* first free snapshot logical volumes */ + for ( i = 0; i < vg[VG_CHR(minor)]->lv_max; i++) { + if ( vg[VG_CHR(minor)]->lv[i] != NULL && + vg[VG_CHR(minor)]->lv[i]->lv_access & LV_SNAPSHOT) { + do_lv_remove ( minor, NULL, i); + current->state = TASK_INTERRUPTIBLE; + schedule_timeout ( 1); + } + } + /* then free the rest */ + for ( i = 0; i < vg[VG_CHR(minor)]->lv_max; i++) { + if ( vg[VG_CHR(minor)]->lv[i] != NULL) { + do_lv_remove ( minor, NULL, i); + current->state = TASK_INTERRUPTIBLE; + schedule_timeout ( 1); + } + } + + /* free PVs */ + for ( i = 0; i < vg[VG_CHR(minor)]->pv_max; i++) { + if ( vg[VG_CHR(minor)]->pv[i] != NULL) { +#ifdef DEBUG_VFREE + printk ( KERN_DEBUG + "%s -- kfree %d\n", lvm_name, __LINE__); +#endif +#ifdef LVM_GET_INODE + lvm_clear_inode ( vg[VG_CHR(minor)]->pv[i]->inode); +#endif + kfree ( vg[VG_CHR(minor)]->pv[i]); + vg[VG_CHR(minor)]->pv[i] = NULL; + } + } + +#ifdef DEBUG_VFREE + printk ( KERN_DEBUG "%s -- kfree %d\n", lvm_name, __LINE__); +#endif + kfree ( vg[VG_CHR(minor)]); + vg[VG_CHR(minor)] = NULL; + + vg_count--; + +#ifdef MODULE + MOD_DEC_USE_COUNT; +#endif + return 0; +} /* do_vg_remove () */ + + +/* + * character device support function logical volume create + */ +static int do_lv_create ( int minor, char *lv_name, lv_t *lv) { + int l, le, l_new, p, size; + ulong lv_status_save; + lv_block_exception_t *lvbe = lv->lv_block_exception; + lv_t *lv_ptr = NULL; + + if ( ( pep = lv->lv_current_pe) == NULL) return -EINVAL; + if ( lv->lv_chunk_size > LVM_SNAPSHOT_MAX_CHUNK) return -EINVAL; + + for ( l = 0; l < vg[VG_CHR(minor)]->lv_max; l++) { + if ( vg[VG_CHR(minor)]->lv[l] != NULL && + lvm_strcmp ( vg[VG_CHR(minor)]->lv[l]->lv_name, lv_name) == 0) + return -EEXIST; + } + + /* in case of lv_remove(), lv_create() pair; for eg. lvrename does this */ + l_new = -1; + if ( vg[VG_CHR(minor)]->lv[lv->lv_number] == NULL) l_new = lv->lv_number; + else { + for ( l = 0; l < vg[VG_CHR(minor)]->lv_max; l++) { + if ( vg[VG_CHR(minor)]->lv[l] == NULL) if ( l_new == -1) l_new = l; + } + } + if ( l_new == -1) return -EPERM; + l = l_new; + + if ( ( lv_ptr = kmalloc ( sizeof ( lv_t), GFP_USER)) == NULL) {; + printk ( KERN_CRIT "%s -- LV_CREATE: kmalloc error LV at line %d\n", + lvm_name, __LINE__); + return -ENOMEM; + } + + /* copy preloaded LV */ + lvm_memcpy ( ( char*) lv_ptr, ( char *) lv, sizeof ( lv_t)); + + lv_status_save = lv_ptr->lv_status; + lv_ptr->lv_status &= ~LV_ACTIVE; + lv_ptr->lv_snapshot_org = \ + lv_ptr->lv_snapshot_prev = \ + lv_ptr->lv_snapshot_next = NULL; + lv_ptr->lv_block_exception = NULL; +#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 3, 4) + lv_ptr->lv_snapshot_sem = MUTEX; +#else + init_MUTEX(&lv_ptr->lv_snapshot_sem); +#endif + vg[VG_CHR(minor)]->lv[l] = lv_ptr; + + /* get the PE structures from user space if this + is no snapshot logical volume */ + if ( ! ( lv_ptr->lv_access & LV_SNAPSHOT)) { + size = lv_ptr->lv_allocated_le * sizeof ( pe_t); + if ( ( lv_ptr->lv_current_pe = vmalloc ( size)) == NULL) { + printk ( KERN_CRIT + "%s -- LV_CREATE: vmalloc error LV_CURRENT_PE of %d Byte " + "at line %d\n", + lvm_name, size, __LINE__); +#ifdef DEBUG_VFREE + printk ( KERN_DEBUG "%s -- vfree %d\n", lvm_name, __LINE__); +#endif + kfree ( lv_ptr); + vg[VG_CHR(minor)]->lv[l] = NULL; + return -ENOMEM; + } + + if ( copy_from_user ( lv_ptr->lv_current_pe, pep, size)) { + vfree ( lv_ptr->lv_current_pe); + kfree ( lv_ptr); + vg[VG_CHR(minor)]->lv[l] = NULL; + return -EFAULT; + } + + /* correct the PE count in PVs */ + for ( le = 0; le < lv_ptr->lv_allocated_le; le++) { + vg[VG_CHR(minor)]->pe_allocated++; + for ( p = 0; p < vg[VG_CHR(minor)]->pv_cur; p++) { + if ( vg[VG_CHR(minor)]->pv[p]->pv_dev == + lv_ptr->lv_current_pe[le].dev) + vg[VG_CHR(minor)]->pv[p]->pe_allocated++; + } + } + } else { + /* Get snapshot exception data and block list */ + if ( lvbe != NULL) { + lv_ptr->lv_snapshot_org = + vg[VG_CHR(minor)]->lv[LV_BLK(lv_ptr->lv_snapshot_minor)]; + if ( lv_ptr->lv_snapshot_org != NULL) { + size = lv_ptr->lv_remap_end * sizeof ( lv_block_exception_t); + if ( ( lv_ptr->lv_block_exception = vmalloc ( size)) == NULL) { + printk ( KERN_CRIT + "%s -- do_lv_create: vmalloc error LV_BLOCK_EXCEPTION " + "of %d byte at line %d\n", + lvm_name, size, __LINE__); +#ifdef DEBUG_VFREE + printk ( KERN_DEBUG "%s -- vfree %d\n", lvm_name, __LINE__); +#endif + kfree ( lv_ptr); + vg[VG_CHR(minor)]->lv[l] = NULL; + return -ENOMEM; + } + + if ( copy_from_user ( lv_ptr->lv_block_exception, lvbe, size)) { + vfree ( lv_ptr->lv_block_exception); + kfree ( lv_ptr); + vg[VG_CHR(minor)]->lv[l] = NULL; + return -EFAULT; + } + + /* get pointer to original logical volume */ + lv_ptr = lv_ptr->lv_snapshot_org; + + lv_ptr->lv_snapshot_minor = 0; + lv_ptr->lv_snapshot_org = lv_ptr; + lv_ptr->lv_snapshot_prev = NULL; + /* walk thrugh the snapshot list */ + while ( lv_ptr->lv_snapshot_next != NULL) + lv_ptr = lv_ptr->lv_snapshot_next; + /* now lv_ptr points to the last existing snapshot in the chain */ + vg[VG_CHR(minor)]->lv[l]->lv_snapshot_prev = lv_ptr; + /* our new one now back points to the previous last in the chain */ + lv_ptr = vg[VG_CHR(minor)]->lv[l]; + /* now lv_ptr points to our new last snapshot logical volume */ + lv_ptr->lv_snapshot_org = lv_ptr->lv_snapshot_prev->lv_snapshot_org; + lv_ptr->lv_snapshot_next = NULL; + lv_ptr->lv_current_pe = lv_ptr->lv_snapshot_org->lv_current_pe; + lv_ptr->lv_allocated_le = lv_ptr->lv_snapshot_org->lv_allocated_le; + lv_ptr->lv_current_le = lv_ptr->lv_snapshot_org->lv_current_le; + lv_ptr->lv_size = lv_ptr->lv_snapshot_org->lv_size; + lv_ptr->lv_stripes = lv_ptr->lv_snapshot_org->lv_stripes; + lv_ptr->lv_stripesize = lv_ptr->lv_snapshot_org->lv_stripesize; + { + int err; + + err = lvm_snapshot_alloc(lv_ptr); + if (err) + { + vfree(lv_ptr->lv_block_exception); + kfree(lv_ptr); + vg[VG_CHR(minor)]->lv[l] = NULL; + return err; + } + } + } else { + vfree ( lv_ptr->lv_block_exception); + kfree ( lv_ptr); + vg[VG_CHR(minor)]->lv[l] = NULL; + return -EFAULT; + } + } else { + kfree ( vg[VG_CHR(minor)]->lv[l]); + vg[VG_CHR(minor)]->lv[l] = NULL; + return -EINVAL; + } + } /* if ( vg[VG_CHR(minor)]->lv[l]->lv_access & LV_SNAPSHOT) */ + + lv_ptr = vg[VG_CHR(minor)]->lv[l]; + lvm_gendisk.part[MINOR(lv_ptr->lv_dev)].start_sect = 0; + lvm_gendisk.part[MINOR(lv_ptr->lv_dev)].nr_sects = lv_ptr->lv_size; + lvm_size[MINOR(lv_ptr->lv_dev)] = lv_ptr->lv_size >> 1; + vg_lv_map[MINOR(lv_ptr->lv_dev)].vg_number = vg[VG_CHR(minor)]->vg_number; + vg_lv_map[MINOR(lv_ptr->lv_dev)].lv_number = lv_ptr->lv_number; + LVM_CORRECT_READ_AHEAD ( lv_ptr->lv_read_ahead); + read_ahead[MAJOR_NR] = lv_ptr->lv_read_ahead; + vg[VG_CHR(minor)]->lv_cur++; + lv_ptr->lv_status = lv_status_save; + + /* optionally add our new snapshot LV */ + if ( lv_ptr->lv_access & LV_SNAPSHOT) { + /* sync the original logical volume */ + fsync_dev ( lv_ptr->lv_snapshot_org->lv_dev); + /* put ourselve into the chain */ + lv_ptr->lv_snapshot_prev->lv_snapshot_next = lv_ptr; + lv_ptr->lv_snapshot_org->lv_access |= LV_SNAPSHOT_ORG; + } + lv_ptr->lv_status |= LV_ACTIVE; + if ( lv_ptr->lv_access & LV_WRITE) + set_device_ro(lv_ptr->lv_dev, 0); + else + set_device_ro(lv_ptr->lv_dev, 1); + + return 0; +} /* do_lv_create () */ + + +/* + * character device support function logical volume remove + */ +static int do_lv_remove ( int minor, char *lv_name, int l) { + uint le, p; + lv_t *lv_ptr; + + if ( l == -1) { + for ( l = 0; l < vg[VG_CHR(minor)]->lv_max; l++) { + if ( vg[VG_CHR(minor)]->lv[l] != NULL && + lvm_strcmp ( vg[VG_CHR(minor)]->lv[l]->lv_name, lv_name) == 0) { + break; + } + } + } + + lv_ptr = vg[VG_CHR(minor)]->lv[l]; + if ( l < vg[VG_CHR(minor)]->lv_max) { +#ifdef LVM_TOTAL_RESET + if ( lv_ptr->lv_open > 0 && lvm_reset_spindown == 0) +#else + if ( lv_ptr->lv_open > 0) +#endif + return -EBUSY; + + /* check for deletion of snapshot source while + snapshot volume still exists */ + if ( ( lv_ptr->lv_access & LV_SNAPSHOT_ORG) && + lv_ptr->lv_snapshot_next != NULL) + return -EPERM; + + lv_ptr->lv_status |= LV_SPINDOWN; + + /* sync the buffers */ + fsync_dev ( lv_ptr->lv_dev); + + lv_ptr->lv_status &= ~LV_ACTIVE; + + /* invalidate the buffers */ + invalidate_buffers ( lv_ptr->lv_dev); + + /* reset generic hd */ + lvm_gendisk.part[MINOR(lv_ptr->lv_dev)].start_sect = -1; + lvm_gendisk.part[MINOR(lv_ptr->lv_dev)].nr_sects = 0; + lvm_size[MINOR(lv_ptr->lv_dev)] = 0; + + /* reset VG/LV mapping */ + vg_lv_map[MINOR(lv_ptr->lv_dev)].vg_number = ABS_MAX_VG; + vg_lv_map[MINOR(lv_ptr->lv_dev)].lv_number = -1; + + /* correct the PE count in PVs if this is no snapshot logical volume */ + if ( ! ( lv_ptr->lv_access & LV_SNAPSHOT)) { + /* only if this is no snapshot logical volume because we share + the lv_current_pe[] structs with the original logical volume */ + for ( le = 0; le < lv_ptr->lv_allocated_le; le++) { + vg[VG_CHR(minor)]->pe_allocated--; + for ( p = 0; p < vg[VG_CHR(minor)]->pv_cur; p++) { + if ( vg[VG_CHR(minor)]->pv[p]->pv_dev == + lv_ptr->lv_current_pe[le].dev) + vg[VG_CHR(minor)]->pv[p]->pe_allocated--; + } + } + vfree ( lv_ptr->lv_current_pe); + /* LV_SNAPSHOT */ + } else { +/* + if ( lv_ptr->lv_block_exception != NULL) { + int i; + kdev_t last_dev; + for ( i = last_dev = 0; i < lv_ptr->lv_remap_ptr; i++) { + if ( lv_ptr->lv_block_exception[i].rdev_new != last_dev) { + last_dev = lv_ptr->lv_block_exception[i].rdev_new; + invalidate_buffers ( last_dev); + current->state = TASK_INTERRUPTIBLE; + schedule_timeout ( 1); + } + } + } +*/ + /* remove this snapshot logical volume from the chain */ + lv_ptr->lv_snapshot_prev->lv_snapshot_next = lv_ptr->lv_snapshot_next; + if ( lv_ptr->lv_snapshot_next != NULL) { + lv_ptr->lv_snapshot_next->lv_snapshot_prev = + lv_ptr->lv_snapshot_prev; + } + /* no more snapshots? */ + if ( lv_ptr->lv_snapshot_org->lv_snapshot_next == NULL) + lv_ptr->lv_snapshot_org->lv_access &= ~LV_SNAPSHOT_ORG; + lvm_snapshot_release(lv_ptr); + } + +#ifdef DEBUG_VFREE + printk ( KERN_DEBUG "%s -- kfree %d\n", lvm_name, __LINE__); +#endif + kfree ( lv_ptr); + vg[VG_CHR(minor)]->lv[l] = NULL; + vg[VG_CHR(minor)]->lv_cur--; + return 0; + } + + return -ENXIO; +} /* do_lv_remove () */ + + +/* + * character device support function logical volume extend / reduce + */ +static int do_lv_extend_reduce ( int minor, char *lv_name, lv_t *lv) { + int l, le, p, size, old_allocated_le; + uint32_t end, lv_status_save; + pe_t *pe; + + if ( ( pep = lv->lv_current_pe) == NULL) return -EINVAL; + + for ( l = 0; l < vg[VG_CHR(minor)]->lv_max; l++) { + if ( vg[VG_CHR(minor)]->lv[l] != NULL && + lvm_strcmp ( vg[VG_CHR(minor)]->lv[l]->lv_name, lv_name) == 0) + break; + } + if ( l == vg[VG_CHR(minor)]->lv_max) return -ENXIO; + + /* check for active snapshot */ + if ( lv->lv_access & ( LV_SNAPSHOT|LV_SNAPSHOT_ORG)) return -EPERM; + + if ( ( pe = vmalloc ( size = lv->lv_current_le * sizeof ( pe_t))) == NULL) { + printk ( KERN_CRIT + "%s -- do_lv_extend_reduce: vmalloc error LV_CURRENT_PE " + "of %d Byte at line %d\n", + lvm_name, size, __LINE__); + return -ENOMEM; + } + + /* get the PE structures from user space */ + if ( copy_from_user ( pe, pep, size)) { + vfree ( pe); + return -EFAULT; + } + +#ifdef DEBUG + printk ( KERN_DEBUG + "%s -- fsync_dev and " + "invalidate_buffers for %s [%s] in %s\n", + lvm_name, vg[VG_CHR(minor)]->lv[l]->lv_name, + kdevname ( vg[VG_CHR(minor)]->lv[l]->lv_dev), + vg[VG_CHR(minor)]->vg_name); +#endif + + vg[VG_CHR(minor)]->lv[l]->lv_status |= LV_SPINDOWN; + fsync_dev ( vg[VG_CHR(minor)]->lv[l]->lv_dev); + vg[VG_CHR(minor)]->lv[l]->lv_status &= ~LV_ACTIVE; + invalidate_buffers ( vg[VG_CHR(minor)]->lv[l]->lv_dev); + + /* reduce allocation counters on PV(s) */ + for ( le = 0; le < vg[VG_CHR(minor)]->lv[l]->lv_allocated_le; le++) { + vg[VG_CHR(minor)]->pe_allocated--; + for ( p = 0; p < vg[VG_CHR(minor)]->pv_cur; p++) { + if ( vg[VG_CHR(minor)]->pv[p]->pv_dev == + vg[VG_CHR(minor)]->lv[l]->lv_current_pe[le].dev) { + vg[VG_CHR(minor)]->pv[p]->pe_allocated--; + break; + } + } + } + +#ifdef DEBUG_VFREE + printk ( KERN_DEBUG "%s -- vfree %d\n", lvm_name, __LINE__); +#endif + + /* save pointer to "old" lv/pe pointer array */ + pep1 = vg[VG_CHR(minor)]->lv[l]->lv_current_pe; + end = vg[VG_CHR(minor)]->lv[l]->lv_current_le; + + /* save open counter */ + lv_open = vg[VG_CHR(minor)]->lv[l]->lv_open; + + /* save # of old allocated logical extents */ + old_allocated_le = vg[VG_CHR(minor)]->lv[l]->lv_allocated_le; + + /* copy preloaded LV */ + lv_status_save = lv->lv_status; + lv->lv_status |= LV_SPINDOWN; + lv->lv_status &= ~LV_ACTIVE; + lvm_memcpy ( ( char*) vg[VG_CHR(minor)]->lv[l], ( char*) lv, sizeof ( lv_t)); + vg[VG_CHR(minor)]->lv[l]->lv_current_pe = pe; + vg[VG_CHR(minor)]->lv[l]->lv_open = lv_open; + + /* save availiable i/o statistic data */ + /* linear logical volume */ + if ( vg[VG_CHR(minor)]->lv[l]->lv_stripes < 2) { + /* Check what last LE shall be used */ + if ( end > vg[VG_CHR(minor)]->lv[l]->lv_current_le) + end = vg[VG_CHR(minor)]->lv[l]->lv_current_le; + for ( le = 0; le < end; le++) { + vg[VG_CHR(minor)]->lv[l]->lv_current_pe[le].reads = pep1[le].reads; + vg[VG_CHR(minor)]->lv[l]->lv_current_pe[le].writes = pep1[le].writes; + } + /* striped logical volume */ + } else { + uint i, j, source, dest, end, old_stripe_size, new_stripe_size; + + old_stripe_size = old_allocated_le / vg[VG_CHR(minor)]->lv[l]->lv_stripes; + new_stripe_size = vg[VG_CHR(minor)]->lv[l]->lv_allocated_le / + vg[VG_CHR(minor)]->lv[l]->lv_stripes; + end = old_stripe_size; + if ( end > new_stripe_size) end = new_stripe_size; + for ( i = source = dest = 0; + i < vg[VG_CHR(minor)]->lv[l]->lv_stripes; i++) { + for ( j = 0; j < end; j++) { + vg[VG_CHR(minor)]->lv[l]->lv_current_pe[dest+j].reads = + pep1[source+j].reads; + vg[VG_CHR(minor)]->lv[l]->lv_current_pe[dest+j].writes = + pep1[source+j].writes; + } + source += old_stripe_size; + dest += new_stripe_size; + } + } + vfree ( pep1); pep1 = NULL; + + + /* extend the PE count in PVs */ + for ( le = 0; le < vg[VG_CHR(minor)]->lv[l]->lv_allocated_le; le++) { + vg[VG_CHR(minor)]->pe_allocated++; + for ( p = 0; p < vg[VG_CHR(minor)]->pv_cur; p++) { + if ( vg[VG_CHR(minor)]->pv[p]->pv_dev == + vg[VG_CHR(minor)]->lv[l]->lv_current_pe[le].dev) { + vg[VG_CHR(minor)]->pv[p]->pe_allocated++; + break; + } + } + } + + lvm_gendisk.part[MINOR(vg[VG_CHR(minor)]->lv[l]->lv_dev)].start_sect = 0; + lvm_gendisk.part[MINOR(vg[VG_CHR(minor)]->lv[l]->lv_dev)].nr_sects = + vg[VG_CHR(minor)]->lv[l]->lv_size; + lvm_size[MINOR(vg[VG_CHR(minor)]->lv[l]->lv_dev)] = + vg[VG_CHR(minor)]->lv[l]->lv_size >> 1; + /* vg_lv_map array doesn't have to be changed here */ + + LVM_CORRECT_READ_AHEAD ( vg[VG_CHR(minor)]->lv[l]->lv_read_ahead); + read_ahead[MAJOR_NR] = vg[VG_CHR(minor)]->lv[l]->lv_read_ahead; + vg[VG_CHR(minor)]->lv[l]->lv_status = lv_status_save; + + return 0; +} /* do_lv_extend_reduce () */ + + +/* + * support function initialize gendisk variables + */ +#ifdef __initfunc +__initfunc ( void lvm_geninit ( struct gendisk *lvm_gdisk)) +#else +void __init lvm_geninit ( struct gendisk *lvm_gdisk) +#endif +{ + int i = 0; + +#ifdef DEBUG_GENDISK + printk ( KERN_DEBUG "%s -- lvm_gendisk\n", lvm_name); +#endif + + for ( i = 0; i < MAX_LV; i++) { + lvm_gendisk.part[i].start_sect = -1; /* avoid partition check */ + lvm_size[i] = lvm_gendisk.part[i].nr_sects = 0; + lvm_blocksizes[i] = BLOCK_SIZE; + } + + blksize_size[MAJOR_NR] = lvm_blocksizes; + blk_size[MAJOR_NR] = lvm_size; + + return; +} /* lvm_gen_init () */ + + +#ifdef LVM_GET_INODE +/* + * support function to get an empty inode + * + * Gets an empty inode to be inserted into the inode hash, + * so that a physical volume can't be mounted. + * This is analog to drivers/block/md.c + * + * Is this the real thing? + * + */ +struct inode *lvm_get_inode ( kdev_t dev) { + struct inode *inode_this = NULL; + + /* Lock the device by inserting a dummy inode. */ + inode_this = get_empty_inode (); + inode_this->i_dev = dev; + insert_inode_hash ( inode_this); + return inode_this; +} + + +/* + * support function to clear an inode + * + */ +void lvm_clear_inode ( struct inode *inode) { +#ifdef I_FREEING + inode->i_state |= I_FREEING; +#endif + clear_inode ( inode); + return; +} +#endif /* #ifdef LVM_GET_INODE */ + + +/* my strlen */ +inline int lvm_strlen ( char *s1) { + int len = 0; + + while ( s1[len] != 0) len++; + return len; +} + + +/* my strcmp */ +inline int lvm_strcmp ( char *s1, char *s2) { + while ( *s1 != 0 && *s2 != 0) { + if ( *s1 != *s2) return -1; + s1++; s2++; + } + if ( *s1 == 0 && *s2 == 0) return 0; + return -1; +} + + +/* my strrchr */ +inline char *lvm_strrchr ( char *s1, char c) { + char *s2 = NULL; + + while ( *s1 != 0) { + if ( *s1 == c) s2 = s1; + s1++; + } + return s2; +} + + +/* my memcpy */ +inline void lvm_memcpy ( char *dest, char *source, int size) { + for ( ;size > 0; size--) *dest++ = *source++; +} diff -urN lvm-ref/include/linux/lvm.h lvm-aa/include/linux/lvm.h --- lvm-ref/include/linux/lvm.h Thu Jan 1 01:00:00 1970 +++ lvm-aa/include/linux/lvm.h Sun Nov 26 05:00:08 2000 @@ -0,0 +1,771 @@ +/* + * include/linux/lvm.h + * kernel/lvm.h + * + * Copyright (C) 1997 - 2000 Heinz Mauelshagen, Germany + * + * February-November 1997 + * May-July 1998 + * January-March,July,September,October,Dezember 1999 + * January,February 2000 + * + * lvm is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * lvm is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with GNU CC; see the file COPYING. If not, write to + * the Free Software Foundation, 59 Temple Place - Suite 330, + * Boston, MA 02111-1307, USA. + * + */ + +/* + * Changelog + * + * 10/10/1997 - beginning of new structure creation + * 12/05/1998 - incorporated structures from lvm_v1.h and deleted lvm_v1.h + * 07/06/1998 - avoided LVM_KMALLOC_MAX define by using vmalloc/vfree + * instead of kmalloc/kfree + * 01/07/1998 - fixed wrong LVM_MAX_SIZE + * 07/07/1998 - extended pe_t structure by ios member (for statistic) + * 02/08/1998 - changes for official char/block major numbers + * 07/08/1998 - avoided init_module() and cleanup_module() to be static + * 29/08/1998 - seprated core and disk structure type definitions + * 01/09/1998 - merged kernel integration version (mike) + * 20/01/1999 - added LVM_PE_DISK_OFFSET macro for use in + * vg_read_with_pv_and_lv(), pv_move_pe(), pv_show_pe_text()... + * 18/02/1999 - added definition of time_disk_t structure for; + * keeps time stamps on disk for nonatomic writes (future) + * 15/03/1999 - corrected LV() and VG() macro definition to use argument + * instead of minor + * 03/07/1999 - define for genhd.c name handling + * 23/07/1999 - implemented snapshot part + * 08/12/1999 - changed LVM_LV_SIZE_MAX macro to reflect current 1TB limit + * 01/01/2000 - extended lv_v2 core structure by wait_queue member + * 12/02/2000 - integrated Andrea Arcagnelli's snapshot work + * 18/02/2000 - seperated user and kernel space parts by + * #ifdef them with __KERNEL__ + * + */ + + +#ifndef _LVM_H_INCLUDE +#define _LVM_H_INCLUDE + +#define _LVM_H_VERSION "LVM 0.8final (22/02/2000)" + +#include + +/* + * preprocessor definitions + */ +/* if you like emergency reset code in the driver */ +#define LVM_TOTAL_RESET + +#ifdef __KERNEL__ +#define LVM_GET_INODE +#define LVM_HD_NAME + +/* lots of debugging output (see driver source) + #define DEBUG_LVM_GET_INFO + #define DEBUG + #define DEBUG_MAP + #define DEBUG_MAP_SIZE + #define DEBUG_IOCTL + #define DEBUG_READ + #define DEBUG_GENDISK + #define DEBUG_VG_CREATE + #define DEBUG_LVM_BLK_OPEN + #define DEBUG_KFREE + */ +#endif /* #ifdef __KERNEL__ */ + +#ifndef __KERNEL__ +#define __KERNEL__ +#include +#include +#undef __KERNEL__ +#else +#include +#include +#endif /* #ifndef __KERNEL__ */ + +#include +#include + +#ifdef __KERNEL__ +#if LINUX_VERSION_CODE >= KERNEL_VERSION ( 2, 3 ,0) +#include +#else +#include +#endif + +#include +#endif /* #ifdef __KERNEL__ */ + +#include + +#if !defined ( LVM_BLK_MAJOR) || !defined ( LVM_CHAR_MAJOR) +#error Bad include/linux/major.h - LVM MAJOR undefined +#endif + + +#define LVM_STRUCT_VERSION 1 /* structure version */ + +#ifndef min +#define min(a,b) (((a)<(b))?(a):(b)) +#endif +#ifndef max +#define max(a,b) (((a)>(b))?(a):(b)) +#endif + +/* set the default structure version */ +#if ( LVM_STRUCT_VERSION == 1) +#define pv_t pv_v1_t +#define lv_t lv_v2_t +#define vg_t vg_v1_t +#define pv_disk_t pv_disk_v1_t +#define lv_disk_t lv_disk_v1_t +#define vg_disk_t vg_disk_v1_t +#define lv_block_exception_t lv_block_exception_v1_t +#endif + + +/* + * i/o protocoll version + * + * defined here for the driver and defined seperate in the + * user land tools/lib/liblvm.h + * + */ +#define LVM_DRIVER_IOP_VERSION 6 + +#define LVM_NAME "lvm" + +/* + * VG/LV indexing macros + */ +/* character minor maps directly to volume group */ +#define VG_CHR(a) ( a) + +/* block minor indexes into a volume group/logical volume indirection table */ +#define VG_BLK(a) ( vg_lv_map[a].vg_number) +#define LV_BLK(a) ( vg_lv_map[a].lv_number) + +/* + * absolute limits for VGs, PVs per VG and LVs per VG + */ +#define ABS_MAX_VG 99 +#define ABS_MAX_PV 256 +#define ABS_MAX_LV 256 /* caused by 8 bit minor */ + +#define MAX_VG ABS_MAX_VG +#define MAX_LV ABS_MAX_LV +#define MAX_PV ABS_MAX_PV + +#if ( MAX_VG > ABS_MAX_VG) +#undef MAX_VG +#define MAX_VG ABS_MAX_VG +#endif + +#if ( MAX_LV > ABS_MAX_LV) +#undef MAX_LV +#define MAX_LV ABS_MAX_LV +#endif + + +/* + * VGDA: default disk spaces and offsets + * + * there's space after the structures for later extensions. + * + * offset what size + * --------------- ---------------------------------- ------------ + * 0 physical volume structure ~500 byte + * + * 1K volume group structure ~200 byte + * + * 5K time stamp structure ~ + * + * 6K namelist of physical volumes 128 byte each + * + * 6k + n * 128byte n logical volume structures ~300 byte each + * + * + m * 328byte m physical extent alloc. structs 4 byte each + * + * End of disk - first physical extent typical 4 megabyte + * PE total * + * PE size + * + * + */ + +/* DONT TOUCH THESE !!! */ +/* base of PV structure in disk partition */ +#define LVM_PV_DISK_BASE 0L + +/* size reserved for PV structure on disk */ +#define LVM_PV_DISK_SIZE 1024L + +/* base of VG structure in disk partition */ +#define LVM_VG_DISK_BASE LVM_PV_DISK_SIZE + +/* size reserved for VG structure */ +#define LVM_VG_DISK_SIZE ( 9 * 512L) + +/* size reserved for timekeeping */ +#define LVM_TIMESTAMP_DISK_BASE ( LVM_VG_DISK_BASE + LVM_VG_DISK_SIZE) +#define LVM_TIMESTAMP_DISK_SIZE 512L /* reserved for timekeeping */ + +/* name list of physical volumes on disk */ +#define LVM_PV_NAMELIST_DISK_BASE ( LVM_TIMESTAMP_DISK_BASE + \ + LVM_TIMESTAMP_DISK_SIZE) + +/* now for the dynamically calculated parts of the VGDA */ +#define LVM_LV_DISK_OFFSET(a, b) ( (a)->lv_on_disk.base + sizeof ( lv_disk_t) * b) +#define LVM_DISK_SIZE(pv) ( (pv)->pe_on_disk.base + \ + (pv)->pe_on_disk.size) +#define LVM_PE_DISK_OFFSET(pe, pv) ( pe * pv->pe_size + \ + ( LVM_DISK_SIZE ( pv) / SECTOR_SIZE)) +#define LVM_PE_ON_DISK_BASE(pv) \ + { int rest; \ + pv->pe_on_disk.base = pv->lv_on_disk.base + pv->lv_on_disk.size; \ + if ( ( rest = pv->pe_on_disk.base % SECTOR_SIZE) != 0) \ + pv->pe_on_disk.base += ( SECTOR_SIZE - rest); \ + } +/* END default disk spaces and offsets for PVs */ + + +/* + * LVM_PE_T_MAX corresponds to: + * + * 8KB PE size can map a ~512 MB logical volume at the cost of 1MB memory, + * + * 128MB PE size can map a 8TB logical volume at the same cost of memory. + * + * Default PE size of 4 MB gives a maximum logical volume size of 256 GB. + * + * Maximum PE size of 16GB gives a maximum logical volume size of 1024 TB. + * + * AFAIK, the actual kernels limit this to 1 TB. + * + * Should be a sufficient spectrum ;*) + */ + +/* This is the usable size of disk_pe_t.le_num !!! v v */ +#define LVM_PE_T_MAX ( ( 1 << ( sizeof ( __u16) * 8)) - 2) + +#define LVM_LV_SIZE_MAX(a) ( ( long long) LVM_PE_T_MAX * (a)->pe_size > ( long long) 2*1024*1024*1024 ? ( long long) 2*1024*1024*1024 : ( long long) LVM_PE_T_MAX * (a)->pe_size) +#define LVM_MIN_PE_SIZE ( 8L * 2) /* 8 KB in sectors */ +#define LVM_MAX_PE_SIZE ( 16L * 1024L * 1024L * 2) /* 16GB in sectors */ +#define LVM_DEFAULT_PE_SIZE ( 4096L * 2) /* 4 MB in sectors */ +#define LVM_DEFAULT_STRIPE_SIZE 16L /* 16 KB */ +#define LVM_MIN_STRIPE_SIZE ( PAGE_SIZE>>9) /* PAGESIZE in sectors */ +#define LVM_MAX_STRIPE_SIZE ( 512L * 2) /* 512 KB in sectors */ +#define LVM_MAX_STRIPES 128 /* max # of stripes */ +#define LVM_MAX_SIZE ( 1024LU * 1024 * 1024 * 2) /* 1TB[sectors] */ +#define LVM_MAX_MIRRORS 2 /* future use */ +#define LVM_MIN_READ_AHEAD 2 /* minimum read ahead sectors */ +#define LVM_MAX_READ_AHEAD 120 /* maximum read ahead sectors */ +#define LVM_MAX_LV_IO_TIMEOUT 60 /* seconds I/O timeout (future use) */ +#define LVM_PARTITION 0xfe /* LVM partition id */ +#define LVM_NEW_PARTITION 0x8e /* new LVM partition id (10/09/1999) */ +#define LVM_PE_SIZE_PV_SIZE_REL 5 /* max relation PV size and PE size */ + +#define LVM_SNAPSHOT_MAX_CHUNK 1024 /* 1024 KB */ +#define LVM_SNAPSHOT_DEF_CHUNK 64 /* 64 KB */ +#define LVM_SNAPSHOT_MIN_CHUNK 1 /* 1 KB */ + +#define UNDEF -1 +#define FALSE 0 +#define TRUE 1 + + +/* + * ioctls + */ +/* volume group */ +#define VG_CREATE _IOW ( 0xfe, 0x00, 1) +#define VG_REMOVE _IOW ( 0xfe, 0x01, 1) + +#define VG_EXTEND _IOW ( 0xfe, 0x03, 1) +#define VG_REDUCE _IOW ( 0xfe, 0x04, 1) + +#define VG_STATUS _IOWR ( 0xfe, 0x05, 1) +#define VG_STATUS_GET_COUNT _IOWR ( 0xfe, 0x06, 1) +#define VG_STATUS_GET_NAMELIST _IOWR ( 0xfe, 0x07, 1) + +#define VG_SET_EXTENDABLE _IOW ( 0xfe, 0x08, 1) + + +/* logical volume */ +#define LV_CREATE _IOW ( 0xfe, 0x20, 1) +#define LV_REMOVE _IOW ( 0xfe, 0x21, 1) + +#define LV_ACTIVATE _IO ( 0xfe, 0x22) +#define LV_DEACTIVATE _IO ( 0xfe, 0x23) + +#define LV_EXTEND _IOW ( 0xfe, 0x24, 1) +#define LV_REDUCE _IOW ( 0xfe, 0x25, 1) + +#define LV_STATUS_BYNAME _IOWR ( 0xfe, 0x26, 1) +#define LV_STATUS_BYINDEX _IOWR ( 0xfe, 0x27, 1) + +#define LV_SET_ACCESS _IOW ( 0xfe, 0x28, 1) +#define LV_SET_ALLOCATION _IOW ( 0xfe, 0x29, 1) +#define LV_SET_STATUS _IOW ( 0xfe, 0x2a, 1) + +#define LE_REMAP _IOW ( 0xfe, 0x2b, 1) + + +/* physical volume */ +#define PV_STATUS _IOWR ( 0xfe, 0x40, 1) +#define PV_CHANGE _IOWR ( 0xfe, 0x41, 1) +#define PV_FLUSH _IOW ( 0xfe, 0x42, 1) + +/* physical extent */ +#define PE_LOCK_UNLOCK _IOW ( 0xfe, 0x50, 1) + +/* i/o protocol version */ +#define LVM_GET_IOP_VERSION _IOR ( 0xfe, 0x98, 1) + +#ifdef LVM_TOTAL_RESET +/* special reset function for testing purposes */ +#define LVM_RESET _IO ( 0xfe, 0x99) +#endif + +/* lock the logical volume manager */ +#define LVM_LOCK_LVM _IO ( 0xfe, 0x100) +/* END ioctls */ + + +/* + * Status flags + */ +/* volume group */ +#define VG_ACTIVE 0x01 /* vg_status */ +#define VG_EXPORTED 0x02 /* " */ +#define VG_EXTENDABLE 0x04 /* " */ + +#define VG_READ 0x01 /* vg_access */ +#define VG_WRITE 0x02 /* " */ + +/* logical volume */ +#define LV_ACTIVE 0x01 /* lv_status */ +#define LV_SPINDOWN 0x02 /* " */ + +#define LV_READ 0x01 /* lv_access */ +#define LV_WRITE 0x02 /* " */ +#define LV_SNAPSHOT 0x04 /* " */ +#define LV_SNAPSHOT_ORG 0x08 /* " */ + +#define LV_BADBLOCK_ON 0x01 /* lv_badblock */ + +#define LV_STRICT 0x01 /* lv_allocation */ +#define LV_CONTIGUOUS 0x02 /* " */ + +/* physical volume */ +#define PV_ACTIVE 0x01 /* pv_status */ +#define PV_ALLOCATABLE 0x02 /* pv_allocatable */ + + +/* + * Structure definitions core/disk follow + * + * conditional conversion takes place on big endian architectures + * in functions * pv_copy_*(), vg_copy_*() and lv_copy_*() + * + */ + +#define NAME_LEN 128 /* don't change!!! */ +#define UUID_LEN 16 /* don't change!!! */ + +/* remap physical sector/rdev pairs including hash */ +typedef struct { + struct list_head hash; + ulong rsector_org; + kdev_t rdev_org; + ulong rsector_new; + kdev_t rdev_new; +} lv_block_exception_v1_t; + +/* disk stored pe information */ +typedef struct { + __u16 lv_num; + __u16 le_num; +} disk_pe_t; + +/* disk stored PV, VG, LV and PE size and offset information */ +typedef struct { + __u32 base; + __u32 size; +} lvm_disk_data_t; + + +/* + * Structure Physical Volume (PV) Version 1 + */ + +/* core */ +typedef struct { + __u8 id[2]; /* Identifier */ + __u16 version; /* HM lvm version */ + lvm_disk_data_t pv_on_disk; + lvm_disk_data_t vg_on_disk; + lvm_disk_data_t pv_namelist_on_disk; + lvm_disk_data_t lv_on_disk; + lvm_disk_data_t pe_on_disk; + __u8 pv_name[NAME_LEN]; + __u8 vg_name[NAME_LEN]; + __u8 system_id[NAME_LEN]; /* for vgexport/vgimport */ + kdev_t pv_dev; + __u32 pv_number; + __u32 pv_status; + __u32 pv_allocatable; + __u32 pv_size; /* HM */ + __u32 lv_cur; + __u32 pe_size; + __u32 pe_total; + __u32 pe_allocated; + __u32 pe_stale; /* for future use */ + + disk_pe_t *pe; /* HM */ + struct inode *inode; /* HM */ +} pv_v1_t; + +/* disk */ +typedef struct { + __u8 id[2]; /* Identifier */ + __u16 version; /* HM lvm version */ + lvm_disk_data_t pv_on_disk; + lvm_disk_data_t vg_on_disk; + lvm_disk_data_t pv_namelist_on_disk; + lvm_disk_data_t lv_on_disk; + lvm_disk_data_t pe_on_disk; + __u8 pv_name[NAME_LEN]; + __u8 vg_name[NAME_LEN]; + __u8 system_id[NAME_LEN]; /* for vgexport/vgimport */ + __u32 pv_major; + __u32 pv_number; + __u32 pv_status; + __u32 pv_allocatable; + __u32 pv_size; /* HM */ + __u32 lv_cur; + __u32 pe_size; + __u32 pe_total; + __u32 pe_allocated; +} pv_disk_v1_t; + + +/* + * Structure Physical Volume (PV) Version 2 (future!) + */ + +typedef struct { + __u8 id[2]; /* Identifier */ + __u16 version; /* HM lvm version */ + lvm_disk_data_t pv_on_disk; + lvm_disk_data_t vg_on_disk; + lvm_disk_data_t pv_uuid_on_disk; + lvm_disk_data_t lv_on_disk; + lvm_disk_data_t pe_on_disk; + __u8 pv_name[NAME_LEN]; + __u8 vg_name[NAME_LEN]; + __u8 system_id[NAME_LEN]; /* for vgexport/vgimport */ + kdev_t pv_dev; + __u32 pv_number; + __u32 pv_status; + __u32 pv_allocatable; + __u32 pv_size; /* HM */ + __u32 lv_cur; + __u32 pe_size; + __u32 pe_total; + __u32 pe_allocated; + __u32 pe_stale; /* for future use */ + disk_pe_t *pe; /* HM */ + struct inode *inode; /* HM */ + /* delta to version 1 starts here */ + __u8 pv_uuid[UUID_LEN]; + __u32 pv_atime; /* PV access time */ + __u32 pv_ctime; /* PV creation time */ + __u32 pv_mtime; /* PV modification time */ +} pv_v2_t; + + +/* + * Structures for Logical Volume (LV) + */ + +/* core PE information */ +typedef struct { + kdev_t dev; + __u32 pe; /* to be changed if > 2TB */ + __u32 reads; + __u32 writes; +} pe_t; + +typedef struct { + __u8 lv_name[NAME_LEN]; + kdev_t old_dev; + kdev_t new_dev; + __u32 old_pe; + __u32 new_pe; +} le_remap_req_t; + +/* + * Structure Logical Volume (LV) Version 1 + */ + +/* + * Structure Logical Volume (LV) Version 2 + */ + +/* core */ +typedef struct lv_v2 { + __u8 lv_name[NAME_LEN]; + __u8 vg_name[NAME_LEN]; + __u32 lv_access; + __u32 lv_status; + __u32 lv_open; /* HM */ + kdev_t lv_dev; /* HM */ + __u32 lv_number; /* HM */ + __u32 lv_mirror_copies; /* for future use */ + __u32 lv_recovery; /* " */ + __u32 lv_schedule; /* " */ + __u32 lv_size; + pe_t *lv_current_pe; /* HM */ + __u32 lv_current_le; /* for future use */ + __u32 lv_allocated_le; + __u32 lv_stripes; + __u32 lv_stripesize; + __u32 lv_badblock; /* for future use */ + __u32 lv_allocation; + __u32 lv_io_timeout; /* for future use */ + __u32 lv_read_ahead; + + /* delta to version 1 starts here */ + struct lv_v2 *lv_snapshot_org; + struct lv_v2 *lv_snapshot_prev; + struct lv_v2 *lv_snapshot_next; + lv_block_exception_t *lv_block_exception; + __u8 __unused; + __u32 lv_remap_ptr; + __u32 lv_remap_end; + __u32 lv_chunk_size; + __u32 lv_snapshot_minor; +#ifdef __KERNEL__ + struct kiobuf *lv_iobuf; + struct semaphore lv_snapshot_sem; + struct list_head *lv_snapshot_hash_table; + ulong lv_snapshot_hash_mask; +#else + char dummy[200]; +#endif +} lv_v2_t; + +/* disk */ +typedef struct { + __u8 lv_name[NAME_LEN]; + __u8 vg_name[NAME_LEN]; + __u32 lv_access; + __u32 lv_status; + __u32 lv_open; /* HM */ + __u32 lv_dev; /* HM */ + __u32 lv_number; /* HM */ + __u32 lv_mirror_copies; /* for future use */ + __u32 lv_recovery; /* " */ + __u32 lv_schedule; /* " */ + __u32 lv_size; + __u32 dummy; + __u32 lv_current_le; /* for future use */ + __u32 lv_allocated_le; + __u32 lv_stripes; + __u32 lv_stripesize; + __u32 lv_badblock; /* for future use */ + __u32 lv_allocation; + __u32 lv_io_timeout; /* for future use */ + __u32 lv_read_ahead; /* HM, for future use */ +} lv_disk_v1_t; + +/* + * Structure Volume Group (VG) Version 1 + */ + +/* core */ +typedef struct { + __u8 vg_name[NAME_LEN]; /* volume group name */ + __u32 vg_number; /* volume group number */ + __u32 vg_access; /* read/write */ + __u32 vg_status; /* active or not */ + __u32 lv_max; /* maximum logical volumes */ + __u32 lv_cur; /* current logical volumes */ + __u32 lv_open; /* open logical volumes */ + __u32 pv_max; /* maximum physical volumes */ + __u32 pv_cur; /* current physical volumes FU */ + __u32 pv_act; /* active physical volumes */ + __u32 dummy; /* was obsolete max_pe_per_pv */ + __u32 vgda; /* volume group descriptor arrays FU */ + __u32 pe_size; /* physical extent size in sectors */ + __u32 pe_total; /* total of physical extents */ + __u32 pe_allocated; /* allocated physical extents */ + __u32 pvg_total; /* physical volume groups FU */ + struct proc_dir_entry *proc; + pv_t *pv[ABS_MAX_PV + 1]; /* physical volume struct pointers */ + lv_t *lv[ABS_MAX_LV + 1]; /* logical volume struct pointers */ +} vg_v1_t; + +typedef struct { + __u8 vg_name[NAME_LEN]; /* volume group name */ + __u32 vg_number; /* volume group number */ + __u32 vg_access; /* read/write */ + __u32 vg_status; /* active or not */ + __u32 lv_max; /* maximum logical volumes */ + __u32 lv_cur; /* current logical volumes */ + __u32 lv_open; /* open logical volumes */ + __u32 pv_max; /* maximum physical volumes */ + __u32 pv_cur; /* current physical volumes FU */ + __u32 pv_act; /* active physical volumes */ + __u32 dummy; + __u32 vgda; /* volume group descriptor arrays FU */ + __u32 pe_size; /* physical extent size in sectors */ + __u32 pe_total; /* total of physical extents */ + __u32 pe_allocated; /* allocated physical extents */ + __u32 pvg_total; /* physical volume groups FU */ +} vg_disk_v1_t; + +/* + * Structure Volume Group (VG) Version 2 + */ + +typedef struct { + __u8 vg_name[NAME_LEN]; /* volume group name */ + __u32 vg_number; /* volume group number */ + __u32 vg_access; /* read/write */ + __u32 vg_status; /* active or not */ + __u32 lv_max; /* maximum logical volumes */ + __u32 lv_cur; /* current logical volumes */ + __u32 lv_open; /* open logical volumes */ + __u32 pv_max; /* maximum physical volumes */ + __u32 pv_cur; /* current physical volumes FU */ + __u32 pv_act; /* future: active physical volumes */ + __u32 max_pe_per_pv; /* OBSOLETE maximum PE/PV */ + __u32 vgda; /* volume group descriptor arrays FU */ + __u32 pe_size; /* physical extent size in sectors */ + __u32 pe_total; /* total of physical extents */ + __u32 pe_allocated; /* allocated physical extents */ + __u32 pvg_total; /* physical volume groups FU */ + struct proc_dir_entry *proc; + pv_t *pv[ABS_MAX_PV + 1]; /* physical volume struct pointers */ + lv_t *lv[ABS_MAX_LV + 1]; /* logical volume struct pointers */ + /* delta to version 1 starts here */ + __u8 vg_uuid[UUID_LEN]; /* volume group UUID */ + time_t vg_atime; /* VG access time */ + time_t vg_ctime; /* VG creation time */ + time_t vg_mtime; /* VG modification time */ +} vg_v2_t; + + +/* + * Timekeeping structure on disk (future) + * + * Holds several timestamps for start/stop time of non + * atomic VGDA disk i/o operations + * + */ + +typedef struct { + __u32 seconds; /* seconds since the epoch */ + __u32 jiffies; /* micro timer */ +} lvm_time_t; + +#define TIMESTAMP_ID_SIZE 2 +typedef struct { + __u8 id[TIMESTAMP_ID_SIZE]; /* Identifier */ + lvm_time_t pv_vg_lv_pe_io_begin; + lvm_time_t pv_vg_lv_pe_io_end; + lvm_time_t pv_io_begin; + lvm_time_t pv_io_end; + lvm_time_t vg_io_begin; + lvm_time_t vg_io_end; + lvm_time_t lv_io_begin; + lvm_time_t lv_io_end; + lvm_time_t pe_io_begin; + lvm_time_t pe_io_end; + lvm_time_t pe_move_io_begin; + lvm_time_t pe_move_io_end; + __u8 dummy[LVM_TIMESTAMP_DISK_SIZE - + TIMESTAMP_ID_SIZE - + 12 * sizeof(lvm_time_t)]; + /* ATTENTION ^^ */ +} timestamp_disk_t; + +/* same on disk and in core so far */ +typedef timestamp_disk_t timestamp_t; + +/* function identifiers for timestamp actions */ +typedef enum { + PV_VG_LV_PE_IO_BEGIN, + PV_VG_LV_PE_IO_END, + PV_IO_BEGIN, + PV_IO_END, + VG_IO_BEGIN, + VG_IO_END, + LV_IO_BEGIN, + LV_IO_END, + PE_IO_BEGIN, + PE_IO_END, + PE_MOVE_IO_BEGIN, + PE_MOVE_IO_END +} ts_fct_id_t; + + +/* + * Request structures for ioctls + */ + +/* Request structure PV_STATUS */ +typedef struct { + char pv_name[NAME_LEN]; + pv_t *pv; +} pv_status_req_t, pv_change_req_t; + +/* Request structure PV_FLUSH */ +typedef struct { + char pv_name[NAME_LEN]; + kdev_t pv_dev; +} pv_flush_req_t; + + +/* Request structure PE_MOVE */ +typedef struct { + enum { + LOCK_PE, UNLOCK_PE + } lock; + struct { + kdev_t lv_dev; + kdev_t pv_dev; + __u32 pv_offset; + } data; +} pe_lock_req_t; + + +/* Request structure LV_STATUS_BYNAME */ +typedef struct { + char lv_name[NAME_LEN]; + lv_t *lv; +} lv_status_byname_req_t, lv_req_t; + +/* Request structure LV_STATUS_BYINDEX */ +typedef struct { + __u32 lv_index; + lv_t *lv; +} lv_status_byindex_req_t; + +#endif /* #ifndef _LVM_H_INCLUDE */ diff -urN lvm-ref/kernel/ksyms.c lvm-aa/kernel/ksyms.c --- lvm-ref/kernel/ksyms.c Sun Nov 26 04:53:34 2000 +++ lvm-aa/kernel/ksyms.c Sun Nov 26 05:00:08 2000 @@ -85,6 +85,14 @@ #endif EXPORT_SYMBOL(get_options); +#ifdef CONFIG_BLK_DEV_LVM_MODULE + extern int (*lvm_map_ptr) ( int, kdev_t *, unsigned long *, + unsigned long, int); + extern void (*lvm_hd_name_ptr) ( char*, int); + EXPORT_SYMBOL(lvm_map_ptr); + EXPORT_SYMBOL(lvm_hd_name_ptr); +#endif + /* process memory management */ EXPORT_SYMBOL(do_mmap); EXPORT_SYMBOL(do_munmap); @@ -112,6 +120,7 @@ EXPORT_SYMBOL(mem_map); EXPORT_SYMBOL(remap_page_range); EXPORT_SYMBOL(max_mapnr); +EXPORT_SYMBOL(num_physpages); EXPORT_SYMBOL(high_memory); EXPORT_SYMBOL(update_vm_cache); EXPORT_SYMBOL(update_vm_cache_conditional);