Re: lvm 0.9.1-beta1 still segfaults vgexport

From: Andrea Arcangeli (andrea@suse.de)
Date: Sun Jan 14 2001 - 11:32:34 EST


On Sun, Jan 14, 2001 at 09:31:29AM -0500, Todd M. Roy wrote:
> Andrea,
> Sorry to say but lvm 0.9.1-beta1 still segfaults
> at the same place, line 140 of pv_read_all_pv_of_vg.c
> pv_this is still null.

BTW, I can easily reproduce. I was near to go into it yesterday but got
interrupted by other issues (like the merging of the 0.9.1-beta1 kernel driver
and extraction of the strictly necessary fixes from the 0.9.1-beta1 userspace
against 0.9).

In the meantime here is the diff between my current kernel lvm driver and
0.9.1-beta1 applied on top of 2.4.1-pre3, Heinz could you merge the below
incremental fixes into your tree before merging with Linus? Thanks!

diff -urN -X /home/andrea/bin/dontdiff 2.4.1pre3-lvm-0.9.1_beta1/Documentation/Configure.help 2.4.1pre3-lvm-aa/Documentation/Configure.help
--- 2.4.1pre3-lvm-0.9.1_beta1/Documentation/Configure.help Fri Jan 12 04:14:36 2001
+++ 2.4.1pre3-lvm-aa/Documentation/Configure.help Sat Jan 13 16:09:24 2001
@@ -1450,15 +1450,6 @@
   want), say M here and read Documentation/modules.txt. The module
   will be called lvm-mod.o.
 
-Logical Volume Manager /proc file system 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 file system
- 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
Binary files 2.4.1pre3-lvm-0.9.1_beta1/ID and 2.4.1pre3-lvm-aa/ID differ
diff -urN -X /home/andrea/bin/dontdiff 2.4.1pre3-lvm-0.9.1_beta1/drivers/block/ll_rw_blk.c 2.4.1pre3-lvm-aa/drivers/block/ll_rw_blk.c
--- 2.4.1pre3-lvm-0.9.1_beta1/drivers/block/ll_rw_blk.c Tue Jan 2 17:41:15 2001
+++ 2.4.1pre3-lvm-aa/drivers/block/ll_rw_blk.c Sun Jan 14 17:20:42 2001
@@ -38,8 +38,6 @@
 extern int mac_floppy_init(void);
 #endif
 
-extern int lvm_init(void);
-
 /*
  * For the allocated request tables
  */
@@ -1267,9 +1265,6 @@
 #endif
 #ifdef CONFIG_SUN_JSFLASH
         jsfd_init();
-#endif
-#ifdef CONFIG_BLK_DEV_LVM
- lvm_init();
 #endif
         return 0;
 };
diff -urN -X /home/andrea/bin/dontdiff 2.4.1pre3-lvm-0.9.1_beta1/drivers/md/lvm-snap.c 2.4.1pre3-lvm-aa/drivers/md/lvm-snap.c
--- 2.4.1pre3-lvm-0.9.1_beta1/drivers/md/lvm-snap.c Sat Jan 13 16:10:33 2001
+++ 2.4.1pre3-lvm-aa/drivers/md/lvm-snap.c Sat Jan 13 16:12:22 2001
@@ -42,10 +42,6 @@
 
 static char *lvm_snap_version __attribute__ ((unused)) = "LVM 0.9 snapshot code (13/11/2000)\n";
 
-#ifndef LockPage
-#define LockPage(map) set_bit(PG_locked, &(map)->flags)
-#endif
-
 extern const char *const lvm_name;
 extern int lvm_blocksizes[];
 
@@ -469,7 +465,7 @@
                 goto out;
 
         err = -ENOMEM;
- iobuf->locked = 1;
+ iobuf->locked = 0;
         iobuf->nr_pages = 0;
         for (i = 0; i < nr_pages; i++)
         {
@@ -480,9 +476,6 @@
                         goto out;
 
                 iobuf->maplist[i] = page;
- /* the only point to lock the page here is to be allowed
- to share unmap_kiobuf() in the fail-path */
- LockPage(page);
                 iobuf->nr_pages++;
         }
         iobuf->offset = 0;
diff -urN -X /home/andrea/bin/dontdiff 2.4.1pre3-lvm-0.9.1_beta1/drivers/md/lvm.c 2.4.1pre3-lvm-aa/drivers/md/lvm.c
--- 2.4.1pre3-lvm-0.9.1_beta1/drivers/md/lvm.c Sat Jan 13 16:10:33 2001
+++ 2.4.1pre3-lvm-aa/drivers/md/lvm.c Sun Jan 14 16:19:03 2001
@@ -148,6 +148,8 @@
  * procfs is always supported now. (JT)
  * 12/01/2001 - avoided flushing logical volume in case of shrinking
  * because of unecessary overhead in case of heavy updates
+ * 07/12/2000 - make sure lvm_make_request_fn returns correct value - 0 or 1 - NeilBrown
+ * 25/12/2000 - fix procfs #defines - Christoph Hellwig
  *
  */
 
@@ -399,7 +401,7 @@
 /*
  * Driver initialization...
  */
-int __init lvm_init(void)
+int lvm_init(void)
 {
         struct gendisk *gendisk_ptr = NULL;
 
@@ -480,7 +482,7 @@
 /*
  * cleanup...
  */
-static void __exit lvm_cleanup(void)
+static void lvm_cleanup(void)
 {
         struct gendisk *gendisk_ptr = NULL, *gendisk_ptr_prev = NULL;
 
@@ -928,6 +930,7 @@
                 P_IOCTL("%s -- lvm_blk_ioctl -- BLKFLSBUF\n", lvm_name);
 
                 fsync_dev(inode->i_rdev);
+ invalidate_buffers(inode->i_rdev);
                 break;
 
 
@@ -1603,6 +1606,8 @@
         if (lv->lv_access & (LV_SNAPSHOT|LV_SNAPSHOT_ORG)) {
                 /* original logical volume */
                 if (lv->lv_access & LV_SNAPSHOT_ORG) {
+ /* Serializes the access to the lv_snapshot_next list */
+ down(&lv->lv_snapshot_sem);
                         if (rw == WRITE || rw == WRITEA)
                         {
                                 lv_t *lv_ptr;
@@ -1613,7 +1618,8 @@
                                      lv_ptr = lv_ptr->lv_snapshot_next) {
                                         /* Check for inactive snapshot */
                                         if (!(lv_ptr->lv_status & LV_ACTIVE)) continue;
- down(&lv->lv_snapshot_org->lv_snapshot_sem);
+ /* Serializes the COW with the accesses to the snapshot device */
+ 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;
@@ -1634,9 +1640,10 @@
                                                 rdev_tmp = rdev_sav;
                                                 rsector_tmp = rsector_sav;
                                         }
- up(&lv->lv_snapshot_org->lv_snapshot_sem);
+ up(&lv_ptr->lv_snapshot_sem);
                                 }
                         }
+ up(&lv->lv_snapshot_sem);
                 } else {
                         /* remap snapshot logical volume */
                         down(&lv->lv_snapshot_sem);
@@ -1698,8 +1705,10 @@
                                int rw,
                                struct buffer_head *bh)
 {
- lvm_map(bh, rw);
- return 1;
+ if (lvm_map(bh, rw)<0)
+ return 0; /* failure, buffer_IO_error has been called, don't recurse */
+ else
+ return 1; /* all ok, mapping done, call lower level driver */
 }
 
 
@@ -2268,18 +2277,10 @@
 
                                 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_ptr->lv[l]->lv_snapshot_prev = lv_ptr;
                                 /* our new one now back points to the previous last in the chain
                                    which can be the original logical volume */
                                 lv_ptr = vg_ptr->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_snapshot_le = lv_ptr->lv_allocated_le;
                                 lv_ptr->lv_allocated_le = lv_ptr->lv_snapshot_org->lv_allocated_le;
@@ -2347,16 +2348,21 @@
 
         /* optionally add our new snapshot LV */
         if (lv_ptr->lv_access & LV_SNAPSHOT) {
+ lv_t * org = lv_ptr->lv_snapshot_org, * last;
                 /* sync the original logical volume */
- fsync_dev(lv_ptr->lv_snapshot_org->lv_dev);
+ fsync_dev(org->lv_dev);
 #ifdef LVM_VFS_ENHANCEMENT
                 /* VFS function call to sync and lock the filesystem */
- fsync_dev_lockfs(lv_ptr->lv_snapshot_org->lv_dev);
+ fsync_dev_lockfs(org->lv_dev);
 #endif
- lv_ptr->lv_snapshot_org->lv_access |= LV_SNAPSHOT_ORG;
- lv_ptr->lv_access &= ~LV_SNAPSHOT_ORG;
- /* put ourselve into the chain */
- lv_ptr->lv_snapshot_prev->lv_snapshot_next = lv_ptr;
+ down(&org->lv_snapshot_sem);
+ org->lv_access |= LV_SNAPSHOT_ORG;
+ lv_ptr->lv_access &= ~LV_SNAPSHOT_ORG; /* this can only hide an userspace bug */
+ /* Link in the list of snapshot volumes */
+ for (last = org; last->lv_snapshot_next; last = last->lv_snapshot_next);
+ lv_ptr->lv_snapshot_prev = last;
+ last->lv_snapshot_next = lv_ptr;
+ up(&org->lv_snapshot_sem);
         }
 
         /* activate the logical volume */
@@ -2412,6 +2418,28 @@
             lv_ptr->lv_snapshot_next != NULL)
                 return -EPERM;
 
+ if (lv_ptr->lv_access & LV_SNAPSHOT) {
+ /*
+ * Atomically make the the snapshot invisible
+ * to the original lv before playing with it.
+ */
+ lv_t * org = lv_ptr->lv_snapshot_org;
+ down(&org->lv_snapshot_sem);
+
+ /* 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;
+ }
+ up(&org->lv_snapshot_sem);
+
+ /* no more snapshots? */
+ if (!org->lv_snapshot_next)
+ org->lv_access &= ~LV_SNAPSHOT_ORG;
+ lvm_snapshot_release(lv_ptr);
+ }
+
         lv_ptr->lv_status |= LV_SPINDOWN;
 
         /* sync the buffers */
@@ -2445,24 +2473,6 @@
                         }
                 }
                 vfree(lv_ptr->lv_current_pe);
- /* LV_SNAPSHOT */
- } else {
- down(&lv_ptr->lv_snapshot_org->lv_snapshot_sem);
-
- /* Update the VG PE(s) used by snapshot reserve space. */
- vg[VG_CHR(minor)]->pe_allocated -= lv_ptr->lv_allocated_snapshot_le;
-
- /* 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);
- up(&lv_ptr->lv_snapshot_org->lv_snapshot_sem);
         }
 
         devfs_unregister(lv_devfs_handle[lv_ptr->lv_number]);
@@ -3088,14 +3098,5 @@
         return uuid;
 }
 
-#ifdef MODULE
-int __init init_module(void)
-{
- return lvm_init();
-}
-
-void __exit cleanup_module(void)
-{
- lvm_cleanup();
-}
-#endif
+module_init(lvm_init);
+module_exit(lvm_cleanup);
diff -urN -X /home/andrea/bin/dontdiff 2.4.1pre3-lvm-0.9.1_beta1/include/linux/lvm.h 2.4.1pre3-lvm-aa/include/linux/lvm.h
--- 2.4.1pre3-lvm-0.9.1_beta1/include/linux/lvm.h Sat Jan 13 16:10:33 2001
+++ 2.4.1pre3-lvm-aa/include/linux/lvm.h Sun Jan 14 16:35:08 2001
@@ -58,6 +58,8 @@
  * 26/06/2000 - implemented snapshot persistency and resizing support
  * 02/11/2000 - added hash table size member to lv structure
  * 12/11/2000 - removed unneeded timestamp definitions
+ * 24/12/2000 - removed LVM_TO_{CORE,DISK}*, use cpu_{from, to}_le*
+ * instead - Christoph Hellwig
  *
  */
 

downloadable from here:

        ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/lvm/kernel-2.4.1pre3-0.9.1_beta1-fixes-1

Andrea
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Mon Jan 15 2001 - 21:00:39 EST