]> sourceware.org Git - lvm2.git/log
lvm2.git
22 years agoo Test program for dbg_malloc unit.
Joe Thornber [Tue, 2 Oct 2001 12:27:55 +0000 (12:27 +0000)]
o Test program for dbg_malloc unit.

  I'm postfixing test programs with _t, and benchmarks with _b

22 years agoAnother step towards consistency & compilation.
Alasdair Kergon [Mon, 1 Oct 2001 22:12:10 +0000 (22:12 +0000)]
Another step towards consistency & compilation.

22 years agoMisc structural changes.
Alasdair Kergon [Mon, 1 Oct 2001 19:36:06 +0000 (19:36 +0000)]
Misc structural changes.

22 years agoMisc structural changes.
Alasdair Kergon [Mon, 1 Oct 2001 19:29:52 +0000 (19:29 +0000)]
Misc structural changes.

22 years agoo changed dev-manager to a dev_filter
Joe Thornber [Mon, 1 Oct 2001 16:21:21 +0000 (16:21 +0000)]
o changed dev-manager to a dev_filter

22 years agoo get block size moved to dev-io.c
Joe Thornber [Mon, 1 Oct 2001 16:07:29 +0000 (16:07 +0000)]
o get block size moved to dev-io.c

22 years agomerge partition code
Alasdair Kergon [Mon, 1 Oct 2001 15:59:40 +0000 (15:59 +0000)]
merge partition code

22 years agoTidy includes
Alasdair Kergon [Mon, 1 Oct 2001 15:53:21 +0000 (15:53 +0000)]
Tidy includes

22 years agoo drop the reference counting in the devices.
Joe Thornber [Mon, 1 Oct 2001 15:43:51 +0000 (15:43 +0000)]
o drop the reference counting in the devices.

22 years agoMerge fixes
Alasdair Kergon [Mon, 1 Oct 2001 15:29:39 +0000 (15:29 +0000)]
Merge fixes

22 years agoo missing *
Joe Thornber [Mon, 1 Oct 2001 15:28:28 +0000 (15:28 +0000)]
o missing *

22 years agoo added filter type.
Joe Thornber [Mon, 1 Oct 2001 15:27:16 +0000 (15:27 +0000)]
o added filter type.

22 years agoTidy include files
Alasdair Kergon [Mon, 1 Oct 2001 15:14:39 +0000 (15:14 +0000)]
Tidy include files

22 years agoCreate symlinks to .h files in an include directory
Alasdair Kergon [Mon, 1 Oct 2001 13:36:54 +0000 (13:36 +0000)]
Create symlinks to .h files in an include directory

22 years agoo I'm splitting dev-manager in two. dev-cache is the bottom layer that
Joe Thornber [Fri, 28 Sep 2001 15:42:25 +0000 (15:42 +0000)]
o I'm splitting dev-manager in two.  dev-cache is the bottom layer that
  handles devices.  Dev-manager will sit on this filtering the view.

22 years agoo list.h from kernel for userland tools to use.
Joe Thornber [Fri, 28 Sep 2001 13:19:17 +0000 (13:19 +0000)]
o list.h from kernel for userland tools to use.

22 years agoo logical data structures
Joe Thornber [Fri, 28 Sep 2001 13:15:30 +0000 (13:15 +0000)]
o logical data structures

22 years agoo generic hash table to store void *'s, not efficient, but adequate for LVM.
Joe Thornber [Fri, 28 Sep 2001 13:08:44 +0000 (13:08 +0000)]
o generic hash table to store void *'s, not efficient, but adequate for LVM.

22 years agoo rewrite of dm_user_bmap, not tested though.
Joe Thornber [Thu, 27 Sep 2001 10:15:02 +0000 (10:15 +0000)]
o rewrite of dm_user_bmap, not tested though.

22 years agoFix prototype for malloc_aux
Patrick Caulfield [Thu, 27 Sep 2001 10:01:17 +0000 (10:01 +0000)]
Fix prototype for malloc_aux

22 years agoDon't include asm/* files
Patrick Caulfield [Thu, 27 Sep 2001 10:00:47 +0000 (10:00 +0000)]
Don't include asm/* files

22 years ago o Kill write funcs for error/status files
Steven Whitehouse [Wed, 26 Sep 2001 20:24:39 +0000 (20:24 +0000)]
 o Kill write funcs for error/status files
 o Redo write logic for table file
 o Relax rules for symlink content by removing the rewriting function

Well I probably won't get a chance to work on this tomorrow, so this is my
changeset to date.

22 years agoo table creation works again.
Joe Thornber [Wed, 26 Sep 2001 19:48:20 +0000 (19:48 +0000)]
o table creation works again.

22 years agoo It should build now
Joe Thornber [Wed, 26 Sep 2001 17:32:57 +0000 (17:32 +0000)]
o It should build now

22 years agoo typos
Joe Thornber [Wed, 26 Sep 2001 17:07:10 +0000 (17:07 +0000)]
o typos

22 years agoo remove steve's insane ramblings from my code.
Joe Thornber [Wed, 26 Sep 2001 14:32:07 +0000 (14:32 +0000)]
o remove steve's insane ramblings from my code.

22 years ago o Lunchtime.
Steven Whitehouse [Wed, 26 Sep 2001 11:47:02 +0000 (11:47 +0000)]
 o Lunchtime.

22 years ago o More updates
Steven Whitehouse [Wed, 26 Sep 2001 09:26:10 +0000 (09:26 +0000)]
 o More updates

22 years ago o Beginnings of new interface.
Steven Whitehouse [Wed, 26 Sep 2001 08:06:46 +0000 (08:06 +0000)]
 o Beginnings of new interface.

22 years agofix memory leak
Alasdair Kergon [Tue, 25 Sep 2001 16:26:38 +0000 (16:26 +0000)]
fix memory leak

22 years agoo quick tidy up
Joe Thornber [Tue, 25 Sep 2001 15:23:20 +0000 (15:23 +0000)]
o quick tidy up

22 years agolvactivate checkpoint commit
Alasdair Kergon [Tue, 25 Sep 2001 12:49:28 +0000 (12:49 +0000)]
lvactivate checkpoint commit

22 years agoAnother missing dependency.
Alasdair Kergon [Mon, 24 Sep 2001 22:44:06 +0000 (22:44 +0000)]
Another missing dependency.

22 years ago o Check in case of setting up volumes before root is mounted.
Steven Whitehouse [Mon, 24 Sep 2001 15:18:45 +0000 (15:18 +0000)]
 o Check in case of setting up volumes before root is mounted.

22 years agoWell when things start looking so complicated that future direction becomes
Steven Whitehouse [Mon, 24 Sep 2001 15:10:33 +0000 (15:10 +0000)]
Well when things start looking so complicated that future direction becomes
non-obvious, its time to simplify :-)

 o Moving towards a simpler and more obviously correct interface
 o Removed some fs operations in directories representing volumes
 o Changed some file names
 o Made things cleaner

more changes to follow...

22 years agoFix dependencies.
Alasdair Kergon [Mon, 24 Sep 2001 12:05:04 +0000 (12:05 +0000)]
Fix dependencies.

22 years agoMakefiles & autoconf.
Alasdair Kergon [Fri, 21 Sep 2001 12:37:43 +0000 (12:37 +0000)]
Makefiles & autoconf.

22 years agoInitialise root node pointer.
Alasdair Kergon [Fri, 21 Sep 2001 12:32:37 +0000 (12:32 +0000)]
Initialise root node pointer.

22 years agoUnused variables.
Alasdair Kergon [Fri, 21 Sep 2001 12:31:57 +0000 (12:31 +0000)]
Unused variables.

22 years ago o Sync up of todays changes .... nothing very important
Steven Whitehouse [Thu, 20 Sep 2001 22:58:06 +0000 (22:58 +0000)]
 o Sync up of todays changes .... nothing very important

22 years ago o Bug fix in error path
Steven Whitehouse [Thu, 20 Sep 2001 20:22:15 +0000 (20:22 +0000)]
 o Bug fix in error path

22 years ago o Use ERR_PTR and PTR_ERR rather than an extra argument.
Steven Whitehouse [Thu, 20 Sep 2001 19:25:58 +0000 (19:25 +0000)]
 o Use ERR_PTR and PTR_ERR rather than an extra argument.

22 years ago o Changed to use table->err_msg rather than passing functions around
Steven Whitehouse [Thu, 20 Sep 2001 18:22:35 +0000 (18:22 +0000)]
 o Changed to use table->err_msg rather than passing functions around

22 years ago o Some new files (also part of new fs interface)
Steven Whitehouse [Wed, 19 Sep 2001 21:28:25 +0000 (21:28 +0000)]
 o Some new files (also part of new fs interface)

22 years ago o Further changes to new file system interface
Steven Whitehouse [Wed, 19 Sep 2001 21:27:46 +0000 (21:27 +0000)]
 o Further changes to new file system interface

22 years ago o Removed the error reporting function from the target constructor function
Steven Whitehouse [Wed, 19 Sep 2001 21:27:15 +0000 (21:27 +0000)]
 o Removed the error reporting function from the target constructor function
   arguments. Errors are now reported by setting a pointer in the table to
   point to an error message.

22 years agohardsect/blksize handling
Alasdair Kergon [Wed, 19 Sep 2001 17:46:27 +0000 (17:46 +0000)]
hardsect/blksize handling

22 years ago o Added ref counting to tables
Steven Whitehouse [Wed, 19 Sep 2001 16:01:27 +0000 (16:01 +0000)]
 o Added ref counting to tables
 o Further changes to new fs interface

22 years ago o Fixed a bug where we were not holding a reference of the block devices
Steven Whitehouse [Wed, 19 Sep 2001 14:54:44 +0000 (14:54 +0000)]
 o Fixed a bug where we were not holding a reference of the block devices
   used by the targets correctly.

22 years ago o Change the deallocation of tables to match the vmalloc changes in my
Steven Whitehouse [Wed, 19 Sep 2001 11:02:02 +0000 (11:02 +0000)]
 o Change the deallocation of tables to match the vmalloc changes in my
   previous commit

22 years ago o Cut down number of vmallocs to increase speed and efficiency
Steven Whitehouse [Wed, 19 Sep 2001 10:59:10 +0000 (10:59 +0000)]
 o Cut down number of vmallocs to increase speed and efficiency

22 years ago o More fs fiddling. Another check point commit.
Steven Whitehouse [Wed, 19 Sep 2001 10:32:51 +0000 (10:32 +0000)]
 o More fs fiddling. Another check point commit.

22 years ago o Bug fix to LV_BMAP ioctl()
Steven Whitehouse [Wed, 19 Sep 2001 10:32:09 +0000 (10:32 +0000)]
 o Bug fix to LV_BMAP ioctl()
 o Account for I/O against tables rather than logical volume devices

22 years agoFix allocation & list-handling.
Alasdair Kergon [Tue, 18 Sep 2001 20:03:00 +0000 (20:03 +0000)]
Fix allocation & list-handling.

22 years agoThese files are now a bit closer towards what I'm aiming at. Still a lot
Steven Whitehouse [Tue, 18 Sep 2001 16:54:14 +0000 (16:54 +0000)]
These files are now a bit closer towards what I'm aiming at. Still a lot
more to do though.

22 years agoShould have been included in the previous commit.
Steven Whitehouse [Tue, 18 Sep 2001 16:53:18 +0000 (16:53 +0000)]
Should have been included in the previous commit.

22 years ago o Changed dm_create() to return a struct mapped_device rather than an int
Steven Whitehouse [Tue, 18 Sep 2001 16:52:50 +0000 (16:52 +0000)]
 o Changed dm_create() to return a struct mapped_device rather than an int
 o Changed dm_remove() to accept a struct mapped_device argument rather than
   a name
 o We no longer have to look up devices by name, the dcache handles that
   nicely for us
 o Fixed a bug where we were freeing a structure before we'd finished with
   it.
 o The name field in struct mapped_device is now only used in a very few
   places in dm.c and will be replaced in future with a back reference to
   the dentry rather than keeping the name in two places.

22 years ago o New file dmfs-super.c
Steven Whitehouse [Tue, 18 Sep 2001 15:38:54 +0000 (15:38 +0000)]
 o New file dmfs-super.c
 o dmfs-dir.c becomes dmfs-lv.c
 o dmfs-file.c becomes dmfs-table.c
 o A few tweeks and updates

The main reason for the slow progress on these files (which are not yet used
by the device mapper) is that we are working out what this interface should
look like as we go along.

Once this has evolved a bit further and in a state where it can be used we'll
announce it on the lists for further comment.

22 years agoError checking: only allow block devices & test for 'nodev'.
Alasdair Kergon [Mon, 17 Sep 2001 21:17:30 +0000 (21:17 +0000)]
Error checking: only allow block devices & test for 'nodev'.

22 years ago o Again, please ignore this for the time being.
Steven Whitehouse [Mon, 17 Sep 2001 19:05:49 +0000 (19:05 +0000)]
 o Again, please ignore this for the time being.

22 years agoo Arbitrary mount path.
Alasdair Kergon [Mon, 17 Sep 2001 16:55:31 +0000 (16:55 +0000)]
o Arbitrary mount path.
o Name length 128.

22 years ago o Work in progress, please ignore these files for a day or two whilst I
Steven Whitehouse [Mon, 17 Sep 2001 15:42:59 +0000 (15:42 +0000)]
 o Work in progress, please ignore these files for a day or two whilst I
   get everything going.

22 years ago o Fixed code where return value of vmalloc wasn't checked
Steven Whitehouse [Mon, 17 Sep 2001 11:23:13 +0000 (11:23 +0000)]
 o Fixed code where return value of vmalloc wasn't checked

22 years ago o Use count should be an atomic_t
Steven Whitehouse [Mon, 17 Sep 2001 09:01:23 +0000 (09:01 +0000)]
 o Use count should be an atomic_t

22 years ago o Targets now get rw passed through so they can do COW for example
Steven Whitehouse [Fri, 14 Sep 2001 16:22:02 +0000 (16:22 +0000)]
 o Targets now get rw passed through so they can do COW for example
 o Added error handler (not sure that this is the "correct" way to do
   this at the moment, so its a bit exprimental for now)

22 years ago o Support /sbin/hotplug
Steven Whitehouse [Fri, 14 Sep 2001 15:35:06 +0000 (15:35 +0000)]
 o Support /sbin/hotplug

22 years ago o Made pending I/O wait uninterruptible
Steven Whitehouse [Fri, 14 Sep 2001 14:03:02 +0000 (14:03 +0000)]
 o Made pending I/O wait uninterruptible

22 years ago o New ioctl(): LV_BMAP which is compatible with LVM so that hopefully LILO
Steven Whitehouse [Fri, 14 Sep 2001 13:45:40 +0000 (13:45 +0000)]
 o New ioctl(): LV_BMAP which is compatible with LVM so that hopefully LILO
   will work. I haven't actually tested that, but this support at least will
   be required.

22 years agoUse dmfs_ function name prefix (in line with other file systems).
Alasdair Kergon [Fri, 14 Sep 2001 13:27:58 +0000 (13:27 +0000)]
Use dmfs_ function name prefix (in line with other file systems).

22 years agolc->in->f_op->read expects its buffer to be in userspace so surround it in
Patrick Caulfield [Fri, 14 Sep 2001 12:27:57 +0000 (12:27 +0000)]
lc->in->f_op->read expects its buffer to be in userspace so surround it in
set_fs() etc calls

22 years ago o As promised earlier, the device registration is now hashed and the
Steven Whitehouse [Fri, 14 Sep 2001 11:25:51 +0000 (11:25 +0000)]
 o As promised earlier, the device registration is now hashed and the
   lists are private to dm-blkdev.c

22 years agoAlways truncate error file.
Alasdair Kergon [Fri, 14 Sep 2001 11:15:54 +0000 (11:15 +0000)]
Always truncate error file.

22 years agoDon't store things in _devs[-1] - it's not nice.
Patrick Caulfield [Fri, 14 Sep 2001 10:54:08 +0000 (10:54 +0000)]
Don't store things in _devs[-1] - it's not nice.

22 years ago o Forgot to create slab caches for dm-blkdev.c
Steven Whitehouse [Fri, 14 Sep 2001 10:40:20 +0000 (10:40 +0000)]
 o Forgot to create slab caches for dm-blkdev.c
 o Misc code tidy

22 years agoo kmalloc error check
Alasdair Kergon [Fri, 14 Sep 2001 10:06:22 +0000 (10:06 +0000)]
o kmalloc error check
o error file mode
The 1st Jan 1970 date I'm seeing in /dev is a devfs issue I think.

22 years agoChanges to device handling;
Steven Whitehouse [Fri, 14 Sep 2001 09:45:35 +0000 (09:45 +0000)]
Changes to device handling;

 o Only one list of block devices for all tables
 o Locking to ensure that block devices only get opened once
 o Block device handling is now in dm-blkdev.c
 o We open block devices when we create the tables and hold them open until
   the table is destroyed (this prevents the module for the device being
   unloaded after table parsing and before the table is used)
 o We compute the hardsect size when the table is created rather than when
   someone requests it.

Still to fix/change:

 o Probably want to hash the device lists in dm-blkdev.c and also remove refs
   to struct dm_bdev outside this file.
 o Need to ensure that hardsect_size doesn't change when new tables are
   swapped in (maybe this ought to be a per volume parameter and the tables
   will only parse if they match the value for the volume?).

Things are changing fast here, so if you want a stable version of thic code
try checking out yesterdays.

22 years ago o New code for handling block device registration. Not yet used but checked
Steven Whitehouse [Fri, 14 Sep 2001 08:06:02 +0000 (08:06 +0000)]
 o New code for handling block device registration. Not yet used but checked
   in for backup purposes.

22 years agois_identifier characters
Alasdair Kergon [Thu, 13 Sep 2001 21:50:38 +0000 (21:50 +0000)]
is_identifier characters

22 years ago o Two fixes which Alasdair pointed out.
Steven Whitehouse [Thu, 13 Sep 2001 20:10:14 +0000 (20:10 +0000)]
 o Two fixes which Alasdair pointed out.

22 years ago o Tidy in dm-fs.c
Steven Whitehouse [Thu, 13 Sep 2001 19:41:46 +0000 (19:41 +0000)]
 o Tidy in dm-fs.c
 o Magic number is really magic
 o Check on directory names

22 years ago o Forgot to add ref to module.
Steven Whitehouse [Thu, 13 Sep 2001 19:36:40 +0000 (19:36 +0000)]
 o Forgot to add ref to module.

22 years agoAdd dm-parse to makefile
Alasdair Kergon [Thu, 13 Sep 2001 19:09:23 +0000 (19:09 +0000)]
Add dm-parse to makefile

22 years ago o Now we handle target modules correctly
Steven Whitehouse [Thu, 13 Sep 2001 18:30:05 +0000 (18:30 +0000)]
 o Now we handle target modules correctly
 o Moved the linear target into its own module (not really because it needs to
   be there, but because its useful to have a simple example so people can see
   what we are doing)

Btw, this needs testing properly.

22 years agoo Remove hard-coded mount point
Alasdair Kergon [Thu, 13 Sep 2001 16:52:50 +0000 (16:52 +0000)]
o Remove hard-coded mount point
o Fix macro for compilation

22 years ago o Some structures change size with different configs, so always include
Steven Whitehouse [Thu, 13 Sep 2001 14:03:42 +0000 (14:03 +0000)]
 o Some structures change size with different configs, so always include
   <linux/config.h> first.

22 years ago o Some ioctl() commands.
Steven Whitehouse [Thu, 13 Sep 2001 14:01:13 +0000 (14:01 +0000)]
 o Some ioctl() commands.

22 years agoFix includes so that string functions get prototyped
Patrick Caulfield [Thu, 13 Sep 2001 12:38:31 +0000 (12:38 +0000)]
Fix includes so that string functions get prototyped

22 years agoFix includes so that string functions get prototyped.
Patrick Caulfield [Thu, 13 Sep 2001 12:38:08 +0000 (12:38 +0000)]
Fix includes so that string functions get prototyped.
Fix cast - repeat after me Joe: "I must not cast pointers to ints"!

22 years ago o I'm afraid that wu and wl etc. is just too confusing.... I've changed it
Steven Whitehouse [Thu, 13 Sep 2001 11:29:38 +0000 (11:29 +0000)]
 o I'm afraid that wu and wl etc. is just too confusing.... I've changed it
   to up_write() and down_write() etc so that you can see what kind of a lock
   it is (otherwise it could be anything.. semaphore, spinlock, spinlock_bh,
   spinlock_irq, br_lock, etc.)

22 years ago o Use kmem_cache_destroy() to remove slab cache.
Steven Whitehouse [Thu, 13 Sep 2001 11:07:08 +0000 (11:07 +0000)]
 o Use kmem_cache_destroy() to remove slab cache.

22 years agoSet DEFAULT_ALIGNMENT to 8 for Alpha.
Patrick Caulfield [Thu, 13 Sep 2001 09:03:42 +0000 (09:03 +0000)]
Set DEFAULT_ALIGNMENT to 8 for Alpha.
If you think this is wasteful on other arches then stick some ifdefs in.

22 years agoAdd dm-parse
Alasdair Kergon [Wed, 12 Sep 2001 13:50:26 +0000 (13:50 +0000)]
Add dm-parse

22 years agoo first sat\7f\7ftab at custom fs. Very rough ATM.
Joe Thornber [Fri, 7 Sep 2001 11:34:46 +0000 (11:34 +0000)]
o first sat\7f\7ftab at custom fs.  Very rough ATM.

Mount the dm-fs filesystem on /device-mapper (will fix later).  mkdir
to create a device, inside that directory every file you create is a table
file.  If there are errors <table>.err will appear automagically.  Mv a table
file to ACTIVE to active\7fate the device.  I'm not happy with mv being the
binding command, symlink would be better.

22 years agoo more tidy ups from Clausen.
Joe Thornber [Wed, 5 Sep 2001 07:48:11 +0000 (07:48 +0000)]
o more tidy ups from Clausen.

22 years agoo Added two new functions get_child [Andrew Clausen] and get_node. I think
Joe Thornber [Tue, 4 Sep 2001 10:17:28 +0000 (10:17 +0000)]
o Added two new functions get_child [Andrew Clausen] and get_node.  I think
  this makes 'high()' a bit more understandable.

22 years agoo added new constant CHILD_PER_NODE to make things clearer
Joe Thornber [Mon, 3 Sep 2001 08:36:41 +0000 (08:36 +0000)]
o added new constant CHILD_PER_NODE to make things clearer

22 years agoo Various tidy ups [Andrew Clausen]
Joe Thornber [Sun, 2 Sep 2001 10:49:20 +0000 (10:49 +0000)]
o Various tidy ups [Andrew Clausen]

22 years agoo added global dm_table_lookup_device(path)
Joe Thornber [Fri, 31 Aug 2001 18:26:27 +0000 (18:26 +0000)]
o added global dm_table_lookup_device(path)

o changed linear target to \7f: <device_path> <start>

22 years agoo added reference counting to the destination devices, make sure that the
Joe Thornber [Fri, 31 Aug 2001 16:36:56 +0000 (16:36 +0000)]
o added reference counting to the destination devices, make sure that the
  destructor for any targets you write call dm_table_remove_device.

22 years agoo split struct mapped_device into mapped_device and dm_table
Joe Thornber [Fri, 31 Aug 2001 15:13:33 +0000 (15:13 +0000)]
o split struct mapped_device into mapped_device and dm_table

o seperated loading of a table from binding a table to the device

These should allow multiple tables to be managed by dm-fs

This page took 0.064106 seconds and 5 git commands to generate.