Commit Graph

691 Commits

Author SHA1 Message Date
gehenna 77a6b82b27 Merge the gehenna-devsw branch into the trunk.
This merge changes the device switch tables from static array to
dynamically generated by config(8).

- All device switches is defined as a constant structure in device drivers.

- The new grammer ``device-major'' is introduced to ``files''.

	device-major <prefix> char <num> [block <num>] [<rules>]

- All device major numbers must be listed up in port dependent majors.<arch>
  by using this grammer.

- Added the new naming convention.
  The name of the device switch must be <prefix>_[bc]devsw for auto-generation
  of device switch tables.

- The backward compatibility of loading block/character device
  switch by LKM framework is broken. This is necessary to convert
  from block/character device major to device name in runtime and vice versa.

- The restriction to assign device major by LKM is completely removed.
  We don't need to reserve LKM entries for dynamic loading of device switch.

- In compile time, device major numbers list is packed into the kernel and
  the LKM framework will refer it to assign device major number dynamically.
2002-09-06 13:18:43 +00:00
thorpej 139cdc3125 Make nbuf, nswbuf, and bufpages unsigned. Make all operations on these
variables unsigned, and update places where their values are printed.
2002-08-25 20:21:33 +00:00
itojun 8dd04cdcd7 correct range check, have overflow check, fix type mismatches,
for cmap args and some other calls.  from openbsd
2002-08-03 00:12:48 +00:00
soren 178d83d503 Die, qaddr_t, die! - mnt_data in struct mount is already effectively
a void *, so stop pretending otherwise.
2002-07-30 07:40:07 +00:00
wiz 645df36eff Spell '[Rr]ight' correctly. From Jim Bernard. 2002-07-26 14:11:34 +00:00
hannken 7de36862a8 Rename bufq_init() to bufq_alloc().
Add bufq_free() to remove a buffer queue.
Avoid MALLOC while holding a spinlock.

From Chuck Silvers.
2002-07-21 15:32:17 +00:00
hannken d4c062b4cc Convert to new device buffer queue interface. 2002-07-19 16:26:01 +00:00
perseant 8f30dc2c9b Remove lying comment on SEGM_PROT seglock. 2002-07-11 21:09:00 +00:00
briggs 77f5558791 Fix a printf format warning. 2002-07-07 14:29:06 +00:00
fredette 10d4232908 Fixed a printf argument type. 2002-07-06 15:39:07 +00:00
perseant 32ae84b188 Deal with fragment size changes better. For each fragment that can
exist on an on-disk inode, we keep a record of its size in struct inode,
which is updated when we write the block to disk.  The cleaner routines
thus have ready access to what size is the correct size for this block,
on disk.

Fixed a related bug: if a file with fragments is being cleaned
(fragments being cleaned) at the same time it is being extended beyond
NDADDR blocks, we could write a bogus FINFO record that has a frag in the
middle; when it was cleaned this would give back bogus file data.  Don't
write the indirect blocks in this case, since there is no need.

lfs_fragextend and lfs_truncate no longer require the seglock, but instead
take a shared lock, which the seglock locks exclusively.
2002-07-06 01:30:11 +00:00
scw 881a4dcac0 Cast pointers first to uintptr_t before casting to register_t.
On SH-5, sizeof(register_t) is always 8, even if sizeof(void *) is 4
as is the case when compiling for ILP32.
2002-07-05 13:49:26 +00:00
yamt d566a58b5e fix printf format for DEBUG_LFS. 2002-07-02 19:07:03 +00:00
perseant 0418a2c352 Fix miscalculation in lfs_fits found by Trevin Beattie <trevin@xmission.com>.
Change some of the variable names from "nb", "db" to "fsb" to reflect their
calling conventions.
2002-06-20 22:10:24 +00:00
perseant ae37d9d186 Don't bomb out of lfs_bmapv if the caller is requesting blocks that
live in the current segment.  There's nothing wrong with this, and
it is necessary for the correct operation of the coaleascer.
2002-06-20 20:43:17 +00:00
jdolecek 20644ff75f clear_inodedeps(): use CIRCLEQ_FOREACH() appropriately 2002-06-18 20:24:31 +00:00
perseant ddfb1dbb92 For synchronous writes, keep separate i/o counters for each write, so
processes don't have to wait for one another to finish (e.g., nfsd seems
to be a little happier now, though I haven't measured the difference).
Synchronous checkpoints, however, must always wait for all i/o to finish.

Take the contents of the callback functions and have them run in thread
context instead (aiodoned thread).  lfs_iocount no longer has to be
protected in splbio(), and quite a bit less of the segment construction
loop needs to be in splbio() as well.

If lfs_markv is handed a block that is not the correct size according to
the inode, refuse to process it.  (Formerly it was extended to the "correct"
size.)  This is possibly more prone to deadlock, but less prone to corruption.

lfs_segclean now outright refuses to clean segments that appear to have live
bytes in them.  Again this may be more prone to deadlock but avoids
corruption.

Replace ufsspec_close and ufsfifo_close with LFS equivalents; this means
that no UFS functions need to know about LFS_ITIMES any more.  Remove
the reference from ufs/inode.h.

Tested on i386, test-compiled on alpha.
2002-06-16 00:13:15 +00:00
chs ea4c4a989f allow read-only mounts even if we can't read the last fragment of the fs.
this enables one to recover data from a failing disk (where the read failure
is a hardware problem) while avoiding corrupting the fs further (in the case
where the read failure is due to a misconfiguration).
2002-06-09 16:46:49 +00:00
perseant c13ae45a2a Let lfs_bmapv fill in the bi_size member of the BLOCK_INFO structure,
as well as bi_daddr.  This lets the cleaner have an idea of what the size
of this block was at the time it was written without having to refer to
a segment header (e.g., in the file coalescing case).

Tested on i386.
2002-06-06 00:46:24 +00:00
chs fffb1de109 get the units right when computing a blkno in the ENOSPC path
for allocations involving indirect blocks.
spotted by Trevin Beattie <trevin@xmission.com>.
2002-06-05 05:23:51 +00:00
thorpej 7903aba812 #if 0 a test that is always false (and the XXX comment above it
indicates so).
2002-05-30 18:54:55 +00:00
perseant d67a5bbb21 Fix a couple of instances where reassignbuf() was not done at splbio.
Tested on i386.
2002-05-24 22:13:57 +00:00
perseant 43ca783b4a Back out rev 1.174 of vfs_subr.c, because the splbio() wasn't protecting
enough to be useful, and broadening it so that it did would have meant
that operations possibly requiring synchronous disk activity would have
to be done in splbio().  This clearly was not going to work.

Worked around this in the LFS case by having lfs_cluster_callback put an
extra hold on the vnode before calling biodone(), and taking the hold
off without HOLDRELE's problematic list swapping.  lfs_vunref() will take
care of that---in thread context---on the next write if need be.

Also, ensure that the list walking in lfs_{writevnodes,segunlock,gather}
takes into account the possibility that the list may change
underneath it (possibly because it itself deleted an element).

Tested on i386, test-compiled on alpha.
2002-05-23 23:05:25 +00:00
perseant ec0ca919be Protect v_freelist with splbio(), since HOLDRELE can be called in
interrupt context (through brelvp).  (LFS may be the only subsystem
affected by this problem.)

Tested on i386.
2002-05-20 22:50:57 +00:00
perseant 36efaa3565 use macros from <sys/queue.h> 2002-05-17 21:42:38 +00:00
thorpej 6c1654256e Fix LP64 printf format warning. 2002-05-16 02:23:55 +00:00
perseant 8886b0f4b2 Phase one of my three-phase plan to make LFS play nice with UBC, and bug-fixes
I found while making sure there weren't any new ones.

* Make the write clusters keep track of the buffers whose blocks they contain.
  This should make it possible to (1) write clusters using a page mapping
  instead of malloc, if desired, and (2) schedule blocks for rewriting
  (somewhere else) if a write error occurs.  Code is present to use
  pagemove() to construct the clusters but that is untested and will go away
  anyway in favor of page mapping.
* DEBUG now keeps a log of Ifile writes, so that any lingering instances of
  the "dirty bufs" problem can be properly debugged.
* Keep track of whether the Ifile has been dirtied by various routines that
  can be called by lfs_segwrite, and loop on that until it is clean, for
  a checkpoint.  Checkpoints need to be squeaky clean.
* Warn the user (once) if the Ifile grows larger than is reasonable for their
  buffer cache.  Both lfs_mountfs and lfs_unmount check since the Ifile can
  grow.
* If an inode is not found in a disk block, try rereading the block, under
  the assumption that the block was copied to a cluster and then freed.
* Protect WRITEINPROG() with splbio() to fix a hang in lfs_update.
2002-05-14 20:03:53 +00:00
mycroft 1523c4c12f In ufs_mkdir(), write the data block *before* updating the inode with the
block pointer, to prevent "DIRECTORY CORRUPTED" errors from fsck(8).
Note: The behavior in the softdep case is unchanged, but needs to be fixed.
2002-05-14 17:37:52 +00:00
matt fed7110558 Commit out code that's no longer used. 2002-05-14 02:46:22 +00:00
matt 0cb85bc7b9 Eliminate commons. 2002-05-12 23:06:27 +00:00
enami b86c56a0b6 Add comment that getblk() in ufs_bmaparray() returns an error only if
we are pagedaemon.
2002-05-11 12:23:53 +00:00
chs e926e6ec99 use the correct size when zeroing an array. 2002-05-05 17:01:41 +00:00
chs dcc6963777 for softdep vnodes, always write together the pages for any block that
might have a dependency , since the accounting doesn't work otherwise.
fixes PRs 15364 16336 16448.
2002-05-05 17:00:06 +00:00
perseant 76d2795556 Make exported LFSes not panic on the first file create. 2002-04-27 01:00:46 +00:00
thorpej 37dc008ca3 Cleanup how file system configuration information is declared, grouping
related information together, with the file system code itself.

This is just low-hanging fruit -- more to come.
2002-04-16 23:14:05 +00:00
mycroft fd303c4dc5 Add a special case for nrpos=1 to cbtorpos(). This massively reduces CPU usage
by newfs(8) -- and fsck_ffs(8) on a relatively empty file system.  There is
still one divide left in the inner loops, to calculate cylno values.
2002-04-10 14:31:07 +00:00
mycroft afc5d40400 Use blkstofrags() and fragstoblks(). Use &(NBBY-1) rather than %NBBY.
Switch off of fs_fragshift rather than fs_frag (generates better jump tables).
2002-04-10 08:05:11 +00:00
mycroft 0a9b835878 Use fsbtodb() rather than multiplying by NSPF(). 2002-04-10 07:46:10 +00:00
enami 89cf6e2727 Hold an extra reference if updating and args.fspec == NULL. 2002-04-01 07:51:58 +00:00
christos e356d686bb Fixes from enami:
- If VOP_ACCESS fails when updating mount, we will vrele() twice.

- The check for update-only flags in mp->mnt_flag when not updating
  case is bogus.  If we really want to check, we need to see flags in
  ufs_args, but I'm not sure if it is really necessary.

- The credential passed to ffs_reload was credential of when looking
  up mount point, but now it is credential of when looking up device
  node.  Anyway, it may be current process's credential.
2002-04-01 01:52:44 +00:00
christos 919d9f5617 PR/16136: Chris Jepeway: Bogus entry in /etc/fstab can panic kernel. 2002-03-31 20:53:25 +00:00
chs fe10bac175 if the size argument to write(2) is 0, do not modify the file in any way,
including updating timestamps.  required for standards conformance.
2002-03-25 02:23:55 +00:00
chs 0f2018fc31 in lfs_write(), flush and invalidate any page cache pages in the range
that we're about to modify.  this weak attempt at coherency is enough
to make some applications (eg. "tail -f") happy, so it's worth having.
2002-03-22 03:57:35 +00:00
wiz 358ed3f6d4 Fix a typo, a KNF-nit, and simplify a printf format string. 2002-03-18 13:38:52 +00:00
chs 79c365e60e don't do any flush-behind for async mounts.
this matches the traditional behaviour.
2002-03-17 23:58:09 +00:00
chs c1d184702f when mounting a filesystem, read the last block in the filesystem
to verify that the device is at least as big as the superblock claims
the filesystem is supposed to be, and if it's not then fail the mount.
this should help reduce the type of confusion reported in PR 13228.
2002-03-17 00:02:34 +00:00
thorpej a180cee23b Pool deals fairly well with physical memory shortage, but it doesn't
deal with shortages of the VM maps where the backing pages are mapped
(usually kmem_map).  Try to deal with this:

* Group all information about the backend allocator for a pool in a
  separate structure.  The pool references this structure, rather than
  the individual fields.
* Change the pool_init() API accordingly, and adjust all callers.
* Link all pools using the same backend allocator on a list.
* The backend allocator is responsible for waiting for physical memory
  to become available, but will still fail if it cannot callocate KVA
  space for the pages.  If this happens, carefully drain all pools using
  the same backend allocator, so that some KVA space can be freed.
* Change pool_reclaim() to indicate if it actually succeeded in freeing
  some pages, and use that information to make draining easier and more
  efficient.
* Get rid of PR_URGENT.  There was only one use of it, and it could be
  dealt with by the caller.

From art@openbsd.org.
2002-03-08 20:48:27 +00:00
simonb 9a942a34e0 Don't use local extern declarations for the mountroot variable or
declare local prototypes for nfs_mountroot() or md_root_setconf().
2002-03-04 02:25:21 +00:00
pooka 360cafaddb Don't add fs->fs_pendingblocks to f_bavail twice. It's already included
in f_bfree, which is added to f_bavail.

Fixes problem with statfs reporting too much free space for filesystems
which have files pending to be freed by softdeps.
2002-02-28 21:59:23 +00:00
enami 70ca5d5195 Record some page cache related information into ubchist. 2002-02-22 08:23:16 +00:00