Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
107,875 Commits 606 Branches 0 Tags 3.1 GiB
a3ff3a3038
Commit Graph

754 Commits

Author SHA1 Message Date
fvdl
a3ff3a3038 Bump daddr_t to 64 bits. Replace it with int32_t in all places where 
it was used on-disk, so that on-disk formats remain the same.
Remove ufs_daddr_t and ufs_lbn_t for the time being.
 2003-01-24 21:55:02 +00:00
thorpej
b78f59b443 Merge the nathanw_sa branch. 2003-01-18 08:51:40 +00:00
yamt
03e1a46833 - zerofill struct lfs when allocating it. 
- use M_ZERO instead of memset after malloc.
 2003-01-12 13:04:52 +00:00
yamt
5f254d46cc backout wrong assertions that i added. 2003-01-08 17:16:52 +00:00
yamt
99d625b53c for lfs_remove/lfs_rmdir, keep removed vnodes marked VDIROP. 
(backout parts of rev.1.40)
otherwise, directory structures can be corrupted because checkpoints can
occur via eg. lfs_vflush before parent directory is written.
 2003-01-08 17:14:58 +00:00
yamt
69f1c0cb29 in set_dirop/endop, use normal vref/vrele instead of lfs versions 
so that we don't miss lfs_inactivate.
 2003-01-08 15:43:29 +00:00
yamt
ee36fccabb add assertions. 2003-01-08 15:40:54 +00:00
yamt
49d2b56b43 use lfs_unmark_vnode instead of duplicated code fragments. 2003-01-08 15:40:04 +00:00
wiz
1035faff1d writable, not writeable. 2003-01-06 20:30:28 +00:00
chs
822a8f2c0f several bugs: 
- move calls to softdep_setup_pagecache() (which can sleep to allocate
   memory) outside the softdep lock.
 - replace the softdep_flush_indir() hack (which tries to find another
   vnode to fsync when we are holding lots of buffer-cache buffers locked
   for long periods of time) with softdep_trackbufs() (which just kicks
   the syncer and sleeps under the same circumstances).  the former method
   had a lock-ordering problem which would occasionally deadlock.
 - relax the assertion in softdep_sync_metadata() which says that we should
   never see D_ALLOCDIRECT deps for VREG vnodes.  it's ok to see those
   attached to indirect blocks.

also, there's no need to splbio() while allocating the buffer headers
to which pagecache dependencies are attached, so remove that.

fixes all the problems in PR 19288.
 2003-01-01 23:08:56 +00:00
yamt
a5bf83bbfc don't set vnode type to VNON in error case of ufs_makeinode. 
(backout rev.1.74)

it seems that there's no need to do it (anymore?) and LFS has trouble with it.
(VNON vnodes marked VDIROP will never reclaimed)

ok'ed by Frank van der Linden.
 2002-12-31 15:00:18 +00:00
yamt
140a8e56ca write ifile only when it has dirty buffers. 2002-12-31 14:54:32 +00:00
yamt
cb9613feef comment and assertions 2002-12-30 05:34:17 +00:00
yamt
6fc496c67a move check of lfs_unlockvp from lfs_reserveavail to lfs_reserve 
because lfs_reservebuf needs same check as well.
 2002-12-30 05:31:53 +00:00
yamt
a05fbf74c0 fix vref/vunref mismatch. 2002-12-29 14:08:12 +00:00
yamt
88ae33f9e0 backout assertions in lfs_inactive. 
they can be false when unmounting forcibly.
 2002-12-29 07:05:55 +00:00
christos
ae2bf40b7e fix compile problem. 2002-12-28 20:08:36 +00:00
yamt
d840722863 avoid warnings without DIAGNOSTIC. 
pointed by Andreas Wrede.
 2002-12-28 17:22:47 +00:00
yamt
a428d8a5af dirop inode can't be passed to lfs_inactivate. 2002-12-28 15:12:26 +00:00
yamt
59be5399b7 - in lfs_reserve, vref vnodes that we're locking so that cleaner doesn't 
try to reclaim them.
  (workaround for deadlock noted in the comment in lfs_reserveavail)
- in lfs_rename, mark vnodes which are being moved as well as directry vnodes.
 2002-12-28 14:39:08 +00:00
hannken
c122326822 Clear IN_SPACECOUNTED on (re-)used inodes. 
This cures the "unmount pending error:" on softdep umounts.

Approved by: Frank van der Linden <fvdl@netbsd.org>
 2002-12-27 16:07:13 +00:00
yamt
4b9c604ba7 - in lfs_reserve, reserve locked buffer count as well. 
- don't wait for locking buf in lfs_bwrite_ext to avoid deadlocks.
- skip lfs_reserve when we're doing dirop.
  reserve more (for lfs_truncate) in set_dirop instead.

this mostly solves PR 18972. (and hopefully PR 19196)
 2002-12-26 13:37:18 +00:00
yamt
e9bd1836a5 don't try to write all blocks passed to lfs_markv at once 
since it likely causes buf starvation.
 2002-12-26 13:04:39 +00:00
yamt
362c57a2d2 add a XXX comment. (description of possible deadlock) 2002-12-22 17:31:52 +00:00
yamt
4370be165c add a XXX comment 2002-12-21 05:35:54 +00:00
yamt
0d95cc5d66 correct/add assertion. 2002-12-18 14:05:50 +00:00
yamt
beff0dd387 #if 0 out vnode unlock/lock in lfs_reserve for now and add a comment about it. 
deadlock is better than corruption (or panic), IMO.
 2002-12-17 15:23:37 +00:00
yamt
a999523301 no need for cleaner to hold vnode locks. 
cleaner and normal vnode operations are synchronized enough by
seglock/fraglock and buf's B_BUSY-ness.
 2002-12-17 14:37:49 +00:00
yamt
b2d5b49e2b use ufs_daddr_t instead of int where appropriate. 2002-12-17 14:28:54 +00:00
yamt
a79cb6db43 - in lfs_bwrite_ext, if we're cleaner, 
mark inode IN_CLEANING rather then IN_MODIFIED.
  otherwise cleaned (indirect) blocks belongs to the inode isn't written
  until next sync.
- add assertions.
 2002-12-14 13:41:25 +00:00
yamt
e5ea55e4ea in lfs_writefile, check v_type==VNON earlier. 
to avoid null dereference with DEBUG_LFS_VERBOSE.
 2002-12-14 11:54:47 +00:00
yamt
8fe8a4ced8 save a segment write when doing checkpoint. 2002-12-13 14:40:02 +00:00
yamt
275b3a47a2 correct DIAGNOSTIC code for duplicated inodes in a segment and su_nbytes. 2002-12-12 12:28:13 +00:00
yamt
9097ffce96 take care of B_CLRBUF in lfs_balloc. 
otherwise you'll see uninitialized blocks.
 2002-12-11 13:34:14 +00:00
matt
60db16d1ff Add multiple inclusion protection for headers. Fix mismatched 
variable declarations (missing const's) as needed.
 2002-12-01 00:12:06 +00:00
kristerw
fa033b67e7 Softdep is mature enough that it shouldn't define DEBUG and DIAGNOSTIC 
unconditionally.
 2002-11-30 20:27:50 +00:00
yamt
2331faab98 more XXX comment. 2002-11-27 11:36:40 +00:00
lukem
0635de35a3 Remove KDIR=, since SYS_INCLUDE=symlinks and KDIR are not supported any more. 2002-11-26 23:30:07 +00:00
yamt
42a8b926f9 eliminate i_ino from in-core inode 
and use local variable instead.

ok'ed by Frank van der Linden.
 2002-11-26 01:23:30 +00:00
thorpej
d6f8cc841d Avoid strict-alias warnings. 2002-11-25 01:55:21 +00:00
thorpej
b409a344b5 Avoid strict-alias warnings. 2002-11-25 01:44:21 +00:00
yamt
3af39ea015 in lfs_fakebuf, make corresponding buffer busy to avoid 
reading blocks that isn't written yet.
it's needed because we'll update metadatas in lfs_updatemeta
before data pointed by them is actually written to disk.

XXX should be solved with fake inode/indirect blocks instead?
 2002-11-24 16:39:13 +00:00
yamt
290fa35864 add a XXX comment to lfs_reserve. 
* it isn't safe to unlock vp here
 * because we're passing data using inode from namei.
 * (eg. i_offset)
 2002-11-24 16:09:50 +00:00
scw
7009056578 Quell an uninitialised variable warning. 2002-11-24 11:09:13 +00:00
yamt
eca07565c3 make sure i_lfs_fragsize is initialized. 
fix panic "lfs_writefile: more than one fragment!"
PR 18974.
 2002-11-24 08:43:26 +00:00
yamt
16a26d41e8 lfs_sync should wait at lfs_writer, not lfs_dirops. 
PR 18973.
 2002-11-24 08:37:43 +00:00
yamt
feacf34c09 lfs_reserve shouldn't block for lfs_unlockvp. 
otherwise cleaner deadlocks.
PR 19134.
 2002-11-24 08:32:22 +00:00
yamt
37b4f42285 blksize() macro shouldn't used for indirect blocks. 
this fixes "getblk: block size invariant failed" panic.
PR 18977.
 2002-11-24 08:27:00 +00:00
yamt
7d0ba73802 correct locking for lfs_rmdir. PR 18976. 2002-11-24 08:23:41 +00:00
wiz
29d58d0333 s/sqiud/squid/ in comment, reported by skrueger at europe com. 2002-11-04 16:59:37 +00:00