d7f9016c8c
Bring back read-write to read-only mount update for ffs.
2017-03-01 10:46:43 +00:00
61b2de1d70
Remove now redundant calls to fstrans_start()/fstrans_done().
2017-03-01 10:41:28 +00:00
a378d58ecb
Enable fstrans on all file systems.
...
Welcome to 7.99.61
2017-02-22 09:50:13 +00:00
326db3aaf6
Add generic genfs_suspendctl() and use it for all file systems.
...
Layered file systems need work.
2017-02-17 08:31:23 +00:00
7599fb1f37
Bring back vrele_flush() to flush deferred vrele() o an suspended file system.
2017-02-17 08:30:00 +00:00
64a4d4bd11
Untangle VFS_SYNC() from VFS_SUSPENDCTL().
2017-02-17 08:29:11 +00:00
1740eca04e
Flush the log to disk when ffs_sync() gets called with MNT_WAIT.
2017-02-17 08:26:41 +00:00
65dda4ca5a
Sprinkle in a pinch of const, not too much, just enough
...
to add a little strength without affecting the overall balance...
2017-02-09 04:37:35 +00:00
ccc9d98e52
Add smaller versions of fsck_ffs(8) and newfs(8) for install media, where
...
support for Endian-Independent FFS and Apple UFS is disabled unless FFS_EI=1
and APPLE_UFS=1 are added to CRUNCHENV, respectively.
This reduces the size of ramdisk image for atari by over 15KB.
Thanks tsutsui and christos for their useful comments.
2017-02-08 16:11:39 +00:00
6a20f05bb9
Fix unsigned
2017-01-13 18:04:36 +00:00
31613a05ed
fix sign confusion
2017-01-12 18:40:02 +00:00
836eaad2e3
Change ufs_truncate_retry() to call UFS_TRUNCATE() at least once.
...
Even with "newsize == ip->i_size" it must set mtime etc.
Adresses PR kern/51762 "mtime not updated by open(O_TRUNC)"
2017-01-04 10:04:17 +00:00
958c9321b9
Fix a bug introduced with Rev. 1.294: use LK_NOWAIT when called with MNT_LAZY.
2016-12-27 10:54:38 +00:00
9a6283711b
KASSERT(mutex_owner(...)) ---> KASSERT(mutex_owned(...))
2016-11-20 21:22:14 +00:00
5f60d884f2
KASSERT(mutex_owner(...)) ---> KASSERT(mutex_owned(...))
...
Fixes part of PR kern/47114. Tested by code inspection.
2016-11-20 21:21:26 +00:00
3d6f4a8817
fix !WAPBL variant of UFS_WAPBL_REGISTER_DEALLOCATION()
2016-11-11 22:59:26 +00:00
3a541ed229
Fix a "slight tweak" from Rev. 1.121: bap1/bap2 must be valid
...
before using BAP_ASSIGN().
Prevents NULL pointer dereference when "lastbn >= 0".
2016-11-11 10:50:16 +00:00
f9c82ee8e7
disable discard when log is enabled to preserve log consistency promise
...
PR kern/50725
2016-11-10 22:19:23 +00:00
86e8a3aae2
during truncate with wapbl, register deallocation for upper indirect block
...
before recursing into lower blocks, to make sure that it will be removed after
all its referenced blocks are removed
fixes 'ffs_blkfree_common: freeing free block' panic triggered by
ufs_truncate_retry() when just the upper indirect block registration failed,
code tried to free the lower blocks again after wapbl flush
problem found by hannken@, thank you
2016-11-10 20:56:32 +00:00
6730f31e00
ffs_indirtrunc(): for !wapbl, restore rev 1.117 behavior of writing the zeroed
...
(indirect) block before freeing the referenced blocks; it's necessary for
fsck to recover the filesystem, if system goes down during truncate
patch courtesy of hannken@ with only sligh tweaks
2016-11-10 19:10:05 +00:00
7b5c072bc8
Apply ufs_extattr.c 1.48:
...
Explain why the lock in here needs to be recursive. Related to PR 46997.
ufs_extattr 1.47 was also committed directly here, so this file is still
fully synced with it.
2016-11-09 05:44:42 +00:00
5cbfa154c2
Explain why the lock in here needs to be recursive. Related to PR 46997.
2016-11-09 05:08:35 +00:00
1fd0461870
ufs_makeinode is declared file-static at the top of the file; mark it
...
at its definition too, for consistency and to avoid misleading casual
passersby.
2016-11-09 04:12:55 +00:00
d16cae2351
fix broken test for partial truncate, introduced in rev 1.118
...
PR kern/51601 kern/51602
2016-11-07 21:14:23 +00:00
84ac775fb4
reduce diff vs 1.117, no functional change
2016-11-07 21:05:38 +00:00
805187f5b9
Tidy up panic messages, no functional change.
2016-10-30 15:01:46 +00:00
b695bc874e
reorganize ffs_truncate()/ffs_indirtrunc() to be able to partially
...
succeed; change wapbl_register_deallocation() to return EAGAIN
rather than panic when code hits the limit
callers changed to either loop calling ffs_truncate() using new
utility ufs_truncate_retry() if their semantics requires it, or
just ignore the failure; remove ufs_wapbl_truncate()
this fixes possible user-triggerable panic during truncate, and
resolves WAPBL performance issue with truncates of large files
PR kern/47146 and kern/49175
2016-10-28 20:38:12 +00:00
3c9f488da8
revert 1.141 - the second ffs_truncate() can't really fail
...
requested by hannken@
2016-10-21 19:28:03 +00:00
e1aea9a285
allow also the snapshot_setup()'s call to ffs_truncate() fail, the code
...
should simply reuse the file blocks in that case; also make sure the
ffs_truncate() call is run within transaction if log is on
2016-10-20 20:17:46 +00:00
7470816b2a
add assertion to ensure ffs_cgupdate() is always called from
...
within a WAPBL transaction (if logging is on)
2016-10-20 19:31:32 +00:00
e433d111ad
Grr, the optimizer on mips64 can't handle this... Use MIN_PAGE_SIZE.
2016-10-04 16:46:20 +00:00
44548039a1
use __func__ and print the filesystem we are printing the message for.
2016-10-02 19:02:57 +00:00
16c7d9d735
allocate wapbl dealloc registration structures via pool, so that there is more
...
flexibility with limit handling
2016-10-01 13:15:45 +00:00
8de0ca1330
wapbl_remove_log(): add missing break; harmless, fallthrough just printed
...
extra debug message
2016-10-01 13:06:20 +00:00
9a0d9b14e1
adjust ffs_realloccg() so that the logic about allocating full
...
contiguous block for future fragment expansion doesn't need to
UFS_WAPBL_REGISTER_DEALLOCATION() or ffs_blkfree(); the free blocks
are now immediatelly available for use by the expanding file in further i/o
primary driver is safe removal of the deallocation registration and
hence failure point, but this also fixes degenerate case for wapbl,
and similar also for discard - if the file would be actually expanded
before wapbl commit, or before discard queue would be processed,
the filesystem would not yet see the contiguous free blocks, and
would be forced to allocate another fragment elsewhere
2016-09-25 17:14:59 +00:00
ae4b9ec77b
fix typo in #ifdef notyet part
2016-09-25 11:45:39 +00:00
7b230d53d8
fix swapped KASSERT()
2016-09-24 21:00:54 +00:00
f082cd3f0b
i/o optimization for wapbl flush - only sync superblock and cgs when
...
they were actually changed
2016-09-24 20:59:51 +00:00
7a68f84aa4
put back second strlcpy; pointed out by dholland.
2016-08-25 07:18:35 +00:00
36b34638d9
CID 1371648: off by one in index checking
...
KNF.
2016-08-23 06:40:54 +00:00
9d3a5ca7ac
KNF, no functional change
2016-08-23 06:40:25 +00:00
d7191ea7c2
CID 1371644: use strlcpy, remove dup copy.
2016-08-23 06:24:30 +00:00
67894636ac
CID 1371645: remove dead code
2016-08-23 06:23:26 +00:00
54bc00f99d
fix code which sets REV1 e2fs_fsmnt, set also mount time and mount count
2016-08-20 21:22:25 +00:00
34f2997571
adjust ext2fs_loadvnode_content() to do the sanity checking before allocating
...
memory, and avoid reallocaing memory on vnode reload
2016-08-20 20:05:28 +00:00
fa4a02b45a
modify the comment to note code needs to brele() to have a shot on actually
...
working
2016-08-20 19:53:43 +00:00
cd053bea0c
#if 0 the check for ext2fs_mapsearch() failure (similar what was done
...
for ffs counterpart), it actually never fails, it panics instead
2016-08-20 19:51:50 +00:00
1a50fd75de
add support for GDT_CSUM AKA uninit_bg feature
2016-08-20 19:47:44 +00:00
7546c749f2
whitespace fix
2016-08-20 19:45:20 +00:00
7139aab724
Remove now obsolete operation vcache_remove().
...
Welcome to 7.99.36
2016-08-20 12:37:06 +00:00
39cc5f5a3f
fix bug introduced in rev 1.82 of ext2fs_lookup.c, when ext2fs_add_entry()
...
was introduced splitting code from ext2fs_direnter() - code used
incorrect new entry size, leading to incomplete entry copy or buffer
overflow; fixed by passing the right size from ext2fs_direnter()
2016-08-19 00:05:43 +00:00
6ab7375654
EXT2F_INCOMPAT_FLEX_BG feature actually doesn't require any explicit
...
code changes, all magic is done by setting the block offsets appropriately
in group descriptors by newfs; add it to the list of supported INCOMPAT flags
2016-08-15 18:46:11 +00:00
a9097f8578
bump link limit to 65000 for files, and add support for EXT2F_ROCOMPAT_DIR_NLINK to make link count unlimited for directories
2016-08-15 18:38:10 +00:00
0471133b8d
adjust ext2fs_makeinode() so that the direnter is optional, use the function (with the direnter off) in ext2fs_mkdir() instead of the code copy; adjust ext2fs_makeinode() to initialize extra_isize and set creation time, if supported by the filesystem
2016-08-15 18:29:34 +00:00
008c8916aa
when converting on-disk direntry, only use the on-disk filetype if the feature flag is present
2016-08-14 11:46:05 +00:00
35875a51bc
switch code to use the EXT2_HAS_{COMPAT|ROCOMPAT|INCOMPAT}_FEATURE() macros instead of open coding the checks
2016-08-14 11:44:54 +00:00
03d15df771
switch ext2fs_htree_has_idx() over to EXT2F_HAS_COMPAT_FEATURE() and remove EXT2F_HAS_COMPAT_FEATURE() - this also fixes it for BE machines, as EXT2F_HAS_COMPAT_FEATURE() did extra byte swap; also remove XXX comment about IN_E3INDEX
2016-08-14 11:42:50 +00:00
6bd2592f36
add EXT2F_HAS_ROCOMPAT_FEATURE() macro, and change the current EXT2F_HAS_{COMPAT|INCOMPAT}_FEATURE() to take fs as first parameter
2016-08-14 11:40:31 +00:00
e393710782
again remove IN_E4EXTENTS; it's not used anywhere any more, and it's better to keep fs-specific flags out of generic headers anyway
2016-08-14 11:31:41 +00:00
437b8a430a
whitespace cleanup
2016-08-14 11:26:35 +00:00
9a22ef9e75
check correct inode extents flag - IN_E4EXTENTS is defined as 0x8000, correct flag EXT2_EXTENTS is 0x80000
2016-08-14 11:25:36 +00:00
01d303e744
KNF, no functional changes...
2016-08-13 07:40:10 +00:00
8378c129b0
sync with hrishi's git
2016-08-13 07:25:29 +00:00
034ad7ec18
cast pointers to uintptr_t before comparing them, also ()s
...
now this at least compiles
2016-08-12 20:30:15 +00:00
fd4db18413
sprinkle ()s in macros with comparisons, shuts up compiler warnings
2016-08-12 20:26:15 +00:00
2058250b3f
add support for extended attributes in ext2fs for ext3/ext4; read-only for now
2016-08-12 19:04:03 +00:00
58de1f8cbe
Undo revert now Christos has added the missing glue...
2016-08-09 21:08:02 +00:00
998fd93a6d
merge missing function.
2016-08-09 20:18:08 +00:00
a25bc2ccdc
Revert previous. This work isn't complete enough to include yet,
...
and the build of current really does need to go back to a working state.
2016-08-09 20:03:05 +00:00
09e85cd4b1
Revert previous - which itself (incorrectly) reverted the previous
...
changes, breaking the build.
2016-08-09 13:18:50 +00:00
50b9c754b0
More htree writing support (Hrishikesh Goyal GSoC 2016)
2016-08-09 07:15:35 +00:00
0bed509624
KNF
2016-08-09 06:40:24 +00:00
96374a6fa7
Remove unused <sys/tree.h>
2016-08-07 05:09:12 +00:00
23132aeaca
Fix stupid thinko.
2016-08-07 02:42:32 +00:00
5a9be9571b
comments
2016-08-07 02:31:03 +00:00
4cf5cc5427
If using constants from dirent.h it ought to be included.
...
Hopefully fixes i386 build.
2016-08-07 01:47:11 +00:00
fbc1063cb6
use static properly
2016-08-07 00:25:22 +00:00
5251c78329
Comments
2016-08-07 00:12:48 +00:00
8e39032fbb
actually pass the d_type from the on-disk directory entry to the lookup results
2016-08-06 21:39:48 +00:00
f048ac5803
typo in comment
2016-08-06 20:42:29 +00:00
118d3ec761
some more inode flags
2016-08-06 09:29:28 +00:00
5f1e3b3190
add defines for the missing ext4 feature flags
2016-08-05 21:22:06 +00:00
954e54d738
PR kern/7867 add support for UF_NODUMP flag to ext2fs
2016-08-05 20:15:41 +00:00
5b3ce9d448
add devel ifndefs for incompat/rocompat features so that it's possible
...
to ignore them and mount the filesystem; default is for the mount to fail
2016-08-05 20:06:55 +00:00
bddc3c1201
make E2MAXSYMLINKLEN just alias for EXT2_MAXSYMLINKLEN, they are the same
2016-08-04 17:50:51 +00:00
2f2ae0277c
move i_e2fs_* defines from ufs/inode.h to ext2fs/ext2fs_dinode.h, where they belong; they don't seem to be used anywhere else then ext2fs code any more
2016-08-04 17:47:47 +00:00
249e8f5f7e
rename struct ext2fs_dinode attribute e2di_dacl to correct
...
e2di_size_high; even Linux ext2 filesystem code actually uses it
unconditionally this way and ext4 code finally also calls it that way
in their struct definition too; if there was any trace of this for other
purpose it's long gone
2016-08-04 17:43:47 +00:00
c3d0e08698
include stddef.h for offsetof.
...
fix newfs_ext2fs build failure on evbppc.
2016-08-04 04:05:14 +00:00
96da721ffb
pass isize to e2fs_i_bswap() if BYTE_ORDER != LITTLE_ENDIAN.
2016-08-04 02:49:50 +00:00
31baab5623
Update previous. Since original format was %llu, replace it with
...
% PRIu64 (unsigned).
2016-08-03 23:33:59 +00:00
fa408750f2
get and set expanded timestamp if the inode contains the extra information, add support for create time
2016-08-03 23:29:05 +00:00
5cc4ca2173
Use correct printf() format for inode (fixes build for me)
2016-08-03 23:28:01 +00:00
966755011f
support arbitrary ext3/ext4 inode size, add all the new ext4 fields ext2fs_dinode, and add support for loading the extra inode data
2016-08-03 21:53:02 +00:00
8601c2c343
adjust the comments for on-disk ext2fs inode to indicate which of the ext* was it implemented for linux kernel; makes it a bit easier to locate
...
split e2di_linux_reserved3 with e2di_extra_isize and e2di_checksum_high, tag as ext4
2016-08-02 17:36:02 +00:00
1f7e7b3174
do not bswap fragment address, support in ext* for them was never actually implemented in linux kernels
2016-08-02 17:24:24 +00:00
f0b8f9c13b
From Michael Plass:
...
The superblock field that distinguishes between 4.2BSD and 4.4BSD
inodes is really only relevant on a UFS1 file system. Make sure that
it is a UFS1 fs before using fs_old_inodefmt.
Note that the NetBSD newfs and mkfs utilities initialize fs_old_inodefmt
even for UFS2, so problems were apparent only on file systems created
by other operating systems, for example, FreeBSD.
2016-07-28 08:24:58 +00:00
90668226c4
Don't do variable stack allocations for systems with non-const PAGE_SIZE;
...
instead assume that the smallest pagesize is 1024.
2016-07-21 18:10:47 +00:00
407bb4505c
Fix a deadlock
...
ok dholland@
2016-07-13 16:26:26 +00:00
8bc54e5be6
KNF. Remove extra spaces. No functional change.
2016-07-07 06:55:38 +00:00
26c7e22f4a
GSoC 2016 (Hrishikesh Goyal): Htree index support from FreeBSD
2016-06-24 17:21:30 +00:00
hannken