"refreshing" the label on the vnode before use, just get the label
right from inception. For single-label file systems, set the label
in the generic VFS getnewvnode() code; for multi-label file systems,
leave the labeling up to the file system. With UFS1/2, this means
reading the extended attribute during vfs_vget() as the inode is
pulled off disk, rather than hitting the extended attributes
frequently during operations later, improving performance. This
also corrects sematics for shared vnode locks, which were not
previously present in the system. This chances the cache
coherrency properties WRT out-of-band access to label data, but in
an acceptable form. With UFS1, there is a small race condition
during automatic extended attribute start -- this is not present
with UFS2, and occurs because EAs aren't available at vnode
inception. We'll introduce a work around for this shortly.
Approved by: re
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories
check for and/or report I/O errors. The result is that a VFS_SYNC
or VOP_FSYNC called with MNT_WAIT could loop infinitely on ufs in
the presence of a hard error writing a disk sector or in a filesystem
full condition. This patch ensures that I/O errors will always be
checked and returned. This patch also ensures that every call to
VFS_SYNC or VOP_FSYNC with MNT_WAIT set checks for and takes
appropriate action when an error is returned.
Sponsored by: DARPA & NAI Labs.
the ffs_copyonwrite routine to avoid a deadlock between the syncer
daemon trying to sync out a snapshot vnode and the bufdaemon
trying to write out a buffer containing the snapshot inode.
With any luck this will be the last snapshot race condition.
Sponsored by: DARPA & NAI Labs.
a full filesystem. Previously, if the allocation failed, we had to
fsync the file before rolling back any partial allocation of indirect
blocks. Most block allocation requests only need to allocate a single
data block and if that allocation fails, there is nothing to unroll.
So, before doing the fsync, we check to see if any rollback will
really be necessary. If none is necessary, then we simply return.
This update eliminates the flurry of disk activity that got triggered
whenever a filesystem would run out of space.
Sponsored by: DARPA & NAI Labs.
that works in the new threaded kernel. It was commented out of
the disksort routine earlier this year for the reasons given in
kern/subr_disklabel.c (which is where this code used to reside
before it moved to kern/subr_disk.c):
----------------------------
revision 1.65
date: 2002/04/22 06:53:20; author: phk; state: Exp; lines: +5 -0
Comment out Kirks io-request priority hack until we can do this in a
civilized way which doesn't cause grief.
The problem is that it is not generally safe to cast a "struct bio
*" to a "struct buf *". Things like ccd, vinum, ata-raid and GEOM
constructs bio's which are not entrails of a struct buf.
Also, curthread may or may not have anything to do with the I/O request
at hand.
The correct solution can either be to tag struct bio's with a
priority derived from the requesting threads nice and have disksort
act on this field, this wouldn't address the "silly-seek syndrome"
where two equal processes bang the diskheads from one edge to the
other of the disk repeatedly.
Alternatively, and probably better: a sleep should be introduced
either at the time the I/O is requested or at the time it is completed
where we can be sure to sleep in the right thread.
The sleep also needs to be in constant timeunits, 1/hz can be practicaly
any sub-second size, at high HZ the current code practically doesn't
do anything.
----------------------------
As suggested in this comment, it is no longer located in the disk sort
routine, but rather now resides in spec_strategy where the disk operations
are being queued by the thread that is associated with the process that
is really requesting the I/O. At that point, the disk queues are not
visible, so the I/O for positively niced processes is always slowed
down whether or not there is other activity on the disk.
On the issue of scaling HZ, I believe that the current scheme is
better than using a fixed quantum of time. As machines and I/O
subsystems get faster, the resolution on the clock also rises.
So, ten years from now we will be slowing things down for shorter
periods of time, but the proportional effect on the system will
be about the same as it is today. So, I view this as a feature
rather than a drawback. Hence this patch sticks with using HZ.
Sponsored by: DARPA & NAI Labs.
Reviewed by: Poul-Henning Kamp <phk@critter.freebsd.dk>
of a file in chunks that are less then the filesystem block size, if the
data is not already cached the system will perform a read-before-write.
The problem is that it does this on a block-by-block basis, breaking up the
I/Os and making clustering impossible for the writes. Programs such
as INN using cyclic file buffers suffer greatly. This problem is only going
to get worse as we use larger and larger filesystem block sizes.
The solution is to extend the sequential heuristic so UFS[2] can perform
a far larger read and readahead when dealing with this case.
(note: maximum disk write bandwidth is 27MB/sec thru filesystem)
(note: filesystem blocksize in test is 8K (1K frag))
dd if=/dev/zero of=test.dat bs=1k count=2m conv=notrunc
Before: (note half of these are reads)
tty da0 da1 acd0 cpu
tin tout KB/t tps MB/s KB/t tps MB/s KB/t tps MB/s us ni sy in id
0 76 14.21 598 8.30 0.00 0 0.00 0.00 0 0.00 0 0 7 1 92
0 76 14.09 813 11.19 0.00 0 0.00 0.00 0 0.00 0 0 9 5 86
0 76 14.28 821 11.45 0.00 0 0.00 0.00 0 0.00 0 0 8 1 91
After: (note half of these are reads)
tty da0 da1 acd0 cpu
tin tout KB/t tps MB/s KB/t tps MB/s KB/t tps MB/s us ni sy in id
0 76 63.62 434 26.99 0.00 0 0.00 0.00 0 0.00 0 0 18 1 80
0 76 63.58 424 26.30 0.00 0 0.00 0.00 0 0.00 0 0 17 2 82
0 76 63.82 438 27.32 0.00 0 0.00 0.00 0 0.00 1 0 19 2 79
Reviewed by: mckusick
Approved by: re
X-MFC after: immediately (was heavily tested in -stable for 4 months)
a common lock. This change avoids a deadlock between snapshots when
separate requests cause them to deadlock checking each other for a
need to copy blocks that are close enough together that they fall
into the same indirect block. Although I had anticipated a slowdown
from contention for the single lock, my filesystem benchmarks show
no measurable change in throughput on a uniprocessor system with
three active snapshots. I conjecture that this result is because
every copy-on-write fault must check all the active snapshots, so
the process was inherently serial already. This change removes the
last of the deadlocks of which I am aware in snapshots.
Sponsored by: DARPA & NAI Labs.
automatically set MNT_MULTILABEL in the mount flags.
If FS_ACLS is set in a UFS or UFS2 superblock, automatically
set MNT_ACLS in the mount flags.
If either of these flags is set, but the appropriate kernel option
to support the features associated with the flag isn't available,
then print a warning at mount-time.
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories
or fifo in UFS2, the normal ufs_strategy routine needs to be used
rather than the spec_strategy or fifo_strategy routine. Thus the
ffsext_strategy routine is interposed in the ffs_vnops vectors for
special devices and fifo's to pick off this special case. Otherwise
it simply falls through to the usual spec_strategy or fifo_strategy
routine.
Submitted by: Robert Watson <rwatson@FreeBSD.org>
Sponsored by: DARPA & NAI Labs.
FS_ACLS Administrative enable/disable of extended ACL support
FS_MULTILABEL Administrative flag to indicate to the MAC Framework
that objects in the file system are individually
labeled using extended attributes.
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories
Reviewed by: (in principal) mckusick, phk
that use it. Specifically, vop_stdlock uses the lock pointed to by
vp->v_vnlock. By default, getnewvnode sets up vp->v_vnlock to
reference vp->v_lock. Filesystems that wish to use the default
do not need to allocate a lock at the front of their node structure
(as some still did) or do a lockinit. They can simply start using
vn_lock/VOP_UNLOCK. Filesystems that wish to manage their own locks,
but still use the vop_stdlock functions (such as nullfs) can simply
replace vp->v_vnlock with a pointer to the lock that they wish to
have used for the vnode. Such filesystems are responsible for
setting the vp->v_vnlock back to the default in their vop_reclaim
routine (e.g., vp->v_vnlock = &vp->v_lock).
In theory, this set of changes cleans up the existing filesystem
lock interface and should have no function change to the existing
locking scheme.
Sponsored by: DARPA & NAI Labs.
Whenever doing a copy-on-write check, first look in the list of
initially allocated blocks to see if it is there. If so, no further
check is needed. If not, fall through and do the full check. This
change eliminates one of two known deadlocks caused by snapshots.
Handling the second deadlock will be the subject of another check-in.
This change also reduces the cost of the copy-on-write check by
speeding up the verification of frequently checked blocks.
Sponsored by: DARPA & NAI Labs.
even when the underlying device has a larger sector size. Therefore,
the filesystem code should not (and with this patch does not) try to
use the underlying sector size when doing disk block address calculations.
This patch fixes problems in -current when using the swap-based
memory-disk device (mdconfig -a -t swap ...). This bugfix is not
relevant to -stable as -stable does not have the memory-disk device.
Sponsored by: DARPA & NAI Labs.
v_tag is now const char * and should only be used for debugging.
Additionally:
1. All users of VT_NTS now check vfsconf->vf_type VFCF_NETWORK
2. The user of VT_PROCFS now checks for the new flag VV_PROCDEP, which
is propagated by pseudofs to all child vnodes if the fs sets PFS_PROCDEP.
Suggested by: phk
Reviewed by: bde, rwatson (earlier version)
layers deep in <sys/proc.h> or <sys/vnode.h>.
Include <sys/vmmeter.h> instead of depending on namespace pollution in
<sys/pcpu.h>.
Sorted includes as much as possible.
pmap_zero_page() and pmap_zero_page_area() were modified to accept
a struct vm_page * instead of a physical address, vm_page_zero_fill()
and vm_page_zero_fill_area() have served no purpose.
parts rather than use vop_{read,write}_args. Access to these
functions will ultimately not be available through the
"vop_{read,write}+IO_EXT" API but this functionality is retained
for debugging purposes for now.
Sponsored by: DARPA & NAI Labs.
UFS only thing, and FFS should in principle not know if it is enabled
or not.
This commit cleans ffs_vnops.c for such knowledge, but not ffs_vfsops.c
Sponsored by: DARPA and NAI Labs.
these in the main filesystems. This does not change the resulting code
but makes the source a little bit more grepable.
Sponsored by: DARPA and NAI Labs.
- v_vflag is protected by the vnode lock and is used when synchronization
with VOP calls is needed.
- v_iflag is protected by interlock and is used for dealing with vnode
management issues. These flags include X/O LOCK, FREE, DOOMED, etc.
- All accesses to v_iflag and v_vflag have either been locked or marked with
mp_fixme's.
- Many ASSERT_VOP_LOCKED calls have been added where the locking was not
clear.
- Many functions in vfs_subr.c were restructured to provide for stronger
locking.
Idea stolen from: BSD/OS
As this code is not actually used by any of the existing
interfaces, it seems unlikely to break anything (famous
last words).
The internal kernel interface to manipulate these attributes
is invoked using two new IO_ flags: IO_NORMAL and IO_EXT.
These flags may be specified in the ioflags word of VOP_READ,
VOP_WRITE, and VOP_TRUNCATE. Specifying IO_NORMAL means that
you want to do I/O to the normal data part of the file and
IO_EXT means that you want to do I/O to the extended attributes
part of the file. IO_NORMAL and IO_EXT are mutually exclusive
for VOP_READ and VOP_WRITE, but may be specified individually
or together in the case of VOP_TRUNCATE. For example, when
removing a file, VOP_TRUNCATE is called with both IO_NORMAL
and IO_EXT set. For backward compatibility, if neither IO_NORMAL
nor IO_EXT is set, then IO_NORMAL is assumed.
Note that the BA_ and IO_ flags have been `merged' so that they
may both be used in the same flags word. This merger is possible
by assigning the IO_ flags to the low sixteen bits and the BA_
flags the high sixteen bits. This works because the high sixteen
bits of the IO_ word is reserved for read-ahead and help with
write clustering so will never be used for flags. This merge
lets us get away from code of the form:
if (ioflags & IO_SYNC)
flags |= BA_SYNC;
For the future, I have considered adding a new field to the
vattr structure, va_extsize. This addition could then be
exported through the stat structure to allow applications to
find out the size of the extended attribute storage and also
would provide a more standard interface for truncating them
(via VOP_SETATTR rather than VOP_TRUNCATE).
I am also contemplating adding a pathconf parameter (for
concreteness, lets call it _PC_MAX_EXTSIZE) which would
let an application determine the maximum size of the extended
atribute storage.
Sponsored by: DARPA & NAI Labs.
direct calls for the two places where the kernel calls into soft
updates code. Set up the hooks in softdep_initialize() and NULL
them out in softdep_uninitialize(). This change allows soft updates
to function correctly when ufs is loaded as a module.
Reviewed by: mckusick
module. This adds an ffs_uninit() function that calls ufs_uninit()
and also calls a new softdep_uninitialize() function. Add a stub
for softdep_uninitialize() to cover the non-SOFTUPDATES case.
Reviewed by: mckusick
imposed by the filesystem structure itself remains. With 16k blocks,
the maximum file size is now just over 128TB.
For now, the UFS1 file size limit is left unchanged so as to remain
consistent with RELENG_4, but it too could be removed in the future.
Reviewed by: mckusick
out of inodes in a cylinder group would fail to check for
free inodes in other cylinder groups. This bug was introduced
in the UFS2 code merge two days ago.
An inode is allocated by calling ffs_valloc which calls
ffs_hashalloc to do the filesystem scan. Ffs_hashalloc
walks around the cylinder groups calling its passed allocator
(ffs_nodealloccg in this case) until the allocator returns a
non-zero result. The bug is that ffs_hashalloc expects the
passed allocator function to return a 64-bit ufs2_daddr_t.
When allocating inodes, it calls ffs_nodealloccg which was
returning a 32-bit ino_t. The ffs_hashalloc code checked
a 64-bit return value and usually found random non-zero bits in
the high 32-bits so decided that the allocation had succeeded
(in this case in the only cylinder group that it checked).
When the result was passed back to ffs_valloc it looked at
only the bottom 32-bits, saw zero and declared the system
out of inodes. But ffs_hashalloc had really only checked
one cylinder group.
The fix is to change ffs_nodealloccg to return 64-bit results.
Sponsored by: DARPA & NAI Labs.
Submitted by: Poul-Henning Kamp <phk@critter.freebsd.dk>
Reviewed by: Maxime Henrion <mux@freebsd.org>
filesystem expands the inode to 256 bytes to make space for 64-bit
block pointers. It also adds a file-creation time field, an ability
to use jumbo blocks per inode to allow extent like pointer density,
and space for extended attributes (up to twice the filesystem block
size worth of attributes, e.g., on a 16K filesystem, there is space
for 32K of attributes). UFS2 fully supports and runs existing UFS1
filesystems. New filesystems built using newfs can be built in either
UFS1 or UFS2 format using the -O option. In this commit UFS1 is
the default format, so if you want to build UFS2 format filesystems,
you must specify -O 2. This default will be changed to UFS2 when
UFS2 proves itself to be stable. In this commit the boot code for
reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c)
as there is insufficient space in the boot block. Once the size of the
boot block is increased, this code can be defined.
Things to note: the definition of SBSIZE has changed to SBLOCKSIZE.
The header file <ufs/ufs/dinode.h> must be included before
<ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and
ufs_lbn_t.
Still TODO:
Verify that the first level bootstraps work for all the architectures.
Convert the utility ffsinfo to understand UFS2 and test growfs.
Add support for the extended attribute storage. Update soft updates
to ensure integrity of extended attribute storage. Switch the
current extended attribute interfaces to use the extended attribute
storage. Add the extent like functionality (framework is there,
but is currently never used).
Sponsored by: DARPA & NAI Labs.
Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
vnode creation globaly, we allow processes to create vnodes concurently.
In case of concurent creation of vnode for the one ino, we allow processes
to race and then check who wins.
Assuming that concurent creation of vnode for same ino is really rare case,
this is belived to be an improvement, as it just allows concurent creation
of vnodes.
Idea by: bp
Reviewed by: dillon
MFC after: 1 month
most cases NULL is passed, but in some cases such as network driver locks
(which use the MTX_NETWORK_LOCK macro) and UMA zone locks, a name is used.
Tested on: i386, alpha, sparc64
general cleanup of the API. The entire API now consists of two functions
similar to the pre-KSE API. The suser() function takes a thread pointer
as its only argument. The td_ucred member of this thread must be valid
so the only valid thread pointers are curthread and a few kernel threads
such as thread0. The suser_cred() function takes a pointer to a struct
ucred as its first argument and an integer flag as its second argument.
The flag is currently only used for the PRISON_ROOT flag.
Discussed on: smp@
provided the latter is nonzero. At this point, the former is a fairly
arbitrary default value (DFTPHYS), so changing it to any reasonable
value specified by the device driver is safe. Using the maximum of
these limits broke ffs clustered i/o for devices whose si_iosize_max
is < DFLTPHYS. Using the minimum would break device drivers' ability
to increase the active limit from DFTLPHYS up to MAXPHYS.
Copied the code for this and the associated (unnecessary?) fixup of
mp_iosize_max to all other filesystems that use clustering (ext2fs and
msdosfs). It was completely missing.
PR: 36309
MFC-after: 1 week
locking flags when acquiring a vnode. The immediate purpose is
to allow polling lock requests (LK_NOWAIT) needed by soft updates
to avoid deadlock when enlisting other processes to help with
the background cleanup. For the future it will allow the use of
shared locks for read access to vnodes. This change touches a
lot of files as it affects most filesystems within the system.
It has been well tested on FFS, loopback, and CD-ROM filesystems.
only lightly on the others, so if you find a problem there, please
let me (mckusick@mckusick.com) know.
the bio and buffer structures to have daddr64_t bio_pblkno,
b_blkno, and b_lblkno fields which allows access to disks
larger than a Terabyte in size. This change also requires
that the VOP_BMAP vnode operation accept and return daddr64_t
blocks. This delta should not affect system operation in
any way. It merely sets up the necessary interfaces to allow
the development of disk drivers that work with these larger
disk block addresses. It also allows for the development of
UFS2 which will use 64-bit block addresses.
read-only.
The trouble here is that we don't reopen the device in read/write mode
when we remount in read/write mode resulting in a filesystem sending
write requests to a device which was only opened read/only.
I'm not quite sure how such a reopen would best be done and defer
the problem to more agile hackers.
and isn't strictly required. However, it lowers the number of false
positives found when grep'ing the kernel sources for p_ucred to ensure
proper locking.
inode'' panic. This change corrects that problem by setting the
fs_active flag when the inode map changes to notify the snapshot
code that the cylinder group must be rescanned.
Submitted by: Robert Watson <rwatson@FreeBSD.org>
without being reclaimed. This bug was introduced in revision 1.95
dealing with filenames placed in newly allocated directory blocks,
thus is not present in 4.X systems. The bug is triggered when a
new entry is made in a directory after the data block containing
the original new entry has been written, but before the inode
that references the data block has been written.
Submitted by: Bill Fenner <fenner@research.att.com>
been unlinked (e.g., with a zero link count). We have to expunge
all trace of these files from the snapshot so that they are neither
reclaimed prematurely by fsck nor saved unnecessarily by dump.
which small and/or nearly full filesystems would fail with `file
system full' messages when trying to replace a number of existing
files (for example during a system installation). When the allocation
routines are about to fail with a file system full condition, they
make a call to softdep_request_cleanup() which attempts to accelerate
the flushing of pending deletion requests in an effort to free up
space. In the face of filesystem I/O requests that exceed the
available disk transfer capacity, the cleanup request could take
an unbounded amount of time. Thus, the softdep_request_cleanup()
routine will only try for tickdelay seconds (default 2 seconds)
before giving up and returning a filesystem full error. Under typical
conditions, the softdep_request_cleanup() routine is able to free
up space in under fifty milliseconds.
which caused incomplete snapshots to be taken. When background
fsck would run on these snapshots, the result would be files
being incorrectly released which would subsequently panic the
kernel with ``handle_workitem_freefile: inodedep survived'',
``handle_written_inodeblock: live inodedep'', and
``handle_workitem_remove: lost inodedep'' errors.
involving file removal or file update were not always being fully
committed to disk. The result was lost files or corrupted file data.
This change ensures that the filesystem is properly synced to disk
before the filesystem is down-graded.
This delta also fixes a long standing bug in which a file open for
reading has been unlinked. When the last open reference to the file
is closed, the inode is reclaimed by the filesystem. Previously,
if the filesystem had been down-graded to read-only, the inode could
not be reclaimed, and thus was lost and had to be later recovered
by fsck. With this change, such files are found at the time of the
down-grade. Normally they will result in the filesystem down-grade
failing with `device busy'. If a forcible down-grade is done, then
the affected files will be revoked causing the inode to be released
and the open file descriptors to begin failing on attempts to read.
Submitted by: "Sam Leffler" <sam@errno.com>
Seigo Tanimura (tanimura) posted the initial delta.
I've polished it quite a bit reducing the need for locking and
adapting it for KSE.
Locks:
1 mutex in each filedesc
protects all the fields.
protects "struct file" initialization, while a struct file
is being changed from &badfileops -> &pipeops or something
the filedesc should be locked.
1 mutex in each struct file
protects the refcount fields.
doesn't protect anything else.
the flags used for garbage collection have been moved to
f_gcflag which was the FILLER short, this doesn't need
locking because the garbage collection is a single threaded
container.
could likely be made to use a pool mutex.
1 sx lock for the global filelist.
struct file * fhold(struct file *fp);
/* increments reference count on a file */
struct file * fhold_locked(struct file *fp);
/* like fhold but expects file to locked */
struct file * ffind_hold(struct thread *, int fd);
/* finds the struct file in thread, adds one reference and
returns it unlocked */
struct file * ffind_lock(struct thread *, int fd);
/* ffind_hold, but returns file locked */
I still have to smp-safe the fget cruft, I'll get to that asap.
lost if some other process uses the lock while we are sleeping. We
restore it after we have slept. This functionality is provided by
a new routine interlocked_sleep() that wraps the interlocking with
functions that sleep. This function is then used in place of the
old ACQUIRE_LOCK_INTERLOCKED() and FREE_LOCK_INTERLOCKED() macros.
Submitted by: Debbie Chu <dchu@juniper.net>
in softdep_sync_metadata(). Otherwise we may miss dependencies
that need to be flushed which will result in a later panic
with the message ``vinvalbuf: dirty bufs''.
Submitted by: Matthew Dillon <dillon@apollo.backplane.com>
MFC after: 1 week
against VM_WAIT in the pageout code. Both fixes involve adjusting
the lockmgr's timeout capability so locks obtained with timeouts do not
interfere with locks obtained without a timeout.
Hopefully MFC: before the 4.5 release
superblock that is already set up to handle pointer types. This
fixes an accidental change in the superblock size on 64-bit platforms
caused by revision 1.24.
