actually in the kernel. This structure is a different size than
what is currently in -CURRENT, but should hopefully be the last time
any application breakage is caused there. As soon as any major
inconveniences are removed, the definition of the in-kernel struct
ucred should be conditionalized upon defined(_KERNEL).
This also changes struct export_args to remove dependency on the
constantly-changing struct ucred, as well as limiting the bounds
of the size fields to the correct size. This means: a) mountd and
friends won't break all the time, b) mountd and friends won't crash
the kernel all the time if they don't know what they're doing wrt
actual struct export_args layout.
Reviewed by: bde
mtx_enter(lock, type) becomes:
mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks)
mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized)
similarily, for releasing a lock, we now have:
mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN.
We change the caller interface for the two different types of locks
because the semantics are entirely different for each case, and this
makes it explicitly clear and, at the same time, it rids us of the
extra `type' argument.
The enter->lock and exit->unlock change has been made with the idea
that we're "locking data" and not "entering locked code" in mind.
Further, remove all additional "flags" previously passed to the
lock acquire/release routines with the exception of two:
MTX_QUIET and MTX_NOSWITCH
The functionality of these flags is preserved and they can be passed
to the lock/unlock routines by calling the corresponding wrappers:
mtx_{lock, unlock}_flags(lock, flag(s)) and
mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN
locks, respectively.
Re-inline some lock acq/rel code; in the sleep lock case, we only
inline the _obtain_lock()s in order to ensure that the inlined code
fits into a cache line. In the spin lock case, we inline recursion and
actually only perform a function call if we need to spin. This change
has been made with the idea that we generally tend to avoid spin locks
and that also the spin locks that we do have and are heavily used
(i.e. sched_lock) do recurse, and therefore in an effort to reduce
function call overhead for some architectures (such as alpha), we
inline recursion for this case.
Create a new malloc type for the witness code and retire from using
the M_DEV type. The new type is called M_WITNESS and is only declared
if WITNESS is enabled.
Begin cleaning up some machdep/mutex.h code - specifically updated the
"optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN
and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently
need those.
Finally, caught up to the interface changes in all sys code.
Contributors: jake, jhb, jasone (in no particular order)
calls returning EACCES instead of EPERM. This patch modifies vaccess()
to return EPERM instead of EACCES if VADMIN is among the requested
rights. This affects functions normally limited to the owners of
a file, such as chmod(), as EPERM is the error indicating that
privilege would allow the operation, rather than a chance in mandatory
or discretionary rights.
Reported by: bde
and numvnodes are longs in the kernel. They should remain longs in systat,
what really needs to change is that they should be using SYSCTL_LONG rather
than SYSCTL_INT. I also changed wantfreevnodes to SYSCTL_LONG because I
happened to notice it.
I wish there was a way to find all of these automatically..
Pointed out by: bde
Removed most of the hacks that were trying to deal with low-memory
situations prior to now.
The new code is based on the concept that I/O must be able to function in
a low memory situation. All major modules related to I/O (except
networking) have been adjusted to allow allocation out of the system
reserve memory pool. These modules now detect a low memory situation but
rather then block they instead continue to operate, then return resources
to the memory pool instead of cache them or leave them wired.
Code has been added to stall in a low-memory situation prior to a vnode
being locked.
Thus situations where a process blocks in a low-memory condition while
holding a locked vnode have been reduced to near nothing. Not only will
I/O continue to operate, but many prior deadlock conditions simply no
longer exist.
Implement a number of VFS/BIO fixes
(found by Ian): in biodone(), bogus-page replacement code, the loop
was not properly incrementing loop variables prior to a continue
statement. We do not believe this code can be hit anyway but we
aren't taking any chances. We'll turn the whole section into a
panic (as it already is in brelse()) after the release is rolled.
In biodone(), the foff calculation was incorrectly
clamped to the iosize, causing the wrong foff to be calculated
for pages in the case of an I/O error or biodone() called without
initiating I/O. The problem always caused a panic before. Now it
doesn't. The problem is mainly an issue with NFS.
Fixed casts for ~PAGE_MASK. This code worked properly before only
because the calculations use signed arithmatic. Better to properly
extend PAGE_MASK first before inverting it for the 64 bit masking
op.
In brelse(), the bogus_page fixup code was improperly throwing
away the original contents of 'm' when it did the j-loop to
fix the bogus pages. The result was that it would potentially
invalidate parts of the *WRONG* page(!), leading to corruption.
There may still be cases where a background bitmap write is
being duplicated, causing potential corruption. We have identified
a potentially serious bug related to this but the fix is still TBD.
So instead this patch contains a KASSERT to detect the problem
and panic the machine rather then continue to corrupt the filesystem.
The problem does not occur very often.. it is very hard to
reproduce, and it may or may not be the cause of the corruption
people have reported.
Review by: (VFS/BIO: mckusick, Ian Dowse <iedowse@maths.tcd.ie>)
Testing by: (VM/Deadlock) Paul Saab <ps@yahoo-inc.com>
getnewvnode(). Otherwise routines called from VOP_INACTIVE() might
attempt to remove the vnode from a free list the vnode isn't on,
causing corruption.
PR: 18012
"administrative" authorization checks. In most cases, the VADMIN test
checks to make sure the credential effective uid is the same as the file
owner.
o Modify vaccess() to set VADMIN as an available right if the uid is
appropriate.
o Modify references to uid-based access control operations such that they
now always invoke VOP_ACCESS() instead of using hard-coded policy checks.
o This allows alternative UFS policies to be implemented by replacing only
ufs_access() (such as mandatory system policies).
o VOP_ACCESS() requires the caller to hold an exclusive vnode lock on the
vnode: I believe that new invocations of VOP_ACCESS() are always called
with the lock held.
o Some direct checks of the uid remain, largely associated with the QUOTA
and SUIDDIR code.
Reviewed by: eivind
Obtained from: TrustedBSD Project
Add lockdestroy() and appropriate invocations, which corresponds to
lockinit() and must be called to clean up after a lockmgr lock is no
longer needed.
separately (nfs, cd9660 etc) or keept as a first element of structure
referenced by v_data pointer(ffs). Such organization leads to known problems
with stacked filesystems.
From this point vop_no*lock*() functions maintain only interlock lock.
vop_std*lock*() functions maintain built-in v_lock structure using lockmgr().
vop_sharedlock() is compatible with vop_stdunlock(), but maintains a shared
lock on vnode.
If filesystem wishes to export lockmgr compatible lock, it can put an address
of this lock to v_vnlock field. This indicates that the upper filesystem
can take advantage of it and use single lock structure for entire (or part)
of stack of vnodes. This field shouldn't be examined or modified by VFS code
except for initialization purposes.
Reviewed in general by: mckusick
* Add lots of comments
* Convert a couple of assertions to KASSERT()
* Minimal whitespace & misapplied {} fixes
* Convert #if 0 to #if COMPILING_LINT for code we presently do not
support, but want to keep available.
Reviewed by: adrian, markm
which hides the 'hole' in the minor bits.
Introduce unit2minor() to do the reverse operation.
Fix some some make_dev() calls which didn't use UID_* or GID_* macros.
Kill the v_hashchain alias macro, it hides the real relationship.
Introduce experimental SI_CHEAPCLONE flag set it on cloned bpfs.
include:
* Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The
alpha port is still in transition and currently uses both.)
* Per-CPU idle processes.
* Interrupts are run in their own separate kernel threads and can be
preempted (i386 only).
Partially contributed by: BSDi (BSD/OS)
Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
object before falling back on privilege. Make vaccess() accept an
additional optional argument, privused, to determine whether
privilege was required for vaccess() to return 0. Add commented
out capability checks for reference. Rename some variables to make
it more clear which modes/uids/etc are associated with the object,
and which with the access mode.
o Update file system use of vaccess() to pass NULL as the optional
privused argument. Once additional patches are applied, suser()
will no longer set ASU, so privused will permit passing of
privilege information up the stack to the caller.
Reviewed by: bde, green, phk, -security, others
Obtained from: TrustedBSD Project
with the new snapshot code.
Update addaliasu to correctly implement the semantics of the old
checkalias function. When a device vnode first comes into existence,
check to see if an anonymous vnode for the same device was created
at boot time by bdevvp(). If so, adopt the bdevvp vnode rather than
creating a new vnode for the device. This corrects a problem which
caused the kernel to panic when taking a snapshot of the root
filesystem.
Change the calling convention of vn_write_suspend_wait() to be the
same as vn_start_write().
Split out softdep_flushworklist() from softdep_flushfiles() so that
it can be used to clear the work queue when suspending filesystem
operations.
Access to buffers becomes recursive so that snapshots can recursively
traverse their indirect blocks using ffs_copyonwrite() when checking
for the need for copy on write when flushing one of their own indirect
blocks. This eliminates a deadlock between the syncer daemon and a
process taking a snapshot.
Ensure that softdep_process_worklist() can never block because of a
snapshot being taken. This eliminates a problem with buffer starvation.
Cleanup change in ffs_sync() which did not synchronously wait when
MNT_WAIT was specified. The result was an unclean filesystem panic
when doing forcible unmount with heavy filesystem I/O in progress.
Return a zero'ed block when reading a block that was not in use at
the time that a snapshot was taken. Normally, these blocks should
never be read. However, the readahead code will occationally read
them which can cause unexpected behavior.
Clean up the debugging code that ensures that no blocks be written
on a filesystem while it is suspended. Snapshots must explicitly
label the blocks that they are writing during the suspension so that
they do not cause a `write on suspended filesystem' panic.
Reorganize ffs_copyonwrite() to eliminate a deadlock and also to
prevent a race condition that would permit the same block to be
copied twice. This change eliminates an unexpected soft updates
inconsistency in fsck caused by the double allocation.
Use bqrelse rather than brelse for buffers that will be needed
soon again by the snapshot code. This improves snapshot performance.
the gating of system calls that cause modifications to the underlying
filesystem. The gating can be enabled by any filesystem that needs
to consistently suspend operations by adding the vop_stdgetwritemount
to their set of vnops. Once gating is enabled, the function
vfs_write_suspend stops all new write operations to a filesystem,
allows any filesystem modifying system calls already in progress
to complete, then sync's the filesystem to disk and returns. The
function vfs_write_resume allows the suspended write operations to
begin again. Gating is not added by default for all filesystems as
for SMP systems it adds two extra locks to such critical kernel
paths as the write system call. Thus, gating should only be added
as needed.
Details on the use and current status of snapshots in FFS can be
found in /sys/ufs/ffs/README.snapshot so for brevity and timelyness
is not included here. Unless and until you create a snapshot file,
these changes should have no effect on your system (famous last words).
This function will probably rewritten/renamed to devpp.
Submitted by: Assar Westerlund <assar@sics.se> on -current
Confirmed to work: Steinar Haug <sthaug@nethelp.no>,
Manfred Antar <mantar@pacbell.net>
Reviewed by: phk
<sys/bio.h>.
<sys/bio.h> is now a prerequisite for <sys/buf.h> but it shall
not be made a nested include according to bdes teachings on the
subject of nested includes.
Diskdrivers and similar stuff below specfs::strategy() should no
longer need to include <sys/buf.> unless they need caching of data.
Still a few bogus uses of struct buf to track down.
Repocopy by: peter
substitute BUF_WRITE(foo) for VOP_BWRITE(foo->b_vp, foo)
substitute BUF_STRATEGY(foo) for VOP_STRATEGY(foo->b_vp, foo)
This patch is machine generated except for the ccd.c and buf.h parts.
return ENXIO (Device not configured). Without this, vn_isdisk()
could (and did in the case of lstat() under fdesc) pass a NULL pointer
to devsw(), which caused a page fault.
Reviewed by: alfred
to recycle full fsids after only 16 mount/unmount's. This is probably
too often for exported fsids. Now we recycle the full fsids only
after 2^16 mount/ umount's and only ensure uniqueness in the lower 16
bits if there have been <= 256 calls to vfs_getnewfsid() since the
system started.
number was packed very wastefully, giving perfect non-uniqeness in
the lower 16 bits of fsids for filesystems with the same vfs type.
This made linux_stat() return perfectly non-unique (broken) 16-bit
st_dev's for nfs mount points, and effectively reduced mntid_base to
8 bits so that the vfs_getnewfsid() looped endlessly when there are
already 256 mounted filesystems with the required vfs type.
Approved by: jkh
the codepath is followed.
From the PR:
vclean calls vrele leading to deadlock (if usecount > 0)
vclean() calls vrele() if v_usecount of the node was higher than one.
But before calling it, it sets the VXLOCK flag, which will make
vn_lock called from vrele dead-lock.
PR: kern/15117
Submitted by: Assar Westerlund <assar@stacken.kth.se>
Reviewed by: rwatson
Obtained from: NetBSD
it only on the buf_daemon process). The problem is that when the
syncer process starts running the worklist, it wants to delete
lots of files. It does this by VFS_VGET'ing the vnodes, clearing
the blocks in them and bdwrite'ing the buffer. It can process close
to a thousand files per second which generates a large number of
dirty buffers. So, giving it special priviledge at the buffer trough
leads to trouble as the buf_daemon does occationally need a free
buffer to proceed and if the syncer has used every last one up,
we are toast.
into vnode dirtyblkhd we append it to the list instead of prepend it to
the list in order to maintain a 'forward' locality of reference, which
is arguably better then 'reverse'. The original algorithm did things this
way to but at a huge time cost.
Enhance the append interlock for NFS writes to handle intr/soft mounts
better.
Fix the hysteresis for NFS async daemon I/O requests to reduce the
number of unnecessary context switches.
Modify handling of NFS mount options. Any given user option that is
too high now defaults to the kernel maximum for that option rather then
the kernel default for that option.
Reviewed by: Alfred Perlstein <bright@wintelcom.net>
madvise().
This feature prevents the update daemon from gratuitously flushing
dirty pages associated with a mapped file-backed region of memory. The
system pager will still page the memory as necessary and the VM system
will still be fully coherent with the filesystem. Modifications made
by other means to the same area of memory, for example by write(), are
unaffected. The feature works on a page-granularity basis.
MAP_NOSYNC allows one to use mmap() to share memory between processes
without incuring any significant filesystem overhead, putting it in
the same performance category as SysV Shared memory and anonymous memory.
Reviewed by: julian, alc, dg
* lockstatus() and VOP_ISLOCKED() gets a new process argument and a new
return value: LK_EXCLOTHER, when the lock is held exclusively by another
process.
* The ASSERT_VOP_(UN)LOCKED family is extended to use what this gives them
* Extend the vnode_if.src format to allow more exact specification than
locked/unlocked.
This commit should not do any semantic changes unless you are using
DEBUG_VFS_LOCKS.
Discussed with: grog, mch, peter, phk
Reviewed by: peter
Alot of the code in sys/kern directly accesses the *Q_HEAD and *Q_ENTRY
structures for list operations. This patch makes all list operations
in sys/kern use the queue(3) macros, rather than directly accessing the
*Q_{HEAD,ENTRY} structures.
Reviewed by: phk
Submitted by: Jake Burkholder <jake@checker.org>
PR: 14914
Correctly lock vnodes when calling VOP_OPEN() from filesystem mount code.
Unify spec_open() for bdev and cdev cases.
Remove the disabled bdev specific read/write code.
Merge the contents (less some trivial bordering the silly comments)
of <vm/vm_prot.h> and <vm/vm_inherit.h> into <vm/vm.h>. This puts
the #defines for the vm_inherit_t and vm_prot_t types next to their
typedefs.
This paves the road for the commit to follow shortly: change
useracc() to use VM_PROT_{READ|WRITE} rather than B_{READ|WRITE}
as argument.
two new functions spec_buf{read|write}.
Add sysctl vfs.bdev_buffered which defaults to 1 == true. This
sysctl can be used to experimentally turn buffered behaviour for
bdevs off. I should not be changed while any blockdevices are
open. Remove the misplaced sysctl vfs.enable_userblk_io.
No other changes in behaviour.
clustering issues (replacing code that used to be in
ufs/ufs/ufs_readwrite.c). vm_fault also now uses the new VM page counter
inlines.
This completes the changeover from vnode->v_lastr to vm_entry_t->v_lastr
for VM, and fp->f_nextread and fp->f_seqcount (which have been in the
tree for a while). Determination of the I/O strategy (sequential, random,
and so forth) is now handled on a descriptor-by-descriptor basis for
base I/O calls, and on a memory-region-by-memory-region and
process-by-process basis for VM faults.
Reviewed by: David Greenman <dg@root.com>, Alan Cox <alc@cs.rice.edu>
addaliasu() into addalias() (no operational change) and clarify comments
relating to a trick that vclean() uses.
The fix to BOOTP is yet another hack. Actually, rootfsid handling
is already a major hack. The whole thing needs to be cleaned up.
Reviewed by: David Greenman <dg@root.com>, Alan Cox <alc@cs.rice.edu>
to buffered block devices are allowed. The default is to be backwards
compatible, i.e. reads and writes are allowed.
The idea is for a larger crowd to start running with this disabled and
see what problems, if any, crop up, and then to change the default to
off and see if any problems crop up in the next 6 months prior to
potentially removing support entirely. There are still a few people,
Julian and myself included, who believe the buffered block device
access from usermode to be useful.
Remove use of vnode->v_lastr from buffered block device I/O in
preparation for removal of vnode->v_lastr field, replacing it with
the already existing seqcount metric to detect sequential operation.
Reviewed by: Alan Cox <alc@cs.rice.edu>, David Greenman <dg@root.com>
Make the alias list a SLIST.
Drop the "fast recycling" optimization of vnodes (including
the returning of a prexisting but stale vnode from checkalias).
It doesn't buy us anything now that we don't hardlimit
vnodes anymore.
Rename checkalias2() and checkalias() to addalias() and
addaliasu() - which takes dev_t and udev_t arg respectively.
Make the revoke syscalls use vcount() instead of VALIASED.
Remove VALIASED flag, we don't need it now and it is faster
to traverse the much shorter lists than to maintain the
flag.
vfs_mountedon() can check the dev_t directly, all the vnodes
point to the same one.
Print the devicename in specfs/vprint().
Remove a couple of stale LFS vnode flags.
Remove unimplemented/unused LK_DRAINED;
In lookup() however it's the other way around as we need to supply the
dev_t for the vnode, so devfs still has a copy of it stashed away.
Sourcing it from the vnode in the vnops however is useful as it makes
a lot of the code almost the same as that in specfs.
have been maintained, and that is still the default. A new sysctl
variable "vfs.timestamp_precision" can be used to enable higher
levels of precision:
0 = seconds only; nanoseconds zeroed (default).
1 = seconds and nanoseconds, accurate within 1/HZ.
2 = seconds and nanoseconds, truncated to microseconds.
>=3 = seconds and nanoseconds, maximum precision.
Level 1 uses getnanotime(), which is fast but can be wrong by up
to 1/HZ. Level 2 uses microtime(). It might be desirable for
consistency with utimes() and friends, which take timeval structures
rather than timespecs. Level 3 uses nanotime() for the higest
precision.
I benchmarked levels 0, 1, and 3 by copying a 550 MB tree with
"cpio -pdu". There was almost negligible difference in the system
times -- much less than 1%, and less than the variation among
multiple runs at the same level. Bruce Evans dreamed up a torture
test involving 1-byte reads with intervening fstat() calls, but
the cpio test seems more realistic to me.
This feature is currently implemented only for the UFS (FFS and
MFS) filesystems. But I think it should be easy to support it in
the others as well.
An earlier version of this was reviewed by Bruce. He's not to
blame for any breakage I've introduced since then.
Reviewed by: bde (an earlier version of the code)
vnodes referencing this device.
Details:
cdevsw->d_parms has been removed, the specinfo is available
now (== dev_t) and the driver should modify it directly
when applicable, and the only driver doing so, does so:
vn.c. I am not sure the logic in checking for "<" was right
before, and it looks even less so now.
An intial pool of 50 struct specinfo are depleted during
early boot, after that malloc had better work. It is
likely that fewer than 50 would do.
Hashing is done from udev_t to dev_t with a prime number
remainder hash, experiments show no better hash available
for decent cost (MD5 is only marginally better) The prime
number used should not be close to a power of two, we use
83 for now.
Add new checkalias2() to get around the loss of info from
dev2udev() in bdevvp();
The aliased vnodes are hung on a list straight of the dev_t,
and speclisth[SPECSZ] is unused. The sharing of struct
specinfo means that the v_specnext moves into the vnode
which grows by 4 bytes.
Don't use a VBLK dev_t which doesn't make sense in MFS, now
we hang a dummy cdevsw on B/Cmaj 253 so that things look sane.
Storage overhead from all of this is O(50k).
Bump __FreeBSD_version to 400009
The next step will add the stuff needed so device-drivers can start to
hang things from struct specinfo
Only know casualy of this is swapinfo/pstat which should be fixes
the right way: Store the actual pathname in the kernel like mount
does. [Volounteers sought for this task]
The road map from here is roughly: expand struct specinfo into struct
based dev_t. Add dev_t registration facilities for device drivers and
start to use them.
large (1G) memory machine configurations. I was able to run 'dbench 32'
on a 32MB system without bring the machine to a grinding halt.
* buffer cache hash table now dynamically allocated. This will
have no effect on memory consumption for smaller systems and
will help scale the buffer cache for larger systems.
* minor enhancement to pmap_clearbit(). I noticed that
all the calls to it used constant arguments. Making
it an inline allows the constants to propogate to
deeper inlines and should produce better code.
* removal of inherent vfs_ioopt support through the emplacement
of appropriate #ifdef's, with John's permission. If we do not
find a use for it by the end of the year we will remove it entirely.
* removal of getnewbufloops* counters & sysctl's - no longer
necessary for debugging, getnewbuf() is now optimal.
* buffer hash table functions removed from sys/buf.h and localized
to vfs_bio.c
* VFS_BIO_NEED_DIRTYFLUSH flag and support code added
( bwillwrite() ), allowing processes to block when too many dirty
buffers are present in the system.
* removal of a softdep test in bdwrite() that is no longer necessary
now that bdwrite() no longer attempts to flush dirty buffers.
* slight optimization added to bqrelse() - there is no reason
to test for available buffer space on B_DELWRI buffers.
* addition of reverse-scanning code to vfs_bio_awrite().
vfs_bio_awrite() will attempt to locate clusterable areas
in both the forward and reverse direction relative to the
offset of the buffer passed to it. This will probably not
make much of a difference now, but I believe we will start
to rely on it heavily in the future if we decide to shift
some of the burden of the clustering closer to the actual
I/O initiation.
* Removal of the newbufcnt and lastnewbuf counters that Kirk
added. They do not fix any race conditions that haven't already
been fixed by the gbincore() test done after the only call
to getnewbuf(). getnewbuf() is a static, so there is no chance
of it being misused by other modules. ( Unless Kirk can think
of a specific thing that this code fixes. I went through it
very carefully and didn't see anything ).
* removal of VOP_ISLOCKED() check in flushbufqueues(). I do not
think this check is necessary, the buffer should flush properly
whether the vnode is locked or not. ( yes? ).
* removal of extra arguments passed to getnewbuf() that are not
necessary.
* missed cluster_wbuild() that had to be a cluster_wbuild_wb() in
vfs_cluster.c
* vn_write() now calls bwillwrite() *PRIOR* to locking the vnode,
which should greatly aid flushing operations in heavy load
situations - both the pageout and update daemons will be able
to operate more efficiently.
* removal of b_usecount. We may add it back in later but for now
it is useless. Prior implementations of the buffer cache never
had enough buffers for it to be useful, and current implementations
which make more buffers available might not benefit relative to
the amount of sophistication required to implement a b_usecount.
Straight LRU should work just as well, especially when most things
are VMIO backed. I expect that (even though John will not like
this assumption) directories will become VMIO backed some point soon.
Submitted by: Matthew Dillon <dillon@backplane.com>
Reviewed by: Kirk McKusick <mckusick@mckusick.com>
QUEUE_AGE, QUEUE_LRU, and QUEUE_EMPTY we instead have QUEUE_CLEAN,
QUEUE_DIRTY, QUEUE_EMPTY, and QUEUE_EMPTYKVA. With this patch clean
and dirty buffers have been separated. Empty buffers with KVM
assignments have been separated from truely empty buffers. getnewbuf()
has been rewritten and now operates in a 100% optimal fashion. That is,
it is able to find precisely the right kind of buffer it needs to
allocate a new buffer, defragment KVM, or to free-up an existing buffer
when the buffer cache is full (which is a steady-state situation for
the buffer cache).
Buffer flushing has been reorganized. Previously buffers were flushed
in the context of whatever process hit the conditions forcing buffer
flushing to occur. This resulted in processes blocking on conditions
unrelated to what they were doing. This also resulted in inappropriate
VFS stacking chains due to multiple processes getting stuck trying to
flush dirty buffers or due to a single process getting into a situation
where it might attempt to flush buffers recursively - a situation that
was only partially fixed in prior commits. We have added a new daemon
called the buf_daemon which is responsible for flushing dirty buffers
when the number of dirty buffers exceeds the vfs.hidirtybuffers limit.
This daemon attempts to dynamically adjust the rate at which dirty buffers
are flushed such that getnewbuf() calls (almost) never block.
The number of nbufs and amount of buffer space is now scaled past the
8MB limit that was previously imposed for systems with over 64MB of
memory, and the vfs.{lo,hi}dirtybuffers limits have been relaxed
somewhat. The number of physical buffers has been increased with the
intention that we will manage physical I/O differently in the future.
reassignbuf previously attempted to keep the dirtyblkhd list sorted which
could result in non-deterministic operation under certain conditions,
such as when a large number of dirty buffers are being managed. This
algorithm has been changed. reassignbuf now keeps buffers locally sorted
if it can do so cheaply, and otherwise gives up and adds buffers to
the head of the dirtyblkhd list. The new algorithm is deterministic but
not perfect. The new algorithm greatly reduces problems that previously
occured when write_behind was turned off in the system.
The P_FLSINPROG proc->p_flag bit has been replaced by the more descriptive
P_BUFEXHAUST bit. This bit allows processes working with filesystem
buffers to use available emergency reserves. Normal processes do not set
this bit and are not allowed to dig into emergency reserves. The purpose
of this bit is to avoid low-memory deadlocks.
A small race condition was fixed in getpbuf() in vm/vm_pager.c.
Submitted by: Matthew Dillon <dillon@apollo.backplane.com>
Reviewed by: Kirk McKusick <mckusick@mckusick.com>
