where the cleaner is trying to write, instead of tying up the "live"
buffers (or pages).
Fix a bug in the LFS_UBC case where oversized buffers would not be
checksummed correctly, causing uncleanable segments.
Make sure that wakeup(fs->lfs_iocount) is done if fs->lfs_iocount is 1
as well as 0, since we wait in some places for it to drop to 1.
Activate all pages that make it into lfs_gop_write without the segment
lock held, since they must have been dirtied very recently, even if
PG_DELWRI is not set.
(there are still some details to work out) but expect that to go
away soon. To support these basic changes (creation of lfs_putpages,
lfs_gop_write, mods to lfs_balloc) several other changes were made, to
wit:
* Create a writer daemon kernel thread whose purpose is to handle page
writes for the pagedaemon, but which also takes over some of the
functions of lfs_check(). This thread is started the first time an
LFS is mounted.
* Add a "flags" parameter to GOP_SIZE. Current values are
GOP_SIZE_READ, meaning that the call should return the size of the
in-core version of the file, and GOP_SIZE_WRITE, meaning that it
should return the on-disk size. One of GOP_SIZE_READ or
GOP_SIZE_WRITE must be specified.
* Instead of using malloc(...M_WAITOK) for everything, reserve enough
resources to get by and use malloc(...M_NOWAIT), using the reserves if
necessary. Use the pool subsystem for structures small enough that
this is feasible. This also obsoletes LFS_THROTTLE.
And a few that are not strictly necessary:
* Moves the LFS inode extensions off onto a separately allocated
structure; getting closer to LFS as an LKM. "Welcome to 1.6O."
* Unified GOP_ALLOC between FFS and LFS.
* Update LFS copyright headers to correct values.
* Actually cast to unsigned in lfs_shellsort, like the comment says.
* Keep track of which segments were empty before the previous
checkpoint; any segments that pass two checkpoints both dirty and
empty can be summarily cleaned. Do this. Right now lfs_segclean
still works, but this should be turned into an effectless
compatibility syscall.
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?
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.
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.
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.
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.
- remove special treatment of pager_map mappings in pmaps. this is
required now, since I've removed the globals that expose the address range.
pager_map now uses pmap_kenter_pa() instead of pmap_enter(), so there's
no longer any need to special-case it.
- eliminate struct uvm_vnode by moving its fields into struct vnode.
- rewrite the pageout path. the pager is now responsible for handling the
high-level requests instead of only getting control after a bunch of work
has already been done on its behalf. this will allow us to UBCify LFS,
which needs tighter control over its pages than other filesystems do.
writing a page to disk no longer requires making it read-only, which
allows us to write wired pages without causing all kinds of havoc.
- use a new PG_PAGEOUT flag to indicate that a page should be freed
on behalf of the pagedaemon when it's unlocked. this flag is very similar
to PG_RELEASED, but unlike PG_RELEASED, PG_PAGEOUT can be cleared if the
pageout fails due to eg. an indirect-block buffer being locked.
this allows us to remove the "version" field from struct vm_page,
and together with shrinking "loan_count" from 32 bits to 16,
struct vm_page is now 4 bytes smaller.
- no longer use PG_RELEASED for swap-backed pages. if the page is busy
because it's being paged out, we can't release the swap slot to be
reallocated until that write is complete, but unlike with vnodes we
don't keep a count of in-progress writes so there's no good way to
know when the write is done. instead, when we need to free a busy
swap-backed page, just sleep until we can get it busy ourselves.
- implement a fast-path for extending writes which allows us to avoid
zeroing new pages. this substantially reduces cpu usage.
- encapsulate the data used by the genfs code in a struct genfs_node,
which must be the first element of the filesystem-specific vnode data
for filesystems which use genfs_{get,put}pages().
- eliminate many of the UVM pagerops, since they aren't needed anymore
now that the pager "put" operation is a higher-level operation.
- enhance the genfs code to allow NFS to use the genfs_{get,put}pages
instead of a modified copy.
- clean up struct vnode by removing all the fields that used to be used by
the vfs_cluster.c code (which we don't use anymore with UBC).
- remove kmem_object and mb_object since they were useless.
instead of allocating pages to these objects, we now just allocate
pages with no object. such pages are mapped in the kernel until they
are freed, so we can use the mapping to find the page to free it.
this allows us to remove splvm() protection in several places.
The sum of all these changes improves write throughput on my
decstation 5000/200 to within 1% of the rate of NetBSD 1.5
and reduces the elapsed time for "make release" of a NetBSD 1.5
source tree on my 128MB pc to 10% less than a 1.5 kernel took.
Kernels and tools understand both v1 and v2 filesystems; newfs_lfs
generates v2 by default. Changes for the v2 layout include:
- Segments of non-PO2 size and arbitrary block offset, so these can be
matched to convenient physical characteristics of the partition (e.g.,
stripe or track size and offset).
- Address by fragment instead of by disk sector, paving the way for
non-512-byte-sector devices. In theory fragments can be as large
as you like, though in reality they must be smaller than MAXBSIZE in size.
- Use serial number and filesystem identifier to ensure that roll-forward
doesn't get old data and think it's new. Roll-forward is enabled for
v2 filesystems, though not for v1 filesystems by default.
- The inode free list is now a tailq, paving the way for undelete (undelete
is not yet implemented, but can be without further non-backwards-compatible
changes to disk structures).
- Inode atime information is kept in the Ifile, instead of on the inode;
that is, the inode is never written *just* because atime was changed.
Because of this the inodes remain near the file data on the disk, rather
than wandering all over as the disk is read repeatedly. This speeds up
repeated reads by a small but noticeable amount.
Other changes of note include:
- The ifile written by newfs_lfs can now be of arbitrary length, it is no
longer restricted to a single indirect block.
- Fixed an old bug where ctime was changed every time a vnode was created.
I need to look more closely to make sure that the times are only updated
during write(2) and friends, not after-the-fact during a segment write,
and certainly not by the cleaner.
on mount, through the newer checkpoint and on through any newer
partial-segments that may have been written but not checkpointed because
of an intervening crash.
LFS_DO_ROLLFORWARD is not defined by default.
in an error case in lfs_markv. Change the vfs_getvfs() error to return
ENOENT, for consistency with failure of vfs_busy().
99% of this patch was from Jesse Off <joff@gci-net.com> (PR #11547).
(PR #11468). In the case of fragment allocation, check to see if enough
space is available before extending a fragment already scheduled for writing.
The locked_queue_* variables indicate the number of buffer headers and bytes,
respectively, that are unavailable to getnewbuf() because they are locked up
waiting for LFS to flush them; make sure that that is actually what we're
counting, i.e., never count malloced buffers, and always use b_bufsize instead
of b_bcount.
If DEBUG is defined, the periodic calls to lfs_countlocked will now complain
if either counter is incorrect. (In the future lfs_countlocked will not need
to be called at all if DEBUG is not defined.)
Kernel:
* Add runtime quantity lfs_ravail, the number of disk-blocks reserved
for writing. Writes to the filesystem first reserve a maximum amount
of blocks before their write is allowed to proceed; after the blocks
are allocated the reserved total is reduced by a corresponding amount.
If the lfs_reserve function cannot immediately reserve the requested
number of blocks, the inode is unlocked, and the thread sleeps until
the cleaner has made enough space available for the blocks to be
reserved. In this way large files can be written to the filesystem
(or, smaller files can be written to a nearly-full but thoroughly
clean filesystem) and the cleaner can still function properly.
* Remove explicit switching on dlfs_minfreeseg from the kernel code; it
is now merely a fs-creation parameter used to compute dlfs_avail and
dlfs_bfree (and used by fsck_lfs(8) to check their accuracy). Its
former role is better assumed by a properly computed dlfs_avail.
* Bounds-check inode numbers submitted through lfs_bmapv and lfs_markv.
This prevents a panic, but, if the cleaner is feeding the filesystem
the wrong data, you are still in a world of hurt.
* Cleanup: remove explicit references of DEV_BSIZE in favor of
btodb()/dbtob().
lfs_cleanerd:
* Make -n mean "send N segments' blocks through a single call to
lfs_markv". Previously it had meant "clean N segments though N calls
to lfs_markv, before looking again to see if more need to be cleaned".
The new behavior gives better packing of direct data on disk with as
little metadata as possible, largely alleviating the problem that the
cleaner can consume more disk through inefficient use of metadata than
it frees by moving dirty data away from clean "holes" to produce
entirely clean segments.
* Make -b mean "read as many segments as necessary to write N segments
of dirty data back to disk", rather than its former meaning of "read
as many segments as necessary to free N segments worth of space". The
new meaning, combined with the new -n behavior described above,
further aids in cleaning storage efficiency as entire segments can be
written at once, using as few blocks as possible for segment summaries
and inode blocks.
* Make the cleaner take note of segments which could not be cleaned due
to error, and not attempt to clean them until they are entirely free
of dirty blocks. This prevents the case in which a cleanerd running
with -n 1 and without -b (formerly the default) would spin trying
repeatedly to clean a corrupt segment, while the remaining space
filled and deadlocked the filesystem.
* Update the lfs_cleanerd manual page to describe all the options,
including the changes mentioned here (in particular, the -b and -n
flags were previously undocumented).
fsck_lfs:
* Check, and optionally fix, lfs_avail (to an exact figure) and
lfs_bfree (within a margin of error) in pass 5.
newfs_lfs:
* Reduce the default dlfs_minfreeseg to 1/20 of the total segments.
* Add a warning if the sgs disklabel field is 16 (the default for FFS'
cpg, but not usually desirable for LFS' sgs: 5--8 is a better range).
* Change the calculation of lfs_avail and lfs_bfree, corresponding to
the kernel changes mentioned above.
mount_lfs:
* Add -N and -b options to pass corresponding -n and -b options to
lfs_cleanerd.
* Default to calling lfs_cleanerd with "-b -n 4".
[All of these changes were largely tested in the 1.5 branch, with the
idea that they (along with previous un-pulled-up work) could be applied
to the branch while it was still in ALPHA2; however my test system has
experienced corruption on another filesystem (/dev/console has gone
missing :^), and, while I believe this unrelated to the LFS changes, I
cannot with good conscience request that the changes be pulled up.]
Make lfs_uinodes a signed quantity for debugging purposes, and set it to
zero as fs mount time.
Enclose setting/clearing of the dirty flags (IN_MODIFIED, IN_ACCESSED,
IN_CLEANING) in macros, and use those macros everywhere. Make
LFS_ITIMES use these macros; updated the ITIMES macro in inode.h to know
about this. Make ufs_getattr use ITIMES instead of FFS_ITIMES.
fixes:
- Write copies of bfree and avail in the CLEANERINFO block, so the
cleaner doesn't have to guess which superblock has the current
information (if indeed any do).
- Tighten up accounting of lfs_avail (more needs to be done).
- When cleansing indirect blocks of UNWRITTEN, make sure not to mark
them clean, since they'll need to be rewritten later.
Change the space computation to appear to change the size of the *disk*
rather than the *bytes used* when more segment summaries and inode
blocks are written. Try to estimate the amount of space that these will
take up when more files are written, so the disk size doesn't change too
much.
Regularize error returns from lfs_valloc, lfs_balloc, lfs_truncate: they
now fail entirely, rather than succeeding half-way and leaving the fs in
an inconsistent state.
Rewrite lfs_truncate, mostly stealing from ffs_truncate. The old
lfs_truncate had difficulty truncating a large file to a non-zero size
(indirect blocks were not handled appropriately).
Unmark VDIROP on fvp after ufs_remove, ufs_rmdir, so these can be
reclaimed immediately: this vnode would not be written to disk again
anyway if the removal succeeded, and if it failed, no directory
operation occurred.
ufs_makeinode and ufs_mkdir now remove IN_ADIROP on error.
references (locked for VOP_INACTIVE at the end of vrele) and it's okay.
Check the return value of lfs_vref where appropriate.
Fixes PR #s 10285 and 10352.
1.4 branch.
* Use a separate per-fs lock, instead of ufs_hashlock, to protect the Inode
free list. This seems to prevent the "lockmgr: %d, not exclusive lock holder
%d, unlocking" message I was mis-attributing last night to an unlocked vnode
being passed to vrele.
* Change calling semantics of lfs_ifind, to give better error reporting:
If fed a struct buf, it can report the block number of the offending inode
block as well as the inode number.
* Back out rev 1.10 of lfs_subr.c, since the replacement code was slightly
uglier while being functionally identical.
* Make lfs_vunref use the same free list convention as vrele/vput, so that
vget does not remove vnodes from a hash list they are not on.
This prevents a rare condition in which Ifile "ifile" blocks, that is, the
blocks of the ifile which point VOP_VGET at the inode block containing the
requested inode, from being "unwritten" when cleaning during intense disk
activity.
system crashed, inodes could be allocated that were not referenced. (Though
not a serious problem, it evidences itself in phase 4 of fsck_lfs.) Fix
this by marking if_daddr with UNASSIGNED before the inodes are actually
written; at mount time the ifile is checked for UNASSIGNED entries and
any that are found are linked back into the free list. (The latter
functionality should move into the roll-forward agent when it materializes.)
conditions. Also change the default setting of lfs_clean_vnhead to 0, which
seems to make the locking problems go away (although this is difficult to
test as I can't reliably reproduce them).
if we are short on vnodes, lfs_vflush from another process can grab a
vnode that lfs_markv has already processed but not yet written; but
lfs_markv holds the seglock. When lfs_vflush gets around to writing it,
the context for copyin is gone. So, now lfs_markv calls copyin itself,
rather than having lfs_writeseg do it.
include:
- DIROP segregation is enabled, and greater care is taken
to make sure that a checkpoint completes. Fsck is not
needed to remount the filesystem.
- Several checks to make sure that the LFS subsystem does not
overuse various resources (memory, in particular).
- The cleaner routines, lfs_markv in particular, are completely
rewritten. A buffer overflow is removed. Greater care is taken
to ensure that inodes come from where lfs_cleanerd say they come
from (so we know nothing has changed since lfs_bmapv was called).
- Fragment allocation is fixed, so that writes beyond end-of-file
do the right thing.
- added an "union inode_ext" to struct inode, for the per-fs extentions.
For now only ext2fs uses it.
- i_din is now an union:
union {
struct dinode ffs_din; /* 128 bytes of the on-disk dinode. */
struct ext2fs_dinode e2fs_din; /* 128 bytes of the on-disk dinode. */
} i_din
Added a lot of #define i_ffs_* and i_e2fs_* to access the fields.
- Added two macros: FFS_ITIMES and EXT2FS_ITIMES. ITIMES calls the rigth
macro, depending on the time of the inode. ITIMES is used where necessary,
FFS_ITIMES and EXT2FS_ITIMES in other places.
perseant