UMA_ZFLAG_INTERNAL zones at all. Apparently, Wilko's alpha
was crashing while entering multi-user because, I think, we
were calculating the garbage cachefree for pcpu caches that
essentially don't exist for at least the 'zones' zone and it so
happened that we were reading from an unmapped location.
Confirmed to fix crash: wilko
Helped debug: wilko, gallatin
compare the zone element size (+1 for the byte of linkage) against
UMA_SLAB_SIZE - sizeof(struct uma_slab), and not just UMA_SLAB_SIZE.
Add a KASSERT in zone_small_init to make sure that the computed
ipers (items per slab) for the zone is not zero, despite the addition
of the check, just to be sure (this part submitted by: silby)
- UMA_ZONE_VM used to imply BUCKETCACHE. Now it implies
CACHEONLY instead. CACHEONLY is like BUCKETCACHE in the
case of bucket allocations, but in addition to that also ensures that
we don't setup the zone with OFFPAGE slab headers allocated from the
slabzone. This means that we're not allowed to have a UMA_ZONE_VM
zone initialized for large items (zone_large_init) because it would
require the slab headers to be allocated from slabzone, and hence
kmem_map. Some of the zones init'd with UMA_ZONE_VM are so init'd
before kmem_map is suballoc'd from kernel_map, which is why this
change is necessary.
- All those diffs to syscalls.master for each architecture *are*
necessary. This needed clarification; the stub code generation for
mlockall() was disabled, which would prevent applications from
linking to this API (suggested by mux)
- Giant has been quoshed. It is no longer held by the code, as
the required locking has been pushed down within vm_map.c.
- Callers must specify VM_MAP_WIRE_HOLESOK or VM_MAP_WIRE_NOHOLES
to express their intention explicitly.
- Inspected at the vmstat, top and vm pager sysctl stats level.
Paging-in activity is occurring correctly, using a test harness.
- The RES size for a process may appear to be greater than its SIZE.
This is believed to be due to mappings of the same shared library
page being wired twice. Further exploration is needed.
- Believed to back out of allocations and locks correctly
(tested with WITNESS, MUTEX_PROFILING, INVARIANTS and DIAGNOSTIC).
PR: kern/43426, standards/54223
Reviewed by: jake, alc
Approved by: jake (mentor)
MFC after: 2 weeks
From alc:
Move pageable pipe memory to a seperate kernel submap to avoid awkward
vm map interlocking issues. (Bad explanation provided by me.)
From me:
Rework pipespace accounting code to handle this new layout, and adjust
our default values to account for the fact that we now have a solid
limit on allocations.
Also, remove the "maxpipes" limit, as it no longer has a purpose.
(The limit on kva usage solves the problem of having two many pipes.)
Eliminate a lot of checkes to make sure requests are not cross-device
which is unnecessary with the new layout. We know a sequential request
cannot possibly be cross-device because there is a reserved page between
the devices.
Remove a couple of comments which no longer are relevant.
to not get any cross-device I/O requests. (The unallocated first page
protecting BSD labels already gave us this, but that hack may go away
at some point in time).
Remove the check for cross-device I/O requests in swap_pager_strategy.
Move the repeated statistics updating into flushchainbuf().
swapbkva. Swapbkva mappings are explicitly managed using pmap_qenter(),
not on-demand by vm_fault(), making kmem_alloc_nofault() more appropriate.
Submitted by: tegge
Use ->bio_children to count child buffers, rather than abuse the
bio_caller1 pointer.
Expand the relevant bits of waitchainbuf() inline, this clarifies
the code a little bit.
striping to a per device round-robin algorithm.
Because of the policy of not attempting to retain previous swap
allocation on page-out, this means that a newly added swap device
almost instantly takes its 1/N share of the I/O load but it takes
somewhat longer for it to assume it's 1/N share of the pages if there
is plenty of space on the other devices.
Change the 8G total swapspace limitation to 8G per device instead
by using a per device blist rather than one global blist. This
reduces the memory footprint by 75% (typically a couple hundred
kilobytes) for the common case with one swapdevice but NSWAPDEV=4.
Remove the compile time constant limit of number of swap devices,
there is no limit now. Instead of a fixed size array, store the
per swapdev structure in a TAILQ.
Total swap space is still addressed by a 32 bit page number and
therefore the upper limit is now 2^42 bytes = 16TB (for i386).
We still do not allocate the first page of each device in order to
give some amount of protection to any bsdlabel at the start of the
device.
A new device is appended after the existing devices in the swap space,
no attempt is made to fill in holes left behind by swapoff (this can
trivially be changed should it ever become a problem).
The sysctl vm.nswapdev now reflects the number of currently configured
swap devices.
Rename vm_swap_size to swap_pager_avail for consistency with other
exported names.
Change argument type for vm_proc_swapin_all() and swap_pager_isswapped()
to be a struct swdevt pointer rather than an index.
Not changed: we are still using blists to manage the free space,
but since the swapspace is no longer fragmented by the striping
different resource managers might fare better.
concurrent invocations from acquiring the same address(es). Also, in case
of an incomplete allocation, free any allocated pages.
In collaboration with: tegge
sure that uma_dbg_free() is called if we're about to call
uma_zfree_internal() but we're asking it to skip the dtor and
uma_dbg_free() call itself. So, if we're about to call
uma_zfree_internal() from uma_zfree_arg() and skip == 1, call
uma_dbg_free() ourselves.
in sync with the backend machdep code. When cpu_thread_init() does not
have the same idea of KSTACK_PAGES as the thing that created the kstack,
all hell breaks loose.
Bad alc! no cookie! :-)
1) The race has to do with zone destruction. From the zone destructor we
would lock the zone, set the working set size to 0, then unlock the zone,
drain it, and then free the structure. Within the window following the
working-set-size set to 0 and unlocking of the zone and the point where
in zone_drain we re-acquire the zone lock, the uma timer routine could
have fired off and changed the working set size to something non-zero,
thereby potentially preventing us from completely freeing slabs before
destroying the zone (and thus leaking them).
2) The leak has to do with zone destruction as well. When destroying a
zone we would take care to free all the buckets cached in the zone, but
although we would drain the pcpu cache buckets, we would not free them.
This resulted in leaking a couple of bucket structures (512 bytes each)
per cpu on SMP during zone destruction.
While I'm here, also silence GCC warnings by turning uma_slab_alloc()
from inline to real function. It's too big to be an inline.
Reviewed by: JeffR
("UMA Zone") carefully, because it does not have pcpu caches allocated
at all. In the UP case, we did not catch this because one pcpu cache
is always allocated with the zone, but for the MP case, we were getting
bogus stats for this zone.
Tested by: Lukas Ertl <le@univie.ac.at>
- In sysctl_vm_zone use the per cpu locks to read the current cache
statistics this makes them more accurate while under heavy load.
Submitted by: tegge
so not only wastes memory but it can also cause a leak in zones that
will be destroyed later. The problem is that the slab allocation code
places newly created slabs on the partially allocated list because it
assumes that the caller will actually allocate some memory from it.
Failure to do so places an otherwise free slab on the partial slab list
where we wont find it later in zone_drain().
Continuously prodded to fix by: phk (Thanks)
contain the filedescriptor number on opens from userland.
The index is used rather than a "struct file *" since it conveys a bit
more information, which may be useful to in particular fdescfs and /dev/fd/*
For now pass -1 all over the place.
uma_zalloc) is called with exactly one of either M_WAITOK or M_NOWAIT and
that it is called with neither M_TRYWAIT or M_DONTWAIT. Print a warning
if anything is wrong. Default to M_WAITOK of no flag is given. This is the
same test as in malloc(9).
the "toss the largest process" emergency handling) from vm_map.c to
swap_pager.c.
The quantity calculated depends strongly on the internals of the
swap_pager and by moving it, we no longer need to expose the
internal metrics of the swap_pager to the world.
paging space and how much of it is in use (in pages).
Use this interface from the Linuxolator instead of groping around in the
internals of the swap_pager.
is not natural and needlessly exposes a lot of dirty laundry.
Move private interfaces between the two from swap_pager.h to swap_pager.c
and staticize as much as possible.
No functional change.
files, so that SWAP_META_PAGES does not vary either.
swap_pager.c ended up with a value of 16, everybody else 8. Go with
the 16 for now.
This should only have any effect in the "kill processes because we
are out of swap" scenario, where it will make some sort of estimate
of something more precise.
- Change vm_pageout_object_deactivate_pages()'s first parameter from a
vm_map_t to a pmap_t.
- Change vm_pageout_object_deactivate_pages()'s and
vm_pageout_map_deactivate_pages()'s last parameter from a vm_pindex_t
to a long. Since the number of pages in an address space doesn't
require 64 bits on an i386, vm_pindex_t is overkill.
order to avoid the overhead of later page faults. In general, it
implements two cases: one for vnode-backed objects and one for
device-backed objects. Only the device-backed case is really
machine-dependent, belonging in the pmap.
This commit moves the vnode-backed case into the (relatively) new
function vm_map_pmap_enter(). On amd64 and i386, this commit only
amounts to code rearrangement. On alpha and ia64, the new machine
independent (MI) implementation of the vnode case is smaller and more
efficient than their pmap-based implementations. (The MI
implementation takes advantage of the fact that objects in -CURRENT
are ordered collections of pages.) On sparc64, pmap_object_init_pt()
hadn't (yet) been implemented.
implementation of a largely MI pmap_object_init_pt() for vnode-backed
objects. pmap_enter_quick() is implemented via pmap_enter() on sparc64
and powerpc.
- Correct a mismatch between pmap_object_init_pt()'s prototype and its
various implementations. (I plan to keep pmap_object_init_pt() as
the MD hook for device-backed objects on i386 and amd64.)
- Correct an error in ia64's pmap_enter_quick() and adjust its interface
to match the other versions. Discussed with: marcel
color in vm_page_alloc(). (This also has small performance benefits.)
- Eliminate vm_page_select_free(); vm_page_alloc() might as well
call vm_pageq_find() directly.
of pcpu locks. This makes uma_zone somewhat smaller (by (LOCKNAME_LEN *
sizeof(char) + sizeof(struct mtx) * maxcpu) bytes, to be exact).
No Objections from jeff.
releasing the lock only if we are about to sleep (e.g., vm_pager_get_pages()
or vm_pager_has_pages()). If we sleep, we have marked the vm object with
the paging-in-progress flag.
Several of the subtypes have an associated vnode which is used for
stuff like the f*() functions.
By giving the vnode a speparate field, a number of checks for the specific
subtype can be replaced simply with a check for f_vnode != NULL, and
we can later free f_data up to subtype specific use.
At this point in time, f_data still points to the vnode, so any code I
might have overlooked will still work.
to the machine-independent parts of the VM. At the same time, this
introduces vm object locking for the non-i386 platforms.
Two details:
1. KSTACK_GUARD has been removed in favor of KSTACK_GUARD_PAGES. The
different machine-dependent implementations used various combinations
of KSTACK_GUARD and KSTACK_GUARD_PAGES. To disable guard page, set
KSTACK_GUARD_PAGES to 0.
2. Remove the (unnecessary) clearing of PG_ZERO in vm_thread_new. In
5.x, (but not 4.x,) PG_ZERO can only be set if VM_ALLOC_ZERO is passed
to vm_page_alloc() or vm_page_grab().
processes in the first pass. Among other things, this will give
us a chance to launder vnode-backed pages before concluding that
we need more swap. This is particularly useful for systems that
have no swap.
While here, update a comment and remove some long-unused code.
Reported by: Lucky Green <shamrock@cypherpunks.to>
Suggested by: dillon
Approved by: re (rwatson)
- Move struct sigacts out of the u-area and malloc() it using the
M_SUBPROC malloc bucket.
- Add a small sigacts_*() API for managing sigacts structures: sigacts_alloc(),
sigacts_free(), sigacts_copy(), sigacts_share(), and sigacts_shared().
- Remove the p_sigignore, p_sigacts, and p_sigcatch macros.
- Add a mutex to struct sigacts that protects all the members of the struct.
- Add sigacts locking.
- Remove Giant from nosys(), kill(), killpg(), and kern_sigaction() now
that sigacts is locked.
- Several in-kernel functions such as psignal(), tdsignal(), trapsignal(),
and thread_stopped() are now MP safe.
Reviewed by: arch@
Approved by: re (rwatson)
trustworthy for vnode-backed objects.
- Restore the old behavior of vm_object_page_remove() when the end
of the given range is zero. Add a comment to vm_object_page_remove()
regarding this behavior.
Reported by: iedowse
- Eliminate an odd, special-case feature:
if start == end == 0 then all pages are removed. Only one caller
used this feature and that caller can trivially pass the object's
size.
- Assert that the vm_object is locked on entry; don't bother testing
for a NULL vm_object.
- Style: Fix lines that are longer than 80 characters.
fork1() and never changes.
- The proc lock is enough to cover reading p_state, so push down sched_lock
into the PRS_NORMAL case of the switch on p_state.
- Remove the Giant required from vm_page_free_toq(). (Any locking
errors will be caught by vm_page_remove().)
This remedies a panic that occurred when kmem_malloc(NOWAIT) performed
without Giant failed to allocate the necessary pages.
Reported by: phk
- Add a parameter to vm_pageout_flush() that tells vm_pageout_flush()
whether its caller has locked the vm_object. (This is a temporary
measure to bootstrap vm_object locking.)
race where a thread could assume that a process was swapped in by
PHOLD() when it actually wasn't fully swapped in yet.
- In faultin(), always msleep() if PS_SWAPPINGIN is set instead of doing
this check after bumping p_lock in the PS_INMEM == 0 case. Also,
sched_lock is only needed for setting and clearning swapping PS_*
flags and the swap thread inhibitor.
- Don't set and clear the thread swap inhibitor in the same loops as the
pmap_swapin/out_thread() since we have to do it under sched_lock.
Instead, mimic the treatment of the PS_INMEM flag and use separate loops
to set the inhibitors when clearing PS_INMEM and clear the inhibitors
when setting PS_INMEM.
- swapout() now returns with the proc lock held as it holds the lock
while adjusting the swapping-related PS_* flags so that the proc lock
can be used to test those flags.
- Only use the proc lock to check the swapping-related PS_* flags in
several places.
- faultin() no longer requires sched_lock to be held by callers.
- Rename PS_SWAPPING to PS_SWAPPINGOUT to be less ambiguous now that we
have PS_SWAPPINGIN.
called without Giant; and obj_alloc() in turn calls vm_page_alloc()
without Giant. This causes an assertion failure in vm_page_alloc().
Fortunately, obj_alloc() is now MPSAFE. So, we need only clean up
some assertions.
- Weaken the assertion in vm_page_lookup() to require Giant only
if the vm_object isn't locked.
- Remove an assertion from vm_page_alloc() that duplicates a check
performed in vm_page_lookup().
In collaboration with: gallatin, jake, jeff
vm_object_pip_add() and vm_object_pip_wakeup().
- Remove GIANT_REQUIRED from vm_object_pip_subtract() and
vm_object_pip_subtract().
- Lock the vm_object when performing vm_object_page_remove().
critical and should not be killed when pageout is looking for more
memory pages in all the wrong places.
Reviewed by: arch@
Sponsored by: St. Bernard Software
