Booting an amd64 kernel on Firecracker with 1 CPU and 128 MB of RAM,
SYSINIT cpu takes roughly 2770 us:
* 2280 us in vm_ksubmap_init
* 535 us in kmem_malloc
* 450 us in pmap_zero_page
* 1720 us in pmap_growkernel
* 1620 us in pmap_zero_page
* 80 us in bufinit
* 480 us in cpu_setregs
* 430 us in cpu_setregs calling load_cr0
Much of this is hypervisor overhead: load_cr0 is slow because it traps
to the hypervisor, and 99% of the time in pmap_zero_page is spent when
we first touch the page, presumably due to the host Linux kernel
faulting in backing pages one by one.
Sponsored by: https://www.patreon.com/cperciva
Differential Revision: https://reviews.freebsd.org/D40327
The wait flag is passed to UMA when allocating boundary tags for the
initial span, and UMA expects either M_WAITOK or M_NOWAIT to be present.
Reported by: cperciva
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
uma_zalloc_arg expects exactly one of the two WAIT flags. A future
commit will assert this.
Reviewed by: rstone
MFC after: 1 month
Sponsored by: Dell EMC Isilon
Differential Revision: https://reviews.freebsd.org/D34450
Replace MAXPHYS by runtime variable maxphys. It is initialized from
MAXPHYS by default, but can be also adjusted with the tunable kern.maxphys.
Make b_pages[] array in struct buf flexible. Size b_pages[] for buffer
cache buffers exactly to atop(maxbcachebuf) (currently it is sized to
atop(MAXPHYS)), and b_pages[] for pbufs is sized to atop(maxphys) + 1.
The +1 for pbufs allow several pbuf consumers, among them vmapbuf(),
to use unaligned buffers still sized to maxphys, esp. when such
buffers come from userspace (*). Overall, we save significant amount
of otherwise wasted memory in b_pages[] for buffer cache buffers,
while bumping MAXPHYS to desired high value.
Eliminate all direct uses of the MAXPHYS constant in kernel and driver
sources, except a place which initialize maxphys. Some random (and
arguably weird) uses of MAXPHYS, e.g. in linuxolator, are converted
straight. Some drivers, which use MAXPHYS to size embeded structures,
get private MAXPHYS-like constant; their convertion is out of scope
for this work.
Changes to cam/, dev/ahci, dev/ata, dev/mpr, dev/mpt, dev/mvs,
dev/siis, where either submitted by, or based on changes by mav.
Suggested by: mav (*)
Reviewed by: imp, mav, imp, mckusick, scottl (intermediate versions)
Tested by: pho
Sponsored by: The FreeBSD Foundation
Differential revision: https://reviews.freebsd.org/D27225
Today, the zone is only used to allocate a trio of kernel maps: the
kernel map itself, and the exec and pipe submaps. Maps for user
processes are dynamically allocated but are embedded in the vmspace
structure, which is allocated from its own zone. Make the
aforementioned kernel maps statically allocated and get rid of the zone.
While here, remove a stale comment above vmspace_alloc() and change the
names of locks initialized in vm_map_init() to match vmspace_zinit().
Reported by: alc
Reviewed by: alc, kib
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26052
ordering to allocate early pages in the same way boot pages were but only
as needed. After the KVA allocator has started up we allocate the KVA that
we consumed during boot. This also makes the boot pages freeable since they
have vm_page structures allocated with the rest of memory.
Parts of this patch were written and tested by markj.
Reviewed by: glebius, markj
Differential Revision: https://reviews.freebsd.org/D23102
- In uma_prealloc(), we need to check for an empty domain before the
first allocation attempt, not after. Fix this by switching
uma_prealloc() to use a vm_domainset iterator, which addresses the
secondary issue of using a signed domain identifier in round-robin
iteration.
- Don't automatically create a page daemon for domain 0.
- In domainset_empty_vm(), recompute ds_cnt and ds_order after
excluding empty domains; otherwise we may frequently specify an empty
domain when calling in to the page allocator, wasting CPU time.
Convert DOMAINSET_PREF() policies for empty domains to round-robin.
- When freeing bootstrap pages, don't count them towards the per-domain
total page counts for now: some vm_phys segments are created before
the SRAT is parsed and are thus always identified as being in domain 0
even when they are not. Then, when bootstrap pages are freed, they
are added to a domain that we had previously thought was empty. Until
this is corrected, we simply exclude them from the per-domain page
count.
Reported and tested by: Rajesh Kumar <rajfbsd@gmail.com>
Reviewed by: gallatin
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17704
This keeps the initialization coupled together with the kmem_* KPI
implementation, which is the main user of these arenas.
No functional change intended.
Reviewed by: alc
Approved by: re (gjb)
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17247
Exposing max_offset and min_offset defines in public headers is
causing clashes with variable names, for example when building QEMU.
Based on the submission by: royger
Reviewed by: alc, markj (previous version)
Sponsored by: The FreeBSD Foundation (kib)
MFC after: 1 week
Approved by: re (marius)
Differential revision: https://reviews.freebsd.org/D16881
became unused in FreeBSD 12.x as a side-effect of the NUMA-related
changes.)
Reviewed by: kib, markj
Discussed with: jeff, re@
Differential Revision: https://reviews.freebsd.org/D16825
Most kernel memory that is allocated after boot does not need to be
executable. There are a few exceptions. For example, kernel modules
do need executable memory, but they don't use UMA or malloc(9). The
BPF JIT compiler also needs executable memory and did use malloc(9)
until r317072.
(Note that a side effect of r316767 was that the "small allocation"
path in UMA on amd64 already returned non-executable memory. This
meant that some calls to malloc(9) or the UMA zone(9) allocator could
return executable memory, while others could return non-executable
memory. This change makes the behavior consistent.)
This change makes malloc(9) return non-executable memory unless the new
M_EXEC flag is specified. After this change, the UMA zone(9) allocator
will always return non-executable memory, and a KASSERT will catch
attempts to use the M_EXEC flag to allocate executable memory using
uma_zalloc() or its variants.
Allocations that do need executable memory have various choices. They
may use the M_EXEC flag to malloc(9), or they may use a different VM
interfact to obtain executable pages.
Now that malloc(9) again allows executable allocations, this change also
reverts most of r317072.
PR: 228927
Reviewed by: alc, kib, markj, jhb (previous version)
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D15691
size of UMA zone allocation is greater than page size. In this case zone
of zones can not use UMA_MD_SMALL_ALLOC, and we need to postpone switch
off of this zone from startup_alloc() until full launch of VM.
o Always supply number of VM zones to uma_startup_count(). On machines
with UMA_MD_SMALL_ALLOC ignore it completely, unless zsize goes over
a page. In the latter case account VM zones for number of allocations
from the zone of zones.
o Rewrite startup_alloc() so that it will immediately switch off from
itself any zone that is already capable of running real alloc.
In worst case scenario we may leak a single page here. See comment
in uma_startup_count().
o Hardcode call to uma_startup2() into vm_mem_init(). Otherwise some
extra SYSINITs, e.g. vm_page_init() may sneak in before.
o While here, remove uma_boot_pages_mtx. With recent changes to boot
pages calculation, we are guaranteed to use all of the boot_pages
in the early single threaded stage.
Reported & tested by: mav
global to per-domain state. Protect reservations with the free lock
from the domain that they belong to. Refactor to make vm domains more
of a first class object.
Reviewed by: markj, kib, gallatin
Tested by: pho
Sponsored by: Netflix, Dell/EMC Isilon
Differential Revision: https://reviews.freebsd.org/D14000
o Call uma_startup1() after initializing kmem, vmem and domains.
o Include 8 eight VM startup pages into uma_startup_count() calculation.
o Account for vmem_startup() and vm_map_startup() preallocating pages.
o Account for extra two allocations done by kmem_init() and vmem_create().
o Hardcode the place of execution of vm_radix_reserve_kva(). Using SYSINIT
allowed several other SYSINITs to sneak in before it, thus bumping
requirement for amount of boot pages.
reservations by giving each memory domain its own KVA space in vmem that
is naturally aligned on superpage boundaries.
Reviewed by: alc, markj, kib (some objections)
Sponsored by: Netflix, Dell/EMC Isilon
Tested by; pho
Differential Revision: https://reviews.freebsd.org/D13289
Mainly focus on files that use BSD 3-Clause license.
The Software Package Data Exchange (SPDX) group provides a specification
to make it easier for automated tools to detect and summarize well known
opensource licenses. We are gradually adopting the specification, noting
that the tags are considered only advisory and do not, in any way,
superceed or replace the license texts.
Special thanks to Wind River for providing access to "The Duke of
Highlander" tool: an older (2014) run over FreeBSD tree was useful as a
starting point.
Renumber cluase 4 to 3, per what everybody else did when BSD granted
them permission to remove clause 3. My insistance on keeping the same
numbering for legal reasons is too pedantic, so give up on that point.
Submitted by: Jan Schaumann <jschauma@stevens.edu>
Pull Request: https://github.com/freebsd/freebsd/pull/96
Upon each execve, we allocate a KVA range for use in copying data to the
new image. Pages must be faulted into the range, and when the range is
freed, the backing pages are freed and their mappings are destroyed. This
is a lot of needless overhead, and the exec_map management becomes a
bottleneck when many CPUs are executing execve concurrently. Moreover, the
number of available ranges is fixed at 16, which is insufficient on large
systems and potentially excessive on 32-bit systems.
The new allocator reduces overhead by making exec_map allocations
persistent. When a range is freed, pages backing the range are marked clean
and made easy to reclaim. With this change, the exec_map is sized based on
the number of CPUs.
Reviewed by: kib
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D8921
On machines with just the wrong amount of physical memory (enough to
have a lot of bufs, but not enough to use VM_FREELIST_DMA32) it is
possible for 32-bit address limited devices to have little to no
memory left when attaching, due to potentially large vfs bio configs
consuming all memory below 4GB not protected by VM_FREELIST_ISADMA.
This causes the 32-bit devices to allocate from VM_FREELIST_ISADMA,
leaving that freelist emtpy when ISA devices need DMAable memory.
Rather than decrease VM_DMA32_NPAGES_THRESHOLD, use the time honored
technique of putting initially allocated kernel data structs
at the end (or at least not the beginning) of memory.
Since this allocation is done at boot and is wired, is not freed,
so the system is low on 32-bit (and ISA) dma'ble memory forever.
So it is a good candidate to move above 4GB.
While here, remove an unneeded round_page() from kmem_malloc's size
argument as suggested by alc. The first thing kmem_malloc() does
is a round_page(size), so there is no need to do it before the call.
Reviewed by: alc
Sponsored by: Netflix
8x performance improvement in a micro benchmark on a 4 socket machine.
- Get buffer headers from a per-cpu uma cache that sits in from of the
free queue.
- Use a per-cpu quantum cache in vmem to eliminate contention for kva.
- Use multiple clean queues according to buffer cache size to eliminate
clean queue lock contention.
- Introduce a bufspace daemon that attempts to prevent getnewbuf() callers
from blocking or doing direct recycling.
- Close some bufspace allocation races that could lead to endless
recycling.
- Further the transition to a more modern style of small functions grouped
by prefix in order to improve growing complexity.
Sponsored by: EMC / Isilon
Reviewed by: kib
Tested by: pho
These changes prevent sysctl(8) from returning proper output,
such as:
1) no output from sysctl(8)
2) erroneously returning ENOMEM with tools like truss(1)
or uname(1)
truss: can not get etype: Cannot allocate memory
there is an environment variable which shall initialize the SYSCTL
during early boot. This works for all SYSCTL types both statically and
dynamically created ones, except for the SYSCTL NODE type and SYSCTLs
which belong to VNETs. A new flag, CTLFLAG_NOFETCH, has been added to
be used in the case a tunable sysctl has a custom initialisation
function allowing the sysctl to still be marked as a tunable. The
kernel SYSCTL API is mostly the same, with a few exceptions for some
special operations like iterating childrens of a static/extern SYSCTL
node. This operation should probably be made into a factored out
common macro, hence some device drivers use this. The reason for
changing the SYSCTL API was the need for a SYSCTL parent OID pointer
and not only the SYSCTL parent OID list pointer in order to quickly
generate the sysctl path. The motivation behind this patch is to avoid
parameter loading cludges inside the OFED driver subsystem. Instead of
adding special code to the OFED driver subsystem to post-load tunables
into dynamically created sysctls, we generalize this in the kernel.
Other changes:
- Corrected a possibly incorrect sysctl name from "hw.cbb.intr_mask"
to "hw.pcic.intr_mask".
- Removed redundant TUNABLE statements throughout the kernel.
- Some minor code rewrites in connection to removing not needed
TUNABLE statements.
- Added a missing SYSCTL_DECL().
- Wrapped two very long lines.
- Avoid malloc()/free() inside sysctl string handling, in case it is
called to initialize a sysctl from a tunable, hence malloc()/free() is
not ready when sysctls from the sysctl dataset are registered.
- Bumped FreeBSD version to indicate SYSCTL API change.
MFC after: 2 weeks
Sponsored by: Mellanox Technologies
an address in the first 2GB of the process's address space. This flag should
have the same semantics as the same flag on Linux.
To facilitate this, add a new parameter to vm_map_find() that specifies an
optional maximum virtual address. While here, fix several callers of
vm_map_find() to use a VMFS_* constant for the findspace argument instead of
TRUE and FALSE.
Reviewed by: alc
Approved by: re (kib)
that don't support superpages. This keeps the number of spans and internal
fragmentation lower.
- When the user asks for alignment from vmem_xalloc adjust the imported size
by 2*align to be certain we can satisfy the allocation. This comes at
the expense of potential failures when the backend can't supply enough
memory but could supply the requested size and alignment.
Sponsored by: EMC / Isilon Storage Division
address alignment of mappings.
- MAP_ALIGNED(n) requests a mapping aligned on a boundary of (1 << n).
Requests for n >= number of bits in a pointer or less than the size of
a page fail with EINVAL. This matches the API provided by NetBSD.
- MAP_ALIGNED_SUPER is a special case of MAP_ALIGNED. It can be used
to optimize the chances of using large pages. By default it will align
the mapping on a large page boundary (the system is free to choose any
large page size to align to that seems best for the mapping request).
However, if the object being mapped is already using large pages, then
it will align the virtual mapping to match the existing large pages in
the object instead.
- Internally, VMFS_ALIGNED_SPACE is now renamed to VMFS_SUPER_SPACE, and
VMFS_ALIGNED_SPACE(n) is repurposed for specifying a specific alignment.
MAP_ALIGNED(n) maps to using VMFS_ALIGNED_SPACE(n), while
MAP_ALIGNED_SUPER maps to VMFS_SUPER_SPACE.
- mmap() of a device object now uses VMFS_OPTIMAL_SPACE rather than
explicitly using VMFS_SUPER_SPACE. All device objects are forced to
use a specific color on creation, so VMFS_OPTIMAL_SPACE is effectively
equivalent.
Reviewed by: alc
MFC after: 1 month
transparent layering and better fragmentation.
- Normalize functions that allocate memory to use kmem_*
- Those that allocate address space are named kva_*
- Those that operate on maps are named kmap_*
- Implement recursive allocation handling for kmem_arena in vmem.
Reviewed by: alc
Tested by: pho
Sponsored by: EMC / Isilon Storage Division
originally inspired by the Solaris vmem detailed in the proceedings
of usenix 2001. The NetBSD version was heavily refactored for bugs
and simplicity.
- Use this resource allocator to allocate the buffer and transient maps.
Buffer cache defrags are reduced by 25% when used by filesystems with
mixed block sizes. Ultimately this may permit dynamic buffer cache
sizing on low KVA machines.
Discussed with: alc, kib, attilio
Tested by: pho
Sponsored by: EMC / Isilon Storage Division
do not map the b_pages pages into buffer_map KVA. The use of the
unmapped buffers eliminate the need to perform TLB shootdown for
mapping on the buffer creation and reuse, greatly reducing the amount
of IPIs for shootdown on big-SMP machines and eliminating up to 25-30%
of the system time on i/o intensive workloads.
The unmapped buffer should be explicitely requested by the GB_UNMAPPED
flag by the consumer. For unmapped buffer, no KVA reservation is
performed at all. The consumer might request unmapped buffer which
does have a KVA reserve, to manually map it without recursing into
buffer cache and blocking, with the GB_KVAALLOC flag.
When the mapped buffer is requested and unmapped buffer already
exists, the cache performs an upgrade, possibly reusing the KVA
reservation.
Unmapped buffer is translated into unmapped bio in g_vfs_strategy().
Unmapped bio carry a pointer to the vm_page_t array, offset and length
instead of the data pointer. The provider which processes the bio
should explicitely specify a readiness to accept unmapped bio,
otherwise g_down geom thread performs the transient upgrade of the bio
request by mapping the pages into the new bio_transient_map KVA
submap.
The bio_transient_map submap claims up to 10% of the buffer map, and
the total buffer_map + bio_transient_map KVA usage stays the
same. Still, it could be manually tuned by kern.bio_transient_maxcnt
tunable, in the units of the transient mappings. Eventually, the
bio_transient_map could be removed after all geom classes and drivers
can accept unmapped i/o requests.
Unmapped support can be turned off by the vfs.unmapped_buf_allowed
tunable, disabling which makes the buffer (or cluster) creation
requests to ignore GB_UNMAPPED and GB_KVAALLOC flags. Unmapped
buffers are only enabled by default on the architectures where
pmap_copy_page() was implemented and tested.
In the rework, filesystem metadata is not the subject to maxbufspace
limit anymore. Since the metadata buffers are always mapped, the
buffers still have to fit into the buffer map, which provides a
reasonable (but practically unreachable) upper bound on it. The
non-metadata buffer allocations, both mapped and unmapped, is
accounted against maxbufspace, as before. Effectively, this means that
the maxbufspace is forced on mapped and unmapped buffers separately.
The pre-patch bufspace limiting code did not worked, because
buffer_map fragmentation does not allow the limit to be reached.
By Jeff Roberson request, the getnewbuf() function was split into
smaller single-purpose functions.
Sponsored by: The FreeBSD Foundation
Discussed with: jeff (previous version)
Tested by: pho, scottl (previous version), jhb, bf
MFC after: 2 weeks
future further optimizations where the vm_object lock will be held
in read mode most of the time the page cache resident pool of pages
are accessed for reading purposes.
The change is mostly mechanical but few notes are reported:
* The KPI changes as follow:
- VM_OBJECT_LOCK() -> VM_OBJECT_WLOCK()
- VM_OBJECT_TRYLOCK() -> VM_OBJECT_TRYWLOCK()
- VM_OBJECT_UNLOCK() -> VM_OBJECT_WUNLOCK()
- VM_OBJECT_LOCK_ASSERT(MA_OWNED) -> VM_OBJECT_ASSERT_WLOCKED()
(in order to avoid visibility of implementation details)
- The read-mode operations are added:
VM_OBJECT_RLOCK(), VM_OBJECT_TRYRLOCK(), VM_OBJECT_RUNLOCK(),
VM_OBJECT_ASSERT_RLOCKED(), VM_OBJECT_ASSERT_LOCKED()
* The vm/vm_pager.h namespace pollution avoidance (forcing requiring
sys/mutex.h in consumers directly to cater its inlining functions
using VM_OBJECT_LOCK()) imposes that all the vm/vm_pager.h
consumers now must include also sys/rwlock.h.
* zfs requires a quite convoluted fix to include FreeBSD rwlocks into
the compat layer because the name clash between FreeBSD and solaris
versions must be avoided.
At this purpose zfs redefines the vm_object locking functions
directly, isolating the FreeBSD components in specific compat stubs.
The KPI results heavilly broken by this commit. Thirdy part ports must
be updated accordingly (I can think off-hand of VirtualBox, for example).
Sponsored by: EMC / Isilon storage division
Reviewed by: jeff
Reviewed by: pjd (ZFS specific review)
Discussed with: alc
Tested by: pho
from being indirectly called via cpu_startup()+vm_ksubmap_init().
The boot order position remains the same at SI_SUB_CPU.
Allocation of the callout array is changed to stardard kernel malloc
from a slightly obscure direct kernel_map allocation.
kern_timeout_callwheel_alloc() is renamed to callout_callwheel_init()
to better describe its purpose.
kern_timeout_callwheel_init() is removed simplifying the per-cpu
initialization.
Reviewed by: davide
details of the string buffer allocation in one place.
Eliminate the portion of the string buffer that was dedicated to storing
the interpreter name. The pointer to the interpreter name can simply be
made to point to the appropriate argument string.
Reviewed by: kib
shell command are stored in exec*()'s demand-paged string buffer. For
a "buildworld" on an 8GB amd64 multiprocessor, the new order reduces
the number of global TLB shootdowns by 31%. It also eliminates about
330k page faults on the kernel address space.
Change exec_shell_imgact() to use "args->begin_argv" consistently as
the start of the argument and environment strings. Previously, it
would sometimes use "args->buf", which is the start of the overall
buffer, but no longer the start of the argument and environment
strings. While I'm here, eliminate unnecessary passing of "&length"
to copystr(), where we don't actually care about the length of the
copied string.
Clean up the initialization of the exec map. In particular, use the
correct size for an entry, and express that size in the same way that
is used when an entry is allocated. The old size was one page too
large. (This discrepancy originated in 2004 when I rewrote
exec_map_first_page() to use sf_buf_alloc() instead of the exec map
for mapping the first page of the executable.)
Reviewed by: kib
no superpage mappings are created within the clean submap, aligning the
start of the clean submap helps to prevent interference with kmem_alloc()'s
use of superpages.
address space sizes to be longs instead of ints. Specifically, the follow
values are now longs: runningbufspace, bufspace, maxbufspace,
bufmallocspace, maxbufmallocspace, lobufspace, hibufspace, lorunningspace,
hirunningspace, maxswzone, maxbcache, and maxpipekva. Previously, a
relatively small number (~ 44000) of buffers set in kern.nbuf would result
in integer overflows resulting either in hangs or bogus values of
hidirtybuffers and lodirtybuffers. Now one has to overflow a long to see
such problems. There was a check for a nbuf setting that would cause
overflows in the auto-tuning of nbuf. I've changed it to always check and
cap nbuf but warn if a user-supplied tunable would cause overflow.
Note that this changes the ABI of several sysctls that are used by things
like top(1), etc., so any MFC would probably require a some gross shims
to allow for that.
MFC after: 1 month
used to request superpage alignment for the submap.
Request superpage alignment for the kmem_map.
Pass VMFS_ANY_SPACE instead of TRUE to vm_map_find(). (They are currently
equivalent but VMFS_ANY_SPACE is the new preferred spelling.)
Remove a stale comment from kmem_malloc().
after each SYSINIT() macro invocation. This makes a number of
lightweight C parsers much happier with the FreeBSD kernel
source, including cflow's prcc and lxr.
MFC after: 1 month
Discussed with: imp, rink
the number of entries in exec_map (maximum number of simultaneous execs
that can be handled by the kernel). The default value of 16 is
insufficient on heavily loaded machines (particularly SMP machines), and
if it is exceeded then executing further processes will generate a SIGABRT.
This is a workaround until a better solution can be implemented.
Reviewed by: alc
MFC after: 3 days
the reduction of the pager map's size by 8M bytes. In other words, eight
megabytes of largely wasted KVA are returned to the kernel map for use
elsewhere.