Read locking is over used in the kernel to guarantee liveness. This API makes
it easy to provide livenes guarantees without atomics.
Includes epoch_test kernel module to stress test the API.
Documentation will follow initial use case.
Test case and improvements to preemption handling in response to discussion
with mjg@
Reviewed by: imp@, shurd@
Approved by: sbruno@
Each malloc/free was testing dtrace_malloc_enabled and forcing
extra reads from the malloc type struct to see if perhaps a
dtmalloc probe was on.
Treat it like lockstat and sdt: have a global bolean.
malloc was showing at the top of profile during while running microbenchmarks.
#define DTMALLOC_PROBE_MAX 2
struct malloc_type_internal {
uint32_t mti_probes[DTMALLOC_PROBE_MAX];
u_char mti_zone;
struct malloc_type_stats mti_stats[MAXCPU];
};
Reading mti_zone it wastes a cacheline to hold mti_probes + mti_zone
(which we know is 0) + part of malloc stats of the first cpu which on top
induces false-sharing.
In particular will-it-scale lock1_processes -t 128 -s 10:
before: average:45879692
after: average:51655596
Note the counters can be padded but the right fix is to move them to
counter(9), leaving the struct read-only after creation (modulo dtrace
probes).
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
for UMA startup.
o Introduce another stage of UMA startup, which is entered after
vm_page_startup() finishes. After this stage we don't yet enable buckets,
but we can ask VM for pages. Rename stages to meaningful names while here.
New list of stages: BOOT_COLD, BOOT_STRAPPED, BOOT_PAGEALLOC, BOOT_BUCKETS,
BOOT_RUNNING.
Enabling page alloc earlier allows us to dramatically reduce number of
boot pages required. What is more important number of zones becomes
consistent across different machines, as no MD allocations are done before
the BOOT_PAGEALLOC stage. Now only UMA internal zones actually need to use
startup_alloc(), however that may change, so vm_page_startup() provides
its need for early zones as argument.
o Introduce uma_startup_count() function, to avoid code duplication. The
functions calculates sizes of zones zone and kegs zone, and calculates how
many pages UMA will need to bootstrap.
It counts not only of zone structures, but also of kegs, slabs and hashes.
o Hide uma_startup_foo() declarations from public file.
o Provide several DIAGNOSTIC printfs on boot_pages usage.
o Bugfix: when calculating zone of zones size use (mp_maxid + 1) instead of
mp_ncpus. Use resulting number not only in the size argument to zone_ctor()
but also as args.size.
Reviewed by: imp, gallatin (earlier version)
Differential Revision: https://reviews.freebsd.org/D14054
domains can be done by the _domain() API variants. UMA also supports a
first-touch policy via the NUMA zone flag.
The slab layer is now segregated by VM domains and is precise. It handles
iteration for round-robin directly. The per-cpu cache layer remains
a mix of domains according to where memory is allocated and freed. Well
behaved clients can achieve perfect locality with no performance penalty.
The direct domain allocation functions have to visit the slab layer and
so require per-zone locks which come at some expense.
Reviewed by: Attilio (a slightly older version)
Tested by: pho
Sponsored by: Netflix, Dell/EMC Isilon
Additionally, move the overflow check logic out to WOULD_OVERFLOW() for
consumers to have a common means of testing for overflowing allocations.
WOULD_OVERFLOW() should be a secondary check -- on 64-bit platforms, just
because an allocation won't overflow size_t does not mean it is a sane size
to request. Callers should be imposing reasonable allocation limits far,
far, below overflow.
Discussed with: emaste, jhb, kp
Sponsored by: Dell EMC Isilon
Similar to calloc() the mallocarray() function checks for integer
overflows before allocating memory.
It does not zero memory, unless the M_ZERO flag is set.
Reviewed by: pfg, vangyzen (previous version), imp (previous version)
Obtained from: OpenBSD
Differential Revision: https://reviews.freebsd.org/D13766
The arena argument to kmem_*() is now only used in an assert. A follow-up
commit will remove the argument altogether before we freeze the API for the
next release.
This replaces the hard limit on kmem size with a soft limit imposed by UMA. When
the soft limit is exceeded we periodically wakeup the UMA reclaim thread to
attempt to shrink KVA. On 32bit architectures this should behave much more
gracefully as we exhaust KVA. On 64bit the limits are likely never hit.
Reviewed by: markj, kib (some objections)
Discussed with: alc
Tested by: pho
Sponsored by: Netflix / Dell EMC Isilon
Differential Revision: https://reviews.freebsd.org/D13187
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.
A panicking thread always executes with a critical section held, so any
attempt to allocate or free memory while dumping will otherwise cause a
second panic. This can occur, for example, if xpt_polled_action() completes
non-dump I/O that was pending at the time of the panic. The fact that this
can occur is itself a bug, but asserting in this case does little but
reduce the reliability of kernel dumps.
Suggested by: kib
Reported by: pho
critical section.
uma_zalloc_arg()/uma_zalloc_free() may acquire a sleepable lock on the
zone. The malloc() family of functions may call uma_zalloc_arg() or
uma_zalloc_free().
The malloc(9) man page currently claims that free() will never sleep.
It also implies that the malloc() family of functions will not sleep
when called with M_NOWAIT. However, it is more correct to say that
these functions will not sleep indefinitely. Indeed, they may acquire
a sleepable lock. However, a developer may overlook this restriction
because the WITNESS check that catches attempts to call the malloc()
family of functions within a critical section is inconsistenly
applied.
This change clarifies the language of the malloc(9) man page to clarify
the restriction against calling the malloc() family of functions
while in a critical section or holding a spin lock. It also adds
KASSERTs at appropriate points to make the enforcement of this
restriction more consistent.
PR: 204633
Differential Revision: https://reviews.freebsd.org/D4197
Reviewed by: markj
Approved by: gnn (mentor)
Sponsored by: Juniper Networks
fragmented conditions currently just wakes up the pagedaemon. The
kmem arena is significantly smaller then the total available physical
memory, which means that there are loads where kmem arena space could
be exhausted, while there is a lot of pages available still. The
woken up pagedaemon sees vm_pages_needed != 0, verifies the condition
vm_paging_needed() which is false, clears the pass and returns back to
sleep, not calling neither uma_reclaim() nor lowmem handler.
To handle low kmem arena conditions, create additional pagedaemon
thread which calls uma_reclaim() directly. The thread sleeps on the
dedicated channel and kmem_reclaim() wakes the thread in addition to
the pagedaemon.
Reported and tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
strings returned to userland include the nulterm byte.
Some uses of sbuf_new_for_sysctl() write binary data rather than strings;
clear the SBUF_INCLUDENUL flag after calling sbuf_new_for_sysctl() in
those cases. (Note that the sbuf code still automatically adds a nulterm
byte in sbuf_finish(), but since it's not included in the length it won't
get copied to userland along with the binary data.)
Remove explicit adding of a nulterm byte in a couple places now that it
gets done automatically by the sbuf drain code.
PR: 195668
footprint systems(32M/64M) and didn't leave enough free memory to load modules
when it was setting up page tables that for sizes that are never used on
these smallish boards.
Set kmem_zmax to PAGE_SIZE on these smaller systems (< 128M) to keep this
from happening. Verified on mips32 h/w.
PR: 193465
Submitted by: delphij
Reviewed by: adrian
Remove duplicate "debug_ktr.mask" sysctl definition.
Remove now unused variable from "kern_ktr.c".
This fixes build of "ktr" which was broken by r267961.
Let the default value for "vm_kmem_size_scale" be zero. It is setup
after that the sysctl has been initialized from "getenv()" in the
"kmeminit()" function to equal the "VM_KMEM_SIZE_MAX" value, if
zero. On Sparc64 the "VM_KMEM_SIZE_MAX" macro is not a constant. This
fixes build of Sparc64 which was broken by r267961.
Add a special macro to dynamically create SYSCTL root nodes, because
root nodes have a special parent. This fixes build of existing OFED
module and CANBUS module for pc98 which was broken by r267961.
Add missing "sysctl.h" includes to get the needed sysctl header file
declarations. This is needed after r267961.
MFC after: 2 weeks
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
To reduce the diff struct pcu.cnt field was not renamed, so
PCPU_OP(cnt.field) is still used. pc_cnt and pcpu are also used in
kvm(3) and vmstat(8). The goal was to not affect externally used KPI.
Bump __FreeBSD_version_ in case some out-of-tree module/code relies on the
the global cnt variable.
Exp-run revealed no ports using it directly.
No objection from: arch@
Sponsored by: EMC / Isilon Storage Division
option, unbreak the lock tracing release semantic by embedding
calls to LOCKSTAT_PROFILE_RELEASE_LOCK() direclty in the inlined
version of the releasing functions for mutex, rwlock and sxlock.
Failing to do so skips the lockstat_probe_func invokation for
unlocking.
- As part of the LOCKSTAT support is inlined in mutex operation, for
kernel compiled without lock debugging options, potentially every
consumer must be compiled including opt_kdtrace.h.
Fix this by moving KDTRACE_HOOKS into opt_global.h and remove the
dependency by opt_kdtrace.h for all files, as now only KDTRACE_FRAMES
is linked there and it is only used as a compile-time stub [0].
[0] immediately shows some new bug as DTRACE-derived support for debug
in sfxge is broken and it was never really tested. As it was not
including correctly opt_kdtrace.h before it was never enabled so it
was kept broken for a while. Fix this by using a protection stub,
leaving sfxge driver authors the responsibility for fixing it
appropriately [1].
Sponsored by: EMC / Isilon storage division
Discussed with: rstone
[0] Reported by: rstone
[1] Discussed with: philip
words, every architecture is now auto-sizing the kmem arena. This revision
changes kmeminit() so that the definition of VM_KMEM_SIZE_SCALE becomes
mandatory and the definition of VM_KMEM_SIZE becomes optional.
Replace or eliminate all existing definitions of VM_KMEM_SIZE. With
auto-sizing enabled, VM_KMEM_SIZE effectively became an alternate spelling
for VM_KMEM_SIZE_MIN on most architectures. Use VM_KMEM_SIZE_MIN for
clarity.
Change kmeminit() so that the effect of defining VM_KMEM_SIZE is similar to
that of setting the tunable vm.kmem_size. Whereas the macros
VM_KMEM_SIZE_{MAX,MIN,SCALE} have had the same effect as the tunables
vm.kmem_size_{max,min,scale}, the effects of VM_KMEM_SIZE and vm.kmem_size
have been distinct. In particular, whereas VM_KMEM_SIZE was overridden by
VM_KMEM_SIZE_{MAX,MIN,SCALE} and vm.kmem_size_{max,min,scale}, vm.kmem_size
was not. Remedy this inconsistency. Now, VM_KMEM_SIZE can be used to set
the size of the kmem arena at compile-time without that value being
overridden by auto-sizing.
Update the nearby comments to reflect the kmem submap being replaced by the
kmem arena. Stop duplicating the auto-sizing formula in every machine-
dependent vmparam.h and place it in kmeminit() where auto-sizing takes
place.
Reviewed by: kib (an earlier version)
Sponsored by: EMC / Isilon Storage Division
kmeminit() runs, so it contributes nothing to 'vm_kmem_size'; update a
comment to reflect that r254025 replaced the kmem submap with the kmem
arena.
Reviewed by: kib
Approved by: re (gjb)
Sponsored by: EMC / Isilon Storage Division
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
VM_KMEM_MAX_SIZE.
The code was not taking into account the size of the kernel_map, which
the kmem_map is allocated from, so it could produce a sub-map size too
large to fit. The simplest solution is to ignore VM_KMEM_MAX entirely
and base the memguard map's size off the kernel_map's size, since this
is always relevant and always smaller.
Found by: Justin Hibbits
long for specifying a boundary constraint.
- Change bus_dma tags to use bus_addr_t instead of bus_size_t for boundary
constraints.
These allow boundary constraints to be fully expressed for cases where
sizeof(bus_addr_t) != sizeof(bus_size_t). Specifically, it allows a
driver to properly specify a 4GB boundary in a PAE kernel.
Note that this cannot be safely MFC'd without a lot of compat shims due
to KBI changes, so I do not intend to merge it.
Reviewed by: scottl
of vm_kmem_size that may occur if the system administrator has specified a
vm.vm_kmem_size tunable value that exceeds the hard cap.
PR: 162741
Submitted by: Adam McDougall
Reviewed by: bde@
MFC after: 3 weeks
The SYSCTL_NODE macro defines a list that stores all child-elements of
that node. If there's no SYSCTL_DECL macro anywhere else, there's no
reason why it shouldn't be static.
more general VM system interfaces. So, their implementation can now
reside in kern_malloc.c alongside the other functions that are declared
in malloc.h.
the unlikely event that sysctl_kmem_map_free() was performed on an
empty kmem map, it would incorrectly report the free space as zero.
Discussed with: avg
MFC after: 1 week
Also, express this new maximum as a fraction of the kernel's address
space size rather than a constant so that increasing KVA_PAGES will
automatically increase this maximum. As a side-effect of this change,
kern.maxvnodes will automatically increase by a proportional amount.
While I'm here ensure that this change doesn't result in an unintended
increase in maxpipekva on i386. Calculate maxpipekva based upon the
size of the kernel address space and the amount of physical memory
instead of the size of the kmem map. The memory backing pipes is not
allocated from the kmem map. It is allocated from its own submap of
the kernel map. In short, it has no real connection to the kmem map.
(In fact, the commit messages for the maxpipekva auto-sizing talk
about using the kernel map size, cf. r117325 and r117391, even though
the implementation actually used the kmem map size.) Although the
calculation is now done differently, the resulting value for
maxpipekva should remain almost the same on i386. However, on amd64,
the value will be reduced by 2/3. This is intentional. The recent
change to VM_KMEM_SIZE_SCALE on amd64 for the benefit of ZFS also had
the unnecessary side-effect of increasing maxpipekva. This change is
effectively restoring maxpipekva on amd64 to its prior value.
Eliminate init_param3() since it is no longer used.
sbuf_new_for_sysctl(9). This allows using an sbuf with a SYSCTL_OUT
drain for extremely large amounts of data where the caller knows that
appropriate references are held, and sleeping is not an issue.
Inspired by: rwatson
Add a drain function for struct sysctl_req, and use it for a variety
of handlers, some of which had to do awkward things to get a large
enough SBUF_FIXEDLEN buffer.
Note that some sysctl handlers were explicitly outputting a trailing
NUL byte. This behaviour was preserved, though it should not be
necessary.
Reviewed by: phk (original patch)
unexpected things in copyout(9) and so wiring the user buffer is not
sufficient to perform a copyout(9) while holding a random mutex.
Requested by: nwhitehorn
handlers, some of which had to do awkward things to get a large enough
FIXEDLEN buffer.
Note that some sysctl handlers were explicitly outputting a trailing NUL
byte. This behaviour was preserved, though it should not be necessary.
Reviewed by: phk
use-after-free over a longer time. Also release the backing pages of
a guarded allocation at free(9) time to reduce the overhead of using
memguard(9). Allow setting and varying the malloc type at run-time.
Add knobs to allow:
- randomly guarding memory
- adding un-backed KVA guard pages to detect underflow and overflow
- a lower limit on the size of allocations that are guarded
Reviewed by: alc
Reviewed by: brueffer, Ulrich Spörlein <uqs spoerlein net> (man page)
Silence from: -arch
Approved by: zml (mentor)
MFC after: 1 month
zones for each malloc bucket size. The purpose is to isolate
different malloc types into hash classes, so that any buffer overruns
or use-after-free will usually only affect memory from malloc types in
that hash class. This is purely a debugging tool; by varying the hash
function and tracking which hash class was corrupted, the intersection
of the hash classes from each instance will point to a single malloc
type that is being misused. At this point inspection or memguard(9)
can be used to catch the offending code.
Add MALLOC_DEBUG_MAXZONES=8 to -current GENERIC configuration files.
The suggestion to have this on by default came from Kostik Belousov on
-arch.
This code is based on work by Ron Steinke at Isilon Systems.
Reviewed by: -arch (mostly silence)
Reviewed by: zml
Approved by: zml (mentor)
a KASSERT to handle it. People are likely to turn off INVARIANTS RSN
and loading an old module can cause garbage-in here.
I saw the issue with an older nvidia driver (x11/nvidia-driver) loading
into a new kernel - a crash wasn't seen 'till sysctl_kern_malloc_stats().
I was lucky that mtp->ks_shortdesc was NULL and not something horrible.
While I'm here, KASSERT that malloc_uninit() isn't passed something that's
not in kmemstatistics.
MFC after: 3 weeks
kern_time.c:
- Unused variable `p'.
kern_thr.c:
- Variable `error' is always caught immediately, so no reason to
initialize it. There is no way that error != 0 at the end of
create_thread().
kern_sig.c:
- Unused variable `code'.
kern_synch.c:
- `rval' is always assigned in all different cases.
kern_rwlock.c:
- `v' is always overwritten with RW_UNLOCKED further on.
kern_malloc.c:
- `size' is always initialized with the proper value before being used.
kern_exit.c:
- `error' is always caught and returned immediately. abort2() never
returns a non-zero value.
kern_exec.c:
- `len' is always assigned inside the if-statement right below it.
tty_info.c:
- `td' is always overwritten by FOREACH_THREAD_IN_PROC().
Found by: LLVM's scan-build
backend kegs so it may source compatible memory from multiple backends.
This is useful for cases such as NUMA or different layouts for the same
memory type.
- Provide a new api for adding new backend kegs to secondary zones.
- Provide a new flag for adjusting the layout of zones to stagger
allocations better across cache lines.
Sponsored by: Nokia
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
Probabilly, a general approach is not the better solution here, so we should
solve the sched_lock protection problems separately.
Requested by: alc
Approved by: jeff (mentor)
vmcnts. This can be used to abstract away pcpu details but also changes
to use atomics for all counters now. This means sched lock is no longer
responsible for protecting counts in the switch routines.
Contributed by: Attilio Rao <attilio@FreeBSD.org>
vm.kmem_size_min. Useful when using ZFS to make sure that vm.kmem size will
be at least 256mb (for example) without forcing a particular value via vm.kmem_size.
Approved by: njl (mentor)
Reviewed by: alc
counters of allocs/frees/use for each malloc type to calculating InUse,
MemUse, and Requests as displayed by the userspace vmstat -m. This is
more useful when debugging malloc(9)-related memory leaks, where the
count of allocs/frees may not usefully reflect that current memory
allocation (i.e., when highly variable size allocations occur with the
same malloc type, such as with contigmalloc).
MFC after: 3 days
Limitations observed by: scottl
the kernel malloc(9) state for vmstat -m. libmemstat is now used to
generate a machine-readable version which is converged by vmstat -m
into a human-readable version.
Not for MFC.
Releasing items from the mt_zone can not be done by a simple
uma_zfree() call since mt_zone is allocated with the UMA_ZONE_MALLOC
flag. Use uma_zfree_arg instead and supply the slab.
This bug caused panics in low memory situations on unloading kernel
modules containing MALLOC_DEFINE(..) statements.
Submitted by: ups
It detects both: buffer underflows and buffer overflows bugs at runtime
(on free(9) and realloc(9)) and prints backtraces from where memory was
allocated and from where it was freed.
Tested by: kris
- Provide tunable vm.memguard.desc, so one can specify memory type without
changing the code and recompiling the kernel.
- Allow to use memguard for kernel modules by providing sysctl
vm.memguard.desc, which can be changed to short description of memory
type before module is loaded.
- Move as much memguard code as possible to memguard.c.
- Add sysctl node vm.memguard. and move memguard-specific sysctl there.
- Add malloc_desc2type() function for finding memory type based on its
short description (ks_shortdesc field).
- Memory type can be changed (via vm.memguard.desc sysctl) only if it
doesn't exist (will be loaded later) or when no memory is allocated yet.
If there is allocated memory for the given memory type, return EBUSY.
- Implement two ways of memory types comparsion and make safer/slower the
default.
UMA_SLAB_MALLOC flag.
In some circumstances (I observed it when I was doing a lot of reallocs)
UMA_SLAB_MALLOC can be set even if us_keg != NULL.
If this is the case we have wonderful, silent data corruption, because less
data is copied to the newly allocated region than should be.
I'm not sure when this bug was introduced, it could be there undetected
for years now, as we don't have a lot of realloc(9) consumers and it was
hard to reproduce it...
...but what I know for sure, is that I don't want to know who introduce
the bug:) It took me two/three days to track it down (of course most of
the time I was looking for the bug in my own code).
available kernel malloc types. Quite useful for post-mortem debugging of
memory leaks without a dump device configured on a panicked box.
MFC after: 2 weeks
- increase number of allocations count only on successfull malloc(9),
so it doesn't confuse people;
- because we need to check if 'size > 0', hide 'mtsp->mts_memalloced += size;'
under the check as well, as for size=0 it is of course a no-op;
- avoid critical_enter()/critical_exit() in case of failure in
malloc_type_allocated() as there will be nothing to do.
OK'ed by: rwatson
MFC after: 2 days
statistics via a binary structure stream:
- Add structure 'malloc_type_stream_header', which defines a stream
version, definition of MAXCPUS used in the stream, and a number of
malloc_type records in the stream.
- Add structure 'malloc_type_header', which defines the name of the
malloc type being reported on.
- When the sysctl is queried, return a stream header, followed by a
series of type descriptions, each consisting of a type header
followed by a series of MAXCPUS malloc_type_stats structures holding
per-CPU allocation information. Typical values of MAXCPUS will be 1
(UP compiled kernel) and 16 (SMP compiled kernel).
This query mechanism allows user space monitoring tools to extract
memory allocation statistics in a machine-readable form, and to do so
at a per-CPU granularity, allowing monitoring of allocation patterns
across CPUs in order to better understand the distribution of work and
memory flow over multiple CPUs.
While here:
- Bump statistics width to uint64_t, and hard code using fixed-width
type in order to be more sure about structure layout in the stream.
We allocate and free a lot of memory.
- Add kmemcount, a counter of the number of registered malloc types,
in order to avoid excessive manual counting of types. Export via a
new sysctl to allow user-space code to better size buffers.
- De-XXX comment on no longer maintaining the high watermark in old
sysctl monitoring code.
A follow-up commit of libmemstat(3), a library to monitor kernel memory
allocation, will occur in the next few days. Likewise, similar changes
to UMA.
allocators: a set of power-of-two UMA zones for small allocations, and the
VM page allocator for large allocations. In order to maintain unified
statistics for specific malloc types, kernel malloc maintains a separate
per-type statistics pool, which can be monitored using vmstat -m. Prior
to this commit, each pool of per-type statistics was protected using a
per-type mutex associated with the malloc type.
This change modifies kernel malloc to maintain per-CPU statistics pools
for each malloc type, and protects writing those statistics using critical
sections. It also moves to unsynchronized reads of per-CPU statistics
when generating coalesced statistics. To do this, several changes are
implemented:
- In the previous world order, the statistics memory was allocated by
the owner of the malloc type structure, allocated statically using
MALLOC_DEFINE(). This embedded the definition of the malloc_type
structure into all kernel modules. Move to a model in which a pointer
within struct malloc_type points at a UMA-allocated
malloc_type_internal data structure owned and maintained by
kern_malloc.c, and not part of the exported ABI/API to the rest of
the kernel. For the purposes of easing a possible MFC, re-use an
existing pointer in 'struct malloc_type', and maintain the current
malloc_type structure size, as well as layout with respect to the
fields reused outside of the malloc subsystem (such as ks_shortdesc).
There are several unused fields as a result of no longer requiring
the mutex in malloc_type.
- Struct malloc_type_internal contains an array of malloc_type_stats,
of size MAXCPU. The structure defined above avoids hard-coding a
kernel compile-time value of MAXCPU into kernel modules that interact
with malloc.
- When accessing per-cpu statistics for a malloc type, surround read -
modify - update requests with critical_enter()/critical_exit() in
order to avoid races during write. The per-CPU fields are written
only from the CPU that owns them.
- Per-CPU stats now maintained "allocated" and "freed" counters for
number of allocations/frees and bytes allocated/freed, since there is
no longer a coherent global notion of the totals. When coalescing
malloc stats, accept a slight race between reading stats across CPUs,
and avoid showing the user a negative allocation count for the type
in the event of a race. The global high watermark is no longer
maintained for a malloc type, as there is no global notion of the
number of allocations.
- While tearing up the sysctl() path, also switch to using sbufs. The
current "export as text" sysctl format is retained with the same
syntax. We may want to change this in the future to export more
per-CPU information, such as how allocations and frees are balanced
across CPUs.
This change results in a substantial speedup of kernel malloc and free
paths on SMP, as critical sections (where usable) out-perform mutexes
due to avoiding atomic/bus-locked operations. There is also a minor
improvement on UP due to the slightly lower cost of critical sections
there. The cost of the change to this approach is the loss of a
continuous notion of total allocations that can be exploited to track
per-type high watermarks, as well as increased complexity when
monitoring statistics.
Due to carefully avoiding changing the ABI, as well as hardening the ABI
against future changes, it is not necessary to recompile kernel modules
for this change. However, MFC'ing this change to RELENG_5 will require
also MFC'ing optimizations for soft critical sections, which may modify
exposed kernel ABIs. The internal malloc API is changed, and
modifications to vmstat in order to restore "vmstat -m" on core dumps will
follow shortly.
Several improvements from: bde
Statistics approach discussed with: ups
Tested by: scottl, others
designed to help detect tamper-after-free scenarios, a problem more
and more common and likely with multithreaded kernels where race
conditions are more prevalent.
Currently MemGuard can only take over malloc()/realloc()/free() for
particular (a) malloc type(s) and the code brought in with this
change manually instruments it to take over M_SUBPROC allocations
as an example. If you are planning to use it, for now you must:
1) Put "options DEBUG_MEMGUARD" in your kernel config.
2) Edit src/sys/kern/kern_malloc.c manually, look for
"XXX CHANGEME" and replace the M_SUBPROC comparison with
the appropriate malloc type (this might require additional
but small/simple code modification if, say, the malloc type
is declared out of scope).
3) Build and install your kernel. Tune vm.memguard_divisor
boot-time tunable which is used to scale how much of kmem_map
you want to allott for MemGuard's use. The default is 10,
so kmem_size/10.
ToDo:
1) Bring in a memguard(9) man page.
2) Better instrumentation (e.g., boot-time) of MemGuard taking
over malloc types.
3) Teach UMA about MemGuard to allow MemGuard to override zone
allocations too.
4) Improve MemGuard if necessary.
This work is partly based on some old patches from Ian Dowse.
improved chance of working despite pressure from running programs.
Instead of trying to throw a bunch of pages out to swap and hope for
the best, only a range that can potentially fulfill contigmalloc(9)'s
request will have its contents paged out (potentially, not forcibly)
at a time.
The new contigmalloc operation still operates in three passes, but it
could potentially be tuned to more or less. The first pass only looks
at pages in the cache and free pages, so they would be thrown out
without having to block. If this is not enough, the subsequent passes
page out any unwired memory. To combat memory pressure refragmenting
the section of memory being laundered, each page is removed from the
systems' free memory queue once it has been freed so that blocking
later doesn't cause the memory laundered so far to get reallocated.
The page-out operations are now blocking, as it would make little sense
to try to push out a page, then get its status immediately afterward
to remove it from the available free pages queue, if it's unlikely to
have been freed. Another change is that if KVA allocation fails, the
allocated memory segment will be freed and not leaked.
There is a sysctl/tunable, defaulting to on, which causes the old
contigmalloc() algorithm to be used. Nonetheless, I have been using
vm.old_contigmalloc=0 for over a month. It is safe to switch at
run-time to see the difference it makes.
A new interface has been used which does not require mapping the
allocated pages into KVA: vm_page.h functions vm_page_alloc_contig()
and vm_page_release_contig(). These are what vm.old_contigmalloc=0
uses internally, so the sysctl/tunable does not affect their operation.
When using the contigmalloc(9) and contigfree(9) interfaces, memory
is now tracked with malloc(9) stats. Several functions have been
exported from kern_malloc.c to allow other subsystems to use these
statistics, as well. This invalidates the BUGS section of the
contigmalloc(9) manpage.
