This significantly reduces contention since chunks get created and removed
all the time. See the review for sample results.
Reviewed by: kib
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21976
No functional change (in program code). Additional DWARF metadata is
generated in the .eh_frame section. Also, it is now a compile-time
requirement that machine/asm.h ENTRY() and END() macros are paired. (This
is subject to ongoing discussion and may change.)
This DWARF metadata allows llvm-libunwind to unwind program stacks when the
program is executing the function. The goal is to collect accurate
userspace stacktraces when programs have entered syscalls.
(The motivation for "Call Frame Information," or CFI for short -- not to be
confused with Control Flow Integrity -- is to sufficiently annotate assembly
functions such that stack unwinders can unwind out of the local frame
without the requirement of a dedicated framepointer register; i.e.,
-fomit-frame-pointer. This is necessary for C++ exception handling or
collecting backtraces.)
For the curious, a more thorough description of the metadata and some
examples may be found at [1] and documentation at [2]. You can also look at
'cc -S -o - foo.c | less' and search for '.cfi_' to see the CFI directives
generated by your C compiler.
[1]: https://www.imperialviolet.org/2017/01/18/cfi.html
[2]: https://sourceware.org/binutils/docs/as/CFI-directives.html
Reviewed by: emaste, kib (with reservations)
Differential Revision: https://reviews.freebsd.org/D22122
Instead of superpages use. The current code employs superpage-wide locking
regardless and the better locking granularity is welcome with NUMA enabled
even when superpage support is not used.
Requested by: alc
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21982
After removing wmb(), vm_set_rendezvous_func() became super trivial, so
there was no point in keeping it.
The wmb (sfence on amd64, lock nop on i386) was not needed. This can be
explained from several points of view.
First, wmb() is used for store-store ordering (although, the primitive
is undocumented). There was no obvious subsequent store that needed the
barrier.
Second, x86 has a memory model with strong ordering including total
store order. An explicit store barrier may be needed only when working
with special memory (device, special caching mode) or using special
instructions (non-temporal stores). That was not the case for this
code.
Third, I believe that there is a misconception that sfence "flushes" the
store buffer in a sense that it speeds up the propagation of stores from
the store buffer to the global visibility. I think that such
propagation always happens as fast as possible. sfence only makes
subsequent stores wait for that propagation to complete. So, sfence is
only useful for ordering of stores and only in the situations described
above.
Reviewed by: jhb
MFC after: 23 days
Differential Revision: https://reviews.freebsd.org/D21978
- We load the kernel at 0x200000. Memory below that address need not
be executable, so do not map it as such.
- Remove references to .ldata and related sections in the kernel linker
script. They come from ld.bfd's default linker script, but are not
used, and we now use ld.lld to link the amd64 kernel. lld does not
contain a default linker script.
- Pad the .bss to a 2MB as we do between .text and .data. This
forces the loader to load additional files starting in the following
2MB page, preserving the use of superpage mappings for kernel data.
- Map memory above the kernel image with NX. The kernel linker now
upgrades protections as needed, and other preloaded file types
(e.g., entropy, microcode) need not be mapped with execute permissions
in the first place.
Reviewed by: kib
MFC after: 1 month
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21859
Move futex_mtx to linux_common.ko for amd64 and aarch64 along
with respective list/mutex init/destroy.
PR: 240989
Reported by: Alex S <iwtcex@gmail.com>
NetGDB(4) is a component of a system using a panic-time network stack to
remotely debug crashed FreeBSD kernels over the network, instead of
traditional serial interfaces.
There are three pieces in the complete NetGDB system.
First, a dedicated proxy server must be running to accept connections from
both NetGDB and gdb(1), and pass bidirectional traffic between the two
protocols.
Second, the NetGDB client is activated much like ordinary 'gdb' and
similarly to 'netdump' in ddb(4) after a panic. Like other debugnet(4)
clients (netdump(4)), the network interface on the route to the proxy server
must be online and support debugnet(4).
Finally, the remote (k)gdb(1) uses 'target remote <proxy>:<port>' (like any
other TCP remote) to connect to the proxy server.
The NetGDB v1 protocol speaks the literal GDB remote serial protocol, and
uses a 1:1 relationship between GDB packets and sequences of debugnet
packets (fragmented by MTU). There is no encryption utilized to keep
debugging sessions private, so this is only appropriate for local
segments or trusted networks.
Submitted by: John Reimer <john.reimer AT emc.com> (earlier version)
Discussed some with: emaste, markj
Relnotes: sure
Differential Revision: https://reviews.freebsd.org/D21568
Debugnet is a simplistic and specialized panic- or debug-time reliable
datagram transport. It can drive a single connection at a time and is
currently unidirectional (debug/panic machine transmit to remote server
only).
It is mostly a verbatim code lift from netdump(4). Netdump(4) remains
the only consumer (until the rest of this patch series lands).
The INET-specific logic has been extracted somewhat more thoroughly than
previously in netdump(4), into debugnet_inet.c. UDP-layer logic and up, as
much as possible as is protocol-independent, remains in debugnet.c. The
separation is not perfect and future improvement is welcome. Supporting
INET6 is a long-term goal.
Much of the diff is "gratuitous" renaming from 'netdump_' or 'nd_' to
'debugnet_' or 'dn_' -- sorry. I thought keeping the netdump name on the
generic module would be more confusing than the refactoring.
The only functional change here is the mbuf allocation / tracking. Instead
of initiating solely on netdump-configured interface(s) at dumpon(8)
configuration time, we watch for any debugnet-enabled NIC for link
activation and query it for mbuf parameters at that time. If they exceed
the existing high-water mark allocation, we re-allocate and track the new
high-water mark. Otherwise, we leave the pre-panic mbuf allocation alone.
In a future patch in this series, this will allow initiating netdump from
panic ddb(4) without pre-panic configuration.
No other functional change intended.
Reviewed by: markj (earlier version)
Some discussion with: emaste, jhb
Objection from: marius
Differential Revision: https://reviews.freebsd.org/D21421
This updates the protection attributes of subranges of the kernel map.
Unlike pmap_protect(), which is typically used for user mappings,
pmap_change_prot() does not perform lazy upgrades of protections.
pmap_change_prot() also updates the aliasing range of the direct map.
Reviewed by: kib
MFC after: 1 month
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21758
After r352110 the page lock no longer protects a page's identity, so
there is no purpose in locking the page in pmap_mincore(). Instead,
if vm.mincore_mapped is set to the non-default value of 0, re-lookup
the page after acquiring its object lock, which holds the page's
identity stable.
The change removes the last callers of vm_page_pa_tryrelock(), so
remove it.
Reviewed by: kib
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21823
callers hold it.
This simplifies pmap code and removes a dependency on the object lock.
Reviewed by: kib, markj
Tested by: pho
Sponsored by: Netflix, Intel
Differential Revision: https://reviews.freebsd.org/D21596
Atomics are used for page busy and valid state when the shared busy is
held. The details of the locking protocol and valid and dirty
synchronization are in the updated vm_page.h comments.
Reviewed by: kib, markj
Tested by: pho
Sponsored by: Netflix, Intel
Differential Revision: https://reviews.freebsd.org/D21594
busy acquires while held.
This allows code that would need to acquire and release a very large number
of page busy locks to use the old mechanism where busy is only checked and
not held. This comes at the cost of false positives but never false
negatives which the single consumer, vm_fault_soft_fast(), handles.
Reviewed by: kib
Tested by: pho
Sponsored by: Netflix, Intel
Differential Revision: https://reviews.freebsd.org/D21592
There are provisions to do it already with pv_dummy, but new locking code
did not account for it. Previous one did not have the problem because
it hashed the address into the lock array.
While here annotate common vars with __read_mostly and __exclusive_cache_line.
Reported by: Thomas Laus
Tesetd by: jkim, Thomas Laus
Fixes: r353149 ("amd64 pmap: implement per-superpage locks")
Sponsored by: The FreeBSD Foundation
starting at the max. domain, and then work down. Then existing FreeBSD
drivers will attach. Interrupt routing from the VMD MSI-X to the NVME
drive is not well known, so any interrupt is sent to all children that
register.
VROC used Intel meta data so graid(8) works with it. However, graid(8)
supports RAID 0,1,10 for read and write. I have some early code to
support writes with RAID 5. Note that RAID 5 can have life issues
with SSDs since it can cause write amplification from updating the parity
data.
Hot plug support needs a change to skip the following check to work:
if (pcib_request_feature(dev, PCI_FEATURE_HP) != 0) {
in sys/dev/pci/pci_pci.c.
Looked at by: imp, rpokala, bcr
Differential Revision: https://reviews.freebsd.org/D21383
ABI already guarantees the direction is forward. Note this does not take care
of i386-specific cld's.
Reviewed by: kib
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21906
This matches the state prior to r353149 and fixes crashes with DRM
modules.
Reported and tested by: cy, garga, Krasznai Andras
Fixes: r353149 ("amd64 pmap: implement per-superpage locks")
Sponsored by: The FreeBSD Foundation
This provides a framework to define a template describing
a set of "variables of interest" and the intended way for
the framework to maintain them (for example the maximum, sum,
t-digest, or a combination thereof). Afterwards the user
code feeds in the raw data, and the framework maintains
these variables inside a user-provided, opaque stats blobs.
The framework also provides a way to selectively extract the
stats from the blobs. The stats(3) framework can be used in
both userspace and the kernel.
See the stats(3) manual page for details.
This will be used by the upcoming TCP statistics gathering code,
https://reviews.freebsd.org/D20655.
The stats(3) framework is disabled by default for now, except
in the NOTES kernel (for QA); it is expected to be enabled
in amd64 GENERIC after a cool down period.
Reviewed by: sef (earlier version)
Obtained from: Netflix
Relnotes: yes
Sponsored by: Klara Inc, Netflix
Differential Revision: https://reviews.freebsd.org/D20477
The current 256-lock sized array is a problem in the following ways:
- it's way too small
- there are 2 locks per cacheline
- it is not NUMA-aware
Solve these issues by introducing per-superpage locks backed by pages
allocated from respective domains.
This significantly reduces contention e.g. during poudriere -j 104.
See the review for results.
Reviewed by: kib
Discussed with: jeff
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21833
Four drivers (hpt27xx, hptmv, hptnr, hptrr, hpt27xx) include precompiled
binary objects; have users load them as modules if they are needed.
Additional work (i.e., integrating devmatch) required before MFC.
Reviewed by: markj
Relnotes: Yes
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21865
pv list lock is the main bottleneck during poudriere -j 104 and
pmap_remove_pages is the most impactful consumer. It frees chunks with the lock
held even though it plays no role in correctness. Moreover chunks are often
freed in groups, sample counts during buildkernel (0-sized frees removed):
value ------------- Distribution ------------- count
0 | 0
1 | 8
2 |@@@@@@@ 19329
4 |@@@@@@@@@@@@@@@@@@@@@@ 58517
8 | 1085
16 | 71
32 |@@@@@@@@@@ 24919
64 | 899
128 | 7
256 | 2
512 | 0
Thus:
1. batch freeing
2. move it past unlocking pv list
Reviewed by: alc (previous version), markj (previous version), kib
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21832
Centralize calculation of signal and ucode delivered on unhandled page
fault in new function vm_fault_trap(). MD trap_pfault() now almost
always uses the signal numbers and error codes calculated in
consistent MI way.
This introduces the protection fault compatibility sysctls to all
non-x86 architectures which did not have that bug, but apparently they
were already much more wrong in selecting delivered signals on
protection violations.
Change the delivered signal for accesses to mapped area after the
backing object was truncated. According to POSIX description for
mmap(2):
The system shall always zero-fill any partial page at the end of an
object. Further, the system shall never write out any modified
portions of the last page of an object which are beyond its
end. References within the address range starting at pa and
continuing for len bytes to whole pages following the end of an
object shall result in delivery of a SIGBUS signal.
An implementation may generate SIGBUS signals when a reference
would cause an error in the mapped object, such as out-of-space
condition.
Adjust according to the description, keeping the existing
compatibility code for SIGSEGV/SIGBUS on protection failures.
For situations where kernel cannot handle page fault due to resource
limit enforcement, SIGBUS with a new error code BUS_OBJERR is
delivered. Also, provide a new error code SEGV_PKUERR for SIGSEGV on
amd64 due to protection key access violation.
vm_fault_hold() is renamed to vm_fault(). Fixed some nits in
trap_pfault()s like mis-interpreting Mach errors as errnos. Removed
unneeded truncations of the fault addresses reported by hardware.
PR: 211924
Reviewed by: alc
Discussed with: jilles, markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D21566
Convert all remaining references to that field to "ref_count" and update
comments accordingly. No functional change intended.
Reviewed by: alc, kib
Sponsored by: Intel, Netflix
Differential Revision: https://reviews.freebsd.org/D21768
The direct map is never used for execution of code, so we might as well
set NX in the direct map's PML4Es. Also clarify the intent of the code
in create_pagetables() that restricts access protections on the region
of the direct map mapping the kernel text.
Reviewed by: alc, kib (previous version)
MFC after: 1 week
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21759
This is required for DPCPU and VNET data variable definitions to work when
KLDs are linked as DSOs. R_X86_64_RELATIVE relocations should not appear
in object files, so assert this in elf_relocaddr().
Reviewed by: kib
MFC after: 1 month
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21755
There does not appear to be any existing need for such mappings to be
executable.
Reviewed by: alc, kib
MFC after: 1 month
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21754
Decode PAT_UNCACHED.
When unknown pat mode is encountered, print the pte bits combination
instead of the index, which is always 8.
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D21738
- Remove a dead variable from the amd64 pmap_extract_and_hold().
- Fix grammar in the vm_page_wire man page.
Reported by: alc
Reviewed by: alc, kib
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21639
- Remove a dead variable from the amd64 pmap_extract_and_hold().
- Fix grammar in the vm_page_wire man page.
Reported by: alc
Reviewed by: alc, kib
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21639
Just return EINVAL if flags != 0. The Linux man page documents one
case of EINVAL as "The filesystem does not support one of the flags in
flags."
After r351723 userland binaries will try using new system calls.
Reported by: mjg
Reviewed by: mjg, trasz
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21590
There are several mechanisms by which a vm_page reference is held,
preventing the page from being freed back to the page allocator. In
particular, holding the page's object lock is sufficient to prevent the
page from being freed; holding the busy lock or a wiring is sufficent as
well. These references are protected by the page lock, which must
therefore be acquired for many per-page operations. This results in
false sharing since the page locks are external to the vm_page
structures themselves and each lock protects multiple structures.
Transition to using an atomically updated per-page reference counter.
The object's reference is counted using a flag bit in the counter. A
second flag bit is used to atomically block new references via
pmap_extract_and_hold() while removing managed mappings of a page.
Thus, the reference count of a page is guaranteed not to increase if the
page is unbusied, unmapped, and the object's write lock is held. As
a consequence of this, the page lock no longer protects a page's
identity; operations which move pages between objects are now
synchronized solely by the objects' locks.
The vm_page_wire() and vm_page_unwire() KPIs are changed. The former
requires that either the object lock or the busy lock is held. The
latter no longer has a return value and may free the page if it releases
the last reference to that page. vm_page_unwire_noq() behaves the same
as before; the caller is responsible for checking its return value and
freeing or enqueuing the page as appropriate. vm_page_wire_mapped() is
introduced for use in pmap_extract_and_hold(). It fails if the page is
concurrently being unmapped, typically triggering a fallback to the
fault handler. vm_page_wire() no longer requires the page lock and
vm_page_unwire() now internally acquires the page lock when releasing
the last wiring of a page (since the page lock still protects a page's
queue state). In particular, synchronization details are no longer
leaked into the caller.
The change excises the page lock from several frequently executed code
paths. In particular, vm_object_terminate() no longer bounces between
page locks as it releases an object's pages, and direct I/O and
sendfile(SF_NOCACHE) completions no longer require the page lock. In
these latter cases we now get linear scalability in the common scenario
where different threads are operating on different files.
__FreeBSD_version is bumped. The DRM ports have been updated to
accomodate the KPI changes.
Reviewed by: jeff (earlier version)
Tested by: gallatin (earlier version), pho
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D20486
- Use ptoa() instead of the archaic ctob().
- Use pagezero() to zero a PDP page.
- Remove PA_MIN_ADDRESS, orphaned by r351742.
- Remove unneeded parens and an unnecessary control flow statement.
Reported by: alc
Reviewed by: alc, kib
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D21495
from recent Ubuntu versions. Without it they segfault on startup.
Reviewed by: emaste
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D20687
