Add a loader tunable, vfs.zfs.arc_meta_min, which controls how much metadata
ZFS should keep in ARC at minimum.
In arc_evict(), when doing recycle, take more factors into account by
applying the following policy:
1. If no evictable data, evict metadata;
2. If no evictable metadata, evict data;
3. If we hit arc_meta_limit, evict metadata;
4. If we haven't hit arc_meta_min, evict data;
5* (Illumos only, not present in new FreeBSD code, yet) evict the oldest
cached element from data and metadata.
(FreeBSD) evict the data type specified by caller, which is the
existing behavior.
Note that because of our splitted locks (implemented in r205231 to improve
scalability by reducing lock contention), implementing the fifth Illumos
behavior will not be cheap, so for now just implement the 1-4 and fall back
to current behavior for 5.
Illumos issue:
5368 ARC should cache more metadata
Provide vnode in memory map info for files on tmpfs
When providing memory map information to userland, populate the vnode pointer
for tmpfs files. Set the memory mapping to appear as a vnode type, to match
FreeBSD 9 behavior.
This fixes the use of tmpfs files with the dtrace pid provider,
procstat -v, procfs, linprocfs, pmc (pmcstat), and ptrace (PT_VM_ENTRY).
Submitted by: Eric Badger <eric@badgerio.us> (initial revision)
Obtained from: Dell Inc.
PR: 198431
dmesg: accommodate message buffer growth between the sysctl calls
Allocate 12.5% extra space to avoid ENOMEM when the message buffer
is growing steadily.
Reported by: Steve Wahl <steve_wahl@dell.com> (and tested)
Sponsored by: Dell Inc.
Always send log(9) messages to the message buffer.
It is truer to the semantics of logging for messages to *always*
go to the message buffer, where they can eventually be collected
and, in fact, be put into a log file.
This restores the behavior prior to r70239, which seems to have
changed it inadvertently.
Submitted by: Eric Badger <eric@badgerio.us>
Obtained from: Dell Inc.
pmcstat.8: fix -a flag description; improve -m flag to match
The -a flag reads a file saved by -O, not -o.
The -m flag requires the -R flag. Copy that paragraph from -a.
Sponsored by: Dell Inc.
Clean up some cosmetic nits in kern_umtx.c, found during recent work
in this area and by the Clang static analyzer.
Remove some dead assignments.
Fix a typo in a panic string.
Use umtx_pi_disown() instead of duplicate code.
Use an existing variable instead of curthread.
Approved by: kib (mentor until recently)
Sponsored by: Dell Inc.
In addition to the ubldr file, also copy ubldr.bin to the
MS-DOS partition. This will help with transitioning to
a single arm/armv6 userland build which could be used for
all FreeBSD/armv6 images without UBLDR_LOADADDR being set
for each board (ultimately requiring a separate buildworld
for each currently).
Sponsored by: The FreeBSD Foundation
Add more mmap tests related to character devices.
- Add cdev-related tests for bad args.
- Add two simple tests cases for mapping /dev/zero that test for
MAP_ANON-like behavior.
Add various tests to ensure that invalid arguments passed to mmap()
trigger failures.
Note: most of the tests that should provoke an EINVAL error do not
pass on stable/10 as stable/10 does not have the changes that added
more strict parameter checking to mmap(). Those changes will not be
merged to stable/10, so I have disabled them via #if 0 with a comment
explaining why.
kgdb uses td_oncpu to determine if a thread is running and should use
a pcb from stoppcbs[] rather than the thread's PCB. However, exited threads
retained td_oncpu from the last time they ran, and newborn threads had their
CPU fields cleared to zero during fork and thread creation since they are
in the set of fields zeroed when threads are setup. To fix, explicitly
update the CPU fields for exiting threads in sched_throw() to reflect the
switch out and reset the CPU fields for new threads in sched_fork_thread()
to NOCPU.
Handle X2APIC entries in the MADT for APICs with an ID < 255. At least one
BIOS has been seen to include such entries even though the relevant specs
require that X2APIC entries only be used for CPUs with an APIC ID >= 255.
This was tested on a system with "plain" local APIC entries in the MADT
to ensure no regressions, but it has not yet been tested on a system with
X2APIC entries in the MADT. Currently such systems do not boot at all,
and with this change they might now boot correctly.
Export a list of VM objects in the system via a sysctl. The list can be
examined via 'vmstat -o'. It can be used to determine which files are
using physical pages of memory and how much each is using.
r288341:
Honor VMFORMATS and VMSIZE if set in release.conf. [1]
r288345:
In followup to r288341, ensure VMFORMATS and VMSIZE are not
set to empty values, which would result in nonintuitive build
errors.
r288347:
Append VMFORMATS and VMSIZE to RELEASE_RMAKEFLAGS only if
WITH_VMIMAGES is set.
PR: 203420 [1]
Sponsored by: The FreeBSD Foundation
Update vmstat usage for last-argument count/wait parameters
Correct the usage in both the manpage and in usage() to indicate
that the wait interval and repetition count may be given either
with the respective -w/-c arguments, or as the final positional
arguments. [0]
The corresponding code to implement the positional arguments has
been conditional on the (always-enabled) BACKWARD_COMPATIBILITY
macro since the original 4.4-lite import. It's no longer reasonable
to remove the functionality, so remove the macro and conditional
instead.
Note that multiple disks may be given on the command line.
While here, sort arguments and apply minor mdoc fixes.
PR: 184755 [0]
Use the cpuset API more consistently:
- Fetch the root set from cpuset_getaffinity() instead of assuming all CPUs
from 0 to hw.ncpu are the root set.
- Use CPU_SETSIZE and CPU_FFS.
- The original notion of halted CPUs the manpage and code refers to is gone.
Use the term "available" instead.
transport specific form of a universal transport address. The
structure is expected to be opaque to consumers. In the current
implementation, the structure contains a pointer to a buffer
that holds the actual address.
In rpcbind(8), netbuf structures are copied directly, which would
result in two netbuf structures that reference to one shared
address buffer. When one of the two netbuf structures is freed,
access to the other netbuf structure would result in an undefined
result that may crash the rpcbind(8) daemon.
Fix this by making a copy of the buffer that is going to be freed
instead of doing a shallow copy.
Security: FreeBSD-SA-15:24.rpcbind
Security: CVE-2015-7236
There is no reason to perform the pmap_remove() on the kernel pmap while
the kmem object lock is held. Do the pmap_remove() before acquiring the
kmem object lock.
Correct a non-fatal error in vm_pageout_worker(). vm_pageout_worker()
should not assume that vm_pages_needed will remain set while it sleeps.
Other threads can clear vm_pages_needed by performing a sufficient
number of vm_page_free() calls, e.g., process termination. The effect
of this error was that vm_pageout_worker() would free and/or launder
pages when, in fact, there was no shortage of free pages.
Rewrite a nearby comment to describe all of the possible cases and not
just the most common case. The problem being that the comment made
the most common case seem like the only case.
The intention of r254304 was to scan the active queue continuously.
However, I've observed the active queue scan stopping when there are
frequent free page shortages and the inactive queue is steadily refilled
by other mechanisms, such as the sequential access heuristic in vm_fault()
or madvise(2). To remedy this problem, record the time of the last active
queue scan, and always scan a number of pages proportional to the time
since the last scan, regardless of whether that last scan was a
timeout-triggered ("pass == 0") or free-page-shortage-triggered ("pass >
0") scan.
Also, on a timeout-triggered scan, allow a full scan of the active queue
when the system is short of inactive pages.
Revise the text about the atomicity of the defined operations across
multiple processors. In particular, clearly state that the operations
are always atomic when they are applied to the default memory type
that is used by the kernel (and applications).
Stop describing an acquire operation as a read barrier and a release
operation as a write barrier. That description has never been correct,
and it has caused confusion.
Also, explicitly say that a thread doesn't see its own accesses being
reordered. The reordering of a thread's accesses is only (potentially)
visible to another thread.
