o Split the compression across several worker threads. By default, "several"
matches number of CPUs, capped at 24 for sanity when running on a very big
hardwares. Provide option to set that number manually;
o Fix bug inherited from the mkulzma (R.I.P) which degraded already slow LZMA
compression even further by calling function to release compression state
after processing each block.
It is neither documented as required nor actually required by the LZMA
library. This caused spree of system calls to release memory and then map
it again for every block. LZMA compression is more than 2x faster after this
change alone;
o Record time it takes to do compression and report throughput achieved.
o Add simple first-level 256 entry hash table for de-dup code, so it's not
becoming a bottleneck at big files.
The purpose of this file was to simply detect the UART speed before
attaching the actual ns8250 driver so that we don't have to specify the
UART speed in DTS files.
However, OpenWRT DTS files specify ns16550a as a compatible string in
their DTS files and this makes the original ns8250 driver attach to
the device. So we would have to edit the DTS files anyway and since this
is only the case for MT7621 and MT7628/MT7688 for now, it's better to
just add the clock-frequency property to those (UART is always clocked
by the same clock in both these SoCs, so that's fine) instead of having
a separate driver and still having to change the DTS files.
Approved by: adrian (mentor)
Sponsored by: Smartcom - Bulgaria AD
Differential Revision: https://reviews.freebsd.org/D6044
The ordering of acquisition of the state and session mutexes was
reversed in two cases executed when an NFSv4.1 client created/freed
a session. Since clients will typically do this only when mounting
and dismounting, the likelyhood of causing a deadlock was low but possible.
This can only occur for NFSv4.1 mounts, since the others do not
use sessions.
This was detected while testing the pNFS server/client where the
client crashed during dismounting.
The patch also reorders the unlocks, although that isn't necessary
for correct operation.
- Added check that the SCOPE ID is only restored for IPv6 linklocal
addresses.
- Changes made by r237263 in the "cma_bind_addr()" function did not
check if the socket address was of type IPv6 and used the IPv4
socket address for IPv6 addresses. This caused the function to
fail. Fixed this.
- In the "rdma_gid2ip()" function and some other places the "sin6_len"
and "sin6_scope_id" fields were not set for IPv6 socket
addresses. Fixed this.
- The scope ID is not stored as part of the GID entries and must be
passed as an argument to "rdma_gid2ip()".
- Added new method to "struct ib_device" which returns a pointer to
the network interface which belongs to the given infiniband
device. This is needed to be able to get the scope ID for IPv6
addresses via the associated ethernet interface.
- Added convenience function, "rdma_get_ipv6_scope_id()", to get the
scope ID for IPv6 addresses.
- Implemented new "get_netdev" method for mlx4ib. Other IB controller
drivers which want to support IPv6 addresses needs to implement this
aswell.
- Bumped the FreeBSD version due to changing "struct ib_device".
In particular,
- avoid dereferencing NULL pointers
- test pointers against NULL, not 0
- test for errout == NULL in the top-level functions (kvm_open, kvm_openfiles,
kvm_open2, etc)
- Replace a realloc and free on failure with reallocf
adrian [Fri, 22 Apr 2016 16:26:53 +0000 (16:26 +0000)]
[bhnd] Add a common bhnd_pci driver shared by both bhnd_pcib and bhnd_pci_hostb
This extracts common code from bhndb_pci, bhnd_pcib, and bhnd_pci_hostb into a
simpler shared bhnd_pci base driver, and should enable SoC-side implementation
of bhnd_pcib root complex support.
While the instructions were not included into the original instruction
set, their support can be indicated by a special feature bit.
For example:
CPU: AMD Phenom(tm) II X4 955 Processor (3214.71-MHz K8-class CPU)
...
AMD Features2=0x37ff<LAHF, ...>
Clang 3.8 uses lahf/sahf as a faster alternative to pushf/popf where
possible.
https://www.illumos.org/issues/6052
At the moment type parameter of lzc_create() is of dmu_objset_type_t type.
That exposes an implementation detail and requires sys/fs/zfs.h to be included
in libzfs_core.h creating unnecessary coupling between libzfs_core interface
and ZFS internals.
I think that dmu_objset_type_t should be replaced with a libzfs_core
enumeration of supported dataset types.
For ABI reasons the new enumeration could be bit-compatible with
dmu_objset_type_t.
For example:
typedef enum {
LZC_DST_ZFS = 2,
LZC_DST_ZVOL
} lzc_dataset_type_t;
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: Andriy Gapon <andriy.gapon@clusterhq.com>
Return `ret` in op_ifentry(..) to mute a -Wunused-but-set-variable warning
This will also now detect error conditions with
value->var.subs[sub - 1] == LEAF_ifPhysAddress where `string_get(..)`
could fail if iifp->physaddr and/or iifp->physaddrlen were deemed
invalid.
Fix for printf() compile warning when fast_reg.length is 64-bit.
Changing fast_reg.length to 64 bits is planned in the future. Krping
uses 32-bit lengths internally.
Fix duplicate TLB entries issue during section promotion/demotion.
Such situation is defined as UNPREDICTABLE by arm arm manual.
This patch fixes all explicit TLB fetches which could cause this issue
and speculative TLB fetches for sections mapped in user address space.
Speculative TLB fetches for sections mapped in kernel address space are
not fixed yet as the break-before-make approach must be implemented for
kernel mappings too. This means that promoted/demoted section will be
unmapped for a while. Either kernel stack the promotion/demotion is
being done on or L1 page table(s) which must be modified may be mapped
by this section. Thus the fix will not be so simple like for userland
mappings.
The issue was detectable only on Cortex-A8 platforms and only very
rarely. It was reported few times. First, it was by Mikael Urankar
in June 2015. He helped to identify the mechanism of this issue, but
we were not sure how to fix it correctly until now.
PR: 208381
Reported by: Mikael Urankar (mikael.urankar at gmail.com)
Reviewed by: kib
Don't use atomic operations for page table entries and handle access
and R/W emulation aborts under pmap lock.
There were two reasons for using of atomic operations:
(1) the pmap code is based on i386 one where they are used,
(2) there was an idea that access and R/W emulation aborts should be
handled as quick as possible, without pmap locking.
However, the atomic operations in i386 pmap code are used only because
page table entries may be modified by hardware. At the beginning, we
were not sure that it's the only reason. So even if arm hardware does
not modify them, we did not risk to not use them at that time. Further,
it turns out after some testing that using of pmap lock for access and
R/W emulation aborts does not bring any extra cost and there was no
measurable difference. Thus, we have decided finally to use pmap lock
for all operations on page table entries and so, there is no reason for
atomic operations on them. This makes the code cleaner and safer.
This decision introduce a question if it's safe to use pmap lock for
access and R/W emulation aborts. Anyhow, there may happen two cases in
general:
(A) Aborts while the pmap lock is locked already - this should not
happen as pmap lock is not recursive. However, under pmap lock only
internal kernel data should be accessed and such data should be mapped
with A bit set and NM bit cleared. If double abort happens, then
a mapping of data which has caused it must be fixed.
(B) Aborts while another lock(s) is/are locked - this already can
happen. There is no difference here if it's either access or R/W
emulation abort, or if it's some other abort.
Add four functions which check a virtual address for stage 1 privileged
(PL1) and unprivileged (PL0) read/write access. As cp15 virtual to
physical address translation operations are used, interrupts must be
disabled to get consistent result when they are called.
These functions should be used only in very specific occasions like
during abort handling or kernel debugging. One of them is going to be
used in pmap_fault(). However, complete function set is added. It cost
nothing, as they are inlined.
- xalloc(..) ensures that e will be non-null via malloc + err.
- `e` is already dereferenced above, so logically it's impossible
to hit the lower test without crashing if it was indeed NULL.
Submitted by: Jun Su <junsu microsoft com>
Reviewed by: sephe, Dexuan Cui <decui microsoft com>
MFC after: 1 week
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D5957
sys: use our roundup2/rounddown2() macros when param.h is available.
rounddown2 tends to produce longer lines than the original code
and when the code has a high indentation level it was not really
advantageous to do the replacement.
This tries to strike a balance between readability using the macros
and flexibility of having the expressions, so not everything is
converted.
Adjust the fdc worker thread startup to work when APs are started earlier.
- Enable the commented out locking in fd_probe(). The worker thread
should not be running yet (even after these changes), but better to be
safe than sorry.
- Defer starting the worker thread until after the child drives have been
probed. The worker thread startup is moved into a fdc_start_worker()
thread that the various front ends call at the end of attach. As a
side effect this fixes a few edge cases that weren't shutting down the
worker thread if attach encountered a late failure.
- When executing the initial reset requested by attach in the worker
thread, use DELAY() instead of a tsleep() if cold is set.
Tested by: Howard Su <howard0su@gmail.com>
Sponsored by: Netflix
Queue the CPU-probing task after all acpi_cpu devices are attached.
Eventually with earlier AP startup this code will change to call the
startup function synchronously instead of queueing the task. Moving
the time we queue the task should be a no-op since taskqueue threads
don't start executing tasks until much later, but this reduces the diff
with the earlier AP startup patches.
IPv6 addresses has a scope ID which sometimes is stored in the
"sin6_scope_id" field of "struct sockaddr_in6" and sometimes as part
of the IPv6 address itself depending on the context. If the scope ID
is not in the expected location, the IPv6 address lookups in the
so-called GID table will fail. Some code factoring has been made to
achieve a clean exit of the "addr_resolve" function via a common
"done" label.
Get rid of rctl_lock; use racct_lock where appropriate. The fast paths
already required both of them, so having a separate rctl_lock didn't
buy us anything.
Reviewed by: mjg@
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D5914
Prevent underflows in tp->snd_wnd if the remote side ACKs more than
tp->snd_wnd. This can happen, for example, when the remote side responds to
a window probe by ACKing the one byte it contains.
ascii.c was removed during r290494 but this introduced an issue with ASCII that
has been fixed in r290618 and lead to a rewrite of ascii.c based on none.c.
ascii.c was wrongly reintroduced in r290620 without proper svn operation which
lost the history.
net80211: enable promiscuous mode state change for non-monitor/ahdemo modes
- Allow to enable/disable promiscuous mode when:
* interface is not a member of bridge, or;
* request was issued by user (ifconfig wlan0 promisc), or;
* interface is in MONITOR or AHDEMO mode.
- Drop local workarounds in mwl(4) and malo(4).
Bring a little more compability with GNU units 2.12
- notionally support a 'history file' flag. This doesn't do much now,
but is there to prevent scripts written against GNU units from
breaking
- correctly gracefully quit rather than exit (this will make it easier
to support a history file in the future)
- remove the "t" flag from fopen which was there to support windows. We
have not supported windows since at the latest, the introduction of
capsicum.
dhclient: Log a warning instead of bailing upon "illegal" options
In Azure, the DHCP servers add private option (id 0xf5), which contains
binary form of an IPv4 address. Once this option is converted to string
form, it could contain '$', e.g.
dhclient bails upon "illegal" options like the above example, thus the
VM bring-up will fail.
Also as a side note, this "illegal" option detection was added in
OpenBSD ~11years ago:
http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sbin/dhclient/dhclient.c?rev=1.50&content-type=text/x-cvsweb-markup
And it was removed along with the removal of script support in OpenBSD
~3years ago:
http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sbin/dhclient/dhclient.c?rev=1.159&content-type=text/x-cvsweb-markup
Reported by: Hongxiong Xian <v-hoxian microsoft com>
Reviewed by: jhb, Dexuan Cui <decui microsoft com>
Tested by: Hongxiong Xian <v-hoxian microsoft com>
Analyzed by: Dong Liu <doliu microsoft com>
MFC after: 1 week
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D5853