Alexander Motin [Sat, 16 Apr 2016 06:36:56 +0000 (06:36 +0000)]
Simplify memory allocation for NS requests.
Since we no longer need additional buffers for request and response IOCBs,
we can increase receive space by 192 bytes, that is enough for fetching 48
more ports. The new limit is 1020 fabric ports per virtual port.
John Baldwin [Sat, 16 Apr 2016 00:01:16 +0000 (00:01 +0000)]
Add a test for cancelling an active AIO request on a socket.
The older AIO code awakened all pending AIO requests on a socket
when any data arrived. This could result in AIO daemons blocking on
an empty socket buffer. These requests could not be cancelled
which led to a deadlock during process exit. This test reproduces
this case. The newer AIO code is able to cancel the pending AIO
request correctly.
Ed Maste [Fri, 15 Apr 2016 19:06:36 +0000 (19:06 +0000)]
elfcopy: fix symbol table handling when sections come after symtab/strtab
Fix a symbol table handling bug in elfcopy: elfcopy puts .symtab,
.strtab and .shstrtab sections in the end of the output object. If
the input objects have more sections after any of these 3 sections,
the section table will be reordered, and in that case the section
symbols should be regenerated for relocations.
The bug is triggered since newer clang puts .strtab section in the
beginning of the object produced.
Phil Shafer [Fri, 15 Apr 2016 18:46:15 +0000 (18:46 +0000)]
Import to 0.6.1
0.5.0:
document "trim" modifier
add xo_emit_field functions
Add xo_set_file{,_h} functions
Fix LIBXO_* variables; add -L and -I as needed
add --disable-silent-rules and an explicit make; s/PACKAGE-NAME/PACKAGE_NAME/; add /download/ to 'url'
fix silliness where xo_flush_h emitted closing tag (html); make the caller (xo_message) do it
flush after transitions; fix flush call in xo_do_emit
mkdir the version-specific packaging dir
use "XO_" instead of LIBXO_
0.6.0:
Add --with-retain-size to set the size (in bits) of the retain hash buckets
Add The Argument Modifier ({a:})
Add retain and no-retain to --libxo
autoconf: Add test for monitor.h
Document quote heuristic
go deep with nroff backslashes
Use "ULL" for 32 bit check
add xo_retain_clear and xo_retain_clear_all
docs: combine two 'handles' section; move command line argument section
handle GETTEXT when msgfmt isn't where it's supposed to be (FreeBSD)
make 'retain' a flag (XOEF_RETAIN) instead of a role; it's simpler, and doesn't feel as tacky. "{R:}" was painful to document, which means it's painful to use.
new xo_emit_f functions
nuke some unused UNUSEDs
test code: path must be static
update test cases
0.6.1:
fix version number (missed a commit during new-release)
Make sure all I/O to the dump device is of the native sector size. To
that end, we keep a native sector sized buffer associated with dump
devices (di->blockbuf) and use it to pad smaller objects as needed (e.g.
kerneldumpheader).
Add dump_write_pad() as a convenience API to dump smaller objects with
zero padding. (Rather than pull in NPM leftpad, we wrote our own.)
Savecore(1) has been updated to deal with these dumps. The format for
512-byte sector dumps should remain backwards compatible.
Minidumps for other architectures are left as an exercise for the
reader.
Alan Somers [Fri, 15 Apr 2016 16:36:17 +0000 (16:36 +0000)]
Don't corrupt ZFS label's physpath attribute when booting while a disk is missing
Prior to this change, vdev_geom_open_by_path would call vdev_geom_attach
prior to verifying the device's GUIDs. vdev_geom_attach calls
vdev_geom_attrchange to set the physpath in the vdev object. The result is
that if the disk could not be found, then the labels for other disks in the
same TLD would overwrite the missing disk's physpath with the physpath of
whichever disk currently has the same devname as the missing one used to
have.
Phil Shafer [Fri, 15 Apr 2016 15:50:13 +0000 (15:50 +0000)]
import libxo-0.4.7
Fix bug w/ {e:} in html, where no default encoding format was built
docs: "t" == "trim" (typo) (cf svn commit: r290445 - head/contrib/libxo/libxo)
Import Mediatek/Ralink dtsi patches against OpenWRT dtsi files
This revision suggests dtsi patches to be used with the original OpenWRT
dtsi files so we can re-use what has already been done in OpenWRT for the
Mediatek/Ralink SoCs.
The only thing that is required after importing this revision should be
the following:
1. Import OpenWRT dts/dtsi files into sys/gnu/dts/mips
2. Run the following script in sys/gnu/dts/mips:
for f in `ls [mr]t*.dtsi`; do
printf "\n#include <fbsd-$f>\n" > $f
done
This will apply our dtsi patches to OpenWRT's dtsi files and will allow us
to re-use dts/dtsi files for ~170 Mediatek/Ralink boards.
Currently our drivers are not 100% compatible with OpenWRT's dts files, but
they're compatible enough.
We can add more functionality in the future that would better leverage the
OpenWRT work as well.
Approved by: adrian (mentor)
Sponsored by: Smartcom - Bulgaria AD
Differential Revision: https://reviews.freebsd.org/D5965
fdt_static_dtb.S dependency in sys/conf/files is currently set as:
$S/boot/fdt/dts/${MACHINE}/${FDT_DTS_FILE}
This is wrong, as what fdt_static_dtb.S actually uses is the DTB file
produced from the FDT_DTS_FILE.
In addition it also makes using DTS files stored in $S/gnu/dts/${MACHINE}/
impossible.
So, change the dependency to "fdt_dtb_file", which seems to be the right
option here anyway.
Approved by: adrian (mentor)
Sponsored by: Smartcom - Bulgaria AD
Differential Revision: https://reviews.freebsd.org/D5963
Mediatek/Ralink: Get our drivers closer to OpenWRT dts definitions
This revision gets our Mediatek/Ralink drivers closer to OpenWRT's dts
definitions, so we can reuse them with less modifications later in order
to bring support for a lot of boards at once.
Approved by: adrian (mentor)
Sponsored by: Smartcom - Bulgaria AD
Differential Revision: https://reviews.freebsd.org/D5961
In order to build a kernel with one of these configs the user should do
the following:
1. Give the appropriate board dts file to be used by either:
1.1. edit the SoC kernel config required (e.g., MT7620A_FDT) and include
the required FDT_DTS_FILE makeoption; or
1.2. simply supply FDT_DTS_FILE="xx.dts" on the command line when building
the kernel
Of course, the user can also create a completely new kernel config to
match the desired board and include the SoC kernel config from within
it.
If required, edit the MEDIATEK config file, which includes optional
drivers and comment out the unneeded ones.
2.1. this would only make sense if kernel size is a concern. Even if we
build the kernel with all drivers, if we lzma it and package it as a uImage,
its size is still around 1.1MiB.
The user will have to choose a dts file (or create a new one) from
sys/gnu/dts/mips , where all Mediatek/Ralink dts files will be imported via
a later revision.
Approved by: adrian (mentor)
Sponsored by: Smartcom - Bulgaria AD
Differential Revision: https://reviews.freebsd.org/D5966
Always calculate divisor for the counter mode of LAPIC timer. Even if
initially configured in the TSC deadline mode, eventtimer subsystem
can be switched to periodic, and then DCR register is loaded with
unitialized value.
Reset the LAPIC eventtimer frequency and min/max periods when changing
between deadline and counted periodic modes.
Reported and tested by: Vladimir Zakharov <zakharov.vv@gmail.com>
Sponsored by: The FreeBSD Foundation
Andrew Turner [Fri, 15 Apr 2016 14:28:34 +0000 (14:28 +0000)]
Add a flag field to struct gic_irqsrc and use it to mark when we should
write to the End of Interrupt (EOI) register before handling the interrupt.
This should be a noop as it will be set for all edge triggered interrupts,
however this will not be the case for MSI interrupts. These are also edge
triggered, however we should not write to the EOI register until later in
arm_gic_pre_ithread.
Obtained from: ABT Systems Ltd
Sponsored by: The FreeBSD Foundation
Andrew Turner [Fri, 15 Apr 2016 14:19:25 +0000 (14:19 +0000)]
Add initial GICv2m support to the arm GIC driver. This will be used to
support MSI and MSI-X interrupts, however intrng needs updates before this
can happen.
For now we just attach the driver until the MSI API is ready.
Obtained from: ABT Systems Ltd
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D5950
Revert r292255 because it can create bounced regions without contiguous
page offsets, which is needed for USB devices.
Another solution would be to force bouncing the full buffer always (even
when only one page requires bouncing), but this seems overly complicated and
unnecessary, and it will probably involve using more bounce pages than the
current code.
Warner Losh [Fri, 15 Apr 2016 05:10:31 +0000 (05:10 +0000)]
Out of an abundance of caution treat
* Samsung 843T Series SSDs (MZ7WD*)
* Samsung PM851 Series SSDs (MZ7TE*)
* Samsung PM853T Series SSDs (MZ7GE*)
as known having broken NCQ TRIM support as they appear to be based on
the same controller technology as the 840 and 850 series.
I've had at least one report of the PM853 being broken, so err on the
side of caution for the above drives. The PM863/SM863 appears to be
based on a newer controller, so give it the benefit of the doubt.
fgetln(3) will returns NULL if cannot get a line from a stream.
strsep(3) it will returns NULL if the end of the string was reached.
jemalloc(3) malloc will returns NULL if it cannot allocate memory.
fgetln(3) it will returns NULL if it cannot get a line from a stream.
John Baldwin [Fri, 15 Apr 2016 03:42:12 +0000 (03:42 +0000)]
Add a new PCI bus interface method to alloc the ivars (dinfo) for a device.
The ACPI and OFW PCI bus drivers as well as CardBus override this to
allocate the larger ivars to hold additional info beyond the stock PCI ivars.
This removes the need to pass the size to functions like pci_add_iov_child()
and pci_read_device() simplifying IOV and bus rescanning implementations.
As a result of this and earlier changes, the ACPI PCI bus driver no longer
needs its own device_attach and pci_create_iov_child methods but can use
the methods in the stock PCI bus driver instead.
fopen(3) returns NULL in case it can't open the STREAM.
fgetln(3) returns NULL if it can't get a line from a STREAM.
malloc returns NULL if it can't allocate memory.
Warner Losh [Fri, 15 Apr 2016 03:10:04 +0000 (03:10 +0000)]
Add FCCT M500 to the NCQ black list. Linux added it in 4.2 (August
2015). Correct the M500 firmware versions. EU07 was the engineering
test version, not the release version with the fix. MU07 is the
release version. It's the only Micron firmware version to actually
work. Remove support for EU07.
This brings the blacklist into parity with the Linux blacklist as of
4.5, except for the Micron M500 MU07 entry. I personally tested the
MU07 firmware on 12 machines running 6 drives each with no corruption
in the past 6 months with Netflix production loads. Prior versions of
the M500 firmware wouldn't last more than a few days.
hyperv: Deprecate HYPERV option by moving Hyper-V IDT vector into vmbus
Submitted by: Jun Su <junsu microsoft com>
Reviewed by: jhb, kib, sephe
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D5910
It allows implementing loadable kernel modules with new actions and
without needing to modify kernel headers and ipfw(8). The module
registers its action handler and keyword string, that will be used
as action name. Using generic syntax user can add rules with this
action. Also ipfw(8) can be easily modified to extend basic syntax
for external actions, that become a part base system.
Sample modules will coming soon.
Enji Cooper [Thu, 14 Apr 2016 22:22:03 +0000 (22:22 +0000)]
Add DEBUG_FLAGS to PROG_VARS and STRIP to PROG_OVERRIDE_VARS
This will allow the variables [*] to be overridden on a per-PROG basis,
which is useful when controlling "stripping" behavior for some tests
that require debug symbols or to be unstripped
DEBUG_FLAGS (similar to CFLAGS) supports appending, whereas STRIP is
an override
*: Due to how STRIP is defined in bsd.own.mk (in addition to
bsd.lib.mk and bsd.prog.mk), and the fact that bsd.test.mk pulls in
bsd.own.mk first, overriding STRIP doesn't work today.
A follow up commit is pending to "rectify" this after additional
testing is done.
When using meta mode with filemon, the build is reliably incremental
safe. Bmake will use the meta files, along with filemon information,
to rebuild targets when their dependencies change, commands change,
or files they generate are missing.
Scott Long [Thu, 14 Apr 2016 21:55:42 +0000 (21:55 +0000)]
Add a devctl/devd notification conduit for CAM errors that happen at the
periph level. When a relevant error is reported to the periph, some
amplifying information is gathered, and the error and information are fed
to devctl with the attributes / keys system=CAM, subsystem=periph. The
'type' key will be either 'error' or 'timeout', and based on this, various
other keys are also populated.
The purpose of this is to provide a concise mechanism for error reporting
that is less noisy than the system console but higher in resolution and
fidelity than simple sysctl counters. We will be using it at Netflix to
populate a structured log and database to track errors and error trends
across our world-wide population of drives.
Warner Losh [Thu, 14 Apr 2016 21:47:58 +0000 (21:47 +0000)]
New CAM I/O scheduler for FreeBSD. The default I/O scheduler is the same
as before. The common scheduling bits have moved from inline code in
each of the CAM periph drivers into a library that implements the
default scheduling.
In addition, a number of rate-limiting and I/O preference options can
be enabled by adding CAM_IOSCHED_NETFLIX to your config file. A number
of extra stats are also maintained. CAM_IOSCHED_NETFLIX isn't on by
default because it uses a separate BIO_READ and BIO_WRITE queue, so
doesn't honor BIO_ORDERED between these two types of operations. We
already didn't honor it for BIO_DELETE, and we don't depend on
BIO_ORDERED between reads and writes anywhere in the system (it is
currently used with BIO_FLUSH in ZFS to make sure some writes are
complete before others start and as a poor-man's soft dependency in
one place in UFS where we won't be issuing READs until after the
operation completes). However, out of an abundance of caution, it
isn't enabled by default.
Plus, this also brings in NCQ TRIM support for those SSDs that support
it. A black list is also provided for known rogues that use NCQ trim
as an excuse to corrupt the drive. It was difficult to separate out
into a separate commit.
This code has run in production at Netflix for over a year now.
o Teach opcode rewriting framework handle several rewriters for
the same opcode.
o Reduce number of times classifier callback is called. It is
redundant to call it just after find_op_rw(), since the last
does call it already and can have all results.
o Do immediately opcode rewrite in the ref_opcode_object().
This eliminates additional classifier lookup later on bulk update.
For unresolved opcodes the behavior still the same, we save information
from classifier callback in the obj_idx array, then perform automatic
objects creation, then perform rewriting for opcodes using indeces
from created objects.
META_MODE: Pass along the default sysroot in bootstrap-tools to avoid rebuilds later.
Some of the clang libraries build in this phase and the cross-tools
phase. Later in the cross-tools phase when they build with a default
TOOLS_PREFIX, they see a changed build command in meta mode due to
the changed DEFAULT_SYSROOT. This is avoided by passing along
TOOLS_PREFIX earlier.
META_MODE: Don't rebuild build-tools targets during normal build.
This avoids 'build command changed' due to CFLAGS/CC changes during the
normal build. Without this the build-tools targets end up rebuilding
for the *target* rather than keeping the native versions built in
build-tools.
Rework META_TARGETS so that it automatically adds META_DEPS to the targets.
This will only be done if the target is defined, so if the target is
defined after bsd.sys.mk is included then it needs to manually add
${META_DEPS} still.
Alexander Motin [Thu, 14 Apr 2016 20:49:01 +0000 (20:49 +0000)]
Extract virtual port address from RQSTYPE_RPT_ID_ACQ.
This should close the race between request arriving on new target mode
virtual port and its scanner thread finally fetch its address for request
routing.
Michael Tuexen [Thu, 14 Apr 2016 19:59:21 +0000 (19:59 +0000)]
Allow the handling of ICMP messages sent in response to SCTP packets
containing an INIT chunk. These need to be handled in case the peer
does not support SCTP and returns an ICMP messages indicating destination
unreachable, protocol unreachable.