Pedro F. Giffuni [Mon, 30 Jun 2014 20:54:25 +0000 (20:54 +0000)]
regex(3): Add support for \< and \> word delimiters
Solaris and other OSs have support for \< and \> as word
delimiters in utilities like sed(1). These are useful to
have for general compatiblity with Solaris but should be
avoided for portability with other systems, including the
traditional BSDs.
Bump __FreeBSD_version as this is likely to affect some
userland utilities.
Adrian Chadd [Mon, 30 Jun 2014 04:38:29 +0000 (04:38 +0000)]
Add initial RSS awareness to the ixgbe(4) driver.
The ixgbe(4) hardware is capable of RSS hashing RX packets and doing RSS
queue selection for up to 8 queues.
However, even if multi-queue is enabled for ixgbe(4), the RX path doesn't use
the RSS flowid from the received descriptor. It just uses the MSIX queue id.
This patch does a handful of things if RSS is enabled:
* Instead of using a random key at boot, fetch the RSS key from the RSS code
and program that in to the RSS redirection table.
That whole chunk of code should be double checked for endian correctness.
* Use the RSS queue mapping to CPU ID to figure out where to thread pin
the RX swi thread and the taskqueue threads for each queue.
* The software queue is now really an "RSS bucket".
* When programming the RSS indirection table, use the RSS code to
figure out which RSS bucket each slot in the indirection table maps
to.
* When transmitting, use the flowid RSS mapping if the mbuf has
an RSS aware hash. The existing method wasn't guaranteed to align
correctly with the destination RSS bucket (and thus CPU ID.)
This code warns if the number of RSS buckets isn't the same as the
automatically configured number of hardware queues. The administrator
will have to tweak one of them for better performance.
There's currently no way to re-balance the RSS indirection table after
startup. I'll worry about that later.
Additionally, it may be worthwhile to always use the full 32 bit flowid if
multi-queue is enabled. It'll make things like lagg(4) behave better with
respect to traffic distribution.
Adrian Chadd [Mon, 30 Jun 2014 04:34:59 +0000 (04:34 +0000)]
Add initial RSS awareness to the igb(4) driver.
The igb(4) hardware is capable of RSS hashing RX packets and doing RSS
queue selection for up to 8 queues. (I believe some hardware is limited
to 4 queues, but I haven't tested on that.)
However, even if multi-queue is enabled for igb(4), the RX path doesn't use
the RSS flowid from the received descriptor. It just uses the MSIX queue id.
This patch does a handful of things if RSS is enabled:
* Instead of using a random key at boot, fetch the RSS key from the RSS code
and program that in to the RSS redirection table.
That whole chunk of code should be double checked for endian correctness.
* Use the RSS queue mapping to CPU ID to figure out where to thread pin
the RX swi thread and the taskqueue threads for each queue.
* The software queue is now really an "RSS bucket".
* When programming the RSS indirection table, use the RSS code to
figure out which RSS bucket each slot in the indirection table maps
to.
* When transmitting, use the flowid RSS mapping if the mbuf has
an RSS aware hash. The existing method wasn't guaranteed to align
correctly with the destination RSS bucket (and thus CPU ID.)
This code warns if the number of RSS buckets isn't the same as the
automatically configured number of hardware queues. The administrator
will have to tweak one of them for better performance.
There's currently no way to re-balance the RSS indirection table after
startup. I'll worry about that later.
Additionally, it may be worthwhile to always use the full 32 bit flowid if
multi-queue is enabled. It'll make things like lagg(4) behave better with
respect to traffic distribution.
Scott Long [Mon, 30 Jun 2014 01:01:54 +0000 (01:01 +0000)]
In rare cases, a SATA drive can stop responding to commands and trigger a
reset device task request from the driver. If the drive fails to respond
with a signature FIS, the driver would previously get into an endless retry
loop, stalling all I/O to the drive and keeping user processes stranded.
Instead, fail the i/o and invalidate the device if the task management
command times out. This is controllable with the sysctl and tunable
hw.isci.fail_on_task_timeout
dev.isci.0.fail_on_task_timeout
The default for these is 1.
Reviewed by: jimharris
Obtained from: Netflix, Inc.
MFC after: 2 days
Ed Maste [Mon, 30 Jun 2014 00:20:12 +0000 (00:20 +0000)]
Rename the WITHOUT_VT_SUPPORT knob to WITHOUT_VT
The _SUPPORT knobs have a consistent meaning which differs from the
behaviour controlled by this knob. As the knob is opt-out and has not
appeared in a release the impact should be low.
Rick Macklem [Sat, 28 Jun 2014 21:47:15 +0000 (21:47 +0000)]
There might be a potential race condition for the NFSv4 client
when a newly created file has another open done on it that
update the open mode. This patch moves the code that updates
the open mode up into the block where the mutex is held to
ensure this cannot happen. No bug caused by this potential
race has been observed, but this fix is a safety belt to ensure
it cannot happen.
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.
Mateusz Guzik [Sat, 28 Jun 2014 05:18:03 +0000 (05:18 +0000)]
Make sure to always clear p_fd for process getting rid of its filetable.
Filetable can be shared with other processes. Previous code failed to
clear the pointer for all but the last process getting rid of the table.
This is mostly cosmetics.
Get rid of 'This should happen earlier' comment. Clearing the pointer in
this place is fine as consumers can reliably check for files availability
by inspecting fd_refcnt and vnodes availabity by NULL-checking them.
Dimitry Andric [Fri, 27 Jun 2014 20:41:12 +0000 (20:41 +0000)]
Pull in r211627 from upstream llvm trunk (by Bill Schmidt):
[PPC64] Fix PR20071 (fctiduz generated for targets lacking that
instruction)
PR20071 identifies a problem in PowerPC's fast-isel implementation
for floating-point conversion to integer. The fctiduz instruction
was added in Power ISA 2.06 (i.e., Power7 and later). However, this
instruction is being generated regardless of which 64-bit PowerPC
target is selected.
The intent is for fast-isel to punt to DAG selection when this
instruction is not available. This patch implements that change.
For testing purposes, the existing fast-isel-conversion.ll test adds
a RUN line for -mcpu=970 and tests for the expected code generation.
Additionally, the existing test fast-isel-conversion-p5.ll was found
to be incorrectly expecting the unavailable instruction to be
generated. I've removed these test variants since we have adequate
coverage in fast-isel-conversion.ll.
This is needed to compile clang with debug+asserts on older powerpc64
and ppc970 targets.
Marius Strobl [Fri, 27 Jun 2014 19:57:57 +0000 (19:57 +0000)]
In order to get vt(4) a bit closer to the feature set provided by sc(4),
implement options TERMINAL_{KERN,NORM}_ATTR. These are aliased to
SC_{KERNEL_CONS,NORM}_ATTR and like these latter, allow to change the
default colors of normal and kernel text respectively.
Note on the naming: Although affecting the output of vt(4), technically
kern/subr_terminal.c is primarily concerned with changing default colors
so it would be inconsistent to term these options VT_{KERN,NORM}_ATTR.
Actually, if the architecture and abstraction of terminal+teken+vt would
be perfect, dev/vt/* wouldn't be touched by this commit at all.
Reviewed by: emaste
MFC after: 3 days
Sponsored by: Bally Wulff Games & Entertainment GmbH
Marius Strobl [Fri, 27 Jun 2014 18:24:20 +0000 (18:24 +0000)]
- SC_NO_SYSMOUSE isn't currently supported by vt(4), so nuke it from vt.4.
- vt_vga(4) is a driver rather than a function so reference it accordingly.
- Uncomment HISTORY section given that vt(4) will first appear in 9.3.
Reviewed by: emaste (modulo last part)
MFC after: 3 days
Sponsored by: Bally Wulff Games & Entertainment GmbH
Ed Maste [Fri, 27 Jun 2014 17:50:33 +0000 (17:50 +0000)]
Use a common tunable to choose between vt(4)/sc(4)
With this change and previous work from ray@ it will be possible to put
both in GENERIC, and have one enabled by default, but allow the other to
be selected via the loader.
(The previous implementation had separate kern.vt.disable and
hw.syscons.disable tunables, and would panic if both drivers were
compiled in and neither was explicitly disabled.)
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Extend the meaning of the CTLFLAG_TUN flag to automatically check if
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.
Peter Grehan [Fri, 27 Jun 2014 05:27:37 +0000 (05:27 +0000)]
Set the version and date to fixed fields rather than using
preprocessor macros that don't allow reproducible builds.
As a side-effect, the date string is now spec-compliant.
Mateusz Guzik [Fri, 27 Jun 2014 05:04:36 +0000 (05:04 +0000)]
Check lower bound of cmsg_len.
If passed cm->cmsg_len was below cmsghdr size the experssion:
datalen = (caddr_t)cm + cm->cmsg_len - (caddr_t)data;
would give negative result. However, in practice it would not
result in a crash because the kernel would try to obtain garbage fds
for given process and would error out with EBADF.
John Baldwin [Thu, 26 Jun 2014 20:12:38 +0000 (20:12 +0000)]
- Document -b to enable the bvmcons console (but mark it as deprecated
similar to -g.)
- Document -U to set the SMBIOS UUID.
- Add missing options to the usage output and to the manpage Synopsis.
- Don't claim that bvmdebug is amd64-only (it is also a device, not an
option).
Alexander Motin [Thu, 26 Jun 2014 20:06:37 +0000 (20:06 +0000)]
Simplify statistics calculation.
Instead of trying to guess size of disk I/O operations (it just won't work
that way for newly added commands, and is equal to data move size for old
ones), account data move traffic. If disk I/Os are that interesting, then
backends have to account and provide that information.
Block backend already exports the information about disk I/Os via devstat,
so having it here too is excessive.
Rui Paulo [Thu, 26 Jun 2014 19:38:16 +0000 (19:38 +0000)]
MFV illumos r266986:
2915 DTrace in a zone should see "cpu", "curpsinfo", et al
2916 DTrace in a zone should be able to access fds[]
2917 DTrace in a zone should have limited provider access
Bjoern A. Zeeb [Thu, 26 Jun 2014 17:10:07 +0000 (17:10 +0000)]
Allow switching between 32bit and 64bit bus width data access at compile
time by setting NF10BMAC_64BIT and using a REGWTYPE #define to set correct
variable and return value widths.
Adjust comments to indicate the 32 or 64bit register widths.
Alan Cox [Thu, 26 Jun 2014 16:04:03 +0000 (16:04 +0000)]
Delay the call to crhold() in vm_map_insert() until we know that we won't
have to undo it by calling crfree(). This reduces the total number of calls
by vm_map_insert() to crhold() and crfree() by 45% in my tests.
Eliminate an unnecessary variable from vm_map_insert().
Hajimu UMEMOTO [Thu, 26 Jun 2014 12:12:18 +0000 (12:12 +0000)]
- Exclude loopback address rather than loopback interface.
- style(9)
TODO: When AI_ADDRCONFIG is specified, getaddrinfo() can
be quite slow for system with many interfaces. We should
have some kernel sysctls to report IPv4/IPv6 status.
Alexander Motin [Thu, 26 Jun 2014 08:56:36 +0000 (08:56 +0000)]
Add READ BUFFER and improve WRITE BUFFER SCSI commands support.
This gives some use to 512KB per-LUN buffers, allocated for Copan-specific
processor code and not used. It allows, for example, to test transport
performance and/or correctness without accessing the media, as supported
by Linux version of sg3_utils.
Davide Italiano [Thu, 26 Jun 2014 05:23:48 +0000 (05:23 +0000)]
Improve r264388 removing namespace pollution previously introduced in
<sys/time.h>. INT64_MAX actually requires __INT64_C() hack to get the
type right on exotic architectures (e.g. on ones with 63-bit ints or long
0x7fffffffffffffff is unsigned int or long). The hardcoded LL suffix is
good enough to avoid these problems for SBT_MAX (it makes the type always
signed long long, without overflow since long long has at least 64 bits).
Many thanks to Bruce Evans for the time spent me to explain this.
Adrian Chadd [Thu, 26 Jun 2014 04:12:41 +0000 (04:12 +0000)]
Retire IP_RSSCPUID ; the right thing to do is query the RSS bucket;
map the bucket to an RSS queue, then map the queue to a CPU ID.
This way the bucket->queue and queue->CPU mapping can change
over time.
Introduce IP_RSSBUCKETID - which instead looks up the RSS bucket.
User applications can then map the RSS bucket to a CPU.
Adrian Chadd [Thu, 26 Jun 2014 02:49:51 +0000 (02:49 +0000)]
Add another RSS method to query the indirection table entries.
There's 128 indirection table entries which correspond to the
low 7 bits of the 32 bit RSS hash. Each value will correspond
to an RSS bucket. (Then each RSS bucket currently will map
to a CPU.)
This is a more explicit way of figuring out which RSS bucket
is in each RSS indirection slot. It can be inferred by the other
methods but I'd rather drivers use something more simplified and
explicit.
vmm_stat.[ch] -
Modify the counter code in bhyve to allow direct setting of a counter
as opposed to incrementing, and providing a callback to fetch a
counter's value.
John Baldwin [Wed, 25 Jun 2014 20:30:47 +0000 (20:30 +0000)]
Expand r261243 even further and ignore any I/O port resources assigned to
PCI root bridges except for the one known-valid case on x86 where bridges
claim the I/O port registers used for PCI config space access.
Alexander Motin [Wed, 25 Jun 2014 17:02:01 +0000 (17:02 +0000)]
Introduce fine-grained CTL locking to improve SMP scalability.
Split global ctl_lock, historically protecting most of CTL context:
- remaining ctl_lock now protects lists of fronends and backends;
- per-LUN lun_lock(s) protect LUN-specific information;
- per-thread queue_lock(s) protect request queues.
This allows to radically reduce congestion on ctl_lock.
Create multiple worker threads, depending on number of CPUs, and assign
each LUN to one of them. This allows to spread load between multiple CPUs,
still avoiging congestion on queues and LUNs locks.
On 40-core server, exporting 5 LUNs, each backed by gstripe of SATA SSDs,
accessed via 6 iSCSI connections, this change improves peak request rate
from 250K to 680K IOPS.
Alexander Motin [Wed, 25 Jun 2014 16:12:14 +0000 (16:12 +0000)]
Allow to use iSCSI immediate data by several ctl_datamove() calls.
While for FreeBSD client that is only a minor optimization, VMWare client
doesn't support additional data requests after all data being sent once as
immediate.