marius [Thu, 30 Jul 2015 02:06:34 +0000 (02:06 +0000)]
MFC: r285839
o Revert the other functional half of r239864 (MFCed to stable/9 in r241681),
i. e. the merge of r134227 from x86 to use smp_ipi_mtx spin lock not only
for smp_rendezvous_cpus() but also for the MD cache invalidation, TLB
demapping and remote register reading IPIs due to the following reasons:
- The cross-IPI SMP deadlock x86 otherwise is subject to can't happen on
sparc64. That's because on sparc64, spin locks don't disable interrupts
completely but only raise the processor interrupt level to PIL_TICK. This
means that IPIs still get delivered and direct dispatch IPIs such as the
cache invalidation etc. IPIs in question are still executed.
- In smp_rendezvous_cpus(), smp_ipi_mtx is held not only while sending an
IPI_RENDEZVOUS, but until all CPUs have processed smp_rendezvous_action().
Consequently, smp_ipi_mtx may be locked for an extended amount of time as
queued IPIs (as opposed to the direct ones) such as IPI_RENDEZVOUS are
scheduled via a soft interrupt. Moreover, given that this soft interrupt
is only delivered at PIL_RENDEZVOUS, processing of smp_rendezvous_action()
on a target may be interrupted by f. e. a tick interrupt at PIL_TICK, in
turn leading to the target in question trying to send an IPI by itself
while IPI_RENDEZVOUS isn't fully handled, yet, and, thus, resulting in a
deadlock.
o As mentioned in the commit message of r245850, on least some sun4u platforms
concurrent sending of IPIs by different CPUs is fatal. Therefore, hold the
reintroduced MD ipi_mtx also while delivering cross-traps via MI helpers,
i. e. ipi_{all_but_self,cpu,selected}().
o Akin to x86, let the last CPU to process cpu_mp_bootstrap() set smp_started
instead of the BSP in cpu_mp_unleash(). This ensures that all APs actually
are started, when smp_started is no longer 0.
o In all MD and MI IPI helpers, check for smp_started == 1 rather than for
smp_cpus > 1 or nothing at all. This avoids races during boot causing IPIs
trying to be delivered to APs that in fact aren't up and running, yet.
While at it, move setting of the cpu_ipi_{selected,single}() pointers to
the appropriate delivery functions from mp_init() to cpu_mp_start() where
it's better suited and allows to get rid of the global isjbus variable.
o Given that now concurrent IPI delivery no longer is possible, also nuke
the delays before completely disabling interrupts again in the CPU-specific
cross-trap delivery functions, previously giving other CPUs a window for
sending IPIs on their part. Actually, we now should be able to entirely get
rid of completely disabling interrupts in these functions. Such a change
needs more testing, though.
o In {s,}tick_get_timecount_mp(), make the {s,}tick variable static. While not
necessary for correctness, this avoids page faults when accessing the stack
of a foreign CPU as {s,}tick now is locked into the TLBs as part of static
kernel data. Hence, {s,}tick_get_timecount_mp() always execute as fast as
possible, avoiding jitter.
marius [Thu, 30 Jul 2015 00:28:32 +0000 (00:28 +0000)]
MFC: r284447, r284552
Merge from NetBSD:
o rev. 1.10: Nuke trailing whitespace.
o rev. 1.15: Fix typo in comment.
o rev. 1.16: Add the following registers from IEEE 802.3-2009 Clause 22:
- PSE control register (0x0b)
- PSE status register (0x0c)
- MMD access control register (0x0d)
- MMD access address data register (0x0e)
o rev. 1.17 (comments only): The bit location of link ability is different
between 1000Base-X and others (see Annex 28B.2 and 28D).
o rev. 1.18: Nuke dupe word.
dim [Wed, 29 Jul 2015 19:37:32 +0000 (19:37 +0000)]
Merge r286033 from stable/10:
Reapply r286007, modified to compile with pre-C++11 compilers:
Pull in r219009 from upstream llvm trunk (by Adam Nemet):
[ISel] Keep matching state consistent when folding during X86 address match
In the X86 backend, matching an address is initiated by the 'addr' complex
pattern and its friends. During this process we may reassociate and-of-shift
into shift-of-and (FoldMaskedShiftToScaledMask) to allow folding of the
shift into the scale of the address.
However as demonstrated by the testcase, this can trigger CSE of not only the
shift and the AND which the code is prepared for but also the underlying load
node. In the testcase this node is sitting in the RecordedNode and MatchScope
data structures of the matcher and becomes a deleted node upon CSE. Returning
from the complex pattern function, we try to access it again hitting an assert
because the node is no longer a load even though this was checked before.
Now obviously changing the DAG this late is bending the rules but I think it
makes sense somewhat. Outside of addresses we prefer and-of-shift because it
may lead to smaller immediates (FoldMaskAndShiftToScale is an even better
example because it create a non-canonical node). We currently don't recognize
addresses during DAGCombiner where arguably this canonicalization should be
performed. On the other hand, having this in the matcher allows us to cover
all the cases where an address can be used in an instruction.
I've also talked a little bit to Dan Gohman on llvm-dev who added the RAUW for
the new shift node in FoldMaskedShiftToScaledMask. This RAUW is responsible
for initiating the recursive CSE on users
(http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-September/076903.html) but it
is not strictly necessary since the shift is hooked into the visited user. Of
course it's safer to keep the DAG consistent at all times (e.g. for accurate
number of uses, etc.).
So rather than changing the fundamentals, I've decided to continue along the
previous patches and detect the CSE. This patch installs a very targeted
DAGUpdateListener for the duration of a complex-pattern match and updates the
matching state accordingly. (Previous patches used HandleSDNode to detect the
CSE but that's not practical here). The listener is only installed on X86.
I tested that there is no measurable overhead due to this while running
through the spec2k BC files with llc. The only thing we pay for is the
creation of the listener. The callback never ever triggers in spec2k since
this is a corner case.
dim [Wed, 29 Jul 2015 13:14:34 +0000 (13:14 +0000)]
Merge r286007 from stable/10:
Pull in r219009 from upstream llvm trunk (by Adam Nemet):
[ISel] Keep matching state consistent when folding during X86 address match
In the X86 backend, matching an address is initiated by the 'addr' complex
pattern and its friends. During this process we may reassociate and-of-shift
into shift-of-and (FoldMaskedShiftToScaledMask) to allow folding of the
shift into the scale of the address.
However as demonstrated by the testcase, this can trigger CSE of not only the
shift and the AND which the code is prepared for but also the underlying load
node. In the testcase this node is sitting in the RecordedNode and MatchScope
data structures of the matcher and becomes a deleted node upon CSE. Returning
from the complex pattern function, we try to access it again hitting an assert
because the node is no longer a load even though this was checked before.
Now obviously changing the DAG this late is bending the rules but I think it
makes sense somewhat. Outside of addresses we prefer and-of-shift because it
may lead to smaller immediates (FoldMaskAndShiftToScale is an even better
example because it create a non-canonical node). We currently don't recognize
addresses during DAGCombiner where arguably this canonicalization should be
performed. On the other hand, having this in the matcher allows us to cover
all the cases where an address can be used in an instruction.
I've also talked a little bit to Dan Gohman on llvm-dev who added the RAUW for
the new shift node in FoldMaskedShiftToScaledMask. This RAUW is responsible
for initiating the recursive CSE on users
(http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-September/076903.html) but it
is not strictly necessary since the shift is hooked into the visited user. Of
course it's safer to keep the DAG consistent at all times (e.g. for accurate
number of uses, etc.).
So rather than changing the fundamentals, I've decided to continue along the
previous patches and detect the CSE. This patch installs a very targeted
DAGUpdateListener for the duration of a complex-pattern match and updates the
matching state accordingly. (Previous patches used HandleSDNode to detect the
CSE but that's not practical here). The listener is only installed on X86.
I tested that there is no measurable overhead due to this while running
through the spec2k BC files with llc. The only thing we pay for is the
creation of the listener. The callback never ever triggers in spec2k since
this is a corner case.
dim [Tue, 28 Jul 2015 09:19:04 +0000 (09:19 +0000)]
MFC r285340:
Fix swapped copyin(9) arguments in cxgb's iwch_arm_cq() function.
Detected by clang 3.7.0 with the warning:
sys/dev/cxgb/ulp/iw_cxgb/iw_cxgb_provider.c:309:18: error: variable
'rptr' is uninitialized when used here [-Werror,-Wuninitialized]
chp->cq.rptr = rptr;
^~~~
MFC r285735:
lseek() allows an offset to be set beyond the end of file. Using
it to check that partition has enough space to write bootcode doesn't
work. Use the known size of provider instead.
MFC r285663, r285664, r285667:
Ensure that locstat_nsecs() has no effect when lockstat probes are not
enabled or when the profiled lock carries the LO_NOPROFILE flag.
MFC r285088:
Fix broken implementation of "kvasprintf()" function by adding missing
kmalloc() call. Make function global instead of static inline to fix
compiler warnings about passing variable argument lists to inline
functions.
MFC r285253 (hrs):
- Add IPv6 support in quota(1). While rpc.rquotad has supported
PF_INET6 for a long time, quota(1) utility supported only PF_INET.
- Clean up confusing changes in f_mntfromname.
- Add an entry for rquotad with rpc/udp6 to inetd.conf.
By default, sendmail 8.15 uses uncompressed IPv6 addresses. Keep
current FreeBSD 10 and earlier behavior of using compressed IPv6
addresses in configuration, maps, rulesets, etc. (FreeBSD 11 and
later will use the new default of uncompressed IPv6 addresses.)
Allow parameters listed on the command line to override the -v option,
instead of crashing.
MFC r279083:
Fix the logic for skipping parameters (with -s) that have "jailsys"
parents (such as host.hostname); these were being skipped all the time.
That it went this long without anyone noticing is a sign that this feature
isn't actually used by anyone, but it's there so it might as well work.
MFC r279123:
Allow for parameters added with the JP_OPT flag to not exist.
That's why the flag exists in the first place.
* jail_setv will leak a parameter name if jailparam_import fails.
* jailparam_all loses the jailparam pointer on realloc error
(a clear freshman mistake).
* If jailparam_init fails, the caller doesn't need to jailparam_free
the buffer. That's not really clear, so set things to NULL allowing
jailparam_free to work without error (though it's still not required).
MFC r284593: MFV r284412: 5911 ZFS "hangs" while deleting file
illumos/illumos-gate@46e1baa6cf6d5432f5fd231bb588df8f9570c858
https://www.illumos.org/issues/5911
Sometimes ZFS appears to hang while deleting a file. It is actually
making slow progress at the file deletion, but other operations
(administrative and writes via the data path) "hang" until the file
removal completes, which can take a long time if the file has many
blocks. The deletion (or most of it) happens in a single txg, and the
sync thread spends most of its time reading indirect blocks...
Reviewed by: Bayard Bell <buffer.g.overflow@gmail.com>
Reviewed by: Alek Pinchuk <alek@nexenta.com>
Reviewed by: Simon Klinkert <simon.klinkert@gmail.com>
Reviewed by: Dan McDonald <danmcd@omniti.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: Matthew Ahrens <mahrens@delphix.com>
marius [Sun, 5 Jul 2015 20:16:45 +0000 (20:16 +0000)]
MFC: r281337
Don't enable RX and TX before their initial configuration is done, i. e.
after setting up interrupt moderation but before turning interrupts on.
This matches what Realtek's r8168 Linux driver does as of version 8.039.00
and fixes problems with certain incarnations of certain MAC revisions
like the interface requiring an extra up/down-cycle after boot to start
working or DMA configuration not being adhered to.
marius [Sun, 5 Jul 2015 19:33:30 +0000 (19:33 +0000)]
MFC: r281751
Refine the workaround for Intel HSD131 [1] added in r269052 (MFCed to
stable/9 in r269593):
- Use the full mask described by the erratum as with a sufficiently high
number of these false-positives, the overflow bit (bit 62) additionally
gets set [7].
- HSD131 has been brought into several other Haswell-derived CPUs including
to the next generation, i. e. Intel Broadwell. Thus, also skip reporting of
these benign errors by default on CPU models affected by HSM142, HSW131 and
BDM48 [2 - 5], describing the HSD131 silicon bug for additional models.
Also, Celeron 2955U with a CPU ID of 0x45 have been reported to be covered
by this fault [6], with the specification update concerned with HSM142 [2]
only referring to 0x3c and 0x46.
Submitted by: David Froehlich [7]
Approved by: re (kib)
dim [Fri, 3 Jul 2015 20:31:27 +0000 (20:31 +0000)]
MFC r284732:
Add __cxa_deleted_virtual to libcxxrt's version map.
This symbol can sometimes be emitted by clang++, and was not yet
exported from libcxxrt. Attempt to be compatible with libsupc++ by
using the same CXXABI_1.3.6 symbol version.
des [Wed, 1 Jul 2015 08:06:15 +0000 (08:06 +0000)]
MFH (r228976): UTFize
MFH (r243076): check range of gid
MFH (r243083): fix possible overflow in error counter
MFH (r275765): clean up, reindent, add special case for + lines
davidcs [Tue, 30 Jun 2015 21:19:27 +0000 (21:19 +0000)]
MFC r284741
Updated Copyright information
Added support for the following:
- iSCSI TLV (requires 64 Tx and 32 Rx rings
- 9K receive buffers for jumbo frames (feature may be enabled/disabled)
- builtin firmware, bootloader and minidump template
- quick stats
- async event handling for SFP insertion/removal and DCBX changes
- Configuring DCBX and interrupt coalescing parameters
davidcs [Tue, 30 Jun 2015 19:59:19 +0000 (19:59 +0000)]
MFC r284470
In bxe_init_mcast_macs_list(): mc_mac->mac needs to point to the multicast mac address
In bxe_set_mc_list(): added missing BXE_MCAST_UNLOCK()
In __ecore_vlan_mac_h_exec_pending(): need to check for ECORE_PENDING
davidcs [Tue, 30 Jun 2015 19:53:59 +0000 (19:53 +0000)]
MFC r284335
PHY LOCK acquires the hardware lock via bxe_acquire_phy_lock() and releases it via bxe_release_phy_lock(). It was simply acquiring a mutex earlier which can cause the PHY to use bogus values. Fixes intermittent link failures.
bxe_ioctl() completes all functions within its context as opposed to a taskqueue earlier.
bxe_handle_rx_mode_tq() no longer required. bxe_set_rx_mode() handles the functionality within its context
pfg [Mon, 29 Jun 2015 17:58:45 +0000 (17:58 +0000)]
MFC r282115, r282152, r282201
MFV r282150
libgomp: Update to version 4.3.5.
bring initial BSD support from upstream.
This was not meant to be MFC'd at first but the original OMP support
for FreeBSD was in very poor shape. The effect of this change should
be minimal as all ports are already using the version of libgomp that
comes with the gcc ports.
gshapiro [Thu, 25 Jun 2015 01:53:45 +0000 (01:53 +0000)]
MFC: An additional fix for the openssl Weak DH remediation:
The import of openssl to address the FreeBSD-SA-15:10.openssl security
advisory includes a change which rejects handshakes with DH parameters
below 768 bits. sendmail releases prior to 8.15.2 (not yet released),
defaulted to a 512 bit DH parameter setting for client connections.
The first fix committed last week changed the default to 1024 bits.
This commit fixes the case where the DHParameters option is set to a
file which doesn't exist, which is the case on newer versions of
FreeBSD which enable STARTTLS by default by auto-creating TLS
certificates.
dim [Sat, 20 Jun 2015 13:25:28 +0000 (13:25 +0000)]
MFC r284167:
Merge r283870 from amd64:
Remove unneeded NULL checks in trap_fatal().
Since td_name is an array member of struct thread, it can never be NULL,
so the check can be removed. In addition, curproc can never be NULL,
so remove the if statement, and splice the two printfs() together.
While here, remove the u_long cast, and use the correct printf format
specifier for curproc->p_pid.
delphij [Wed, 17 Jun 2015 19:13:13 +0000 (19:13 +0000)]
MFC r264038 (theraven)
Fix an issue where the locale and rune locale could become out of sync,
causing mb* functions (and similar) to be called with the wrong data
(possibly a null pointer, causing a crash).
hselasky [Wed, 17 Jun 2015 07:43:20 +0000 (07:43 +0000)]
MFC r284012:
Disable VGA PCI interrupts until a chipset driver is loaded for VGA
PCI devices. Else unhandled display adapter interrupts might freeze
the CPU or consume a lot of CPU.
hselasky [Wed, 17 Jun 2015 07:28:51 +0000 (07:28 +0000)]
MFC r280991:
Extend fixes made in r278103 and r38754 by copying the complete packet
header and not only partial flags and fields. Firewalls can attach
classification tags to the outgoing mbufs which should be copied to
all the new fragments. Else only the first fragment will be let
through by the firewall. This can easily be tested by sending a large
ping packet through a firewall. It was also discovered that VLAN
related flags and fields should be copied for packets traversing
through VLANs. This is all handled by "m_dup_pkthdr()".
Regarding the MAC policy check in ip_fragment(), the tag provided by
the originating mbuf is copied instead of using the default one
provided by m_gethdr().
gshapiro [Wed, 17 Jun 2015 03:11:25 +0000 (03:11 +0000)]
MFC: The import of openssl to address the FreeBSD-SA-15:10.openssl security
advisory includes a change which rejects handshakes with DH parameters
below 768 bits. sendmail releases prior to 8.15.2 (not yet released),
defaulted to a 512 bit DH parameter setting for client connections.
This commit chages that default to 1024 bits. sendmail 8.15.2, when
released well use a default of 2048 bits.
dim [Mon, 8 Jun 2015 19:44:04 +0000 (19:44 +0000)]
MFC r283870:
Remove unneeded NULL checks in amd64's trap_fatal().
Since td_name is an array member of struct thread, it can never be NULL,
so the check can be removed. In addition, curproc can never be NULL,
so remove the if statement, and splice the two printfs() together.
While here, remove the u_long cast, and use the correct printf format
specifier curproc->p_pid.
jhb [Mon, 8 Jun 2015 18:59:14 +0000 (18:59 +0000)]
MFC 233665: (just to ps.1)
mandoc complains loudly when <TAB>s are misused in columnated lists. Fix
this syntax violation and while I'm here also convert <TAB> to Ta and adjust
quotation marks in order to prevent this problem in the future.
jhb [Sat, 6 Jun 2015 20:37:40 +0000 (20:37 +0000)]
MFC 261811,282660,282706:
Place VM objects on the object list when created and never remove them.
261811:
Fix function name in KASSERT().
282660:
Place VM objects on the object list when created and never remove them.
This is ok since objects come from a NOFREE zone and allows objects to
be locked while traversing the object list without triggering a LOR.
Ensure that objects on the list are marked DEAD while free or stillborn,
and that they have a refcount of zero. This required updating most of
the pagers to explicitly mark an object as dead when deallocating it.
(Only the vnode pager did this previously.)
282706:
Satisfy vm_object uma zone destructor requirements after r282660 when
vnode object creation raced.
jhb [Fri, 5 Jun 2015 20:38:22 +0000 (20:38 +0000)]
MFC 281932:
Rename the kld for oce(4) to if_oce.ko. ifconfig(8) has special knowledge
about kld filenames for network drivers that requires them to follow the
pattern of if_<foo>. This also fixes the existing documentation in the
manpage which says to use if_oce_load=YES in loader.conf.
jhb [Tue, 2 Jun 2015 15:12:33 +0000 (15:12 +0000)]
MFC 282417:
Various updates to the ftruncate(2) documentation:
- Note that ftruncate(2) can operate on shared memory objects and cross
reference shm_open(2).
- Note that ftruncate(2) does not change the file position pointer (aka
seek pointer) of the file descriptor.
- ftruncate(2) will fail with EINVAL for all sorts of other fd types than
just sockets, so instead note that it fails for all but regular files and
shared memory objects.
- Note that ftruncate(2) also appeared in 4.2BSD along with truncate(2).
(Or at least the manpage for both appeared in 4.2, I did not check the
kernel code itself to see if either predated 4.2.)
jhb [Mon, 1 Jun 2015 18:08:56 +0000 (18:08 +0000)]
MFC 283123:
Fix two bugs that could result in PMC sampling effectively stopping.
In both cases, the the effect of the bug was that a very small positive
number was written to the counter. This means that a large number of
events needed to occur before the next sampling interrupt would trigger.
Even with very frequently occurring events like clock cycles wrapping all
the way around could take a long time. Both bugs occurred when updating
the saved reload count for an outgoing thread on a context switch.
First, the counter-independent code compares the current reload count
against the count set when the thread switched in and generates a delta
to apply to the saved count. If this delta causes the reload counter
to go negative, it would add a full reload interval to wrap it around to
a positive value. The fix is to add the full reload interval if the
resulting counter is zero.
Second, occasionally the raw counter value read during a context switch
has actually wrapped, but an interrupt has not yet triggered. In this
case the existing logic would return a very large reload count (e.g.
2^48 - 2 if the counter had overflowed by a count of 2). This was seen
both for fixed-function and programmable counters on an E5-2643.
Workaround this case by returning a reload count of zero.