hselasky [Wed, 16 Oct 2019 08:55:29 +0000 (08:55 +0000)]
Replace rdma_is_upper_dev_rcu() with rdma_vlan_dev_real_dev() in ibcore.
This reduces the number of references to VLAN_TRUNKDEV() in ibcore.
Currently only VLAN is supported as a child interface in FreeBSD.
Remove superfluous RCU locking.
kib [Wed, 16 Oct 2019 07:09:15 +0000 (07:09 +0000)]
Fix assert in PowerPC pmaps after introduction of object busy.
The VM_PAGE_OBJECT_BUSY_ASSERT() in pmap_enter() implementation should
be only asserted when the code is executed as result of pmap_enter(),
not when the same code is entered from e.g. pmap_enter_quick(). This
is relevant for all PowerPC pmap variants, because mmu_*_enter() is
used as the backend, and assert is located there.
Add a PowerPC private pmap_enter() PMAP_ENTER_QUICK_LOCKED flag to
indicate that the call is not from pmap_enter(). For non-quick-locked
calls, assert that the object is locked.
Reported and tested by: bdragon
Reviewed by: alc, bdragon, markj
Sponsored by: The FreeBSD Foundation
Differential revision: https://reviews.freebsd.org/D22041
https://www.illumos.org/issues/9691
When iterating over a ZAP object, we're almost always certain to
iterate over the entire object. If there are multiple leaf blocks, we
can realize a performance win by issuing reads for all the leaf blocks
in parallel when the iteration begins.
For example, if we have 10,000 snapshots, "zfs destroy -nv
pool/fs@1%9999" can take 30 minutes when the cache is cold. This
change provides a >3x performance improvement, by issuing the reads
for all ~64 blocks of each ZAP object in parallel.
https://www.illumos.org/issues/9425
Problem Statement
ZFS Channel program scripts currently require a timeout, so that hung
or long-running scripts return a timeout error instead of causing ZFS
to get wedged. This limit can currently be set up to 100 million Lua
instructions. Even with a limit in place, it would be desirable to
have a sys admin (support engineer) be able to cancel a script that is
taking a long time.
Proposed Solution
Make it possible to abort a channel program by sending an interrupt
signal.In the underlying txg_wait_sync function, switch the cv_wait to
a cv_wait_sig to catch the signal. Once a signal is encountered, the
dsl_sync_task function can install a Lua hook that will get called
before the Lua interpreter executes a new line of code. The
dsl_sync_task can resume with a standard txg_wait_sync call and wait
for the txg to complete. Meanwhile, the hook will abort the script and
indicate that the channel program was canceled. The kernel returns a
EINTR to indicate that the channel program run was canceled.
FreeBSD note: the return value of cv_wait_sig() has inverted meaning
between us and illumos.
https://www.illumos.org/issues/10330
3 recent ZoL changes in the vdev and metaslab code which we can pull over:
PR 8324 c853f382db 8324 Change target size of metaslabs from 256GB to 16GB
PR 8290 b194fab0fb 8290 Factor metaslab_load_wait() in metaslab_load()
PR 8286 419ba59145 8286 Update vdev_is_spacemap_addressable() for new spacemap
encoding
jhibbits [Wed, 16 Oct 2019 00:38:50 +0000 (00:38 +0000)]
powerpc: Add AmigaOne platform, a subclass of MPC85xx
Summary:
The AmigaOne platform, encompassing the X5000 and A1222 at this time, is
based on the mpc85xx platform, but includes some things not listed in
the device tree. Some custom devices, like CPLD, could be added to the
device tree with an overlay, or other means. However, some cannot
easily be done, such as the power button interrupt.
The directory will also become a location to add AmigaOne platform drivers,
such as the aforementioned CPLD, and its children.
erj [Tue, 15 Oct 2019 21:56:19 +0000 (21:56 +0000)]
ixgbe: Disable EEE for backplane X550EM_X
From Zach:
Intel documentation indicates that backplane X550EM_X KR devices do not
support Energy Efficient Ethernet. Prior to this patch, X552 devices
(device ID 0x15AB) will crash the system when transitioning EEE state
via sysctl.
brooks [Tue, 15 Oct 2019 21:27:06 +0000 (21:27 +0000)]
Add the ability to link programs against a compat ABI.
Linkage is controlled by two make knobs:
WANT_COMPAT - Prefer to link against the compat ABI.
NEED_COMPAT - Link against the compat ABI or fail to build.
Supported values are "32", "soft", and "any". The latter meaning pick
the first[0] supported compat ABI.
This can be used to provide test binaries for compat ABIs or to link
ABI-specific programs.
[0] We currently support only one compat ABI at a time, but this may
change in the future and some code in this commit is structured to ease
that change.
jhb [Tue, 15 Oct 2019 19:04:39 +0000 (19:04 +0000)]
Support hot insertion and removal of PCI devices on EC2.
Install ACPI notify handlers on PCI devices with an _EJ0 method. This
handler is invoked when devices are added or removed.
- When an ACPI_NOTIFY_DEVICE_CHECK event posts, rescan the parent bus
device. Note that strictly speaking we only need to rescan the
specified device, but BUS_RESCAN is what is available, so we rescan
the entire bus.
- When an ACPI_NOTIFY_EJECT_REQUEST event posts, detach the device
associated with the ACPI handle, invoke the _EJ0 method, and then
delete the device.
Eventually this might be changed to vector notify events to devd in
userspace where devctl can be used instead to permit more complex
actions such as graceful unmounting of filesystems.
jhb [Tue, 15 Oct 2019 18:16:10 +0000 (18:16 +0000)]
Use __FreeBSD_version to determine if gets() has been removed.
GCC compilers set __FreeBSD__ statically to a build-time determined
targeted version (which in ports always matches the build host's
version). This means that when building any version (12 or 13, etc.)
of riscv or some other architecture via GCC on a 12.x host,
__FreeBSD__ will always be set to 12. As a result, __FreeBSD__ cannot
be used to reliably detect the target FreeBSD version being built.
Instead, __FreeBSD_version from either <sys/param.h> (in the kernel)
or <osreldate.h> (in userland) should be used.
This changes the gets() test in libc++ to use __FreeBSD_version from
<osreldate.h>.
jhb [Tue, 15 Oct 2019 17:11:42 +0000 (17:11 +0000)]
Update MIPS kernel builds to work with mips-gcc.
- Use a default -march of mips64 on N64 and N32 kernels.
- Set the endianness (via MIPS_ENDIAN) and ABI (via MIPS_ABI) in
CFLAGS from MACHINE_ARCH. ARCH_FLAGS now only sets a different
-march value if needed.
- TRAMP_ARCH_FLAGS inherits MIPS_ENDIAN from MACHINE_ARCH but does
not set the ABI since XLPN32 needs an N64 ABI for the trampoline
loader. When TRAMP_ARCH_FLAGS is used it must set both -march
and -mabi.
https://www.illumos.org/issues/10343
On the openzfs feature/porting matrix, this is listed as:
prefix to refcount funcs/types
Having these changes will make it easier to share other work across the
different ZFS operating systems.
PR 7963 424fd7c3e Prefix all refcount functions with zfs_
PR 7885 & 7932 c13060e47 Linux 4.19-rc3+ compat: Remove refcount_t compat
PR 5823 & 5842 4859fe796 Linux 4.11 compat: avoid refcount_t name conflict
Author: Tim Schumacher <timschumi@gmx.de>
Obtained from: illumos, ZoL
MFC after: 3 weeks
10572 Fix race in dnode_check_slots_free()
https://www.illumos.org/issues/10572
The Fix from ZoL:
Currently, dnode_check_slots_free() works by checking dn->dn_type
in the dnode to determine if the dnode is reclaimable. However,
there is a small window of time between dnode_free_sync() in the
first call to dsl_dataset_sync() and when the useraccounting code
is run when the type is set DMU_OT_NONE, but the dnode is not yet
evictable, leading to crashes. This patch adds the ability for
dnodes to track which txg they were last dirtied in and adds a
check for this before performing the reclaim.
This patch also corrects several instances when dn_dirty_link was
treated as a list_node_t when it is technically a multilist_node_t.
10579 Don't allow dnode allocation if dn_holds != 0
https://www.illumos.org/issues/10579
The fix from ZoL:
This patch simply fixes a small bug where dnode_hold_impl() could
attempt to allocate a dnode that was in the process of being freed,
but which still had active references. This patch simply adds the
required check.
hselasky [Tue, 15 Oct 2019 12:08:09 +0000 (12:08 +0000)]
The two functions ifnet_byindex() and ifnet_byindex_locked() are exactly the
same after the network stack was epochified. Merge the two into one function
and cleanup all uses of ifnet_byindex_locked().
While at it:
- Add branch prediction macros.
- Make sure the ifnet pointer is only deferred once,
also when code optimisation is disabled.
hselasky [Tue, 15 Oct 2019 11:20:16 +0000 (11:20 +0000)]
Exclude the network link eventhandler from epochification after r353292.
This fixes the following assert when "options RATELIMIT" is used:
panic()
malloc()
sysctl_add_oid()
tcp_rl_ifnet_link()
do_link_state_change()
taskqueue_run_locked()
ae [Tue, 15 Oct 2019 09:50:02 +0000 (09:50 +0000)]
Explicitly initialize the memory buffer to store O_ICMP6TYPE opcode.
By default next_cmd() initializes only first u32 of opcode. O_ICMP6TYPE
opcode has array of bit masks to store corresponding ICMPv6 types.
An opcode that precedes O_ICMP6TYPE, e.g. O_IP6_DST, can have variable
length and during opcode filling it can modify memory that will be used
by O_ICMP6TYPE opcode. Without explicit initialization this leads to
creation of wrong opcode.
Reported by: Boris N. Lytochkin
Obtained from: Yandex LLC
MFC after: 3 days
jeff [Tue, 15 Oct 2019 03:45:41 +0000 (03:45 +0000)]
(4/6) Protect page valid with the busy lock.
Atomics are used for page busy and valid state when the shared busy is
held. The details of the locking protocol and valid and dirty
synchronization are in the updated vm_page.h comments.
jeff [Tue, 15 Oct 2019 03:41:36 +0000 (03:41 +0000)]
(3/6) Add a shared object busy synchronization mechanism that blocks new page
busy acquires while held.
This allows code that would need to acquire and release a very large number
of page busy locks to use the old mechanism where busy is only checked and
not held. This comes at the cost of false positives but never false
negatives which the single consumer, vm_fault_soft_fast(), handles.
jeff [Tue, 15 Oct 2019 03:35:11 +0000 (03:35 +0000)]
(1/6) Replace busy checks with acquires where it is trival to do so.
This is the first in a series of patches that promotes the page busy field
to a first class lock that no longer requires the object lock for
consistency.
manu [Mon, 14 Oct 2019 21:53:53 +0000 (21:53 +0000)]
arm64: allwinner: Add Allwinner H6 Support
This adds support for H6 SoC.
Add a CCU driver for H6 that support all PLLs and most of the clocks
that we are intersted in for now (i2c, mmc, usb, etc ...)
jhb [Mon, 14 Oct 2019 21:22:42 +0000 (21:22 +0000)]
Disconnect powerpc from the default tinderbox for now.
The wrong toolchain was set in MAKE_PARAMS_powerpc, however, there are
some other issues preventing powerpc from building in tinderbox:
1) There is no powerpc-gcc and powerpc-xtoolchain-gcc port that
provides an external 32-bit powerpc GCC toolchain.
2) On other targets, the same toolchain can build all of the
architectures for a given target. MIPS achieves this by always
setting -mabi and -EB/-EL explicitly instead of relying on the
compiler's default architecture. PowerPC might be able to do the
same thing, but as of today, powerpc-gcc would be required for
powerpc and powerpcspe and powerpc64-gcc would be required for
powerpc64. Our existing logic for make universe does not permit
per-MACHINE_ARCH toolchains.
I tried hacking TARGETS_powerpc to only include powerpc64 when
powerpc64-gcc was present, and while that skipped the 32-bit worlds,
it tried to build all the kernels.
cem [Mon, 14 Oct 2019 20:55:01 +0000 (20:55 +0000)]
x86: Use canonical spelling of MOVDIR64B feature/instruction
The former spelling probably confused MOVDIR64B with MOVDIRI64.
MOVDIR_64B is the 64-*byte* direct store instruction; MOVDIR_I64 is the
64-*bit* direct store instruction (underscores added here for clarity; they are
not part of the canonical instruction name).
dougm [Mon, 14 Oct 2019 17:15:42 +0000 (17:15 +0000)]
Move the definition of _vm_map_assert_consistent so that it can use
vm_map_free_{left,right} rather than re-implementing them. Use the
VM_MAP_FOREACH macro where applicable. Fix some indentation.
Suggested by: kib (in a comment on D21964)
Tested by: pho (as part of D21964)
Differential Revision: https://reviews.freebsd.org/D22011