We generate the installed objcopy man page from ELF Tool Chain's
elfcopy, but the sed expresion used for this ended up producing
"objcopy, objcopy - copy and translate object files".
Instead of replacing the first "elfcopy" with objcopy, just remove it.
Switch reproducible builds to unmodified src tree mode
newvers.sh supports two modes for reproducible builds:
-r Reproducible build. Do not embed directory names, user
names, time stamps or other dynamic information into
the output file. This is intended to allow two builds
done at different times and even by different people on
different hosts to produce identical output.
-R Reproducible build if the tree represents an unmodified
checkout from a version control system. Metadata is
included if the tree is modified.
Switch to the second mode when reproducible builds are enabled.
The value of a reproducible build is much less when building from an
uncontrolled, modified src tree, and -R likely provides the best
compromise in allowing the REPRODUCIBLE_BUILD knob to be enabled by
default for the release.
Approved by: re (kib)
Sponsored by: The FreeBSD Foundation
ix(4), ixv(4): VLAN tag stripping fixes for Amazon EC2 Enhanced Networking
From Piotr:
ix(4), ixv(4): Add VLAN tag strip check when receiving packets
ixv(4): Fix support for VLAN_HWTAGGING and VLAN_HWFILTER flags
This change will prevent driver from passing VLAN tags when
interface configuration is not expecting them. VF driver will
check for VLAN_HWTAGGING and VLAN_HWFILTER flags and act adequately.
This patch resolves problem occuring on EC2 platforms.
des [Mon, 10 Sep 2018 19:39:20 +0000 (19:39 +0000)]
Through a combination of insufficient variable initialization and
imprudent reuse of static buffers, the end-of-transfer statistics
displayed when stdout is not a tty always ended up as 0 B / 0 Bps.
Reorganize the code to use caller-provided buffers, tweak the ETA
display a bit, and reduce the visual differences between the tty and
non-tty end-of-transfer displays.
Lookups are protected by an epoch section, so the LB group linkage must
be a CK_LIST rather than a plain LIST. Furthermore, we were not
deferring LB group frees, so in_pcbremlbgrouphash() could race with
readers and cause a use-after-free.
Reviewed by: sbruno, Johannes Lundberg <johalun0@gmail.com>
Tested by: gallatin
Approved by: re (gjb)
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17031
geli init with multiple providers - fix init and fix -B "none"
Apply some fixes post rS336659, which allowed multiple provders to be
initialized in a single command.
- Fix issue where second and subsequent providers would fail init.
This was due to the metadata struct being zeroed after the first
provider init was completed, despite containing common data required
for subsequent providers.
- Fix issue where -B "none" would still result in the metadata being
backed-up if multiple providers had been specified. This was due to
the backupfile of "none" being incorrectly made unique for each
provider by appending "-<prov>".
Approved by: asomers
Approved by: re (gjb)
Differential Revision: https://reviews.freebsd.org/D17096
Allow dhclient and ping to build WITHOUT_DYNAMICROOT
dhclient and ping normally use libcasper services. These are not
available in statically-linked binaries, so when WITHOUT_DYNAMICROOT is
set disable libcasper use, as with rescue builds. Also emit a warning
as it's undesirable to build this way.
Reported by: Michael Dexter
Reviewed by: rgrimes
Tested by: Michael Dexter
Approved by: re (kib)
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17074
Introduce and use sgid_index in CM requests in ibcore.
For RoCE, when CM requests are received for RC and UD connections,
netdevice of the incoming request is unavailable. Because of that CM
requests are always forwarded to init_net namespace.
Now that we have the GID index available, introduce SGID index in
incoming CM requests and refer to the netdevice of it.
While at it fix some incorrect uses of init_net and make sure
the rdma_create_id() function stores the VNET it is passed.
Exclude the EFI framebuffer from phys_avail[] on arm64.
On the ThunderX the region occupied by the framebuffer is included in
the EFI map, so explicitly add it to the set of regions that aren't
managed by the physical memory allocator.
PR: 231064
Reviewed by: andrew
Approved by: re (gjb)
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17073
These limits are hit on the ThunderX. Also make
arm_physmem_exclude_region() panic rather than fail silently if the
limit on excluded regions is reached.
PR: 231064
Reviewed by: andrew
Approved by: re (kib)
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17073
While executing vm_pqbatch_process_page(m), m->queue may change to
PQ_NONE if the page daemon is concurrently freeing the page. In this
case m's queue state flags must be clear, so vm_pqbatch_process_page()
will be a no-op, but the race could cause spurious assertion failures.
Correct the assertion which assumed that m->queue's value does not
change while the page queue lock is held.
Reviewed by: alc, kib
Reported and tested by: pho
Approved by: re (gjb)
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17027
Reviewed by: bz, Johannes Lundberg <johalun0@gmail.com>
Approved by: re (kib)
Sponsored by: The FreeBSD Foundation
Differential revision: https://reviews.freebsd.org/D17065
Implement get network interface by params function in ipoib.
Also fix the validate_ipv4_net_dev() and validate_ipv6_net_dev() functions
which had source and destination addresses swapped, and didn't set the
scope ID for IPv6 link-local addresses.
This allows applications like krping to work using IPoIB devices.
MFC after: 3 days
Approved by: re (gjb)
Sponsored by: Mellanox Technologies
The Call For Testing had no reports of operational problems and
found that performance was no worse and usually better when running
with TRIM consolidation. Performance improvement was most noticable
when multiple large files are released in a short period of time.
Thus, TRIM consolidation is being enabled by default. Should
operational problems be found, it can be disabled using the command
`sysctl vfs.ffs.dotrimcons=0'. This variable can also be set as a
tunable if early disabling is necessary.
marius [Thu, 6 Sep 2018 21:09:54 +0000 (21:09 +0000)]
- Explicitly compare a pointer to NULL. The __builtin_expect() of clang
3.4.1 otherwise isn't able to cope with the expression.
- Fix a nearby whitespace bug.
bhyve: Use MAP_GUARD when mapping guest memory ranges.
Instead of relying on PROT_NONE mappings with MAP_ANON, use MAP_GUARD
to reserve address space around guest memory ranges including the
guard ranges of address space around mappings.
The inp_lle field to struct inpcb, along with two "valid" flags
for the rt and lle cache were added in r191129 (2009).
To my best knowledge they have never been used and route caching
has converted the inp_rt field from that commit to inp_route
rendering this field and these flags obsolete.
Convert the pointer into a spare pointer to not change the size of
the structure anymore (and to have a spare pointer) and mark the
two fields as unused.
The stac/clac combo around each byte copy is causing a measurable
slowdown in benchmarks. Do it only before and after all data is
copied. While here reorder the code to avoid a forward branch in
the common case.
Note the copying loop (originating from copyinstr) is avoidably slow
and will be fixed later.
Reviewed by: kib
Approved by: re (gjb)
Differential Revision: https://reviews.freebsd.org/D17063
Avoid resource deadlocks when one domain has exhausted its memory. Attempt
other allowed domains if the requested domain is below the minimum paging
threshold. Block in fork only if all domains available to the forking
thread are below the severe threshold rather than any.
Submitted by: jeff
Reported by: mjg
Reviewed by: alc, kib, markj
Approved by: re (rgrimes)
Differential Revision: https://reviews.freebsd.org/D16191
Re-enable kernel modules for the MALTA64EL kernel configuration.
Update the BOOTSTRAPPING check for libelf to require the fix for
mips64el object files committed in r338478 and re-enable kernel
modules in the MALTA64EL config file.
Reviewed by: emaste
Approved by: re (gjb)
Differential Revision: https://reviews.freebsd.org/D17054
Remove sysctls:
txq_drain_encapfail - now a duplicate of encap_txd_encap_fail
intr_link - was never incremented
intr_msix - was never incremented
rx_zero_len - was never incremented
The following were not incremented in all code-paths that apply:
m_pullups, mbuf_defrag, rxd_flush, tx_encap, rx_intr_enables, tx_frees,
encap_txd_encap_fail.
Fixes:
Replace the broken collapse_pkthdr() implementation with an MPASS().
fl_refills and fl_refills_large were not incremented when using netmap.
andrew [Thu, 6 Sep 2018 17:25:50 +0000 (17:25 +0000)]
Fix the GIC ACPI cross reference value.
To support INTRNG with ACPI we need to set a non-zero cross reference value
for the interrupt controller. The GICv3 driver already had this value set,
however it was missed in the GICv2 driver. Fix this by setting xref to the
correct value.
andrew [Thu, 6 Sep 2018 17:25:01 +0000 (17:25 +0000)]
Remove the check that the Arm generic interrupt controller variant is
non-zero. This is the case on qemu, so remove it to allow us to boot there.
This change is needed to boot on qemu with ACPI.
This patch adds the very initial support for HTM that might come at FreeBSD
version 12.1. This basic support defines a new kABI, so, we do not need to change
it later during 12.1 time frame, when the full implementation will come.
Testing m->queue != PQ_NONE is not sufficient; see the commit log
message for r338276. As of r332974 vm_page_dequeue() handles
already-dequeued pages, so just replace vm_page_remque() calls with
vm_page_dequeue() calls.
Reviewed by: kib
Tested by: pho
Approved by: re (marius)
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17025
Make tcp_hpts.c compile a LINT kernel with options RSS and PCBGROUPS added by
adding the missing include files and changing a the type of cpuid which
would otherwise cause a false comparison with NETISR_CPUID_NONE.
Reviewed by: rrs
Approved by: re (marius)
Differential Revision: https://reviews.freebsd.org/D16891
Add support for receive side scaling stride, RSSS, in mlx5en(4).
The receive side scaling stride parameter is a value which define the interval
between active receive side queues. The traffic for the inactive queues is
redirected to the nearest active queue by use of modulus. The default value
of this parameter is one, which means all receive side queues are used.
The point of this feature is to redirect more traffic to fewer receive side
queues in order to take more advantage of sorted large receive offload,
sorted LRO. The sorted LRO works better when more packets are accumulated
per service interval.
MFC after: 3 days
Approved by: re (marius)
Sponsored by: Mellanox Technologies
ibcore: Fix endless loop in searching for matching VLAN device
In r337943 ifnet's if_pcp was set to the PCP value in use
instead of IFNET_PCP_NONE.
Current ibcore code assumes that if_pcp is IFNET_PCP_NONE with
VLAN interfaces so it can identify prio-tagged traffic.
Fix that by explicitly verifying that that the if_type is IFT_ETHER
and not IFT_L2VLAN.
MFC after: 3 days
Approved by: re (Marius), hselasky (mentor), kib (mentor)
Sponsored by: Mellanox Technologies
Rework rtld's TLS Variant I implementation to match r326794
The above commit fixed handling overaligned TLS segments in libc's
TLS Variant I implementation, but rtld provides its own implementation
for dynamically-linked executables which lacks these fixes. Thus,
port these changes to rtld.
This was previously commited as r337978 and reverted in r338149 due to
exposing a bug the ARM rtld. This bug was fixed in r338317 by mmel.
MIPS64 does not store the 'r_info' field of a relocation table entry as
a 64-bit value consisting of a 32-bit symbol index in the high 32 bits
and a 32-bit type in the low 32 bits as on other architectures. Instead,
the 64-bit 'r_info' field is really a 32-bit symbol index followed by four
individual byte type fields. For big-endian MIPS64, treating this as a
64-bit integer happens to be compatible with the layout expected by other
architectures (symbol index in upper 32-bits of resulting "native" 64-bit
integer). However, for little-endian MIPS64 the parsed 64-bit integer
contains the symbol index in the low 32 bits and the 4 individual byte
type fields in the upper 32-bits (but as if the upper 32-bits were
byte-swapped).
To cope, add two helper routines in gelf_getrel.c to translate between the
correct native 'r_info' value and the value obtained after the normal
byte-swap translation. Use these routines in gelf_getrel(), gelf_getrela(),
gelf_update_rel(), and gelf_update_rela(). This fixes 'readelf -r' on
little-endian MIPS64 objects which was previously decoding incorrect
relocations as well as 'objcopy: invalid symbox index' warnings from
objcopy when extracting debug symbols from kernel modules.
Even with this fixed, objcopy was still crashing when trying to extract
debug symbols from little-endian MIPS64 modules. The workaround in
gelf_*rel*() depends on the current ELF object having a valid ELF header
so that the 'e_machine' field can be compared against EM_MIPS. objcopy
was parsing the relocation entries to possibly rewrite the 'r_info' fields
in the update_relocs() function before writing the initial ELF header to
the destination object file. Move the initial write of the ELF header
earlier before copy_contents() so that update_relocs() uses the correct
symbol index values.
Note that this change should really go upstream. The binutils readelf
source has a similar hack for MIPS64EL though I implemented this version
from scratch using the MIPS64 ABI PDF as a reference.
Be a little conservative about when to force size optimizations.
Reports have come in that there's issue with powerpc and sparc64 since
we've switched to using -Oz / -Os. We don't strictly need them for
!x86, so be conservative about when we enable them.
Correct the condition under which we allocate a terminator node.
We will have last_block < blocks if the block count is divisible
by BLIST_BMAP_RADIX, but a terminator node is still needed if the
tree isn't balanced. In this case we were overruning the blist
array by 16 bytes during initialization.
While here, add a check for the invalid blocks == 0 case.
PR: 231116
Reviewed by: alc, kib (previous version), Doug Moore <dougm@rice.edu>
Approved by: re (gjb)
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17020
Reviewed by: bz, Johannes Lundberg <johalun0@gmail.com>
Approved by: re (rgrimes)
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D17030
eugen [Wed, 5 Sep 2018 13:59:36 +0000 (13:59 +0000)]
Fix "ipfw fwd" to work for incoming IPv4 packets when ip_tryforward() chooses
fast forwarding path, as it already works for IPv6 and for both of them
on old slow path.
amd64: For non-PTI mode, do not initialize PCPU kcr3 to KPML4phys.
Non-PTI mode does not switch kcr3, which means that kcr3 is almost
always stale. This is important for the NMI handler, which reloads
%cr3 with PCPU(kcr3) if the value is different from PMAP_NO_CR3.
The end result is that curpmap in NMI handler does not match the page
table loaded into hardware. The manifestation was copyin(9) looping
forever when a usermode access page fault cannot be resolved by
vm_fault() updating a different page table.
Reported by: mmacy
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
MFC after: 3 days
Approved by: re (gjb)
r337289 has a side effect of reducing usb frame 0 buffer size down to
touch report size. That broke some devices e.g. "Raydium Touch System"
which are capable of generating non-touch frames of bigger length.
Fix it with enlarging frame 0 buffer up to internal wmt(4) buffer size.
As discussed in D6262 post-commit review, change inp_route to
inp_route6 for IPv6 code after r301217.
This was most likely a c&p error from the legacy IP code, which
did not matter as it is a union and both structures have the same
layout at the beginning.
No functional changes.
This was disabled recently due to lack of support in KDB disassembler
and DTrace FBT provider. Support for 'C'-extension to both of these was
added, so we can now enable 'C'-extension.
This reduces size of the kernel important for low-end embedded devices,
and saves cache footprint for high perfomance machines.
The kernel DTrace audit provider (dtaudit) relies on auditd(8) to load
/etc/security/audit_event to provide a list of audit event-number <->
name mappings. However, this occurs too late for anonymous tracing.
With this change, adding 'audit_event_load="YES"' to /boot/loader.conf
will cause the boot loader to preload the file, and then the kernel
audit code will parse it to register an initial set of audit event-number
<-> name mappings. Those mappings can later be updated by auditd(8) if
the configuration file changes.
This appeared to be required to have EFI RT support and EFI RTC
enabled by default, because there are too many reports of faulting
calls on many different machines. The knob is added to leave the
exceptions unhandled to allow to debug the actual bugs.
Reviewed by: kevans
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Approved by: re (rgrimes)
Differential revision: https://reviews.freebsd.org/D16972
Improve error messages from clock_if.m method failures.
Print error message in verbose mode when CLOCK_SETTIME() clock_if.m
method failed. For EFIRT RTC clock, add error code for the failure of
CLOCK_GETTIME() report.
Reviewed by: kevans
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Approved by: re (rgrimes)
Differential revision: https://reviews.freebsd.org/D16972
Swap order of dererencing PCPU curpmap and checking for usermode in
trap_pfault() KPTI violation check.
EFI RT may set curpmap to NULL for the duration of the call for some
machines (PCID but no INVPCID). Since apparently EFI RT code must be
ready for exceptions from the calls, avoid dereferencing curpmap until
we know that this call does not come from usermode.
Reviewed by: kevans
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Approved by: re (rgrimes)
Differential revision: https://reviews.freebsd.org/D16972
Recent changes have created, for the first time, physical memory segments
that can be coalesced. To be clear, fragmentation of phys_avail[] is not
the cause. This fragmentation of vm_phys_segs[] arises from the "special"
calls to vm_phys_add_seg(), in other words, not those that derive directly
from phys_avail[], but those that we create for the initial kernel page
table pages and now for the kernel and modules loaded at boot time. Since
we sometimes iterate over the physical memory segments, coalescing these
segments at initialization time is a worthwhile change.