asomers@ reported a crash on an NFSv4.0 server with a backtrace of:
kdb_backtrace
vpanic
panic
nfsrv_docallback
nfsrv_checkgetattr
nfsrvd_getattr
nfsrvd_dorpc
nfssvc_program
svc_run_internal
svc_thread_start
fork_exit
fork_trampoline
where the panic message was "docallb", which indicates that a callback
was attempted when the ClientID is unconfirmed.
This would not normally occur, but it is possible to have an unconfirmed
ClientID structure with delegation structure(s) chained off it if the
client were to issue a SetClientID with the same "id" but different
"verifier" after acquiring delegations on the previously confirmed ClientID.
The bug appears to be that nfsrv_checkgetattr() failed to check for
this uncommon case of an unconfirmed ClientID with a delegation structure
that no longer refers to a delegation the client knows about.
This patch adds a check for this case, handling it as if no delegation
exists, which is the case when the above occurs.
Although difficult to reproduce, this change should avoid the panic().
PR: 249127
Reported by: asomers
Reviewed by: asomers
MFC after: 1 week
Differential Revision: https://reviews.freebbsd.org/D26342
To make it easier to work with this in the future, convert to c99
designated initializer syntax.
Tested on powerpc, powerpc64, and powerpc64le. No functional change.
Sponsored by: Tag1 Consulting, Inc.
The intention of the bus_be naming was for those to be the no-endian-swapping
and for the bus_le to be endian-swapping in all the functions.
This naming breaks down when we're actually are running in LE and need to
use the opposite sense.
As such, rename bs_be_* to native_bs_* and rename bs_le_* to swapped_bs_*.
No functional change.
Sponsored by: Tag1 Consulting, Inc.
Swap the BE and LE bus_space tags when on LE, and adjust the nexus tag
to match.
This is prep for a a followup that makes the powerpc bus_space macros easier
to maintain in the future.
Sponsored by: Tag1 Consulting, Inc.
Prior to this commit, EC2 AMIs used a "dual-dhclient" tool which was
launched in place of dhclient and spawned both the base system dhclient
for IPv4 and the ISC dhclient from ports for IPv6.
Now that rtsold supports the "M bit" (managed configuration), we can go
back to having the base system dhclient spawned normally, and provide a
script to rtsold which spawns the ISC dhclient from ports when rtsold
decides that it is appropriate.
Thanks to: bz
MFC after: 1 week
Sponsored by: https://www.patreon.com/cperciva
The FreeBSD base system continues to expand. 4GB is now insufficient;
we passed 3 GB in May 2019; we passed 2 GB in August 2017. Over half
of the disk space used is in /usr/lib/debug/.
Without this change, instances boot but are unusable, since the first
thing which breaks when VM filesystems are too small is the "pkg install"
in the VM building process.
Implement pmap_mincore() for moea64.
This will need some slight tweaks when large page support in HPT lands.
Submitted by: Fernando Eckhardt Valle <fernando.valle@eldorado.org.br>
Reviewed by: bdragon
Differential Revision: https://reviews.freebsd.org/D26314
Fixes for Raspberry Pi 4B PCIe / USB:
- Pass through a DMA tag for the controller.
- In theory the controller can access the lower 3 GB, but testing found
that unreliable. OpenBSD also restricts DMA to the lowest 960 MiB.
- Rename some constants to be a bit more meaningful.
Submitted by: Robert Crowston, crowston at protonmail.com
Reviewed by: mkarels, outside reviewers
Differential Revision: https://reviews.freebsd.org/D26344
Teach clang that powerpc64le-*-freebsd* is a valid triple.
This is already in upstream clang, but was too late for llvm 11.0.0.
Apply it directly for now, until it can be backported to llvm 11.0.1.
See upstream https://reviews.llvm.org/D73425 for details.
Reviewed by: dim (upstream version), emaste
Approved by: emaste
Differential Revision: https://reviews.freebsd.org/D26400
The floppy test passes with this. The others fail due to 'integrity
checks' failing in GPART. It's not at all clear those integrity
checks are legit or if the test samples were bogusly generated
by FreeBSD.
For configurations without x2APIC support (guests, older hardware), the global
LAPIC MMIO mapping will trigger false-positive KCSan reports as it will appear
that multiple CPUs are concurrently reading and writing the same address.
This isn't actually true, as the underlying physical access will be performed
on the local CPU's APIC. Additionally, because LAPIC access can happen during
event timer configuration, the resulting KCSan printf can produce a panic due
to attempted recursion on event timer resources.
Add a __nosanitizethread preprocessor define to prevent the compiler from
inserting TSan hooks, and apply it to the x86 LAPIC accessors.
PR: 249149
Reported by: gbe
Reviewed by: andrew, kib
Tested by: gbe
Differential Revision: https://reviews.freebsd.org/D26354
This update adds support for:
HW VLAN tagging
HW checksum offload for IPv4 and IPv6
tx and rx aggreegation (for full gige speeds)
multiple transactions
In my testing, I am able to get 900-950Mbps depending upon
TCP or UDP, which is a significant improvement over the previous
91Mbps (~8kint/sec*1500bytes/packet*1packet/int).
Reviewed by: hselasky
MFC after: 2 months
Differential Revision: https://reviews.freebsd.org/D25809
freebsd-update(8) builds, where BRANCH is suffixed with -p0 for
builds.
Noticed by: gordon
With help from: cperciva
MFC after: 3 days
MFC note: before 12.2-BETA2
Sponsored by: Rubicon Communications, LLC (netgate.com)
These images are no longer relevant... However, I've also not tested
the regression test here to see if it still works or not... It needs
a lot of love regardless...
Even though copy_file_range has a file-system agnostic version, it still
fails on devfs (perhaps because the file descriptor is non-seekable?) In
that case, fallback to old-fashioned read/write. Fixes
"cp /dev/null /tmp/null"
PR: 249248
Reported by: Michael Butler
Reviewed by: mjg
MFC-With: 365549
Differential Revision: https://reviews.freebsd.org/D26395
Two more cases of explicitly marking globals for internal linkage where they
need not be shared. Committed upstream as of a38e62314a1f.
MFC after: 1 week
In r365419 ieee80211_media_change() callers were updated to not longer
act on the obselete ENETRESET return code.
While in the old days iwm has done a stop/init cycle in these cases,
this was not executed since r193340.
As a consequence simplify iwm code as well by passing ieee80211_media_change()
right to ieee80211_vap_attach() as there is no more need for a local
implementation.
Reported by: Tomoaki AOKI (junchoon dec.sakura.ne.jp)
Tested by: Tomoaki AOKI (junchoon dec.sakura.ne.jp)
MFC after: 3 days
X-MFC: fix is already in stable/12
PR: 248955
This would allow interested parties to do experimental runs with an
environment set appropriately to raise all the warnings throughout the
build; e.g. env WARNS=6 NO_WERROR=yes buildworld.
Not currently touching the numerous instances in ^/tools.
MFC after: 1 week
Properly handle the case, when some opcode keywords follow after
the `frag` opcode without additional options.
Reported by: Evgeniy Khramtsov <evgeniy at khramtsov org>
Return values are passed in a0, so read it from there. We also pass a1 through
to userspace, as the ABI allows small structs to be returned in registers
a0/a1. While here read the register values directly from the trapframe rather
than rtval, and remove the now unneeded argument from dtrace_invop().
Set fbtp_roffset so that we get the correct return location in arg0.
Reviewed by: markj
Sponsored by: Axiado
Differential Revision: https://reviews.freebsd.org/D26389
This also fixes a minor issue that was missed in the initial review; the
layout of the MFD_HUGE_* flags is actually not 1:1 bit:flag -- it instead
borrowed the Linux convention of how this is laid out since it was
originally implemented on Linux, the top 6 bits represent the shift required
for the requested page size.
This allows us to remove the flag <-> pgsize mapping table and simplify the
logic just prior to validation of the requested page size.
While we're here, fix two small nits:
- HUGETLB memfd shouldn't exhibit the SHM_GROW_ON_WRITE behavior. We can
only grow largepage shm by appropriately aligned (i.e. requested pagesize)
sizes, so it can't work in the typical/sane fashion. Furthermore, Linux
does the same, so let's be compatible.
- We don't allow MFD_HUGETLB without specifying a pagesize, so no need to
check for that later.
Reviewed by: kib (slightly earlier version)
Summary of changes:
- Assorted bug fixes
- Support for newer versions of the device firmware
- Suspend/resume support
- Support for Lenient Link Mode for E82X devices (e.g. can try to link with
SFP/QSFP modules with bad EEPROMs)
- Adds port-level rx_discards sysctl, similar to ixl(4)'s
This version of the driver is intended to be used with DDP package 1.3.16.0,
which has already been updated in a previous commit.
Tested by: Jeffrey Pieper <jeffrey.e.pieper@intel.com>
MFC after: 3 days
MFC with: r365332, r365550
Sponsored by: Intel Corporation
Differential Revision: https://reviews.freebsd.org/D26322
Literally returning EINVAL from a function designed to return an fd makes
for interesting scenarios.
I cannot assign enough pointy hats to cover this one.
On write with SHM_GROW_ON_WRITE, use proper truncate.
Do not allow to grow largepage shm if F_SEAL_GROW is set. Note that
shrinks are not supported at all due to unmanaged mappings.
Call to vm_pager_update_writecount() is only valid for swap objects,
skip it for unmanaged largepages.
Largepages cannot support write sealing.
Do not writecnt largepage mappings.
Reported by: kevans
Reviewed by: kevans, markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D26394