If a type=dir entry exists and all contents are directories, files
added with contents=, or symlinks with link= attributes then it doesn't
need to exist. Just let openat fail in that case. It's conceivable
this will make debugging some cases weird, but it's sufficent to handle
the way we add /root/.ssh in CheriBSD VM images.
This is a recommit of 794154149f with
bugfixes.
Reviewed by: markj
Differential Revision: https://reviews.freebsd.org/D38029
If a type=dir entry exists and all contents are directories, files
added with contents=, or symlinks with link= attributes then it doesn't
need to exist. Just let openat fail in that case. It's conceivable
this will make debugging some cases weird, but it's sufficent to handle
the way we add /root/.ssh in CheriBSD VM images.
Reviewed by: markj
Differential Revision: https://reviews.freebsd.org/D38029
When a link target is specified use it rather than attempting to read
a potentially non-existant file.
Reviewed by: markj
Differential Revision: https://reviews.freebsd.org/D38028
This reverts commit c33509d49a.
It turns out that the long 27 second delay I saw in the
gss_acquire_cred() call was caused by a (mis)configured
DNS. Although I did not specify "dns" in /etc/nsswitch.conf,
I did have a /etc/resolv.conf file on the system (left
there by wpa_supplicant). As such, with no route, it was
somehow trying to contact the DNS server, although there was none.
Once I got rid of the /etc/resolv.conf file, it worked
as expected.
Since there is now a large 5 minute timeout on the
kernel to gssd(8) upcalls, the gssd(8) daemon will not
get terminated when this delay occurs and the only affect
is a 30 second delay during the mount.
Discussed with: bjk
* Replay 2010[acflm] which had been merged but not recorded.
* Merge 2010n.
* Reorganize (unsplit) the code to match the upstream layout.
* Merge 2022[cdefg].
MFC after: 1 week
Sponsored by: Klara, Inc.
If an NFS mount using "sec=krb5[ip],gssname=<name>" is
done, the gssd daemon fails. There is a long delay
(several seconds) in the gss_acquire_cred() call and then
it returns success, but the credentials returned are
junk.
I have no idea how long this has been broken, due to some
change in the Heimdal gssapi library call, but I suspect
it has been quite some time.
Anyhow, it turns out that replacing the "desired_name"
argument with GSS_C_NO_NAME fixes the problem.
Replacing the argument should not be a problem, since the
TGT for the host based initiator credential in the default
keytab file should be the only TGT in the gssd'd credential
cache (which is not the one for uid 0).
I will try and determine if FreeBSD13 and/or FreeBSD12
needs this same fix and will MFC if they need the fix.
This problem only affected Kerberized NFS mounts when the
"gssname" mount option was used. Other Kerberized NFS
mount cases already used GSS_C_NO_NAME and work ok.
A workaround if you do not have this patch is to do a
"kinit -k host/FQDN" as root on the machine, followed by
the Kerberized NFS mount without the "gssname" mount
option.
MFC after: 1 month
The intention of /etc/passwd was to support legacy applications that are
not yet converted to use modern API like getpwent(3). Comments are not
defined in the legacy format, so copying them could break these
applications. Plus, it could leak sensitive information (e.g. encrypted
form of password of an user that was commented out instead of deleted
or disabled).
PR: bin/144652
MFC after: 1 month
Although the tls syscall to set up the upcall should
not normally fail, the daemon should check for such
a failure. This patch adds a check for that failure.
MFC after: 1 week
Although the tls syscall to set up the upcall should
not normally fail, the daemon should check for such
a failure. This patch adds a check for that failure.
MFC after: 1 week
Set the thread affinity in fbsdrun_start_thread next to where the
thread name is set. This keeps all the pthread initialization
operations at the start of a thread in one place.
Reviewed by: corvink, markj
Differential Revision: https://reviews.freebsd.org/D37646
The global vmexit[] array is no longer needed to smuggle the rip
value from fbsdrun_addcpu() to vm_loop().
Reviewed by: corvink, markj
Differential Revision: https://reviews.freebsd.org/D37644
fbsdrun_addcpu() read the current vCPU's RIP register from the kernel
via vm_get_register() to pass along through some layers to vm_loop()
which then set the register via vm_set_register(). However, this is
just always setting the value back to itself.
Reviewed by: corvink
Differential Revision: https://reviews.freebsd.org/D37643
- Enable VM_CAP_IPI_EXIT in fbsdrun_set_capabilities along with other
capabilities enabled on all vCPUs.
- Don't call fbsdrun_set_capabilities a second time on the BSP in
spinup_vcpu.
- To preserve previous behavior, don't unconditionally enable
unrestricted guest mode on the BSP (this unbreaks single-vCPU guests
on Nehalem systems, though supporting such setups is of dubious
value). Other places that enbale UG on the BSP are careful to check
the result of the operation and fail if it is not available.
- Don't set any capabilities in spinup_ap(). These are now all
redundant with earlier settings from spinup_vcpu().
- While here, axe a stale comment from fbsdrun_addcpu(). This
function is now always called from the main thread for all vCPUs.
Reviewed by: corvink, markj
Differential Revision: https://reviews.freebsd.org/D37642
Since commit 0bda8d3e9f, bhyve always enables VM_EXITCODE_IPI exits
instead, so this handler is no longer used.
Reviewed by: corvink, markj
Differential Revision: https://reviews.freebsd.org/D37640
Commit bba7a2e896 added "allow.nfsd" to optionally allow
mountd/nfsd to be run inside a vnet prison when the kernel
is built with "options VNET_NFSD".
This patch updates the man page for this change.
This is a content change.
Reviewed by: jamie, bcr (manpages)
MFC after: 4 months
Differential Revision: https://reviews.freebsd.org/D37665
dma(8) supports mailq, but not mailq -Ac to print the submission
queue. Don't try to print that queue from the daily script if
mailq -Ac returns an error.
Reviewed by: bapt
Differential Revision: https://reviews.freebsd.org/D37713
dma(8) does not have hoststat or purgestat, so this script produces
an error from the daily script. We could disable this script, but
that would mean yet another change to switch back to sendmail. Check
for purgestat in mailer.conf before attempting either hoststat or
purgestat.
Reviewed by: pstef, bapt
Differential Revision: https://reviews.freebsd.org/D37712
These ioctls are not vCPU-specific and the ioctl now ignores the vCPU
ID. 0 is used instead of -1 to provide limited forwards
compatibility.
Reviewed by: corvink, markj
Differential Revision: https://reviews.freebsd.org/D37651
This allows the use of chroot and/or jail environments which depend on
interpreters registed with imgact_binmisc to use emulator binaries from
the host to emulate programs inside the chroot.
Reviewed by: imp
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D37432
ppp supports MSS clamping for TCP/IPv4. This patch
* improves MSS clamping for TCP/IPv4 by using the MSS as specified
in RFC 6691.
* adds support for MSS clamping for TCP/IPv6.
Reported by: Timo Voelker
Reviewed by: thj
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D37624
OVMF ships some static ACPI tables. This worked in the past but won't
work in the future when we support devices like tpms. They require a TPM
ACPI table. So, we have to dynamically create ACPI tables depending on
the bhyve configuration.
Bhyve has much more information about the system than OVMF. Therefore,
it's easier for bhyve to build up some ACPI tables. For that reason, it
would be much better to use the ACPI tables provided by bhyve instead of
building some tables by OVMF.
At the moment, OVMF always creates a SPCR table. Maybe someone depends
on it. So, we have to build it by bhyve too before we can patch OVMF to
install the tables provided by bhyve.
Reviewed by: markj
MFC after: 1 week
Sponsored by: Beckhoff Automation GmbH & Co. KG
Differential Revision: https://reviews.freebsd.org/D37591
When a boot environment has been booted via the bootonce feature,
userboot clears the bootonce value from an nvlist but fails to write the
updated nvlist back to disk.
The failure occurs because bhyveload opens the guest boot disk image
O_RDONLY, fix this by opening it O_RDWR.
Reviewed by: imp, markj, jhb
Differential Revision: https://reviews.freebsd.org/D37274
firewire.h includes zero length arrays in unions that trigger this
warning.
Reviewed by: imp, emaste
Differential Revision: https://reviews.freebsd.org/D37538
This function intentionally saves a pointer to an on-stack variable in
a global as a dubious way of reading the stack pointer.
Reviewed by: imp, emaste
Differential Revision: https://reviews.freebsd.org/D37535
Make 'line' static to move it to .bss instead as that pattern is used
elsewhere in pw(8) (e.g. the static buffer in pw_pwcrypt).
Reported by: GCC -Wdangling-pointer
Reviewed by: imp, emaste
Differential Revision: https://reviews.freebsd.org/D37534
Before this ioctl frontend always replaced tags with sequential ones.
It was done for ctladm, that can not keep track of global tag list.
But in case of virtio-scsi in bhyve we can pass provided tags as-is.
It should be on virtio-scsi initiator to provide us valid tags. It
should allow proper task management, error reporting, etc. In case
of several virtio-scsi devices, they should use different CTL ports
or initiator IDs to avoid conflicts, but this is expected by design.
PR: 267539
SAM-5 specification states maximum size of command identifier (tag),
defined by specific transports, should not be larger than 64 bits.
While most of supported transports use 32 bits or less, it was
reported that virtio-scsi uses 64 bits. Truncation to 32 bits in
bhyve code caused false tag conflict errors reported and possibly
other issues.
This changes CTL ABI and HA protocol, so CTL_HA_VERSION is bumped.
While we make HA protocol incompatible, increase default maximum
number of ports in CTL from 256 to 1024, matching number of LUNs.
There are many reports from people who need many iSCSI targets with
only one LUN each. Increased memory consumption should be less of
a problem these days.
PR: 267539
trpt(8) was utility to pull TCP debugging data from the kernel
originating back from 4.2BSD. It is not used nowadays by TCP
developers. We have more powerful debugging facilities, e.g.
the Dtrace probing, the TCP black box logging and siftr.
Discussed with: rscheff, tuexen, rrs, jtl and others
The variable used for the checklist's default value needs to correspond
to the rc.conf variable as that's what's being parsed to determine them.
In the case of local_unbound it's missing the _enable suffix and thus
always defaults to off on revisit.
Fixes: 58eb9abb31 ("Add a line to the post-installation configuration dialog to enable the local_unbound service.")
There are a few issues here, some of which are hiding others. The first
is that we don't use double quotes around the command substitution so
every word in the conf file is treated as a separate argument to eval,
resulting in spaces being used in place of newlines and thus comments in
the file commenting out the rest of the file, not just to the end of
their line. In particular, we insert one comment just before the dumpdev
entry (the final one in the file) and so we never see dumpdev as set,
and thus set a default value of on for the menu.
The second issue is that, for dumpdev, it takes a value of AUTO not YES
when set, but we don't replace this with on when eval'ing, so then end
up giving AUTO to bsddialog which is interpreted the same as off (which
seems to match GPL dialog). Thus handle AUTO like YES otherwise it will
always appear as unchecked on revisit.
The final issue is that our case-insensitive YES/NO (and now AUTO)
replacements have no word boundaries around them so match the middle of
words too. As it happens this doesn't matter in practice at the moment,
but it could in future; currently the only effect is that it rewrites
moused_nondefault_enable to moused_offndefault_enable, but since this
variable is never read, only written based on moused(_enable) this is
harmless, but we should fix it in case a service comes along in future
that does get affected by it.
This installer option is currently totally useless, as it ends up
creating an ntpd_sync_on_start_enable="YES" entry in rc.conf, not an
ntpd_sync_on_start="YES" entry, as is the correct name. This can also be
noticed by revisiting the services menu, which parses the previously
written rc.conf.services file to set variables governing the default
menu entry values so that selecting OK regenerates the same file, as the
menu entry will use the correct variable name and thus think the entry
was not selected last time, defaulting back to off and losing the
setting.
Thus, add a special case in the loop for this option. The only other
entry that doesn't follow the *_enable pattern is dumpdev (even moused
does, it just also sets a second variable), but that also deviates in
terms of being explicitly set either way and using AUTO rather than YES,
hence why ntpd_sync_on_start follows a different pattern here and is
special-cased rather than introducing a whole new variable that governs
behaviour outside the loop.
Fixes: c153a35bfd ("bsdinstall: replace ntpdate by ntpd_sync_on_start")
All of the error paths in pci_vtcon_sock_add free the sock pointer.
However, sock is not initialized until part way through the function.
An early error would pass stack garbage to free().
Reviewed by: corvink, markj
Differential Revision: https://reviews.freebsd.org/D37491