textproc/uim: Update to 1.9.7
- Remove textproc/uim-gtk2 because the upstream dropped gtk2 support
- Add textproc/uim-gtk4
- Bump PORTREVISION of textproc/uim consumers
powerpc/radix: acquire the pmap lock in mmu_radix_extract()
mmu_radix_extract() walks the page tables without holding the pmap lock,
unlike its hash MMU counterpart moea64_extract(). A concurrent unmap can
free and recycle the page table page being walked, so the read returns
whatever now occupies that memory and the caller gets a physical address
that never existed.
That is how mmu_radix_sync_icache() came to hand a bogus address to
__syncicache() and panic the machine. Commit 1574ca1955f5 worked around
it by taking the pmap lock in mmu_radix_sync_icache(), but the machine
independent callers of pmap_extract() - vm_sync_icache(), proc_rwmem()
and the vslock() paths - remain exposed to the same failure.
Rename the existing body to mmu_radix_extract_locked(), which asserts the
lock, and make mmu_radix_extract() a thin wrapper that acquires it.
mmu_radix_sync_icache() already holds the pmap lock, so it calls the
locked variant directly and neither recurses nor reacquires the lock once
per page.
[7 lines not shown]
powerpc/radix: take the pmap lock in mmu_radix_sync_icache()
mmu_radix_sync_icache() walked the page tables with an unlocked
pmap_extract() and passed the result straight to PHYS_TO_DMAP(),
checking only that it was non-zero. Nothing keeps the mapping - or the
page table page holding it - alive across that window: if another thread
of the same process tears a mapping down concurrently, the page table
page can be freed and reused, so pmap_extract() reads arbitrary memory
and returns a bogus physical address. __syncicache() then dereferences
an unmapped direct map address and the kernel takes a data storage
interrupt:
fatal kernel trap:
exception = 0x300 (data storage interrupt)
virtual address = 0xc003317ca6022a00
dsisr = 0x40000000
srr0 = 0xc000000000f59460 (__syncicache)
lr = 0xc000000000f23588 (mmu_radix_sync_icache)
pid = 23878, comm = skyframe-evaluator-
[27 lines not shown]
powerpc/radix: acquire the pmap lock in mmu_radix_extract()
mmu_radix_extract() walks the page tables without holding the pmap lock,
unlike its hash MMU counterpart moea64_extract(). A concurrent unmap can
free and recycle the page table page being walked, so the read returns
whatever now occupies that memory and the caller gets a physical address
that never existed.
That is how mmu_radix_sync_icache() came to hand a bogus address to
__syncicache() and panic the machine. Commit 1574ca1955f5 worked around
it by taking the pmap lock in mmu_radix_sync_icache(), but the machine
independent callers of pmap_extract() - vm_sync_icache(), proc_rwmem()
and the vslock() paths - remain exposed to the same failure.
Rename the existing body to mmu_radix_extract_locked(), which asserts the
lock, and make mmu_radix_extract() a thin wrapper that acquires it.
mmu_radix_sync_icache() already holds the pmap lock, so it calls the
locked variant directly and neither recurses nor reacquires the lock once
per page.
[7 lines not shown]
powerpc/radix: take the pmap lock in mmu_radix_sync_icache()
mmu_radix_sync_icache() walked the page tables with an unlocked
pmap_extract() and passed the result straight to PHYS_TO_DMAP(),
checking only that it was non-zero. Nothing keeps the mapping - or the
page table page holding it - alive across that window: if another thread
of the same process tears a mapping down concurrently, the page table
page can be freed and reused, so pmap_extract() reads arbitrary memory
and returns a bogus physical address. __syncicache() then dereferences
an unmapped direct map address and the kernel takes a data storage
interrupt:
fatal kernel trap:
exception = 0x300 (data storage interrupt)
virtual address = 0xc003317ca6022a00
dsisr = 0x40000000
srr0 = 0xc000000000f59460 (__syncicache)
lr = 0xc000000000f23588 (mmu_radix_sync_icache)
pid = 23878, comm = skyframe-evaluator-
[27 lines not shown]
net-mgmt/nagios-plugins: fix SIGSEGV
The code in /usr/local/libexec/nagios/check_http calls
die(STATE_CRITICAL, NULL) but die() can cause SIGSEGV due to
missing check for NULL. Add the check to not dereference NULL.
PR: 298716
misc/clifm: update CliFM to version 1.29 (Stan Stanman)
Bump color schemes to the latest commit, drop no longer
needed MANDIR override, and amend CATEGORIES.
Reported by: portscout
sysutils/fsearch: the port had been updated to version 0.3.2
This release attempts to fix most of the reported bugs so far.
It's also improving performance drastically when the database
index is being updated. While here, drop LICENSE_FILE which
is a standard boilerplate without real copyright information.
Reported by: portscout
devel/py-tavern: new port
Tavern is a pytest plugin and command-line tool for automated testing of
RESTful APIs. Tests are written in YAML, describing the requests to send
and the responses to expect, so API behaviour can be verified without
writing Python code.
It also supports MQTT, GraphQL and gRPC testing through optional extras.
Co-authored-by: Michael Osipov <michaelo at FreeBSD.org>
PR: 298695
x11-clocks/buici-clock: update 0.4.9.4 -> 0.4.10
Major changes:
- integrate remaining debian patches
- use timeout instead of signal to make the clock tick
- switch upstream repo for my GH repo
For the record, buici-clock was originally created by a Debian committer,
and its source has since been maintained in Debian’s Git repository.
The FreeBSD port has been using that repository as its upstream source.
My GitHub repository is now the authoritative upstream for the Debian package.
Therefore, the FreeBSD port should be updated to use it as well, rather than
continuing to track the Debian Git repository.
vfs_lookup_cross_mount(): fix missing LK_CANRECURSE
This was a rather dumb miss on my part in commit 42442d7a6e.
LK_CANRECURSE is clearly needed in any case in which the covered vnode
is held exclusive across the call to VFS_ROOT(), regardless of whether
it was initially held exclusive or upgraded. The commit message for
that change also noted that unionfs lookup only worked without
LK_CANRECURSE due to a coincidence of the then-current unionfs
implementation. As it happens, said coincidence was recently removed
in commit b952606b4f ("unionfs_lock(): eliminate LK_CANRECURSE special-
case").
PR: 298201
Reported by: olivier
Fixes: 42442d7a6e "Generalize the VV_CROSSLOCK logic in
vfs_lookup"
Reviewed by: kib, markj, pho
Tested by: pho
Differential Revision: https://reviews.freebsd.org/D59494
[2 lines not shown]
vfs_lookup_cross_mount(): fix missing LK_CANRECURSE
This was a rather dumb miss on my part in commit 42442d7a6e.
LK_CANRECURSE is clearly needed in any case in which the covered vnode
is held exclusive across the call to VFS_ROOT(), regardless of whether
it was initially held exclusive or upgraded. The commit message for
that change also noted that unionfs lookup only worked without
LK_CANRECURSE due to a coincidence of the then-current unionfs
implementation. As it happens, said coincidence was recently removed
in commit b952606b4f ("unionfs_lock(): eliminate LK_CANRECURSE special-
case").
PR: 298201
Reported by: olivier
Fixes: 42442d7a6e "Generalize the VV_CROSSLOCK logic in
vfs_lookup"
Reviewed by: kib, markj, pho
Tested by: pho
Differential Revision: https://reviews.freebsd.org/D59494
[2 lines not shown]
www/py-binarycookies: update the port to version 2.5.0
- Fix serializer so written files match Apple's format
- Preserve full cookie fidelity on round trips
Reported by: portscout
net-p2p/libtorrent-rasterbar2: revert "net-p2p/libtorrent-rasterbar2: update 2.0.13 → 2.1.1"
There are run-time issues that need to be investigated.
This reverts commit 7c412d80f232eb225c1d48ff72d5fab1dd8769e9.
www/immich: Set the machine learning URL
Immich defaults to the Docker hostname immich-machine-learning, which
never resolves outside a container, so the service ran unnoticed while
the server reported it unhealthy. The rc script now sets
IMMICH_MACHINE_LEARNING_URL, adjustable with immich_server_ml_url.
Sponsored by: Netzkommune GmbH
www/linux-brave: DRM out of the box, GPU auto-detection
Add a WIDEVINE option (default on) so DRM playback works with no
browser-side setup once the CDM is present. The option links
${PREFIX}/lib/WidevineCdm into the Brave install directory, where the
component updater discovers it as a preinstalled component, and runs a
pre-launch seed script that sets brave.widevine_opted_in and maintains
the Widevine component hint file: Chromium registers the CDM at
startup from the hint and never repairs it itself, so the script
plants the hint before the first launch and repoints it when its
target no longer holds a CDM (e.g. after the CDM port is reinstalled),
while leaving intact hints -- including one naming a newer
component-updated CDM -- alone.
The option adds no dependency: Google's license permits fetching the
CDM but not redistributing it (www/linux-widevine-cdm is
no-pkg-mirror), so a RUN_DEPENDS would keep linux-brave itself off the
package mirrors. Everything the option installs is inert until the
user builds www/linux-widevine-cdm from ports -- the symlink dangles
[39 lines not shown]
nfsuserd.c: Fix handling where pw_name/gr_name differ from lookup name
When an NSS backend returns a canonical pw_name or gr_name that differs from
the lookup name supplied by the NFSv4 upcall, nfsuserd stores the successful
mapping in the kernel cache under the canonical name instead of the requested
name.
This causes the retry lookup performed by nfsv4_strtouid() or
nfsv4_strtogid() to miss the newly inserted cache entry, resulting in the
default UID/GID being returned although the NSS lookup itself succeeded.
PR: 296753
(cherry picked from commit 1771ab245c2341033f0cee3bd098c76888cf3515)
