stand: Accept any device name in disk_parsedev()
disk_parsedev() only accepted names starting with "disk" or "vdisk", so other
DEVT_DISK devices using disk_fmtdev() could print a name they could not parse
back, e.g. "md0s2a:".
Skip any alphabetic prefix instead; devparse() has already matched it to the
devsw.
Reviewed by: imp
Approved by: imp
Sponsored by: Netflix
Differential Revision: https://ron-dev.freebsd.org/FreeBSD/src/pulls/97
www/nginx-module-ip-anon: Update to 0.2.0
Adds $remote_addr_hash, a keyed one-way pseudonym of the client
address. Logs can still tell clients apart and limit_req can act per
client, without an address being written anywhere. The hash is
HMAC-MD5 so it also works in builds without OpenSSL, and without a
secret the variable is not defined at all.
Sponsored by: Netzkommune GmbH
www/immich: Update to 3.4.1
Shared link cover queries no longer drag, .hif files are recognised as
HEIF, and the Prometheus exporter listens on all interfaces.
Changelog:
https://github.com/immich-app/immich/releases/tag/v3.4.1
Sponsored by: Netzkommune GmbH
loader.efi: Allocate ram disks as EfiReservedMemoryType
The kernel adds EfiLoaderData ranges to physmap and reuses the pages, so a
memdisk=root is overwritten while the system runs. The kernel skips
EfiReservedMemoryType, which ACPI requires for a registered RAM disk.
Fixes: 85fb40bad584
Reviewed by: imp
Approved by: imp
Sponsored by: Netflix
Differential Revision: https://ron-dev.freebsd.org/FreeBSD/src/pulls/96
net/tigervnc-viewer: Bump again to trigger rebuild against FLTK 1.4
PORTREVISION had been bumped at PR 296190 (320e46a) in the master port,
bumping to PORTREVISION=1 did not take effect. Bumping again.
snd_envy24ht: Fix out-of-bounds DMA buffer indexing
378e0c5c8967 reduced the number of playback channels from 5 to 1. This
introduced a regression, because the global DMA buffer indexing was not
adapted accordingly, which now results in kernel panics because of
out-of-bounds reads into the buffer. Additionally, because of wrong
indexing, recording channels ended up being serviced by the playback
part of envy24ht_intr().
Register 1 recording channel instead of 2 + sc->adcn, as the recording
DMA buffer only has room for 1 stereo stream. With 1 channel per
direction, give each channel the first index of its direction
(ENVY24HT_CHAN_PLAY_DAC1 and ENVY24HT_CHAN_REC_ADC1).
While here, fix the buffer wrap in envy24ht_r16sl(). It has to be
ch->size / ch->unit, but was copied verbatim from the 32-bit equivalent,
whose samples are twice as large, so 16-bit recording wrapped halfway
through the DMA buffer and never read the second half.
[8 lines not shown]
sysutils/fjord: Update to 0.4.3
- Require cni-epair 1.2.1, which no longer sends the host's name to
the DHCP server for every container
- Install sbin/fjord, a link to fjordd that runs its new CLI
- rc.d: start after appjail
Changelog: https://github.com/daemonless/fjord/releases/tag/v0.4.3
Reviewed by: dtxdf
Approved by: dtxdf (mentor)
Differential Revision: https://reviews.freebsd.org/D61442
devel/vectorscan: update to 5.4.14
- Update 5.4.13 to 5.4.14
- Drop patch for cmake/build_wrapper.sh; upstream rewrote it to use
nm -f posix, which works with llvm-nm
- Patch cmake/build_wrapper.cmake so the new libcsyms.sh step uses
/lib/libc.so.7 instead of libc.so.6 (otherwise the libc symbol list
is silently empty) and runs the script with sh instead of bash
- aarch64: add elf_aux_info()-based SVE/SVE2 detection for the fat
runtime dispatcher; upstream only implements it on Linux, so SVE
variants were never selected on FreeBSD. Also require SVE2 bitperm
for the sve2 variant, since it is built with +sve2-bitperm.
Verified to fall back to NEON on a non-SVE CPU (eMAG); not yet
tested on SVE hardware.
