v0.9.3 (2025-09-23)

• Fix test_ssp on OpenBSD (@omegametabroccolo, #599)
• Fix build of hvt on aarch64 (#601, @cmainas)
• Adapt OpenBSD hvt tender with the new vmm(4) (@omegametabroccolo, @hannesm, #600)

v0.9.2 (2025-07-15)

• Send FPU info to GDB in HVT server (@greydot, #571)
• Support of OpenBSD 7.5 (@shym, @adamsteen, @omegametabroccolo, #573)

v0.9.1 (2025-03-06)

• Fetch stdalign.h from system on FreeBSD by @shym in #590
• Fix the documentation about clocks (sorry for my french) by @dinosaure in #591

v0.9.0 (2024-10-11)

• Fetch stdadomic.h from system on FreeBSD (#574, @shym, @hannesm)
• Improve documentation about NAT & deployement (#575, @panglesd, @reynir)
• Ensure reproducibility of installed objects (#576, @dinosaure)
• Allow 4GB of memory for the unikernel (#577, @mato, @reynir, @Kensan, @hannesm)
• Fix error path execution about unstable TSC from system (#578, @reynir, @hannesm)
• -no-acpi is deprecated, use -machine acpi=off (#579, @wegank)
• Fix yield on spt target (#582, @reynir, @Firobe)
• Delete redundant check (#583, @hannesm, @PizieDust)
• Multiple net devices for virtio targets (#581, @ricarkol, @palainp, @reynir, @hannesm)

v0.8.1 (2024-04-05)

• Update the documentation about how to use Solo5 (@fabbing, @dinosaure, @Kensan, #558)
• Fix few warnings about the new introduced TLS support (@greydot, #563)
• Prevent the release if we get some errors with -Werror (@hannesm, @palainp, #565)
• Fix the gdb support and the EFLAGS register (@greydot, @reynir, #567)
• Fix few warnings and errors on FreeBSD 14 (@hannesm, #564)
• Allow to pass HOSTAR and HOST_PKG_CONFIG for the NixOS support (@sternenseemann, #508)

v0.8.0 (2023-04-25)

• Be able to build spt, virtio, muen and xen targets on OpenBSD
  (@adamsteen, #544). This change does not allow us to "run" these targets on
  OpenBSD

• Fix linker scripts with TLS (Thread Local Storage) sections
  (@palainp, @hannesm, @dinosaure, #542)

• Export TLS symbols (@palainp, @hannesm, @dinosaure, #546)
  breaking change due to #542 & #546, tenders must be upgraded. Indeed,
  solo5.0.7.* tenders will not be able to load correctly unikernels compiled
  with solo5.0.8.0. The internal ABI version for solo5-hvt/solo5-spt was
  upgraded accordingly.

  This version implements Thread Local Storage. The user can initialise a TLS
  block with solo5_tls_init on a pointer to solo5_tls_size() free bytes.
  Then, the user is able to set the tp (Thread Pointer) pointer via
  solo5_set_tls_base(solo5_tls_tp_offset(tls_block)). More details are
  available into solo5.h.

  Note: this change does not allow a Solo5 unikernel to use multiple cores! It
  only provides an API required by OCaml 5 / pthread to launch, at most, one
  thread.

• Fix tests reported by NixOS (@greydot, @ehmry, #547)

• Split out the time.c implementation between Muen and HVT
  (@dinosaure, @Kensan, #552)

• User hypercall instead of TSC-based clock when the user asks for the
  wall-clock (@dinosaure, @reynir, #549, #550)

  Note: only hvt & virtio are updated to avoid a clock drift on the wall-clock.
  Indeed, when the unikernel is suspended, the wall-clock is not updated. Muen
  & Xen still use a TSC-based wall-clock. The spt target was already in sync
  with the host's wall-clock.

• Improve the muen clock (@Kensan, #553)

• Fix the .bss section according to #542 & #546. The .bss section is tagged
  with PT_LOAD. Tenders are available to load this section properly.
  (@Kensan, @dinosaure, #551, #554)

• Fix the cross-compilation of Solo5 for aarch64
  (@dinosaure, @palainp, @hannesm, #555)

• Increase the Muen ABI (2 to 3) due to TLS changes (@Kensan, #557)

• Support lifecycle management for Muen (@Kensan, #557)
  The user is able to configure automatic restarting of unikernels that invokes
  solo5_ext()

• Fix the test_fpu test & ensure the alignment of variables (@Kensan, #557)

v0.7.5 (2022-12-07)

• Since MirageOS moved from PV mode to PVH on Xen, and thus replacing Mini-OS
  with solo5, there was an issue in the solo5 code which failed to properly
  account the already written bytes on the console. This only occurs if the
  output to be performed does not fit in a single output buffer (2048 bytes on
  Xen).

  The code in question set the number of bytes written to the last written count
  written = output_some(buf), instead of increasing the written count
  written += output_some(buf).

  Thanks to Krzysztof Burghardt, Pierre Alain, Thomas Leonard & Hannes Mehnert
  for the issue, review and proposal. The fix is available here: #538.

v0.7.4 (2022-11-04)

• Mark .text execute-only, currently only on OpenBSD (@adamsteen, #450)
• Allow all log levels to be passed as command line parameter to the tender
  (added --solo5:error, --solo5:warn, --solo5:info) (@reynir, #532)
• Add tender command line argument --block-sector-size:=int. This allows
  to specify the desired block sector size. The default if not provided is 512,
  the same that was used before (@reynir, #528, addresses partially #325)
• Check that the file passed as block device is aligned to the block sector size
  (@reynir, #527)
• Use realpath to determine toolchain paths - allowing tools being symlinks
  as they are on NixOS (@greydot, #526)
• Allow slack in sleep in test_time (@greydot, #525, #535)
• Fix build when using git worktree (.git being a file) (@reynir, #531)
• Fix tests on OpenBSD 7.2 (@dinosaure, #535)
• Add x-ci-failures on our OPAM files about CentOS 7 (@dinosaure, #535)

v0.7.3 (2022-07-20)

• Fix broken API doc link in the README.md (@YuseiIto, #521)
• Be more informative when tap_attach fails (@reynir, #522)
• Ignore up variable to avoid compiler warning on FreeBSD (@hannesm, #520)

• Suppress gcc array bounds warning in test_zeropage (@felixmoebius, #515)
• Xen: retrieve mem_size uniformly via XENMEM_memory_map hypercall.
  Previously, the memory map was extracted from the HVM start info (if
  available and booting directly via PVH), or multiboot info (if booting via
  multiboot). The fallback for direct PVH booting was the XENMEM_memory_map
  hypercall (which retrieves an E820 memory map). This lead to three distinct
  paths, with no fallback for the memory map not being present in the multiboot
  info. With QubesOS 4.1 (Xen 4.14), this didn't work anymore (it worked with
  QubesOS 4.0 (Xen 4.8)).
  Now, there is a single path of the code, which uses the hypercall. Since this
  is only executed once at startup, the overhead is negligible (@hannesm, #516,
  review and discussions with @marmarek @xaki23 @palainp)
• Xen: do not skip first token of command line when booted via multiboot.
  This code originated from the virtio binding, but when booting on xen via
  multiboot there is no additional token.
  (@hannesm, #517, review and testing with @palainp @xaki23)