1. 02 Sep, 2018 1 commit
    • Ben Hutchings's avatar
      Import Debian version 0.132 · 21af342d
      Ben Hutchings authored
      initramfs-tools (0.132) unstable; urgency=medium
      
        * [4e1f80e] mkinitramfs: Don't try to remove nonexistent /var/cache/ldconfig
          (Closes: #904065)
        * [d0212d3] hook-functions: Quote exclusions in copy_modules_dir
        * [8c806b4] hook-functions: Add support for compressed module files
          (Closes: #891560)
        * [71f5d61] mkinitramfs: Create usr-merged filesystem layout
          (Closes: #874523)
        * [4c0d86c] hooks/resume: Loop over all swap devices when selecting
          automatically
        * [312393b] hooks/resume: Reject zram swap devices when selecting
          automatically (Closes: #861228)
        * [c0ae42b] update-initramfs: Always take over other initramfs images
          (Closes: #765750, #804295)
        * [f39625a] update-initramfs: Make "-k all" take over other initramfs images
        * [f273172] scripts/functions: Fix parsing of device name from ip parameter
        * [abae896] scripts/functions: Merge information from BOOTIF/DEVICE and ip
          parameters (Closes: #721088)
      
      initramfs-tools (0.131) unstable; urgency=medium
      
        [ Ben Hutchings ]
        * [34bf250] debian/control: Remove maximilian attems and Adam Conrad from
          Uploaders field
        * [5ad3976] hooks/resume: Generalise logic to replace dev name with UUID when
          RESUME=auto
        * [0170b7f] hooks/resume: Store /dev/mapper path for swap if possible when
          RESUMEauto (Closes: #883735)
        * [5bf4c29] debian/control: Point Vcs URLs to Salsa
        * [e8377b3] hooks/fsck: Fix invocation of fstype at build time.
          Thanks to Simon Tatham <anakin@pobox.com> (Closes: #865691)
        * [492763a] docs: Convert maintainer notes to Markdown format
        * [0214976] docs/maintainer-notes.md: Update references to Alioth, to Salsa
        * [b71c6c9] docs/maintainer-notes.md: Delete mention of git-core package
        * [89daebf] docs/maintainer-notes.md: Change BTS link to specify source
          package
        * [8e539dc] docs/maintainer-notes.md: Link to tracker.debian.org instead of
          PTS
        * [3f27197] docs/maintainer-notes.md: Remove stale link to Debian bug count
          graph
        * [7dece27] docs/maintainer-notes.md: Use HTTP-S for Debian wiki and popcon
          links
        * [838b6ec] docs/maintainer-notes.md: Link directly to Ubuntu bug list
        * [eaa61d4] docs/maintainer-notes.md: Add myself as co-author
        * [fa7df8f] hook-functions: Correct the comment above hidden_dep_add_modules
        * [0ffbeb9] hook-functions: Remove directory information from
          hidden_dep_add_modules
        * [78e95c1] hook-functions: Check modules.builtin in hidden_dep_add_modules
        * [976aeb9] hook-functions: Handle i8042 hidden dependency on psmouse
          (Closes: #878829)
        * [211e050] hook-functions: Handle nvme hidden dependency on vmd.
          Thanks to Mario Limonciello <Mario.Limonciello@dell.com> (Closes: #886349)
        * [28536f6] unmkinitramfs, lsinitramfs: Split multiple early initramfs
          sections (Closes: #886424)
        * [2a9d97d] debian/control: Add e2fsprogs to Depends for initramfs-tools-core
          (Closes: #887211)
        * [577c94e] initramfs.conf: Update documentation of the DEVICE variable
        * [f1cbe6f] mkinitramfs: Build reproducible images when SOURCE_DATE_EPOCH is
          set. Thanks to Chris Lamb <lamby@debian.org> (Closes: #845034)
        * [717e6bd] mkinitramfs: Exclude ldconfig auxiliary cache from initramfs
          image. Thanks to Kristian Klausen <klausenbusk@hotmail.com>
        * [8bddfe4] mkinitramfs(8): Remove description of nonexistent options
        * [ecd6ce2] mkinitramfs(8), hook-functions: Correct references to "ramdisk"
        * [63cec7b] docs: Move initramfs-tools man page to section 7
        * [562d86e] docs: Update man page header dates
        * [9bd960e] docs: Set man page header source and manual parameters correctly
        * [97cc7e7] docs: Don't package maintainer documentation in
          initramfs-tools-core
        * [4e19a6f] docs: Remove outdated "HACKING" documentation file
        * [6c02622] debian/control: Remove obsolete Breaks and Depends relations
        * [749751c] debian/control: Update Standards-Version to 4.1.5
          - Set Rules-Requires-Root: no
        * [db469f0] initramfs-tools: Override lintian warning "breaks-without-version
          upstart"
      
        [ Riku Voipio ]
        * [72b6419] auto_add_modules: add mfd for MODULES==most (Closes: #880584)
      
        [ Benjamin Drung ]
        * [7ca4d0d] Add mlx4 and mlx5 InfiniBand modules to hidden_dep_add_modules
          (Closes: #871595)
        * [df7d7e8] Add option to clear screen at the beginning (Closes: #856929)
        * [a33cb62] Fix indentation (spaces to tabs)
        * [3962c4f] Fix config directory in error message (Closes: #856589)
        * [be2b91c] Double quote to prevent globbing and word splitting
        * [ae339ff] Make size of /run configurable (Closes: #862013)
      
        [ Steve Langasek ]
        * [c198331] Skip dmsetup resume devices if they are randomly encrypted.
      
        [ Paul Menzel ]
        * [e28b865] mkinitramfs: Add LZ4 support (Closes: #893845)
        * [5091a5e] control: Remove suggestion of *lz4-tools*
      
        [ Colin Watson ]
        * [b4804dd] Only sync the filesystem containing the initramfs
          (Closes: #882380) (LP: #1667512)
      
        [ Karsten Merker ]
        * [281ad9a] Include the axp20x_usb_power module in the base modules list for
          MODULES=most (Closes: #866521)
      
        [ Ivan Kozik ]
        * [54367c0] init: export DEVICE so that when configured, it is actually set
          in configure_networking (Closes: #883664)
      21af342d
  2. 08 Apr, 2017 4 commits
  3. 06 Apr, 2017 1 commit
  4. 16 Apr, 2016 1 commit
    • Ben Hutchings's avatar
      parse_numeric: Do not attempt to canonicalise root device name · eff18063
      Ben Hutchings authored
      LILO allows specifying an md-RAID device as root, e.g. root=/dev/md0,
      but as usual translates them into device numbers, e.g. root=900.
      
      parsen_numeric translates device numbers back into e.g. /dev/block/9:0
      but then tries to canonicalise them using readlink -f.  This works OK
      for simple disk devices whose drivers will have initialised by the
      point it runs, but not for stacked disk devices which will be created
      later.
      
      We could move parse_numeric later, but I think it's preferable to
      handle this weird special case before running external scripts that
      use $ROOT.
      
      Closes: #815555
      Signed-off-by: default avatarBen Hutchings <ben@decadent.org.uk>
      eff18063
  5. 21 Jan, 2016 7 commits
  6. 18 Jan, 2016 1 commit
  7. 22 Dec, 2015 4 commits
  8. 12 Dec, 2015 1 commit
  9. 09 Dec, 2015 4 commits
  10. 08 Dec, 2015 3 commits
  11. 06 Dec, 2015 4 commits
  12. 12 Apr, 2015 3 commits
  13. 01 Mar, 2015 3 commits
    • Ben Hutchings's avatar
      local: Use max(30, rootdelay) as timeout for block device to appear · 84023324
      Ben Hutchings authored
      The rootdelay kernel parameter is specified as a fixed delay before
      attempting to mount the root filesystem, not a timeout.  In the
      initramfs it is processed by udev's init-top script.  This is being
      used to workaround bug #678696, but as most users will set the
      shortest value that seems to work, it is fragile.  Using the same
      value as a timeout makes our fix for #678696 fragile too.
      
      We shouldn't really use $ROOTDELAY here at all, but as some users may
      have a good reason for wanting a longer timeout and may depend on the
      current behaviour, make our current default timeout of 30 seconds a
      minimum.
      Signed-off-by: default avatarBen Hutchings <ben@decadent.org.uk>
      84023324
    • Ben Hutchings's avatar
      local: Invoke local-block scripts from the loop in local_device_setup · d4e56010
      Ben Hutchings authored
      Currently we invoke the local-top script once (and possibly
      local-block once for /usr) and then poll for block devices to appear
      if that wasn't sufficient.  This works for synchronously discovered
      devices, devices stacked on top of them (using lvm2/mdadm/cryptsetup)
      and asynchronously discovered devices like USB storage.  It doesn't
      work for devices stacked on top of USB unless they are set up entirely
      via udev hooks.
      
      We need to invoke local-block again before each time we poll.
      As the local-block scripts might not be that quick and we don't
      have a good way to include their run-time in the timeout check,
      wait 1 second rather than 0.1 second between polls.
      
      Closes: #678696
      Signed-off-by: default avatarBen Hutchings <ben@decadent.org.uk>
      d4e56010
    • Ben Hutchings's avatar
  14. 27 Jan, 2015 1 commit
  15. 17 Jan, 2015 2 commits
    • Ben Hutchings's avatar
      checkfs: Correct error messages and emergency shell in case of failure · e5526cda
      Ben Hutchings authored
      The current error messages and some code are taken from initscripts,
      which will fsck an already-mounted root filesystem and then must
      reboot if it is modified.  Here we have not mounted the filesystem and
      can continue if the fsck was successful, or retry after running an
      emergency shell.
      
      The emergency shell also doesn't work (sulogin crashes); we must use
      the panic function for this instead.
      
      log_warning_msg does not print to the console if the 'quiet' parameter
      was used, so is not suitable for warning messages that must appear
      unconditionally.  Pass a terse explanatory message to the panic
      function.
      Signed-off-by: default avatarBen Hutchings <ben@decadent.org.uk>
      e5526cda
    • Ben Hutchings's avatar
      checkfs: Continue unless fsck reports uncorrectable errors · 1780a669
      Ben Hutchings authored
      The current checking of the fsck exit code is extremely fragile.  We
      should never see the reboot flag (2) set as that is only used when
      checking a mounted root filesystem, so don't bother checking for it.
      So we should only need to check whether the uncorrectable error flag
      (4) is set, and should continue otherwise.
      
      This mitigates bug #766448.
      
      Closes: #765614
      Signed-off-by: default avatarBen Hutchings <ben@decadent.org.uk>
      1780a669