Log the board and version when entering the recovery shell. Extract
the firmware version logic from init.
Currently this is the only way to get the debug log. If we add a way
from the GUI, we may want to log the board and version somewhere else
too.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
LOG() is added to log to the log only (not kmsg, more verbose than
TRACE).
DO_WITH_DEBUG only captures stdout/stderr to the log with LOG().
kexec-boot silences stderr from kexec, we don't want it on the console.
No need to repeat the kexec command when asking in debug to continue
boot, it's no longer hidden behind verbose output from kexec.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
`eval "$kexeccmd"` should become `DO_WITH_DEBUG eval "$kexeccmd"` when
adding DO_WITH_DEBUG, command invocation is still the same, still needs
eval.
Restore DO_WITH_DEBUG in front of kexec-parse-boot that had been
removed.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
"$1 err:" looked like an error, but often there's output on stderr
that's diagnostic (like kexec -d). "$1 stderr:" is clearer.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
DO_WITH_DEBUG traces command exit status (if failed), stdout/stderr (if
not empty), and PATH (if command was not found). The caller still
observes the exit status, and stdout/stderr still go to the caller as
well.
This way, DO_WITH_DEBUG can be inserted anywhere with minimal spam in
the logs and without affecting the script.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
Otherwise we get ehci-pci and xhci_hcd kernel messages in dmesg debug AFTER "Verifying presence of GPG card" which explains why dongle might not be found in time and fails in oem-factory-reset
Fixes https://github.com/Nitrokey/heads/issues/48
Signed-off-by: Thierry Laurion <insurgo@riseup.net>
TODO:
- $(pcrs) call sometimes fail in DEBUG call, outputting too many chars to be inserted in kmesg. Call removed here since redundant (PCR6 already extended with LUKS header)
- Notes added for TPM2 simplification over TPM1 in code as TODO
Signed-off-by: Thierry Laurion <insurgo@riseup.net>
Fixes linuxboot#1092.
Supersedes linuxboot#1093
- Cherry-picks ed1c23a (credit to @hardened-vault) thank you!)
- Addresses and correct self-review under linuxboot#1093 (@hardened-vault: you don't answer often here!)
- kexec-unseal-key: Warn a user who attempts to default boot while his Disk Unlock Key passphrase fails to unseal because LUKS headers changed.
(linuxboot#1093 (comment))
- kexec-seal-key: Identical as in ed1c23a
- kexec-add-key: Tell the user that the Headers did not change when changing TPM released Disk Unlock Key
(Through changing default boot at Options->Boot Options -> Show OS boot options: select a new boot option
and set a Disk Unlock Key in TPM, accept to modify disk and sign /boot options)
- Here, we cancel the diff output shown on screen linuxboot#1093 (comment)
- And we change the warning given to the user to past tense "Headers of LUKS containers to be unlocked via TPM Disk Unlock Key passphrase did not change."
Signed-off-by: Thierry Laurion <insurgo@riseup.net>
All boards with CONFIG_LINUX_USB=y ship ehci-* and xhci-*, they are
not controlled by CONFIG_LINUX_USB_COMPANION_CONTROLLER. Always
insert them when initializing USB. Fixes commit 35de2348
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
Also remove output of attempted module loading since DEBUG will show if needed
Remove timeout after 30 seconds to unify UX and block
Change UX wording
Should address all PR review comments
Signed-off-by: Thierry Laurion <insurgo@riseup.net>
- Add additional requirements to linux config
- Add additional CONFIG_MOBILE_TETHERING=y to all maximized board configs
- Fix issue under network-recovery-init to NTP sync against NTP server pool
- Extend network-recovery-init to first try NTP sync against DNS server returned by DHCP answer
- Remove network-recovery-init earlytty and tty0 redirection (console should be setuped properly by init in all cases)
- If CONFIG_MOBILE_TETHERING=y added to board config and network-recovery-init called, wait to user input on instructions and warning 30 secs before proceeding (non-blocking)
- Machines having STATIC_IP under board config won't benefit of autoatic NTP sync
Since network-recovery-init can only be called from recovery shell now, and recovery shell can be guarded by GPG auth, this is PoC code to be used to complement TOTP being out of sync
TODO(Future PR):
- Refactor into functions and reuse into TOTP/HOTP being out of sync automatically.
Signed-off-by: Thierry Laurion <insurgo@riseup.net>
When testing a possible boot device, detect its partition type and
skip grub, LUKS, and LVM partitions. These aren't mountable as /boot,
this silences spurious exFAT errors.
In detect_boot_device, skip testing CONFIG_BOOT_DEV a second time if it
is found as a block device. This avoids doubling any errors shown from
checking this device, no sense trying it twice.
Refactor some logic to avoid duplication - extract
device_has_partitions and use it in detect_boot_device, extract
mount_possible_boot_device and use it instead of duplicating the logic.
Move find_lvm_vg_name() to /etc/functions.
Avoid mixing up similarly-named devices like 'nvme0n1'/'nvme0n10' or
'sda'/'sdaa' - it's probably unlikely that many devices will appear,
but looking for partitions in '/sys/class/block/<device>/' instead of
'/dev/' would avoid any collisions.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
Since exFAT support was enabled, mounting an iso9660 filesystem prints
spurious exFAT errors to the console. That is because busybox mount
tries all filesystems in the order listed, and exfat precedes iso9660
(those are the last two in our config). Most filesystems are silent
when used on the wrong type of filesystem, but exFAT logs errors, which
appear on the console.
Move exFAT after iso9660, so iso9660 filesystems won't show these
errors. The errors will still appear if the filesystem is actually
exFAT but cannot be mounted.
There's no significant risk of misdetecting a remnant iso9660
superblock here either. Although an iso9660 superblock could fall in
the unused space between the exFAT boot region and the FAT itself,
mkfs.exfat does zero this space so it is unlikely such a remnant
superblock would exist.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
Exception: scripts sourcing/calls within etc/ash_functions continues to use old TRACE functions until we switch to bash completely getting rid of ash.
This would mean getting rid of legacy boards (flash + legacy boards which do not have enough space for bash in flash boards) once and for all.
Signed-off-by: Thierry Laurion <insurgo@riseup.net>
Reword the notice shown when enabling USB keyboards based on feedback.
Remove the notice when disabling USB keyboard support, show it only
when enabling.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
Next steps on this is introspection and PCRs reconstruction helpers, which will output in DEBUG and be usable from recovery shell.
We have to keep in mind that providing those tools is useful in DEBUG mode and for users having access to Recovery Shell.
But currently, having access to cbmem -L output and final PCRs content is making it too easy for Evil Maid to know what needs to be hardcoded to pass measured boot.
Signed-off-by: Thierry Laurion <insurgo@riseup.net>
When playing with long fbwhiptail/whiptail messages, this commit played around the long string using fold.
'''
echo -e "This will replace the encrypted container content and its LUKS Disk Recovery Key.\n\nThe passphrase associated with this key will be asked from the user under the following conditions:\n 1-Every boot if no Disk Unlock Key was added to the TPM\n 2-If the TPM fails (hardware failure)\n 3-If the firmware has been tampered with/modified by the user\n\nThis process requires you to type the current LUKS Disk Recovery Key passphrase and will delete the LUKS TPM Disk Unlock Key slot, if set up, by setting a default boot LUKS key slot (1) if present.\n\nAt the next prompt, you may be asked to select which file corresponds to the LUKS device container.\n\nHit Enter to continue." | fold -w 70 -s
'''
Which gave the exact output of what will be inside of the fbwhiptail prompt, fixed to 70 chars width:
'''
This will replace the encrypted container content and its LUKS Disk
Recovery Key.
The passphrase associated with this key will be asked from the user
under the following conditions:
1-Every boot if no Disk Unlock Key was added to the TPM
2-If the TPM fails (hardware failure)
3-If the firmware has been tampered with/modified by the user
This process requires you to type the current LUKS Disk Recovery Key
passphrase and will delete the LUKS TPM Disk Unlock Key slot, if set
up, by setting a default boot LUKS key slot (1) if present.
At the next prompt, you may be asked to select which file corresponds
to the LUKS device container.
Hit Enter to continue.
'''
Therefore, for long prompts in the future, one can just deal with "\n 1-" alignments to be respected in prompts and have fold deal with cutting the length of strings properly.
Signed-off-by: Thierry Laurion <insurgo@riseup.net>
Don't spew the root hashes to the console when creating the hash file.
This speeds up hash creation significantly. A basic Qubes install on a
cheap (slow) SATA SSD reduced from about 1.5 minutes to just under 1
minute, and a PureOS install on a fast NVMe disk reduced from 2.5
minutes to 1 minute.
Support opening LVM volume groups to find the root disk. If an LVM PV
is found, its group is opened and the 'root' volume is used. There is
no way to set the volume name in this iteration; this is the default
name used by Qubes and probably common to many LVM OS installations.
LUKS and LVM can be mixed. Tested LUKS (PureOS) and LUKS+LVM (Qubes).
Always cd to "$ROOT_MOUNT" in a subshell, improves robustness of
scripts (previously some functions only worked if they were called
after another function had cd'd to "$ROOT_MOUNT").
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
Say the action to take in the menu (enable or disable) instead of just
"Check root hashes at boot".
Clean up some use of load_config_value, set_config, combine_configs.
Get config values from the environment directly. set_user_config does
set_config and combine_configs.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
Add TRACE_FUNC to trace the file, line, and name of the calling
function. File and function names don't have to be duplicated in a
TRACE statement with this (they tend to become inaccurate as functions
are renamed and the TRACE statement is forgotten).
Add DEBUG_STACK to dump the bash stack to debug output.
Configure bash with --enable-debugger. Bash doesn't actually include
the entire debugger, this is just some supporting variables for it.
Evidently, BASH_SOURCE[n] is only set within a function if this is
enabled. I couldn't find this indicated in any documentation, but it
happened in practice.
Compressed initrd size only increased by 2560 bytes for librem_mini_v2,
I think that is fine. This also gives us BASH_ARGC/BASH_ARGV which
might be useful for diagnostics.
Signed-off-by: Jonathon Hall <jonathon.hall@puri.sm>
