heads/initrd/bin/oem-factory-reset

1354 lines
53 KiB
Plaintext
Raw Normal View History

#!/bin/bash
# Automated setup of TPM, GPG keys, and disk
set -o pipefail
## External files sourced
. /etc/functions
. /etc/gui_functions
. /etc/luks-functions
. /tmp/config
TRACE_FUNC
# use TERM to exit on error
trap "exit 1" TERM
export TOP_PID=$$
## Static local variables
CLEAR="--clear"
CONTINUE="--yes-button Continue"
CANCEL="--no-button Cancel"
HEIGHT="0"
WIDTH="80"
USER_PIN_DEF=123456
ADMIN_PIN_DEF=12345678
TPM_PASS_DEF=12345678
USER_PIN=""
ADMIN_PIN=""
TPM_PASS=""
GPG_GEN_KEY_IN_MEMORY="n"
GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD="n"
#Circumvent Librem Key/Nitrokey HOTP firmware bug https://github.com/osresearch/heads/issues/1167
MAX_HOTP_GPG_PIN_LENGTH=25
# What are the Security components affected by custom passwords
CUSTOM_PASS_AFFECTED_COMPONENTS=""
# Default GPG Algorithm is RSA
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
# p256 also supported (TODO: nk3 supports RSA 4096 in secure element in firmare v1.7.1. Switch!?
GPG_ALGO="RSA"
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
# Default RSA key length is 3072 bits for OEM key gen. 4096 are way longer to generate in smartcard
RSA_KEY_LENGTH=3072
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
#Override RSA_KEY_LENGTH to 2048 bits for Canokey under qemu testing boards until canokey fixes
if [[ "$CONFIG_BOARD_NAME" == qemu-* ]]; then
DEBUG "Overriding RSA_KEY_LENGTH to 2048 bits for Canokey under qemu testing boards"
RSA_KEY_LENGTH=2048
fi
GPG_USER_NAME="OEM Key"
GPG_KEY_NAME=$(date +%Y%m%d%H%M%S)
GPG_USER_MAIL="oem-${GPG_KEY_NAME}@example.com"
GPG_USER_COMMENT="OEM-generated key"
SKIP_BOOT="n"
2020-07-07 08:32:22 +00:00
## functions
die() {
local msg=$1
if [ -n "$msg" ]; then
echo -e "\n$msg"
fi
kill -s TERM $TOP_PID
exit 1
}
local_whiptail_error() {
local msg=$1
if [ "$msg" = "" ]; then
die "whiptail error: An error msg is required"
fi
whiptail_error --msgbox "${msg}\n\n" $HEIGHT $WIDTH --title "Error"
}
whiptail_error_die() {
local_whiptail_error "$@"
die
}
mount_boot() {
TRACE_FUNC
# Mount local disk if it is not already mounted.
# Added so that 'o' can be typed early at boot to enter directly into OEM Factory Reset
if ! grep -q /boot /proc/mounts; then
# try to mount if CONFIG_BOOT_DEV exists
if [ -e "$CONFIG_BOOT_DEV" ]; then
mount -o ro $CONFIG_BOOT_DEV /boot || die "Failed to mount $CONFIG_BOOT_DEV. Please change boot device under Configuration > Boot Device"
fi
fi
}
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
#Generate a gpg master key: no expiration date, ${RSA_KEY_LENGTH} bits
#This key will be used to sign 3 subkeys: encryption, authentication and signing
#The master key and subkeys will be copied to backup, and the subkeys moved from memory keyring to the smartcard
generate_inmemory_RSA_master_and_subkeys() {
TRACE_FUNC
echo "Generating GPG RSA ${RSA_KEY_LENGTH} bits master key..."
# Generate GPG master key
{
echo "Key-Type: RSA" # RSA key
echo "Key-Length: ${RSA_KEY_LENGTH}" # RSA key length
echo "Key-Usage: sign" # RSA key usage
echo "Name-Real: ${GPG_USER_NAME}" # User name
echo "Name-Comment: ${GPG_USER_COMMENT}" # User comment
echo "Name-Email: ${GPG_USER_MAIL}" # User email
echo "Expire-Date: 0" # No expiration date
echo "Passphrase: ${ADMIN_PIN}" # Admin PIN
echo "%commit" # Commit changes
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --expert --batch --command-fd=0 --status-fd=1 --pinentry-mode=loopback --generate-key >/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "GPG Key generation failed!\n\n$ERROR"
fi
echo "Generating GPG RSA ${RSA_KEY_LENGTH} bits signing subkey..."
# Add signing subkey
{
echo addkey # add key in --edit-key mode
echo 4 # RSA (sign only)
echo ${RSA_KEY_LENGTH} # Signing key size set to RSA_KEY_LENGTH
echo 0 # No expiration date
echo ${ADMIN_PIN} # Local keyring admin pin
echo y # confirm
echo save # save changes and commit to keyring
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --command-fd=0 --status-fd=1 --pinentry-mode=loopback --edit-key "${GPG_USER_MAIL}" \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "GPG Key signing subkey generation failed!\n\n$ERROR"
fi
echo "Generating GPG RSA ${RSA_KEY_LENGTH} bits encryption subkey..."
#Add encryption subkey
{
echo addkey # add key in --edit-key mode
echo 6 # RSA (encrypt only)
echo ${RSA_KEY_LENGTH} # Encryption key size set to RSA_KEY_LENGTH
echo 0 # No expiration date
echo ${ADMIN_PIN} # Local keyring admin pin
echo y # confirm
echo save # save changes and commit to keyring
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --command-fd=0 --status-fd=1 --pinentry-mode=loopback --edit-key "${GPG_USER_MAIL}" \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "GPG Key encryption subkey generation failed!\n\n$ERROR"
fi
echo "Generating GPG RSA ${RSA_KEY_LENGTH} bits authentication subkey..."
#Add authentication subkey
{
#Authentication subkey needs gpg in expert mode to select RSA custom mode (8)
# in order to disable encryption and signing capabilities of subkey
# and then enable authentication capability
echo addkey # add key in --edit-key mode
echo 8 # RSA (set your own capabilities)
echo S # disable sign capability
echo E # disable encryption capability
echo A # enable authentication capability
echo Q # Quit
echo ${RSA_KEY_LENGTH} # Authentication key size set to RSA_KEY_LENGTH
echo 0 # No expiration date
echo ${ADMIN_PIN} # Local keyring admin pin
echo y # confirm
echo save # save changes and commit to keyring
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --command-fd=0 --status-fd=1 --pinentry-mode=loopback --expert --edit-key "${GPG_USER_MAIL}" \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "GPG Key authentication subkey generation failed!\n\n$ERROR"
fi
}
#Generate a gpg master key: no expiration date, p256 key (ECC)
#This key will be used to sign 3 subkeys: encryption, authentication and signing
#The master key and subkeys will be copied to backup, and the subkeys moved from memory keyring to the smartcard
generate_inmemory_p256_master_and_subkeys() {
TRACE_FUNC
echo "Generating GPG p256 bits master key..."
{
echo "Key-Type: ECDSA" # ECDSA key
echo "Key-Curve: nistp256" # ECDSA key curve
echo "Key-Usage: cert" # ECDSA key usage
echo "Name-Real: ${GPG_USER_NAME}" # User name
echo "Name-Comment: ${GPG_USER_COMMENT}" # User comment
echo "Name-Email: ${GPG_USER_MAIL}" # User email
echo "Passphrase: ${ADMIN_PIN}" # Local keyring admin pin
echo "Expire-Date: 0" # No expiration date
echo "%commit" # Commit changes
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --expert --batch --command-fd=0 --status-fd=1 --pinentry-mode=loopback --generate-key \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "GPG p256 Key generation failed!\n\n$ERROR"
fi
#Keep Master key fingerprint for add key calls
MASTER_KEY_FP=$(gpg --list-secret-keys --with-colons | grep fpr | cut -d: -f10)
echo "Generating GPG nistp256 signing subkey..."
{
echo addkey # add key in --edit-key mode
echo 11 # ECC own set capability
echo Q # sign already present, do not modify
echo 3 # P-256
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
echo 0 # No validity/expiration date
echo ${ADMIN_PIN} # Local keyring admin pin
echo save # save changes and commit to keyring
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --expert --command-fd=0 --status-fd=1 --pinentry-mode=loopback --edit-key ${MASTER_KEY_FP} >/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR_MSG=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "Failed to add ECC nistp256 signing key to master key\n\n${ERROR_MSG}"
fi
echo "Generating GPG nistp256 encryption subkey..."
{
echo addkey
echo 12 # ECC own set capability
echo Q # Quit
echo 3 # P-256
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
echo 0 # No validity/expiration date
echo ${ADMIN_PIN} # Local keyring admin pin
echo save # save changes and commit to keyring
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --expert --command-fd=0 --status-fd=1 --pinentry-mode=loopback --edit-key ${MASTER_KEY_FP} >/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR_MSG=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "Failed to add ECC nistp256 encryption key to master key\n\n${ERROR_MSG}"
fi
echo "Generating GPG nistp256 authentication subkey..."
{
echo addkey # add key in --edit-key mode
echo 11 # ECC own set capability
echo S # deactivate sign
echo A # activate auth
echo Q # Quit
echo 3 # P-256
echo 0 # no expiration
echo ${ADMIN_PIN} # Local keyring admin pin
echo save # save changes and commit to keyring
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --expert --command-fd=0 --status-fd=1 --pinentry-mode=loopback --edit-key ${MASTER_KEY_FP} >/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR_MSG=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "Failed to add ECC nistp256 authentication key to master key\n\n${ERROR_MSG}"
fi
}
#Function to move current gpg keyring subkeys to card (keytocard)
# This is aimed to be used after having generated master key and subkeys in memory and having backed up them to a LUKS container
# This function will keytocard the subkeys from the master key in the keyring
# The master key will be kept in the keyring
# The master key was already used to sign the subkeys, so it is not needed anymore
# Delete the master key from the keyring once key to card is done (already backed up on LUKS private partition)
keytocard_subkeys_to_smartcard() {
TRACE_FUNC
#make sure usb ready and USB Security Dongle ready to communicate with
enable_usb
enable_usb_storage
gpg --card-status >/dev/null 2>&1 || die "Error getting GPG card status"
gpg_key_factory_reset
echo "Moving subkeys to smartcard..."
{
echo "key 1" #Toggle on Signature key in --edit-key mode on local keyring
echo "keytocard" #Move Signature key to smartcard
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
echo "1" #Select Signature key key slot on smartcard
echo "${ADMIN_PIN}" #Local keyring Subkey PIN
echo "${ADMIN_PIN_DEF}" #Smartcard Admin PIN
echo "0" #No expiration date
echo "key 1" #Toggle off Signature key
echo "key 2" #Toggle on Encryption key
echo "keytocard" #Move Encryption key to smartcard
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
echo "2" #Select Encryption key key slot on smartcard
echo "${ADMIN_PIN}" #Local keyring Subkey PIN
echo "${ADMIN_PIN_DEF}" #Smartcard Admin PIN
echo "key 2" #Toggle off Encryption key
echo "key 3" #Toggle on Authentication key
echo "keytocard" #Move Authentication key to smartcard
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
echo "3" #Select Authentication key slot on smartcard
echo "${ADMIN_PIN}" #Local keyring Subkey PIN
echo "${ADMIN_PIN_DEF}" #Smartcard Admin PIN
echo "key 3" #Toggle off Authentication key
echo "save" #Save changes and commit to keyring
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --expert --command-fd=0 --status-fd=1 --pinentry-mode=loopback --edit-key "${GPG_USER_MAIL}" \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "GPG Key moving subkeys to smartcard failed!\n\n$ERROR"
fi
TRACE_FUNC
}
#Whiptail prompt to insert to be wiped thumb drive
prompt_insert_to_be_wiped_thumb_drive() {
TRACE_FUNC
#Whiptail warning about having only desired to be wiped thumb drive inserted
whiptail_warning --title 'WARNING: Please insert the thumb drive to be wiped' \
--msgbox "The thumb drive will be WIPED next.\n\nPlease connect only the thumb drive to be wiped and disconnect others." 0 80 ||
die "Error displaying warning about having only desired to be wiped thumb drive inserted"
}
#export master key and subkeys to thumbdrive's private LUKS contained partition
export_master_key_subkeys_and_revocation_key_to_private_LUKS_container() {
TRACE_FUNC
#Sanity check on passed arguments
while [ $# -gt 0 ]; do
case "$1" in
--mode)
mode="$2"
shift
shift
;;
--device)
device="$2"
shift
shift
;;
--mountpoint)
mountpoint="$2"
shift
shift
;;
--pass)
pass="${2}"
shift
shift
;;
*)
die "Error: unknown argument: $1"
;;
esac
done
mount-usb --mode "$mode" --device "$device" --mountpoint "$mountpoint" --pass "$pass" || die "Error mounting thumb drive's private partition"
#Export master key and subkeys to thumb drive
DEBUG "Exporting master key and subkeys to private LUKS container's partition..."
gpg --export-secret-key --armor --pinentry-mode loopback --passphrase="${pass}" "${GPG_USER_MAIL}" >"$mountpoint"/privkey.sec ||
die "Error exporting master key to private LUKS container's partition"
gpg --export-secret-subkeys --armor --pinentry-mode loopback --passphrase="${pass}" "${GPG_USER_MAIL}" >"$mountpoint"/subkeys.sec ||
die "Error exporting subkeys to private LUKS container's partition"
#copy whole keyring to thumb drive, including revocation key and trust database
cp -af ~/.gnupg "$mountpoint"/.gnupg || die "Error copying whole keyring to private LUKS container's partition"
#Unmount private LUKS container's mount point
umount "$mountpoint" || die "Error unmounting private LUKS container's mount point"
TRACE_FUNC
}
#Export public key to thumb drive's public partition
export_public_key_to_thumbdrive_public_partition() {
TRACE_FUNC
#Sanity check on passed arguments
while [ $# -gt 0 ]; do
case "$1" in
--mode)
mode="$2"
shift
shift
;;
--device)
device="$2"
shift
shift
;;
--mountpoint)
mountpoint="$2"
shift
shift
;;
*)
die "Error: unknown argument: $1"
;;
esac
done
#pass non-empty arguments to --pass, --mountpoint, --device, --mode
mount-usb --device "$device" --mode "$mode" --mountpoint "$mountpoint" || die "Error mounting thumb drive's public partition"
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
#TODO: reuse "Obtain GPG key ID" so that pubkey on public thumb drive partition is named after key ID
gpg --export --armor "${GPG_USER_MAIL}" >"$mountpoint"/pubkey.asc || die "Error exporting public key to thumb drive's public partition"
umount "$mountpoint" || die "Error unmounting thumb drive's public partition"
TRACE_FUNC
}
# Select thumb drive and LUKS container size for GPG key export
# Sets variables containing selections:
# - thumb_drive
# - thumb_drive_luks_percent
select_thumb_drive_for_key_material() {
TRACE_FUNC
#enable usb storage
enable_usb
enable_usb_storage
prompt_insert_to_be_wiped_thumb_drive
#loop until user chooses a disk
thumb_drive=""
while [ -z "$thumb_drive" ]; do
#list usb storage devices
list_usb_storage disks >/tmp/usb_disk_list
# Abort if:
# - no disks found (prevent file_selector's nonsense prompt)
# - file_selector fails for any reason
# - user aborts (file_selector succeeds but FILE is empty)
if [ $(cat /tmp/usb_disk_list | wc -l) -gt 0 ] &&
file_selector --show-size "/tmp/usb_disk_list" "Select USB device to partition" &&
[ -n "$FILE" ]; then
# Obtain size of thumb drive to be wiped with fdisk
disk_size_bytes="$(blockdev --getsize64 "$FILE")"
if [ "$disk_size_bytes" -lt "$((128*1024*1024))" ]; then
warn "Thumb drive size is less than 128MB!"
warn "LUKS container needs to be at least 8MB!"
warn "If the next operation fails, try with a bigger thumb drive"
fi
select_luks_container_size_percent
thumb_drive_luks_percent="$(cat /tmp/luks_container_size_percent)"
if ! confirm_thumb_drive_format "$FILE" "$thumb_drive_luks_percent"; then
warn "Thumb drive wipe aborted by user!"
continue
fi
#User chose and confirmed a thumb drive and its size to be wiped
thumb_drive=$FILE
else
#No USB storage device detected
warn "No USB storage device detected! Aborting OEM Factory Reset / Re-Ownership"
sleep 3
die "No USB storage device detected! User decided to not wipe any thumb drive"
fi
done
thumb_drive_luks_percent="$(cat /tmp/luks_container_size_percent)"
}
#Wipe a thumb drive and export master key and subkeys to it
# $1 - thumb drive block device
# $2 - LUKS container percentage [1-99]
wipe_thumb_drive_and_copy_gpg_key_material() {
TRACE_FUNC
local thumb_drive thumb_drive_luks_percent
thumb_drive="$1"
thumb_drive_luks_percent="$2"
#Wipe thumb drive with a LUKS container of size $(cat /tmp/luks_container_size_percent)
prepare_thumb_drive "$thumb_drive" "$thumb_drive_luks_percent" "${ADMIN_PIN}"
#Export master key and subkeys to thumb drive first partition
export_master_key_subkeys_and_revocation_key_to_private_LUKS_container --mode rw --device "$thumb_drive"1 --mountpoint /media --pass "${ADMIN_PIN}"
#Export public key to thumb drive's public partition
export_public_key_to_thumbdrive_public_partition --mode rw --device "$thumb_drive"2 --mountpoint /media
TRACE_FUNC
}
gpg_key_factory_reset() {
TRACE_FUNC
#enable usb storage
enable_usb
# Factory reset GPG card
echo "GPG factory reset of USB Security Dongle's smartcard..."
{
echo admin # admin menu
echo factory-reset # factory reset smartcard
echo y # confirm
echo yes # confirm
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --command-fd=0 --status-fd=1 --pinentry-mode=loopback --card-edit \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "GPG Key factory reset failed!\n\n$ERROR"
fi
2020-08-05 09:49:06 +00:00
# If Nitrokey Storage is inserted, reset AES keys as well
if lsusb | grep -q "20a0:4109" && [ -x /bin/hotp_verification ]; then
DEBUG "Nitrokey Storage detected, resetting AES keys..."
2020-08-05 09:49:06 +00:00
/bin/hotp_verification regenerate ${ADMIN_PIN_DEF}
DEBUG "Restarting scdaemon to remove possible exclusive lock of dongle"
killall -9 scdaemon
fi
# Toggle forced sig (good security practice, forcing PIN request for each signature request)
if gpg --card-status | grep "Signature PIN" | grep -q "not forced"; then
DEBUG "GPG toggling forcesig on since off..."
{
echo admin # admin menu
echo forcesig # toggle forcesig
echo ${ADMIN_PIN_DEF} # local keyring PIN
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --command-fd=0 --status-fd=1 --pinentry-mode=loopback --card-edit \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "GPG Key forcesig toggle on failed!\n\n$ERROR"
fi
fi
# use p256 for key generation if requested
if [ "$GPG_ALGO" = "p256" ]; then
{
echo admin # admin menu
echo key-attr # key attributes
echo 2 # ECC
echo 3 # P-256
echo ${ADMIN_PIN_DEF} # local keyring PIN
echo 2 # ECC
echo 3 # P-256
echo ${ADMIN_PIN_DEF} # local keyring PIN
echo 2 # ECC
echo 3 # P-256
echo ${ADMIN_PIN_DEF} # local keyring PIN
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --expert --command-fd=0 --status-fd=1 --pinentry-mode=loopback --card-edit \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "Setting key to NIST-P256 in USB Security Dongle failed."
fi
# fallback to RSA key generation by default
elif [ "$GPG_ALGO" = "RSA" ]; then
DEBUG "GPG setting RSA key length to ${RSA_KEY_LENGTH} bits..."
# Set RSA key length
{
echo admin
echo key-attr
echo 1 # RSA
echo ${RSA_KEY_LENGTH} #Signing key size set to RSA_KEY_LENGTH
echo ${ADMIN_PIN_DEF} #Local keyring PIN
echo 1 # RSA
echo ${RSA_KEY_LENGTH} #Encryption key size set to RSA_KEY_LENGTH
echo ${ADMIN_PIN_DEF} #Local keyring PIN
echo 1 # RSA
echo ${RSA_KEY_LENGTH} #Authentication key size set to RSA_KEY_LENGTH
echo ${ADMIN_PIN_DEF} #Local keyring PIN
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --command-fd=0 --status-fd=1 --pinentry-mode=loopback --card-edit \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "Setting key attributed to RSA ${RSA_KEY_LENGTH} bits in USB Security Dongle failed."
fi
else
#Unknown GPG_ALGO
whiptail_error_die "Unknown GPG_ALGO: $GPG_ALGO"
fi
TRACE_FUNC
}
generate_OEM_gpg_keys() {
TRACE_FUNC
#This function simply generates subkeys in smartcard following smarcard config from gpg_key_factory_reset
echo "Generating GPG keys in USB Security Dongle's smartcard..."
{
echo admin # admin menu
echo generate # generate keys
echo n # Do not export keys
echo ${ADMIN_PIN_DEF} # Default admin PIN since we just factory reset
echo ${USER_PIN_DEF} # Default user PIN since we just factory reset
echo 0 # No key expiration
echo ${GPG_USER_NAME} # User name
echo ${GPG_USER_MAIL} # User email
echo ${GPG_USER_COMMENT} # User comment
echo ${USER_PIN_DEF} # Default user PIN since we just factory reset
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --command-fd=0 --status-fd=2 --pinentry-mode=loopback --card-edit \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output)
whiptail_error_die "GPG Key automatic keygen failed!\n\n$ERROR"
fi
TRACE_FUNC
}
gpg_key_change_pin() {
TRACE_FUNC
DEBUG "Changing GPG key PIN"
# 1 = user PIN, 3 = admin PIN
PIN_TYPE=$1
PIN_ORIG=${2}
PIN_NEW=${3}
# Change PIN
{
echo admin # admin menu
echo passwd # change PIN
echo ${PIN_TYPE} # 1 = user PIN, 3 = admin PIN
echo ${PIN_ORIG} # old PIN
echo ${PIN_NEW} # new PIN
echo ${PIN_NEW} # confirm new PIN
echo q # quit
echo q
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
} | DO_WITH_DEBUG gpg --command-fd=0 --status-fd=2 --pinentry-mode=loopback --card-edit \
>/tmp/gpg_card_edit_output 2>&1
if [ $? -ne 0 ]; then
ERROR=$(cat /tmp/gpg_card_edit_output | fold -s)
whiptail_error_die "GPG Key PIN change failed!\n\n$ERROR"
fi
TRACE_FUNC
}
generate_checksums() {
TRACE_FUNC
# ensure /boot mounted
if ! grep -q /boot /proc/mounts; then
mount -o rw /boot || whiptail_error_die "Unable to mount /boot"
else
mount -o remount,rw /boot || whiptail_error_die "Unable to mount /boot"
fi
Uniformize vocabulary: LUKS TPM Disk Unlock Key & LUKS Disk Recovery Key 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>
2024-01-19 17:32:04 +00:00
#Check if previous LUKS TPM Disk Unlock Key was set
if [ -e /boot/kexec_key_devices.txt ]; then
TPM_DISK_ENCRYPTION_KEY_SET=1
fi
# clear any existing checksums/signatures
rm /boot/kexec* 2>/dev/null
# create Heads TPM counter
if [ "$CONFIG_TPM" = "y" ]; then
if [ "$CONFIG_IGNORE_ROLLBACK" != "y" ]; then
tpmr counter_create \
-pwdc '' \
-la -3135106223 |
tee /tmp/counter ||
whiptail_error_die "Unable to create TPM counter"
TPM_COUNTER=$(cut -d: -f1 </tmp/counter)
# increment TPM counter
increment_tpm_counter $TPM_COUNTER >/dev/null 2>&1 ||
whiptail_error_die "Unable to increment tpm counter"
# create rollback file
sha256sum /tmp/counter-$TPM_COUNTER >/boot/kexec_rollback.txt 2>/dev/null ||
whiptail_error_die "Unable to create rollback file"
else
## needs to exist for initial call to unseal-hotp
echo "0" >/boot/kexec_hotp_counter
fi
Addition of qemu-(fb)whiptail-tpm2(-hotp) boards -coreboot support of TPM v2.0 (shared config for TPM2 support across all 4 previous variations) -swtpm set to be launched under TPM v2.0 mode under board config -Documentation file under each board.md softlinks to qemu-coreboot-fbwhiptail-tpm1.md (which has been generalized) This is skeleton for TPM v2 integration under Heads ------------- WiP TODO: - libcurl cannot be built as a tpm2-tools dependency as of now not sure why. curl currently needs to be added in board config to be built - Note: tpm-reset (master and here) needs some review, no handle of no tpm use case. Caller is responsible to not call it otherwise does nothing - init tries to bind fd and fails currently - Note: Check if whiptail is different of fbwhiptail in clearing screen. As of now every clear seems to be removed, still whiptail clears previous console output - When no OS' /boot can be mounted, do not try to TPM reset (will fail) - seal-hotpkey is not working properly - setting disk unlock key asks for TPM ownership passphrase (sealing in NV requires ownership, but text is misleading user as if reowning TPM) - We should cache input, feed tpm behind the scene and wipe passphrase and state clearly that this is TPM disk unlock kye passphrase. - primary key from TPM2 is invalid most of the time from kexec-select-boot and verifying global hashes but is setuped correctly at disk unlock key setup - would be nice to take advantage of bash function tracing to understand where we are for debugging purposes, code takes ash in consideration only - tpmr says it implements nv calls but actually doesn't. Removing those falsely wrapped functions would help. - Implementing them would be better - REVIEW TODOS IN CODE - READD CIRCLECI CONFIG Current state: - TPM unseal works without disk unlock key and generates TOTP properly (was missing die condition at unseal to not produce always good TOTP even if invalid) - TPM disk encryption key fails. Hypothesis is that sealing with USB drivers loaded and measures in inconsistent with sealed with/without. - TPM disk unsealing happens without USB modules being loaded in non-HOTP setup. This fails. - Current tests are with fbwhiptail (no clear called so having traces on command line of what happens) - Testing with HOTP implementation for sealing/unsealing since that forces USB module loads on each boot to remove this from failing possibilities
2022-08-25 18:43:31 +00:00
fi
Uniformize vocabulary: LUKS TPM Disk Unlock Key & LUKS Disk Recovery Key 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>
2024-01-19 17:32:04 +00:00
# set default boot option only if no LUKS TPM Disk Unlock Key previously set
if [ -z "$TPM_DISK_ENCRYPTION_KEY_SET" ]; then
set_default_boot_option
fi
DEBUG "Generating hashes"
(
set -e -o pipefail
cd /boot
find ./ -type f ! -path './kexec*' -print0 |
xargs -0 sha256sum >/boot/kexec_hashes.txt 2>/dev/null
print_tree >/boot/kexec_tree.txt
)
[ $? -eq 0 ] || whiptail_error_die "Error generating kexec hashes"
param_files=$(find /boot/kexec*.txt)
[ -z "$param_files" ] &&
whiptail_error_die "No kexec parameter files to sign"
if [ "$GPG_GEN_KEY_IN_MEMORY" = "y" -a "$GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD" = "n" ]; then
#The local keyring used to generate in memory subkeys is still valid since no key has been moved to smartcard
#Local keyring passwd is ADMIN_PIN. We need to set USER_PIN to ADMIN_PIN to be able to sign next in this boot session
DEBUG "Setting GPG User PIN to GPG Admin PIN so local keyring can be used to detach-sign kexec files next"
USER_PIN=$ADMIN_PIN
fi
DEBUG "Detach-signing boot files under kexec.sig: ${param_files}"
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
if sha256sum $param_files 2>/dev/null | DO_WITH_DEBUG gpg \
--pinentry-mode loopback \
--passphrase "${USER_PIN}" \
--digest-algo SHA256 \
--detach-sign \
-a \
>/boot/kexec.sig 2>/tmp/error; then
# successful - update the validated params
if ! check_config /boot >/dev/null 2>/tmp/error; then
cat /tmp/error
ret=1
else
ret=0
fi
else
cat /tmp/error
ret=1
fi
# done writing to /boot, switch back to RO
mount -o ro,remount /boot
if [ $ret = 1 ]; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Error signing kexec boot files:\n\n$ERROR"
fi
TRACE_FUNC
}
set_default_boot_option() {
TRACE_FUNC
option_file="/tmp/kexec_options.txt"
tmp_menu_file="/tmp/kexec/kexec_menu.txt"
hash_file="/boot/kexec_default_hashes.txt"
mkdir -p /tmp/kexec/
rm $option_file 2>/dev/null
# parse boot options from grub.cfg
for i in $(find /boot -name "grub.cfg"); do
kexec-parse-boot "/boot" "$i" >>$option_file
done
# FC29/30+ may use BLS format grub config files
# https://fedoraproject.org/wiki/Changes/BootLoaderSpecByDefault
# only parse these if $option_file is still empty
if [ ! -s $option_file ] && [ -d "/boot/loader/entries" ]; then
for i in $(find /boot -name "grub.cfg"); do
kexec-parse-bls "/boot" "$i" "/boot/loader/entries" >>$option_file
done
fi
[ ! -s $option_file ] &&
whiptail_error_die "Failed to parse any boot options"
# sort boot options
sort -r $option_file | uniq >$tmp_menu_file
## save first option as default
entry=$(head -n 1 $tmp_menu_file | tail -1)
# clear existing default configs
rm "/boot/kexec_default.*.txt" 2>/dev/null
# get correct index for entry
index=$(grep -n "$entry" $option_file | cut -f1 -d ':')
# write new config
echo "$entry" >/boot/kexec_default.$index.txt
# validate boot option
(cd /boot && /bin/kexec-boot -b "/boot" -e "$entry" -f |
xargs sha256sum >$hash_file 2>/dev/null) ||
whiptail_error_die "Failed to create hashes of boot files"
TRACE_FUNC
}
report_integrity_measurements() {
TRACE_FUNC
#check for GPG key in keyring
GPG_KEY_COUNT=$(gpg -k 2>/dev/null | wc -l)
if [ "$GPG_KEY_COUNT" -ne 0 ]; then
# Check and report TOTP
# update the TOTP code every thirty seconds
date=$(date "+%Y-%m-%d %H:%M:%S %Z")
seconds=$(date "+%s")
half=$(expr \( "$seconds" % 60 \) / 30)
if [ "$CONFIG_TPM" != "y" ]; then
TOTP="NO TPM"
elif [ "$half" != "$last_half" ]; then
last_half=$half
TOTP=$(unseal-totp) >/dev/null 2>&1
fi
# Check and report on HOTP status
if [ -x /bin/hotp_verification ]; then
HOTP="Unverified"
enable_usb
for attempt in 1 2 3; do
if ! hotp_verification info >/dev/null 2>&1; then
whiptail_warning --title "WARNING: Please insert your HOTP enabled USB Security Dongle (Attempt $attempt/3)" --msgbox "Your HOTP enabled USB Security Dongle was not detected.\n\nPlease remove it and insert it again." 0 80
else
break
fi
done
if [ $attempt -eq 3 ]; then
die "No HOTP enabled USB Security Dongle detected. Please disable 'CONFIG_HOTPKEY' in the board config and rebuild."
fi
# Don't output HOTP codes to screen, so as to make replay attacks harder
HOTP=$(unseal-hotp) >/dev/null 2>&1
hotp_verification check $HOTP
case "$?" in
0)
HOTP="Success"
;;
4)
HOTP="Invalid code"
BG_COLOR_MAIN_MENU="error"
;;
*)
HOTP="Error checking code, Insert USB Security Dongle and retry"
BG_COLOR_MAIN_MENU="warning"
;;
esac
else
HOTP='N/A'
fi
# Check for detached signed digest and report on /boot integrity status
check_config /boot force
TMP_HASH_FILE="/tmp/kexec/kexec_hashes.txt"
if (cd /boot && sha256sum -c "$TMP_HASH_FILE" >/tmp/hash_output); then
HASH="OK"
else
HASH="ALTERED"
fi
#Show results
whiptail_type $BG_COLOR_MAIN_MENU --title "Measured Integrity Report" --msgbox "$date\nTOTP: $TOTP | HOTP: $HOTP\n/BOOT INTEGRITY: $HASH\n\nPress OK to continue or Ctrl+Alt+Delete to reboot" 0 80
fi
TRACE_FUNC
}
usb_security_token_capabilities_check() {
TRACE_FUNC
enable_usb
# ... first set board config preference
if [ -n "$CONFIG_GPG_ALGO" ]; then
GPG_ALGO=$CONFIG_GPG_ALGO
DEBUG "Setting GPG_ALGO to (board-)configured: $CONFIG_GPG_ALGO"
fi
# ... overwrite with usb-token capability
if lsusb | grep -q "20a0:42b2"; then
GPG_ALGO="p256"
DEBUG "Nitrokey 3 detected: Setting GPG_ALGO to: $GPG_ALGO"
fi
}
## main script start
# check for args
if [ "$1" != "" ]; then
title_text=$1
else
title_text="OEM Factory Reset / Re-Ownership"
fi
if [ "$2" != "" ]; then
bg_color=$2
else
bg_color=""
fi
# show warning prompt
if [ "$CONFIG_TPM" = "y" ]; then
TPM_STR=" * ERASE the TPM and own it with a password\n"
else
TPM_STR=""
fi
if ! whiptail_warning --yesno "
This operation will automatically:\n
$TPM_STR
* ERASE any keys or passwords on the GPG smart card,\n
reset it to a factory state, generate new keys\n
and optionally set custom PIN(s)\n
* Add the new GPG key to the firmware and reflash it\n
* Sign all of the files in /boot with the new GPG key\n\n
It requires that you already have an OS installed on a\n
dedicated /boot partition. Do you wish to continue?" \
$HEIGHT $WIDTH $CONTINUE $CANCEL $CLEAR --title "$title_text"; then
exit 1
fi
#Make sure /boot is mounted if board config defines default
mount_boot
# We show current integrity measurements status and time
report_integrity_measurements
# Clear the screen
clear
#Prompt user for use of default configuration options
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
TRACE_FUNC
echo -e -n "Would you like to use default configuration options?\nIf N, you will be prompted for each option [Y/n]: "
read -n 1 use_defaults
if [ "$use_defaults" == "n" -o "$use_defaults" == "N" ]; then
#Give general guidance to user on how to answer prompts
echo
echo "****************************************************"
echo "**** Factory Reset / Re-Ownership Questionnaire ****"
echo "****************************************************"
echo "The following questionnaire will help you configure the security components of your system."
echo "Each prompt requires a single letter answer: eg. (Y/n)."
echo -e "If you don't know what to answer, pressing Enter will select the default answer for that prompt: eg. Y, above.\n"
# Re-ownership of LUKS encrypted Disk: key, content and passphrase
echo -e -n "\n\nWould you like to change the current LUKS Disk Recovery Key passphrase?\n (Highly recommended if you didn't install the Operating System yourself, so that past configured passphrase would not permit to access content.\n Note that without re-encrypting disk, a backed up header could be restored to access encrypted content with old passphrase) [y/N]: "
read -n 1 prompt_output
echo
if [ "$prompt_output" == "y" \
-o "$prompt_output" == "Y" ]; then
luks_new_Disk_Recovery_Key_passphrase_desired=1
echo -e "\n"
fi
Uniformize vocabulary: LUKS TPM Disk Unlock Key & LUKS Disk Recovery Key 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>
2024-01-19 17:32:04 +00:00
echo -e -n "Would you like to re-encrypt LUKS encrypted container and generate new LUKS Disk Recovery Key?\n (Highly recommended if you didn't install the operating system yourself: this would prevent any LUKS backed up header to be restored to access encrypted data) [y/N]: "
read -n 1 prompt_output
echo
if [ "$prompt_output" == "y" \
-o "$prompt_output" == "Y" ]; then
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
TRACE_FUNC
test_luks_current_disk_recovery_key_passphrase
luks_new_Disk_Recovery_Key_desired=1
echo -e "\n"
fi
#Prompt to ask if user wants to generate GPG key material in memory or on smartcard
echo -e -n "Would you like to format an encrypted USB Thumb drive to store GPG key material?\n (Required to enable GPG authentication) [y/N]: "
read -n 1 prompt_output
echo
if [ "$prompt_output" == "y" \
-o "$prompt_output" == "Y" ] \
; then
GPG_GEN_KEY_IN_MEMORY="y"
echo " ++++ Master key and subkeys will be generated in memory, backed up to dedicated LUKS container +++"
echo -e -n "Would you like in-memory generated subkeys to be copied to USB Security Dongle's smartcard?\n (Highly recommended so the smartcard is used on daily basis and backup is kept safe, but not required) [Y/n]: "
read -n 1 prompt_output
echo
if [ "$prompt_output" == "n" \
-o "$prompt_output" == "N" ]; then
warn "Subkeys will NOT be copied to USB Security Dongle's smartcard"
warn "Your GPG key material backup thumb drive should be cloned to a second thumb drive for redundancy for production environements"
GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD="n"
else
echo "++++ Subkeys will be copied to USB Security Dongle's smartcard ++++"
warn "Please keep your GPG key material backup thumb drive safe"
GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD="y"
fi
else
echo "GPG key material will be generated on USB Security Dongle's smartcard without backup"
GPG_GEN_KEY_IN_MEMORY="n"
GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD="n"
fi
# Dynamic messages to be given to user in terms of security components that will be applied
# based on previous answers
CUSTOM_PASS_AFFECTED_COMPONENTS="\n"
# Adapt message to be given to user in terms of security components that will be applied.
if [ -n "$luks_new_Disk_Recovery_Key_passphrase_desired" -o -n "$luks_new_Disk_Recovery_Key_passphrase" ]; then
CUSTOM_PASS_AFFECTED_COMPONENTS+="LUKS Disk Recovery Key passphrase\n"
fi
if [ "$CONFIG_TPM" = "y" ]; then
CUSTOM_PASS_AFFECTED_COMPONENTS+="TPM Owner Password\n"
fi
if [ "$GPG_GEN_KEY_IN_MEMORY" = "y" ]; then
CUSTOM_PASS_AFFECTED_COMPONENTS+="GPG Key material backup passphrase (Same as GPG Admin PIN)\n"
fi
CUSTOM_PASS_AFFECTED_COMPONENTS+="GPG Admin PIN\n"
# Only show GPG User PIN as affected component if GPG_GEN_KEY_IN_MEMORY not requested or GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD is
if [ "$GPG_GEN_KEY_IN_MEMORY" = "n" -o "$GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD" = "y" ]; then
CUSTOM_PASS_AFFECTED_COMPONENTS+="GPG User PIN\n"
fi
# Inform user of security components affected for the following prompts
echo
echo -e "The following Security Components will be configured with defaults or further chosen PINs/passwords:
$CUSTOM_PASS_AFFECTED_COMPONENTS\n"
# Prompt to change default passwords
echo -e -n "Would you like to set a single custom password to all previously stated security components? [y/N]: "
read -n 1 prompt_output
echo
if [ "$prompt_output" == "y" \
-o "$prompt_output" == "Y" ]; then
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
echo -e "\nThe chosen custom password must be between 8 and $MAX_HOTP_GPG_PIN_LENGTH characters in length."
while [[ ${#CUSTOM_SINGLE_PASS} -lt 8 ]] || [[ ${#CUSTOM_SINGLE_PASS} -gt $MAX_HOTP_GPG_PIN_LENGTH ]]; do
echo -e -n "Enter the custom password: "
read CUSTOM_SINGLE_PASS
done
echo
TPM_PASS=${CUSTOM_SINGLE_PASS}
USER_PIN=${CUSTOM_SINGLE_PASS}
ADMIN_PIN=${CUSTOM_SINGLE_PASS}
# Only set if user said desired
if [ -n "$luks_new_Disk_Recovery_Key_passphrase_desired" ]; then
luks_new_Disk_Recovery_Key_passphrase=${CUSTOM_SINGLE_PASS}
fi
else
echo -e -n "Would you like to set distinct PINs/passwords to configure previously stated security components? [y/N]: "
read -n 1 prompt_output
echo
if [ "$prompt_output" == "y" \
-o "$prompt_output" == "Y" ]; then
echo -e "\nThe TPM Owner Password and Admin PIN must be at least 8, the User PIN at least 6 characters in length.\n"
echo
if [ "$CONFIG_TPM" = "y" ]; then
while [[ ${#TPM_PASS} -lt 8 ]]; do
echo -e -n "Enter desired TPM Owner Password: "
read TPM_PASS
done
fi
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
while [[ ${#ADMIN_PIN} -lt 6 ]] || [[ ${#ADMIN_PIN} -gt $MAX_HOTP_GPG_PIN_LENGTH ]]; do
echo -e -n "\nThis PIN should be between 6 to $MAX_HOTP_GPG_PIN_LENGTH characters in length.\n"
echo -e -n "Enter desired GPG Admin PIN: "
read ADMIN_PIN
done
#USER PIN not required in case of GPG_GEN_KEY_IN_MEMORY not requested of if GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD is
# That is, if keys were NOT generated in memory (on smartcard only) or
# if keys were generated in memory but are to be moved from local keyring to smartcard
if [ "$GPG_GEN_KEY_IN_MEMORY" = "n" -o "$GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD" = "y" ]; then
while [[ ${#USER_PIN} -lt 6 ]] || [[ ${#USER_PIN} -gt $MAX_HOTP_GPG_PIN_LENGTH ]]; do
echo -e -n "\nThis PIN should be between 6 to $MAX_HOTP_GPG_PIN_LENGTH characters in length.\n"
echo -e -n "Enter desired GPG User PIN: "
read USER_PIN
done
fi
echo
fi
fi
if [ -n "$luks_new_Disk_Recovery_Key_passphrase_desired" -a -z "$luks_new_Disk_Recovery_Key_passphrase" ]; then
# We catch here if changing LUKS Disk Recovery Key passphrase was desired
# but yet undone. This is if not being covered by the single password
Uniformize vocabulary: LUKS TPM Disk Unlock Key & LUKS Disk Recovery Key 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>
2024-01-19 17:32:04 +00:00
echo -e "\nEnter desired replacement for current LUKS Disk Recovery Key passphrase (At least 8 characters long):"
while [[ ${#luks_new_Disk_Recovery_Key_passphrase} -lt 8 ]]; do
{
read -r luks_new_Disk_Recovery_Key_passphrase
}
done
Uniformize vocabulary: LUKS TPM Disk Unlock Key & LUKS Disk Recovery Key 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>
2024-01-19 17:32:04 +00:00
#We test that current LUKS Disk Recovery Key passphrase is known prior of going further
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
TRACE_FUNC
test_luks_current_disk_recovery_key_passphrase
echo -e "\n"
fi
# Prompt to change default GnuPG key information
echo -e -n "Would you like to set custom user information for the GnuPG key? [y/N]: "
read -n 1 prompt_output
echo
if [ "$prompt_output" == "y" \
-o "$prompt_output" == "Y" ]; then
echo -e "\n\n"
echo -e "We will generate a GnuPG (PGP) keypair identifiable with the following text form:"
echo -e "Real Name (Comment) email@address.org"
echo -e "\nEnter your Real Name (Optional):"
read -r GPG_USER_NAME
echo -e "\nEnter your email@adress.org:"
read -r GPG_USER_MAIL
while ! $(expr "$GPG_USER_MAIL" : '.*@' >/dev/null); do
{
echo -e "\nEnter your email@address.org:"
read -r GPG_USER_MAIL
}
done
echo -e "\nEnter Comment (Optional, to distinguish this key from others with same previous attributes. Must be smaller then 60 characters):"
read -r GPG_USER_COMMENT
while [[ ${#GPG_USER_COMMENT} -gt 60 ]]; do
{
echo -e "\nEnter Comment (Optional, to distinguish this key from others with same previous attributes. Must be smaller then 60 characters):"
read -r GPG_USER_COMMENT
}
done
fi
if [ "$GPG_GEN_KEY_IN_MEMORY" = "y" ]; then
select_thumb_drive_for_key_material
fi
fi
# If nothing is stored in custom variables, we set them to their defaults
if [ "$TPM_PASS" == "" ]; then TPM_PASS=${TPM_PASS_DEF}; fi
if [ "$USER_PIN" == "" ]; then USER_PIN=${USER_PIN_DEF}; fi
if [ "$ADMIN_PIN" == "" ]; then ADMIN_PIN=${ADMIN_PIN_DEF}; fi
2020-07-07 08:32:22 +00:00
## sanity check the USB, GPG key, and boot device before proceeding further
if [ "$GPG_GEN_KEY_IN_MEMORY" = "n" ]; then
# Prompt to insert USB drive if desired
echo -e -n "\nWould you like to export your public key to an USB drive? [y/N]: "
read -n 1 prompt_output
echo
if [ "$prompt_output" == "y" \
-o "$prompt_output" == "Y" ] \
; then
GPG_EXPORT=1
# mount USB over /media only if not already mounted
if ! grep -q /media /proc/mounts; then
# mount USB in rw
if ! mount-usb --mode rw 2>/tmp/error; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Unable to mount USB on /media:\n\n${ERROR}"
fi
else
#/media already mounted, make sure it is in r+w mode
if ! mount -o remount,rw /media 2>/tmp/error; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Unable to remount in read+write USB on /media:\n\n${ERROR}"
fi
fi
else
GPG_EXPORT=0
# needed for USB Security Dongle below and is ensured via mount-usb in case of GPG_EXPORT=1
enable_usb
2020-07-07 08:32:22 +00:00
fi
fi
# ensure USB Security Dongle connected if GPG_GEN_KEY_IN_MEMORY=n or if GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD=y
if [ "$GPG_GEN_KEY_IN_MEMORY" = "n" -o "$GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD" = "y" ]; then
echo -e "\nChecking for USB Security Dongle...\n"
enable_usb
if ! gpg --card-status >/dev/null 2>&1; then
local_whiptail_error "Can't access USB Security Dongle; \nPlease remove and reinsert, then press Enter."
if ! gpg --card-status >/dev/null 2>/tmp/error; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Unable to detect USB Security Dongle:\n\n${ERROR}"
fi
fi
#Now that USB Security Dongle is detected, we can check its capabilities and limitations
usb_security_token_capabilities_check
fi
assert_signable
# Action time...
# clear gpg-agent cache so that next gpg calls doesn't have past keyring in memory
killall gpg-agent >/dev/null 2>&1 || true
# clear local keyring
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
rm -rf /.gnupg/*.kbx /.gnupg/*.gpg >/dev/null 2>&1 || true
# detect and set /boot device
echo -e "\nDetecting and setting boot device...\n"
if ! detect_boot_device; then
SKIP_BOOT="y"
else
echo -e "Boot device set to $CONFIG_BOOT_DEV\n"
fi
# update configs
if [[ "$SKIP_BOOT" == "n" ]]; then
replace_config /etc/config.user "CONFIG_BOOT_DEV" "$CONFIG_BOOT_DEV"
combine_configs
fi
if [ -n "$luks_new_Disk_Recovery_Key_desired" -a -n "$luks_new_Disk_Recovery_Key_passphrase_desired" ]; then
Uniformize vocabulary: LUKS TPM Disk Unlock Key & LUKS Disk Recovery Key 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>
2024-01-19 17:32:04 +00:00
#Reencryption of disk, LUKS Disk Recovery Key and LUKS Disk Recovery Key passphrase change is requested
luks_reencrypt
cryptsetup2 toolstack version bump and script fixes to support multi-LUKS containers (BTRFS QubesOS 4.2) cryptsetup2 2.6.1 is a new release that supports reencryption of Q4.2 release LUKS2 volumes created at installation. This is a critical feature for the Qubes OS 4.2 release for added data at rest protection Cryptsetup 2.6.x internal changes: - Argon2 used externally and internally: requires a lot of RAM and CPU to derivate passphrase to key validated in key slots. - This is used to rate limit efficiently bruteforcing of LUKS key slots, requiring each offline brute force attempt to consume ~15-30 seconds per attempt - OF course, strong passphrases are still recommended, but bruteforcing LUKSv2 containers with Argon2 would require immense time, ram and CPU even to bruteforce low entropy passphrase/PINs. - passphrase change doesn't permit LUKS key slot specification anymore: key slot rotates (new one consusumed per op: then old one wiped internally. EG: LUKS key slot 1 created, then 0 deleted) - reencryption doesn't permit old call arguments. No more direct-io; inadmissively slow through AIO (async) calls, need workarounds for good enough perfs (arguments + newer kernel with cloudfare fixes in tree) cryptsetup 2.6.1 requires: - lvm2 2.03.23, which is also included in this PR. - requires libaio, which is also included in this PR (could be hacked out but deep dependency at first sight: left in) - requires util-linux 2.39 - patches for reproducible builds are included for above 3 packages. luks-functions was updated to support the new cryptsetup2 version calls/changes - reencryption happen in direct-io, offline mode and without locking, requiring linux 5.10.9+ to bypass linux queues - from tests, this is best for performance and reliability in single-user mode - LUKS container ops now validate Disk Recovery Key (DRK) passphrase prior and DRK key slot prior of going forward if needed, failing early. - Heads don't expect DRK to be in static key slot anymore, and finds the DRK key slot dynamically. - If reencrytipn/passphrase change: make sure all LUKS containers on same block device can be unlocked with same DRK - Reencryption: requires to know which key slot to reencrypt. - Find LUKS key slot that unlocks with DRK passphrase unlock prior of reencrypt call - Passphrase change: no slot can be passed, but key slot of DRK rotates. kexec-seal-key - TPM LUKS Disk Unlock Key key slots have changed to be set in max slots per LUKS version (LUKSv1:7 /LUKSv2: 31) - If key slot != default LUKS version's keyslot outside of DRK key slot: prompt the user before wiping that key slot, otherwise wipe automatically - This takes for granted that the DRK key slot alone is needed on the system and Heads controls the LUKS key slots. - If user has something else going on, ie: Using USB Security dongle + TPM DUK, then the user will need to say no when wiping keys. - It was suggested to leave LUKS key slots outside of DRK alone, but then: what to do when all key slots would be used? - Alternative implementation could be to only prompt users to wipe keyslots other then DRK when key slots are all used (LUKSv1: 0-7, LUKSv2: 0-31) - But then cleanup would need to happen prior of operations (LUKS passphrase change, TPM DUK setup) and could be problematic. - LUKS containers now checked to be same LUKS version prior of permitting to set TPM DUK and will refuse to go forward of different versions. TODO: - async (AIO) calls are not used. direct-io is used instead. libaio could be hacked out - this could be subject to future work Notes: - time to deprecated legacy boards the do not enough space for the new space requirements - x230-legacy, x230-legacy-flash, x230-hotp-legacy - t430-legacy, t430-legacy-flash, t430-hotp-legacy already deprecated Unrelated: - typos fixes found along the way Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-08-16 19:20:11 +00:00
luks_change_passphrase
elif [ -n "$luks_new_Disk_Recovery_Key_desired" -a -z "$luks_new_Disk_Recovery_Key_passphrase_desired" ]; then
#Reencryption of disk was requested but not passphrase change
luks_reencrypt
elif [ -z "$luks_new_Disk_Recovery_Key_desired" -a -n "$luks_new_Disk_Recovery_Key_passphrase_desired" ]; then
#Passphrase change is requested without disk reencryption
luks_change_passphrase
fi
## reset TPM and set password
if [ "$CONFIG_TPM" = "y" ]; then
echo -e "\nResetting TPM...\n"
tpmr reset "$TPM_PASS" >/dev/null 2>/tmp/error
Addition of qemu-(fb)whiptail-tpm2(-hotp) boards -coreboot support of TPM v2.0 (shared config for TPM2 support across all 4 previous variations) -swtpm set to be launched under TPM v2.0 mode under board config -Documentation file under each board.md softlinks to qemu-coreboot-fbwhiptail-tpm1.md (which has been generalized) This is skeleton for TPM v2 integration under Heads ------------- WiP TODO: - libcurl cannot be built as a tpm2-tools dependency as of now not sure why. curl currently needs to be added in board config to be built - Note: tpm-reset (master and here) needs some review, no handle of no tpm use case. Caller is responsible to not call it otherwise does nothing - init tries to bind fd and fails currently - Note: Check if whiptail is different of fbwhiptail in clearing screen. As of now every clear seems to be removed, still whiptail clears previous console output - When no OS' /boot can be mounted, do not try to TPM reset (will fail) - seal-hotpkey is not working properly - setting disk unlock key asks for TPM ownership passphrase (sealing in NV requires ownership, but text is misleading user as if reowning TPM) - We should cache input, feed tpm behind the scene and wipe passphrase and state clearly that this is TPM disk unlock kye passphrase. - primary key from TPM2 is invalid most of the time from kexec-select-boot and verifying global hashes but is setuped correctly at disk unlock key setup - would be nice to take advantage of bash function tracing to understand where we are for debugging purposes, code takes ash in consideration only - tpmr says it implements nv calls but actually doesn't. Removing those falsely wrapped functions would help. - Implementing them would be better - REVIEW TODOS IN CODE - READD CIRCLECI CONFIG Current state: - TPM unseal works without disk unlock key and generates TOTP properly (was missing die condition at unseal to not produce always good TOTP even if invalid) - TPM disk encryption key fails. Hypothesis is that sealing with USB drivers loaded and measures in inconsistent with sealed with/without. - TPM disk unsealing happens without USB modules being loaded in non-HOTP setup. This fails. - Current tests are with fbwhiptail (no clear called so having traces on command line of what happens) - Testing with HOTP implementation for sealing/unsealing since that forces USB module loads on each boot to remove this from failing possibilities
2022-08-25 18:43:31 +00:00
fi
if [ $? -ne 0 ]; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Error resetting TPM:\n\n${ERROR}"
fi
# clear local keyring
rm /.gnupg/*.gpg 2>/dev/null
rm /.gnupg/*.kbx 2>/dev/null
# initialize gpg wth empty keyring
gpg --list-keys >/dev/null 2>&1
#Generate keys in memory and copy to smartcard
if [ "$GPG_GEN_KEY_IN_MEMORY" = "y" ]; then
if [ "$GPG_ALGO" == "RSA" ]; then
# Generate GPG master key
generate_inmemory_RSA_master_and_subkeys
elif [ "$GPG_ALGO" == "p256" ]; then
generate_inmemory_p256_master_and_subkeys
else
die "Unsupported GPG_ALGO: $GPG_ALGO"
fi
wipe_thumb_drive_and_copy_gpg_key_material "$thumb_drive" "$thumb_drive_luks_percent"
set_user_config "CONFIG_HAVE_GPG_KEY_BACKUP" "y"
if [ "$GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD" = "y" ]; then
keytocard_subkeys_to_smartcard
fi
else
#Generate GPG key and subkeys on smartcard only
echo -e "\nResetting USB Security Dongle's GPG smartcard...\n(this will take around 3 minutes...)\n"
gpg_key_factory_reset
generate_OEM_gpg_keys
fi
# Obtain GPG key ID
GPG_GEN_KEY=$(gpg --list-keys --with-colons | grep "^fpr" | cut -d: -f10 | head -n1)
#Where to export the public key
PUBKEY="/tmp/${GPG_GEN_KEY}.asc"
# export pubkey to file
if ! gpg --export --armor "$GPG_GEN_KEY" >"${PUBKEY}" 2>/tmp/error; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "GPG Key gpg export to file failed!\n\n$ERROR"
fi
#Applying custom GPG PINs to the smartcard if they were provided
if [ "$GPG_GEN_KEY_IN_MEMORY" = "n" -o "$GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD" = "y" ]; then
#Only apply smartcard PIN change if smartcard only or if keytocard op is expected next
if [ "${USER_PIN}" != "" -o "${ADMIN_PIN}" != "" ]; then
echo -e "\nChanging default GPG Admin PIN\n"
gpg_key_change_pin "3" "${ADMIN_PIN_DEF}" "${ADMIN_PIN}"
echo -e "\nChanging default GPG User PIN\n"
gpg_key_change_pin "1" "${USER_PIN_DEF}" "${USER_PIN}"
fi
fi
2020-07-07 09:16:18 +00:00
## export pubkey to USB
if [ "$GPG_EXPORT" != "0" ]; then
echo -e "\nExporting generated key to USB...\n"
# copy to USB
if ! cp "${PUBKEY}" "/media/${GPG_GEN_KEY}.asc" 2>/tmp/error; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Key export error: unable to copy ${GPG_GEN_KEY}.asc to /media:\n\n$ERROR"
2020-07-07 08:32:22 +00:00
fi
mount -o remount,ro /media 2>/dev/null
fi
# ensure key imported locally
Fix key to card failing with invalid time when moving keys to smartcard on master (Opt: Authenticated Heads) - Revert gnupg toolstack version bump to prior of #1661 merge (2.4.2 -> 2.4.0). Version bump not needed for reproducibility. - Investigation and upstream discussions will take their time resolving invalid time issue introduced by between 2.4.0 and latest gnupg, fix regression first under master) - oem-factory-reset - Adding DO_WITH_DEBUG to oem-factory-reset for all its gpg calls. If failing in debug mode, /tmp/debug.txt contains calls and errors - Wipe keyrings only (*.gpg, *.kbx) not conf files under gpg homedir (keep initrd/.gnupg/*.conf) - flake.nix - switch build derivative from qemu and qemu_kvm to qemu_full to have qemu-img tool which was missing to run qemu boards (v0.1.8 docker) - add gnupg so that qemu boards can call inject_gpg to inject public key in absence of flashrom+pflash support for internal flashing - flake.lock: Updated nix pinned package list under flake.lock with 'nix flake update' so qemu_full builds - README.md: have consistent docker testing + release (push) notes - .circleci/config.yml: depend on docker v0.1.8 (qemu_full built with canokey-qemu lib support, diffoscopeMinimal and gnupg for proper qemu testing) TODO: - some fd2 instead of fd1?! - oem-factory-resest has whiptail_or_die which sets whiptail box to HEIGHT 0. This doesn't show a scrolling window on gpg errors which is problematic with fbwhiptail, not whiptail Signed-off-by: Thierry Laurion <insurgo@riseup.net>
2024-05-14 16:44:11 +00:00
if ! cat "$PUBKEY" | DO_WITH_DEBUG gpg --import >/dev/null 2>/tmp/error; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Error importing GPG key:\n\n$ERROR"
fi
# update /.gnupg/trustdb.gpg to ultimately trust all user provided public keys
if ! gpg --list-keys --fingerprint --with-colons 2>/dev/null |
sed -E -n -e 's/^fpr:::::::::([0-9A-F]+):$/\1:6:/p' |
gpg --import-ownertrust >/dev/null 2>/tmp/error; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Error importing GPG ownertrust:\n\n$ERROR"
fi
if ! gpg --update-trust >/dev/null 2>/tmp/error; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Error updating GPG ownertrust:\n\n$ERROR"
fi
# Do not attempt to flash the key to ROM if we are running in QEMU based on CONFIG_BOARD_NAME matching glob pattern containing qemu-*
# We check for qemu-* instead of ^qemu- because CONFIG_BOARD_NAME could be renamed to UNTESTED-qemu-* in a probable future
if [[ "$CONFIG_BOARD_NAME" == qemu-* ]]; then
warn "Skipping flash of GPG key to ROM because we are running in QEMU without internal flashing support."
warn "Please review boards/qemu*/qemu*.md documentation to extract public key from raw disk and inject at build time"
warn "Also review boards/qemu*/qemu*.config to tweak CONFIG_* options you might need to turn on/off manually at build time"
else
#We are not running in QEMU, so flash the key to ROM
## flash generated key to ROM
echo -e "\nReading current firmware...\n(this will take a minute or two)\n"
/bin/flash.sh -r /tmp/oem-setup.rom >/dev/null 2>/tmp/error
if [ ! -s /tmp/oem-setup.rom ]; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Error reading current firmware:\n\n$ERROR"
fi
# clear any existing heads/gpg files from current firmware
for i in $(cbfs.sh -o /tmp/oem-setup.rom -l | grep -e "heads/"); do
cbfs.sh -o /tmp/oem-setup.rom -d "$i"
done
# add heads/gpg files to current firmware
if [ -e /.gnupg/pubring.kbx ]; then
cbfs.sh -o /tmp/oem-setup.rom -a "heads/initrd/.gnupg/pubring.kbx" -f /.gnupg/pubring.kbx
if [ -e /.gnupg/pubring.gpg ]; then
rm /.gnupg/pubring.gpg
fi
elif [ -e /.gnupg/pubring.gpg ]; then
cbfs.sh -o /tmp/oem-setup.rom -a "heads/initrd/.gnupg/pubring.gpg" -f /.gnupg/pubring.gpg
fi
if [ -e /.gnupg/trustdb.gpg ]; then
cbfs.sh -o /tmp/oem-setup.rom -a "heads/initrd/.gnupg/trustdb.gpg" -f /.gnupg/trustdb.gpg
fi
# persist user config changes (boot device)
if [ -e /etc/config.user ]; then
cbfs.sh -o /tmp/oem-setup.rom -a "heads/initrd/etc/config.user" -f /etc/config.user
fi
# flash updated firmware image
echo -e "\nAdding generated key to current firmware and re-flashing...\n"
if ! /bin/flash.sh /tmp/oem-setup.rom 2>/tmp/error; then
ERROR=$(tail -n 1 /tmp/error | fold -s)
whiptail_error_die "Error flashing updated firmware image:\n\n$ERROR"
fi
fi
## sign files in /boot and generate checksums
if [[ "$SKIP_BOOT" == "n" ]]; then
echo -e "\nSigning boot files and generating checksums...\n"
generate_checksums
fi
# passphrases set to be empty first
passphrases="\n"
# Prepare whiptail output of configured secrets
if [ -n "$luks_new_Disk_Recovery_Key_passphrase" -o -n "$luks_new_Disk_Recovery_Key_passphrase_desired" ]; then
passphrases+="LUKS Disk Recovery Key passphrase: ${luks_new_Disk_Recovery_Key_passphrase}\n"
fi
if [ "$CONFIG_TPM" = "y" ]; then
passphrases+="TPM Owner Password: ${TPM_PASS}\n"
fi
#GPG PINs output
passphrases+="GPG Admin PIN: ${ADMIN_PIN}\n"
#USER PIN was configured if GPG_GEN_KEY_IN_MEMORY is not active or if GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD is active
if [ "$GPG_GEN_KEY_IN_MEMORY" = "n" -o "$GPG_GEN_KEY_IN_MEMORY_COPY_TO_SMARTCARD" = "y" ]; then
passphrases+="GPG User PIN: ${USER_PIN}\n"
fi
#If user decided to generate keys in memory, we add the thumb drive passphrase
if [ "$GPG_GEN_KEY_IN_MEMORY" = "y" ]; then
passphrases+="GPG key material backup passphrase: ${ADMIN_PIN}\n"
fi
## Show to user current configured secrets prior of rebooting
whiptail --msgbox "
$(echo -e "$passphrases" | fold -w $((WIDTH-5)))" \
$HEIGHT $WIDTH --title "Configured secrets"
## all done -- reboot
whiptail --msgbox "
OEM Factory Reset / Re-Ownership has completed successfully\n\n
After rebooting, you will need to generate new TOTP/HOTP secrets\n
when prompted in order to complete the setup process.\n\n
Press Enter to reboot.\n" \
$HEIGHT $WIDTH --title "OEM Factory Reset / Re-Ownership Complete"
# Clean LUKS secrets
luks_secrets_cleanup
unset luks_passphrase_changed
unset tpm_owner_password_changed
reboot