heads/initrd/bin/kexec-seal-key

228 lines
6.9 KiB
Plaintext
Raw Normal View History

#!/bin/bash
# This will generate a disk encryption key and seal / ecncrypt
# with the current PCRs and then store it in the TPM NVRAM.
# It will then need to be bundled into initrd that is booted.
set -e -o pipefail
. /etc/functions
TPM_INDEX=3
TPM_SIZE=312
KEY_FILE="/tmp/secret/secret.key"
TPM_SEALED="/tmp/secret/secret.sealed"
RECOVERY_KEY="/tmp/secret/recovery.key"
. /etc/functions
. /tmp/config
paramsdir=$1
if [ -z "$paramsdir" ]; then
die "Usage $0 /boot"
fi
KEY_DEVICES="$paramsdir/kexec_key_devices.txt"
KEY_LVM="$paramsdir/kexec_key_lvm.txt"
if [ ! -r "$KEY_DEVICES" ]; then
die "No devices defined for disk encryption"
fi
if [ -r "$KEY_LVM" ]; then
# Activate the LVM volume group
VOLUME_GROUP=`cat $KEY_LVM`
if [ -z "$VOLUME_GROUP" ]; then
die "No LVM volume group defined for activation"
fi
lvm vgchange -a y $VOLUME_GROUP \
|| die "$VOLUME_GROUP: unable to activate volume group"
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
echo "DEBUG: CONFIG_TPM: $CONFIG_TPM"
echo "DEBUG: CONFIG_TPM2_TOOLS: $CONFIG_TPM2_TOOLS"
echo "DEBUG: Show PCRs"
pcrs
# LUKS Key slot 0 is the manual recovery pass phrase
# that they user entered when they installed OS,
# key slot 1 is the one that we've generated.
read -s -p "Enter disk recovery key: " disk_password
echo -n "$disk_password" > "$RECOVERY_KEY"
echo
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
if [ "$CONFIG_AUTO_UNLOCK" != y ]; then
read -s -p "New disk unlock password for booting: " key_password
echo
read -s -p "Repeat unlock code: " key_password2
echo
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
if [ "$key_password" != "$key_password2" ]; then
die "Key passwords do not match"
fi
fi
# Generate key file
dd \
if=/dev/urandom \
of="$KEY_FILE" \
bs=1 \
count=128 \
2>/dev/null \
|| die "Unable to generate 128 random bytes"
# Remove all the old keys from slot 1
for dev in `cat "$KEY_DEVICES" | cut -d\ -f1`; do
echo "++++++ $dev: Removing old key slot"
cryptsetup luksKillSlot \
--key-file "$RECOVERY_KEY" \
$dev 1 \
|| warn "$dev: ignoring problem"
echo "++++++ $dev: Adding key"
cryptsetup luksAddKey \
--key-file "$RECOVERY_KEY" \
--key-slot 1 \
$dev "$KEY_FILE" \
|| die "$dev: Unable to add key"
done
# Now that we have setup the new keys, measure the PCRs
# We don't care what ends up in PCR 6; we just want
# to get the /tmp/luksDump.txt file. We use PCR16
# since it should still be zero
cat "$KEY_DEVICES" | cut -d\ -f1 | xargs /bin/qubes-measure-luks \
|| die "Unable to measure the LUKS headers"
if [ "$CONFIG_TPM" = "y" ] && [ "$CONFIG_TPM2_TOOLS" != "y" ]; then
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
luks_pcr=`tpm calcfuturepcr -ix 16 -if /tmp/luksDump.txt`
# HOTP USB Secrity Dongle loads USB modules which changes PCR5.
# In the event HOTP USB Security Dongle is enabled, skip verification of PCR5
# If modules should be loaded during normal boot, skip verification of PCR5
#
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
#TODO: check condition, no libdata.ko
if [ "$CONFIG_USB_KEYBOARD" = "y" -o -r /lib/modules/libata.ko -o -x /bin/hotp_verification ]; then
pcr_5="X"
else
pcr_5="0000000000000000000000000000000000000000"
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
# Note that PCR 4 needs to be set with the "normal-boot"
# path value, which we do not have right now since we are
# in a "init" mode
# used to be -ix 4 f8fa3b6e32e7c6fe04c366e74636e505b28f3b0d \
# now just all zeros in a normal boot
# PCR 5 must be all zero since no kernel modules should have
# been loaded during a normal boot, but might have been
# loaded in the recovery shell.
# Otherwise use the current values of the PCRs, which will be read
# from the TPM as part of the sealing ("X").
if [ "$CONFIG_AUTO_UNLOCK" != y ]; then
tpm sealfile2 \
-if "$KEY_FILE" \
-of "$TPM_SEALED" \
-pwdd "$key_password" \
-hk 40000000 \
-ix 0 X \
-ix 1 X \
-ix 2 X \
-ix 3 X \
-ix 4 0000000000000000000000000000000000000000 \
-ix 5 $pcr_5 \
-ix 6 $luks_pcr \
-ix 7 X
else
tpm sealfile2 \
-if "$KEY_FILE" \
-of "$TPM_SEALED" \
-hk 40000000 \
-ix 0 X \
-ix 1 X \
-ix 2 X \
-ix 3 X \
-ix 4 0000000000000000000000000000000000000000 \
-ix 5 $pcr_5 \
-ix 6 $luks_pcr \
-ix 7 X
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
if [ $? -ne 0 ]; then
die "Unable to seal secret"
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
shred -n 10 -z -u "$KEY_FILE" 2> /dev/null \
|| die "Failed to delete key file"
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
# try it without the owner password first
if ! tpm nv_writevalue \
-in $TPM_INDEX \
-if "$TPM_SEALED" \
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
; then
# to create an nvram space we need the TPM owner password
# and the TPM physical presence must be asserted.
#
# The permissions are 0 since there is nothing special
# about the sealed file
tpm physicalpresence -s \
|| warn "Warning: Unable to assert physical presence"
read -s -p "TPM Owner password: " tpm_password
echo
tpm nv_definespace \
-in $TPM_INDEX \
-sz $TPM_SIZE \
-pwdo "$tpm_password" \
-per 0 \
|| warn "Warning: Unable to define NVRAM space; trying anyway"
tpm nv_writevalue \
-in $TPM_INDEX \
-if "$TPM_SEALED" \
|| die "Unable to write sealed secret to NVRAM"
fi
elif [ "$CONFIG_TPM2_TOOLS" = "y" ]; then
pcrf="/tmp/secret/pcrf.bin"
#TODO: verify logic below against tpm2
tpm2 pcrread -o "$pcrf" sha256:0,1,2,3
# pcr 4 is expected to be zero (init mode)
dd if=/dev/zero bs=32 count=1 >> "$pcrf"
if [ "$CONFIG_USB_KEYBOARD" = "y" -o -r /lib/modules/libata.ko -o -x /bin/hotp_verification ]; then
# Here, we take pcr 5 into consideration if modules are expected to be measured+loaded
#
# binary pcr dump can only go via stderr for redirection
# sadly this busybox has buggy support for "2>>" and "2|"
# so we need workaround...
tpm2 pcrread -o /dev/stderr sha256:5 2>&1 >/dev/console | cat >> "$pcrf"
else
#no kernel modules are expected to be measured+loaded
dd if=/dev/zero bs=32 count=1 >> "$pcrf"
fi
# Use pcr 23 to precompute the value for pcr 6
tpmr extend -ix 23 -if /tmp/luksDump.txt
tpm2 pcrread -o /dev/stderr sha256:23 2>&1 >/dev/console | cat >> "$pcrf"
#TODO: delete the following pcrs output on screen
# goal is to validate that what is in pcr 23 is at pcr 6 at unseal
pcrs
tpm2 pcrreset 23
# We take into consideration user files in cbfs
tpm2 pcrread -o /dev/stderr sha256:7 2>&1 >/dev/console | cat >> "$pcrf"
if [ "$CONFIG_AUTO_UNLOCK" = "y" ]; then
#TODO: There is no CONFIG_AUTO_UNLOCK. Should be removed
tpmr seal "$KEY_FILE" "0x8100000$TPM_INDEX" sha256:0,1,2,3,4,5,6,7 "$pcrf"
else
#TODO: wrap TPM disk encryption key passphrase, otherwise prompt to user looks like if we were asking for TPM ownership passphrase
#TODO: everything is supposed to be under $pcrf, why considering them twice?
# TODO: review syntax to not duplicate expending pcr 2 times with pcr0-7: find a way to only use $pcrf? : sha256 "$pcrf" "$key_password"
tpmr seal "$KEY_FILE" "0x8100000$TPM_INDEX" sha256:0,1,2,3,4,5,6,7 "$pcrf" "$key_password"
fi
if [ $? -eq 0 ]; then
# should be okay if this fails
shred -n 10 -z -u "$pcrf".* 2> /dev/null || true
fi
fi
shred -n 10 -z -u "$TPM_SEALED" 2> /dev/null \
|| warn "Failed to delete the sealed secret - continuing"
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
cp /tmp/luksDump.txt "$paramsdir/kexec_lukshdr_hash.txt" \
|| warn "Failed to have hashes of LUKS header - continuing"