Re: [PATCH v3 08/11] PM: hibernate: Use TPM-backed keys to encrypt image

From: Jarkko Sakkinen
Date: Sun Oct 23 2022 - 17:55:30 EST


On Fri, Oct 21, 2022 at 12:56:50PM -0700, Evan Green wrote:
> On Fri, Sep 30, 2022 at 2:35 PM Jarkko Sakkinen <jarkko@xxxxxxxxxx> wrote:
> >
> > On Tue, Sep 27, 2022 at 09:49:19AM -0700, Evan Green wrote:
> > > When using encrypted hibernate images, have the TPM create a key for us
> > > and seal it. By handing back a sealed blob instead of the raw key, we
> > > prevent usermode from being able to decrypt and tamper with the
> > > hibernate image on a different machine.
> > >
> > > We'll also go through the motions of having PCR23 set to a known value at
> > > the time of key creation and unsealing. Currently there's nothing that
> > > enforces the contents of PCR23 as a condition to unseal the key blob,
> > > that will come in a later change.
> > >
> > > Sourced-from: Matthew Garrett <mjg59@xxxxxxxxxx>
> > > Signed-off-by: Evan Green <evgreen@xxxxxxxxxxxx>
> > >
> > > ---
> > > Matthew's incarnation of this patch is at:
> > > https://patchwork.kernel.org/project/linux-pm/patch/20210220013255.1083202-9-matthewgarrett@xxxxxxxxxx/
> > >
> > > Changes in v3:
> > > - ENCRYPTED_HIBERNATION needs TRUSTED_KEYS builtin for
> > > key_type_trusted.
> > > - Remove KEYS dependency since it's covered by TRUSTED_KEYS (Kees)
> > >
> > > Changes in v2:
> > > - Rework load/create_kernel_key() to eliminate a label (Andrey)
> > > - Call put_device() needed from calling tpm_default_chip().
> > >
> > > kernel/power/Kconfig | 1 +
> > > kernel/power/snapenc.c | 207 +++++++++++++++++++++++++++++++++++++++--
> > > kernel/power/user.h | 1 +
> > > 3 files changed, 200 insertions(+), 9 deletions(-)
> > >
> > > diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig
> > > index cd574af0b43379..2f8acbd87b34dc 100644
> > > --- a/kernel/power/Kconfig
> > > +++ b/kernel/power/Kconfig
> > > @@ -96,6 +96,7 @@ config ENCRYPTED_HIBERNATION
> > > bool "Encryption support for userspace snapshots"
> > > depends on HIBERNATION_SNAPSHOT_DEV
> > > depends on CRYPTO_AEAD2=y
> > > + depends on TRUSTED_KEYS=y
> > > default n
> > > help
> > > Enable support for kernel-based encryption of hibernation snapshots
> > > diff --git a/kernel/power/snapenc.c b/kernel/power/snapenc.c
> > > index cb90692d6ab83a..90079f6d4f184b 100644
> > > --- a/kernel/power/snapenc.c
> > > +++ b/kernel/power/snapenc.c
> > > @@ -4,13 +4,23 @@
> > > #include <linux/crypto.h>
> > > #include <crypto/aead.h>
> > > #include <crypto/gcm.h>
> > > +#include <keys/trusted-type.h>
> > > +#include <linux/key-type.h>
> > > #include <linux/random.h>
> > > #include <linux/mm.h>
> > > +#include <linux/tpm.h>
> > > #include <linux/uaccess.h>
> > >
> > > #include "power.h"
> > > #include "user.h"
> > >
> > > +/* sha256("To sleep, perchance to dream") */
> > > +static struct tpm_digest known_digest = { .alg_id = TPM_ALG_SHA256,
> > > + .digest = {0x92, 0x78, 0x3d, 0x79, 0x2d, 0x00, 0x31, 0xb0, 0x55, 0xf9,
> > > + 0x1e, 0x0d, 0xce, 0x83, 0xde, 0x1d, 0xc4, 0xc5, 0x8e, 0x8c,
> > > + 0xf1, 0x22, 0x38, 0x6c, 0x33, 0xb1, 0x14, 0xb7, 0xec, 0x05,
> > > + 0x5f, 0x49}};
> > > +
> > > /* Encrypt more data from the snapshot into the staging area. */
> > > static int snapshot_encrypt_refill(struct snapshot_data *data)
> > > {
> > > @@ -313,6 +323,12 @@ void snapshot_teardown_encryption(struct snapshot_data *data)
> > > {
> > > int i;
> > >
> > > + if (data->key) {
> >
> > Would be a helpful to have perhaps inline comment before the check.
> >
> > Just stating this because I did not exactly follow why the null
> > check was needed (but do believe that there are good reasons to
> > do it).
> >
> > > + key_revoke(data->key);
> > > + key_put(data->key);
> > > + data->key = NULL;
> > > + }
> > > +
> > > if (data->aead_req) {
> > > aead_request_free(data->aead_req);
> > > data->aead_req = NULL;
> > > @@ -381,11 +397,83 @@ static int snapshot_setup_encryption_common(struct snapshot_data *data)
> > > return rc;
> > > }
> > >
> > > +static int snapshot_create_kernel_key(struct snapshot_data *data)
> > > +{
> > > + const struct cred *cred = current_cred();
> > > + struct tpm_digest *digests = NULL;
> > > + struct tpm_chip *chip;
> > > + struct key *key = NULL;
> > > + int ret, i;
> > > + /* Create a key sealed by the SRK. */
> > > + char *keyinfo = "new\t32\tkeyhandle=0x81000000";
> >
> > Again, I'd consider put this declaration as first.
> >
> > > +
> > > + chip = tpm_default_chip();
> > > + if (!chip)
> > > + return -ENODEV;
> > > +
> > > + if (!(tpm_is_tpm2(chip))) {
> > > + ret = -ENODEV;
> > > + goto out_dev;
> > > + }
> > > +
> > > + ret = tpm_pcr_reset(chip, 23);
> > > + if (ret)
> > > + goto out;
> > > +
> > > + digests = kcalloc(chip->nr_allocated_banks, sizeof(struct tpm_digest),
> > > + GFP_KERNEL);
> > > + if (!digests) {
> > > + ret = -ENOMEM;
> > > + goto out;
> > > + }
> > > +
> > > + for (i = 0; i <= chip->nr_allocated_banks; i++) {
> > > + digests[i].alg_id = chip->allocated_banks[i].alg_id;
> > > + if (digests[i].alg_id == known_digest.alg_id)
> > > + memcpy(&digests[i], &known_digest, sizeof(known_digest));
> > > + }
> > > +
> > > + ret = tpm_pcr_extend(chip, 23, digests);
> > > + if (ret != 0)
> > > + goto out;
> > > +
> > > + key = key_alloc(&key_type_trusted, "swsusp", GLOBAL_ROOT_UID,
> > > + GLOBAL_ROOT_GID, cred, 0, KEY_ALLOC_NOT_IN_QUOTA,
> > > + NULL);
> > > +
> > > + if (IS_ERR(key)) {
> > > + ret = PTR_ERR(key);
> > > + key = NULL;
> > > + goto out;
> > > + }
> > > +
> > > + ret = key_instantiate_and_link(key, keyinfo, strlen(keyinfo) + 1, NULL,
> > > + NULL);
> >
> > Generally speaking, even if it somehow would be "safe", not strlen()
> > thank you.
> >
> > AFAIK, keyinfo is a constant so you could just as well use sizeof().
> > And then you would not need "+ 1".
>
> Ack, I'm changing this one to sizeof(keyinfo), but...
>
> >
> > > + if (ret != 0)
> > > + goto out;
> > > +
> > > + data->key = key;
> > > + key = NULL;
> > > +
> > > +out:
> > > + if (key) {
> > > + key_revoke(key);
> > > + key_put(key);
> > > + }
> > > +
> > > + kfree(digests);
> > > + tpm_pcr_reset(chip, 23);
> > > +
> > > +out_dev:
> > > + put_device(&chip->dev);
> > > + return ret;
> > > +}
> > > +
> > > int snapshot_get_encryption_key(struct snapshot_data *data,
> > > struct uswsusp_key_blob __user *key)
> > > {
> > > - u8 aead_key[SNAPSHOT_ENCRYPTION_KEY_SIZE];
> > > u8 nonce[USWSUSP_KEY_NONCE_SIZE];
> > > + struct trusted_key_payload *payload;
> > > int rc;
> > > /* Don't pull a random key from a world that can be reset. */
> > > if (data->ready)
> > > @@ -399,21 +487,28 @@ int snapshot_get_encryption_key(struct snapshot_data *data,
> > > get_random_bytes(nonce, sizeof(nonce));
> > > memcpy(&data->nonce_low, &nonce[0], sizeof(data->nonce_low));
> > > memcpy(&data->nonce_high, &nonce[8], sizeof(data->nonce_high));
> > > - /* Build a random key */
> > > - get_random_bytes(aead_key, sizeof(aead_key));
> > > - rc = crypto_aead_setkey(data->aead_tfm, aead_key, sizeof(aead_key));
> > > +
> > > + /* Create a kernel key, and set it. */
> > > + rc = snapshot_create_kernel_key(data);
> > > + if (rc)
> > > + goto fail;
> > > +
> > > + payload = data->key->payload.data[0];
> > > + /* Install the key */
> > > + rc = crypto_aead_setkey(data->aead_tfm, payload->key, SNAPSHOT_ENCRYPTION_KEY_SIZE);
> > > if (rc)
> > > goto fail;
> > >
> > > - /* Hand the key back to user mode (to be changed!) */
> > > - rc = put_user(sizeof(struct uswsusp_key_blob), &key->blob_len);
> > > + /* Hand the key back to user mode in sealed form. */
> > > + rc = put_user(payload->blob_len, &key->blob_len);
> > > if (rc)
> > > goto fail;
> > >
> > > - rc = copy_to_user(&key->blob, &aead_key, sizeof(aead_key));
> > > + rc = copy_to_user(&key->blob, &payload->blob, payload->blob_len);
> > > if (rc)
> > > goto fail;
> > >
> > > + /* The nonce just gets handed back in the clear. */
> > > rc = copy_to_user(&key->nonce, &nonce, sizeof(nonce));
> > > if (rc)
> > > goto fail;
> > > @@ -425,10 +520,99 @@ int snapshot_get_encryption_key(struct snapshot_data *data,
> > > return rc;
> > > }
> > >
> > > +static int snapshot_load_kernel_key(struct snapshot_data *data,
> > > + struct uswsusp_key_blob *blob)
> >
> > Bad alignment.
> >
> > > +{
> > > +
> > > + const struct cred *cred = current_cred();
> > > + char *keytemplate = "load\t%s\tkeyhandle=0x81000000";
> >
> > Ditto.
> >
> > > + struct tpm_digest *digests = NULL;
> > > + char *blobstring = NULL;
> > > + char *keyinfo = NULL;
> > > + struct tpm_chip *chip;
> > > + struct key *key = NULL;
> > > + int i, ret;
> > > +
> > > + chip = tpm_default_chip();
> > > + if (!chip)
> > > + return -ENODEV;
> > > +
> > > + if (!(tpm_is_tpm2(chip))) {
> > > + ret = -ENODEV;
> > > + goto out_dev;
> > > + }
> > > +
> > > + ret = tpm_pcr_reset(chip, 23);
> > > + if (ret)
> > > + goto out;
> > > +
> > > + digests = kcalloc(chip->nr_allocated_banks, sizeof(struct tpm_digest),
> > > + GFP_KERNEL);
> > > + if (!digests)
> > > + goto out;
> > > +
> > > + for (i = 0; i <= chip->nr_allocated_banks; i++) {
> > > + digests[i].alg_id = chip->allocated_banks[i].alg_id;
> > > + if (digests[i].alg_id == known_digest.alg_id)
> > > + memcpy(&digests[i], &known_digest, sizeof(known_digest));
> > > + }
> > > +
> > > + ret = tpm_pcr_extend(chip, 23, digests);
> > > + if (ret != 0)
> > > + goto out;
> > > +
> > > + blobstring = kmalloc(blob->blob_len * 2, GFP_KERNEL);
> > > + if (!blobstring) {
> > > + ret = -ENOMEM;
> > > + goto out;
> > > + }
> > > +
> > > + bin2hex(blobstring, blob->blob, blob->blob_len);
> > > + keyinfo = kasprintf(GFP_KERNEL, keytemplate, blobstring);
> > > + if (!keyinfo) {
> > > + ret = -ENOMEM;
> > > + goto out;
> > > + }
> > > +
> > > + key = key_alloc(&key_type_trusted, "swsusp", GLOBAL_ROOT_UID,
> > > + GLOBAL_ROOT_GID, cred, 0, KEY_ALLOC_NOT_IN_QUOTA,
> > > + NULL);
> > > +
> > > + if (IS_ERR(key)) {
> > > + ret = PTR_ERR(key);
> > > + key = NULL;
> > > + goto out;
> > > + }
> > > +
> > > + ret = key_instantiate_and_link(key, keyinfo, strlen(keyinfo) + 1, NULL,
> > > + NULL);
> >
> > Ditto.
>
> ... I can't change this one to sizeof. Since this came out of
> kasprintf() and we already checked against null, strlen() seemed safe
> here. Is there a different pattern I should be following?

You're right. Let's strlen() here givent that as long as kasprintf()
is working correctly there's no risks involved.

BR, Jarkko