linux-integrity.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jarkko Sakkinen <jarkko@kernel.org>
To: Evan Green <evgreen@chromium.org>
Cc: linux-kernel@vger.kernel.org, linux-integrity@vger.kernel.org,
	apronin@chromium.org, dlunev@google.com,
	Pavel Machek <pavel@ucw.cz>, Ben Boeckel <me@benboeckel.net>,
	rjw@rjwysocki.net, corbet@lwn.net, linux-pm@vger.kernel.org,
	zohar@linux.ibm.com, Kees Cook <keescook@chromium.org>,
	Eric Biggers <ebiggers@kernel.org>,
	jejb@linux.ibm.com, gwendal@chromium.org,
	Matthew Garrett <mgarrett@aurora.tech>,
	Len Brown <len.brown@intel.com>,
	"Rafael J. Wysocki" <rafael@kernel.org>
Subject: Re: [PATCH v3 08/11] PM: hibernate: Use TPM-backed keys to encrypt image
Date: Mon, 24 Oct 2022 00:55:07 +0300	[thread overview]
Message-ID: <Y1W4O2bQIQMyuT2+@kernel.org> (raw)
In-Reply-To: <CAE=gft45p4QFqe0E0X_1XGeRB2kLgH3p9ZfNNTvMk2H9GwbhMw@mail.gmail.com>

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@kernel.org> 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@google.com>
> > > Signed-off-by: Evan Green <evgreen@chromium.org>
> > >
> > > ---
> > > Matthew's incarnation of this patch is at:
> > > https://patchwork.kernel.org/project/linux-pm/patch/20210220013255.1083202-9-matthewgarrett@google.com/
> > >
> > > 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

  reply	other threads:[~2022-10-23 21:55 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-27 16:49 [PATCH v3 00/11] Encrypted Hibernation Evan Green
2022-09-27 16:49 ` [PATCH v3 01/11] tpm: Add support for in-kernel resetting of PCRs Evan Green
2022-09-30 20:52   ` Jarkko Sakkinen
2022-09-27 16:49 ` [PATCH v3 02/11] tpm: Export and rename tpm2_find_and_validate_cc() Evan Green
2022-09-27 16:49 ` [PATCH v3 03/11] tpm: Allow PCR 23 to be restricted to kernel-only use Evan Green
2022-09-30 20:57   ` Jarkko Sakkinen
2022-09-27 16:49 ` [PATCH v3 04/11] security: keys: trusted: Include TPM2 creation data Evan Green
2022-09-27 16:49 ` [PATCH v3 05/11] security: keys: trusted: Allow storage of PCR values in " Evan Green
2022-09-27 16:58   ` Ben Boeckel
2022-09-27 16:49 ` [PATCH v3 06/11] security: keys: trusted: Verify " Evan Green
2022-09-27 16:49 ` [PATCH v3 07/11] PM: hibernate: Add kernel-based encryption Evan Green
2022-09-30 21:30   ` Jarkko Sakkinen
2022-09-27 16:49 ` [PATCH v3 08/11] PM: hibernate: Use TPM-backed keys to encrypt image Evan Green
2022-09-30 21:35   ` Jarkko Sakkinen
2022-10-21 19:56     ` Evan Green
2022-10-23 21:55       ` Jarkko Sakkinen [this message]
2022-09-27 16:49 ` [PATCH v3 09/11] PM: hibernate: Mix user key in encrypted hibernate Evan Green
2022-09-27 16:49 ` [PATCH v3 10/11] PM: hibernate: Verify the digest encryption key Evan Green
2022-09-27 16:49 ` [PATCH v3 11/11] PM: hibernate: seal the encryption key with a PCR policy Evan Green

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Y1W4O2bQIQMyuT2+@kernel.org \
    --to=jarkko@kernel.org \
    --cc=apronin@chromium.org \
    --cc=corbet@lwn.net \
    --cc=dlunev@google.com \
    --cc=ebiggers@kernel.org \
    --cc=evgreen@chromium.org \
    --cc=gwendal@chromium.org \
    --cc=jejb@linux.ibm.com \
    --cc=keescook@chromium.org \
    --cc=len.brown@intel.com \
    --cc=linux-integrity@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pm@vger.kernel.org \
    --cc=me@benboeckel.net \
    --cc=mgarrett@aurora.tech \
    --cc=pavel@ucw.cz \
    --cc=rafael@kernel.org \
    --cc=rjw@rjwysocki.net \
    --cc=zohar@linux.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).