Re: [PATCH RFC v11 14/19] ipe: add support for dm-verity as a trust provider

From: Fan Wu
Date: Thu Nov 02 2023 - 18:40:54 EST

Next message: syzbot: "[syzbot] [btrfs?] KASAN: slab-use-after-free Read in btrfs_qgroup_account_extent"
Previous message: Nhat Pham: "Re: [RFC PATCH v2] zswap: memcontrol: implement zswap writeback disabling"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On 10/23/2023 8:52 PM, Paul Moore wrote:

On Oct 4, 2023 Fan Wu <wufan@xxxxxxxxxxxxxxxxxxx> wrote:

Allows author of IPE policy to indicate trust for a singular dm-verity
volume, identified by roothash, through "dmverity_roothash" and all
signed dm-verity volumes, through "dmverity_signature".

Signed-off-by: Deven Bowers <deven.desai@xxxxxxxxxxxxxxxxxxx>
Signed-off-by: Fan Wu <wufan@xxxxxxxxxxxxxxxxxxx>
---

...

security/ipe/Kconfig | 18 +++++
security/ipe/Makefile | 1 +
security/ipe/audit.c | 31 +++++++-
security/ipe/digest.c | 142 +++++++++++++++++++++++++++++++++++
security/ipe/digest.h | 26 +++++++
security/ipe/eval.c | 101 ++++++++++++++++++++++++-
security/ipe/eval.h | 13 ++++
security/ipe/hooks.c | 51 +++++++++++++
security/ipe/hooks.h | 8 ++
security/ipe/ipe.c | 15 ++++
security/ipe/ipe.h | 4 +
security/ipe/policy.h | 3 +
security/ipe/policy_parser.c | 24 +++++-
13 files changed, 433 insertions(+), 4 deletions(-)
create mode 100644 security/ipe/digest.c
create mode 100644 security/ipe/digest.h

...

diff --git a/security/ipe/audit.c b/security/ipe/audit.c
index 0dd5f10c318f..b5c58655ac74 100644
--- a/security/ipe/audit.c
+++ b/security/ipe/audit.c
@@ -13,6 +13,7 @@
#include "hooks.h"
#include "policy.h"
#include "audit.h"
+#include "digest.h"
#define ACTSTR(x) ((x) == IPE_ACTION_ALLOW ? "ALLOW" : "DENY")
@@ -40,8 +41,29 @@ static const char *const audit_op_names[__IPE_OP_MAX] = {
static const char *const audit_prop_names[__IPE_PROP_MAX] = {
"boot_verified=FALSE",
"boot_verified=TRUE",
+#ifdef CONFIG_IPE_PROP_DM_VERITY
+ "dmverity_roothash=",
+ "dmverity_signature=FALSE",
+ "dmverity_signature=TRUE",
+#endif /* CONFIG_IPE_PROP_DM_VERITY */
};
+#ifdef CONFIG_IPE_PROP_DM_VERITY
+/**
+ * audit_dmv_roothash - audit a roothash of a dmverity volume.
+ * @ab: Supplies a pointer to the audit_buffer to append to.
+ * @r: Supplies a pointer to the digest structure.
+ */
+static void audit_dmv_roothash(struct audit_buffer *ab, const void *rh)
+{
+ ipe_digest_audit(ab, rh);
+}
+#else
+static void audit_dmv_roothash(struct audit_buffer *ab, const void *rh)
+{
+}
+#endif /* CONFIG_IPE_PROP_DM_VERITY */

Since the "dmverity_roothash=" field name is going to be written to
the audit record regardless of CONFIG_IPE_PROP_DM_VERITY we should
ensure that it has a value of "?" instead of nothing. To fix that
I would suggest something like this:

#else
static void audit_dmv_roothash(struct audit_buffer *ab, const void *rh)
{
audit_log_format(ab, "?");
}
#endif /* CONFIG_IPE_PROP_DM_VERITY */

/**
* audit_rule - audit an IPE policy rule approximation.
* @ab: Supplies a pointer to the audit_buffer to append to.
@@ -53,8 +75,13 @@ static void audit_rule(struct audit_buffer *ab, const struct ipe_rule *r)
audit_log_format(ab, "rule=\"op=%s ", audit_op_names[r->op]);
- list_for_each_entry(ptr, &r->props, next)
- audit_log_format(ab, "%s ", audit_prop_names[ptr->type]);
+ list_for_each_entry(ptr, &r->props, next) {
+ audit_log_format(ab, "%s", audit_prop_names[ptr->type]);
+ if (ptr->type == IPE_PROP_DMV_ROOTHASH)
+ audit_dmv_roothash(ab, ptr->value);

If you wanted to avoid the roothash change above, you could change the
code here so it didn't always write "dmverity_roothash=" into the audit
record.

Yes this sounds more reasonable to me, I will change this part to a switch case statement to avoid unnessary logging.

+ audit_log_format(ab, " ");
+ }
audit_log_format(ab, "action=%s\"", ACTSTR(r->action));
}
diff --git a/security/ipe/digest.c b/security/ipe/digest.c
new file mode 100644
index 000000000000..7a42ca71880c
--- /dev/null
+++ b/security/ipe/digest.c
@@ -0,0 +1,142 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) Microsoft Corporation. All rights reserved.
+ */
+
+#include "digest.h"
+
+/**
+ * ipe_digest_parse - parse a digest in IPE's policy.
+ * @valstr: Supplies the string parsed from the policy.
+ * @value: Supplies a pointer to be populated with the result.
+ *
+ * Digests in IPE are defined in a standard way:
+ * <alg_name>:<hex>
+ *
+ * Use this function to create a property to parse the digest
+ * consistently. The parsed digest will be saved in @value in IPE's
+ * policy.
+ *
+ * Return:
+ * * 0 - OK
+ * * !0 - Error
+ */
+int ipe_digest_parse(const char *valstr, void **value)

Why is @value void? You should make it a digest_info type or simply
skip the second parameter and return a digest_info pointer.

Yes it should better just return a digest_info type pointer. I will update this part.

+{
+ char *sep, *raw_digest;
+ size_t raw_digest_len;
+ int rc = 0;
+ u8 *digest = NULL;
+ struct digest_info *info = NULL;
+
+ info = kzalloc(sizeof(*info), GFP_KERNEL);
+ if (!info)
+ return -ENOMEM;
+
+ sep = strchr(valstr, ':');
+ if (!sep) {
+ rc = -EBADMSG;
+ goto err;
+ }
+
+ info->alg = kstrndup(valstr, sep - valstr, GFP_KERNEL);
+ if (!info->alg) {
+ rc = -ENOMEM;
+ goto err;
+ }
+
+ raw_digest = sep + 1;
+ raw_digest_len = strlen(raw_digest);
+ info->raw_digest = kstrndup(raw_digest, raw_digest_len, GFP_KERNEL);

Since you're running a strlen() over @raw_digest, you might as well
just use kstrdup() instead of kstrndup().

Thanks I will update this part.

+ if (!info->raw_digest) {
+ rc = -ENOMEM;
+ goto err_free_alg;
+ }
+
+ info->digest_len = (raw_digest_len + 1) / 2;
+ digest = kzalloc(info->digest_len, GFP_KERNEL);
+ if (!digest) {
+ rc = -ENOMEM;
+ goto err_free_raw;
+ }
+
+ rc = hex2bin(digest, raw_digest, info->digest_len);
+ if (rc < 0) {
+ rc = -EINVAL;
+ goto err_free_raw;
+ }
+
+ info->digest = digest;
+ *value = info;
+ return 0;
+
+err_free_raw:
+ kfree(info->raw_digest);
+err_free_alg:
+ kfree(info->alg);
+err:
+ kfree(digest);
+ kfree(info);
+ return rc;
+}
+
+/**
+ * ipe_digest_eval - evaluate an IPE digest against another digest.
+ * @expect: Supplies the policy-provided digest value.
+ * @digest: Supplies the digest to compare against the policy digest value.
+ * @digest_len: The length of @digest.
+ * @alg: Supplies the name of the algorithm used to calculated @digest.
+ *
+ * Return:
+ * * true - digests match
+ * * false - digests do not match
+ */
+bool ipe_digest_eval(const void *expect, const u8 *digest, size_t digest_len,
+ const char *alg)

Similar to the above, why not make the @expect parameter a digest_info
type? Also, why not pass a second digest_info parameter instead of
a separate @digest, @digest_len, and @alg?

The digest_info was inteneded to be used for policy only, which also contains a raw string of digest in the policy that can be used during audit. After second thought I start to think the raw_digest seems unnecessary and I can refactor them all to use digest_info instead.

+{
+ const struct digest_info *info = (struct digest_info *)expect;
+
+ return (digest_len == info->digest_len) && !strcmp(alg, info->alg) &&
+ (!memcmp(info->digest, digest, info->digest_len));
+}
+
+/**
+ * ipe_digest_free - free an IPE digest.
+ * @value: Supplies a pointer the policy-provided digest value to free.
+ */
+void ipe_digest_free(void **value)

Another digest_info parameter/type issue.

+{
+ struct digest_info *info = (struct digest_info *)(*value);
+
+ if (IS_ERR_OR_NULL(info))
+ return;
+
+ kfree(info->alg);
+ kfree(info->raw_digest);
+ kfree(info->digest);
+ kfree(info);
+}
+
+/**
+ * ipe_digest_audit - audit a digest that was sourced from IPE's policy.
+ * @ab: Supplies the audit_buffer to append the formatted result.
+ * @val: Supplies a pointer to source the audit record from.
+ *
+ * Digests in IPE are defined in a standard way:
+ * <alg_name>:<hex>
+ *
+ * Use this function to create a property to audit the digest
+ * consistently.
+ *
+ * Return:
+ * 0 - OK
+ * !0 - Error
+ */
+void ipe_digest_audit(struct audit_buffer *ab, const void *val)

Another digest_info parameter/type issue.

+{
+ const struct digest_info *info = (struct digest_info *)val;
+
+ audit_log_untrustedstring(ab, info->alg);
+ audit_log_format(ab, ":");
+ audit_log_untrustedstring(ab, info->raw_digest);
+}

...

diff --git a/security/ipe/eval.c b/security/ipe/eval.c
index 78c54ff1fdd3..82ad48d7aa3d 100644
--- a/security/ipe/eval.c
+++ b/security/ipe/eval.c
@@ -69,15 +88,89 @@ void build_eval_ctx(struct ipe_eval_ctx *ctx,
const struct file *file,
enum ipe_op_type op)
{
+ struct inode *ino = NULL;
+
if (op == IPE_OP_EXEC && file)
pin_sb(FILE_SUPERBLOCK(file));
ctx->file = file;
ctx->op = op;
- if (file)
+ if (file) {
ctx->from_init_sb = from_pinned(FILE_SUPERBLOCK(file));
+ ino = d_real_inode(file->f_path.dentry);
+ build_ipe_bdev_ctx(ctx, ino);

You don't need @ino.

build_ipe_bdev_ctx(ctx, d_real_inode(file->f_path.dentry));

This variable will make sense in the fsverity patch, I do agree it doesn't make sense for this one. Will remove it.

+ }
+}
+
+#ifdef CONFIG_IPE_PROP_DM_VERITY
+/**
+ * evaluate_dmv_roothash - Evaluate @ctx against a dmv roothash property.
+ * @ctx: Supplies a pointer to the context being evaluated.
+ * @p: Supplies a pointer to the property being evaluated.
+ *
+ * Return:
+ * * true - The current @ctx match the @p
+ * * false - The current @ctx doesn't match the @p
+ */
+static bool evaluate_dmv_roothash(const struct ipe_eval_ctx *const ctx,
+ struct ipe_prop *p)
+{
+ return !!ctx->ipe_bdev &&
+ ipe_digest_eval(p->value,
+ ctx->ipe_bdev->digest,
+ ctx->ipe_bdev->digest_len,
+ ctx->ipe_bdev->digest_algo);

Building on my comments above in ipe_digest_eval(), if you convert it
to use digest_info structs this is simplified:

ipe_digest_eval(p->vlaue, ctx->ipe_bdev)

Yep but the seucrity blob also contains signature information. I will make it like ipe_digest_eval(p->vlaue, ctx->ipe_bdev->dmv_digest_info)

+}
+
+/**
+ * evaluate_dmv_sig_false: Analyze @ctx against a dmv sig false property.
+ * @ctx: Supplies a pointer to the context being evaluated.
+ * @p: Supplies a pointer to the property being evaluated.
+ *
+ * Return:
+ * * true - The current @ctx match the @p
+ * * false - The current @ctx doesn't match the @p
+ */
+static bool evaluate_dmv_sig_false(const struct ipe_eval_ctx *const ctx,
+ struct ipe_prop *p)
+{
+ return !ctx->ipe_bdev || (!ctx->ipe_bdev->dm_verity_signed);
+}
+
+/**
+ * evaluate_dmv_sig_true: Analyze @ctx against a dmv sig true property.
+ * @ctx: Supplies a pointer to the context being evaluated.
+ * @p: Supplies a pointer to the property being evaluated.
+ *
+ * Return:
+ * * true - The current @ctx match the @p
+ * * false - The current @ctx doesn't match the @p
+ */
+static bool evaluate_dmv_sig_true(const struct ipe_eval_ctx *const ctx,
+ struct ipe_prop *p)
+{
+ return ctx->ipe_bdev && (!!ctx->ipe_bdev->dm_verity_signed);
+}

Do you need both evaluate_dmv_sig_true() and evaluate_dmv_sig_false()?
If yes, you should make one call the other and return the inverse to
help ensure there are no oddities.

Yes this code is redundant. I also found the ipe_prop *p variable is necssary. Will clean them up in the next version.

diff --git a/security/ipe/hooks.c b/security/ipe/hooks.c
index e9386762a597..8b8031e66f36 100644
--- a/security/ipe/hooks.c
+++ b/security/ipe/hooks.c
@@ -193,3 +196,51 @@ void ipe_sb_free_security(struct super_block *mnt_sb)
{
ipe_invalidate_pinned_sb(mnt_sb);
}
+
+#ifdef CONFIG_IPE_PROP_DM_VERITY
+/**
+ * ipe_bdev_free_security - free IPE's LSM blob of block_devices.
+ * @bdev: Supplies a pointer to a block_device that contains the structure
+ * to free.
+ */
+void ipe_bdev_free_security(struct block_device *bdev)
+{
+ struct ipe_bdev *blob = ipe_bdev(bdev);
+
+ kfree(blob->digest);
+ kfree(blob->digest_algo);
+}
+
+/**
+ * ipe_bdev_setsecurity - save data from a bdev to IPE's LSM blob.
+ * @bdev: Supplies a pointer to a block_device that contains the LSM blob.
+ * @key: Supplies the string key that uniquely identifies the value.
+ * @value: Supplies the value to store.
+ * @len: The length of @value.
+ */
+int ipe_bdev_setsecurity(struct block_device *bdev, const char *key,
+ const void *value, size_t len)
+{
+ struct ipe_bdev *blob = ipe_bdev(bdev);
+
+ if (!strcmp(key, DM_VERITY_ROOTHASH_SEC_NAME)) {
+ const struct dm_verity_digest *digest = value;
+
+ blob->digest = kmemdup(digest->digest, digest->digest_len, GFP_KERNEL);
+ if (!blob->digest)
+ return -ENOMEM;
+
+ blob->digest_algo = kstrdup_const(digest->algo, GFP_KERNEL);
+ if (!blob->digest_algo)
+ return -ENOMEM;

You need to cleanup @blob on error so you don't leak ipe_bdev::digest.

Thanks for catching that. I will fix this.

-Fan

+ blob->digest_len = digest->digest_len;
+ return 0;
+ } else if (!strcmp(key, DM_VERITY_SIGNATURE_SEC_NAME)) {
+ blob->dm_verity_signed = true;
+ return 0;
+ }
+
+ return -EOPNOTSUPP;
+}
+#endif /* CONFIG_IPE_PROP_DM_VERITY */

--
paul-moore.com

Next message: syzbot: "[syzbot] [btrfs?] KASAN: slab-use-after-free Read in btrfs_qgroup_account_extent"
Previous message: Nhat Pham: "Re: [RFC PATCH v2] zswap: memcontrol: implement zswap writeback disabling"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]