On Fri, May 03, 2019 at 04:23:51PM +0200, Ard Biesheuvel wrote:

On Fri, 3 May 2019 at 10:59, joeyli [off-list ref] wrote:

quoted

On Fri, May 03, 2019 at 10:07:59AM +0200, Ard Biesheuvel wrote:

quoted

On Fri, 3 May 2019 at 09:18, joeyli [off-list ref] wrote:

quoted

Hi Ard,

On Thu, May 02, 2019 at 11:04:34AM +0200, Ard Biesheuvel wrote:

quoted

On Thu, 2 May 2019 at 06:04, Lee, Chun-Yi [off-list ref] wrote:

quoted

When loading certificates list from UEFI variable, the original error
message direct shows the efi status code from UEFI firmware. It looks
ugly:

[    2.335031] Couldn't get size: 0x800000000000000e
[    2.335032] Couldn't get UEFI MokListRT
[    2.339985] Couldn't get size: 0x800000000000000e
[    2.339987] Couldn't get UEFI dbx list

So, this patch shows the status string instead of status code.

On the other hand, the "Couldn't get UEFI" message doesn't need
to be exposed when db/dbx/mok variable do not exist. So, this
patch set the message level to debug.

v3.
- Print messages similar to db/mok when loading dbx hash to blacklist:
[    1.500952] EFI: Blacklisting hash of an executable: UEFI:dbx
[    1.501773] blacklist: Loaded blacklisting hash
'bin:80b4d96931bf0d02fd91a61e19d14f1da452e66db2408ca8604d411f92659f0a'

- Setting messages for the existence of db/mok/dbx lists to debug level.

v2.
Setting the MODSIGN messages level to debug.

Link:
https://forums.opensuse.org/showthread.php/535324-MODSIGN-Couldn-t-get-UEFI-db-list?p=2897516#post2897516
Cc: James Morris <jmorris@namei.org>
Cc: Serge E. Hallyn" <serge@hallyn.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Nayna Jain <nayna@linux.ibm.com>
Cc: Josh Boyer <redacted>
Cc: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: "Lee, Chun-Yi" <jlee@suse.com>
---
 certs/blacklist.c                             |  3 +-
 security/integrity/platform_certs/load_uefi.c | 40 +++++++++++++++++++--------
 2 files changed, 31 insertions(+), 12 deletions(-)

diff --git a/certs/blacklist.c b/certs/blacklist.c
index 3a507b9e2568..f91437e39e44 100644
--- a/certs/blacklist.c
+++ b/certs/blacklist.c

@@ -100,7 +100,8 @@ int mark_hash_blacklisted(const char *hash)
        if (IS_ERR(key)) {
                pr_err("Problem blacklisting hash (%ld)\n", PTR_ERR(key));
                return PTR_ERR(key);
-       }
+       } else
+               pr_notice("Loaded blacklisting hash '%s'\n", hash);
        return 0;
 }

diff --git a/security/integrity/platform_certs/load_uefi.c b/security/integrity/platform_certs/load_uefi.c
index 81b19c52832b..6b6996e5bc27 100644
--- a/security/integrity/platform_certs/load_uefi.c
+++ b/security/integrity/platform_certs/load_uefi.c

@@ -1,5 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0

+#define pr_fmt(fmt) "EFI: "fmt
+
 #include <linux/kernel.h>
 #include <linux/sched.h>
 #include <linux/cred.h>

@@ -35,6 +37,18 @@ static __init bool uefi_check_ignore_db(void)
        return status == EFI_SUCCESS;
 }

+static void str16_to_str(efi_char16_t *str16, char *str, int str_size)
+{
+       int i = 0;
+
+       while (str16[i] != '\0' && i < (str_size - 1)) {
+               str[i] = str16[i];
+               i++;
+       }
+
+       str[i] = '\0';
+}
+
 /*
  * Get a certificate list blob from the named EFI variable.
  */

@@ -44,13 +58,20 @@ static __init void *get_cert_list(efi_char16_t *name, efi_guid_t *guid,
        efi_status_t status;
        unsigned long lsize = 4;
        unsigned long tmpdb[4];
+       char namestr[16];
        void *db;

+       str16_to_str(name, namestr, ARRAY_SIZE(namestr));

Please drop this (and the function above) - instead, just return NULL
if the variable is not found (without reporting an error).

This name string is for printing debug level message, not error message.
This function already returns NULL when EFI_NOT_FOUND be returned by
firmware.

quoted

quoted

        status = efi.get_variable(name, guid, NULL, &lsize, &tmpdb);
        if (status != EFI_BUFFER_TOO_SMALL) {
-               pr_err("Couldn't get size: 0x%lx\n", status);
+               if (status == EFI_NOT_FOUND)
+                       pr_debug("UEFI %s list doesn't exist\n", namestr);
+               else
+                       pr_err("Couldn't get size for UEFI %s list: %s\n",
+                               namestr, efi_status_to_str(status));
                return NULL;

here returns NULL when EFI_NOT_FOUND. The message of existence is for
debugging.

I understand that. But I don't think we need it.

OK. I will remove the debug message.

quoted

quoted

quoted

quoted

        }
+       pr_debug("UEFI %s list exists\n", namestr);

        db = kmalloc(lsize, GFP_KERNEL);
        if (!db)

@@ -59,7 +80,8 @@ static __init void *get_cert_list(efi_char16_t *name, efi_guid_t *guid,
        status = efi.get_variable(name, guid, NULL, &lsize, db);
        if (status != EFI_SUCCESS) {
                kfree(db);
-               pr_err("Error reading db var: 0x%lx\n", status);
+               pr_err("Error reading UEFI %s list: %s\n",
+                       namestr, efi_status_to_str(status));
                return NULL;
        }

@@ -95,6 +117,7 @@ static __init void uefi_blacklist_hash(const char *source, const void *data,
 static __init void uefi_blacklist_x509_tbs(const char *source,
                                           const void *data, size_t len)
 {
+       pr_info("Blacklisting X.509 TBS hash: %s\n", source);
        uefi_blacklist_hash(source, data, len, "tbs:", 4);
 }

@@ -104,6 +127,7 @@ static __init void uefi_blacklist_x509_tbs(const char *source,
 static __init void uefi_blacklist_binary(const char *source,
                                         const void *data, size_t len)
 {
+       pr_info("Blacklisting hash of an executable: %s\n", source);
        uefi_blacklist_hash(source, data, len, "bin:", 4);
 }

These are separate changes - I don't have an opinion whether they are
appropriate or not, but they should be in a separate patch.

I will move the message of blacklising hash to other patch. Thanks!

quoted

quoted

@@ -154,9 +178,7 @@ static int __init load_uefi_certs(void)
         */
        if (!uefi_check_ignore_db()) {
                db = get_cert_list(L"db", &secure_var, &dbsize);
-               if (!db) {
-                       pr_err("MODSIGN: Couldn't get UEFI db list\n");
-               } else {
+               if (db) {
                        rc = parse_efi_signature_list("UEFI:db",
                                        db, dbsize, get_handler_for_db);
                        if (rc)

@@ -167,9 +189,7 @@ static int __init load_uefi_certs(void)
        }

        mok = get_cert_list(L"MokListRT", &mok_var, &moksize);
-       if (!mok) {
-               pr_info("Couldn't get UEFI MokListRT\n");
-       } else {
+       if (mok) {
                rc = parse_efi_signature_list("UEFI:MokListRT",
                                              mok, moksize, get_handler_for_db);
                if (rc)

@@ -178,9 +198,7 @@ static int __init load_uefi_certs(void)
        }

        dbx = get_cert_list(L"dbx", &secure_var, &dbxsize);
-       if (!dbx) {
-               pr_info("Couldn't get UEFI dbx list\n");
-       } else {
+       if (dbx) {
                rc = parse_efi_signature_list("UEFI:dbx",
                                              dbx, dbxsize,
                                              get_handler_for_dbx);
--

2.16.4

I think we should consider carefully what it means if some of these
variables don't exist:
- if secure boot is enabled, db and dbx must exist, so if they don't,
something is wrong

The existence of db/dbx is not related to secure boot. If manufacturer/user
enrolled certificate/hash to db or dbx, then the variable will be created.
If user didn't enroll anything to db/dbx, then variables will not show up.

Yes, but if secure boot is enabled and db is empty, how could you have
booted in the first place?

I agree. When secure boot enabled, kernel can not be booted without db.

quoted

And what about the converse case: if secure boot is not enabled, why
should we trust the contents of db?

The db and dbx are authenticated variables that it protected by KEK.
So it can be trusted even secure boot is disabled. Unless manufacturer
or user's KEK is leaked.

Is that true for non-secureboot capable firmware? Of course, in that
case, we cannot be sure that the kernel itself is trusted, but we
shouldn't open up another hole either.

I agree with you that we should filer-out non-secureboot firmware. I
think that we can check the existence of SecureBoot variable before loading
db/dbx. On the other hand, the SetupMode should also be checked because
db/dbx can be changed in setup mode. 

quoted

quoted

quoted

quoted

- secure boot might be enabled but we may be booting without shim.

Shim always creates MokListRT no matter secure boot enabled or disabled.

That is not my point. What happens if you booted with secure boot
enabled but without the help of shim?

Without shim, the signed EFI stub can still be booted by EFI boot manager.
But the MokListRT will not be created for runtime. So MOK signed kernel
module can not be verified. (or IMA can not verify MOK signed kernel image
for kexec...)

So if we opt out of using shim, we lose kexec capability as well? That
doesn't make any sense.

MOK is just one of options that it can be used to sign crash kernel image.
We can also use the key in db or the kernel embedded keys.
 

In general, I am very uncomfortable with the way we have painted
ourselves into a corner with shim. If I install Suse on a system that
has the Suse cert in db or KEK, why would I need shim in the first
place?

Yes! We don't need MOK if distro's certificates are preloaded in most
machines' db. Unfortunately it's not the case on most x86 machines.
 

quoted

quoted

quoted

quoted

- secure boot might be disabled.

It's not about secure boot, db/dbx/MokListRT are always available at
runtime if user was enrolled something to those list.

Yes, but again, depending on whether shim was involved, and/or whether
secure boot was enabled or not, the way we interpret these things may
be very different.

I want the reasoning to be sound before merging any patches that deal
with these variables.

Here is a simple summary:

When secure boot is enabled:
    - db/dbx: Can be trusted because they are authenticated variables.
              (unless end user doesn't want to trust db/dbx)
    - MokListRT:
        - with shim: MokListRT will be created. It can be trusted.
        - without shim: MokListRT will not be created.
                        MOK protected kernel module or file can not be
                        verified.

So who polices that MoKlistRT is not created?

At boot time, shim copies MokList (boot variable) to MokListRT (runtime
variable). The MokListRT is a read-only runtime volatile variable. So,
without shim, the MokListRT will not be created for runtime. 
 

quoted

When secure boot is disabled:
    - db/dbx: Can be trusted because they are authenticated variables.
              (unless end user doesn't want to trust db/dbx)

Is this true? Does non-secure boot capable firmware still guarantee
that db/dbx can't be created?

I agree. So we should check the existence of SecureBoot variable as I
mentioned above. 
 

quoted

    - MokListRT:
        - with shim: MokListRT will be created. But it can not be trusted.
                     MOK protected kernel module or file can not be
                     verified.
        - without shim: MokListRT will not be created.
                        MOK protected kernel module or file can not be
                        verified.

Enabling this securely involves more than connecting the dots when it
comes to db, MokListRT etc. I would like for someone to convince me as
a maintainer, as well as the IMA maintainers, that the security
requirements are still met in all cases we care about. Just copying db
and MokListRT into a kernel keychain doesn't appear to do that, and if
it does, it needs more documentation to clarify that.

Welcome for more input from other maintainers. But I hope that my cosmetic
can be accepted because current warning messages looks ugly. 

Thanks a lot!
Joey Lee