Add functionality for external signatures #2
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
erichte-ibm
reviewed
Jan 26, 2021
| @@ -816,7 +817,7 @@ int to_pkcs7_already_signed_data(unsigned char **pkcs7, size_t *pkcs7Size, const | |||
| if (sigs[i]) free(sigs[i]); | |||
| } | |||
| if (sigs) free (sigs); | |||
|
|
|||
| if (sig_sizes) free(sig_sizes); | |||
There was a problem hiding this comment.
Are these fixes to a bug/leak introduced in the first commit of this PR? If so you can just squash this commit into that one, no need to have a separate commit for it.
If this is fixing an unrelated bug, then nevermind.
erichte-ibm
reviewed
Jan 26, 2021
| @@ -49,13 +49,14 @@ extern int verbose; | |||
| typedef struct PKCS7Info { | |||
| unsigned char **crts; // signing crt DER | |||
| size_t *crtSizes; | |||
| unsigned char **keys; // signing key DER | |||
| unsigned char **keys; // signing key DER or signatures (depends on alreadySignedFlag) | |||
There was a problem hiding this comment.
So, I'm not fully up on the pkcs7 implementation for mbedtls, but make sure if you are adding features to the pkcs7 generation code for new features in secvarctl, that those changes make sense for mbedtls as well.
It may be suggested that rolling pre-signed data into a pkcs7 blob should be handled in glue code outside of the library (i.e. in secvarctl) instead of in the library itself. Not suggest you change this now, but more of a warning. Maybe look at other implementations of similar algos/specs/etc and see if there's precedence for this?
|
Looks good to me, I'm marking as approved, but will wait for an answer on the squash commit question before merging. |
erichte-ibm
approved these changes
Jan 26, 2021
…vate key
Previously, if a user wanted to make a signed auth file, they would need to supply secvarctl with public and private key pairs for signing the data. Recently, it has come to our attention that some users will not have access to their private keys as it may be under the guard of a signing framework server. In this case, the most secvarctl can do is to generate the hash to be signed. From their, it is up to the user to generate the signature with their signing framework. Then the user can give secvarctl the signature and the generation of PKCS7/auth files can resume. It is important the user uses the same timestamp in both the presignature generation and auth/PKCS7 generation. Now, with this commit, the user can use a signing framework to generate an auth file using the below commands:
- $ secvarctl generate c:x -n <varName> -t <timeStamp> -i <certificate> -o <sha256.hash>
- $ <user sends hash to be signed -in <sha256.hash> -out <signature.bin>
- $ secvarctl generate c:a -n <varName> <timeStamp> -i <certificate> -o <newAuth> -c <signingCert> -s <signature.bin>
In the previous commit, abilities to create auth and PKCS7 files with a raw signed data files in replacement of private key arguments was introduced. This commit adds those changes to the documentation and help messages
In the two previus commits, documentation and functionality were added to allow for secvarctl to accept an external signature rather than generating the signature internally. Now, we add tests to ensure that the enhanement was executed appropriately. To test we generate a signed auth file to work as our control. Then we generate an unsigned hash digest and sign it externally with openssl. We then use the signature to generate another signed auth file that will be our experiment. If the control and experiment contain the exact same data then the test passes.
nick-child-ibm
added a commit
to nick-child-ibm/secvarctl
that referenced
this pull request
Oct 5, 2023
There are several things wrong with get_hash_data:
1. it doesn't work, all returned hashes are all 0x00
2. memory leak of hash allocated by crypto_md_generate_hash
3. memory leak of der data buffer from is_x509certificate
4. useless casting of hash_ids to x509_hash_ids
Issues 1-3 can be seen below
$ bin/secvarctl-dbg -m guest generate f:e -i ./test/testdata/guest/x509certs/dbx.crt -o dbx.esl -n dbx
error: invalid crypto alg key 8
error: invalid crypto alg key 11
error: invalid crypto alg key 8
RESULT: SUCCESS
=================================================================
==88342==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 813 byte(s) in 1 object(s) allocated from:
#0 0x7fa153bdebb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8)
#1 0x7fa1537a42da in PEM_read_bio_ex (/lib64/libcrypto.so.1.1+0x19f2da)
Direct leak of 32 byte(s) in 1 object(s) allocated from:
#0 0x7fa153bdebb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8)
#1 0x463e27 in crypto_md_generate_hash src/crypto_openssl.c:768
open-power#2 0x445b4a in get_hash_data backends/guest/common/generate.c:477
open-power#3 0x43234a in generate_esl backends/guest/guest_svc_generate.c:73
open-power#4 0x43505e in generate_data backends/guest/guest_svc_generate.c:365
open-power#5 0x439382 in guest_generate_command backends/guest/guest_svc_generate.c:690
open-power#6 0x4069b4 in main /home/nchild/IBM/secvarctl-stuff/erichte-guest-devel/secvarctl/secvarctl.c:249
open-power#7 0x7fa152558d84 in __libc_start_main (/lib64/libc.so.6+0x3ad84)
SUMMARY: AddressSanitizer: 845 byte(s) leaked in 2 allocation(s).
$ hexdump dbx.esl
0000000 1626 c1c4 504c 4092 a9ac f941 9336 2843
0000010 004c 0000 0000 0000 0030 0000 0000 0000
0000020 0000 0000 0000 0000 0000 0000 0000 0000
*
0000040 0000 0000 0000 0000 0000 0000
000004c
Fixing issue 1 and 2:
get_hash_data() acts as a middle function during hash generation.
First generate_esl() allocates an empty hash buffer and hands it to
get_hash_data(). get_hash_data() then hands the address to that buffer
to crypto_md_generate_hash(). crypto_md_generate_hash() allocates its
own buffer and overwrites our pointer to our buffer. get_hash_data()
has no way of updating generate_esl() of the new address. This results
in a successful hash generation but a memoryleaks and an empty hash
in the esl:
To avoid this, simply copy the returned buffer into the existing buffer.
Fixing issue 3:
is_x509certificate will allocate and return data if it finds the given
buffer contains an x509.
To fix this, simply free the der data after generating the hash.
Fixing issue 4:
This is only a semi-issue, but previously, if the buffer was a
certificate then we would cast our hash_function parameter into an
x509 hash function ID. Later on this value get passed to
get_crypto_alg_id() to get the crypto library specific MD identifier.
get_crypto_alg_id() does not keep track of x509 hash function ID (and
tbh the difference between them is not clear) and defaults to sha256.
The point of converting the given hash id to x509 hash id only to then
convert that value into crypto library specific ID is unnecessary.
To fix just stick with the hash function passed to the function.
Lastly, add a test case for all this.
Signed-off-by: Nick Child <nick.child@ibm.com>
nick-child-ibm
added a commit
to nick-child-ibm/secvarctl
that referenced
this pull request
Oct 6, 2023
There are several things wrong with get_hash_data:
1. it doesn't work, all returned hashes are all 0x00
2. memory leak of hash allocated by crypto_md_generate_hash
3. memory leak of der data buffer from is_x509certificate
4. useless casting of hash_ids to x509_hash_ids
Issues 1-3 can be seen below
$ bin/secvarctl-dbg -m guest generate f:e -i ./test/testdata/guest/x509certs/dbx.crt -o dbx.esl -n dbx
error: invalid crypto alg key 8
error: invalid crypto alg key 11
error: invalid crypto alg key 8
RESULT: SUCCESS
=================================================================
==88342==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 813 byte(s) in 1 object(s) allocated from:
#0 0x7fa153bdebb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8)
#1 0x7fa1537a42da in PEM_read_bio_ex (/lib64/libcrypto.so.1.1+0x19f2da)
Direct leak of 32 byte(s) in 1 object(s) allocated from:
#0 0x7fa153bdebb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8)
#1 0x463e27 in crypto_md_generate_hash src/crypto_openssl.c:768
open-power#2 0x445b4a in get_hash_data backends/guest/common/generate.c:477
open-power#3 0x43234a in generate_esl backends/guest/guest_svc_generate.c:73
open-power#4 0x43505e in generate_data backends/guest/guest_svc_generate.c:365
open-power#5 0x439382 in guest_generate_command backends/guest/guest_svc_generate.c:690
open-power#6 0x4069b4 in main /home/nchild/IBM/secvarctl-stuff/erichte-guest-devel/secvarctl/secvarctl.c:249
open-power#7 0x7fa152558d84 in __libc_start_main (/lib64/libc.so.6+0x3ad84)
SUMMARY: AddressSanitizer: 845 byte(s) leaked in 2 allocation(s).
$ hexdump dbx.esl
0000000 1626 c1c4 504c 4092 a9ac f941 9336 2843
0000010 004c 0000 0000 0000 0030 0000 0000 0000
0000020 0000 0000 0000 0000 0000 0000 0000 0000
*
0000040 0000 0000 0000 0000 0000 0000
000004c
Fixing issue 1 and 2:
get_hash_data() acts as a middle function during hash generation.
First generate_esl() allocates an empty hash buffer and hands it to
get_hash_data(). get_hash_data() then hands the address to that buffer
to crypto_md_generate_hash(). crypto_md_generate_hash() allocates its
own buffer and overwrites our pointer to our buffer. get_hash_data()
has no way of updating generate_esl() of the new address. This results
in a successful hash generation but a memoryleaks and an empty hash
in the esl:
To avoid this, simply copy the returned buffer into the existing buffer.
Fixing issue 3:
is_x509certificate will allocate and return data if it finds the given
buffer contains an x509.
To fix this, simply free the der data after generating the hash.
Fixing issue 4:
This is only a semi-issue, but previously, if the buffer was a
certificate then we would cast our hash_function parameter into an
x509 hash function ID. Later on this value get passed to
get_crypto_alg_id() to get the crypto library specific MD identifier.
get_crypto_alg_id() does not keep track of x509 hash function ID (and
tbh the difference between them is not clear) and defaults to sha256.
The point of converting the given hash id to x509 hash id only to then
convert that value into crypto library specific ID is unnecessary.
To fix just stick with the hash function passed to the function.
Lastly, add a test case for all this.
Signed-off-by: Nick Child <nick.child@ibm.com>
erichte-ibm
pushed a commit
that referenced
this pull request
Oct 6, 2023
There are several things wrong with get_hash_data:
1. it doesn't work, all returned hashes are all 0x00
2. memory leak of hash allocated by crypto_md_generate_hash
3. memory leak of der data buffer from is_x509certificate
4. useless casting of hash_ids to x509_hash_ids
Issues 1-3 can be seen below
$ bin/secvarctl-dbg -m guest generate f:e -i ./test/testdata/guest/x509certs/dbx.crt -o dbx.esl -n dbx
error: invalid crypto alg key 8
error: invalid crypto alg key 11
error: invalid crypto alg key 8
RESULT: SUCCESS
=================================================================
==88342==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 813 byte(s) in 1 object(s) allocated from:
#0 0x7fa153bdebb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8)
#1 0x7fa1537a42da in PEM_read_bio_ex (/lib64/libcrypto.so.1.1+0x19f2da)
Direct leak of 32 byte(s) in 1 object(s) allocated from:
#0 0x7fa153bdebb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8)
#1 0x463e27 in crypto_md_generate_hash src/crypto_openssl.c:768
#2 0x445b4a in get_hash_data backends/guest/common/generate.c:477
#3 0x43234a in generate_esl backends/guest/guest_svc_generate.c:73
#4 0x43505e in generate_data backends/guest/guest_svc_generate.c:365
#5 0x439382 in guest_generate_command backends/guest/guest_svc_generate.c:690
#6 0x4069b4 in main /home/nchild/IBM/secvarctl-stuff/erichte-guest-devel/secvarctl/secvarctl.c:249
#7 0x7fa152558d84 in __libc_start_main (/lib64/libc.so.6+0x3ad84)
SUMMARY: AddressSanitizer: 845 byte(s) leaked in 2 allocation(s).
$ hexdump dbx.esl
0000000 1626 c1c4 504c 4092 a9ac f941 9336 2843
0000010 004c 0000 0000 0000 0030 0000 0000 0000
0000020 0000 0000 0000 0000 0000 0000 0000 0000
*
0000040 0000 0000 0000 0000 0000 0000
000004c
Fixing issue 1 and 2:
get_hash_data() acts as a middle function during hash generation.
First generate_esl() allocates an empty hash buffer and hands it to
get_hash_data(). get_hash_data() then hands the address to that buffer
to crypto_md_generate_hash(). crypto_md_generate_hash() allocates its
own buffer and overwrites our pointer to our buffer. get_hash_data()
has no way of updating generate_esl() of the new address. This results
in a successful hash generation but a memoryleaks and an empty hash
in the esl:
To avoid this, simply copy the returned buffer into the existing buffer.
Fixing issue 3:
is_x509certificate will allocate and return data if it finds the given
buffer contains an x509.
To fix this, simply free the der data after generating the hash.
Fixing issue 4:
This is only a semi-issue, but previously, if the buffer was a
certificate then we would cast our hash_function parameter into an
x509 hash function ID. Later on this value get passed to
get_crypto_alg_id() to get the crypto library specific MD identifier.
get_crypto_alg_id() does not keep track of x509 hash function ID (and
tbh the difference between them is not clear) and defaults to sha256.
The point of converting the given hash id to x509 hash id only to then
convert that value into crypto library specific ID is unnecessary.
To fix just stick with the hash function passed to the function.
Lastly, add a test case for all this.
Signed-off-by: Nick Child <nick.child@ibm.com>
nick-child-ibm
added a commit
that referenced
this pull request
Feb 6, 2024
There are several things wrong with get_hash_data:
1. it doesn't work, all returned hashes are all 0x00
2. memory leak of hash allocated by crypto_md_generate_hash
3. memory leak of der data buffer from is_x509certificate
4. useless casting of hash_ids to x509_hash_ids
Issues 1-3 can be seen below
$ bin/secvarctl-dbg -m guest generate f:e -i ./test/testdata/guest/x509certs/dbx.crt -o dbx.esl -n dbx
error: invalid crypto alg key 8
error: invalid crypto alg key 11
error: invalid crypto alg key 8
RESULT: SUCCESS
=================================================================
==88342==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 813 byte(s) in 1 object(s) allocated from:
#0 0x7fa153bdebb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8)
#1 0x7fa1537a42da in PEM_read_bio_ex (/lib64/libcrypto.so.1.1+0x19f2da)
Direct leak of 32 byte(s) in 1 object(s) allocated from:
#0 0x7fa153bdebb8 in __interceptor_malloc (/lib64/libasan.so.5+0xefbb8)
#1 0x463e27 in crypto_md_generate_hash src/crypto_openssl.c:768
#2 0x445b4a in get_hash_data backends/guest/common/generate.c:477
#3 0x43234a in generate_esl backends/guest/guest_svc_generate.c:73
#4 0x43505e in generate_data backends/guest/guest_svc_generate.c:365
#5 0x439382 in guest_generate_command backends/guest/guest_svc_generate.c:690
#6 0x4069b4 in main /home/nchild/IBM/secvarctl-stuff/erichte-guest-devel/secvarctl/secvarctl.c:249
#7 0x7fa152558d84 in __libc_start_main (/lib64/libc.so.6+0x3ad84)
SUMMARY: AddressSanitizer: 845 byte(s) leaked in 2 allocation(s).
$ hexdump dbx.esl
0000000 1626 c1c4 504c 4092 a9ac f941 9336 2843
0000010 004c 0000 0000 0000 0030 0000 0000 0000
0000020 0000 0000 0000 0000 0000 0000 0000 0000
*
0000040 0000 0000 0000 0000 0000 0000
000004c
Fixing issue 1 and 2:
get_hash_data() acts as a middle function during hash generation.
First generate_esl() allocates an empty hash buffer and hands it to
get_hash_data(). get_hash_data() then hands the address to that buffer
to crypto_md_generate_hash(). crypto_md_generate_hash() allocates its
own buffer and overwrites our pointer to our buffer. get_hash_data()
has no way of updating generate_esl() of the new address. This results
in a successful hash generation but a memoryleaks and an empty hash
in the esl:
To avoid this, simply copy the returned buffer into the existing buffer.
Fixing issue 3:
is_x509certificate will allocate and return data if it finds the given
buffer contains an x509.
To fix this, simply free the der data after generating the hash.
Fixing issue 4:
This is only a semi-issue, but previously, if the buffer was a
certificate then we would cast our hash_function parameter into an
x509 hash function ID. Later on this value get passed to
get_crypto_alg_id() to get the crypto library specific MD identifier.
get_crypto_alg_id() does not keep track of x509 hash function ID (and
tbh the difference between them is not clear) and defaults to sha256.
The point of converting the given hash id to x509 hash id only to then
convert that value into crypto library specific ID is unnecessary.
To fix just stick with the hash function passed to the function.
Lastly, add a test case for all this.
Signed-off-by: Nick Child <nick.child@ibm.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR includes the implementation, documentation, and the testing for secvarctl ability to accept externally signed data. Rather than only supporting signatures computed internally. This will let secvarctl produce the presigned digest, allowing the user to use an external signing framework for the actual signature generation. From there, secvarctl can use this raw signed data when generating an auth or PKCS7 file. The process will look like this:
$ secvarctl generate c:x -n <varName> -t <timestamp> -i <newCrt> -o <digest.bin>
$<user generates the signature using another tool, resulting in file <signature.bin>
$secvarctl generate c:a -n <varName> -t <timestamp> -i <newCrt> -o <update.auth> -s <signature.bin> -c <signingCrt.pem>
NOTE: The timestamp and varName must be the exact same for both secvarctl commands