summaryrefslogtreecommitdiff
path: root/src/nss/README
blob: 884ade9909023e9176bcab731315738362d30544 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
WHAT VERSION OF NSS?
------------------------------------------------------------------------
NSS 3.8 or greater and NSPR 4.3 or greater are required.

KEYS MANAGER
------------------------------------------------------------------------

xmlsec-nss key manager uses a custom Keys Store, and a custom X509 Store.
The custom Keys Store and the X509 Store use the NSS database as the underlying 
store for public/private keys, Certs and CRLs.

The NSS Keys store uses the Simple Keys Store on top of the NSS repository.
The reason for this is that XMLSEC's generic adoptkey/getKey functions use a
XMLSEC key object that contains more attributes than the raw NSS key object,
and the getkey function may use a combination of one or more of these attributes
(name, type, usage, Id) to find a key. There is no straightforward 1-1 mapping 
between XMLSEC's adoptkey/getkey and NSS's APIs. 

For example, the store may be asked to adopt a symmetric key, and later asked 
to find it just by name. Or the store may be asked to adopt a private key 
just by its type, and later asked to find it just by type. The key returned 
by getKey is expected to contain all the attributes that were present at the 
time of adoptkey - NSS store does not provide a way to store app-specific 
attributes.

When a key is adopted by the NSS Keys Store, it is simply saved in the 
Simple Keys Store. It is not saved into the NSS database. The only
way to load keys into the NSS database is with a load operation through
the XMLSEC API or via an administrator operation.

When a getKey is done on the NSS Keys Store, it first checks the Simple
Keys Store. If the key is found there, it is returned. If not, the key
is searched in the NSS database. If found, the key is stored in the
Simple Keys Store before it is returned.


Thus, the various sources for keys/certs/crls for an XMLSEC-NSS application
are:
- elements in XML documents
- PKCS12 and DER files
- NSS Database


KNOWN ISSUES.
------------------------------------------------------------------------
1)  NSS needs to provide a way to convert a DER integer string to an ASCII 
decimal string. Once NSS is fixed, the function xmlSecNssASN1IntegerWrite 
in src/nss/x509.c needs to be implemented. 
    NSS bug: http://bugzilla.mozilla.org/show_bug.cgi?id=212864
    xmlsec bug: http://bugzilla.gnome.org/show_bug.cgi?id=118633

2) RSA Encryption/Decryption using PKCS#1 v1.5 padding not currently exposed 
in NSS. This causes some tests to fail.
          
    NSS bug: http://bugzilla.mozilla.org/show_bug.cgi?id=214236
    xmlsec bug: http://bugzilla.gnome.org/show_bug.cgi?id=118628

3) RSA-OAEP is not yet implemented in NSS. This is the only REQUIRED algorithm 
that is missing from xmlsec-nss.
           
    NSS bug: http://bugzilla.mozilla.org/show_bug.cgi?id=158747
    xmlsec bug: http://bugzilla.gnome.org/show_bug.cgi?id=118629    

4) There are 2 fixes in NSS that will be rolled into the next release (3.9). 
They are:
    - http://bugzilla.mozilla.org/show_bug.cgi?id=208194 
    Once available, need to change src/nss/x509.c. Replace 
        PR_AtomicIncrement(&(crlSrc->referenceCount));
    with
        SEC_DupCrl(crlSrc);
    (there is a comment there that already has the correct line)

    - http://bugzilla.mozilla.org/show_bug.cgi?id=211384
    No changes required. The "NSS_Shutdown failed" in some of the
    current tests will go away

xmlsec bug: http://bugzilla.gnome.org/show_bug.cgi?id=118630

5) CERT_FindCertByNameString does not work in all cases

   NSS bug: http://bugzilla.mozilla.org/show_bug.cgi?id=210709
   xmlsec bug: http://bugzilla.gnome.org/show_bug.cgi?id=118631

6) CERT_FindCertBySubjectKeyID does not work in all cases
	   
   NSS bug: http://bugzilla.mozilla.org/show_bug.cgi?id=211051
   xmlsec bug: http://bugzilla.gnome.org/show_bug.cgi?id=118632

7) Finding a cert by Issuer & Serial Number needs the ability to
convert an ASCII decimal string to a DER integer string. Filed
an RFE against NSS. Once fixed, xmlSecNumToItem in x509vfy.c 
needs to be changed to use the new function(s) provided

    NSS bug: http://bugzilla.mozilla.org/show_bug.cgi?id=212864
    xmlsec bug: http://bugzilla.gnome.org/show_bug.cgi?id=118633

8) RIPEMD160 Digest and RIPEMD160 HMAC is not supported by NSS

   xmlsec bug: http://bugzilla.gnome.org/show_bug.cgi?id=118634
	    
9) AES Key wrap algorithm is implemented in NSS but not exposed due to 
some bug src/nss/kw_aes.c uses a workaround which should be removed
when the bug is fixed

   NSS bug: http://bugzilla.mozilla.org/show_bug.cgi?id=213795
   xmlsec bug: http://bugzilla.gnome.org/show_bug.cgi?id=118635

10) Not all file formats are supported

- xmlSecNssAppKeyLoad(): This function loads a PKI key from a file. 
    The following formats are supported:
	. xmlSecKeyDataFormatDer: This expects the private key to be in 
		PrivateKeyInfo format. Note that the DER files containing 
                private keys in the xmlsec test suite aren't in that format
	. xmlsecKeyDataFormatPkcs12 

    The following formats are not supported:
	. xmlSecKeyDataFormatPkcs8Pem
	. xmlSecKeyDataFormatPkcs8Der


- xmlSecNssAppCertLoad(): This function loads a cert from a file. 
    The following formats are supported:
	xmlSecKeyDataFormatDer

    The following formats are not supported:
	xmlSecKeyDataFormatPem

11) "Trusted" vs "Untrusted" certificates: 
The distinction between "trusted" and "untrusted" certificates in 
xmlsec-openssl is maintained because the OPENSSL application (and
not the OPENSSL library) has to maintain a cert store and verify 
certificates. With NSS, no such distinction is necessary in the
application.

Aleksey: Not sure that I understand this point but thats what Tej wrote.