Michael Brown [Thu, 22 Mar 2012 13:39:45 +0000 (13:39 +0000)]
[image] Add concept of trusted images
Trusted images may always be executed. Untrusted images may be
executed only if the current image trust requirement allows untrusted
images.
Images can be marked as trusted using image_trust(), and marked as
untrusted using image_untrust().
The current image trust requirement can be changed using
image_set_trust(). It is possible to make the change permanent, in
which case any future attempts to change the image trust requirement
will fail.
Michael Brown [Thu, 22 Mar 2012 12:22:38 +0000 (12:22 +0000)]
[bios] Set character attributes only when necessary
There is no INT 10 call for "display character with attribute,
advancing the cursor and scrolling the screen as necessary". We
therefore make two INT 10 calls: INT 10,09 to write the character with
its attribute at the current cursor position, and then INT 10,0e to
(re)write the character (leaving the attribute unchanged), advance the
cursor position and scroll as necessary.
This confuses the serial-over-LAN console redirection feature provided
by some BIOSes.
Fix by performing the INT10,09 only when necessary to change the
existing attribute.
Reported-by: Itay Gazit <itaygazit@gmail.com> Tested-by: Itay Gazit <itaygazit@gmail.com> Signed-off-by: Michael Brown <mcb30@ipxe.org>
Michael Brown [Tue, 20 Mar 2012 23:54:16 +0000 (23:54 +0000)]
[crypto] Add support for Cryptographic Message Syntax (PKCS #7)
The Cryptographic Message Syntax (PKCS#7) provides a format for
encapsulating digital signatures of arbitrary binary blobs. A
signature can be generated using
Michael Brown [Wed, 21 Mar 2012 13:57:32 +0000 (13:57 +0000)]
[crypto] Generalise X.509 OID-identified algorithm to asn1.c
The concept of an OID-identified algorithm as defined in X.509 is used
in some other standards (e.g. PKCS#7). Generalise this functionality
and provide it as part of the ASN.1 core.
Michael Brown [Wed, 21 Mar 2012 16:09:58 +0000 (16:09 +0000)]
[crypto] Treat ASN.1 OIDs as opaque
OIDs are theoretically part of a global hierarchy. However, the
hierarchy is sufficiently disorganised as to be essentially
meaningless for all purposes other than guaranteeing uniqueness.
Ignore the hierarchical nature of OIDs and treat them as opaque.
Michael Brown [Tue, 20 Mar 2012 13:32:20 +0000 (13:32 +0000)]
[build] Allow a client certificate to be specified at build time
Allow a client certificate and corresponding private key to be
specified at build time using the syntax
make CERT=/path/to/certificate KEY=/path/to/key
The build process uses openssl to convert the files into DER format,
and includes them within the client certificate store in
clientcert.c. The build process will prompt for the private key
password if applicable.
Note that the private key is stored unencrypted, and so the resulting
iPXE binary (and the temporary files created during the build process)
should be treated as being equivalent to an unencrypted private key
file.
Michael Brown [Tue, 20 Mar 2012 17:05:37 +0000 (17:05 +0000)]
[tls] Treat handshake digest algorithm as a session parameter
Simplify code by recording the active handshake digest algorithm as a
session parameter. (Note that we must still accumulate digests for
all supported algorithms, since we don't know which digest will
eventually be used until we receive the Server Hello.)
Michael Brown [Tue, 20 Mar 2012 04:07:53 +0000 (04:07 +0000)]
[tls] Use hybrid MD5+SHA1 algorithm
TLSv1.1 and earlier use a hybrid of MD5 and SHA-1 to generate digests
over the handshake messages. Formalise this as a separate digest
algorithm "md5+sha1".
Michael Brown [Sun, 18 Mar 2012 20:17:41 +0000 (20:17 +0000)]
[tls] Validate server certificate
Validate the server certificate against the trusted root certificate
store. The server must provide a complete certificate chain, up to
and including the trusted root certificate that is embedded into iPXE.
Note that the date and time are not yet validated.
Michael Brown [Sun, 18 Mar 2012 22:55:29 +0000 (22:55 +0000)]
[build] Allow trusted root certificates to be specified at build time
Allow trusted root certificates to be specified at build time using
the syntax
make TRUST=/path/to/certificate1,/path/to/certificate2,...
The build process uses openssl to calculate the SHA-256 fingerprints
of the specified certificates, and adds them to the root certificate
store in rootcert.c. The certificates can be in any format understood
by openssl.
The certificates may be server certificates or (more usefully) CA
certificates.
If no trusted certificates are specified, then the default "iPXE root
CA" certificate will be used.
Michael Brown [Sun, 18 Mar 2012 20:02:25 +0000 (20:02 +0000)]
[crypto] Force caller to provide temporary storage for modular calculations
bigint_mod_multiply() and bigint_mod_exp() require a fixed amount of
temporary storage for intermediate results. (The amount of temporary
storage required depends upon the size of the integers involved.)
When performing calculations for 4096-bit RSA the amount of temporary
storage space required will exceed 2.5kB, which is too much to
allocate on the stack. Avoid this problem by forcing the caller to
allocate temporary storage.
Michael Brown [Tue, 13 Mar 2012 16:35:21 +0000 (16:35 +0000)]
[crypto] Add big-integer library for RSA calculations
RSA requires modular exponentiation using arbitrarily large integers.
Given the sizes of the modulus and exponent, all required calculations
can be done without any further dynamic storage allocation. The x86
architecture allows for efficient large integer support via inline
assembly using the instructions that take advantage of the carry flag
(e.g. "adcl", "rcrl").
This implemention is approximately 80% smaller than the (more generic)
AXTLS implementation.
Michael Brown [Fri, 9 Mar 2012 15:45:56 +0000 (15:45 +0000)]
[crypto] Upgrade AES and RSA code to upstream axTLS version 1.4.5
All axTLS files are now vanilla versions of the upstream axTLS files,
with one minor exception: the unused "ctx" parameter of
bi_int_divide() has been marked with "__unused" to avoid a compilation
error.
Michael Brown [Tue, 6 Mar 2012 13:42:06 +0000 (13:42 +0000)]
[rng] Choose HMAC_DRBG using SHA-256 as the DRBG algorithm
Both HMAC_DRBG using SHA-1 and HMAC_DRBG using SHA-256 are Approved
algorithms in ANS X9.82 for our chosen security strength of 128 bits.
However, general recommendations (see e.g. NIST SP800-57) are to use a
larger hash function in preference to SHA-1.
Since SHA-256 is required anyway for TLSv1.2 support, there is no code
size penalty for switching HMAC_DRBG to also use SHA-256.
Michael Brown [Mon, 5 Mar 2012 16:21:49 +0000 (16:21 +0000)]
[rng] Use SHA-256 for Hash_df, and validate the hash function strength
ANS X9.82 Part 4 (April 2011 Draft) Section 13.3.4.2 states that "When
using the derivation function based on a hash function, the output
length of the hash function shall meet or exceed the security strength
indicated by the min_entropy parameter in the Get_entropy_input call",
although this criteria is missing from the pseudocode provided in the
same section.
Add a test for this condition, and upgrade from SHA-1 to SHA-256 since
SHA-1 has an output length of 160 bits, which is insufficient for
generating the (128 * 3/2 = 192) bits required when instantiating the
128-bit strength DRBG.
Michael Brown [Sun, 4 Mar 2012 21:41:24 +0000 (21:41 +0000)]
[i386] Use memory address constraints in __bswap_16s() and __bswap_64s()
Minimise code size by forcing the use of memory addresses for
__bswap_16s() and __bswap_64s(). (__bswap_32s() cannot avoid loading the
value into a register.)
Michael Brown [Sun, 4 Mar 2012 15:13:54 +0000 (15:13 +0000)]
[crypto] Replace SHA-1 implementation
Replace SHA-1 implementation from AXTLS with a dedicated iPXE
implementation which is around 40% smaller. This implementation has
been verified using the existing SHA-1 self-tests (including the NIST
SHA-1 test vectors).
Michael Brown [Sun, 4 Mar 2012 15:06:45 +0000 (15:06 +0000)]
[i386] Optimise byte-swapping functions and provide __bswap_{16,32,64}s()
Use the "bswap" instruction to shrink the size of byte-swapping code,
and provide the in-place variants __bswap_{16,32,64}s.
"bswap" is available only on 486 and later processors. (We already
assume the presence of "cpuid" and "rdtsc", which are available only
on Pentium and later processors.)
Michael Brown [Sat, 3 Mar 2012 14:11:46 +0000 (14:11 +0000)]
[tls] Support TLS version 1.1
Advertise support for TLS version 1.1, and be prepared to downgrade to
TLS version 1.0. Tested against Apache with mod_gnutls, using the
GnuTLSPriorities directive to force specific protocol versions.
Michael Brown [Fri, 2 Mar 2012 20:12:10 +0000 (20:12 +0000)]
[arp] Maintain an ARP transmission queue
Allow packet transmission to be deferred pending successful ARP
resolution. This avoids the time spent waiting for a higher-level
protocol (e.g. TCP or TFTP) to attempt retransmission.
Michael Brown [Fri, 2 Mar 2012 18:02:03 +0000 (18:02 +0000)]
[undi] Allow underlying PXE stack to construct link-layer header
Some PXE stacks (observed with a QLogic 8242) will always try to
prepend a link-layer header, even if the caller uses P_UNKNOWN to
indicate that the link-layer header has already been filled in. This
results in an invalid packet being transmitted.
Work around these faulty PXE stacks where possible by stripping the
existing link-layer header and allowing the PXE stack to (re)construct
the link-layer header itself.
Originally-fixed-by: Buck Huppmann <buckh@pobox.com> Signed-off-by: Michael Brown <mcb30@ipxe.org>
Michael Brown [Thu, 1 Mar 2012 16:26:38 +0000 (16:26 +0000)]
[iscsi] Send any padding inline with the data segment
Some iSCSI targets respond to a PDU before receiving the padding
bytes. If the target responds quickly enough, this can cause iPXE to
start processing a new TX PDU before the padding bytes have been sent,
which results in a protocol violation.
Fix by always transmitting the padding bytes along with the data
segment.
Originally-fixed-by: Shyam Iyer <shyam_iyer@dell.com> Signed-off-by: Michael Brown <mcb30@ipxe.org>
projects / thirdparty / ipxe.git / log
