<feed xmlns='http://www.w3.org/2005/Atom'>
<title>openbsd/src/lib/libcrypto/evp, branch OPENBSD_7_2</title>
<subtitle>A mirror of https://github.com/libressl/openbsd.git
</subtitle>
<id>https://git.lua4.win/openbsd/atom?h=OPENBSD_7_2</id>
<link rel='self' href='https://git.lua4.win/openbsd/atom?h=OPENBSD_7_2'/>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/'/>
<updated>2022-09-15T07:04:19+00:00</updated>
<entry>
<title>Use LONG_MAX as the limit for ciphers with long based APIs.</title>
<updated>2022-09-15T07:04:19+00:00</updated>
<author>
<name>jsing</name>
<email></email>
</author>
<published>2022-09-15T07:04:19+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=69a6645367fe0e98f414f8ce038c6a4c2e3fb102'/>
<id>urn:sha1:69a6645367fe0e98f414f8ce038c6a4c2e3fb102</id>
<content type='text'>
These ciphers have long based APIs, while EVP has a size_t based API. The
intent of these loops is to handle sizes that are bigger than LONG_MAX.
Rather than using the rather crazy EVP_MAXCHUNK construct, use LONG_MAX
rounded down to a large block size, ensuring that it is a block size
multiple. Revert the recently added overflow checks now that this is
handled more appropriately.

ok tb@
</content>
</entry>
<entry>
<title>Stop pretending that EVP_CIPHER cleanup can fail.</title>
<updated>2022-09-13T04:59:18+00:00</updated>
<author>
<name>jsing</name>
<email></email>
</author>
<published>2022-09-13T04:59:18+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=2ec5e8a2e012afe68c1a580604d754b7e8cc73ee'/>
<id>urn:sha1:2ec5e8a2e012afe68c1a580604d754b7e8cc73ee</id>
<content type='text'>
Now that EVP_CIPHER is opaque, stop pretending that EVP_CIPHER cleanup can
fail.

ok tb@
</content>
</entry>
<entry>
<title>Expose EVP_chacha20_poly1305()</title>
<updated>2022-09-11T17:29:24+00:00</updated>
<author>
<name>tb</name>
<email></email>
</author>
<published>2022-09-11T17:29:24+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=994f306172b8bb6dce5e7d2ba813e435994c7bab'/>
<id>urn:sha1:994f306172b8bb6dce5e7d2ba813e435994c7bab</id>
<content type='text'>
ok jsing
</content>
</entry>
<entry>
<title>Expose various EVP AEAD constants for EVP ChaCha and QUIC</title>
<updated>2022-09-11T17:28:33+00:00</updated>
<author>
<name>tb</name>
<email></email>
</author>
<published>2022-09-11T17:28:33+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=66a87f09c5b56ec4b967f10cac7a0eb7aa5e5d59'/>
<id>urn:sha1:66a87f09c5b56ec4b967f10cac7a0eb7aa5e5d59</id>
<content type='text'>
ok jsing
</content>
</entry>
<entry>
<title>Increment the input and output position for EVP AES CFB1.</title>
<updated>2022-09-10T17:45:10+00:00</updated>
<author>
<name>jsing</name>
<email></email>
</author>
<published>2022-09-10T17:45:10+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=8b1f87ff17c15812707feb063a073472bba18199'/>
<id>urn:sha1:8b1f87ff17c15812707feb063a073472bba18199</id>
<content type='text'>
The length is decremented, however the input is repeatedly read from and
output written to the same position. Correct this by actually incrementing
the input and output pointers.

Found via OpenSSL 604e591ed7,

ok tb@
</content>
</entry>
<entry>
<title>Use correct length for EVP CFB mode ciphers.</title>
<updated>2022-09-10T17:39:47+00:00</updated>
<author>
<name>jsing</name>
<email></email>
</author>
<published>2022-09-10T17:39:47+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=de6c106921914263e126f0a65ff53578716ce425'/>
<id>urn:sha1:de6c106921914263e126f0a65ff53578716ce425</id>
<content type='text'>
The BLOCK_CIPHER_* macros contained a bug where the total length is passed
to the underlying cipher implementation, rather than the length of the
current chunk. Correct this and use the chunk length instead.

Should address the remaining issues reported by Coverity.

ok tb@
</content>
</entry>
<entry>
<title>Replace aes_{ccm,xts}_cleanup usage with NULL.</title>
<updated>2022-09-06T06:38:26+00:00</updated>
<author>
<name>jsing</name>
<email></email>
</author>
<published>2022-09-06T06:38:26+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=8915991783ef245f913d2d2d5b64d91f6a179d6d'/>
<id>urn:sha1:8915991783ef245f913d2d2d5b64d91f6a179d6d</id>
<content type='text'>
Only change in generated assembly is due to line numbers.
</content>
</entry>
<entry>
<title>Stop casting a size_t to a long and then passing it as a size_t.</title>
<updated>2022-09-06T06:17:11+00:00</updated>
<author>
<name>jsing</name>
<email></email>
</author>
<published>2022-09-06T06:17:11+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=1a2b24a50a3e7801e8ad635598d6079f62b970cb'/>
<id>urn:sha1:1a2b24a50a3e7801e8ad635598d6079f62b970cb</id>
<content type='text'>
These cipher implementations take a size_t length argument, so stop
casting it to a long.

Found by Coverity.

ok tb@
</content>
</entry>
<entry>
<title>Remove dead code.</title>
<updated>2022-09-04T15:56:51+00:00</updated>
<author>
<name>jsing</name>
<email></email>
</author>
<published>2022-09-04T15:56:51+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=6e1462350b37c8a46335178d23fa0738f688825f'/>
<id>urn:sha1:6e1462350b37c8a46335178d23fa0738f688825f</id>
<content type='text'>
No change in generated assembly.
</content>
</entry>
<entry>
<title>Add bounds checks for various EVP cipher implementations.</title>
<updated>2022-09-04T15:45:25+00:00</updated>
<author>
<name>jsing</name>
<email></email>
</author>
<published>2022-09-04T15:45:25+00:00</published>
<link rel='alternate' type='text/html' href='https://git.lua4.win/openbsd/commit/?id=8869e441d0c3b744fcf47865bc3a28770c66b2b5'/>
<id>urn:sha1:8869e441d0c3b744fcf47865bc3a28770c66b2b5</id>
<content type='text'>
The EVP cipher API uses size_t, however a number of the underlying
implementations use long in their API. This means that an input with
size &gt; LONG_MAX will go negative.

Found by Coverity, hiding under a large pile of macros.

ok tb@
</content>
</entry>
</feed>
