ktls/af_ktls

Issues

• test suite

  #17 opened 9 years ago by nmav
  2

• Paper might also wish to cite Plan 9's devtls as related work

  #96 opened 7 years ago by nwf
  1

• OpenConnect protocol support

  #19 opened 9 years ago by fridex
  1

• Support async crypto API

  #94 opened 8 years ago by djwatson
  0

• Review variable types

  #12 opened 8 years ago by fridex
  1

• Investigate using a single FD

  #92 opened 8 years ago by djwatson
  0

• DTLS window check needs to happen after decrypt

  #93 opened 8 years ago by djwatson
  0

• Crypto API scatterwalk copy

  #58 opened 8 years ago by djwatson
  2

• DTLS window handling from user space

  #13 opened 9 years ago by fridex
  2

• Issues when bind(2) is called more than once

  #44 opened 8 years ago by fridex
  0

• tls_sendmsg should send partial data for TCP

  #29 opened 8 years ago by djwatson
  6

• Design MSG_DONTWAIT for AF_KTLS

  #38 opened 8 years ago by fridex
  5

• Buffer changes

  #55 opened 8 years ago by djwatson
  6

• DTLS sliding window should always advance

  #87 opened 8 years ago by fridex
  0

• Don't depend on af_alg

  #27 opened 8 years ago by djwatson
  3

• Avoid kernel_sendmsg()

  #22 opened 8 years ago by fridex
  6

• MTU handling

  #61 opened 8 years ago by lancerchao
  8

• possible lock inversion surrounding splice

  #68 opened 8 years ago by lancerchao
  2

• check sock_queue_rcv_skb return value

  #75 opened 8 years ago by lancerchao
  4

• Review socket locking

  #10 opened 8 years ago by fridex
  1

• Fix indentation to conform Linux kernel requirements

  #8 opened 8 years ago by fridex
  1

• Copy data from multiple TLS records to userspace buffer

  #53 opened 8 years ago by lancerchao
  9

• peek tcp data using tcp_read_sock

  #74 opened 8 years ago by lancerchao
  2

• kernel fault when decrypting to user buffer

  #56 opened 8 years ago by lancerchao
  7

• Asynchronous cache should propagate error

  #26 opened 8 years ago by fridex
  2

• Documentation: no parallel socket operations on AF_KTLS and bound socket are possible without explicit synchronization

  #25 opened 8 years ago by fridex
  3

• Do MSG_PEEK call only once

  #37 opened 8 years ago by fridex
  1

• Per socket worker

  #20 opened 8 years ago by fridex
  1

• possibly dangerous cast

  #14 opened 8 years ago by nmav
  3

• Review async cache handling

  #9 opened 8 years ago by fridex
  2

• Race condition in KTLS_RECV_READY

  #71 opened 8 years ago by lancerchao
  2

• Remove rx async work

  #69 opened 8 years ago by ilyalesokhin
  6

• Allocating too many pages

  #65 opened 8 years ago by lancerchao
  1

• incorrect header length checks

  #23 opened 8 years ago by nmav
  3

• Avoid kernel_recvmsg()

  #21 opened 8 years ago by fridex
  3

• support poll

  #33 opened 8 years ago by djwatson
  5

• Sleeping with lock held

  #66 opened 8 years ago by lancerchao
  1

• sendmsg doesn't work with multiple iovecs

  #42 opened 8 years ago by djwatson
  4

• license is GPLv3

  #51 opened 8 years ago by nmav
  0

• new ciphersuite: chacha20-poly1305

  #50 opened 9 years ago by nmav
  1

• Test suite

  #47 opened 9 years ago by lancerchao
  6

• Pass all key material in one setsockopt(2) call

  #46 opened 9 years ago by fridex
  0

• DTLS over TCP is not possible

  #15 opened 9 years ago by nmav
  3

• rmmod crashes kernel

  #31 opened 9 years ago by djwatson
  3

• Error when a cipher is not present

  #7 opened 9 years ago by fridex
  1

• tls_bind issues

  #40 opened 9 years ago by djwatson
  1

• release & tls_rx_async_work race

  #32 opened 9 years ago by djwatson
  1

• Recv nonce and sequence number are reversed

  #30 opened 9 years ago by djwatson
  2

• sa_ktls and sa_cipher (KTLS_CIPHER_AES_GCM_128)

  #16 opened 9 years ago by nmav
  0

• Consider add likely() and unlikely() to meaningful parts

  #11 opened 9 years ago by fridex
  0