summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2007-08-01Stop making de win over dc on alpha, and include both dc and de on allMiod Vallat
installation media now - it fits. THIS MEANS YOU NEED TO RENAME /etc/hostname.de0 TO /etc/hostname.dc0 BEFORE BOOTING A NEW KERNEL. ok deraadt@
2007-08-01make sure the MIC field is set in group key handshake messages beforeDamien Bergamini
passing the frame to the handlers.
2007-08-01make ieee80211_recv_wpa_group_msg1() install the GTK in ic_nw_keys[] too.Damien Bergamini
i only converted ieee80211_recv_rsn_group_msg1() in my previous commit.
2007-08-01- -T before -tJason McIntyre
- use .Bk/.Ek
2007-08-01in (unused function) free_core(), unmap the core; from veins@evilkittens.orgTheo de Raadt
2007-08-01close core file fd after mmap; veins@evilkittens.orgTheo de Raadt
2007-08-01add generic ieee80211_encrypt() and ieee80211_decrypt() functions thatDamien Bergamini
can handle multiple ciphers (the key to use is determined automatically by these functions based on the frame's destination address). add ieee80211_ccmp_encrypt() and ieee80211_ccmp_decrypt(). those two functions only do encapsulation/decapsulation of CCMP frames for now (they don't do SW crypto). they will help to test things with drivers that can do HW crypto. add a ni_pairwise_key field to struct ieee80211_node to actually install the pairwise transient key. install the GTK in ic_nw_keys[].
2007-08-01unbreak.Damien Bergamini
2007-08-01Refactor realloc() + checks into erealloc()Todd C. Miller
Fix a NULL dereference when the -t flag is given and a line consists solely of delimiters. OK deraadt@, fixes PR 5555
2007-08-01check the key length field in message 3 of the 4-way handshake.Damien Bergamini
change ieee80211_recv_eapol() so that it is called without the ethernet header striped.
2007-08-01switch i386 to use the MI i8253 header file and remove the now obsolete MDMartin Reindl
timerreg.h ok miod@
2007-08-01set key length field in message 1 & 3 of the 4-way handshake.Damien Bergamini
2007-08-01The wdt char device was removed during c2k6 (or shortly after, don'tMichael Knudsen
remember), so we no longer need wdt.h in the build dir for NWDT, which means wdt can stop being needs-flag. ok dlg
2007-08-01wdt doesn't malloc so no reason to include malloc.h.Michael Knudsen
ok dlg
2007-08-01group key handshake message 1 is very different between RSN and WPA1.Damien Bergamini
RSN uses a GTK KDE while WPA1 stores the GTK in the EAPOL-Key frame data field (encrypted) and uses some bits in the info field. split ieee80211_recv_group_msg1() in two separate functions.
2007-08-01remove the need to inspect the output and check everything programmaticallyKurt Miller
okay kettenis@
2007-08-01pass the supplicant's nonce generated on reception of message 1 of theDamien Bergamini
4-way handshake to ieee80211_send_4way_msg2().
2007-08-01in 4-way handshake messages 2 and 3, parse vendor-specific WPA1 IEs too.Damien Bergamini
fix a test for ni_replaycnt_ok while i'm here.
2007-08-01add a TSC field to the key structure. it will be used to store theDamien Bergamini
TKIP sequence counter (TKIP) or the PN (packet number, CCMP).
2007-08-01PMKID KDE is optionnal in message 1 of the 4-way handshake (at leastDamien Bergamini
when using WPA-PSK). add a ni_replaycnt_ok flag to struct ieee80211_node to mark the replay counter as valid. the replay counter is marked valid only when message 3 of the 4-way handshake is received since message 1 contains no MIC.
2007-08-01fix parsing and construction of RSN/WPA IEs.Damien Bergamini
2007-08-01all WPA implementations i have tested use EAPOL-Key frames version 1,Damien Bergamini
so use that too and remove a check in ieee80211_recv_eapol(). WPA1 stores the group key id into bits 4-5 of the EAPOL-Key frame info field and uses bit 6 to indicate if the key is Rx/Tx or Rx only. remove a check in ieee80211_eapol_key_decrypt() because WPA1 encrypts the payload of message 1 of the group-key handshake without setting the encrypted bit in the info field.
2007-08-01add cvs id now that this file has diverged from the sudo cvs versionTodd C. Miller
2007-08-01Correct a 10-year old typo inherited from NetBSD by renaming psm.c,Mark Kettenis
psm_intelli.c and psmreg.h into pms.c, pms_intelli.c and pmsreg.h. ok deraadt@
2007-08-01the EAPOL-Key MIC must be computed with the MIC bit set.Damien Bergamini
this simplifies ieee80211_eapol_key_mic() and ieee80211_eapol_key_check_mic() quite a bit. set the EAPOL-Key body length before computing the MIC since the MIC is computed with the 802.1X header too. add a missing htons() while i'm here.
2007-08-01add pkg_add and make release environment variables to env_keepTodd C. Miller
2007-08-01KEYACK bit is not set for 4-way handshake messages 2 and 4.Damien Bergamini
it is set in message 3.
2007-08-01cipher suite type for CCMP is 4 not 3.Damien Bergamini
2007-08-01unused, and Miod says this file was superseded by sys/dev/ic/i8253reg.h, soMartin Reindl
remove
2007-08-01Improve state switching and fix some debug output.Marcus Glocker
2007-08-01escape [, ], and ? as well. from cbiere@netbsd.Federico G. Schwindt
millert@ ok.
2007-08-01allow ftp-proxy to add tag statements to teh rules it insertsHenning Brauer
clever, nice and easy diff from bsd@openbsd.rutgers.edu, ok pyr reyk
2007-07-31Add first version to enable auto association (best signal AP). ThereforeMarcus Glocker
remove experimental, hard-coded MAC address block. Also tested by claudio@
2007-07-31Add missing splx(s); calls. OK mglocker@Claudio Jeker
2007-07-31Plug a mbuf leak in the TX path and replace a handbuilt m_copydata()Claudio Jeker
implementation with m_copydata(). Remove the CMALO_DEBUG that slipped in somehow. OK mglocker@
2007-07-31don't duplicate .numt in .desc; 'In this driver, whatever you want' deraadt@Constantine A. Murenin
2007-07-31change sensor type: if .value can only ever be 0 or 1, .type should be indicatorConstantine A. Murenin
2007-07-31make rip code work like mmc code for how it does progress reports (useTheo de Raadt
a timer, so that stderr does not get splattered); tested by xsa and others
2007-07-31remove magic numbers: use PCAGPIO_NPINS instead (no binary change); ok deraadt@Constantine A. Murenin
2007-07-31Use correct function name in log message. Noticed by Igor ZinovkHans-Joerg Hoexer
zinovik@cs.karelia.ru. Thanks!
2007-07-31+ hidden subdirKurt Miller
2007-07-31add regress to test __asm(".hidden symbol") works as expected. for gcc3Kurt Miller
arch only. okay kettenis@
2007-07-31typo fixes from joel@ionix.com.auTheo de Raadt
2007-07-31Our audio layer doesn't know anything about mixer names likeDeanna Phillips
"selector" and "front", so instead create the items that it actually needs. This unbreaks the mute button in wscons, and allows setting the volume in programs like xmms and xine. This is for STAC9200 only - similiar fixes for other codecs coming soon. Tested and ok krw@
2007-07-31In the mixer to device level conversions, round down to the nearestDeanna Phillips
valid step, and keep them in sync. This lets audio(4) set exact gain levels based on the mixer delta, and unbreaks volume up/down buttons in wscons. This also avoids a divide by zero that could occur if the generic mixer init found an amplifier with zero steps. Problem noticed and tested by jmc@, similar diff tested by krw@
2007-07-31Start to improve TX/RX path; don't accept further packets from TX queueMarcus Glocker
unless the last packet has been processed by the device. Protect RX routine with splnet(). This changes let me scp with ~575KB/s in 11b for now.
2007-07-31Simplify TX path.Marcus Glocker
2007-07-31something went wrong w/ previous;Jason McIntyre
2007-07-31document the -d option of sensorsd; input and ok from jmc@, deraadt@Robert Nagy
2007-07-31oops, forgot to include the man page tweaks from jmc@Pierre-Yves Ritschard