| Age | Commit message (Collapse) | Author |
|---|
|
ok miod@ millert@
|
|
For constant strings we don't actually need to use auth_mkvalue(3).
Problem reported by Ross L Richardson.
|
|
value < 0. errno is only updated in this case. Change all (most?)
callers of syscalls to follow this better, and let's see if this strictness
helps us in the future.
|
|
|
|
NULL pointer dereference. It tried to pass a file descriptor that
did not exist. This has to be done conditionally.
bug found by Raimund Specht with process accounting; OK millert@
|
|
use pledge and file locking. OK deraadt@
|
|
possible. Annotate <sys/param.h> lines with their current reasons. Switch
to PATH_MAX, NGROUPS_MAX, HOST_NAME_MAX+1, LOGIN_NAME_MAX, etc. Change
MIN() and MAX() to local definitions of MINIMUM() and MAXIMUM() where
sensible to avoid pulling in the pollution. These are the files confirmed
through binary verification.
ok guenther, millert, doug (helped with the verification protocol)
|
|
|
|
each cmsg_len (ie. msg_controllen = sum of CMSG_ALIGN(cmsg_len). This
works now that kernel fd passing has been fixed to accept a bit of
sloppiness because of this ABI repair.
lots of discussion with kettenis
|
|
CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
of alignment; ok kettenis hshoexer
|
|
an extensive discussion with otto, kettenis, millert, and hshoexer
|
|
turned off. This is consistent with historic UNIX behavior.
|
|
|
|
|
|
Fixes fd passing problems on sparc and sparc64. OK henning@
|
|
|
|
|
|
forth between login_skey and the invoking process. This allows us
to keep the record locked between an invocation of login_skey that
receives the challenge and another that verifies the response,
preventing an interloper from sniffing the challenge and beating
the legitimate user to the response.
|
|
|
|
|
|
|
|
|
|
|
|
hand editing to make comments line up correctly. Another pass is forthcoming that handles the cases that could not be done automatically.
|
|
update. We have to be careful and drop our lock if we are suspended
and then regain the lock on resume. This is necessary because the
user must not be allowed to keep a record locked for a long period
of time to avoid a DoS. We must be sure to re-lock when we resume
because otherwise an attacker could suspend us until a user starts
to login and then resume and then race the user for login using
the challenge response from the user.
|
|
getpass()/readpassphrase() from being able to restore the tty mode
on keyboard interrupt. Along with the recent readpassphrase.c commit
this means that if you ^C things that use login scripts (like su(1))
with a non-CBREAK shell your tty mode will be restored nicely.
TODO:
The various login scripts need to install handlers to avoid leaving
turd files or otherwise ending in a bad state. It would also be
nice to send BI_REJECT to the back channel.
|
|
millert@ ok
|
|
the invoking user.
|
|
|
|
will be used when BSD authentication is enabled
|
