glibc versions 2.06 and earlier did not set a umask on files created by
mkstemp created the world-writable. Wrap mkstemp to set and restore
the umask. From Coverity (CIDs 291826 291886 291891), ok djm@
Since we don't use getopt_long() nothing outside the getopt()
implementation itself uses this structure, so move it into the
source to remove it from visibility and clashes with libc's
ok dtucker@
POSIX doesn't specify the internal layout of the fd_set object, so let's
not assume it is just a bit mask. This increases compatibility with
systems that have a different layout.
The assumption is also worthless as we already refuse to use file
descriptors over FD_SETSIZE anyway. Meaning that the default size of
fd_set is quite sufficient.
Have it call native getentropy and fall back as required. Should fix
issues of platforms where libc has getentropy but it is not implemented
in the kernel. Based on github PR#354 from simsergey.
This brings us up to current, including djm's random-reseeding change,
as prompted by logan at cyberstorm.mu in bz#3467. It brings the
platform-specific hooks from LibreSSL Portable, simplified to match our
use case. ok djm@.
Putting this after the copyright statement (which doesn't change)
instead of before the version identifier (which does) prevents merge
conflicts when resyncing changes.
Since arc4random seeds from getentropy, and we use OpenSSL for that
if enabled, there's the possibility that if we build on a system that
does not have getentropy then run on a system that does have it, then
OpenSSL could end up calling our getentropy and getting stuck in a loop.
Pointed out by deraadt@, ok djm@
Factor out the arc4random seeding into its own file and change the
interface to match getentropy. Use native getentropy if available.
This will make it easier to resync OpenBSD changes to arc4random.
Prompted by bz#3467, ok djm@.
We have some compatibility hacks that were added to support OpenSSL
versions that do not support AES CTR mode. Since that time, however,
the minimum OpenSSL version that we support has moved to 1.0.1 which
*does* have CTR, so this is no longer needed. ok djm@
We have some compatibility hacks that were added to support OpenSSL
versions that do not support AES GCM mode. Since that time, however,
the minimum OpenSSL version that we support has moved to 1.0.1 which
*does* have GCM, so this is no longer needed. ok djm@
If getcwd() is supplied a buffer size of exactly 1 and a path of "/", it
could result in a nul byte being written out of array bounds. POSIX says
it should return ERANGE if the path will not fit in the available buffer
(with terminating nul). 1 byte cannot fit any possible path with its nul,
so immediately return ERANGE in that case.
OpenSSH never uses getcwd() with this buffer size, and all current
(and even quite old) platforms that we are currently known to work
on have a native getcwd() so this code is not used on those anyway.
Reported by Qualys, ok djm@
If the underlying system's select() returns bits that were not in the
request set, our ppoll() implementation can return revents for events
not requested, which can apparently cause a hang. Only return revents
for activity in the requested event set. bz#3416, analysis and fix by
yaroslav.kuzmin at vmssoftware com, ok djm@
