Mercurial > dropbear
view INSTALL @ 1306:34e6127ef02e
merge fixes from PuTTY import.c
toint() from misc.c
(revids are from hggit conversion)
changeset: 4620:60a336a6c85c
user: Simon Tatham <[email protected]>
date: Thu Feb 25 20:26:33 2016 +0000
files: import.c
description:
Fix potential segfaults in reading OpenSSH's ASN.1 key format.
The length coming back from ber_read_id_len might have overflowed, so
treat it as potentially negative. Also, while I'm here, accumulate it
inside ber_read_id_len as an unsigned, so as to avoid undefined
behaviour on integer overflow, and toint() it before return.
Thanks to Hanno Böck for spotting this, with the aid of AFL.
(cherry picked from commit 5b7833cd474a24ec098654dcba8cb9509f3bf2c1)
Conflicts:
import.c
(cherry-picker's note: resolving the conflict involved removing an
entire section of the original commit which fixed ECDSA code not
present on this branch)
changeset: 4619:9c6c638d98d8
user: Simon Tatham <[email protected]>
date: Sun Jul 14 10:45:54 2013 +0000
files: import.c ssh.c sshdss.c sshpubk.c sshrsa.c
description:
Tighten up a lot of casts from unsigned to int which are read by one
of the GET_32BIT macros and then used as length fields. Missing bounds
checks against zero have been added, and also I've introduced a helper
function toint() which casts from unsigned to int in such a way as to
avoid C undefined behaviour, since I'm not sure I trust compilers any
more to do the obviously sensible thing.
[originally from svn r9918]
changeset: 4618:3957829f24d3
user: Simon Tatham <[email protected]>
date: Mon Jul 08 22:36:04 2013 +0000
files: import.c sshdss.c sshrsa.c
description:
Add an assortment of extra safety checks.
[originally from svn r9896]
changeset: 4617:2cddee0bce12
user: Jacob Nevins <[email protected]>
date: Wed Dec 07 00:24:45 2005 +0000
files: import.c
description:
Institutional failure to memset() things pointed at rather than pointers.
Things should now be zeroed and memory not leaked. Spotted by Brant Thomsen.
[originally from svn r6476]
changeset: 4616:24ac78a9c71d
user: Simon Tatham <[email protected]>
date: Wed Feb 11 13:58:27 2004 +0000
files: import.c
description:
Jacob's last-minute testing found a couple of trivial bugs in
import.c, and my attempts to reproduce them in cmdgen found another
one there :-)
[originally from svn r3847]
changeset: 4615:088d39a73db0
user: Simon Tatham <[email protected]>
date: Thu Jan 22 18:52:49 2004 +0000
files: import.c
description:
Placate some gcc warnings.
[originally from svn r3761]
changeset: 4614:e4288bad4d93
parent: 1758:108b8924593d
user: Simon Tatham <[email protected]>
date: Fri Oct 03 21:21:23 2003 +0000
files: import.c
description:
My ASN.1 decoder returned wrong IDs for anything above 0x1E! Good
job it's never had to yet. Ahem.
[originally from svn r3479]
author | Matt Johnston <matt@ucc.asn.au> |
---|---|
date | Tue, 12 Jul 2016 23:00:01 +0800 |
parents | 2d943453cecf |
children | 8f88f4290b22 |
line wrap: on
line source
Basic Dropbear build instructions: - Edit options.h to set which features you want. - Edit debug.h if you want any debug options (not usually required). (If using a non-tarball copy, "autoconf; autoheader") ./configure (optionally with --disable-zlib or --disable-syslog, or --help for other options) Now compile: make PROGRAMS="dropbear dbclient dropbearkey dropbearconvert scp" And install (/usr/local/bin is usual default): make PROGRAMS="dropbear dbclient dropbearkey dropbearconvert scp" install (you can leave items out of the PROGRAMS list to avoid compiling them. If you recompile after changing the PROGRAMS list, you *MUST* "make clean" before recompiling - bad things will happen otherwise) See MULTI for instructions on making all-in-one binaries. If you want to compile statically, add "STATIC=1" to the make command-line. Binaries can be stripped with "make strip" ============================================================================ If you're compiling for a 386-class CPU, you will probably need to add CFLAGS=-DLTC_NO_BSWAP so that libtomcrypt doesn't use 486+ instructions. ============================================================================ Compiling with uClibc: Firstly, make sure you have at least uclibc 0.9.17, as getusershell() in prior versions is broken. Also note that you may get strange issues if your uClibc headers don't match the library you are running with, ie the headers might say that shadow password support exists, but the libraries don't have it. Compiling for uClibc should be the same as normal, just set CC to the magic uClibc toolchain compiler (ie export CC=i386-uclibc-gcc or whatever). You can use "make STATIC=1" to make statically linked binaries, and it is advisable to strip the binaries too. If you're looking to make a small binary, you should remove unneeded ciphers and MD5, by editing options.h It is possible to compile zlib in, by copying zlib.h and zconf.h into a subdirectory (ie zlibincludes), and export CFLAGS="-Izlibincludes -I../zlibincludes" export LDFLAGS=/usr/lib/libz.a before ./configure and make. If you disable zlib, you must explicitly disable compression for the client - OpenSSH is possibly buggy in this regard, it seems you need to disable it globally in ~/.ssh/config, not just in the host entry in that file. You may want to manually disable lastlog recording when using uClibc, configure with --disable-lastlog. One common problem is pty allocation. There are a number of types of pty allocation which can be used -- if they work properly, the end result is the same for each type. Running configure should detect the best type to use automatically, however for some systems, this may be incorrect. Some things to note: If your system expects /dev/pts to be mounted (this is a uClibc option), make sure that it is. Make sure that your libc headers match the library version you are using. If openpty() is being used (HAVE_OPENPTY defined in config.h) and it fails, you can try compiling with --disable-openpty. You will probably then need to create all the /dev/pty?? and /dev/tty?? devices, which can be problematic for devfs. In general, openpty() is the best way to allocate PTYs, so it's best to try and get it working.