simplesshd/NOTES

BUILD INSTRUCTIONS:


To build this, you'll need to remove my key.* settings in ant.properties.
You'll have to build a "debug" apk instead of a "release" one.  Android
will not let you install a debug apk on top of a release one, so you have
to remove stock SimpleSSHD first before installing the debug build.

Then follow these steps (roughly the "doit" script):

   ndk-build -j8 &&
   mv libs/armeabi/scp libs/armeabi/libscp.so &&
   mv libs/armeabi/sftp-server libs/armeabi/libsftp-server.so &&
   mv libs/armeabi/rsync libs/armeabi/librsync.so &&
   mv libs/armeabi/buffersu libs/armeabi/libbuffersu.so &&
   ant debug

The mv steps are very important, because ant will only package the
necessary binaries if they have a .so extension (even though they are
stand alone executables).



DEVELOPMENT JOURNAL:


December 6, 2014.

The idea is to make a proper ssh implementation for Android.  Important
features:

   * it should run happily without root (on a non-root port)

   * it should be a regular android app requiring no special permissions,
     and not requiring any 'magic' executable files

   * should not rely on busybox

   * preferably support sftp

   * open source

The existing apps are either expensive, don't work, need root, or too
complicated, or a mix of all of the above.  And none of them are open
source.

I figure I'll start with dropbear, which I will run through JNI instead
of putting it in its own binary (because making such a binary executable
is a bit of a hack).

So that's the plan........


December 14, 2014.

I got dropbear to compile under the Android NDK, so now it's time to work
on the Android side of it.

I need:
  * a Service that can be started, stopped, and queried for whether it's
    running or not
  * a Thread to implement the Service's work (by calling into dropbear's
    main()), which can also be stopped.
  * a config UI with at least these choices:
    - bool: start on boot (def: false)
    - number: port number (def: 2222)
    - string: path to authorized_keys file (def: /sdcard/ssh)
    - string: name of default shell (def: /system/bin/sh -l)
    - string: default path for HOME (def: /sdcard/ssh)
    - button: start or (if it's running) stop


December 15, 2014.

Getting to the fun part.  Process management...

To start sshd, it seems like I can startService().  Then in the Service's
onStartCommand(), call startForeground() so it won't be killed (return
START_STICKY too?).

The question is if dropbear's main() should run under a separate Thread,
or a separate Process.  The trouble with a Thread is that it might be
hard to kill.  The trouble with a process is that there is no way to
report back status (such as a failure to start sshd).

Connectbot starts a new process for its shell -- it really doesn't have a
choice because the shell binary isn't linked with Connectbot, and exec()
in a thread stinks.  To stop it, it just closes stdin/stdout!!!  So
zombies can (and do) linger.

I suppose dropbear could be in its own process if it had something like
stdin/stdout to communicate failure?  Or it could just write error
messages to (i.e.) /sdcard/ssh/log.  To stop the service, it would just
use kill().

I am curious how the main waiting-for-connections loop looks, but even if
it uses select(), I'm not sure how I would honor Thread.interrupt() or
whatever.  It's not guaranteed to interrupt select(), and I'm not keen on
adding an arbitrary timeout/polling feature to it.


December 20, 2014.

So, I added a builtin scp endpoint.  It was pretty straight forward,
except dropbear defaults to vfork(), which blocks the parent until the
child runs execve()!!

Anyways, I noticed that scp doesn't quote its arguments to the remote
scp. That means you can't conveniently copy a remote file with a space in
its name (it becomes two files).  But the upside is that this is where
wildcards are handled -- by the shell!

So I need to either run it as a separate executable launched through the
shell, or make my own implementation of wildcards.

It is easy, using a $(BUILD_EXECUTABLE) script, to get ndk to build an
executable.  But it is only packaged up if it is named "gdbserver" (and
debug apk), or "libfoo.so".  The good news is that libfoo.so can be
executed in /data/data/org.galexander.sshd/lib/libfoo.so, so that is a
viable option.

Doing the expansion myself is not necessarily hard either, though.  I
need a library function called glob(), which is apparently not part of
bionic.  But I have the idea some cut and paste would resolve that with
very little extra work on my part.


December 21, 2014.

Well, bionic libc *does* provide fnmatch(), and even scandir() (a
shortcut for readdir).  In the best case, though, that still leaves me
with a bit of a path parsing conundrum (I have to tell scandir which
directory to operate on).  And also a bit of an escape character
conundrum -- \* and "*" should not act like wildcards.

Those are not insurmountable but I think I've talked myself out of it.
So then the question is, do I figure out how to ship an executable, or
do I do some hack like open a pipe to "/system/bin/sh echo filespec" and
use the shell solely for expansion?

I'm developing the idea that it's actually pretty easy to ship an
executable, I just need to find some -pre-package step where I can do
"mv scp libscp.so" and then it will ship.  ndk-build will not let me make
a target with a "." in it directly.

...

Now scp, sftp-server, and rsync work as separate executables... rsync
does fail at -z because it needs it's own custom zlib...The stock
"external" one seems to lack the "old-style compress" method.  There is
another commandline option for the new (deflate?) technique, but I think
the normal -z ought to work too.  But that is literally the last feature!
Then just release details.


December 29, 2014.

First problem report from a user.  Lollipop (Android 5.0) requires "PIE"
executables -- position independent code.  I think that is a modern
equivalent to -fpic that Android is now requiring so that it can
randomize addresses to try to obscure stack smashing attacks that rely on
fixed addresses.  It is epicly lame.

Anyways, the big fuck-you from Google is that Ice Cream Sandwich (Android
4.1) and earlier require fixed-position code.  So one binary will not
generally work on both.

Here is a good summary:
   https://code.google.com/p/android-developer-preview/issues/detail?id=888

There is something called "run_pie" which you can wrap your executables
in that lets older Android run PIE executables.  It would require a
relatively small change to the exec() call to prepend it with "run_pie".
That seems like a hack.

The suggested remedy is to build two different apks!  Yuck!

Anyways, it is only executables (not libraries -- they are position
independent already) that are affected.  And apparently static
executables don't care one way or the other.

So that is my remedy -- static executables for the moment.  I tested them
and it is only a little bit bigger -- 904kB of binaries instead of 668kB.


January 18, 2015.

Markus Ethen suggested it display the current IP address so you know
where to ssh to, in case it isn't convenient to use static dhcp or to
remember the address.  That seems to be easier said than done.  You can
use WifiManager, but that won't give your IP address unless you're on
wifi.  That is probably "good enough", but it is certainly not ideal.
There is also java.net.NetworkInterface, which seems to return a random
ipv6 address.

Ah-hah!  It is fe80::macaddr, which is a bogus "local-connection only"
ipv6 address, like 192.168, but automatically-generated without dhcp.
So if I skip that, it finds the proper ipv4 address!

Pfew!  I was thinking I'd have to directly use /proc/net/dev and
SIOCGIFCONF, just like ifconfig does, but it works fine with
java.net.NetworkInterface.


June 20, 2015.

At some points, rsync is only write()ing, and assumes that the other end
will receive it all.  The other end does a little bit of write()ing of
its own, and then is happy to read() all it wants.  So this written stuff
may sit in a buffer somewhere indefinitely.  If that happens, an
infelicity in the design of SuperSU causes everything to wedge.

Of course, this is only if you set the shell to /system/xbin/su as a way
of having root access for rsync.

Anyways, I made a new program, "buffersu", which is just a
stdin/stdout-buffering wrapper for rsync that is guaranteed to always
perform any read() that is possible at any time, no matter how many
write()s are outstanding.  That seems to do the trick.


June 21, 2016.

Chris Moore reports that rsync and sftp do not like files larger than 2GB.

rsync was easy - it just needed an additional #define in rsync/config.h
to enable its builtin support for using stat64/lseek64/off64_t/etc.

Now that I'm investigating sftp, I find this surprising fact about bionic
(though the glibc man page for stat(2) tried to tell me this) - stat64
and stat are the same thing!  But off64_t and lseek64 are significant.
That should make converting sftp pretty convenient.  Especially since
sftp already uses "u_int64_t" instead of off_t.

p.s. Chris Moore gave me this command to test sftp, which turned out to
be useful:

   curl -v --pubkey .ssh/id_rsa.pub -r 2147482624-2147484672 -k sftp://mushroom:2222/sdcard/ssh/buh -o buh-new

As for scp, it's not as clear what needs to be done.  It doesn't use
lseek.  But it does use off_t a bit, including on an index in a for loop
that is compared against st_size (which is 64-bit).  So I'll just change
all of the off_t to off64_t and hope for the best.

sftp and rsync work!  Not gonna bother testing scp on big files...


October 1, 2016.

Jared Stafford told me startForeground() improves responsiveness on
Nougat.  There had been a comment suggesting startForeground(), but I
never got around to trying it because it has worked "well enough".  With
Nougat, though, there is a definite tendency for SimpleSSHD to be
non-responsive.  I'm not sure exactly what its cause is, but the symptom
I notice most frequently is that the first ssh connection after a while
will be delayed "a long time" - on the order of 10-30 seconds, or maybe
indefinitely sometimes.  Oddly, a second connection can sometimes get
through undelayed, even before the first connection does.  It is as if
the fork() of process for the new connection is where the delay is, not
in the listen() call.

That's not overall too surprising, Nougat is a lot harsher about
background processes as part of a Google push to reduce power consumption
on idle devices.

Another concern is related to a change back in July - sometimes the
system will kill the sshd process for no good reason (maybe because they
removed it from the recent apps list).  The remedy I settled on was to
monitor for the sshd process dying from within the regular
Android-managed process, and restart it.  I guess I didn't write down
where in the documentation I found it, but Android seems expressly
antagonistic to non-system-managed processes.  If they ever get more
hard-assed about that, this whole idea goes out the window.

Anyways, I implemented startForeground(), and I am unhappy that it
requires a Notification. At API 7 (Android 2.1) those are really
primitive.  For example, the PRIORITY_MIN behavior which will sometimes
hide the notification is added in API 16 (Android 4.1).  So, I've got it
with this stupid old-style notification, and it really doesn't look good,
and it is always present.  That is not awesome.

On the other hand, it's easy to block the notifications, and a few people
have expressed an interest in a notification.

It doesn't seem worth it to me to upgrade to a newer API just for the
better notifications...  On the other hand, Google Play shows that I have
862 users:
    Android 7+  :  3.13%
    Android 6+  : 37.47%
    Android 5+  : 67.86%
    Android 4.1+: 96.62%
    Android 4.0+: 98.01%

The oldest reported version is Android 2.3.

So, there are a few people on very old versions, but actually SimpleSSHD
is used on newer devices than the average app, which is the reverse of my
typical trend.  So a few people would be negatively impacted, but not a
very large number.  I could switch to multi-APK mode so that legacy users
are just stuck with an unsupported back-version, which is probably what
they truly want anyways..

Anyways, I'm gonna use it with startForeground() and the notifications
disabled for a while, and if I find it to be an improvement then I'll
just make it so clicking on the notification goes to the app, update the
doc, and then publish it.

...

Looking up doze mode details for improving a different app, I stumbled
across this:

   https://www.bignerdranch.com/blog/diving-into-doze-mode-for-developers/

   One last case which is not mentioned in the Android documentation is
   an app using a foreground service. Any process using a foreground
   service is exempt from Doze Mode effects, which is key for
   long-running applications that you may want to run even if the user
   has put their phone down for a long period of time -- like a music app,
   for example.

   There is, however, a bug in Android Marshmallow which requires a
   foreground service to be in a separate process from all other activity
   code, documented here. It is planned to be fixed in Nougat.

Not exactly confidence-inspiring.  And I haven't come across a
description of how exactly doze mode screws up SimpleSSHD, either.

It might be worthwhile to do the work to find out where things actually
get wedged, it might even be a flaw in Dropbear that makes it so
unreliable with Nougat.


October 16, 2016.

I've been using a foreground service for a while, and it has not caused
me any troubles.  The thing I have been most anxious about is that it
might disable doze mode entirely...but there has not been any noteworthy
overnight drain.

So, in case someone does have problems with it, I am making it an option
that defaults to enabled.

Looking through email, I haven't seen any, but I have the idea several
users have asked for a notification icon in the past.  And now that that
is finally implemented, I am curious about other things people have
requested that I have not been keen on.  And Jan Ondrej's requests come
to mind:

  o) setting to start the service automatically when the application is
     launched
  o) QUIT button that stops the service and the activity at once
  o) not allow wifi to power down when the activity is open

I'm really not crazy about integrating any kind of wakelock.  And having
two buttons still seems silly to me.  But with the notification there,
the idea that someone will micromanage whether the service is running or
not does not seem so far fetched.

So I'll go ahead and add "Start on Open" setting.

...

Looking at reviews on Google Play Store, I finally found a couple reviews
in the last couple months who would have enjoyed password login.  I'm
still pretty opposed to it because it seems like the usage model would be
to enable a default password, login with it to transfer an
authorized_keys file, and then disable the default password.  That is
fine, and would even be kind of convenient (I might use it), but it seems
more often than not the default password would be left enabled
indefinitely, which is not awesome.  And I can't imagine more than 1% of
users ever typing a strong password into their phones.

But I've thought of a compromise.  What if, if there is no
authorized_keys file, it accepts passwordless logins but with some sort
of obnoxious alert dialog sort of thing to interrupt the user?  That way,
you would be able to login once to copy the authorized keys file, and the
nuissance alert would be no big deal.  Then once the authorized keys
exists, no further action is necessary.

It's a pity there is no convenient way to interact between the Android
GUI thread and the sshd thread.

Ah-hah!  I've got the least effort idea in my head!  Under situations
that I'm not sure what they should be, it should generate a random
password and display it on the phone's screen.  Probably it should
generate it iff there is no authorized_keys file at all.  Since that
detection happens for each client connection, then probably all of this
logic should go in dropbear itself, and the display should come through
dropbear.err.

Then we might even want a UI way to delete authorized_keys, perhaps even
as a replacement for the current awkward UI.