caps: check uid and euid

When we are running inside of a user namespace getuid() will return a
non-zero uid. So let's check euid as well to make sure we correctly drop
capabilities
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: ondra <ondrak@localhost.localdomain>

Signed-off-by: Ondrej Kubik <ondrej.kubik@canonical.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

s/strlcpy/strncpy/g
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

custom backport
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

shutdown() affects sockets even across forked processes. The
attached child process doesn't have any interest in using the
IPC socket, so just close it in the child process and let the
intermediate process handle shutting it down.

This fixes a bug seen with lxc exec in crbug.com/884244
Signed-off-by: Stephen Barber <smbarber@chromium.org>

Fail the build if --enable-thread-safety is passed and the environment cannot
guarantee thread-safety.
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Currently for section headings we use fourth level markdown heading
level (####).  We do not have levels two or three.

We can use standard incremental levels for heading adornments i.e

1) =========
2) ##
3) ###
ect.

Since this document is likely referenced by maintainers when guiding new
contributors it can save maintainer time to be able to quickly reference
a section in the coding stlye guide.  If we add numbers to each heading
(like the kernel stlye guide) then maintainers can say:

   Nice patch, please see section 3 of the coding style guide and ...

So, this patch makes two changes

- Use incremental level heading adornments
- Add a number to each section heading
Signed-off-by: Tobin C. Harding <me@tobin.cc>

Heading uses only 3 level header (###) but the rest of the file uses
four (####).  We should be uniform.

Use uniform heading level in line with the rest of the file.
Signed-off-by: Tobin C. Harding <me@tobin.cc>

We explicitly state that LXC uses coding style based on Linux kernel
style.  It is redundant to then document obvious, and well known, kernel
style rules.  Identifier names certainly fall into this category as does
usage of braces.

Remove sections implied by 'kernel style'.  Naming conventions and brace
placement conventions.
Signed-off-by: Tobin C. Harding <me@tobin.cc>

Current explanation of rules around usage of 'extern' are overly
verbose.  It is not necessary to state that functions should be declared
in header files, the compiler already enforces this.  These rules are
simple, they are better described with simple statements.  An example is
not necessary for such simple rules and serves only to make the document
longer.

Use two simple statements describing the rules that govern function
declaration and the usage of the 'extern' keyword.
Signed-off-by: Tobin C. Harding <me@tobin.cc>

Fix minor grammatical issue.
Signed-off-by: Tobin C. Harding <me@tobin.cc>

Currently the coding style guide does not mention that we use kernel
coding style as a base style for LXC.  We have just linked to
CODING_STLYE.md from CONTRIBUTING (which mentions use of kernel coding
style).  We can increase documentation congruence and completeness by
mentioning kernel coding style guide in the introduction to our style
guide.

Add heading and introduction to coding style guide informing readers
that we follow kernel coding style as a base before explaining our style
additions.
Signed-off-by: Tobin C. Harding <me@tobin.cc>

Currently the 'Coding Style' section mentions only the kernel coding
style.  We have additions on top on this outlined in CODING_STYLE.md.
We should direct readers to this document as well as the kernel docs.

Direct readers to CODING_STLYE.md in the 'Coding Style' section.
Signed-off-by: Tobin C. Harding <me@tobin.cc>

Currently we link to a URL for v4.10 of the kernel docs.  Since we
already mention the kernel tree we should link to the _latest_ kernel
docs online instead of a fixed past version.

Link to latest online kernel docs tracking the mainline instead of past
fixed version.
Signed-off-by: Tobin C. Harding <me@tobin.cc>

Kernel coding style guide filename is stale, this file has been renamed
in the kernel tree.  While this file still exists we should use the new
filename.

Update reference to kernel coding style guide to use the new file name.
Signed-off-by: Tobin C. Harding <me@tobin.cc>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Serge Hallyn <serge@hallyn.com>

String not null terminated
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

We don't want to link caps.{c,h} against utils.{c,h} for the sake of our static
builds init.lxc.static. This means lxc_write_nointr() will not be available. So
handle it EINTR.
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

We don't want to link log.{c,h} against utils.{c,h} for the sake of our static
builds init.lxc.static. This means lxc_write_nointr() will not be available. So
handle it EINTR.
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

It's a valid case to call nla_put() with NULL data and 0 len. It's done e.g. in
the nla_put_attr().

There has to be a check for data in nla_put() as passing NULL to the memcpy()
is not allowed. Even if length is 0, both pointers have to be valid.

For a reference see C99 standard (7.21.1/2), it says: "pointer arguments on
such a call shall still have valid values".
Reported-by: Daniel Gimpelevich <daniel@gimpelevich.san-francisco.ca.us>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
[christian.brauner@ubuntu.com: adapted commit message]
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>