Commit graph

422 commits

Author SHA1 Message Date
Eelco Dolstra 1511aa9f48 Allow remote builds without sending the derivation closure
Previously, to build a derivation remotely, we had to copy the entire
closure of the .drv file to the remote machine, even though we only
need the top-level derivation. This is very wasteful: the closure can
contain thousands of store paths, and in some Hydra use cases, include
source paths that are very large (e.g. Git/Mercurial checkouts).

So now there is a new operation, StoreAPI::buildDerivation(), that
performs a build from an in-memory representation of a derivation
(BasicDerivation) rather than from a on-disk .drv file. The only files
that need to be in the Nix store are the sources of the derivation
(drv.inputSrcs), and the needed output paths of the dependencies (as
described by drv.inputDrvs). "nix-store --serve" exposes this
interface.

Note that this is a privileged operation, because you can construct a
derivation that builds any store path whatsoever. Fixing this will
require changing the hashing scheme (i.e., the output paths should be
computed from the other fields in BasicDerivation, allowing them to be
verified without access to other derivations). However, this would be
quite nice because it would allow .drv-free building (e.g. "nix-env
-i" wouldn't have to write any .drv files to disk).

Fixes #173.
2015-07-17 17:57:40 +02:00
Eelco Dolstra dd48c06bb6 Typo 2015-07-02 00:30:16 +02:00
Ludovic Courtès 9aed117395 Preserve supplementary groups of build users
The following patch is an attempt to address this bug (see
<http://bugs.gnu.org/18994>) by preserving the supplementary groups of
build users in the build environment.

In practice, I would expect that supplementary groups would contain only
one or two groups: the build users group, and possibly the “kvm” group.

[Changed &at(0) to data() and removed tabs - Eelco]
2015-07-01 14:57:48 +02:00
Eelco Dolstra b65875f859 Export outputPaths function
This is useful for the new hydra-queue-runner.
2015-06-10 16:17:06 +02:00
Eelco Dolstra 7c4501886d Use std::vector::data() 2015-06-09 10:54:46 +02:00
Eelco Dolstra b64988bb35 Allow substitutes for builds that have preferLocalBuild set
Not substituting builds with "preferLocalBuild = true" was a bad idea,
because it didn't take the cost of dependencies into account. For
instance, if we can't substitute a fetchgit call, then we have to
download/build git and all its dependencies.

Partially reverts 5558652709 and adds a
new derivation attribute "allowSubstitutes" to specify whether a
derivation may be substituted.
2015-06-04 16:30:22 +02:00
Eelco Dolstra 07d7e7df84 Chown files created for passAsFile
Nixpkgs' writeTextAsFile does this:

  mv "$textPath" "$n"

Since $textPath was owned by root, if $textPath is on the same
filesystem as $n, $n will be owned as root. As a result, the build
result was rejected as having suspicious ownership.

http://hydra.nixos.org/build/22836807
2015-06-04 14:07:43 +02:00
Rok Garbas 000de699e9 cygwin: explicitly include required c headers 2015-05-13 09:37:12 +02:00
Shea Levy 71083f9e5e Don't try to map /bin/sh to a store path on non-Linux 2015-05-12 16:36:15 -04:00
Shea Levy 4d652875bd Add the pre-build hook.
This hook can be used to set system-specific per-derivation build
settings that don't fit into the derivation model and are too complex or
volatile to be hard-coded into nix. Currently, the pre-build hook can
only add chroot dirs/files through the interface, but it also has full
access to the chroot root.

The specific use case for this is systems where the operating system ABI
is more complex than just the kernel-support system calls. For example,
on OS X there is a set of system-provided frameworks that can reliably
be accessed by any program linked to them, no matter the version the
program is running on. Unfortunately, those frameworks do not
necessarily live in the same locations on each version of OS X, nor do
their dependencies, and thus nix needs to know the specific version of
OS X currently running in order to make those frameworks available. The
pre-build hook is a perfect mechanism for doing just that.
2015-04-18 16:56:02 -04:00
Shea Levy fd6774e285 Revert "Add the pre-build hook."
Going to reimplement differently.

This reverts commit 1e4a4a2e9f.
2015-04-18 14:59:58 -04:00
Shea Levy 1e4a4a2e9f Add the pre-build hook.
This hook can be used to set system specific per-derivation build
settings that don't fit into the derivation model and are too complex or
volatile to be hard-coded into nix. Currently, the pre-build hook can
only add chroot dirs/files.

The specific use case for this is systems where the operating system ABI
is more complex than just the kernel-supported system calls. For
example, on OS X there is a set of system-provided frameworks that can
reliably be accessed by any program linked to them, no matter the
version the program is running on. Unfortunately, those frameworks do
not necessarily live in the same locations on each version of OS X, nor
do their dependencies, and thus nix needs to know the specific version
of OS X currently running in order to make those frameworks available.
The pre-build hook is a perfect mechanism for doing just that.
2015-04-12 12:56:38 -04:00
Eelco Dolstra 1711679ea5 Revert /nix/store permission back to 01775
This broke NixOS VM tests.

Mostly reverts 27b7b94923,
5ce50cd99e,
afa433e58c.
2015-04-07 13:21:26 +02:00
Eelco Dolstra afa433e58c Chroot builds: Provide world-readable /nix/store
This was causing NixOS VM tests to fail mysteriously since
5ce50cd99e. Nscd could (sometimes) no
longer read /etc/hosts:

open("/etc/hosts", O_RDONLY|O_CLOEXEC) = -1 EACCES (Permission denied)

Probably there was some wacky interaction between the guest kernel and
the 9pfs implementation in QEMU.
2015-04-02 16:59:40 +02:00
Eelco Dolstra 5114a07d95 Improve setting the default chroot dirs 2015-03-24 11:57:46 +01:00
Eelco Dolstra fd89f97be9 Add the closure of store paths to the chroot
Thus, for example, to get /bin/sh in a chroot, you only need to
specify /bin/sh=${pkgs.bash}/bin/sh in build-chroot-dirs. The
dependencies of sh will be added automatically.
2015-03-24 11:52:34 +01:00
Eelco Dolstra 5ce50cd99e Tighten permissions on chroot directories 2015-03-24 11:35:53 +01:00
Eelco Dolstra 75ede65e3d Don't use vfork() before clone()
I'm seeing hangs in Glibc's setxid_mark_thread() again. This is
probably because the use of an intermediate process to make clone()
safe from a multi-threaded program (see
524f89f139) is defeated by the use of
vfork(), since the intermediate process will have a copy of Glibc's
threading data structures due to the vfork(). So use a regular fork()
again.
2015-03-04 15:13:10 +01:00
Shea Levy c2699be93b Merge branch 'allow-system-library' of git://github.com/copumpkin/nix
Make the default impure prefix include all of /System/Library
2015-03-03 15:01:09 -05:00
Dan Peebles 336c4270c6 Make the default impure prefix (not actual allowed impurities!) include all of /System/Library, since we also want PrivateFrameworks from there and (briefly) TextEncodings, and who knows what else. Yay infectious impurities? 2015-03-02 23:01:24 -05:00
Dan Peebles 66d612f1da Allow local networking in the darwin sandbox to appease tests 2015-03-02 22:55:42 -05:00
Eelco Dolstra 885bebf13b More graceful fallback for chroots on Linux < 2.13 2015-02-23 15:54:31 +01:00
Eelco Dolstra 99897f6979 Use chroots for all derivations
If ‘build-use-chroot’ is set to ‘true’, fixed-output derivations are
now also chrooted. However, unlike normal derivations, they don't get
a private network namespace, so they can still access the
network. Also, the use of the ‘__noChroot’ derivation attribute is
no longer allowed.

Setting ‘build-use-chroot’ to ‘relaxed’ gives the old behaviour.
2015-02-23 15:54:31 +01:00
Eelco Dolstra bd91064150 Use $<attr>Path instead of $<attr> for passAsFile 2015-02-17 16:42:54 +01:00
Eelco Dolstra a70d275f3d Allow passing attributes via files instead of environment variables
Closes #473.
2015-02-17 14:42:15 +01:00
Harald van Dijk 5451b8db9d Use pivot_root in addition to chroot when possible
chroot only changes the process root directory, not the mount namespace root
directory, and it is well-known that any process with chroot capability can
break out of a chroot "jail". By using pivot_root as well, and unmounting the
original mount namespace root directory, breaking out becomes impossible.

Non-root processes typically have no ability to use chroot() anyway, but they
can gain that capability through the use of clone() or unshare(). For security
reasons, these syscalls are limited in functionality when used inside a normal
chroot environment. Using pivot_root() this way does allow those syscalls to be
put to their full use.
2015-02-16 12:18:19 +01:00
Shea Levy d66d9e8425 Require linux 3.13 or later for chroot
Fixes #453
2015-02-04 18:15:56 +01:00
Daniel Peebles f46e329a13 Make inputs writeable in the sandbox (builds still can’t actually write due to user permissions) 2015-01-18 23:25:29 -05:00
Shea Levy 79ca503332 Allow using /bin and /usr/bin as impure prefixes on non-darwin by default
These directories are generally world-readable anyway, and give us the two
most common linux impurities (env and sh)
2015-01-13 15:41:46 +01:00
Eelco Dolstra fcf57aad27 SysError -> Error 2015-01-13 11:17:56 +01:00
Eelco Dolstra 100961e370 Don't resolve symlinks while checking __impureHostDeps
Since these come from untrusted users, we shouldn't do any I/O on them
before we've checked that they're in an allowed prefix.
2015-01-13 11:16:32 +01:00
Daniel Peebles f1151a3373 Add basic Apple sandbox support 2015-01-12 12:00:01 +01:00
Eelco Dolstra 27b7b94923 Set /nix/store permission to 1737
I.e., not readable to the nixbld group. This improves purity a bit for
non-chroot builds, because it prevents a builder from enumerating
store paths (i.e. it can only access paths it knows about).
2015-01-08 16:39:07 +01:00
Eelco Dolstra df05f49dcd Fix building on Darwin
Fixes #433.
2015-01-06 10:49:44 +01:00
Eelco Dolstra bd0f362d2f Revive running builds in a PID namespace 2014-12-23 17:25:06 +01:00
Eelco Dolstra b77037b8fd Silence some warnings on GCC 4.9 2014-12-12 17:14:28 +01:00
Eelco Dolstra f52b6c944e Fix some memory leaks 2014-12-12 15:01:16 +01:00
Eelco Dolstra 28f22b4653 Ensure we're writing to stderr in the builder
http://hydra.nixos.org/build/17862041
2014-12-12 14:35:44 +01:00
Eelco Dolstra 4acca1a5b9 Remove chatty message
This broke building with "-vv", because the builder is not allowed to
write to stderr at this point.
2014-12-12 13:43:59 +01:00
Eelco Dolstra ad790022fd Doh 2014-12-12 13:41:00 +01:00
Eelco Dolstra 851b47bd7d Don't do vfork in conjunction with setuid 2014-12-10 18:01:01 +01:00
Eelco Dolstra b5ed5b6e66 Rename function 2014-12-10 17:25:38 +01:00
Eelco Dolstra b7b6e3ddec Build derivations in a more predictable order
Derivations are now built in order of derivation name, so a package
named "aardvark" is built before "baboon".

Fixes #399.
2014-11-24 16:50:46 +01:00
Eelco Dolstra 9e3389c337 Don't create unnecessary substitution goals for derivations 2014-11-24 16:50:46 +01:00
Eelco Dolstra 35aad73bb6 Fix message 2014-11-17 01:00:39 +01:00
Eelco Dolstra 8cfe939b0f Don't use ADDR_LIMIT_3GB
This gives 32-bit builds on x86_64-linux more memory.
2014-11-14 14:16:20 +01:00
Eelco Dolstra bab8d9b52a Make ~DerivationGoal more reliable 2014-11-12 11:35:53 +01:00
Eelco Dolstra 11aef17a77 Remove comments claiming we use a private PID namespace
This is no longer the case since
524f89f139.
2014-10-29 15:49:34 +01:00
Eelco Dolstra d98bfcbf81 On Linux, disable address space randomization 2014-09-17 17:21:13 +02:00
Eelco Dolstra d37d012774 Settings: Add bool get() 2014-09-17 15:18:13 +02:00
Eelco Dolstra b72e93bca8 Add disallowedReferences / disallowedRequisites
For the "stdenv accidentally referring to bootstrap-tools", it seems
easier to specify the path that we don't want to depend on, e.g.

  disallowedRequisites = [ bootstrapTools ];
2014-08-28 18:57:13 +02:00
Gergely Risko fd61069a42 Introduce allowedRequisites feature 2014-08-28 18:23:55 +02:00
Joel Taylor b224ac1520 fix disappearing bash arguments 2014-08-21 23:17:50 +02:00
Eelco Dolstra 524f89f139 Use unshare() instead of clone()
It turns out that using clone() to start a child process is unsafe in
a multithreaded program. It can cause the initialisation of a build
child process to hang in setgroups(), as seen several times in the
build farm:

The reason is that Glibc thinks that the other threads of the parent
exist in the child, so in setxid_mark_thread() it tries to get a futex
that has been acquired by another thread just before the clone(). With
fork(), Glibc runs pthread_atfork() handlers that take care of this
(in particular, __reclaim_stacks()). But clone() doesn't do that.

Fortunately, we can use fork()+unshare() instead of clone() to set up
private namespaces.

See also https://www.mail-archive.com/lxc-devel@lists.linuxcontainers.org/msg03434.html.
2014-08-21 14:08:09 +02:00
Eelco Dolstra 11849a320e Use proper quotes everywhere 2014-08-20 18:03:48 +02:00
Eelco Dolstra 373fad75e1 Add some color 2014-08-20 16:50:17 +02:00
Eelco Dolstra 954188af27 Filter Nix-specific ANSI escape sequences from stderr
The Nixpkgs stdenv prints some custom escape sequences to denote
nesting and stuff like that. Most terminals (e.g. xterm, konsole)
ignore them, but some do not (e.g. xfce4-terminal). So for the benefit
of the latter, filter them out.
2014-08-20 14:30:01 +02:00
Eelco Dolstra 029424d17d Make hook shutdown more reliable 2014-08-19 17:47:30 +02:00
Eelco Dolstra 632f989b51 Doh 2014-08-18 11:35:50 +02:00
Eelco Dolstra c160ead82f Reduce verbosity 2014-08-17 19:11:50 +02:00
Eelco Dolstra 42e9ad8fd1 Propagate remote timeouts properly 2014-08-17 19:09:03 +02:00
Eelco Dolstra daccd68999 Refactor 2014-08-04 18:02:29 +02:00
Eelco Dolstra d5a076c36f Add option ‘build-extra-chroot-dirs’
This is useful for extending (rather than overriding) the default set
of chroot paths.
2014-08-04 18:00:00 +02:00
Eelco Dolstra 4d73e2e893 Get rid of "killing <pid>" message for unused build hooks 2014-08-04 17:27:45 +02:00
Eelco Dolstra 51485dcda2 Remove ugly hack for detecting build environment setup errors 2014-08-01 19:38:21 +02:00
Eelco Dolstra 4c4b219c07 Call commonChildInit() before doing chroot init
This ensures that daemon clients see error messages from the chroot
setup.
2014-08-01 19:29:03 +02:00
Eelco Dolstra 50dc1f5b71 Restore default SIGPIPE handler before invoking ‘man’
Fixes NixOS/nixpkgs#3410.
2014-07-31 10:31:17 +02:00
Eelco Dolstra ee3c5d7916 Revert old useBuildHook behaviour 2014-07-19 02:25:47 +02:00
Eelco Dolstra e196eecbe6 Allow $NIX_BUILD_HOOK to be relative to Nix libexec directory 2014-07-11 13:55:06 +02:00
Eelco Dolstra 43b64f5038 Remove tabs 2014-07-10 17:32:21 +02:00
Eelco Dolstra 8e9140cfde Refactoring: Move all fork handling into a higher-order function
C++11 lambdas ftw.
2014-07-10 16:58:09 +02:00
Eelco Dolstra 7911e4c27a Remove maybeVfork 2014-07-10 13:35:44 +02:00
Danny Wilson ae6b631dc4 Fix compile errors on Illumos 2014-04-03 17:39:57 +02:00
Ludovic Courtès e7720aa10a Make sure /dev/pts/ptmx is world-writable
While running Python 3’s test suite, we noticed that on some systems
/dev/pts/ptmx is created with permissions 0 (that’s the case with my
Nixpkgs-originating 3.0.43 kernel, but someone with a Debian-originating
3.10-3 reported not having this problem.)

There’s still the problem that people without
CONFIG_DEVPTS_MULTIPLE_INSTANCES=y are screwed (as noted in build.cc),
but I don’t see how we could work around it.
2014-04-03 13:42:45 +02:00
Eelco Dolstra 1c2550a2ae boost::shared_ptr -> std::shared_ptr 2014-03-30 00:49:23 +01:00
Eelco Dolstra acb8facbbc Fix potential segfault in waitForInput()
Since the addition of build-max-log-size, a call to
handleChildOutput() can result in cancellation of a goal.  This
invalidated the "j" iterator in the waitForInput() loop, even though
it was still used afterwards.  Likewise for the maxSilentTime
handling.

Probably fixes #231.  At least it gets rid of the valgrind warnings.
2014-03-29 22:14:11 +01:00
Eelco Dolstra 90dc50b07c restoreSIGPIPE(): Fill in sa_mask
Issue #231.
2014-03-29 20:20:14 +01:00
Ludovic Courtès 24cb65efc3 Make /dev/kvm optional
The daemon now creates /dev deterministically (thanks!).  However, it
expects /dev/kvm to be present.

The patch below restricts that requirement (1) to Linux-based systems,
and (2) to systems where /dev/kvm already exists.

I’m not sure about the way to handle (2).  We could special-case
/dev/kvm and create it (instead of bind-mounting it) in the chroot, so
it’s always available; however, it wouldn’t help much since most likely,
if /dev/kvm missing, then KVM support is missing.
2014-03-21 17:27:19 +01:00
Eelco Dolstra 1017bd68ea Set up a private /dev/pts in the chroot 2014-02-27 23:35:23 +01:00
Eelco Dolstra 3fd01b171a Set up a minimal /dev in chroots
Not bind-mounting the /dev from the host also solves the problem with
/dev/shm being a symlink to something not in the chroot.
2014-02-27 23:17:53 +01:00
Eelco Dolstra 29cde917fe Fix deadlock in SubstitutionGoal
We were relying on SubstitutionGoal's destructor releasing the lock,
but if a goal is a top-level goal, the destructor won't run in a
timely manner since its reference count won't drop to zero.  So
release it explicitly.

Fixes #178.
2014-02-27 13:34:13 +01:00
Eelco Dolstra 8129cf33d9 Slight simplification 2014-02-18 10:46:30 +01:00
Eelco Dolstra 1aa19b24b2 Add a flag ‘--check’ to verify build determinism
The flag ‘--check’ to ‘nix-store -r’ or ‘nix-build’ will cause Nix to
redo the build of a derivation whose output paths are already valid.
If the new output differs from the original output, an error is
printed.  This makes it easier to test if a build is deterministic.
(Obviously this cannot catch all sources of non-determinism, but it
catches the most common one, namely the current time.)

For example:

  $ nix-build '<nixpkgs>' -A patchelf
  ...
  $ nix-build '<nixpkgs>' -A patchelf --check
  error: derivation `/nix/store/1ipvxsdnbhl1rw6siz6x92s7sc8nwkkb-patchelf-0.6' may not be deterministic: hash mismatch in output `/nix/store/4pc1dmw5xkwmc6q3gdc9i5nbjl4dkjpp-patchelf-0.6.drv'

The --check build fails if not all outputs are valid.  Thus the first
call to nix-build is necessary to ensure that all outputs are valid.

The current outputs are left untouched: the new outputs are either put
in a chroot or diverted to a different location in the store using
hash rewriting.
2014-02-18 01:01:14 +01:00
Eelco Dolstra b6def5b542 Make --repair work on Darwin
Mac OS X doesn't allow renaming a read-only directory.

http://hydra.nixos.org/build/9113895
2014-02-17 23:09:48 +01:00
Eelco Dolstra dfbcb7c403 Refactoring 2014-02-17 23:04:52 +01:00
Eelco Dolstra 71adb090f0 When using a build hook, only copy missing paths 2014-02-17 22:58:21 +01:00
Eelco Dolstra 69fe6c58fa Move some code around
In particular, do replacing of valid paths during repair later.  This
prevents us from replacing a valid path after the build fails.
2014-02-17 22:25:15 +01:00
Eelco Dolstra 00d30496ca Heuristically detect if a build may have failed due to a full disk
This will allow Hydra to detect that a build should not be marked as
"permanently failed", allowing it to be retried later.
2014-02-17 14:15:56 +01:00
Eelco Dolstra 94f9c14d52 Fix some clang warnings 2014-01-21 18:29:55 +01:00
Eelco Dolstra 11cb4bfb25 Fix checking of NAR hashes
*headdesk*
*headdesk*
*headdesk*

So since commit 22144afa8d, Nix hasn't
actually checked whether the content of a downloaded NAR matches the
hash specified in the manifest / NAR info file.  Urghhh...
2014-01-08 17:35:49 +01:00
Domen Kožar 485f4740ee wording 2014-01-06 11:38:24 +01:00
Eelco Dolstra c5b8fe3151 Print a trace message if a build fails due to the platform being unknown 2013-12-05 14:31:57 -05:00
Eelco Dolstra a478e8a7bb Remove nix-setuid-helper
AFAIK, nobody uses it, it's not maintained, and it has no tests.
2013-11-14 11:57:37 +01:00
Eelco Dolstra b29d3f4aee Only show trace messages when tracing is enabled 2013-09-02 12:01:04 +02:00
Eelco Dolstra efe4289464 Add an option to limit the log output of builders
This is mostly useful for Hydra to deal with builders that get stuck
in an infinite loop writing data to stdout/stderr.
2013-09-02 11:58:18 +02:00
Gergely Risko c6c024ca6f Fix personality switching from x86_64 to i686
On Linux, Nix can build i686 packages even on x86_64 systems.  It's not
enough to recognize this situation by settings.thisSystem, we also have
to consult uname().  E.g. we can be running on a i686 Debian with an
amd64 kernel.  In that situation settings.thisSystem is i686-linux, but
we still need to change personality to i686 to make builds consistent.
2013-08-26 11:12:35 +02:00
Eelco Dolstra a583a2bc59 Run the daemon worker on the same CPU as the client
On a system with multiple CPUs, running Nix operations through the
daemon is significantly slower than "direct" mode:

$ NIX_REMOTE= nix-instantiate '<nixos>' -A system
real    0m0.974s
user    0m0.875s
sys     0m0.088s

$ NIX_REMOTE=daemon nix-instantiate '<nixos>' -A system
real    0m2.118s
user    0m1.463s
sys     0m0.218s

The main reason seems to be that the client and the worker get moved
to a different CPU after every call to the worker.  This patch adds a
hack to lock them to the same CPU.  With this, the overhead of going
through the daemon is very small:

$ NIX_REMOTE=daemon nix-instantiate '<nixos>' -A system
real    0m1.074s
user    0m0.809s
sys     0m0.098s
2013-08-07 14:02:04 +02:00
Eelco Dolstra a4921b8ceb Revert "build-remote.pl: Enforce timeouts locally"
This reverts commit 69b8f9980f.

The timeout should be enforced remotely.  Otherwise, if the garbage
collector is running either locally or remotely, if will block the
build or closure copying for some time.  If the garbage collector
takes too long, the build may time out, which is not what we want.
Also, on heavily loaded systems, copying large paths to and from the
remote machine can take a long time, also potentially resulting in a
timeout.
2013-07-18 12:52:29 +02:00
Shea Levy 16591eb3cc Allow bind-mounting regular files into the chroot
mount(2) with MS_BIND allows mounting a regular file on top of a regular
file, so there's no reason to only bind directories. This allows finer
control over just which files are and aren't included in the chroot
without having to build symlink trees or the like.

Signed-off-by: Shea Levy <shea@shealevy.com>
2013-07-15 16:01:33 +02:00
Eelco Dolstra 7ccd946407 Don't set $preferLocalBuild and $requiredSystemFeatures in builders
With C++ std::map, doing a comparison like ‘map["foo"] == ...’ has the
side-effect of adding a mapping from "foo" to the empty string if
"foo" doesn't exist in the map.  So we ended up setting some
environment variables by accident.
2013-06-20 18:07:27 +00:00