Commit graph

515 commits

Author SHA1 Message Date
Eelco Dolstra c2b0d8749f
exportReferencesGraph: Export more complete info in JSON format
This writes info about every path in the closure in the same format as
‘nix path-info --json’. Thus it also includes NAR hashes and sizes.

Example:

  [
    {
      "path": "/nix/store/10h6li26i7g6z3mdpvra09yyf10mmzdr-hello-2.10",
      "narHash": "sha256:0ckdc4z20kkmpqdilx0wl6cricxv90lh85xpv2qljppcmz6vzcxl",
      "narSize": 197648,
      "references": [
        "/nix/store/10h6li26i7g6z3mdpvra09yyf10mmzdr-hello-2.10",
        "/nix/store/27binbdy296qvjycdgr1535v8872vz3z-glibc-2.24"
      ],
      "closureSize": 20939776
    },
    {
      "path": "/nix/store/27binbdy296qvjycdgr1535v8872vz3z-glibc-2.24",
      "narHash": "sha256:1nfn3m3p98y1c0kd0brp80dn9n5mycwgrk183j17rajya0h7gax3",
      "narSize": 20742128,
      "references": [
        "/nix/store/27binbdy296qvjycdgr1535v8872vz3z-glibc-2.24"
      ],
      "closureSize": 20742128
    }
  ]

Fixes #1134.
2017-01-26 20:41:08 +01:00
Eelco Dolstra 6de33a9c67
Add support for passing structured data to builders
Previously, all derivation attributes had to be coerced into strings
so that they could be passed via the environment. This is lossy
(e.g. lists get flattened, necessitating configureFlags
vs. configureFlagsArray, of which the latter cannot be specified as an
attribute), doesn't support attribute sets at all, and has size
limitations (necessitating hacks like passAsFile).

This patch adds a new mode for passing attributes to builders, namely
encoded as a JSON file ".attrs.json" in the current directory of the
builder. This mode is activated via the special attribute

  __structuredAttrs = true;

(The idea is that one day we can set this in stdenv.mkDerivation.)

For example,

  stdenv.mkDerivation {
    __structuredAttrs = true;
    name = "foo";
    buildInputs = [ pkgs.hello pkgs.cowsay ];
    doCheck = true;
    hardening.format = false;
  }

results in a ".attrs.json" file containing (sans the indentation):

  {
    "buildInputs": [],
    "builder": "/nix/store/ygl61ycpr2vjqrx775l1r2mw1g2rb754-bash-4.3-p48/bin/bash",
    "configureFlags": [
      "--with-foo",
      "--with-bar=1 2"
    ],
    "doCheck": true,
    "hardening": {
      "format": false
    },
    "name": "foo",
    "nativeBuildInputs": [
      "/nix/store/10h6li26i7g6z3mdpvra09yyf10mmzdr-hello-2.10",
      "/nix/store/4jnvjin0r6wp6cv1hdm5jbkx3vinlcvk-cowsay-3.03"
    ],
    "propagatedBuildInputs": [],
    "propagatedNativeBuildInputs": [],
    "stdenv": "/nix/store/f3hw3p8armnzy6xhd4h8s7anfjrs15n2-stdenv",
    "system": "x86_64-linux"
  }

"passAsFile" is ignored in this mode because it's not needed - large
strings are included directly in the JSON representation.

It is up to the builder to do something with the JSON
representation. For example, in bash-based builders, lists/attrsets of
string values could be mapped to bash (associative) arrays.
2017-01-26 20:40:33 +01:00
Eelco Dolstra 951357e5fb
UserLock: Fix multi-threaded access to a global variable 2017-01-26 20:40:33 +01:00
Eelco Dolstra c0f2f4eeef
UserLock: Make more RAII-ish 2017-01-26 20:40:33 +01:00
Eelco Dolstra a529c740d2
Moving more code out of DerivationGoal::startBuilder() 2017-01-26 20:40:33 +01:00
Eelco Dolstra 4425a5c547
Move exportReferencesGraph into a separate method
startBuilder() is getting rather obese.
2017-01-26 20:40:32 +01:00
Eelco Dolstra 21948deed9
Kill builds when we get EOF on the log FD
This closes a long-time bug that allowed builds to hang Nix
indefinitely (regardless of timeouts) simply by doing

  exec > /dev/null 2>&1; while true; do true; done

Now, on EOF, we just send SIGKILL to the child to make sure it's
really gone.
2017-01-19 17:16:14 +01:00
Eelco Dolstra 2579e32c2b
Use std::unique_ptr for HookInstance 2017-01-19 17:06:04 +01:00
Eelco Dolstra 7adb986e35 Merge pull request #1139 from Mic92/master
Simplify remouting with MS_PRIVATE in sandbox build
2017-01-09 16:46:59 +01:00
Eelco Dolstra b6b142b4b1
Provide /var/run/nscd/socket in the sandbox
Otherwise sandbox builds can fail, e.g.

  $ NIX_REMOTE=local?root=/tmp/nix nix-build '<nixpkgs>' -A hello --option build-use-substitutes false
  ...
  downloading ‘http://ftpmirror.gnu.org/bash/bash-4.3-patches/bash43-047’...
  error: unable to download ‘http://ftpmirror.gnu.org/bash/bash-4.3-patches/bash43-047’: Couldn't resolve host name (6)
2017-01-02 14:46:37 +01:00
Eelco Dolstra 786ee585b8
Add comment 2016-12-22 17:39:49 +01:00
Eelco Dolstra 3a4bd320c2
Revert "Merge branch 'seccomp' of https://github.com/aszlig/nix"
This reverts commit 9f3f2e21ed, reversing
changes made to 47f587700d.
2016-12-19 11:52:57 +01:00
Eelco Dolstra 05862209de
Revert "Give root a valid home directory"
This reverts commit ec7d498b72.
2016-12-19 11:49:03 +01:00
Eelco Dolstra ec7d498b72
Give root a valid home directory
Some programs barf if the current user has a non-writable home
directory, e.g. http://hydra.nixos.org/build/44818144.
2016-12-15 15:56:08 +01:00
Eelco Dolstra 9f3f2e21ed
Merge branch 'seccomp' of https://github.com/aszlig/nix 2016-12-15 12:04:45 +01:00
Eelco Dolstra b30d1e7ada
Don't delete .check directories of running builds
We need to keep them around for diffoscope.
2016-12-08 21:38:58 +01:00
Eelco Dolstra e629a17cc1
Fix build
http://hydra.nixos.org/build/44628517
2016-12-08 20:36:14 +01:00
Eelco Dolstra 9a313469a4
Add a hook to run diffoscope when non-determinism is detected 2016-12-07 17:57:35 +01:00
Eelco Dolstra b07060688a
Keep track of the exact build start/stop times 2016-12-07 16:09:38 +01:00
Eelco Dolstra dadfddfa7c
Bail out early when non-determinism is detected 2016-12-07 15:31:18 +01:00
Eelco Dolstra 8bdf83f936
Add an option to make non-determinism non-fatal
That is, when build-repeat > 0, and the output of two rounds differ,
then print a warning rather than fail the build. This is primarily to
let Hydra check reproducibility of all packages.
2016-12-07 13:16:06 +01:00
Eelco Dolstra ceeedb58d2
Use a steady clock for timeouts
Fixes #1146.
2016-12-06 21:58:04 +01:00
Eelco Dolstra 7a3e7d0e61
nix-store --serve: Suppress log output on stderr when repeating a build 2016-12-06 17:43:39 +01:00
Eelco Dolstra 215b70f51e
Revert "Get rid of unicode quotes (#1140)"
This reverts commit f78126bfd6. There
really is no need for such a massive change...
2016-11-26 00:38:01 +01:00
Guillaume Maudoux f78126bfd6 Get rid of unicode quotes (#1140) 2016-11-25 15:48:27 +01:00
Jörg Thalheim 855abd85d8
Simplify remouting with MS_PRIVATE in sandbox build
also fix race condition if mounts are added after mountinfo is read.
2016-11-25 00:15:39 +01:00
Shea Levy 8bf378e999 Update darwin build for optional sandbox paths
Fixes #1132
2016-11-17 08:06:32 -05:00
aszlig 4e1a2cd537
seccomp: Forge return values for *chown32
These syscalls are only available in 32bit architectures, but libseccomp
should handle them correctly even if we're on native architectures that
do not have these syscalls.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2016-11-16 17:29:36 +01:00
aszlig ed64976cec
seccomp: Forge return codes for POSIX ACL syscalls
Commands such as "cp -p" also use fsetxattr() in addition to fchown(),
so we need to make sure these syscalls always return successful as well
in order to avoid nasty "Invalid value" errors.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2016-11-16 17:29:14 +01:00
aszlig b90a435332
libstore/build: Forge chown() to return success
What we basically want is a seccomp mode 2 BPF program like this but for
every architecture:

  BPF_STMT(BPF_LD+BPF_W+BPF_ABS, offsetof(struct seccomp_data, nr)),
  BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, __NR_chown, 4, 0),
  BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, __NR_fchown, 3, 0),
  BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, __NR_fchownat, 2, 0),
  BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, __NR_lchown, 1, 0),
  BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_ALLOW),
  BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_ERRNO)

However, on 32 bit architectures we do have chown32, lchown32 and
fchown32, so we'd need to add all the architecture blurb which
libseccomp handles for us.

So we only need to make sure that we add the 32bit seccomp arch while
we're on x86_64 and otherwise we just stay at the native architecture
which was set during seccomp_init(), which more or less replicates
setting 32bit personality during runChild().

The FORCE_SUCCESS() macro here could be a bit less ugly but I think
repeating the seccomp_rule_add() all over the place is way uglier.

Another way would have been to create a vector of syscalls to iterate
over, but that would make error messages uglier because we can either
only print the (libseccomp-internal) syscall number or use
seccomp_syscall_resolve_num_arch() to get the name or even make the
vector a pair number/name, essentially duplicating everything again.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2016-11-16 16:48:30 +01:00
aszlig e8838713df
Run builds as root in user namespace again
This reverts commit ff0c0b645c.

We're going to use seccomp to allow "cp -p" and force chown-related
syscalls to always return 0.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2016-11-16 16:48:08 +01:00
Eelco Dolstra 10ae8fabf1 buildPaths(): Handle ecIncompleteClosure
buildPaths() on a non-derivation would incorrectly not throw an error
if the path didn't have a substitute.
2016-11-14 15:00:17 +01:00
Eelco Dolstra dd77f7d593 Store::computeFSClosure(): Support a set of paths
This way, callers can exploits the parallelism of computeFSClosure()
when they have multiple paths that they need the (combined) closure of.
2016-11-10 17:45:04 +01:00
Eelco Dolstra a83b10f84c Typo 2016-11-08 20:19:02 +01:00
Eelco Dolstra 18b7363a69 Support optional sandbox paths
For example, you can now set

  build-sandbox-paths = /dev/nvidiactl?

to specify that /dev/nvidiactl should only be mounted in the sandbox
if it exists in the host filesystem. This is useful e.g. for EC2
images that should support both CUDA and non-CUDA instances.
2016-10-31 17:09:52 +01:00
Eelco Dolstra 5c0bd51d49 Fix build log output in nix-store --serve 2016-10-26 17:13:38 +02:00
Eelco Dolstra ae8884b949 Fix /dev/ptmx in sandboxes
This was broken since ff0c0b645c. Since
I can't figure out how to mount a devpts instance in the sandbox,
let's just bind-mount the host devpts.
2016-10-14 15:42:13 +02:00
Eelco Dolstra 818ab58cc6 Add sandbox-dev-shm-size option
Fixes #1069.
2016-09-21 16:54:53 +02:00
Eelco Dolstra c55bf085eb printMsg(lvlError, ...) -> printError(...) etc. 2016-09-21 16:54:53 +02:00
Eelco Dolstra 4036185cb4 Some notational convenience for formatting strings
We can now write

  throw Error("file '%s' not found", path);

instead of

  throw Error(format("file '%s' not found") % path);

and similarly

  printError("file '%s' not found", path);

instead of

  printMsg(lvlError, format("file '%s' not found") % path);
2016-09-21 16:54:53 +02:00
Eelco Dolstra ff0c0b645c Don't run builds as root in the user namespace
This largely reverts c68e5913c7. Running
builds as root breaks "cp -p", since when running as root, "cp -p"
assumes that it can succesfully chown() files. But that's not actually
the case since the user namespace doesn't provide a complete uid
mapping. So it barfs with a fatal error message ("cp: failed to
preserve ownership for 'foo': Invalid argument").
2016-09-08 18:16:23 +02:00
Eelco Dolstra e4bdd49022 Give a better error message if a path in build-sandbox-paths is invalid 2016-09-08 17:30:25 +02:00
Eelco Dolstra 0babc4ff31 Fix assertion failure after early build setup failure 2016-09-08 17:29:50 +02:00
Eelco Dolstra 97b1af1cbe Fix --fallback
This fixes an assertion failure in "assert(goal);" in
Worker::waitForInput() after a substitution goal is cancelled by the
termination of another goal. The problem was the line

  //worker.childTerminated(shared_from_this()); // FIXME

in the SubstitutionGoal destructor. This was disabled because
shared_from_this() obviously doesn't work from a destructor. So we now
use a real pointer for object identity.
2016-08-30 15:45:39 +02:00
Eelco Dolstra 22d6e31fc6 Add a mechanism for derivation attributes to reference the derivation's outputs
For example, you can now say:

  configureFlags = "--prefix=${placeholder "out"} --includedir=${placeholder "dev"}";

The strings returned by the ‘placeholder’ builtin are replaced at
build time by the actual store paths corresponding to the specified
outputs.

Previously, you had to work around the inability to self-reference by doing stuff like:

  preConfigure = ''
    configureFlags+=" --prefix $out --includedir=$dev"
  '';

or rely on ad-hoc variable interpolation semantics in Autoconf or Make
(e.g. --prefix=\$(out)), which doesn't always work.
2016-08-17 17:19:32 +02:00
Eelco Dolstra d961c29c9c Mark content-addressed paths in the Nix database and in .narinfo
This allows such paths to be imported without signatures.
2016-08-10 18:05:35 +02:00
Eelco Dolstra 3eb6217508 Add a "root" parameter to local stores
This makes it easier to create a diverted store, i.e.

  NIX_REMOTE="local?root=/tmp/root"

instead of

  NIX_REMOTE="local?real=/tmp/root/nix/store&state=/tmp/root/nix/var/nix" NIX_LOG_DIR=/tmp/root/nix/var/log
2016-08-10 18:05:35 +02:00
Eelco Dolstra ee22a91ab8 makeFixedOutputPath(): Drop superfluous HashType argument 2016-07-26 21:25:52 +02:00
Shea Levy 18b0808475 Respect --keep-going when a substituter fails.
Fixes #977
2016-07-23 09:16:08 -04:00
Shea Levy cb5e7254b6 Modernize AutoCloseFD 2016-07-11 15:44:44 -04:00