Commit graph

26 commits

Author SHA1 Message Date
Eelco Dolstra 6f0d050324 * Tricky: child processes should not send data to the client since
that might mess up the protocol.  And besides, the socket file
  descriptor is probably closed.
2006-12-05 18:21:16 +00:00
Eelco Dolstra 8d1854c3f1 * Oops! In daemon mode, we can't run as root either if build-users is empty. 2006-12-05 17:44:19 +00:00
Eelco Dolstra 99655245ae * Use an explicit handler for SIGCHLD, since SIG_IGN doesn't do the
right thing on FreeBSD 4 (it leaves zombies).
2006-12-05 17:21:42 +00:00
Eelco Dolstra c808e6252f * Ugly hack to handle spurious SIGPOLLs. 2006-12-05 15:36:31 +00:00
Eelco Dolstra fd4a9db91f * Some renaming. 2006-12-05 14:15:51 +00:00
Eelco Dolstra a9c4f66cfb * Allow unprivileged users to run the garbage collector and to do
`nix-store --delete'.  But unprivileged users are not allowed to
  ignore liveness.
* `nix-store --delete --ignore-liveness': ignore the runtime roots as
  well.
2006-12-05 02:18:46 +00:00
Eelco Dolstra 29cf434a35 * The determination of the root set should be made by the privileged
process, so forward the operation.
* Spam the user about GC misconfigurations (NIX-71).
* findRoots: skip all roots that are unreadable - the warnings with
  which we spam the user should be enough.
2006-12-05 01:31:45 +00:00
Eelco Dolstra 74033a844f * Add indirect root registration to the protocol so that unprivileged
processes can register indirect roots.  Of course, there is still
  the problem that the garbage collector can only read the targets of
  the indirect roots when it's running as root...
2006-12-04 23:29:16 +00:00
Eelco Dolstra 0d40f6d7bb * Not every OS knows about SIGPOLL. 2006-12-04 22:58:44 +00:00
Eelco Dolstra 40c3529909 * Handle exceptions and stderr for all protocol functions.
* SIGIO -> SIGPOLL (POSIX calls it that).
* Use sigaction instead of signal to register the SIGPOLL handler.
  Sigaction is better defined, and a handler registered with signal
  appears not to interrupt fcntl(..., F_SETLKW, ...), which is bad.
2006-12-04 17:55:14 +00:00
Eelco Dolstra 0130ef88ea * Daemon mode (`nix-worker --daemon'). Clients connect to the server
via the Unix domain socket in /nix/var/nix/daemon.socket.  The
  server forks a worker process per connection.
* readString(): use the heap, not the stack.
* Some protocol fixes.
2006-12-04 17:17:13 +00:00
Eelco Dolstra f4279bcde0 * Don't run setuid root when build-users is empty.
* Send startup errors to the client.
2006-12-03 16:25:19 +00:00
Eelco Dolstra a9f9241054 * Handle a subtle race condition: the client closing the socket
between the last worker read/write and the enabling of the signal
  handler.
2006-12-03 03:16:27 +00:00
Eelco Dolstra 3ed9e4ad9b * Some hardcore magic to handle asynchronous client disconnects.
The problem is that when we kill the client while the worker is
  building, and the builder is not writing anything to stderr, then
  the worker never notice that the socket is closed on the other side,
  so it just continues indefinitely.  The solution is to catch SIGIO,
  which is sent when the far side of the socket closes, and simulate
  an normal interruption.  Of course, SIGIO is also sent every time
  the client sends data over the socket, so we only enable the signal
  handler when we're not expecting any data...
2006-12-03 03:03:36 +00:00
Eelco Dolstra 7951c3c546 * Some hackery to propagate the worker's stderr and exceptions to the
client.
2006-12-03 02:08:13 +00:00
Eelco Dolstra 714fa24cfb * Run the worker in a separate session to prevent terminal signals
from interfering.
2006-12-03 00:52:27 +00:00
Eelco Dolstra e25fad691a * Move addTempRoot() to the store API, and add another function
syncWithGC() to allow clients to register GC roots without needing
  write access to the global roots directory or the GC lock.
2006-12-02 16:41:36 +00:00
Eelco Dolstra 8ba5d32769 * Remove queryPathHash().
* Help for nix-worker.
2006-12-02 14:27:24 +00:00
Eelco Dolstra a824d58b56 * Merge addToStore and addToStoreFixed.
* addToStore now adds unconditionally, it doesn't use readOnlyMode.
  Read-only operation is up to the caller (who can call
  computeStorePathForPath).
2006-12-01 20:51:18 +00:00
Eelco Dolstra ceb982a1be * Right name. 2006-12-01 18:02:05 +00:00
Eelco Dolstra b0d8e05be1 * More operations.
* addToStore() and friends: don't do a round-trip to the worker if
  we're only interested in the path (i.e., in read-only mode).
2006-12-01 18:00:01 +00:00
Eelco Dolstra 0565b5f2b3 * More remote operations.
* Added new operation hasSubstitutes(), which is more efficient than
  querySubstitutes().size() > 0.
2006-11-30 22:43:55 +00:00
Eelco Dolstra 0263279071 * More operations. 2006-11-30 20:45:20 +00:00
Eelco Dolstra a711689368 * First remote operation: isValidPath(). 2006-11-30 20:13:59 +00:00
Eelco Dolstra 765bdfe542 * When NIX_REMOTE is set to "slave", fork off nix-worker in slave
mode.  Presumably nix-worker would be setuid to the Nix store user.
  The worker performs all operations on the Nix store and database, so
  the caller can be completely unprivileged.

  This is already much more secure than the old setuid scheme, since
  the worker doesn't need to do Nix expression evaluation and so on.
  Most importantly, this means that it doesn't need to access any user
  files, with all resulting security risks; it only performs pure
  store operations.

  Once this works, it is easy to move to a daemon model that forks off
  a worker for connections established through a Unix domain socket.
  That would be even more secure.
2006-11-30 19:54:43 +00:00
Eelco Dolstra 40b3f64b55 * Skeleton of the privileged worker program.
* Some refactoring: put the NAR archive integer/string serialisation
  code in a separate file so it can be reused by the worker protocol
  implementation.
2006-11-30 19:19:59 +00:00