about summary refs log tree commit diff
path: root/src
AgeCommit message (Collapse)AuthorFilesLines
2006-12-05 * FreeBSD returns ESRCH when there are no processes to kill.Eelco Dolstra1-4/+8
2006-12-05 * Oops! In daemon mode, we can't run as root either if build-users is empty.Eelco Dolstra1-2/+2
2006-12-05 * Use an explicit handler for SIGCHLD, since SIG_IGN doesn't do the Eelco Dolstra1-3/+15
right thing on FreeBSD 4 (it leaves zombies).
2006-12-05 * Better message.Eelco Dolstra1-1/+1
2006-12-05 * Ugly hack to handle spurious SIGPOLLs.Eelco Dolstra1-20/+50
2006-12-05 * Some renaming.Eelco Dolstra1-6/+6
2006-12-05 * Allow unprivileged users to run the garbage collector and to doEelco Dolstra16-56/+106
`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 * The determination of the root set should be made by the privilegedEelco Dolstra9-77/+138
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 * findRoots: return a map from the symlink (outside of the store) toEelco Dolstra1-11/+16
the store path (inside the store).
2006-12-05 * In addPermRoot, check that the root that we just registered can beEelco Dolstra2-18/+42
found by the garbage collector. This addresses NIX-71 and is a particular concern in multi-user stores.
2006-12-04 * Add indirect root registration to the protocol so that unprivilegedEelco Dolstra7-6/+42
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 * Not every OS knows about SIGPOLL.Eelco Dolstra1-0/+5
2006-12-04 * Don't redirect stderr.Eelco Dolstra1-6/+0
2006-12-04 * Handle exceptions and stderr for all protocol functions.Eelco Dolstra2-19/+64
* 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 * Daemon mode (`nix-worker --daemon'). Clients connect to the serverEelco Dolstra7-53/+182
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 * When NIX_REMOTE=daemon, connect to /nix/var/nix/daemon.socketEelco Dolstra3-8/+37
instead of forking a worker.
2006-12-04 * Refactoring.Eelco Dolstra3-17/+33
2006-12-04 * Pass the verbosity level to the worker.Eelco Dolstra1-2/+8
2006-12-04 * Install the worker in bindir, not libexecdir.Eelco Dolstra6-2/+11
* Allow the worker path to be overriden through the NIX_WORKER environment variable.
2006-12-03 * Doh.Eelco Dolstra2-1/+4
2006-12-03 * Don't run setuid root when build-users is empty.Eelco Dolstra3-26/+51
* Send startup errors to the client.
2006-12-03 * Removed `build-allow-root'.Eelco Dolstra4-21/+34
* Added `build-users-group', the group under which builds are to be performed. * Check that /nix/store has 1775 permission and is owner by the build-users-group.
2006-12-03 * Use setreuid if setresuid is not available.Eelco Dolstra1-6/+12
2006-12-03 * Handle a subtle race condition: the client closing the socketEelco Dolstra1-2/+30
between the last worker read/write and the enabling of the signal handler.
2006-12-03 * Some hardcore magic to handle asynchronous client disconnects.Eelco Dolstra1-1/+37
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 * Use a Unix domain socket instead of pipes.Eelco Dolstra2-15/+18
2006-12-03 * Better error message if the worker doesn't start.Eelco Dolstra1-4/+8
2006-12-03 * Pid::kill() should be interruptable.Eelco Dolstra1-1/+3
2006-12-03 * Some hackery to propagate the worker's stderr and exceptions to theEelco Dolstra7-104/+206
client.
2006-12-03 * Run the worker in a separate session to prevent terminal signalsEelco Dolstra1-0/+6
from interfering.
2006-12-02 * Move addTempRoot() to the store API, and add another functionEelco Dolstra11-17/+81
syncWithGC() to allow clients to register GC roots without needing write access to the global roots directory or the GC lock.
2006-12-02 * Doh.Eelco Dolstra1-1/+1
2006-12-02 * Remove most of the old setuid code.Eelco Dolstra3-138/+57
* Much simpler setuid code for the worker in slave mode.
2006-12-02 * Remove SwitchToOriginalUser, we're not going to need it anymore.Eelco Dolstra8-63/+4
2006-12-02 * Remove queryPathHash().Eelco Dolstra4-2/+32
* Help for nix-worker.
2006-12-01 * Replace read-only calls to addTextToStore.Eelco Dolstra2-3/+9
2006-12-01 * Merge addToStore and addToStoreFixed.Eelco Dolstra11-96/+40
* addToStore now adds unconditionally, it doesn't use readOnlyMode. Read-only operation is up to the caller (who can call computeStorePathForPath).
2006-12-01 * Right name.Eelco Dolstra1-2/+2
2006-12-01 * More operations.Eelco Dolstra5-33/+116
* 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).
2006-11-30 * More remote operations.Eelco Dolstra12-46/+138
* Added new operation hasSubstitutes(), which is more efficient than querySubstitutes().size() > 0.
2006-11-30 * Doh.Eelco Dolstra1-1/+1
2006-11-30 * More operations.Eelco Dolstra3-9/+50
2006-11-30 * First remote operation: isValidPath().Eelco Dolstra4-7/+63
2006-11-30 * When NIX_REMOTE is set to "slave", fork off nix-worker in slaveEelco Dolstra5-24/+87
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 * Skeleton of the privileged worker program.Eelco Dolstra11-126/+254
* 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 * Oops.Eelco Dolstra2-0/+135
2006-11-30 * Skeleton of remote store implementation.Eelco Dolstra4-6/+14
2006-11-30 * Put building in the store API.Eelco Dolstra11-48/+31
2006-11-30 * Refactoring. There is now an abstract interface class StoreAPIEelco Dolstra17-338/+458
containing functions that operate on the Nix store. One implementation is LocalStore, which operates on the Nix store directly. The next step, to enable secure multi-user Nix, is to create a different implementation RemoteStore that talks to a privileged daemon process that uses LocalStore to perform the actual operations.
2006-11-29 * Don't spam.Eelco Dolstra1-0/+2