nix-super/src/libstore/build.cc

2399 lines
68 KiB
C++
Raw Normal View History

#include "references.hh"
#include "pathlocks.hh"
#include "misc.hh"
#include "globals.hh"
#include "local-store.hh"
#include "db.hh"
#include "util.hh"
2003-07-20 22:29:38 +03:00
#include <map>
#include <iostream>
#include <sstream>
#include <boost/shared_ptr.hpp>
#include <boost/weak_ptr.hpp>
#include <boost/enable_shared_from_this.hpp>
2003-07-20 22:29:38 +03:00
#include <time.h>
2006-05-30 14:37:21 +03:00
#include <sys/time.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
2004-07-01 14:11:16 +03:00
#include <errno.h>
#include <pwd.h>
#include <grp.h>
2003-07-20 22:29:38 +03:00
namespace nix {
using std::map;
2003-07-20 22:29:38 +03:00
2005-01-20 18:01:07 +02:00
/* !!! TODO derivationFromPath shouldn't be used here */
static string pathNullDevice = "/dev/null";
static const uid_t rootUserId = 0;
/* Forward definition. */
class Worker;
/* A pointer to a goal. */
class Goal;
typedef boost::shared_ptr<Goal> GoalPtr;
typedef boost::weak_ptr<Goal> WeakGoalPtr;
/* Set of goals. */
typedef set<GoalPtr> Goals;
typedef set<WeakGoalPtr> WeakGoals;
/* A map of paths to goals (and the other way around). */
typedef map<Path, WeakGoalPtr> WeakGoalMap;
class Goal : public boost::enable_shared_from_this<Goal>
{
public:
typedef enum {ecBusy, ecSuccess, ecFailed} ExitCode;
protected:
/* Backlink to the worker. */
Worker & worker;
/* Goals that this goal is waiting for. */
Goals waitees;
/* Goals waiting for this one to finish. Must use weak pointers
here to prevent cycles. */
WeakGoals waiters;
/* Number of goals we are/were waiting for that have failed. */
unsigned int nrFailed;
/* Name of this goal for debugging purposes. */
string name;
/* Whether the goal is finished. */
ExitCode exitCode;
Goal(Worker & worker) : worker(worker)
{
nrFailed = 0;
exitCode = ecBusy;
}
virtual ~Goal()
{
trace("goal destroyed");
}
public:
virtual void work() = 0;
void addWaitee(GoalPtr waitee);
virtual void waiteeDone(GoalPtr waitee, bool success);
virtual void handleChildOutput(int fd, const string & data)
{
abort();
}
virtual void handleEOF(int fd)
2004-06-29 12:41:50 +03:00
{
abort();
}
void trace(const format & f);
string getName()
{
return name;
}
ExitCode getExitCode()
{
return exitCode;
}
protected:
void amDone(bool success = true);
};
/* A mapping used to remember for each child process to what goal it
belongs, and file descriptors for receiving log data and output
path creation commands. */
2004-06-20 00:45:04 +03:00
struct Child
{
WeakGoalPtr goal;
set<int> fds;
2004-06-20 00:45:04 +03:00
bool inBuildSlot;
time_t lastOutput; /* time we last got output on stdout/stderr */
2004-06-20 00:45:04 +03:00
};
typedef map<pid_t, Child> Children;
/* The worker class. */
class Worker
{
private:
/* Note: the worker should only have strong pointers to the
top-level goals. */
/* The top-level goals of the worker. */
Goals topGoals;
/* Goals that are ready to do some work. */
WeakGoals awake;
/* Goals waiting for a build slot. */
WeakGoals wantingToBuild;
/* Child processes currently running. */
Children children;
2004-06-20 00:45:04 +03:00
/* Number of build slots occupied. Not all child processes
(namely build hooks) count as occupied build slots. */
unsigned int nrChildren;
/* Maps used to prevent multiple instantiations of a goal for the
2005-01-20 18:01:07 +02:00
same derivation / path. */
WeakGoalMap derivationGoals;
WeakGoalMap substitutionGoals;
public:
Worker();
~Worker();
/* Make a goal (with caching). */
GoalPtr makeDerivationGoal(const Path & drvPath);
GoalPtr makeSubstitutionGoal(const Path & storePath);
/* Remove a dead goal. */
void removeGoal(GoalPtr goal);
/* Wake up a goal (i.e., there is something for it to do). */
void wakeUp(GoalPtr goal);
/* Can we start another child process? */
bool canBuildMore();
/* Registers / unregisters a running child process. */
void childStarted(GoalPtr goal, pid_t pid,
const set<int> & fds, bool inBuildSlot);
void childTerminated(pid_t pid, bool wakeSleepers = true);
/* Add a goal to the set of goals waiting for a build slot. */
void waitForBuildSlot(GoalPtr goal, bool reallyWait = false);
/* Loop until the specified top-level goals have finished. */
void run(const Goals & topGoals);
/* Wait for input to become available. */
void waitForInput();
};
class SubstError : public Error
{
public:
SubstError(const format & f) : Error(f) { };
};
class BuildError : public Error
{
public:
BuildError(const format & f) : Error(f) { };
};
//////////////////////////////////////////////////////////////////////
void Goal::addWaitee(GoalPtr waitee)
{
waitees.insert(waitee);
waitee->waiters.insert(shared_from_this());
}
void Goal::waiteeDone(GoalPtr waitee, bool success)
{
assert(waitees.find(waitee) != waitees.end());
waitees.erase(waitee);
trace(format("waitee `%1%' done; %2% left") %
waitee->name % waitees.size());
if (!success) ++nrFailed;
if (waitees.empty() || (!success && !keepGoing)) {
/* If we failed and keepGoing is not set, we remove all
remaining waitees. */
for (Goals::iterator i = waitees.begin(); i != waitees.end(); ++i) {
GoalPtr goal = *i;
WeakGoals waiters2;
for (WeakGoals::iterator j = goal->waiters.begin();
j != goal->waiters.end(); ++j)
if (j->lock() != shared_from_this())
waiters2.insert(*j);
goal->waiters = waiters2;
}
waitees.clear();
worker.wakeUp(shared_from_this());
}
}
void Goal::amDone(bool success)
{
trace("done");
assert(exitCode == ecBusy);
exitCode = success ? ecSuccess : ecFailed;
for (WeakGoals::iterator i = waiters.begin(); i != waiters.end(); ++i) {
GoalPtr goal = i->lock();
if (goal) goal->waiteeDone(shared_from_this(), success);
}
waiters.clear();
worker.removeGoal(shared_from_this());
}
void Goal::trace(const format & f)
{
debug(format("%1%: %2%") % name % f);
}
2004-06-22 13:21:44 +03:00
//////////////////////////////////////////////////////////////////////
/* Common initialisation performed in child processes. */
void commonChildInit(Pipe & logPipe)
{
/* Put the child in a separate session (and thus a separate
process group) so that it has no controlling terminal (meaning
that e.g. ssh cannot open /dev/tty) and it doesn't receive
terminal signals. */
if (setsid() == -1)
throw SysError(format("creating a new session"));
2004-06-22 13:21:44 +03:00
/* Dup the write side of the logger pipe into stderr. */
if (dup2(logPipe.writeSide, STDERR_FILENO) == -1)
throw SysError("cannot pipe standard error into log file");
logPipe.readSide.close();
/* Dup stderr to stdout. */
2004-06-22 13:21:44 +03:00
if (dup2(STDERR_FILENO, STDOUT_FILENO) == -1)
throw SysError("cannot dup stderr into stdout");
/* Reroute stdin to /dev/null. */
int fdDevNull = open(pathNullDevice.c_str(), O_RDWR);
if (fdDevNull == -1)
throw SysError(format("cannot open `%1%'") % pathNullDevice);
if (dup2(fdDevNull, STDIN_FILENO) == -1)
throw SysError("cannot dup null device into stdin");
}
/* Convert a string list to an array of char pointers. Careful: the
string list should outlive the array. */
2004-06-22 13:21:44 +03:00
const char * * strings2CharPtrs(const Strings & ss)
{
const char * * arr = new const char * [ss.size() + 1];
const char * * p = arr;
for (Strings::const_iterator i = ss.begin(); i != ss.end(); ++i)
*p++ = i->c_str();
*p = 0;
return arr;
}
//////////////////////////////////////////////////////////////////////
class UserLock
{
private:
/* POSIX locks suck. If we have a lock on a file, and we open and
close that file again (without closing the original file
descriptor), we lose the lock. So we have to be *very* careful
not to open a lock file on which we are holding a lock. */
static PathSet lockedPaths; /* !!! not thread-safe */
Path fnUserLock;
AutoCloseFD fdUserLock;
string user;
uid_t uid;
gid_t gid;
public:
UserLock();
~UserLock();
void acquire();
void release();
void kill();
string getUser() { return user; }
uid_t getUID() { return uid; }
uid_t getGID() { return gid; }
bool enabled() { return uid != 0; }
};
PathSet UserLock::lockedPaths;
UserLock::UserLock()
{
uid = gid = 0;
}
UserLock::~UserLock()
{
release();
}
void UserLock::acquire()
{
assert(uid == 0);
string buildUsersGroup = querySetting("build-users-group", "");
assert(buildUsersGroup != "");
/* Get the members of the build-users-group. */
struct group * gr = getgrnam(buildUsersGroup.c_str());
if (!gr)
throw Error(format("the group `%1%' specified in `build-users-group' does not exist")
% buildUsersGroup);
gid = gr->gr_gid;
/* Copy the result of getgrnam. */
Strings users;
for (char * * p = gr->gr_mem; *p; ++p) {
debug(format("found build user `%1%'") % *p);
users.push_back(*p);
}
if (users.empty())
throw Error(format("the build users group `%1%' has no members")
% buildUsersGroup);
/* Find a user account that isn't currently in use for another
build. */
for (Strings::iterator i = users.begin(); i != users.end(); ++i) {
2006-12-05 21:01:19 +02:00
debug(format("trying user `%1%'") % *i);
struct passwd * pw = getpwnam(i->c_str());
if (!pw)
throw Error(format("the user `%1%' in the group `%2%' does not exist")
% *i % buildUsersGroup);
2006-12-07 02:19:27 +02:00
fnUserLock = (format("%1%/userpool/%2%") % nixStateDir % pw->pw_uid).str();
if (lockedPaths.find(fnUserLock) != lockedPaths.end())
/* We already have a lock on this one. */
continue;
AutoCloseFD fd = open(fnUserLock.c_str(), O_RDWR | O_CREAT, 0600);
if (fd == -1)
throw SysError(format("opening user lock `%1%'") % fnUserLock);
if (lockFile(fd, ltWrite, false)) {
fdUserLock = fd.borrow();
lockedPaths.insert(fnUserLock);
user = *i;
uid = pw->pw_uid;
2006-12-07 02:19:27 +02:00
/* Sanity check... */
if (uid == getuid() || uid == geteuid())
throw Error(format("the Nix user should not be a member of `%1%'")
% buildUsersGroup);
return;
}
}
throw Error(format("all build users are currently in use; "
"consider creating additional users and adding them to the `%1%' group")
% buildUsersGroup);
}
void UserLock::release()
{
if (uid == 0) return;
fdUserLock.close(); /* releases lock */
assert(lockedPaths.find(fnUserLock) != lockedPaths.end());
lockedPaths.erase(fnUserLock);
fnUserLock = "";
uid = 0;
}
static void runSetuidHelper(const string & command,
const string & arg)
{
string program = nixLibexecDir + "/nix-setuid-helper";
/* Fork. */
Pid pid;
pid = fork();
switch (pid) {
case -1:
throw SysError("unable to fork");
case 0: /* child */
try {
std::vector<const char *> args; /* careful with c_str()!
*/
args.push_back(program.c_str());
args.push_back(command.c_str());
args.push_back(arg.c_str());
args.push_back(0);
execve(program.c_str(), (char * *) &args[0], 0);
throw SysError(format("executing `%1%'") % program);
}
catch (std::exception & e) {
std::cerr << "error: " << e.what() << std::endl;
}
quickExit(1);
}
/* Parent. */
/* Wait for the child to finish. */
int status = pid.wait(true);
if (!statusOk(status))
throw Error(format("program `%1%' %2%")
% program % statusToString(status));
}
void UserLock::kill()
{
assert(enabled());
if (amPrivileged())
killUser(uid);
else
runSetuidHelper("kill", user);
}
bool amPrivileged()
{
return geteuid() == 0;
}
bool haveBuildUsers()
{
return querySetting("build-users-group", "") != "";
}
void getOwnership(const Path & path)
{
runSetuidHelper("get-ownership", path);
}
static void deletePathWrapped(const Path & path)
{
/* When using build users and we're not root, we may not have
sufficient permission to delete the path. So use the setuid
helper to change ownership to us. */
if (haveBuildUsers() && !amPrivileged())
getOwnership(path);
deletePath(path);
}
//////////////////////////////////////////////////////////////////////
class DerivationGoal : public Goal
{
private:
2005-01-20 18:01:07 +02:00
/* The path of the derivation. */
Path drvPath;
2005-01-20 18:01:07 +02:00
/* The derivation stored at drvPath. */
Derivation drv;
/* The remainder is state held during the build. */
/* Locks on the output paths. */
PathLocks outputLocks;
/* All input paths (that is, the union of FS closures of the
immediate input paths). */
PathSet inputPaths;
/* Referenceable paths (i.e., input and output paths). */
PathSet allPaths;
/* User selected for running the builder. */
UserLock buildUser;
/* The process ID of the builder. */
2004-06-22 12:51:44 +03:00
Pid pid;
/* The temporary directory. */
Path tmpDir;
/* File descriptor for the log file. */
2004-06-15 16:49:42 +03:00
AutoCloseFD fdLogFile;
/* Pipe for the builder's standard output/error. */
Pipe logPipe;
/* Pipes for talking to the build hook (if any). */
Pipe toHook;
Pipe fromHook;
typedef void (DerivationGoal::*GoalState)();
GoalState state;
public:
DerivationGoal(const Path & drvPath, Worker & worker);
~DerivationGoal();
void work();
2004-06-20 00:45:04 +03:00
Path getDrvPath()
{
return drvPath;
}
2004-06-20 00:45:04 +03:00
private:
/* The states. */
void init();
2006-12-08 01:58:36 +02:00
void haveDerivation();
void outputsSubstituted();
void inputsRealised();
void tryToBuild();
void buildDone();
2004-06-20 00:45:04 +03:00
/* Is the build hook willing to perform the build? */
typedef enum {rpAccept, rpDecline, rpPostpone, rpDone} HookReply;
HookReply tryBuildHook();
/* Synchronously wait for a build hook to finish. */
void terminateBuildHook();
/* Acquires locks on the output paths and gathers information
about the build (e.g., the input closures). During this
process its possible that we find out that the build is
unnecessary, in which case we return false (this is not an
error condition!). */
bool prepareBuild();
/* Start building a derivation. */
void startBuilder();
/* Must be called after the output paths have become valid (either
due to a successful build or hook, or because they already
were). */
void computeClosure();
/* Open a log file and a pipe to it. */
void openLogFile();
/* Common initialisation to be performed in child processes (i.e.,
both in builders and in build hooks. */
void initChild();
/* Delete the temporary directory, if we have one. */
void deleteTmpDir(bool force);
2004-06-29 12:41:50 +03:00
/* Callback used by the worker to write to the log. */
void handleChildOutput(int fd, const string & data);
void handleEOF(int fd);
2004-06-29 12:41:50 +03:00
/* Return the set of (in)valid paths. */
PathSet checkPathValidity(bool returnValid);
};
DerivationGoal::DerivationGoal(const Path & drvPath, Worker & worker)
: Goal(worker)
{
this->drvPath = drvPath;
state = &DerivationGoal::init;
name = (format("building of `%1%'") % drvPath).str();
trace("created");
}
DerivationGoal::~DerivationGoal()
{
/* Careful: we should never ever throw an exception from a
destructor. */
try {
if (pid != -1) {
worker.childTerminated(pid);
if (buildUser.enabled()) {
/* Note that we can't let pid's destructor kill the
the child process, since it may not have the
appropriate privilege (i.e., the setuid helper
should do it).
However, if we're using a build user, then there is
a tricky race condition: if we kill the build user
before the child has done its setuid() to the build
user uid, then it won't be killed, and we'll
potentially lock up in pid.wait(). So also send a
conventional kill to the child. */
::kill(-pid, SIGKILL); /* ignore the result */
buildUser.kill();
pid.wait(true);
assert(pid == -1);
}
}
deleteTmpDir(false);
} catch (Error & e) {
printMsg(lvlError, format("error (ignored): %1%") % e.msg());
}
}
void DerivationGoal::work()
{
(this->*state)();
}
void DerivationGoal::init()
{
trace("init");
2005-01-20 18:01:07 +02:00
/* The first thing to do is to make sure that the derivation
exists. If it doesn't, it may be created through a
substitute. */
addWaitee(worker.makeSubstitutionGoal(drvPath));
2006-12-08 01:58:36 +02:00
state = &DerivationGoal::haveDerivation;
}
2006-12-08 01:58:36 +02:00
void DerivationGoal::haveDerivation()
{
trace("loading derivation");
if (nrFailed != 0) {
printMsg(lvlError,
format("cannot build missing derivation `%1%'")
% drvPath);
amDone(false);
return;
}
assert(store->isValidPath(drvPath));
2003-07-20 22:29:38 +03:00
/* Get the derivation. */
drv = derivationFromPath(drvPath);
for (DerivationOutputs::iterator i = drv.outputs.begin();
i != drv.outputs.end(); ++i)
store->addTempRoot(i->second.path);
/* Check what outputs paths are not already valid. */
PathSet invalidOutputs = checkPathValidity(false);
/* If they are all valid, then we're done. */
if (invalidOutputs.size() == 0) {
amDone(true);
return;
}
/* We are first going to try to create the invalid output paths
through substitutes. If that doesn't work, we'll build
them. */
for (PathSet::iterator i = invalidOutputs.begin();
i != invalidOutputs.end(); ++i)
/* Don't bother creating a substitution goal if there are no
substitutes. */
if (store->hasSubstitutes(*i))
addWaitee(worker.makeSubstitutionGoal(*i));
if (waitees.empty()) /* to prevent hang (no wake-up event) */
outputsSubstituted();
else
state = &DerivationGoal::outputsSubstituted;
}
void DerivationGoal::outputsSubstituted()
{
trace("all outputs substituted (maybe)");
if (nrFailed > 0 && !tryFallback)
throw Error(format("some substitutes for the outputs of derivation `%1%' failed; try `--fallback'") % drvPath);
nrFailed = 0;
if (checkPathValidity(false).size() == 0) {
amDone(true);
return;
}
/* Otherwise, at least one of the output paths could not be
produced using a substitute. So we have to build instead. */
/* The inputs must be built before we can build this goal. */
/* !!! but if possible, only install the paths that we need */
for (DerivationInputs::iterator i = drv.inputDrvs.begin();
i != drv.inputDrvs.end(); ++i)
addWaitee(worker.makeDerivationGoal(i->first));
for (PathSet::iterator i = drv.inputSrcs.begin();
i != drv.inputSrcs.end(); ++i)
addWaitee(worker.makeSubstitutionGoal(*i));
state = &DerivationGoal::inputsRealised;
}
void DerivationGoal::inputsRealised()
{
trace("all inputs realised");
if (nrFailed != 0) {
printMsg(lvlError,
format("cannot build derivation `%1%': "
"%2% inputs could not be realised")
% drvPath % nrFailed);
amDone(false);
return;
}
2004-06-20 00:45:04 +03:00
/* Okay, try to build. Note that here we don't wait for a build
slot to become available, since we don't need one if there is a
build hook. */
state = &DerivationGoal::tryToBuild;
2004-06-20 00:45:04 +03:00
worker.wakeUp(shared_from_this());
}
void DerivationGoal::tryToBuild()
{
trace("trying to build");
try {
/* Is the build hook willing to accept this job? */
switch (tryBuildHook()) {
case rpAccept:
/* Yes, it has started doing so. Wait until we get
EOF from the hook. */
state = &DerivationGoal::buildDone;
return;
case rpPostpone:
/* Not now; wait until at least one child finishes. */
worker.waitForBuildSlot(shared_from_this(), true);
return;
case rpDecline:
/* We should do it ourselves. */
break;
case rpDone:
/* Somebody else did it. */
amDone();
return;
}
/* Make sure that we are allowed to start a build. */
if (!worker.canBuildMore()) {
2004-06-20 00:45:04 +03:00
worker.waitForBuildSlot(shared_from_this());
return;
}
/* Acquire locks and such. If we then see that the build has
been done by somebody else, we're done. */
if (!prepareBuild()) {
2004-06-20 00:45:04 +03:00
amDone();
return;
}
2004-06-20 00:45:04 +03:00
/* Okay, we have to build. */
startBuilder();
} catch (BuildError & e) {
printMsg(lvlError, e.msg());
amDone(false);
2004-06-20 00:45:04 +03:00
return;
}
/* This state will be reached when we get EOF on the child's
log pipe. */
state = &DerivationGoal::buildDone;
2004-06-20 00:45:04 +03:00
}
void DerivationGoal::buildDone()
2004-06-20 00:45:04 +03:00
{
trace("build done");
2004-06-20 00:45:04 +03:00
/* Since we got an EOF on the logger pipe, the builder is presumed
to have terminated. In fact, the builder could also have
simply have closed its end of the pipe --- just don't do that
:-) */
/* !!! this could block! security problem! solution: kill the
child */
2004-06-22 12:51:44 +03:00
pid_t savedPid = pid;
int status = pid.wait(true);
2004-06-20 00:45:04 +03:00
debug(format("builder process for `%1%' finished") % drvPath);
2004-06-20 00:45:04 +03:00
/* So the child is gone now. */
2004-06-22 12:51:44 +03:00
worker.childTerminated(savedPid);
2004-06-20 00:45:04 +03:00
/* Close the read side of the logger pipe. */
logPipe.readSide.close();
/* Close the log file. */
fdLogFile.close();
/* When running under a build user, make sure that all processes
running under that uid are gone. This is to prevent a
malicious user from leaving behind a process that keeps files
open and modifies them after they have been chown'ed to
root. */
if (buildUser.enabled())
buildUser.kill();
/* Some cleanup per path. We do this here and not in
computeClosure() for convenience when the build has failed. */
for (DerivationOutputs::iterator i = drv.outputs.begin();
i != drv.outputs.end(); ++i)
{
Path path = i->second.path;
if (!pathExists(path)) continue;
2004-06-20 00:45:04 +03:00
struct stat st;
if (lstat(path.c_str(), &st))
throw SysError(format("getting attributes of path `%1%'") % path);
#ifndef __CYGWIN__
/* Check that the output is not group or world writable, as
that means that someone else can have interfered with the
build. Also, the output should be owned by the build
user. */
if ((st.st_mode & (S_IWGRP | S_IWOTH)) ||
(buildUser.enabled() && st.st_uid != buildUser.getUID()))
throw Error(format("suspicious ownership or permission on `%1%'; rejecting this build output") % path);
#endif
2004-06-20 00:45:04 +03:00
/* Gain ownership of the build result using the setuid wrapper
if we're not root. If we *are* root, then
canonicalisePathMetaData() will take care of this later
on. */
if (buildUser.enabled() && !amPrivileged())
getOwnership(path);
}
2004-06-20 00:45:04 +03:00
/* Check the exit status. */
if (!statusOk(status)) {
2004-06-20 00:45:04 +03:00
deleteTmpDir(false);
printMsg(lvlError, format("builder for `%1%' %2%")
% drvPath % statusToString(status));
amDone(false);
return;
}
deleteTmpDir(true);
2004-06-20 00:45:04 +03:00
/* Compute the FS closure of the outputs and register them as
being valid. */
try {
computeClosure();
} catch (BuildError & e) {
printMsg(lvlError, e.msg());
amDone(false);
return;
}
2004-06-20 00:45:04 +03:00
/* Release the build user, if applicable. */
buildUser.release();
2004-06-20 00:45:04 +03:00
amDone();
}
static string readLine(int fd)
{
string s;
while (1) {
char ch;
ssize_t rd = read(fd, &ch, 1);
if (rd == -1) {
if (errno != EINTR)
throw SysError("reading a line");
} else if (rd == 0)
throw Error("unexpected EOF reading a line");
else {
if (ch == '\n') return s;
s += ch;
}
}
}
static void writeLine(int fd, string s)
{
s += '\n';
writeFull(fd, (const unsigned char *) s.c_str(), s.size());
}
/* !!! ugly hack */
static void drain(int fd)
{
unsigned char buffer[1024];
while (1) {
ssize_t rd = read(fd, buffer, sizeof buffer);
if (rd == -1) {
if (errno != EINTR)
throw SysError("draining");
} else if (rd == 0) break;
else writeToStderr(buffer, rd);
2004-06-20 00:45:04 +03:00
}
}
* Removed the `id' attribute hack. * Formalise the notion of fixed-output derivations, i.e., derivations for which a cryptographic hash of the output is known in advance. Changes to such derivations should not propagate upwards through the dependency graph. Previously this was done by specifying the hash component of the output path through the `id' attribute, but this is insecure since you can lie about it (i.e., you can specify any hash and then produce a completely different output). Now the responsibility for checking the output is moved from the builder to Nix itself. A fixed-output derivation can be created by specifying the `outputHash' and `outputHashAlgo' attributes, the latter taking values `md5', `sha1', and `sha256', and the former specifying the actual hash in hexadecimal or in base-32 (auto-detected by looking at the length of the attribute value). MD5 is included for compatibility but should be considered deprecated. * Removed the `drvPath' pseudo-attribute in derivation results. It's no longer necessary. * Cleaned up the support for multiple output paths in derivation store expressions. Each output now has a unique identifier (e.g., `out', `devel', `docs'). Previously there was no way to tell output paths apart at the store expression level. * `nix-hash' now has a flag `--base32' to specify that the hash should be printed in base-32 notation. * `fetchurl' accepts parameters `sha256' and `sha1' in addition to `md5'. * `nix-prefetch-url' now prints out a SHA-1 hash in base-32. (TODO: a flag to specify the hash.)
2005-01-17 18:55:19 +02:00
PathSet outputPaths(const DerivationOutputs & outputs)
{
PathSet paths;
for (DerivationOutputs::const_iterator i = outputs.begin();
i != outputs.end(); ++i)
paths.insert(i->second.path);
return paths;
}
string showPaths(const PathSet & paths)
{
string s;
for (PathSet::const_iterator i = paths.begin();
i != paths.end(); ++i)
{
if (s.size() != 0) s += ", ";
s += "`" + *i + "'";
* Removed the `id' attribute hack. * Formalise the notion of fixed-output derivations, i.e., derivations for which a cryptographic hash of the output is known in advance. Changes to such derivations should not propagate upwards through the dependency graph. Previously this was done by specifying the hash component of the output path through the `id' attribute, but this is insecure since you can lie about it (i.e., you can specify any hash and then produce a completely different output). Now the responsibility for checking the output is moved from the builder to Nix itself. A fixed-output derivation can be created by specifying the `outputHash' and `outputHashAlgo' attributes, the latter taking values `md5', `sha1', and `sha256', and the former specifying the actual hash in hexadecimal or in base-32 (auto-detected by looking at the length of the attribute value). MD5 is included for compatibility but should be considered deprecated. * Removed the `drvPath' pseudo-attribute in derivation results. It's no longer necessary. * Cleaned up the support for multiple output paths in derivation store expressions. Each output now has a unique identifier (e.g., `out', `devel', `docs'). Previously there was no way to tell output paths apart at the store expression level. * `nix-hash' now has a flag `--base32' to specify that the hash should be printed in base-32 notation. * `fetchurl' accepts parameters `sha256' and `sha1' in addition to `md5'. * `nix-prefetch-url' now prints out a SHA-1 hash in base-32. (TODO: a flag to specify the hash.)
2005-01-17 18:55:19 +02:00
}
return s;
}
/* Return a string accepted by `nix-store --register-validity' that
registers the specified paths as valid. Note: it's the
responsibility of the caller to provide a closure. */
static string makeValidityRegistration(const PathSet & paths,
bool showDerivers)
{
string s = "";
for (PathSet::iterator i = paths.begin(); i != paths.end(); ++i) {
s += *i + "\n";
Path deriver = showDerivers ? queryDeriver(noTxn, *i) : "";
s += deriver + "\n";
PathSet references;
store->queryReferences(*i, references);
s += (format("%1%\n") % references.size()).str();
for (PathSet::iterator j = references.begin();
j != references.end(); ++j)
s += *j + "\n";
}
return s;
}
DerivationGoal::HookReply DerivationGoal::tryBuildHook()
2004-06-20 00:45:04 +03:00
{
Path buildHook = getEnv("NIX_BUILD_HOOK");
if (buildHook == "") return rpDecline;
buildHook = absPath(buildHook);
/* Create a directory where we will store files used for
communication between us and the build hook. */
tmpDir = createTempDir();
/* Create the log file and pipe. */
openLogFile();
/* Create the communication pipes. */
toHook.create();
fromHook.create();
/* Fork the hook. */
2004-06-22 12:51:44 +03:00
pid = fork();
switch (pid) {
2004-06-20 00:45:04 +03:00
case -1:
throw SysError("unable to fork");
case 0:
try { /* child */
initChild();
execl(buildHook.c_str(), buildHook.c_str(),
(worker.canBuildMore() ? (string) "1" : "0").c_str(),
thisSystem.c_str(),
drv.platform.c_str(),
drvPath.c_str(), NULL);
2004-06-20 00:45:04 +03:00
throw SysError(format("executing `%1%'") % buildHook);
} catch (std::exception & e) {
std::cerr << format("build hook error: %1%\n") % e.what();
2004-06-20 00:45:04 +03:00
}
quickExit(1);
2004-06-20 00:45:04 +03:00
}
/* parent */
pid.setSeparatePG(true);
2004-06-20 00:45:04 +03:00
logPipe.writeSide.close();
worker.childStarted(shared_from_this(),
pid, singleton<set<int> >(logPipe.readSide), false);
2004-06-20 00:45:04 +03:00
fromHook.writeSide.close();
toHook.readSide.close();
/* Read the first line of input, which should be a word indicating
whether the hook wishes to perform the build. !!! potential
for deadlock here: we should also read from the child's logger
pipe. */
string reply;
try {
reply = readLine(fromHook.readSide);
} catch (Error & e) {
drain(logPipe.readSide);
throw;
}
debug(format("hook reply is `%1%'") % reply);
if (reply == "decline" || reply == "postpone") {
/* Clean up the child. !!! hacky / should verify */
drain(logPipe.readSide);
terminateBuildHook();
return reply == "decline" ? rpDecline : rpPostpone;
}
else if (reply == "accept") {
/* Acquire locks and such. If we then see that the output
paths are now valid, we're done. */
2004-06-20 00:45:04 +03:00
if (!prepareBuild()) {
/* Tell the hook to exit. */
2004-06-20 00:45:04 +03:00
writeLine(toHook.writeSide, "cancel");
terminateBuildHook();
return rpDone;
}
printMsg(lvlInfo, format("running hook to build path(s) %1%")
% showPaths(outputPaths(drv.outputs)));
2004-06-20 00:45:04 +03:00
/* Write the information that the hook needs to perform the
2005-01-20 18:01:07 +02:00
build, i.e., the set of input paths, the set of output
paths, and the references (pointer graph) in the input
paths. */
2004-06-20 00:45:04 +03:00
Path inputListFN = tmpDir + "/inputs";
Path outputListFN = tmpDir + "/outputs";
Path referencesFN = tmpDir + "/references";
2004-06-20 00:45:04 +03:00
/* The `inputs' file lists all inputs that have to be copied
to the remote system. This unfortunately has to contain
the entire derivation closure to ensure that the validity
invariant holds on the remote system. (I.e., it's
unfortunate that we have to list it since the remote system
*probably* already has it.) */
PathSet allInputs;
allInputs.insert(inputPaths.begin(), inputPaths.end());
computeFSClosure(drvPath, allInputs);
2004-06-20 00:45:04 +03:00
string s;
for (PathSet::iterator i = allInputs.begin();
i != allInputs.end(); ++i)
2004-06-20 00:45:04 +03:00
s += *i + "\n";
writeStringToFile(inputListFN, s);
/* The `outputs' file lists all outputs that have to be copied
from the remote system. */
2004-06-20 00:45:04 +03:00
s = "";
for (DerivationOutputs::iterator i = drv.outputs.begin();
i != drv.outputs.end(); ++i)
* Removed the `id' attribute hack. * Formalise the notion of fixed-output derivations, i.e., derivations for which a cryptographic hash of the output is known in advance. Changes to such derivations should not propagate upwards through the dependency graph. Previously this was done by specifying the hash component of the output path through the `id' attribute, but this is insecure since you can lie about it (i.e., you can specify any hash and then produce a completely different output). Now the responsibility for checking the output is moved from the builder to Nix itself. A fixed-output derivation can be created by specifying the `outputHash' and `outputHashAlgo' attributes, the latter taking values `md5', `sha1', and `sha256', and the former specifying the actual hash in hexadecimal or in base-32 (auto-detected by looking at the length of the attribute value). MD5 is included for compatibility but should be considered deprecated. * Removed the `drvPath' pseudo-attribute in derivation results. It's no longer necessary. * Cleaned up the support for multiple output paths in derivation store expressions. Each output now has a unique identifier (e.g., `out', `devel', `docs'). Previously there was no way to tell output paths apart at the store expression level. * `nix-hash' now has a flag `--base32' to specify that the hash should be printed in base-32 notation. * `fetchurl' accepts parameters `sha256' and `sha1' in addition to `md5'. * `nix-prefetch-url' now prints out a SHA-1 hash in base-32. (TODO: a flag to specify the hash.)
2005-01-17 18:55:19 +02:00
s += i->second.path + "\n";
2004-06-20 00:45:04 +03:00
writeStringToFile(outputListFN, s);
/* The `references' file has exactly the format accepted by
`nix-store --register-validity'. */
writeStringToFile(referencesFN,
makeValidityRegistration(allInputs, true));
/* Tell the hook to proceed. */
2004-06-20 00:45:04 +03:00
writeLine(toHook.writeSide, "okay");
return rpAccept;
}
else throw Error(format("bad hook reply `%1%'") % reply);
}
void DerivationGoal::terminateBuildHook()
2004-06-20 00:45:04 +03:00
{
/* !!! drain stdout of hook */
debug("terminating build hook");
2004-06-22 12:51:44 +03:00
pid_t savedPid = pid;
pid.wait(true);
worker.childTerminated(savedPid, false);
2004-06-20 00:45:04 +03:00
fromHook.readSide.close();
toHook.writeSide.close();
fdLogFile.close();
logPipe.readSide.close();
deleteTmpDir(true); /* get rid of the hook's temporary directory */
2004-06-20 00:45:04 +03:00
}
bool DerivationGoal::prepareBuild()
2004-06-20 00:45:04 +03:00
{
/* Obtain locks on all output paths. The locks are automatically
released when we exit this function or Nix crashes. */
/* !!! BUG: this could block, which is not allowed. */
outputLocks.lockPaths(outputPaths(drv.outputs),
(format("waiting for lock on %1%") % showPaths(outputPaths(drv.outputs))).str());
/* Now check again whether the outputs are valid. This is because
another process may have started building in parallel. After
it has finished and released the locks, we can (and should)
reuse its results. (Strictly speaking the first check can be
omitted, but that would be less efficient.) Note that since we
now hold the locks on the output paths, no other process can
2005-01-20 18:01:07 +02:00
build this derivation, so no further checks are necessary. */
PathSet validPaths = checkPathValidity(true);
if (validPaths.size() == drv.outputs.size()) {
debug(format("skipping build of derivation `%1%', someone beat us to it")
% drvPath);
outputLocks.setDeletion(true);
2004-06-20 00:45:04 +03:00
return false;
}
2003-07-20 22:29:38 +03:00
if (validPaths.size() > 0) {
/* !!! fix this; try to delete valid paths */
throw Error(
format("derivation `%1%' is blocked by its output paths")
% drvPath);
}
/* Gather information necessary for computing the closure and/or
running the build hook. */
/* The outputs are referenceable paths. */
for (DerivationOutputs::iterator i = drv.outputs.begin();
i != drv.outputs.end(); ++i)
{
* Removed the `id' attribute hack. * Formalise the notion of fixed-output derivations, i.e., derivations for which a cryptographic hash of the output is known in advance. Changes to such derivations should not propagate upwards through the dependency graph. Previously this was done by specifying the hash component of the output path through the `id' attribute, but this is insecure since you can lie about it (i.e., you can specify any hash and then produce a completely different output). Now the responsibility for checking the output is moved from the builder to Nix itself. A fixed-output derivation can be created by specifying the `outputHash' and `outputHashAlgo' attributes, the latter taking values `md5', `sha1', and `sha256', and the former specifying the actual hash in hexadecimal or in base-32 (auto-detected by looking at the length of the attribute value). MD5 is included for compatibility but should be considered deprecated. * Removed the `drvPath' pseudo-attribute in derivation results. It's no longer necessary. * Cleaned up the support for multiple output paths in derivation store expressions. Each output now has a unique identifier (e.g., `out', `devel', `docs'). Previously there was no way to tell output paths apart at the store expression level. * `nix-hash' now has a flag `--base32' to specify that the hash should be printed in base-32 notation. * `fetchurl' accepts parameters `sha256' and `sha1' in addition to `md5'. * `nix-prefetch-url' now prints out a SHA-1 hash in base-32. (TODO: a flag to specify the hash.)
2005-01-17 18:55:19 +02:00
debug(format("building path `%1%'") % i->second.path);
allPaths.insert(i->second.path);
}
/* Determine the full set of input paths. */
/* First, the input derivations. */
for (DerivationInputs::iterator i = drv.inputDrvs.begin();
i != drv.inputDrvs.end(); ++i)
{
/* Add the relevant output closures of the input derivation
`*i' as input paths. Only add the closures of output paths
that are specified as inputs. */
assert(store->isValidPath(i->first));
Derivation inDrv = derivationFromPath(i->first);
for (StringSet::iterator j = i->second.begin();
j != i->second.end(); ++j)
if (inDrv.outputs.find(*j) != inDrv.outputs.end())
computeFSClosure(inDrv.outputs[*j].path, inputPaths);
else
throw Error(
format("derivation `%1%' requires non-existent output `%2%' from input derivation `%3%'")
% drvPath % *j % i->first);
2003-07-20 22:29:38 +03:00
}
/* Second, the input sources. */
for (PathSet::iterator i = drv.inputSrcs.begin();
i != drv.inputSrcs.end(); ++i)
computeFSClosure(*i, inputPaths);
debug(format("added input paths %1%") % showPaths(inputPaths));
allPaths.insert(inputPaths.begin(), inputPaths.end());
2004-06-20 00:45:04 +03:00
return true;
}
void DerivationGoal::startBuilder()
{
startNest(nest, lvlInfo,
format("building path(s) %1%") % showPaths(outputPaths(drv.outputs)))
/* Right platform? */
if (drv.platform != thisSystem)
throw BuildError(
format("a `%1%' is required to build `%3%', but I am a `%2%'")
% drv.platform % thisSystem % drvPath);
/* If any of the outputs already exist but are not registered,
delete them. */
for (DerivationOutputs::iterator i = drv.outputs.begin();
i != drv.outputs.end(); ++i)
{
* Removed the `id' attribute hack. * Formalise the notion of fixed-output derivations, i.e., derivations for which a cryptographic hash of the output is known in advance. Changes to such derivations should not propagate upwards through the dependency graph. Previously this was done by specifying the hash component of the output path through the `id' attribute, but this is insecure since you can lie about it (i.e., you can specify any hash and then produce a completely different output). Now the responsibility for checking the output is moved from the builder to Nix itself. A fixed-output derivation can be created by specifying the `outputHash' and `outputHashAlgo' attributes, the latter taking values `md5', `sha1', and `sha256', and the former specifying the actual hash in hexadecimal or in base-32 (auto-detected by looking at the length of the attribute value). MD5 is included for compatibility but should be considered deprecated. * Removed the `drvPath' pseudo-attribute in derivation results. It's no longer necessary. * Cleaned up the support for multiple output paths in derivation store expressions. Each output now has a unique identifier (e.g., `out', `devel', `docs'). Previously there was no way to tell output paths apart at the store expression level. * `nix-hash' now has a flag `--base32' to specify that the hash should be printed in base-32 notation. * `fetchurl' accepts parameters `sha256' and `sha1' in addition to `md5'. * `nix-prefetch-url' now prints out a SHA-1 hash in base-32. (TODO: a flag to specify the hash.)
2005-01-17 18:55:19 +02:00
Path path = i->second.path;
if (store->isValidPath(path))
throw Error(format("obstructed build: path `%1%' exists") % path);
if (pathExists(path)) {
debug(format("removing unregistered path `%1%'") % path);
deletePathWrapped(path);
}
}
/* Construct the environment passed to the builder. */
typedef map<string, string> Environment;
Environment env;
/* Most shells initialise PATH to some default (/bin:/usr/bin:...) when
PATH is not set. We don't want this, so we fill it in with some dummy
value. */
env["PATH"] = "/path-not-set";
/* Set HOME to a non-existing path to prevent certain programs from using
/etc/passwd (or NIS, or whatever) to locate the home directory (for
example, wget looks for ~/.wgetrc). I.e., these tools use /etc/passwd
if HOME is not set, but they will just assume that the settings file
they are looking for does not exist if HOME is set but points to some
non-existing path. */
env["HOME"] = "/homeless-shelter";
/* Tell the builder where the Nix store is. Usually they
shouldn't care, but this is useful for purity checking (e.g.,
the compiler or linker might only want to accept paths to files
in the store or in the build directory). */
env["NIX_STORE"] = nixStore;
2005-01-20 18:01:07 +02:00
/* Add all bindings specified in the derivation. */
for (StringPairs::iterator i = drv.env.begin();
i != drv.env.end(); ++i)
2003-07-20 22:29:38 +03:00
env[i->first] = i->second;
/* Create a temporary directory where the build will take
place. */
tmpDir = createTempDir();
/* For convenience, set an environment pointing to the top build
directory. */
env["NIX_BUILD_TOP"] = tmpDir;
/* Also set TMPDIR and variants to point to this directory. */
env["TMPDIR"] = env["TEMPDIR"] = env["TMP"] = env["TEMP"] = tmpDir;
/* Compatibility hack with Nix <= 0.7: if this is a fixed-output
derivation, tell the builder, so that for instance `fetchurl'
can skip checking the output. On older Nixes, this environment
variable won't be set, so `fetchurl' will do the check. */
bool fixedOutput = true;
for (DerivationOutputs::iterator i = drv.outputs.begin();
i != drv.outputs.end(); ++i)
if (i->second.hash == "") fixedOutput = false;
if (fixedOutput)
env["NIX_OUTPUT_CHECKED"] = "1";
/* *Only* if this is a fixed-output derivation, propagate the
values of the environment variables specified in the
`impureEnvVars' attribute to the builder. This allows for
instance environment variables for proxy configuration such as
`http_proxy' to be easily passed to downloaders like
`fetchurl'. Passing such environment variables from the caller
to the builder is generally impure, but the output of
fixed-output derivations is by definition pure (since we
already know the cryptographic hash of the output). */
if (fixedOutput) {
Strings varNames = tokenizeString(drv.env["impureEnvVars"]);
for (Strings::iterator i = varNames.begin(); i != varNames.end(); ++i)
env[*i] = getEnv(*i);
}
/* The `exportReferencesGraph' feature allows the references graph
to be passed to a builder. This attribute should be a list of
pairs [name1 path1 name2 path2 ...]. The references graph of
each `pathN' will be stored in a text file `nameN' in the
temporary build directory. The text files have the format used
by `nix-store --register-validity'. However, the deriver
fields are left empty. */
string s = drv.env["exportReferencesGraph"];
Strings ss = tokenizeString(s);
if (ss.size() % 2 != 0)
throw Error(format("odd number of tokens in `exportReferencesGraph': `%1%'") % s);
for (Strings::iterator i = ss.begin(); i != ss.end(); ) {
string fileName = *i++;
Path storePath = *i++;
if (!store->isValidPath(storePath))
throw Error(format("`exportReferencesGraph' refers to an invalid path `%1%'")
% storePath);
checkStoreName(fileName); /* !!! abuse of this function */
PathSet refs;
computeFSClosure(storePath, refs);
/* !!! in secure Nix, the writing should be done on the
build uid for security (maybe). */
writeStringToFile(tmpDir + "/" + fileName,
makeValidityRegistration(refs, false));
}
/* If `build-users-group' is not empty, then we have to build as
one of the members of that group. */
if (haveBuildUsers()) {
buildUser.acquire();
assert(buildUser.getUID() != 0);
assert(buildUser.getGID() != 0);
/* Make sure that no other processes are executing under this
uid. */
buildUser.kill();
/* Change ownership of the temporary build directory, if we're
root. If we're not root, then the setuid helper will do it
just before it starts the builder. */
if (amPrivileged()) {
if (chown(tmpDir.c_str(), buildUser.getUID(), buildUser.getGID()) == -1)
throw SysError(format("cannot change ownership of `%1%'") % tmpDir);
}
/* Check that the Nix store has the appropriate permissions,
i.e., owned by root and mode 1775 (sticky bit on so that
the builder can create its output but not mess with the
outputs of other processes). */
struct stat st;
if (stat(nixStore.c_str(), &st) == -1)
throw SysError(format("cannot stat `%1%'") % nixStore);
if (!(st.st_mode & S_ISVTX) ||
((st.st_mode & S_IRWXG) != S_IRWXG) ||
(st.st_gid != buildUser.getGID()))
throw Error(format(
"builder does not have write permission to `%2%'; "
"try `chgrp %1% %2%; chmod 1775 %2%'")
% buildUser.getGID() % nixStore);
}
/* Run the builder. */
printMsg(lvlChatty, format("executing builder `%1%'") %
drv.builder);
/* Create the log file and pipe. */
openLogFile();
/* Fork a child to build the package. Note that while we
currently use forks to run and wait for the children, it
shouldn't be hard to use threads for this on systems where
fork() is unavailable or inefficient. */
2004-06-22 12:51:44 +03:00
pid = fork();
switch (pid) {
case -1:
throw SysError("unable to fork");
case 0:
/* Warning: in the child we should absolutely not make any
Berkeley DB calls! */
try { /* child */
initChild();
/* Fill in the environment. */
Strings envStrs;
for (Environment::const_iterator i = env.begin();
2004-06-22 13:21:44 +03:00
i != env.end(); ++i)
envStrs.push_back(i->first + "=" + i->second);
const char * * envArr = strings2CharPtrs(envStrs);
Path program = drv.builder.c_str();
std::vector<const char *> args; /* careful with c_str()! */
string user; /* must be here for its c_str()! */
/* If we are running in `build-users' mode, then switch to
the user we allocated above. Make sure that we drop
all root privileges. Note that initChild() above has
closed all file descriptors except std*, so that's
safe. Also note that setuid() when run as root sets
the real, effective and saved UIDs. */
if (buildUser.enabled()) {
2006-12-07 23:43:35 +02:00
printMsg(lvlChatty, format("switching to user `%1%'") % buildUser.getUser());
if (amPrivileged()) {
if (setgroups(0, 0) == -1)
throw SysError("cannot clear the set of supplementary groups");
if (setgid(buildUser.getGID()) == -1 ||
getgid() != buildUser.getGID() ||
getegid() != buildUser.getGID())
throw SysError("setgid failed");
if (setuid(buildUser.getUID()) == -1 ||
getuid() != buildUser.getUID() ||
geteuid() != buildUser.getUID())
throw SysError("setuid failed");
} else {
/* Let the setuid helper take care of it. */
program = nixLibexecDir + "/nix-setuid-helper";
args.push_back(program.c_str());
args.push_back("run-builder");
user = buildUser.getUser().c_str();
args.push_back(user.c_str());
args.push_back(drv.builder.c_str());
}
}
/* Fill in the arguments. */
string builderBasename = baseNameOf(drv.builder);
args.push_back(builderBasename.c_str());
for (Strings::iterator i = drv.args.begin();
i != drv.args.end(); ++i)
args.push_back(i->c_str());
args.push_back(0);
/* Execute the program. This should not return. */
execve(program.c_str(), (char * *) &args[0], (char * *) envArr);
throw SysError(format("executing `%1%'")
% drv.builder);
} catch (std::exception & e) {
std::cerr << format("build error: %1%\n") % e.what();
2003-07-20 22:29:38 +03:00
}
quickExit(1);
}
/* parent */
2004-06-22 12:51:44 +03:00
pid.setSeparatePG(true);
logPipe.writeSide.close();
worker.childStarted(shared_from_this(), pid,
singleton<set<int> >(logPipe.readSide), true);
}
/* Parse a list of reference specifiers. Each element must either be
a store path, or the symbolic name of the output of the derivation
(such as `out'). */
PathSet parseReferenceSpecifiers(const Derivation & drv, string attr)
{
PathSet result;
Paths paths = tokenizeString(attr);
for (Strings::iterator i = paths.begin(); i != paths.end(); ++i) {
if (isStorePath(*i))
result.insert(*i);
else if (drv.outputs.find(*i) != drv.outputs.end())
result.insert(drv.outputs.find(*i)->second.path);
else throw Error(
format("derivation contains an illegal reference specifier `%1%'")
% *i);
}
return result;
}
void DerivationGoal::computeClosure()
{
map<Path, PathSet> allReferences;
map<Path, Hash> contentHashes;
/* Check whether the output paths were created, and grep each
output path to determine what other paths it references. Also make all
output paths read-only. */
for (DerivationOutputs::iterator i = drv.outputs.begin();
i != drv.outputs.end(); ++i)
{
* Removed the `id' attribute hack. * Formalise the notion of fixed-output derivations, i.e., derivations for which a cryptographic hash of the output is known in advance. Changes to such derivations should not propagate upwards through the dependency graph. Previously this was done by specifying the hash component of the output path through the `id' attribute, but this is insecure since you can lie about it (i.e., you can specify any hash and then produce a completely different output). Now the responsibility for checking the output is moved from the builder to Nix itself. A fixed-output derivation can be created by specifying the `outputHash' and `outputHashAlgo' attributes, the latter taking values `md5', `sha1', and `sha256', and the former specifying the actual hash in hexadecimal or in base-32 (auto-detected by looking at the length of the attribute value). MD5 is included for compatibility but should be considered deprecated. * Removed the `drvPath' pseudo-attribute in derivation results. It's no longer necessary. * Cleaned up the support for multiple output paths in derivation store expressions. Each output now has a unique identifier (e.g., `out', `devel', `docs'). Previously there was no way to tell output paths apart at the store expression level. * `nix-hash' now has a flag `--base32' to specify that the hash should be printed in base-32 notation. * `fetchurl' accepts parameters `sha256' and `sha1' in addition to `md5'. * `nix-prefetch-url' now prints out a SHA-1 hash in base-32. (TODO: a flag to specify the hash.)
2005-01-17 18:55:19 +02:00
Path path = i->second.path;
if (!pathExists(path)) {
throw BuildError(
2004-07-06 14:21:34 +03:00
format("builder for `%1%' failed to produce output path `%2%'")
% drvPath % path);
}
struct stat st;
if (lstat(path.c_str(), &st))
throw SysError(format("getting attributes of path `%1%'") % path);
2005-04-11 11:07:41 +03:00
startNest(nest, lvlTalkative,
format("scanning for references inside `%1%'") % path);
/* Check that fixed-output derivations produced the right
outputs (i.e., the content hash should match the specified
hash). */
if (i->second.hash != "") {
bool recursive = false;
string algo = i->second.hashAlgo;
if (string(algo, 0, 2) == "r:") {
recursive = true;
algo = string(algo, 2);
}
if (!recursive) {
/* The output path should be a regular file without
execute permission. */
if (!S_ISREG(st.st_mode) || (st.st_mode & S_IXUSR) != 0)
throw Error(
format("output path `%1% should be a non-executable regular file")
% path);
}
/* Check the hash. */
HashType ht = parseHashType(algo);
if (ht == htUnknown)
throw Error(format("unknown hash algorithm `%1%'") % algo);
Hash h = parseHash(ht, i->second.hash);
Hash h2 = recursive ? hashPath(ht, path) : hashFile(ht, path);
if (h != h2)
throw Error(
2005-05-10 17:21:46 +03:00
format("output path `%1%' should have %2% hash `%3%', instead has `%4%'")
% path % algo % printHash(h) % printHash(h2));
}
/* Get rid of all weird permissions. */
canonicalisePathMetaData(path);
/* For this output path, find the references to other paths contained
in it. */
PathSet references = scanForReferences(path, allPaths);
/* For debugging, print out the referenced and unreferenced
paths. */
for (PathSet::iterator i = inputPaths.begin();
i != inputPaths.end(); ++i)
{
PathSet::iterator j = references.find(*i);
if (j == references.end())
debug(format("unreferenced input: `%1%'") % *i);
else
debug(format("referenced input: `%1%'") % *i);
}
allReferences[path] = references;
/* If the derivation specifies an `allowedReferences'
attribute (containing a list of paths that the output may
refer to), check that all references are in that list. !!!
allowedReferences should really be per-output. */
if (drv.env.find("allowedReferences") != drv.env.end()) {
PathSet allowed = parseReferenceSpecifiers(drv, drv.env["allowedReferences"]);
for (PathSet::iterator i = references.begin(); i != references.end(); ++i)
if (allowed.find(*i) == allowed.end())
throw Error(format("output is not allowed to refer to path `%1%'") % *i);
}
/* Hash the contents of the path. The hash is stored in the
database so that we can verify later on whether nobody has
messed with the store. !!! inefficient: it would be nice
if we could combine this with filterReferences(). */
contentHashes[path] = hashPath(htSHA256, path);
}
/* Register each output path as valid, and register the sets of
paths referenced by each of them. This is wrapped in one
database transaction to ensure that if we crash, either
everything is registered or nothing is. This is for
recoverability: unregistered paths in the store can be deleted
arbitrarily, while registered paths can only be deleted by
running the garbage collector.
The reason that we do the transaction here and not on the fly
while we are scanning (above) is so that we don't hold database
locks for too long. */
Transaction txn;
createStoreTransaction(txn);
for (DerivationOutputs::iterator i = drv.outputs.begin();
i != drv.outputs.end(); ++i)
{
registerValidPath(txn, i->second.path,
contentHashes[i->second.path],
allReferences[i->second.path],
drvPath);
}
txn.commit();
/* It is now safe to delete the lock files, since all future
lockers will see that the output paths are valid; they will not
create new lock files with the same names as the old (unlinked)
lock files. */
outputLocks.setDeletion(true);
}
string drvsLogDir = "drvs";
void DerivationGoal::openLogFile()
{
/* Create a log file. */
Path dir = (format("%1%/%2%") % nixLogDir % drvsLogDir).str();
createDirs(dir);
Path logFileName = (format("%1%/%2%") % dir % baseNameOf(drvPath)).str();
fdLogFile = open(logFileName.c_str(),
O_CREAT | O_WRONLY | O_TRUNC, 0666);
if (fdLogFile == -1)
throw SysError(format("creating log file `%1%'") % logFileName);
/* Create a pipe to get the output of the child. */
logPipe.create();
}
void DerivationGoal::initChild()
{
2004-06-22 13:21:44 +03:00
commonChildInit(logPipe);
if (chdir(tmpDir.c_str()) == -1)
throw SysError(format("changing into `%1%'") % tmpDir);
/* When running a hook, dup the communication pipes. */
bool inHook = fromHook.writeSide.isOpen();
if (inHook) {
fromHook.readSide.close();
if (dup2(fromHook.writeSide, 3) == -1)
throw SysError("dupping from-hook write side");
toHook.writeSide.close();
if (dup2(toHook.readSide, 4) == -1)
throw SysError("dupping to-hook read side");
}
/* Close all other file descriptors. */
int maxFD = 0;
maxFD = sysconf(_SC_OPEN_MAX);
for (int fd = 0; fd < maxFD; ++fd)
if (fd != STDIN_FILENO && fd != STDOUT_FILENO && fd != STDERR_FILENO
&& (!inHook || (fd != 3 && fd != 4)))
close(fd); /* ignore result */
}
void DerivationGoal::deleteTmpDir(bool force)
{
if (tmpDir != "") {
if (keepFailed && !force) {
printMsg(lvlError,
format("builder for `%1%' failed; keeping build directory `%2%'")
% drvPath % tmpDir);
if (buildUser.enabled() && !amPrivileged())
getOwnership(tmpDir);
}
else
deletePathWrapped(tmpDir);
tmpDir = "";
}
}
void DerivationGoal::handleChildOutput(int fd, const string & data)
2004-06-29 12:41:50 +03:00
{
if (fd == logPipe.readSide) {
if (verbosity >= buildVerbosity)
writeToStderr((unsigned char *) data.c_str(), data.size());
writeFull(fdLogFile, (unsigned char *) data.c_str(), data.size());
}
else abort();
}
void DerivationGoal::handleEOF(int fd)
{
if (fd == logPipe.readSide) worker.wakeUp(shared_from_this());
2004-06-29 12:41:50 +03:00
}
PathSet DerivationGoal::checkPathValidity(bool returnValid)
{
PathSet result;
for (DerivationOutputs::iterator i = drv.outputs.begin();
i != drv.outputs.end(); ++i)
if (store->isValidPath(i->second.path)) {
if (returnValid) result.insert(i->second.path);
} else {
if (!returnValid) result.insert(i->second.path);
}
return result;
}
2003-07-20 22:29:38 +03:00
//////////////////////////////////////////////////////////////////////
2003-07-20 22:29:38 +03:00
class SubstitutionGoal : public Goal
{
private:
/* The store path that should be realised through a substitute. */
Path storePath;
/* The remaining substitutes for this path. */
Substitutes subs;
/* The current substitute. */
Substitute sub;
/* Outgoing references for this path. */
PathSet references;
/* Pipe for the substitute's standard output/error. */
Pipe logPipe;
/* The process ID of the builder. */
2004-06-22 12:51:44 +03:00
Pid pid;
/* Lock on the store path. */
boost::shared_ptr<PathLocks> outputLock;
typedef void (SubstitutionGoal::*GoalState)();
GoalState state;
public:
SubstitutionGoal(const Path & storePath, Worker & worker);
~SubstitutionGoal();
void work();
/* The states. */
void init();
void referencesValid();
void tryNext();
void tryToRun();
void finished();
2004-06-22 20:04:10 +03:00
2004-06-29 12:41:50 +03:00
/* Callback used by the worker to write to the log. */
void handleChildOutput(int fd, const string & data);
void handleEOF(int fd);
};
SubstitutionGoal::SubstitutionGoal(const Path & storePath, Worker & worker)
: Goal(worker)
{
this->storePath = storePath;
state = &SubstitutionGoal::init;
name = (format("substitution of `%1%'") % storePath).str();
trace("created");
}
SubstitutionGoal::~SubstitutionGoal()
{
if (pid != -1) worker.childTerminated(pid);
}
void SubstitutionGoal::work()
{
(this->*state)();
}
void SubstitutionGoal::init()
{
2004-06-22 20:04:10 +03:00
trace("init");
store->addTempRoot(storePath);
/* If the path already exists we're done. */
if (store->isValidPath(storePath)) {
amDone();
return;
}
/* !!! race condition; should get the substitutes and the
references in a transaction (in case a clearSubstitutes() is
done simultaneously). */
/* Read the substitutes. */
subs = store->querySubstitutes(storePath);
2005-09-20 19:14:00 +03:00
/* To maintain the closure invariant, we first have to realise the
paths referenced by this one. */
store->queryReferences(storePath, references);
for (PathSet::iterator i = references.begin();
i != references.end(); ++i)
if (*i != storePath) /* ignore self-references */
addWaitee(worker.makeSubstitutionGoal(*i));
if (waitees.empty()) /* to prevent hang (no wake-up event) */
referencesValid();
else
state = &SubstitutionGoal::referencesValid;
}
void SubstitutionGoal::referencesValid()
{
trace("all referenced realised");
if (nrFailed > 0)
throw Error(format("some references of path `%1%' could not be realised") % storePath);
for (PathSet::iterator i = references.begin();
i != references.end(); ++i)
if (*i != storePath) /* ignore self-references */
assert(store->isValidPath(*i));
tryNext();
}
void SubstitutionGoal::tryNext()
{
2004-06-22 20:04:10 +03:00
trace("trying next substitute");
if (subs.size() == 0) {
/* None left. Terminate this goal and let someone else deal
with it. */
printMsg(lvlError,
format("path `%1%' is required, but it has no (remaining) substitutes")
% storePath);
amDone(false);
return;
}
sub = subs.front();
subs.pop_front();
/* Wait until we can run the substitute program. */
state = &SubstitutionGoal::tryToRun;
worker.waitForBuildSlot(shared_from_this());
}
void SubstitutionGoal::tryToRun()
{
2004-06-22 20:04:10 +03:00
trace("trying to run");
/* Make sure that we are allowed to start a build. */
if (!worker.canBuildMore()) {
worker.waitForBuildSlot(shared_from_this());
return;
}
/* Acquire a lock on the output path. */
outputLock = boost::shared_ptr<PathLocks>(new PathLocks);
outputLock->lockPaths(singleton<PathSet>(storePath),
(format("waiting for lock on `%1%'") % storePath).str());
/* Check again whether the path is invalid. */
if (store->isValidPath(storePath)) {
debug(format("store path `%1%' has become valid") % storePath);
outputLock->setDeletion(true);
amDone();
return;
}
printMsg(lvlInfo,
format("substituting path `%1%' using substituter `%2%'")
% storePath % sub.program);
2004-06-22 20:04:10 +03:00
logPipe.create();
/* Remove the (stale) output path if it exists. */
if (pathExists(storePath))
deletePathWrapped(storePath);
/* Fork the substitute program. */
2004-06-22 12:51:44 +03:00
pid = fork();
switch (pid) {
case -1:
throw SysError("unable to fork");
case 0:
try { /* child */
logPipe.readSide.close();
/* !!! close other handles */
2004-06-22 13:21:44 +03:00
commonChildInit(logPipe);
2004-06-22 13:21:44 +03:00
/* Fill in the arguments. */
Strings args(sub.args);
args.push_front(storePath);
args.push_front(baseNameOf(sub.program));
2004-06-22 13:21:44 +03:00
const char * * argArr = strings2CharPtrs(args);
execv(sub.program.c_str(), (char * *) argArr);
throw SysError(format("executing `%1%'") % sub.program);
} catch (std::exception & e) {
std::cerr << format("substitute error: %1%\n") % e.what();
}
quickExit(1);
}
/* parent */
2004-06-22 12:51:44 +03:00
pid.setSeparatePG(true);
logPipe.writeSide.close();
worker.childStarted(shared_from_this(),
pid, singleton<set<int> >(logPipe.readSide), true);
state = &SubstitutionGoal::finished;
}
void SubstitutionGoal::finished()
{
2004-06-22 20:04:10 +03:00
trace("substitute finished");
/* Since we got an EOF on the logger pipe, the substitute is
presumed to have terminated. */
/* !!! this could block! */
2004-06-22 12:51:44 +03:00
pid_t savedPid = pid;
int status = pid.wait(true);
/* So the child is gone now. */
2004-06-22 12:51:44 +03:00
worker.childTerminated(savedPid);
/* Close the read side of the logger pipe. */
logPipe.readSide.close();
debug(format("substitute for `%1%' finished") % storePath);
/* Check the exit status and the build result. */
try {
if (!statusOk(status))
throw SubstError(format("builder for `%1%' %2%")
% storePath % statusToString(status));
if (!pathExists(storePath))
throw SubstError(
format("substitute did not produce path `%1%'")
% storePath);
} catch (SubstError & e) {
printMsg(lvlInfo,
format("substitution of path `%1%' using substituter `%2%' failed: %3%")
% storePath % sub.program % e.msg());
/* Try the next substitute. */
state = &SubstitutionGoal::tryNext;
worker.wakeUp(shared_from_this());
return;
}
canonicalisePathMetaData(storePath);
Hash contentHash = hashPath(htSHA256, storePath);
Transaction txn;
createStoreTransaction(txn);
registerValidPath(txn, storePath, contentHash,
references, sub.deriver);
txn.commit();
outputLock->setDeletion(true);
printMsg(lvlChatty,
format("substitution of path `%1%' succeeded") % storePath);
amDone();
}
void SubstitutionGoal::handleChildOutput(int fd, const string & data)
2004-06-29 12:41:50 +03:00
{
assert(fd == logPipe.readSide);
if (verbosity >= buildVerbosity)
writeToStderr((unsigned char *) data.c_str(), data.size());
2004-06-29 12:41:50 +03:00
/* Don't write substitution output to a log file for now. We
probably should, though. */
}
void SubstitutionGoal::handleEOF(int fd)
{
if (fd == logPipe.readSide) worker.wakeUp(shared_from_this());
}
//////////////////////////////////////////////////////////////////////
static bool working = false;
Worker::Worker()
{
/* Debugging: prevent recursive workers. */
if (working) abort();
working = true;
nrChildren = 0;
}
Worker::~Worker()
{
working = false;
/* Explicitly get rid of all strong pointers now. After this all
goals that refer to this worker should be gone. (Otherwise we
are in trouble, since goals may call childTerminated() etc. in
their destructors). */
topGoals.clear();
}
template<class T>
static GoalPtr addGoal(const Path & path,
Worker & worker, WeakGoalMap & goalMap)
{
GoalPtr goal = goalMap[path].lock();
if (!goal) {
goal = GoalPtr(new T(path, worker));
goalMap[path] = goal;
worker.wakeUp(goal);
2003-07-20 22:29:38 +03:00
}
return goal;
}
2003-07-20 22:29:38 +03:00
GoalPtr Worker::makeDerivationGoal(const Path & nePath)
{
return addGoal<DerivationGoal>(nePath, *this, derivationGoals);
}
GoalPtr Worker::makeSubstitutionGoal(const Path & storePath)
{
return addGoal<SubstitutionGoal>(storePath, *this, substitutionGoals);
}
static void removeGoal(GoalPtr goal, WeakGoalMap & goalMap)
{
/* !!! inefficient */
for (WeakGoalMap::iterator i = goalMap.begin();
i != goalMap.end(); )
if (i->second.lock() == goal) {
WeakGoalMap::iterator j = i; ++j;
goalMap.erase(i);
i = j;
}
else ++i;
}
void Worker::removeGoal(GoalPtr goal)
{
nix::removeGoal(goal, derivationGoals);
nix::removeGoal(goal, substitutionGoals);
if (topGoals.find(goal) != topGoals.end()) {
topGoals.erase(goal);
/* If a top-level goal failed, then kill all other goals
(unless keepGoing was set). */
if (goal->getExitCode() == Goal::ecFailed && !keepGoing)
topGoals.clear();
}
}
2003-11-21 18:05:19 +02:00
void Worker::wakeUp(GoalPtr goal)
{
goal->trace("woken up");
awake.insert(goal);
}
bool Worker::canBuildMore()
{
2004-06-20 00:45:04 +03:00
return nrChildren < maxBuildJobs;
2003-07-20 22:29:38 +03:00
}
2004-06-20 00:45:04 +03:00
void Worker::childStarted(GoalPtr goal,
pid_t pid, const set<int> & fds, bool inBuildSlot)
2003-07-20 22:29:38 +03:00
{
2004-06-20 00:45:04 +03:00
Child child;
child.goal = goal;
child.fds = fds;
child.lastOutput = time(0);
2004-06-20 00:45:04 +03:00
child.inBuildSlot = inBuildSlot;
children[pid] = child;
if (inBuildSlot) nrChildren++;
}
2003-07-20 22:29:38 +03:00
void Worker::childTerminated(pid_t pid, bool wakeSleepers)
{
2004-06-20 00:45:04 +03:00
Children::iterator i = children.find(pid);
assert(i != children.end());
if (i->second.inBuildSlot) {
assert(nrChildren > 0);
nrChildren--;
}
children.erase(pid);
if (wakeSleepers) {
/* Wake up goals waiting for a build slot. */
for (WeakGoals::iterator i = wantingToBuild.begin();
i != wantingToBuild.end(); ++i)
{
GoalPtr goal = i->lock();
if (goal) wakeUp(goal);
}
wantingToBuild.clear();
}
}
void Worker::waitForBuildSlot(GoalPtr goal, bool reallyWait)
{
debug("wait for build slot");
if (reallyWait && children.size() == 0)
throw Error("waiting for a build slot, yet there are no children - "
"maybe the build hook gave an inappropriate `postpone' reply?");
if (!reallyWait && canBuildMore())
wakeUp(goal); /* we can do it right away */
else
wantingToBuild.insert(goal);
}
void Worker::run(const Goals & _topGoals)
{
for (Goals::iterator i = _topGoals.begin();
i != _topGoals.end(); ++i)
topGoals.insert(*i);
startNest(nest, lvlDebug, format("entered goal loop"));
while (1) {
checkInterrupt();
/* Call every wake goal. */
while (!awake.empty() && !topGoals.empty()) {
WeakGoals awake2(awake);
awake.clear();
for (WeakGoals::iterator i = awake2.begin(); i != awake2.end(); ++i) {
checkInterrupt();
GoalPtr goal = i->lock();
if (goal) goal->work();
if (topGoals.empty()) break;
}
}
if (topGoals.empty()) break;
/* !!! not when we're polling */
assert(!children.empty());
/* Wait for input. */
waitForInput();
}
/* If --keep-going is not set, it's possible that the main goal
exited while some of its subgoals were still active. But if
--keep-going *is* set, then they must all be finished now. */
assert(!keepGoing || awake.empty());
assert(!keepGoing || wantingToBuild.empty());
assert(!keepGoing || children.empty());
2003-07-20 22:29:38 +03:00
}
void Worker::waitForInput()
2003-07-20 22:29:38 +03:00
{
printMsg(lvlVomit, "waiting for children");
2003-07-20 22:29:38 +03:00
/* Process output from the file descriptors attached to the
children, namely log output and output path creation commands.
We also use this to detect child termination: if we get EOF on
the logger pipe of a build, we assume that the builder has
terminated. */
/* If we're monitoring for silence on stdout/stderr, sleep until
the first deadline for any child. */
struct timeval timeout;
if (maxSilentTime != 0) {
time_t oldest = 0;
for (Children::iterator i = children.begin();
i != children.end(); ++i)
{
oldest = oldest == 0 || i->second.lastOutput < oldest
? i->second.lastOutput : oldest;
}
time_t now = time(0);
timeout.tv_sec = (time_t) (oldest + maxSilentTime) <= now ? 0 :
oldest + maxSilentTime - now;
timeout.tv_usec = 0;
printMsg(lvlVomit, format("sleeping %1% seconds") % timeout.tv_sec);
}
/* Use select() to wait for the input side of any logger pipe to
become `available'. Note that `available' (i.e., non-blocking)
includes EOF. */
fd_set fds;
FD_ZERO(&fds);
int fdMax = 0;
for (Children::iterator i = children.begin();
i != children.end(); ++i)
{
for (set<int>::iterator j = i->second.fds.begin();
j != i->second.fds.end(); ++j)
{
FD_SET(*j, &fds);
if (*j >= fdMax) fdMax = *j + 1;
}
}
2003-07-20 22:29:38 +03:00
if (select(fdMax, &fds, 0, 0, maxSilentTime != 0 ? &timeout : 0) == -1) {
if (errno == EINTR) return;
throw SysError("waiting for input");
}
2003-07-20 22:29:38 +03:00
time_t now = time(0);
/* Process all available file descriptors. */
for (Children::iterator i = children.begin();
i != children.end(); ++i)
{
checkInterrupt();
GoalPtr goal = i->second.goal.lock();
assert(goal);
set<int> fds2(i->second.fds);
for (set<int>::iterator j = fds2.begin(); j != fds2.end(); ++j) {
if (FD_ISSET(*j, &fds)) {
unsigned char buffer[4096];
ssize_t rd = read(*j, buffer, sizeof(buffer));
if (rd == -1) {
if (errno != EINTR)
throw SysError(format("reading from %1%")
% goal->getName());
} else if (rd == 0) {
debug(format("%1%: got EOF") % goal->getName());
goal->handleEOF(*j);
i->second.fds.erase(*j);
} else {
printMsg(lvlVomit, format("%1%: read %2% bytes")
% goal->getName() % rd);
string data((char *) buffer, rd);
goal->handleChildOutput(*j, data);
i->second.lastOutput = now;
}
}
}
if (maxSilentTime != 0 &&
now - i->second.lastOutput >= (time_t) maxSilentTime)
throw Error(format("%1% timed out after %2% seconds of silence")
% goal->getName() % maxSilentTime);
}
2003-07-20 22:29:38 +03:00
}
//////////////////////////////////////////////////////////////////////
2006-11-30 20:02:04 +02:00
void LocalStore::buildDerivations(const PathSet & drvPaths)
{
startNest(nest, lvlDebug,
format("building %1%") % showPaths(drvPaths));
Worker worker;
Goals goals;
for (PathSet::const_iterator i = drvPaths.begin();
i != drvPaths.end(); ++i)
goals.insert(worker.makeDerivationGoal(*i));
worker.run(goals);
PathSet failed;
for (Goals::iterator i = goals.begin(); i != goals.end(); ++i)
if ((*i)->getExitCode() == Goal::ecFailed) {
DerivationGoal * i2 = dynamic_cast<DerivationGoal *>(i->get());
assert(i2);
failed.insert(i2->getDrvPath());
}
if (!failed.empty())
throw Error(format("build of %1% failed") % showPaths(failed));
}
2006-11-30 20:02:04 +02:00
void LocalStore::ensurePath(const Path & path)
2003-07-20 22:29:38 +03:00
{
/* If the path is already valid, we're done. */
if (store->isValidPath(path)) return;
2004-06-22 12:51:44 +03:00
Worker worker;
GoalPtr goal = worker.makeSubstitutionGoal(path);
Goals goals = singleton<Goals>(goal);
worker.run(goals);
if (goal->getExitCode() != Goal::ecSuccess)
throw Error(format("path `%1%' does not exist and cannot be created") % path);
2003-07-20 22:29:38 +03:00
}
}