mklockrank.go

     1  // Copyright 2022 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  //go:build ignore
     6  
     7  // mklockrank records the static rank graph of the locks in the
     8  // runtime and generates the rank checking structures in lockrank.go.
     9  package main
    10  
    11  import (
    12  	"bytes"
    13  	"flag"
    14  	"fmt"
    15  	"go/format"
    16  	"internal/dag"
    17  	"io"
    18  	"log"
    19  	"os"
    20  	"strings"
    21  )
    22  
    23  // ranks describes the lock rank graph. See "go doc internal/dag" for
    24  // the syntax.
    25  //
    26  // "a < b" means a must be acquired before b if both are held
    27  // (or, if b is held, a cannot be acquired).
    28  //
    29  // "NONE < a" means no locks may be held when a is acquired.
    30  //
    31  // If a lock is not given a rank, then it is assumed to be a leaf
    32  // lock, which means no other lock can be acquired while it is held.
    33  // Therefore, leaf locks do not need to be given an explicit rank.
    34  //
    35  // Ranks in all caps are pseudo-nodes that help define order, but do
    36  // not actually define a rank.
    37  //
    38  // TODO: It's often hard to correlate rank names to locks. Change
    39  // these to be more consistent with the locks they label.
    40  const ranks = `
    41  # Sysmon
    42  NONE
    43  < sysmon
    44  < scavenge, forcegc, computeMaxProcs, updateMaxProcsG;
    45  
    46  # Defer
    47  NONE < defer;
    48  
    49  # GC
    50  NONE <
    51    sweepWaiters,
    52    assistQueue,
    53    strongFromWeakQueue,
    54    cleanupQueue,
    55    sweep;
    56  
    57  # Test only
    58  NONE < testR, testW;
    59  
    60  # vgetrandom
    61  NONE < vgetrandom;
    62  
    63  NONE < timerSend;
    64  
    65  # Scheduler, timers, netpoll
    66  NONE < allocmW, execW, cpuprof, pollCache, pollDesc, wakeableSleep;
    67  scavenge, sweep, testR, wakeableSleep, timerSend < hchan;
    68  assistQueue,
    69    cleanupQueue,
    70    computeMaxProcs,
    71    cpuprof,
    72    forcegc,
    73    updateMaxProcsG,
    74    hchan,
    75    pollDesc, # pollDesc can interact with timers, which can lock sched.
    76    scavenge,
    77    strongFromWeakQueue,
    78    sweep,
    79    sweepWaiters,
    80    testR,
    81    wakeableSleep
    82  # Above SCHED are things that can call into the scheduler.
    83  < SCHED
    84  # Below SCHED is the scheduler implementation.
    85  < allocmR,
    86    execR;
    87  allocmR, execR, hchan < sched;
    88  sched < allg, allp;
    89  
    90  # Channels
    91  NONE < notifyList;
    92  hchan, notifyList < sudog;
    93  
    94  hchan, pollDesc, wakeableSleep < timers;
    95  timers, timerSend < timer < netpollInit;
    96  
    97  # Semaphores
    98  NONE < root;
    99  
   100  # Itabs
   101  NONE
   102  < itab
   103  < reflectOffs;
   104  
   105  # Synctest
   106  hchan,
   107    notifyList,
   108    reflectOffs,
   109    root,
   110    strongFromWeakQueue,
   111    sweepWaiters,
   112    timer,
   113    timers
   114  < synctest;
   115  
   116  # User arena state
   117  NONE < userArenaState;
   118  
   119  # Tracing without a P uses a global trace buffer.
   120  scavenge
   121  # Above TRACEGLOBAL can emit a trace event without a P.
   122  < TRACEGLOBAL
   123  # Below TRACEGLOBAL manages the global tracing buffer.
   124  # Note that traceBuf eventually chains to MALLOC, but we never get that far
   125  # in the situation where there's no P.
   126  < traceBuf;
   127  # Starting/stopping tracing traces strings.
   128  traceBuf < traceStrings;
   129  
   130  # Malloc
   131  allg,
   132    allocmR,
   133    allp, # procresize
   134    execR, # May grow stack
   135    execW, # May allocate after BeforeFork
   136    hchan,
   137    notifyList,
   138    reflectOffs,
   139    timer,
   140    traceStrings,
   141    userArenaState,
   142    vgetrandom
   143  # Above MALLOC are things that can allocate memory.
   144  < MALLOC
   145  # Below MALLOC is the malloc implementation.
   146  < fin,
   147    spanSetSpine,
   148    mspanSpecial,
   149    traceTypeTab,
   150    MPROF;
   151  
   152  # We can acquire gcBitsArenas for pinner bits, and
   153  # it's guarded by mspanSpecial.
   154  MALLOC, mspanSpecial < gcBitsArenas;
   155  
   156  # Memory profiling
   157  MPROF < profInsert, profBlock, profMemActive;
   158  profMemActive < profMemFuture;
   159  
   160  # Stack allocation and copying
   161  gcBitsArenas,
   162    netpollInit,
   163    profBlock,
   164    profInsert,
   165    profMemFuture,
   166    spanSetSpine,
   167    synctest,
   168    fin,
   169    root
   170  # Anything that can grow the stack can acquire STACKGROW.
   171  # (Most higher layers imply STACKGROW, like MALLOC.)
   172  < STACKGROW
   173  # Below STACKGROW is the stack allocator/copying implementation.
   174  < gscan;
   175  gscan < stackpool;
   176  gscan < stackLarge;
   177  # Generally, hchan must be acquired before gscan. But in one case,
   178  # where we suspend a G and then shrink its stack, syncadjustsudogs
   179  # can acquire hchan locks while holding gscan. To allow this case,
   180  # we use hchanLeaf instead of hchan.
   181  gscan < hchanLeaf;
   182  
   183  # Write barrier
   184  defer,
   185    gscan,
   186    mspanSpecial,
   187    pollCache,
   188    sudog,
   189    timer
   190  # Anything that can have write barriers can acquire WB.
   191  # Above WB, we can have write barriers.
   192  < WB
   193  # Below WB is the write barrier implementation.
   194  < wbufSpans;
   195  
   196  # Span allocator
   197  stackLarge,
   198    stackpool,
   199    wbufSpans
   200  # Above mheap is anything that can call the span allocator.
   201  < mheap;
   202  # Below mheap is the span allocator implementation.
   203  #
   204  # Specials: we're allowed to allocate a special while holding
   205  # an mspanSpecial lock, and they're part of the malloc implementation.
   206  # Pinner bits might be freed by the span allocator.
   207  mheap, mspanSpecial < mheapSpecial;
   208  mheap, mheapSpecial < globalAlloc;
   209  
   210  # Execution tracer events (with a P)
   211  hchan,
   212    mheap,
   213    root,
   214    sched,
   215    traceStrings,
   216    notifyList,
   217    fin
   218  # Above TRACE is anything that can create a trace event
   219  < TRACE
   220  < trace
   221  < traceStackTab;
   222  
   223  # panic is handled specially. It is implicitly below all other locks.
   224  NONE < panic;
   225  # deadlock is not acquired while holding panic, but it also needs to be
   226  # below all other locks.
   227  panic < deadlock;
   228  # raceFini is only held while exiting.
   229  panic < raceFini;
   230  
   231  # RWMutex internal read lock
   232  
   233  allocmR,
   234    allocmW
   235  < allocmRInternal;
   236  
   237  execR,
   238    execW
   239  < execRInternal;
   240  
   241  testR,
   242    testW
   243  < testRInternal;
   244  `
   245  
   246  // cyclicRanks lists lock ranks that allow multiple locks of the same
   247  // rank to be acquired simultaneously. The runtime enforces ordering
   248  // within these ranks using a separate mechanism.
   249  var cyclicRanks = map[string]bool{
   250  	// Multiple timers are locked simultaneously in destroy().
   251  	"timers": true,
   252  	// Multiple hchans are acquired in hchan.sortkey() order in
   253  	// select.
   254  	"hchan": true,
   255  	// Multiple hchanLeafs are acquired in hchan.sortkey() order in
   256  	// syncadjustsudogs().
   257  	"hchanLeaf": true,
   258  	// The point of the deadlock lock is to deadlock.
   259  	"deadlock": true,
   260  }
   261  
   262  func main() {
   263  	flagO := flag.String("o", "", "write to `file` instead of stdout")
   264  	flagDot := flag.Bool("dot", false, "emit graphviz output instead of Go")
   265  	flag.Parse()
   266  	if flag.NArg() != 0 {
   267  		fmt.Fprintf(os.Stderr, "too many arguments")
   268  		os.Exit(2)
   269  	}
   270  
   271  	g, err := dag.Parse(ranks)
   272  	if err != nil {
   273  		log.Fatal(err)
   274  	}
   275  
   276  	var out []byte
   277  	if *flagDot {
   278  		var b bytes.Buffer
   279  		g.TransitiveReduction()
   280  		// Add cyclic edges for visualization.
   281  		for k := range cyclicRanks {
   282  			g.AddEdge(k, k)
   283  		}
   284  		// Reverse the graph. It's much easier to read this as
   285  		// a "<" partial order than a ">" partial order. This
   286  		// ways, locks are acquired from the top going down
   287  		// and time moves forward over the edges instead of
   288  		// backward.
   289  		g.Transpose()
   290  		generateDot(&b, g)
   291  		out = b.Bytes()
   292  	} else {
   293  		var b bytes.Buffer
   294  		generateGo(&b, g)
   295  		out, err = format.Source(b.Bytes())
   296  		if err != nil {
   297  			log.Fatal(err)
   298  		}
   299  	}
   300  
   301  	if *flagO != "" {
   302  		err = os.WriteFile(*flagO, out, 0666)
   303  	} else {
   304  		_, err = os.Stdout.Write(out)
   305  	}
   306  	if err != nil {
   307  		log.Fatal(err)
   308  	}
   309  }
   310  
   311  func generateGo(w io.Writer, g *dag.Graph) {
   312  	fmt.Fprintf(w, `// Code generated by mklockrank.go; DO NOT EDIT.
   313  
   314  package runtime
   315  
   316  type lockRank int
   317  
   318  `)
   319  
   320  	// Create numeric ranks.
   321  	topo := g.Topo()
   322  	for i, j := 0, len(topo)-1; i < j; i, j = i+1, j-1 {
   323  		topo[i], topo[j] = topo[j], topo[i]
   324  	}
   325  	fmt.Fprintf(w, `
   326  // Constants representing the ranks of all non-leaf runtime locks, in rank order.
   327  // Locks with lower rank must be taken before locks with higher rank,
   328  // in addition to satisfying the partial order in lockPartialOrder.
   329  // A few ranks allow self-cycles, which are specified in lockPartialOrder.
   330  const (
   331  	lockRankUnknown lockRank = iota
   332  
   333  `)
   334  	for _, rank := range topo {
   335  		if isPseudo(rank) {
   336  			fmt.Fprintf(w, "\t// %s\n", rank)
   337  		} else {
   338  			fmt.Fprintf(w, "\t%s\n", cname(rank))
   339  		}
   340  	}
   341  	fmt.Fprintf(w, `)
   342  
   343  // lockRankLeafRank is the rank of lock that does not have a declared rank,
   344  // and hence is a leaf lock.
   345  const lockRankLeafRank lockRank = 1000
   346  `)
   347  
   348  	// Create string table.
   349  	fmt.Fprintf(w, `
   350  // lockNames gives the names associated with each of the above ranks.
   351  var lockNames = []string{
   352  `)
   353  	for _, rank := range topo {
   354  		if !isPseudo(rank) {
   355  			fmt.Fprintf(w, "\t%s: %q,\n", cname(rank), rank)
   356  		}
   357  	}
   358  	fmt.Fprintf(w, `}
   359  
   360  func (rank lockRank) String() string {
   361  	if rank == 0 {
   362  		return "UNKNOWN"
   363  	}
   364  	if rank == lockRankLeafRank {
   365  		return "LEAF"
   366  	}
   367  	if rank < 0 || int(rank) >= len(lockNames) {
   368  		return "BAD RANK"
   369  	}
   370  	return lockNames[rank]
   371  }
   372  `)
   373  
   374  	// Create partial order structure.
   375  	fmt.Fprintf(w, `
   376  // lockPartialOrder is the transitive closure of the lock rank graph.
   377  // An entry for rank X lists all of the ranks that can already be held
   378  // when rank X is acquired.
   379  //
   380  // Lock ranks that allow self-cycles list themselves.
   381  var lockPartialOrder [][]lockRank = [][]lockRank{
   382  `)
   383  	for _, rank := range topo {
   384  		if isPseudo(rank) {
   385  			continue
   386  		}
   387  		list := []string{}
   388  		for _, before := range g.Edges(rank) {
   389  			if !isPseudo(before) {
   390  				list = append(list, cname(before))
   391  			}
   392  		}
   393  		if cyclicRanks[rank] {
   394  			list = append(list, cname(rank))
   395  		}
   396  
   397  		fmt.Fprintf(w, "\t%s: {%s},\n", cname(rank), strings.Join(list, ", "))
   398  	}
   399  	fmt.Fprintf(w, "}\n")
   400  }
   401  
   402  // cname returns the Go const name for the given lock rank label.
   403  func cname(label string) string {
   404  	return "lockRank" + strings.ToUpper(label[:1]) + label[1:]
   405  }
   406  
   407  func isPseudo(label string) bool {
   408  	return strings.ToUpper(label) == label
   409  }
   410  
   411  // generateDot emits a Graphviz dot representation of g to w.
   412  func generateDot(w io.Writer, g *dag.Graph) {
   413  	fmt.Fprintf(w, "digraph g {\n")
   414  
   415  	// Define all nodes.
   416  	for _, node := range g.Nodes {
   417  		fmt.Fprintf(w, "%q;\n", node)
   418  	}
   419  
   420  	// Create edges.
   421  	for _, node := range g.Nodes {
   422  		for _, to := range g.Edges(node) {
   423  			fmt.Fprintf(w, "%q -> %q;\n", node, to)
   424  		}
   425  	}
   426  
   427  	fmt.Fprintf(w, "}\n")
   428  }
   429
View as plain text