gvisor/pkg/refsvfs2/refs_map.go

139 lines
4.1 KiB
Go
Raw Normal View History

Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
// Copyright 2020 The gVisor Authors.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package refsvfs2
import (
"fmt"
"gvisor.dev/gvisor/pkg/log"
refs_vfs1 "gvisor.dev/gvisor/pkg/refs"
"gvisor.dev/gvisor/pkg/sync"
)
var (
// liveObjects is a global map of reference-counted objects. Objects are
// inserted when leak check is enabled, and they are removed when they are
// destroyed. It is protected by liveObjectsMu.
liveObjects map[CheckedObject]struct{}
liveObjectsMu sync.Mutex
)
// CheckedObject represents a reference-counted object with an informative
// leak detection message.
type CheckedObject interface {
// RefType is the type of the reference-counted object.
RefType() string
Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
// LeakMessage supplies a warning to be printed upon leak detection.
LeakMessage() string
// LogRefs indicates whether reference-related events should be logged.
LogRefs() bool
Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
}
func init() {
liveObjects = make(map[CheckedObject]struct{})
}
// leakCheckEnabled returns whether leak checking is enabled. The following
Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
// functions should only be called if it returns true.
func leakCheckEnabled() bool {
Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
return refs_vfs1.GetLeakMode() != refs_vfs1.NoLeakChecking
}
// Register adds obj to the live object map.
func Register(obj CheckedObject) {
if leakCheckEnabled() {
liveObjectsMu.Lock()
if _, ok := liveObjects[obj]; ok {
panic(fmt.Sprintf("Unexpected entry in leak checking map: reference %p already added", obj))
}
liveObjects[obj] = struct{}{}
liveObjectsMu.Unlock()
if leakCheckEnabled() && obj.LogRefs() {
logEvent(obj, "registered")
}
Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
}
}
// Unregister removes obj from the live object map.
func Unregister(obj CheckedObject) {
if leakCheckEnabled() {
liveObjectsMu.Lock()
defer liveObjectsMu.Unlock()
if _, ok := liveObjects[obj]; !ok {
panic(fmt.Sprintf("Expected to find entry in leak checking map for reference %p", obj))
Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
}
delete(liveObjects, obj)
if leakCheckEnabled() && obj.LogRefs() {
logEvent(obj, "unregistered")
}
}
}
// LogIncRef logs a reference increment.
func LogIncRef(obj CheckedObject, refs int64) {
if leakCheckEnabled() && obj.LogRefs() {
logEvent(obj, fmt.Sprintf("IncRef to %d", refs))
}
}
// LogTryIncRef logs a successful TryIncRef call.
func LogTryIncRef(obj CheckedObject, refs int64) {
if leakCheckEnabled() && obj.LogRefs() {
logEvent(obj, fmt.Sprintf("TryIncRef to %d", refs))
}
}
// LogDecRef logs a reference decrement.
func LogDecRef(obj CheckedObject, refs int64) {
if leakCheckEnabled() && obj.LogRefs() {
logEvent(obj, fmt.Sprintf("DecRef to %d", refs))
}
}
// logEvent logs a message for the given reference-counted object.
//
// obj.LogRefs() should be checked before calling logEvent, in order to avoid
// calling any text processing needed to evaluate msg.
func logEvent(obj CheckedObject, msg string) {
log.Infof("[%s %p] %s:\n%s", obj.RefType(), obj, msg, refs_vfs1.FormatStack(refs_vfs1.RecordStack()))
Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
}
// checkOnce makes sure that leak checking is only done once. DoLeakCheck is
// called from multiple places (which may overlap) to cover different sandbox
// exit scenarios.
var checkOnce sync.Once
Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
// DoLeakCheck iterates through the live object map and logs a message for each
// object. It is called once no reference-counted objects should be reachable
// anymore, at which point anything left in the map is considered a leak.
func DoLeakCheck() {
if leakCheckEnabled() {
checkOnce.Do(func() {
liveObjectsMu.Lock()
defer liveObjectsMu.Unlock()
leaked := len(liveObjects)
if leaked > 0 {
msg := fmt.Sprintf("Leak checking detected %d leaked objects:\n", leaked)
for obj := range liveObjects {
msg += obj.LeakMessage() + "\n"
}
log.Warningf(msg)
}
})
Rewrite reference leak checker without finalizers. Our current reference leak checker uses finalizers to verify whether an object has reached zero references before it is garbage collected. There are multiple problems with this mechanism, so a rewrite is in order. With finalizers, there is no way to guarantee that a finalizer will run before the program exits. When an unreachable object with a finalizer is garbage collected, its finalizer will be added to a queue and run asynchronously. The best we can do is run garbage collection upon sandbox exit to make sure that all finalizers are enqueued. Furthermore, if there is a chain of finalized objects, e.g. A points to B points to C, garbage collection needs to run multiple times before all of the finalizers are enqueued. The first GC run will register the finalizer for A but not free it. It takes another GC run to free A, at which point B's finalizer can be registered. As a result, we need to run GC as many times as the length of the longest such chain to have a somewhat reliable leak checker. Finally, a cyclical chain of structs pointing to one another will never be garbage collected if a finalizer is set. This is a well-known issue with Go finalizers (https://github.com/golang/go/issues/7358). Using leak checking on filesystem objects that produce cycles will not work and even result in memory leaks. The new leak checker stores reference counted objects in a global map when leak check is enabled and removes them once they are destroyed. At sandbox exit, any remaining objects in the map are considered as leaked. This provides a deterministic way of detecting leaks without relying on the complexities of finalizers and garbage collection. This approach has several benefits over the former, including: - Always detects leaks of objects that should be destroyed very close to sandbox exit. The old checker very rarely detected these leaks, because it relied on garbage collection to be run in a short window of time. - Panics if we forgot to enable leak check on a ref-counted object (we will try to remove it from the map when it is destroyed, but it will never have been added). - Can store extra logging information in the map values without adding to the size of the ref count struct itself. With the size of just an int64, the ref count object remains compact, meaning frequent operations like IncRef/DecRef are more cache-efficient. - Can aggregate leak results in a single report after the sandbox exits. Instead of having warnings littered in the log, which were non-deterministically triggered by garbage collection, we can print all warning messages at once. Note that this could also be a limitation--the sandbox must exit properly for leaks to be detected. Some basic benchmarking indicates that this change does not significantly affect performance when leak checking is enabled, which is understandable since registering/unregistering is only done once for each filesystem object. Updates #1486. PiperOrigin-RevId: 338685972
2020-10-23 16:14:52 +00:00
}
}