nuclei/pkg/core/executors.go

249 lines
7.6 KiB
Go
Raw Normal View History

package core
import (
"context"
"sync"
"sync/atomic"
"time"
Fuzzing layer enhancements + input-types support (#4477) * feat: move fuzz package to root directory * feat: added support for input providers like openapi,postman,etc * feat: integration of new fuzzing logic in engine * bugfix: use and instead of or * fixed lint errors * go mod tidy * add new reqresp type + bump utils * custom http request parser * use new struct type RequestResponse * introduce unified input/target provider * abstract input formats via new inputprovider * completed input provider refactor * remove duplicated code * add sdk method to load targets * rename component url->path * add new yaml format + remove duplicated code * use gopkg.in/yaml.v3 for parsing * update .gitignore * refactor/move + docs fuzzing in http protocol * fuzz: header + query integration test using fuzzplayground * fix integration test runner in windows * feat add support for filter in http fuzz * rewrite header/query integration test with filter * add replace regex rule * support kv fuzzing + misc updates * add path fuzzing example + misc improvements * fix matchedURL + skip httpx on multi formats * cookie fuzz integration test * add json body + params body tests * feat add multipart/form-data fuzzing support * add all fuzz body integration test * misc bug fixes + minor refactor * add multipart form + body form unit tests * only run fuzzing templates if -fuzz flag is given * refactor/move fuzz playground server to pkg * fix integration test + refactor * add auth types and strategies * add file auth provider * start implementing auth logic in http * add logic in http protocol * static auth implemented for http * default :80,:443 normalization * feat: dynamic auth init * feat: dynamic auth using templates * validate targets count in openapi+swagger * inputformats: add support to accept variables * fix workflow integration test * update lazy cred fetch logic * fix unit test * drop postman support * domain related normalization * update secrets.yaml file format + misc updates * add auth prefetch option * remove old secret files * add fuzzing+auth related sdk options * fix/support multiple mode in kv header fuzzing * rename 'headers' -> 'header' in fuzzing rules * fix deadlock due to merge conflict resolution * misc update * add bool type in parsed value * add openapi validation+override+ new flags * misc updates * remove optional path parameters when unavailable * fix swagger.yaml file * misc updates * update print msg * multiple openapi validation enchancements + appMode * add optional params in required_openapi_vars.yaml file * improve warning/verbose msgs in format * fix skip-format-validation not working * use 'params/parameter' instead of 'variable' in openapi * add retry support for falky tests * fix nuclei loading ignored templates (#4849) * fix tag include logic * fix unit test * remove quoting in extractor output * remove quote in debug code command * feat: issue tracker URLs in JSON + misc fixes (#4855) * feat: issue tracker URLs in JSON + misc fixes * misc changes * feat: status update support for issues * feat: report metadata generation hook support * feat: added CLI summary of tickets created * misc changes * introduce `disable-unsigned-templates` flag (#4820) * introduce `disable-unsigned-templates` flag * minor * skip instead of exit * remove duplicate imports * use stats package + misc enhancements * force display warning + adjust skipped stats in unsigned count * include unsigned skipped templates without -dut flag --------- Co-authored-by: Tarun Koyalwar <tarun@projectdiscovery.io> * Purge cache on global callback set (#4840) * purge cache on global callback set * lint * purging cache * purge cache in runner after loading templates * include internal cache from parsers + add global cache register/purge via config * remove disable cache purge option --------- Co-authored-by: Tarun Koyalwar <tarun@projectdiscovery.io> * misc update * add application/octet-stream support * openapi: support path specific params * misc option + readme update --------- Co-authored-by: Sandeep Singh <sandeep@projectdiscovery.io> Co-authored-by: sandeep <8293321+ehsandeep@users.noreply.github.com> Co-authored-by: Tarun Koyalwar <tarun@projectdiscovery.io> Co-authored-by: Tarun Koyalwar <45962551+tarunKoyalwar@users.noreply.github.com> Co-authored-by: Dogan Can Bakir <65292895+dogancanbakir@users.noreply.github.com> Co-authored-by: Mzack9999 <mzack9999@protonmail.com>
2024-03-14 03:08:53 +05:30
"github.com/projectdiscovery/nuclei/v3/pkg/input/provider"
"github.com/projectdiscovery/nuclei/v3/pkg/output"
"github.com/projectdiscovery/nuclei/v3/pkg/protocols/common/contextargs"
"github.com/projectdiscovery/nuclei/v3/pkg/scan"
"github.com/projectdiscovery/nuclei/v3/pkg/templates"
"github.com/projectdiscovery/nuclei/v3/pkg/templates/types"
generalTypes "github.com/projectdiscovery/nuclei/v3/pkg/types"
2024-04-03 17:50:57 +02:00
syncutil "github.com/projectdiscovery/utils/sync"
)
2023-03-05 20:07:40 +01:00
// Executors are low level executors that deals with template execution on a target
// executeAllSelfContained executes all self contained templates that do not use `target`
func (e *Engine) executeAllSelfContained(ctx context.Context, alltemplates []*templates.Template, results *atomic.Bool, sg *sync.WaitGroup) {
for _, v := range alltemplates {
sg.Add(1)
go func(template *templates.Template) {
defer sg.Done()
var err error
var match bool
ctx := scan.NewScanContext(ctx, contextargs.New(ctx))
if e.Callback != nil {
if results, err := template.Executer.ExecuteWithResults(ctx); err == nil {
for _, result := range results {
e.Callback(result)
}
}
match = true
} else {
match, err = template.Executer.Execute(ctx)
}
if err != nil {
Remove singletons from Nuclei engine (continuation of #6210) (#6296) * introducing execution id * wip * . * adding separate execution context id * lint * vet * fixing pg dialers * test ignore * fixing loader FD limit * test * fd fix * wip: remove CloseProcesses() from dev merge * wip: fix merge issue * protocolstate: stop memguarding on last dialer delete * avoid data race in dialers.RawHTTPClient * use shared logger and avoid race conditions * use shared logger and avoid race conditions * go mod * patch executionId into compiled template cache * clean up comment in Parse * go mod update * bump echarts * address merge issues * fix use of gologger * switch cmd/nuclei to options.Logger * address merge issues with go.mod * go vet: address copy of lock with new Copy function * fixing tests * disable speed control * fix nil ExecuterOptions * removing deprecated code * fixing result print * default logger * cli default logger * filter warning from results * fix performance test * hardcoding path * disable upload * refactor(runner): uses `Warning` instead of `Print` for `pdcpUploadErrMsg` Signed-off-by: Dwi Siswanto <git@dw1.io> * Revert "disable upload" This reverts commit 114fbe6663361bf41cf8b2645fd2d57083d53682. * Revert "hardcoding path" This reverts commit cf12ca800e0a0e974bd9fd4826a24e51547f7c00. --------- Signed-off-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Mzack9999 <mzack9999@protonmail.com> Co-authored-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Dwi Siswanto <25837540+dwisiswant0@users.noreply.github.com>
2025-07-09 14:47:26 -05:00
e.options.Logger.Warning().Msgf("[%s] Could not execute step (self-contained): %s\n", e.executerOpts.Colorizer.BrightBlue(template.ID), err)
}
results.CompareAndSwap(false, match)
}(v)
}
}
// executeTemplateWithTargets executes a given template on x targets (with a internal targetpool(i.e concurrency))
func (e *Engine) executeTemplateWithTargets(ctx context.Context, template *templates.Template, target provider.InputProvider, results *atomic.Bool) {
cache, goroutine and unbounded workers management (#6420) * Enhance matcher compilation with caching for regex and DSL expressions to improve performance. Update template parsing to conditionally retain raw templates based on size constraints. * Implement caching for regex and DSL expressions in extractors and matchers to enhance performance. Introduce a buffer pool in raw requests to reduce memory allocations. Update template cache management for improved efficiency. * feat: improve concurrency to be bound * refactor: replace fmt.Sprintf with fmt.Fprintf for improved performance in header handling * feat: add regex matching tests and benchmarks for performance evaluation * feat: add prefix check in regex extraction to optimize matching process * feat: implement regex caching mechanism to enhance performance in extractors and matchers, along with tests and benchmarks for validation * feat: add unit tests for template execution in the core engine, enhancing test coverage and reliability * feat: enhance error handling in template execution and improve regex caching logic for better performance * Implement caching for regex and DSL expressions in the cache package, replacing previous sync.Map usage. Add unit tests for cache functionality, including eviction by capacity and retrieval of cached items. Update extractors and matchers to utilize the new cache system for improved performance and memory efficiency. * Add tests for SetCapacities in cache package to ensure cache behavior on capacity changes - Implemented TestSetCapacities_NoRebuildOnZero to verify that setting capacities to zero does not clear existing caches. - Added TestSetCapacities_BeforeFirstUse to confirm that initial cache settings are respected and not overridden by subsequent capacity changes. * Refactor matchers and update load test generator to use io package - Removed maxRegexScanBytes constant from match.go. - Replaced ioutil with io package in load_test.go for NopCloser usage. - Restored TestValidate_AllowsInlineMultiline in load_test.go to ensure inline validation functionality. * Add cancellation support in template execution and enhance test coverage - Updated executeTemplateWithTargets to respect context cancellation. - Introduced fakeTargetProvider and slowExecuter for testing. - Added Test_executeTemplateWithTargets_RespectsCancellation to validate cancellation behavior during template execution.
2025-09-15 23:48:02 +05:30
if e.workPool == nil {
e.workPool = e.GetWorkPool()
}
// Bounded worker pool using input concurrency
pool := e.workPool.InputPool(template.Type())
workerCount := 1
if pool != nil && pool.Size > 0 {
workerCount = pool.Size
}
var (
index uint32
)
e.executerOpts.ResumeCfg.Lock()
currentInfo, ok := e.executerOpts.ResumeCfg.Current[template.ID]
if !ok {
currentInfo = &generalTypes.ResumeInfo{}
e.executerOpts.ResumeCfg.Current[template.ID] = currentInfo
}
if currentInfo.InFlight == nil {
currentInfo.InFlight = make(map[uint32]struct{})
}
resumeFromInfo, ok := e.executerOpts.ResumeCfg.ResumeFrom[template.ID]
if !ok {
resumeFromInfo = &generalTypes.ResumeInfo{}
e.executerOpts.ResumeCfg.ResumeFrom[template.ID] = resumeFromInfo
}
e.executerOpts.ResumeCfg.Unlock()
// track progression
cleanupInFlight := func(index uint32) {
currentInfo.Lock()
delete(currentInfo.InFlight, index)
currentInfo.Unlock()
}
cache, goroutine and unbounded workers management (#6420) * Enhance matcher compilation with caching for regex and DSL expressions to improve performance. Update template parsing to conditionally retain raw templates based on size constraints. * Implement caching for regex and DSL expressions in extractors and matchers to enhance performance. Introduce a buffer pool in raw requests to reduce memory allocations. Update template cache management for improved efficiency. * feat: improve concurrency to be bound * refactor: replace fmt.Sprintf with fmt.Fprintf for improved performance in header handling * feat: add regex matching tests and benchmarks for performance evaluation * feat: add prefix check in regex extraction to optimize matching process * feat: implement regex caching mechanism to enhance performance in extractors and matchers, along with tests and benchmarks for validation * feat: add unit tests for template execution in the core engine, enhancing test coverage and reliability * feat: enhance error handling in template execution and improve regex caching logic for better performance * Implement caching for regex and DSL expressions in the cache package, replacing previous sync.Map usage. Add unit tests for cache functionality, including eviction by capacity and retrieval of cached items. Update extractors and matchers to utilize the new cache system for improved performance and memory efficiency. * Add tests for SetCapacities in cache package to ensure cache behavior on capacity changes - Implemented TestSetCapacities_NoRebuildOnZero to verify that setting capacities to zero does not clear existing caches. - Added TestSetCapacities_BeforeFirstUse to confirm that initial cache settings are respected and not overridden by subsequent capacity changes. * Refactor matchers and update load test generator to use io package - Removed maxRegexScanBytes constant from match.go. - Replaced ioutil with io package in load_test.go for NopCloser usage. - Restored TestValidate_AllowsInlineMultiline in load_test.go to ensure inline validation functionality. * Add cancellation support in template execution and enhance test coverage - Updated executeTemplateWithTargets to respect context cancellation. - Introduced fakeTargetProvider and slowExecuter for testing. - Added Test_executeTemplateWithTargets_RespectsCancellation to validate cancellation behavior during template execution.
2025-09-15 23:48:02 +05:30
// task represents a single target execution unit
type task struct {
index uint32
skip bool
value *contextargs.MetaInput
}
tasks := make(chan task)
var workersWg sync.WaitGroup
workersWg.Add(workerCount)
for i := 0; i < workerCount; i++ {
go func() {
defer workersWg.Done()
for t := range tasks {
func() {
defer cleanupInFlight(t.index)
select {
case <-ctx.Done():
return
default:
}
if t.skip {
return
}
match, err := e.executeTemplateOnInput(ctx, template, t.value)
if err != nil {
e.options.Logger.Warning().Msgf("[%s] Could not execute step on %s: %s\n", e.executerOpts.Colorizer.BrightBlue(template.ID), t.value.Input, err)
}
results.CompareAndSwap(false, match)
}()
}
}()
}
Fuzzing layer enhancements + input-types support (#4477) * feat: move fuzz package to root directory * feat: added support for input providers like openapi,postman,etc * feat: integration of new fuzzing logic in engine * bugfix: use and instead of or * fixed lint errors * go mod tidy * add new reqresp type + bump utils * custom http request parser * use new struct type RequestResponse * introduce unified input/target provider * abstract input formats via new inputprovider * completed input provider refactor * remove duplicated code * add sdk method to load targets * rename component url->path * add new yaml format + remove duplicated code * use gopkg.in/yaml.v3 for parsing * update .gitignore * refactor/move + docs fuzzing in http protocol * fuzz: header + query integration test using fuzzplayground * fix integration test runner in windows * feat add support for filter in http fuzz * rewrite header/query integration test with filter * add replace regex rule * support kv fuzzing + misc updates * add path fuzzing example + misc improvements * fix matchedURL + skip httpx on multi formats * cookie fuzz integration test * add json body + params body tests * feat add multipart/form-data fuzzing support * add all fuzz body integration test * misc bug fixes + minor refactor * add multipart form + body form unit tests * only run fuzzing templates if -fuzz flag is given * refactor/move fuzz playground server to pkg * fix integration test + refactor * add auth types and strategies * add file auth provider * start implementing auth logic in http * add logic in http protocol * static auth implemented for http * default :80,:443 normalization * feat: dynamic auth init * feat: dynamic auth using templates * validate targets count in openapi+swagger * inputformats: add support to accept variables * fix workflow integration test * update lazy cred fetch logic * fix unit test * drop postman support * domain related normalization * update secrets.yaml file format + misc updates * add auth prefetch option * remove old secret files * add fuzzing+auth related sdk options * fix/support multiple mode in kv header fuzzing * rename 'headers' -> 'header' in fuzzing rules * fix deadlock due to merge conflict resolution * misc update * add bool type in parsed value * add openapi validation+override+ new flags * misc updates * remove optional path parameters when unavailable * fix swagger.yaml file * misc updates * update print msg * multiple openapi validation enchancements + appMode * add optional params in required_openapi_vars.yaml file * improve warning/verbose msgs in format * fix skip-format-validation not working * use 'params/parameter' instead of 'variable' in openapi * add retry support for falky tests * fix nuclei loading ignored templates (#4849) * fix tag include logic * fix unit test * remove quoting in extractor output * remove quote in debug code command * feat: issue tracker URLs in JSON + misc fixes (#4855) * feat: issue tracker URLs in JSON + misc fixes * misc changes * feat: status update support for issues * feat: report metadata generation hook support * feat: added CLI summary of tickets created * misc changes * introduce `disable-unsigned-templates` flag (#4820) * introduce `disable-unsigned-templates` flag * minor * skip instead of exit * remove duplicate imports * use stats package + misc enhancements * force display warning + adjust skipped stats in unsigned count * include unsigned skipped templates without -dut flag --------- Co-authored-by: Tarun Koyalwar <tarun@projectdiscovery.io> * Purge cache on global callback set (#4840) * purge cache on global callback set * lint * purging cache * purge cache in runner after loading templates * include internal cache from parsers + add global cache register/purge via config * remove disable cache purge option --------- Co-authored-by: Tarun Koyalwar <tarun@projectdiscovery.io> * misc update * add application/octet-stream support * openapi: support path specific params * misc option + readme update --------- Co-authored-by: Sandeep Singh <sandeep@projectdiscovery.io> Co-authored-by: sandeep <8293321+ehsandeep@users.noreply.github.com> Co-authored-by: Tarun Koyalwar <tarun@projectdiscovery.io> Co-authored-by: Tarun Koyalwar <45962551+tarunKoyalwar@users.noreply.github.com> Co-authored-by: Dogan Can Bakir <65292895+dogancanbakir@users.noreply.github.com> Co-authored-by: Mzack9999 <mzack9999@protonmail.com>
2024-03-14 03:08:53 +05:30
target.Iterate(func(scannedValue *contextargs.MetaInput) bool {
select {
case <-ctx.Done():
return false // exit
default:
}
// Best effort to track the host progression
// skips indexes lower than the minimum in-flight at interruption time
var skip bool
if resumeFromInfo.Completed { // the template was completed
Remove singletons from Nuclei engine (continuation of #6210) (#6296) * introducing execution id * wip * . * adding separate execution context id * lint * vet * fixing pg dialers * test ignore * fixing loader FD limit * test * fd fix * wip: remove CloseProcesses() from dev merge * wip: fix merge issue * protocolstate: stop memguarding on last dialer delete * avoid data race in dialers.RawHTTPClient * use shared logger and avoid race conditions * use shared logger and avoid race conditions * go mod * patch executionId into compiled template cache * clean up comment in Parse * go mod update * bump echarts * address merge issues * fix use of gologger * switch cmd/nuclei to options.Logger * address merge issues with go.mod * go vet: address copy of lock with new Copy function * fixing tests * disable speed control * fix nil ExecuterOptions * removing deprecated code * fixing result print * default logger * cli default logger * filter warning from results * fix performance test * hardcoding path * disable upload * refactor(runner): uses `Warning` instead of `Print` for `pdcpUploadErrMsg` Signed-off-by: Dwi Siswanto <git@dw1.io> * Revert "disable upload" This reverts commit 114fbe6663361bf41cf8b2645fd2d57083d53682. * Revert "hardcoding path" This reverts commit cf12ca800e0a0e974bd9fd4826a24e51547f7c00. --------- Signed-off-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Mzack9999 <mzack9999@protonmail.com> Co-authored-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Dwi Siswanto <25837540+dwisiswant0@users.noreply.github.com>
2025-07-09 14:47:26 -05:00
e.options.Logger.Debug().Msgf("[%s] Skipping \"%s\": Resume - Template already completed", template.ID, scannedValue.Input)
skip = true
} else if index < resumeFromInfo.SkipUnder { // index lower than the sliding window (bulk-size)
Remove singletons from Nuclei engine (continuation of #6210) (#6296) * introducing execution id * wip * . * adding separate execution context id * lint * vet * fixing pg dialers * test ignore * fixing loader FD limit * test * fd fix * wip: remove CloseProcesses() from dev merge * wip: fix merge issue * protocolstate: stop memguarding on last dialer delete * avoid data race in dialers.RawHTTPClient * use shared logger and avoid race conditions * use shared logger and avoid race conditions * go mod * patch executionId into compiled template cache * clean up comment in Parse * go mod update * bump echarts * address merge issues * fix use of gologger * switch cmd/nuclei to options.Logger * address merge issues with go.mod * go vet: address copy of lock with new Copy function * fixing tests * disable speed control * fix nil ExecuterOptions * removing deprecated code * fixing result print * default logger * cli default logger * filter warning from results * fix performance test * hardcoding path * disable upload * refactor(runner): uses `Warning` instead of `Print` for `pdcpUploadErrMsg` Signed-off-by: Dwi Siswanto <git@dw1.io> * Revert "disable upload" This reverts commit 114fbe6663361bf41cf8b2645fd2d57083d53682. * Revert "hardcoding path" This reverts commit cf12ca800e0a0e974bd9fd4826a24e51547f7c00. --------- Signed-off-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Mzack9999 <mzack9999@protonmail.com> Co-authored-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Dwi Siswanto <25837540+dwisiswant0@users.noreply.github.com>
2025-07-09 14:47:26 -05:00
e.options.Logger.Debug().Msgf("[%s] Skipping \"%s\": Resume - Target already processed", template.ID, scannedValue.Input)
skip = true
} else if _, isInFlight := resumeFromInfo.InFlight[index]; isInFlight { // the target wasn't completed successfully
Remove singletons from Nuclei engine (continuation of #6210) (#6296) * introducing execution id * wip * . * adding separate execution context id * lint * vet * fixing pg dialers * test ignore * fixing loader FD limit * test * fd fix * wip: remove CloseProcesses() from dev merge * wip: fix merge issue * protocolstate: stop memguarding on last dialer delete * avoid data race in dialers.RawHTTPClient * use shared logger and avoid race conditions * use shared logger and avoid race conditions * go mod * patch executionId into compiled template cache * clean up comment in Parse * go mod update * bump echarts * address merge issues * fix use of gologger * switch cmd/nuclei to options.Logger * address merge issues with go.mod * go vet: address copy of lock with new Copy function * fixing tests * disable speed control * fix nil ExecuterOptions * removing deprecated code * fixing result print * default logger * cli default logger * filter warning from results * fix performance test * hardcoding path * disable upload * refactor(runner): uses `Warning` instead of `Print` for `pdcpUploadErrMsg` Signed-off-by: Dwi Siswanto <git@dw1.io> * Revert "disable upload" This reverts commit 114fbe6663361bf41cf8b2645fd2d57083d53682. * Revert "hardcoding path" This reverts commit cf12ca800e0a0e974bd9fd4826a24e51547f7c00. --------- Signed-off-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Mzack9999 <mzack9999@protonmail.com> Co-authored-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Dwi Siswanto <25837540+dwisiswant0@users.noreply.github.com>
2025-07-09 14:47:26 -05:00
e.options.Logger.Debug().Msgf("[%s] Repeating \"%s\": Resume - Target wasn't completed", template.ID, scannedValue.Input)
// skip is already false, but leaving it here for clarity
skip = false
} else if index > resumeFromInfo.DoAbove { // index above the sliding window (bulk-size)
// skip is already false - but leaving it here for clarity
skip = false
}
currentInfo.Lock()
currentInfo.InFlight[index] = struct{}{}
currentInfo.Unlock()
// Skip if the host has had errors
if e.executerOpts.HostErrorsCache != nil && e.executerOpts.HostErrorsCache.Check(e.executerOpts.ProtocolType.String(), contextargs.NewWithMetaInput(ctx, scannedValue)) {
skipEvent := &output.ResultEvent{
TemplateID: template.ID,
TemplatePath: template.Path,
Info: template.Info,
Type: e.executerOpts.ProtocolType.String(),
Host: scannedValue.Input,
MatcherStatus: false,
Error: "host was skipped as it was found unresponsive",
Timestamp: time.Now(),
}
if e.Callback != nil {
e.Callback(skipEvent)
} else if e.executerOpts.Output != nil {
_ = e.executerOpts.Output.Write(skipEvent)
}
return true
}
cache, goroutine and unbounded workers management (#6420) * Enhance matcher compilation with caching for regex and DSL expressions to improve performance. Update template parsing to conditionally retain raw templates based on size constraints. * Implement caching for regex and DSL expressions in extractors and matchers to enhance performance. Introduce a buffer pool in raw requests to reduce memory allocations. Update template cache management for improved efficiency. * feat: improve concurrency to be bound * refactor: replace fmt.Sprintf with fmt.Fprintf for improved performance in header handling * feat: add regex matching tests and benchmarks for performance evaluation * feat: add prefix check in regex extraction to optimize matching process * feat: implement regex caching mechanism to enhance performance in extractors and matchers, along with tests and benchmarks for validation * feat: add unit tests for template execution in the core engine, enhancing test coverage and reliability * feat: enhance error handling in template execution and improve regex caching logic for better performance * Implement caching for regex and DSL expressions in the cache package, replacing previous sync.Map usage. Add unit tests for cache functionality, including eviction by capacity and retrieval of cached items. Update extractors and matchers to utilize the new cache system for improved performance and memory efficiency. * Add tests for SetCapacities in cache package to ensure cache behavior on capacity changes - Implemented TestSetCapacities_NoRebuildOnZero to verify that setting capacities to zero does not clear existing caches. - Added TestSetCapacities_BeforeFirstUse to confirm that initial cache settings are respected and not overridden by subsequent capacity changes. * Refactor matchers and update load test generator to use io package - Removed maxRegexScanBytes constant from match.go. - Replaced ioutil with io package in load_test.go for NopCloser usage. - Restored TestValidate_AllowsInlineMultiline in load_test.go to ensure inline validation functionality. * Add cancellation support in template execution and enhance test coverage - Updated executeTemplateWithTargets to respect context cancellation. - Introduced fakeTargetProvider and slowExecuter for testing. - Added Test_executeTemplateWithTargets_RespectsCancellation to validate cancellation behavior during template execution.
2025-09-15 23:48:02 +05:30
tasks <- task{index: index, skip: skip, value: scannedValue}
index++
return true
})
cache, goroutine and unbounded workers management (#6420) * Enhance matcher compilation with caching for regex and DSL expressions to improve performance. Update template parsing to conditionally retain raw templates based on size constraints. * Implement caching for regex and DSL expressions in extractors and matchers to enhance performance. Introduce a buffer pool in raw requests to reduce memory allocations. Update template cache management for improved efficiency. * feat: improve concurrency to be bound * refactor: replace fmt.Sprintf with fmt.Fprintf for improved performance in header handling * feat: add regex matching tests and benchmarks for performance evaluation * feat: add prefix check in regex extraction to optimize matching process * feat: implement regex caching mechanism to enhance performance in extractors and matchers, along with tests and benchmarks for validation * feat: add unit tests for template execution in the core engine, enhancing test coverage and reliability * feat: enhance error handling in template execution and improve regex caching logic for better performance * Implement caching for regex and DSL expressions in the cache package, replacing previous sync.Map usage. Add unit tests for cache functionality, including eviction by capacity and retrieval of cached items. Update extractors and matchers to utilize the new cache system for improved performance and memory efficiency. * Add tests for SetCapacities in cache package to ensure cache behavior on capacity changes - Implemented TestSetCapacities_NoRebuildOnZero to verify that setting capacities to zero does not clear existing caches. - Added TestSetCapacities_BeforeFirstUse to confirm that initial cache settings are respected and not overridden by subsequent capacity changes. * Refactor matchers and update load test generator to use io package - Removed maxRegexScanBytes constant from match.go. - Replaced ioutil with io package in load_test.go for NopCloser usage. - Restored TestValidate_AllowsInlineMultiline in load_test.go to ensure inline validation functionality. * Add cancellation support in template execution and enhance test coverage - Updated executeTemplateWithTargets to respect context cancellation. - Introduced fakeTargetProvider and slowExecuter for testing. - Added Test_executeTemplateWithTargets_RespectsCancellation to validate cancellation behavior during template execution.
2025-09-15 23:48:02 +05:30
close(tasks)
workersWg.Wait()
// on completion marks the template as completed
currentInfo.Lock()
currentInfo.Completed = true
currentInfo.Unlock()
}
// executeTemplatesOnTarget execute given templates on given single target
func (e *Engine) executeTemplatesOnTarget(ctx context.Context, alltemplates []*templates.Template, target *contextargs.MetaInput, results *atomic.Bool) {
// all templates are executed on single target
// wp is workpool that contains different waitgroups for
// headless and non-headless templates
// global waitgroup should not be used here
wp := e.GetWorkPool()
defer wp.Wait()
for _, tpl := range alltemplates {
select {
case <-ctx.Done():
return
default:
}
2024-04-03 18:50:46 +02:00
// resize check point - nop if there are no changes
wp.RefreshWithConfig(e.GetWorkPoolConfig())
2024-04-03 17:50:57 +02:00
var sg *syncutil.AdaptiveWaitGroup
if tpl.Type() == types.HeadlessProtocol {
sg = wp.Headless
} else {
sg = wp.Default
}
sg.Add()
2024-04-03 17:50:57 +02:00
go func(template *templates.Template, value *contextargs.MetaInput, wg *syncutil.AdaptiveWaitGroup) {
defer wg.Done()
cache, goroutine and unbounded workers management (#6420) * Enhance matcher compilation with caching for regex and DSL expressions to improve performance. Update template parsing to conditionally retain raw templates based on size constraints. * Implement caching for regex and DSL expressions in extractors and matchers to enhance performance. Introduce a buffer pool in raw requests to reduce memory allocations. Update template cache management for improved efficiency. * feat: improve concurrency to be bound * refactor: replace fmt.Sprintf with fmt.Fprintf for improved performance in header handling * feat: add regex matching tests and benchmarks for performance evaluation * feat: add prefix check in regex extraction to optimize matching process * feat: implement regex caching mechanism to enhance performance in extractors and matchers, along with tests and benchmarks for validation * feat: add unit tests for template execution in the core engine, enhancing test coverage and reliability * feat: enhance error handling in template execution and improve regex caching logic for better performance * Implement caching for regex and DSL expressions in the cache package, replacing previous sync.Map usage. Add unit tests for cache functionality, including eviction by capacity and retrieval of cached items. Update extractors and matchers to utilize the new cache system for improved performance and memory efficiency. * Add tests for SetCapacities in cache package to ensure cache behavior on capacity changes - Implemented TestSetCapacities_NoRebuildOnZero to verify that setting capacities to zero does not clear existing caches. - Added TestSetCapacities_BeforeFirstUse to confirm that initial cache settings are respected and not overridden by subsequent capacity changes. * Refactor matchers and update load test generator to use io package - Removed maxRegexScanBytes constant from match.go. - Replaced ioutil with io package in load_test.go for NopCloser usage. - Restored TestValidate_AllowsInlineMultiline in load_test.go to ensure inline validation functionality. * Add cancellation support in template execution and enhance test coverage - Updated executeTemplateWithTargets to respect context cancellation. - Introduced fakeTargetProvider and slowExecuter for testing. - Added Test_executeTemplateWithTargets_RespectsCancellation to validate cancellation behavior during template execution.
2025-09-15 23:48:02 +05:30
match, err := e.executeTemplateOnInput(ctx, template, value)
if err != nil {
Remove singletons from Nuclei engine (continuation of #6210) (#6296) * introducing execution id * wip * . * adding separate execution context id * lint * vet * fixing pg dialers * test ignore * fixing loader FD limit * test * fd fix * wip: remove CloseProcesses() from dev merge * wip: fix merge issue * protocolstate: stop memguarding on last dialer delete * avoid data race in dialers.RawHTTPClient * use shared logger and avoid race conditions * use shared logger and avoid race conditions * go mod * patch executionId into compiled template cache * clean up comment in Parse * go mod update * bump echarts * address merge issues * fix use of gologger * switch cmd/nuclei to options.Logger * address merge issues with go.mod * go vet: address copy of lock with new Copy function * fixing tests * disable speed control * fix nil ExecuterOptions * removing deprecated code * fixing result print * default logger * cli default logger * filter warning from results * fix performance test * hardcoding path * disable upload * refactor(runner): uses `Warning` instead of `Print` for `pdcpUploadErrMsg` Signed-off-by: Dwi Siswanto <git@dw1.io> * Revert "disable upload" This reverts commit 114fbe6663361bf41cf8b2645fd2d57083d53682. * Revert "hardcoding path" This reverts commit cf12ca800e0a0e974bd9fd4826a24e51547f7c00. --------- Signed-off-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Mzack9999 <mzack9999@protonmail.com> Co-authored-by: Dwi Siswanto <git@dw1.io> Co-authored-by: Dwi Siswanto <25837540+dwisiswant0@users.noreply.github.com>
2025-07-09 14:47:26 -05:00
e.options.Logger.Warning().Msgf("[%s] Could not execute step on %s: %s\n", e.executerOpts.Colorizer.BrightBlue(template.ID), value.Input, err)
}
results.CompareAndSwap(false, match)
}(tpl, target, sg)
}
}
cache, goroutine and unbounded workers management (#6420) * Enhance matcher compilation with caching for regex and DSL expressions to improve performance. Update template parsing to conditionally retain raw templates based on size constraints. * Implement caching for regex and DSL expressions in extractors and matchers to enhance performance. Introduce a buffer pool in raw requests to reduce memory allocations. Update template cache management for improved efficiency. * feat: improve concurrency to be bound * refactor: replace fmt.Sprintf with fmt.Fprintf for improved performance in header handling * feat: add regex matching tests and benchmarks for performance evaluation * feat: add prefix check in regex extraction to optimize matching process * feat: implement regex caching mechanism to enhance performance in extractors and matchers, along with tests and benchmarks for validation * feat: add unit tests for template execution in the core engine, enhancing test coverage and reliability * feat: enhance error handling in template execution and improve regex caching logic for better performance * Implement caching for regex and DSL expressions in the cache package, replacing previous sync.Map usage. Add unit tests for cache functionality, including eviction by capacity and retrieval of cached items. Update extractors and matchers to utilize the new cache system for improved performance and memory efficiency. * Add tests for SetCapacities in cache package to ensure cache behavior on capacity changes - Implemented TestSetCapacities_NoRebuildOnZero to verify that setting capacities to zero does not clear existing caches. - Added TestSetCapacities_BeforeFirstUse to confirm that initial cache settings are respected and not overridden by subsequent capacity changes. * Refactor matchers and update load test generator to use io package - Removed maxRegexScanBytes constant from match.go. - Replaced ioutil with io package in load_test.go for NopCloser usage. - Restored TestValidate_AllowsInlineMultiline in load_test.go to ensure inline validation functionality. * Add cancellation support in template execution and enhance test coverage - Updated executeTemplateWithTargets to respect context cancellation. - Introduced fakeTargetProvider and slowExecuter for testing. - Added Test_executeTemplateWithTargets_RespectsCancellation to validate cancellation behavior during template execution.
2025-09-15 23:48:02 +05:30
// executeTemplateOnInput performs template execution for a single input and returns match status and error
func (e *Engine) executeTemplateOnInput(ctx context.Context, template *templates.Template, value *contextargs.MetaInput) (bool, error) {
ctxArgs := contextargs.New(ctx)
ctxArgs.MetaInput = value
scanCtx := scan.NewScanContext(ctx, ctxArgs)
switch template.Type() {
case types.WorkflowProtocol:
return e.executeWorkflow(scanCtx, template.CompiledWorkflow), nil
default:
if e.Callback != nil {
results, err := template.Executer.ExecuteWithResults(scanCtx)
if err != nil {
return false, err
}
for _, result := range results {
e.Callback(result)
}
return len(results) > 0, nil
}
return template.Executer.Execute(scanCtx)
}
}