community
/
dns
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
dns/msg_generate.go

352 lines
10 KiB
Go
Raw Normal View History

Run gofmt -w -s *.go (#1408) Signed-off-by: Miek Gieben <miek@miek.nl> Signed-off-by: Miek Gieben <miek@miek.nl>
2023-01-05 06:53:57 +00:00
//go:build ignore
// +build ignore
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
// msg_generate.go is meant to run with go generate. It will use
// go/{importer,types} to track down all the RR struct types. Then for each type
// it will generate pack/unpack methods based on the struct tags. The generated source is
// written to zmsg.go, and is meant to be checked into git.
package main
import (
"bytes"
"fmt"
"go/format"
"go/types"
"log"
"os"
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
"strings"
code generation compatible with go modules (#1050) * code generation compatible with go modules * build: ensure go generate makes no changes
2019-12-30 11:25:57 +00:00
"golang.org/x/tools/go/packages"
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
)
var packageHdr = `
Lint: use ignore-this on generated files (#596) Make golint not complain about generated file - it's a bit of cheating, but I think that is OK.
2017-12-06 10:02:47 +00:00
// Code generated by "go run msg_generate.go"; DO NOT EDIT.
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
package dns
`
// getTypeStruct will take a type and the package scope, and return the
// (innermost) struct if the type is considered a RR type (currently defined as
// those structs beginning with a RR_Header, could be redefined as implementing
// the RR interface). The bool return value indicates if embedded structs were
// resolved.
func getTypeStruct(t types.Type, scope *types.Scope) (*types.Struct, bool) {
st, ok := t.Underlying().(*types.Struct)
if !ok {
return nil, false
}
Implement SVCB (#1067) * Implement SVCB * Fix serialization and deserialization of double quotes * More effort (?) 4 months old commit * DEBUG * _ * Presentation format serialization/deserialization * _ Remove generated * Progress on presentation format parse & write * _ * Finish parsing presentation format * Regenerate * Pack unpack * Move to svcb.go Scan_rr.go and types.go should be untouched now * :bug: Thanks ghedo * Definitions * TypeHTTPSSVC * Generated and isDuplicate * Goodbye lenient functions Now private key=value pairs have to be defined as structs too. They are no longer automatically named as KeyNNNNN * Encode/decode * Experimental svc * Read method * Implement some of the methods, use trick... to report where the error is while reading it. This should be applied to EDNS too. Todo: Find if case can only contain e := new(SVC_ALPN) and rest moved out Also fix two compile errors * Add SVC_LOCAL methods, reorder, remove alpn value, bugs * Errors * Alpn, make it build * Correct testsuite * Fully implement parser Change from keeping a state variable to reading in one iteration until the key=value pair is fully consumed * Simplify and document EDNS should be simplified too * Attempt to fix fuzzer And Alpn bug * A bug and change type values to match @ghedo's implementation * IP bug Also there are two ip duplicating patterns, one with copy, one with append. Maybe change it to be consistent. * Check for strictly increasing keys as required * Don't panic on invalid alpn * Redundant check, don't modify original array * Size calculation * Fix the fuzzer, match the style * 65535 is reserved too, don't delay errors * Check keyNNN, check for aliasform having values * IPvNHint is an array * Fix ipvNHint * Rename everything * Unrecognized keys according to the updated specification * Skip zero-length structs in generators. Fix CI * Doc cleanup * Off by one * Add parse tests * Check if private key doesn't collide with known key, invalid tests * Disallow IPv4 as IPv6. More tests. Related #1107 * Style fixes * More consistency, more tests * :bug: Deep copy as in the documentation a := make([]net.IP, 1) a[0] = net.ParseIP("1.1.1.1").To4() b := append(make([]net.IP, 0, 1), a...) b[0] = net.ParseIP("3.1.1.1").To4() fmt.Println(a[0][0]) * Make tests readable * Move valid parse tests to different file * :bug: One of previous commits not fully committed * Test binary single value encoding/decoding and full encode/decode * Add worst-case grows to builders, :bug: Wrong visible character range, redundant tests * Testing improvements And don't convert to IPv4 twice * Doc update only * Document worst case allocations and ipv6 can be at most of length 39, not 40 * Redundant IP copy, consistent IPv6 behavior, fix deep copy * isDuplicate for SVCB * Optimizations * echoconfig * Svc => SVCB * Fix CI * Regenerate after REBASE (2) Rebased twice on 15th and 20th May * Rename svc, use escapeByte. * Fix parsing whitespaces between quotes, rename ECHOHOConfig * resolve Remove svcbFieldLen Use reverseInt Uppercase SVCB Rename key_value "invalid" => bad Alpn comments > 65535 check Unneeded slices * a little more read => parse IP array meaning Force pushed because forgot to change read in svcb_test.go * HTTPSSVC -> HTTPS * Use new values * mandatory code https://github.com/MikeBishop/dns-alt-svc/pull/205 * Resolve comments Rename svcb-pairs Remove SVCB_PRIVATE ranges Comment on SVCB_KEY65535 ParseError return l.token rename svcbKeyToString and svcbStringToKey privatize SVCBKeyToString, SVCBStringToKey * Refactor 1 Rename sorted, originalPairs Use append instead of copy Use svcb_RESERVED instead of 65535, with it now being private "type SVCBKey uint16" * Refactor 2 svcbKeyToString as method svcbStringToKey updated after key 0 :bug: mandatory has missing key Rename str idx < 0 * Refactor 3 Use l.token as z var key, value string Comment wrap 0: Sentences with '.' keyValue => kv * Refactor 4 * Refactor 5 len() int * Refactor 6 * Refactor 7 * Test remove parsing * Error messages * Rewrite two estimate comments * parse shouldn't modify original array :bug: * Remove two unneeded comments * Address review comments Push 2 because can't build fuzzer python Push 3 to try again * Simplify argument duplication as per tmthrgd's suggestion And add the relevant test Force push edit: Make sorting code fit into one line * Rewrite ECHConfig and address the review * Remove the optional tab * Add To4() Check * More cleanup and fix mandatory not sorting bug
2020-10-11 07:09:36 +00:00
if st.NumFields() == 0 {
return nil, false
}
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
if st.Field(0).Type() == scope.Lookup("RR_Header").Type() {
return st, false
}
if st.Field(0).Anonymous() {
st, _ := getTypeStruct(st.Field(0).Type(), scope)
return st, true
}
return nil, false
}
code generation compatible with go modules (#1050) * code generation compatible with go modules * build: ensure go generate makes no changes
2019-12-30 11:25:57 +00:00
// loadModule retrieves package description for a given module.
func loadModule(name string) (*types.Package, error) {
conf := packages.Config{Mode: packages.NeedTypes | packages.NeedTypesInfo}
pkgs, err := packages.Load(&conf, name)
if err != nil {
return nil, err
}
return pkgs[0].Types, nil
}
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
func main() {
// Import and type-check the package
code generation compatible with go modules (#1050) * code generation compatible with go modules * build: ensure go generate makes no changes
2019-12-30 11:25:57 +00:00
pkg, err := loadModule("github.com/miekg/dns")
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
fatalIfErr(err)
scope := pkg.Scope()
// Collect actual types (*X)
var namedTypes []string
for _, name := range scope.Names() {
o := scope.Lookup(name)
if o == nil || !o.Exported() {
continue
}
if st, _ := getTypeStruct(o.Type(), scope); st == nil {
continue
}
Remove reflection (#376) Everything is generated. Remove all uses of packStruct/unpackStruct and make the library reflectionless.
2016-06-12 20:06:46 +00:00
if name == "PrivateRR" {
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
continue
}
// Check if corresponding TypeX exists
if scope.Lookup("Type"+o.Name()) == nil && o.Name() != "RFC3597" {
log.Fatalf("Constant Type%s does not exist.", o.Name())
}
namedTypes = append(namedTypes, o.Name())
}
b := &bytes.Buffer{}
b.WriteString(packageHdr)
fmt.Fprint(b, "// pack*() functions\n\n")
for _, name := range namedTypes {
o := scope.Lookup(name)
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
st, _ := getTypeStruct(o.Type(), scope)
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
Move RR header packing out of generated code (#885)
2019-01-03 23:39:14 +00:00
fmt.Fprintf(b, "func (rr *%s) pack(msg []byte, off int, compression compressionMap, compress bool) (off1 int, err error) {\n", name)
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
for i := 1; i < st.NumFields(); i++ {
o := func(s string) {
fmt.Fprintf(b, s, st.Field(i).Name())
fmt.Fprint(b, `if err != nil {
Move RR header packing out of generated code (#885)
2019-01-03 23:39:14 +00:00
return off, err
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
}
`)
}
Generate pack/unpack for all RRs (#360) Add dns:txt parsing helper to prevent compile errors. This allows us to generate all unpack/pack function. Add pack to the RR interface definition and add this method to PrivateRR. We still use typeToUnpack to select which types don't use reflection.
2016-06-05 06:53:12 +00:00
if _, ok := st.Field(i).Type().(*types.Slice); ok {
switch st.Tag(i) {
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
case `dns:"-"`: // ignored
Generate pack/unpack for all RRs (#360) Add dns:txt parsing helper to prevent compile errors. This allows us to generate all unpack/pack function. Add pack to the RR interface definition and add this method to PrivateRR. We still use typeToUnpack to select which types don't use reflection.
2016-06-05 06:53:12 +00:00
case `dns:"txt"`:
o("off, err = packStringTxt(rr.%s, msg, off)\n")
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
case `dns:"opt"`:
o("off, err = packDataOpt(rr.%s, msg, off)\n")
case `dns:"nsec"`:
o("off, err = packDataNsec(rr.%s, msg, off)\n")
Implement SVCB (#1067) * Implement SVCB * Fix serialization and deserialization of double quotes * More effort (?) 4 months old commit * DEBUG * _ * Presentation format serialization/deserialization * _ Remove generated * Progress on presentation format parse & write * _ * Finish parsing presentation format * Regenerate * Pack unpack * Move to svcb.go Scan_rr.go and types.go should be untouched now * :bug: Thanks ghedo * Definitions * TypeHTTPSSVC * Generated and isDuplicate * Goodbye lenient functions Now private key=value pairs have to be defined as structs too. They are no longer automatically named as KeyNNNNN * Encode/decode * Experimental svc * Read method * Implement some of the methods, use trick... to report where the error is while reading it. This should be applied to EDNS too. Todo: Find if case can only contain e := new(SVC_ALPN) and rest moved out Also fix two compile errors * Add SVC_LOCAL methods, reorder, remove alpn value, bugs * Errors * Alpn, make it build * Correct testsuite * Fully implement parser Change from keeping a state variable to reading in one iteration until the key=value pair is fully consumed * Simplify and document EDNS should be simplified too * Attempt to fix fuzzer And Alpn bug * A bug and change type values to match @ghedo's implementation * IP bug Also there are two ip duplicating patterns, one with copy, one with append. Maybe change it to be consistent. * Check for strictly increasing keys as required * Don't panic on invalid alpn * Redundant check, don't modify original array * Size calculation * Fix the fuzzer, match the style * 65535 is reserved too, don't delay errors * Check keyNNN, check for aliasform having values * IPvNHint is an array * Fix ipvNHint * Rename everything * Unrecognized keys according to the updated specification * Skip zero-length structs in generators. Fix CI * Doc cleanup * Off by one * Add parse tests * Check if private key doesn't collide with known key, invalid tests * Disallow IPv4 as IPv6. More tests. Related #1107 * Style fixes * More consistency, more tests * :bug: Deep copy as in the documentation a := make([]net.IP, 1) a[0] = net.ParseIP("1.1.1.1").To4() b := append(make([]net.IP, 0, 1), a...) b[0] = net.ParseIP("3.1.1.1").To4() fmt.Println(a[0][0]) * Make tests readable * Move valid parse tests to different file * :bug: One of previous commits not fully committed * Test binary single value encoding/decoding and full encode/decode * Add worst-case grows to builders, :bug: Wrong visible character range, redundant tests * Testing improvements And don't convert to IPv4 twice * Doc update only * Document worst case allocations and ipv6 can be at most of length 39, not 40 * Redundant IP copy, consistent IPv6 behavior, fix deep copy * isDuplicate for SVCB * Optimizations * echoconfig * Svc => SVCB * Fix CI * Regenerate after REBASE (2) Rebased twice on 15th and 20th May * Rename svc, use escapeByte. * Fix parsing whitespaces between quotes, rename ECHOHOConfig * resolve Remove svcbFieldLen Use reverseInt Uppercase SVCB Rename key_value "invalid" => bad Alpn comments > 65535 check Unneeded slices * a little more read => parse IP array meaning Force pushed because forgot to change read in svcb_test.go * HTTPSSVC -> HTTPS * Use new values * mandatory code https://github.com/MikeBishop/dns-alt-svc/pull/205 * Resolve comments Rename svcb-pairs Remove SVCB_PRIVATE ranges Comment on SVCB_KEY65535 ParseError return l.token rename svcbKeyToString and svcbStringToKey privatize SVCBKeyToString, SVCBStringToKey * Refactor 1 Rename sorted, originalPairs Use append instead of copy Use svcb_RESERVED instead of 65535, with it now being private "type SVCBKey uint16" * Refactor 2 svcbKeyToString as method svcbStringToKey updated after key 0 :bug: mandatory has missing key Rename str idx < 0 * Refactor 3 Use l.token as z var key, value string Comment wrap 0: Sentences with '.' keyValue => kv * Refactor 4 * Refactor 5 len() int * Refactor 6 * Refactor 7 * Test remove parsing * Error messages * Rewrite two estimate comments * parse shouldn't modify original array :bug: * Remove two unneeded comments * Address review comments Push 2 because can't build fuzzer python Push 3 to try again * Simplify argument duplication as per tmthrgd's suggestion And add the relevant test Force push edit: Make sorting code fit into one line * Rewrite ECHConfig and address the review * Remove the optional tab * Add To4() Check * More cleanup and fix mandatory not sorting bug
2020-10-11 07:09:36 +00:00
case `dns:"pairs"`:
o("off, err = packDataSVCB(rr.%s, msg, off)\n")
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
case `dns:"domain-name"`:
Move compress=false out of packDataDomainNames into caller (#849) This makes packDataDomainNames more consistent with PackDomainName where it will only be called with compress = true for `dns:"cdomain-name"`.
2018-11-30 07:49:06 +00:00
o("off, err = packDataDomainNames(rr.%s, msg, off, compression, false)\n")
APL record support (#1058) * APL record: add structure and code point * APL record: add wire format support * APL record: add presentation format support * APL record: add isDuplicate implementation * APL record: add copy implementation * APL record: add len implementation * APL record: run go generate * APL record: fix condition checking for equality * APL record: use switches to map family to address length * APL record: check bounds of individual fields rather than whole header * APL record: stylistic changes * APL record: remove APLPrefix methods from public interface * APL record: update README * APL record: additional cleanup for code review * APL record: change return type from pointer to struct * APL record: refactor of pack and unpack to eliminate extra variables
2020-01-03 12:41:45 +00:00
case `dns:"apl"`:
o("off, err = packDataApl(rr.%s, msg, off)\n")
Generate pack/unpack for all RRs (#360) Add dns:txt parsing helper to prevent compile errors. This allows us to generate all unpack/pack function. Add pack to the RR interface definition and add this method to PrivateRR. We still use typeToUnpack to select which types don't use reflection.
2016-06-05 06:53:12 +00:00
default:
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
log.Fatalln(name, st.Field(i).Name(), st.Tag(i))
Generate pack/unpack for all RRs (#360) Add dns:txt parsing helper to prevent compile errors. This allows us to generate all unpack/pack function. Add pack to the RR interface definition and add this method to PrivateRR. We still use typeToUnpack to select which types don't use reflection.
2016-06-05 06:53:12 +00:00
}
continue
}
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
switch {
case st.Tag(i) == `dns:"-"`: // ignored
case st.Tag(i) == `dns:"cdomain-name"`:
Stop using packDomainName in IsDomainName (#873) * Fork packDomainName for IsDomainName * Eliminate msg buffer from packDomainName2 * Eliminate compression code from packDomainName2 * Remove off argument and return from packDomainName2 * Remove bs buffer from packDomainName2 * Merge packDomainName2 into IsDomainName * Eliminate root label special case from IsDomainName * Remove ls variable from IsDomainName * Fixup comments in IsDomainName * Remove msg == nil special cases from packDomainName * Eliminate lenmsg variable from packDomainName * Eliminate label counting from packDomainName * Change off length check in IsDomainName * Fix IsDomainName for escaped names * Use strings.HasSuffix for IsFqdn * Revert "Use strings.HasSuffix for IsFqdn" I'll submit this as a seperate PR. This reverts commit 80bf8c83700d121ea45edac0f00db52817498166. * Cross reference IsDomainName and packDomainName * Correct IsDomainName max length comment
2019-01-03 16:39:37 +00:00
o("off, err = packDomainName(rr.%s, msg, off, compression, compress)\n")
Fix packing compression according to struct tagging (#433)
2016-12-27 20:21:54 +00:00
case st.Tag(i) == `dns:"domain-name"`:
Stop using packDomainName in IsDomainName (#873) * Fork packDomainName for IsDomainName * Eliminate msg buffer from packDomainName2 * Eliminate compression code from packDomainName2 * Remove off argument and return from packDomainName2 * Remove bs buffer from packDomainName2 * Merge packDomainName2 into IsDomainName * Eliminate root label special case from IsDomainName * Remove ls variable from IsDomainName * Fixup comments in IsDomainName * Remove msg == nil special cases from packDomainName * Eliminate lenmsg variable from packDomainName * Eliminate label counting from packDomainName * Change off length check in IsDomainName * Fix IsDomainName for escaped names * Use strings.HasSuffix for IsFqdn * Revert "Use strings.HasSuffix for IsFqdn" I'll submit this as a seperate PR. This reverts commit 80bf8c83700d121ea45edac0f00db52817498166. * Cross reference IsDomainName and packDomainName * Correct IsDomainName max length comment
2019-01-03 16:39:37 +00:00
o("off, err = packDomainName(rr.%s, msg, off, compression, false)\n")
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
case st.Tag(i) == `dns:"a"`:
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
o("off, err = packDataA(rr.%s, msg, off)\n")
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
case st.Tag(i) == `dns:"aaaa"`:
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
o("off, err = packDataAAAA(rr.%s, msg, off)\n")
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
case st.Tag(i) == `dns:"uint48"`:
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
o("off, err = packUint48(rr.%s, msg, off)\n")
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
case st.Tag(i) == `dns:"txt"`:
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
o("off, err = packString(rr.%s, msg, off)\n")
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
case strings.HasPrefix(st.Tag(i), `dns:"size-base32`): // size-base32 can be packed just like base32
fallthrough
case st.Tag(i) == `dns:"base32"`:
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
o("off, err = packStringBase32(rr.%s, msg, off)\n")
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
case strings.HasPrefix(st.Tag(i), `dns:"size-base64`): // size-base64 can be packed just like base64
fallthrough
case st.Tag(i) == `dns:"base64"`:
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
o("off, err = packStringBase64(rr.%s, msg, off)\n")
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
Only pack NSEC3/NSEC3PARAM salt when it is not empty (#474) * Only pack when salt is not empty * Actually generate things properly * Use string literal
2017-03-21 19:35:21 +00:00
case strings.HasPrefix(st.Tag(i), `dns:"size-hex:SaltLength`):
// directly write instead of using o() so we get the error check in the correct place
field := st.Field(i).Name()
fmt.Fprintf(b, `// Only pack salt if value is not "-", i.e. empty
if rr.%s != "-" {
off, err = packStringHex(rr.%s, msg, off)
if err != nil {
Move RR header packing out of generated code (#885)
2019-01-03 23:39:14 +00:00
return off, err
Only pack NSEC3/NSEC3PARAM salt when it is not empty (#474) * Only pack when salt is not empty * Actually generate things properly * Use string literal
2017-03-21 19:35:21 +00:00
}
}
`, field, field)
Handle empty salt value (#392) When removing the reflection we inadvertely also removed the code for handling empty salt values in NSEC3 and NSEC3PARAM. These are somewhat annoying because the text representation is '-', which is not valid hex.
2016-07-26 03:20:27 +00:00
continue
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
case strings.HasPrefix(st.Tag(i), `dns:"size-hex`): // size-hex can be packed just like hex
fallthrough
case st.Tag(i) == `dns:"hex"`:
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
o("off, err = packStringHex(rr.%s, msg, off)\n")
Add NULL record (#840) Sorely missing from this library. Add it. As there is no presentation format the String method for this type puts a comment in front of it. Signed-off-by: Miek Gieben <miek@miek.nl>
2018-12-30 09:45:18 +00:00
case st.Tag(i) == `dns:"any"`:
o("off, err = packStringAny(rr.%s, msg, off)\n")
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
case st.Tag(i) == `dns:"octet"`:
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
o("off, err = packStringOctet(rr.%s, msg, off)\n")
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
case st.Tag(i) == "":
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
switch st.Field(i).Type().(*types.Basic).Kind() {
case types.Uint8:
o("off, err = packUint8(rr.%s, msg, off)\n")
case types.Uint16:
o("off, err = packUint16(rr.%s, msg, off)\n")
case types.Uint32:
o("off, err = packUint32(rr.%s, msg, off)\n")
case types.Uint64:
o("off, err = packUint64(rr.%s, msg, off)\n")
case types.String:
o("off, err = packString(rr.%s, msg, off)\n")
default:
log.Fatalln(name, st.Field(i).Name())
}
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
default:
log.Fatalln(name, st.Field(i).Name(), st.Tag(i))
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
}
}
Move RR header packing out of generated code (#885)
2019-01-03 23:39:14 +00:00
fmt.Fprintln(b, "return off, nil }\n")
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
}
fmt.Fprint(b, "// unpack*() functions\n\n")
for _, name := range namedTypes {
o := scope.Lookup(name)
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
st, _ := getTypeStruct(o.Type(), scope)
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
Use an interface method for unpacking records (#884) * Use an interface method for unpacking records * Eliminate err var declaration from unpack functions * Remove pointless r.Data assignment in PrivateRR.unpack
2019-01-03 16:35:32 +00:00
fmt.Fprintf(b, "func (rr *%s) unpack(msg []byte, off int) (off1 int, err error) {\n", name)
fmt.Fprint(b, `rdStart := off
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
_ = rdStart
`)
for i := 1; i < st.NumFields(); i++ {
o := func(s string) {
fmt.Fprintf(b, s, st.Field(i).Name())
fmt.Fprint(b, `if err != nil {
Use an interface method for unpacking records (#884) * Use an interface method for unpacking records * Eliminate err var declaration from unpack functions * Remove pointless r.Data assignment in PrivateRR.unpack
2019-01-03 16:35:32 +00:00
return off, err
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
}
`)
}
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
// size-* are special, because they reference a struct member we should use for the length.
if strings.HasPrefix(st.Tag(i), `dns:"size-`) {
structMember := structMember(st.Tag(i))
structTag := structTag(st.Tag(i))
switch structTag {
case "hex":
fmt.Fprintf(b, "rr.%s, off, err = unpackStringHex(msg, off, off + int(rr.%s))\n", st.Field(i).Name(), structMember)
case "base32":
fmt.Fprintf(b, "rr.%s, off, err = unpackStringBase32(msg, off, off + int(rr.%s))\n", st.Field(i).Name(), structMember)
case "base64":
fmt.Fprintf(b, "rr.%s, off, err = unpackStringBase64(msg, off, off + int(rr.%s))\n", st.Field(i).Name(), structMember)
default:
log.Fatalln(name, st.Field(i).Name(), st.Tag(i))
}
fmt.Fprint(b, `if err != nil {
Use an interface method for unpacking records (#884) * Use an interface method for unpacking records * Eliminate err var declaration from unpack functions * Remove pointless r.Data assignment in PrivateRR.unpack
2019-01-03 16:35:32 +00:00
return off, err
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
}
`)
continue
}
Generate pack/unpack for all RRs (#360) Add dns:txt parsing helper to prevent compile errors. This allows us to generate all unpack/pack function. Add pack to the RR interface definition and add this method to PrivateRR. We still use typeToUnpack to select which types don't use reflection.
2016-06-05 06:53:12 +00:00
if _, ok := st.Field(i).Type().(*types.Slice); ok {
switch st.Tag(i) {
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
case `dns:"-"`: // ignored
Generate pack/unpack for all RRs (#360) Add dns:txt parsing helper to prevent compile errors. This allows us to generate all unpack/pack function. Add pack to the RR interface definition and add this method to PrivateRR. We still use typeToUnpack to select which types don't use reflection.
2016-06-05 06:53:12 +00:00
case `dns:"txt"`:
o("rr.%s, off, err = unpackStringTxt(msg, off)\n")
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
case `dns:"opt"`:
o("rr.%s, off, err = unpackDataOpt(msg, off)\n")
case `dns:"nsec"`:
o("rr.%s, off, err = unpackDataNsec(msg, off)\n")
Implement SVCB (#1067) * Implement SVCB * Fix serialization and deserialization of double quotes * More effort (?) 4 months old commit * DEBUG * _ * Presentation format serialization/deserialization * _ Remove generated * Progress on presentation format parse & write * _ * Finish parsing presentation format * Regenerate * Pack unpack * Move to svcb.go Scan_rr.go and types.go should be untouched now * :bug: Thanks ghedo * Definitions * TypeHTTPSSVC * Generated and isDuplicate * Goodbye lenient functions Now private key=value pairs have to be defined as structs too. They are no longer automatically named as KeyNNNNN * Encode/decode * Experimental svc * Read method * Implement some of the methods, use trick... to report where the error is while reading it. This should be applied to EDNS too. Todo: Find if case can only contain e := new(SVC_ALPN) and rest moved out Also fix two compile errors * Add SVC_LOCAL methods, reorder, remove alpn value, bugs * Errors * Alpn, make it build * Correct testsuite * Fully implement parser Change from keeping a state variable to reading in one iteration until the key=value pair is fully consumed * Simplify and document EDNS should be simplified too * Attempt to fix fuzzer And Alpn bug * A bug and change type values to match @ghedo's implementation * IP bug Also there are two ip duplicating patterns, one with copy, one with append. Maybe change it to be consistent. * Check for strictly increasing keys as required * Don't panic on invalid alpn * Redundant check, don't modify original array * Size calculation * Fix the fuzzer, match the style * 65535 is reserved too, don't delay errors * Check keyNNN, check for aliasform having values * IPvNHint is an array * Fix ipvNHint * Rename everything * Unrecognized keys according to the updated specification * Skip zero-length structs in generators. Fix CI * Doc cleanup * Off by one * Add parse tests * Check if private key doesn't collide with known key, invalid tests * Disallow IPv4 as IPv6. More tests. Related #1107 * Style fixes * More consistency, more tests * :bug: Deep copy as in the documentation a := make([]net.IP, 1) a[0] = net.ParseIP("1.1.1.1").To4() b := append(make([]net.IP, 0, 1), a...) b[0] = net.ParseIP("3.1.1.1").To4() fmt.Println(a[0][0]) * Make tests readable * Move valid parse tests to different file * :bug: One of previous commits not fully committed * Test binary single value encoding/decoding and full encode/decode * Add worst-case grows to builders, :bug: Wrong visible character range, redundant tests * Testing improvements And don't convert to IPv4 twice * Doc update only * Document worst case allocations and ipv6 can be at most of length 39, not 40 * Redundant IP copy, consistent IPv6 behavior, fix deep copy * isDuplicate for SVCB * Optimizations * echoconfig * Svc => SVCB * Fix CI * Regenerate after REBASE (2) Rebased twice on 15th and 20th May * Rename svc, use escapeByte. * Fix parsing whitespaces between quotes, rename ECHOHOConfig * resolve Remove svcbFieldLen Use reverseInt Uppercase SVCB Rename key_value "invalid" => bad Alpn comments > 65535 check Unneeded slices * a little more read => parse IP array meaning Force pushed because forgot to change read in svcb_test.go * HTTPSSVC -> HTTPS * Use new values * mandatory code https://github.com/MikeBishop/dns-alt-svc/pull/205 * Resolve comments Rename svcb-pairs Remove SVCB_PRIVATE ranges Comment on SVCB_KEY65535 ParseError return l.token rename svcbKeyToString and svcbStringToKey privatize SVCBKeyToString, SVCBStringToKey * Refactor 1 Rename sorted, originalPairs Use append instead of copy Use svcb_RESERVED instead of 65535, with it now being private "type SVCBKey uint16" * Refactor 2 svcbKeyToString as method svcbStringToKey updated after key 0 :bug: mandatory has missing key Rename str idx < 0 * Refactor 3 Use l.token as z var key, value string Comment wrap 0: Sentences with '.' keyValue => kv * Refactor 4 * Refactor 5 len() int * Refactor 6 * Refactor 7 * Test remove parsing * Error messages * Rewrite two estimate comments * parse shouldn't modify original array :bug: * Remove two unneeded comments * Address review comments Push 2 because can't build fuzzer python Push 3 to try again * Simplify argument duplication as per tmthrgd's suggestion And add the relevant test Force push edit: Make sorting code fit into one line * Rewrite ECHConfig and address the review * Remove the optional tab * Add To4() Check * More cleanup and fix mandatory not sorting bug
2020-10-11 07:09:36 +00:00
case `dns:"pairs"`:
o("rr.%s, off, err = unpackDataSVCB(msg, off)\n")
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
case `dns:"domain-name"`:
o("rr.%s, off, err = unpackDataDomainNames(msg, off, rdStart + int(rr.Hdr.Rdlength))\n")
APL record support (#1058) * APL record: add structure and code point * APL record: add wire format support * APL record: add presentation format support * APL record: add isDuplicate implementation * APL record: add copy implementation * APL record: add len implementation * APL record: run go generate * APL record: fix condition checking for equality * APL record: use switches to map family to address length * APL record: check bounds of individual fields rather than whole header * APL record: stylistic changes * APL record: remove APLPrefix methods from public interface * APL record: update README * APL record: additional cleanup for code review * APL record: change return type from pointer to struct * APL record: refactor of pack and unpack to eliminate extra variables
2020-01-03 12:41:45 +00:00
case `dns:"apl"`:
o("rr.%s, off, err = unpackDataApl(msg, off)\n")
Generate pack/unpack for all RRs (#360) Add dns:txt parsing helper to prevent compile errors. This allows us to generate all unpack/pack function. Add pack to the RR interface definition and add this method to PrivateRR. We still use typeToUnpack to select which types don't use reflection.
2016-06-05 06:53:12 +00:00
default:
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
log.Fatalln(name, st.Field(i).Name(), st.Tag(i))
Generate pack/unpack for all RRs (#360) Add dns:txt parsing helper to prevent compile errors. This allows us to generate all unpack/pack function. Add pack to the RR interface definition and add this method to PrivateRR. We still use typeToUnpack to select which types don't use reflection.
2016-06-05 06:53:12 +00:00
}
continue
}
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
switch st.Tag(i) {
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
case `dns:"-"`: // ignored
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
case `dns:"cdomain-name"`:
fallthrough
case `dns:"domain-name"`:
o("rr.%s, off, err = UnpackDomainName(msg, off)\n")
case `dns:"a"`:
o("rr.%s, off, err = unpackDataA(msg, off)\n")
case `dns:"aaaa"`:
o("rr.%s, off, err = unpackDataAAAA(msg, off)\n")
case `dns:"uint48"`:
o("rr.%s, off, err = unpackUint48(msg, off)\n")
case `dns:"txt"`:
o("rr.%s, off, err = unpackString(msg, off)\n")
case `dns:"base32"`:
o("rr.%s, off, err = unpackStringBase32(msg, off, rdStart + int(rr.Hdr.Rdlength))\n")
case `dns:"base64"`:
o("rr.%s, off, err = unpackStringBase64(msg, off, rdStart + int(rr.Hdr.Rdlength))\n")
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
case `dns:"hex"`:
o("rr.%s, off, err = unpackStringHex(msg, off, rdStart + int(rr.Hdr.Rdlength))\n")
Add NULL record (#840) Sorely missing from this library. Add it. As there is no presentation format the String method for this type puts a comment in front of it. Signed-off-by: Miek Gieben <miek@miek.nl>
2018-12-30 09:45:18 +00:00
case `dns:"any"`:
o("rr.%s, off, err = unpackStringAny(msg, off, rdStart + int(rr.Hdr.Rdlength))\n")
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
case `dns:"octet"`:
o("rr.%s, off, err = unpackStringOctet(msg, off)\n")
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
case "":
switch st.Field(i).Type().(*types.Basic).Kind() {
case types.Uint8:
o("rr.%s, off, err = unpackUint8(msg, off)\n")
case types.Uint16:
o("rr.%s, off, err = unpackUint16(msg, off)\n")
case types.Uint32:
o("rr.%s, off, err = unpackUint32(msg, off)\n")
case types.Uint64:
o("rr.%s, off, err = unpackUint64(msg, off)\n")
case types.String:
o("rr.%s, off, err = unpackString(msg, off)\n")
default:
log.Fatalln(name, st.Field(i).Name())
}
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
default:
log.Fatalln(name, st.Field(i).Name(), st.Tag(i))
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
}
// If we've hit len(msg) we return without error.
if i < st.NumFields()-1 {
fmt.Fprintf(b, `if off == len(msg) {
Use an interface method for unpacking records (#884) * Use an interface method for unpacking records * Eliminate err var declaration from unpack functions * Remove pointless r.Data assignment in PrivateRR.unpack
2019-01-03 16:35:32 +00:00
return off, nil
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
}
`)
}
}
Use an interface method for unpacking records (#884) * Use an interface method for unpacking records * Eliminate err var declaration from unpack functions * Remove pointless r.Data assignment in PrivateRR.unpack
2019-01-03 16:35:32 +00:00
fmt.Fprintf(b, "return off, nil }\n\n")
Skip reflection for most types (#369) Make the reflection types a black list (these types use (or should use) the tag 'size-xxx' in their struct definition.s HIP, IPSECKEY, NSEC3, TSIG All other types don't use reflection anymore. * Return a pointer to the header when there is no rdata, this restores old behavior. The rest of the conversion mostly hangs on getting size-hex right, but then packStruct and packStructValue and the unpack variant can be killed. * Generate pack and unpack for all embedded types as well. * Fix PrivateRRs, register an unpack function as well, when you register a new PrivateRR. * Add the tag octet, nsec, []domains and more to msg_helper.go
2016-06-12 15:09:37 +00:00
}
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
// gofmt
res, err := format.Source(b.Bytes())
if err != nil {
b.WriteTo(os.Stderr)
log.Fatal(err)
}
// write result
f, err := os.Create("zmsg.go")
fatalIfErr(err)
defer f.Close()
f.Write(res)
}
Kill all reflection when packing/unpacking RR (#372) Update the size-xxx-member tags to point to another field in the struct that should be used for the length in that field. Fix NSEC3/HIP and TSIG to use to this and generate the correct pack/unpack functions for them. Remove IPSECKEY from the lib and handle it as an unknown record - it is such a horrible RR, needed kludges before - now just handle it as an unknown RR. All types now use generated pack and unpack functions. The blacklist is removed.
2016-06-12 17:31:50 +00:00
// structMember will take a tag like dns:"size-base32:SaltLength" and return the last part of this string.
func structMember(s string) string {
fields := strings.Split(s, ":")
if len(fields) == 0 {
return ""
}
f := fields[len(fields)-1]
// f should have a closing "
if len(f) > 1 {
return f[:len(f)-1]
}
return f
}
// structTag will take a tag like dns:"size-base32:SaltLength" and return base32.
func structTag(s string) string {
fields := strings.Split(s, ":")
if len(fields) < 2 {
return ""
}
return fields[1][len("\"size-"):]
}
Remove (most) reflection Remove the use of reflection when packing and unpacking, instead generate all the pack and unpack functions using msg_generate. This will generate zmsg.go which in turn calls the helper functions from msg_helper.go. This increases the speed by about ~30% while cutting back on memory usage. Not all RRs are using it, but that will be rectified in upcoming PR. Most of the speed increase is in the header/question section parsing. These functions *are* not generated, but straight forward enough. The implementation can be found in msg.go. The new code has been fuzzed by go-fuzz, which turned up some issues. All files that started with 'z', and not autogenerated were renamed, i.e. zscan.go is now scan.go. Reflection is still used, in subsequent PRs it will be removed entirely.
2016-05-14 16:56:20 +00:00
func fatalIfErr(err error) {
if err != nil {
log.Fatal(err)
}
}