Tracking issue for problems on RISC-V Linux platform

While RISC-V is still an unofficial platform, there are lots of interest to push it to experimental tier(#2876) .

This issue tracks bugs that are closely related to linux riscv64 platform.

Build Errors

(main, 24.x) node-v24.1.0 fails to build on riscv / gcc-15 node#58485 (should be fixed by deps: V8: cherry-pick 0ce2edb7adfd node#58773)

Bugs

riscv64: strange behavior of exception handling node#47522. This one is caused by a GCC regression and affects arm64 as well.
(v24.x) [RISCV] Check failed: (trampoline_pos - fixup_pos) <= kMaxBranchOffset. node#60895 [Regression] V8 crashed while building AdGuard Home with webpack riscv-forks/electron#9. Should be fixed by ~~[riscv] Fix check failed in bind_to node#58746~~ [v24.x][riscv] Fix check failed in bind_to node#60962
Node.js 23 wasm related error (reproducer: 389-ds-base) riscv-forks/electron#7. This one also appears to be a GCC specific bug as node compiled with clang is not affected.
(v22.x, v23.x) [Regression] Zettlr crashes with illegal instruction riscv-forks/electron#6. Fixed by deps: V8: cherry-pick 9ab40592f697 node#56781
(all versions after v20.x) [Regression] v8 wasm error when running yarn install with yarn v3+ riscv-forks/electron#3. WebAssembly's trap handler implementation uses a lot of virtual memory for guard pages, where risc-v systems with Sv39 only have 256GB virtual memory and would quickly run out of address space when multiple wasm are running.
- Disabling trap handler would solve it but also leads to other new bugs like Unreachable error in wasm rollup when building on riscv64 riscv-forks/electron#8
(all versions from at least v16.x) Sporadic Nodejs segfault on sg2042 revyos/revyos#27. This one only triggers on SG2042, which has 64 cores. It is mysterious as it happens by chance and when gdb is attached, the call stack is already corrupted so the cause is still unknown.
- This one also appears to affect loongarch64. (fixed in https://chromium-review.googlesource.com/c/v8/v8/+/6786372)
- Should be fixed with deps: V8: backport 6a0a25abaed3 node#61666
Node.js test-snapshot-reproducible.js failing on riscv64

Performance Regressions

(v21.x, v22.x) [electron26] Severe performance regression when running Code OSS riscv-forks/electron#1. Fixed by [RISC-V] Fix performance regression node#53412

Thanks for collating these links. I'm dumping this info here just now so I have it logged and shared. I'm running various builds at the moment natively (trying to populate my ccaches!) and with the cross-compilers. At the moment the cross compilers are having problems building, and since those are how the unofficial builds need to run it's something we'll want to resolve.

For reference in 24.x I'm getting:

  CXX(host) /home/node/node/out/Release/obj.host/v8_base_without_compiler/deps/v8/src/api/api-arguments.o
In file included from ../deps/v8/src/execution/vm-state-inl.h:12,
                 from ../deps/v8/src/api/api-arguments-inl.h:13,
                 from ../deps/v8/src/api/api-arguments.cc:7:
../deps/v8/src/execution/simulator.h: In static member function 'static int v8::internal::SimulatorStack::JSStackLimitMargin()':
../deps/v8/src/execution/simulator.h:66:55: error: 'JSStackLimitMargin' is not a member of 'v8::internal::Simulator'
   66 |   static int JSStackLimitMargin() { return Simulator::JSStackLimitMargin(); }
      |                                                       ^~~~~~~~~~~~~~~~~~
make: *** [tools/v8_gypfiles/v8_base_without_compiler.host.mk:1140: /home/node/node/out/Release/obj.host/v8_base_without_compiler/deps/v8/src/api/api-arguments.o] Error 1
rm f41dec747f7d07bc6189224d714bfe6f70ce7700.intermediate 4b9faa377f69f07a06bcf45eb52d6b09bd3dc776.intermediate 45471a7c870f238f988ac4c73ccc531d343b1d2d.intermediate 5cf52c47b2c7a3d1debdf7e64552adabc181a77d.intermediate 91e5efd5c01b316d5563811267cb2f34bf83226a.intermediate
make: Leaving directory '/home/node/node/out'

and for main I'm getting:

In file included from ../deps/v8/third_party/highway/src/hwy/ops/scalar-inl.h:24,
                 from ../deps/v8/third_party/highway/src/hwy/highway.h:722,
                 from ../deps/v8/src/json/json-stringifier.cc:10:
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h: In instantiation of 'struct hwy::N_SCALAR::detail::FixedTagChecker<unsigned char, 16>':
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h:423:7:   required by substitution of 'template<class T, long unsigned int kNumLanes> using hwy::N_SCALAR::FixedTag = typename hwy::N_SCALAR::detail::FixedTagChecker::type [with T = unsigned char; long unsigned int kNumLanes = 16]'
  423 | using FixedTag = typename detail::FixedTagChecker<T, kNumLanes>::type;
      |       ^~~~~~~~
../deps/v8/src/json/json-stringifier.cc:3379:29:   required from 'bool v8::internal::FastJsonStringifier<Char>::AppendStringSIMD(const SrcChar*, size_t, const v8::internal::DisallowGarbageCollection&) [with SrcChar = unsigned char; Char = unsigned char; size_t = long unsigned int; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'
 3379 |   hw::FixedTag<SrcChar, 16> tag;
      |                             ^~~
../deps/v8/src/json/json-stringifier.cc:3316:28:   required from 'bool v8::internal::FastJsonStringifier<Char>::AppendString(const SrcChar*, size_t, const v8::internal::DisallowGarbageCollection&) [with SrcChar = unsigned char; Char = unsigned char; size_t = long unsigned int; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'
 3316 |     return AppendStringSIMD(chars, length, no_gc);
      |            ~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~
../deps/v8/src/json/json-stringifier.cc:2531:17:   required from 'v8::internal::FastJsonStringifierResult v8::internal::FastJsonStringifier<Char>::SerializeString(v8::internal::Tagged<v8::internal::HeapObject>, const v8::internal::DisallowGarbageCollection&) [with StringT = v8::internal::SeqOneByteString; Char = unsigned char; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'
 2531 |     AppendString(chars, length, no_gc);
      |     ~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~
../deps/v8/src/json/json-stringifier.cc:2554:47:   required from 'v8::internal::FastJsonStringifierResult v8::internal::FastJsonStringifier<Char>::TrySerializeSimpleObject(v8::internal::Tagged<v8::internal::Union<v8::internal::Smi, v8::internal::HeapNumber, v8::internal::BigInt, v8::internal::String, v8::internal::Symbol, v8::internal::Boolean, v8::internal::Null, v8::internal::Undefined, v8::internal::JSReceiver> >) [with Char = unsigned char]'
 2554 |       return SerializeString<SeqOneByteString>(obj, no_gc);
      |              ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~
../deps/v8/src/json/json-stringifier.cc:3167:38:   required from 'v8::internal::FastJsonStringifierResult v8::internal::FastJsonStringifier<Char>::SerializeObject(v8::internal::Tagged<v8::internal::Union<v8::internal::Smi, v8::internal::HeapNumber, v8::internal::BigInt, v8::internal::String, v8::internal::Symbol, v8::internal::Boolean, v8::internal::Null, v8::internal::Undefined, v8::internal::JSReceiver> >, const v8::internal::DisallowGarbageCollection&) [with Char = unsigned char; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'
 3167 |   FastJsonStringifierResult result = TrySerializeSimpleObject(object);
      |                                      ^~~~~~~~~~~~~~~~~~~~~~~~
../deps/v8/src/json/json-stringifier.cc:3509:43:   required from here
 3509 |       one_byte_stringifier.SerializeObject(*object, no_gc);
      |       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h:369:27: error: static assertion failed: Too many lanes
  369 |   static_assert(kNumLanes <= HWY_LANES(T), "Too many lanes");
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h:369:27: note: the comparison reduces to '(16 <= 1)'
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h: In instantiation of 'struct hwy::N_SCALAR::Simd<unsigned char, 16, 0>':
../deps/v8/src/json/json-stringifier.cc:3379:29:   required from 'bool v8::internal::FastJsonStringifier<Char>::AppendStringSIMD(const SrcChar*, size_t, const v8::internal::DisallowGarbageCollection&) [with SrcChar = unsigned char; Char = unsigned char; size_t = long unsigned int; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'

and for main I'm getting:

In file included from ../deps/v8/third_party/highway/src/hwy/ops/scalar-inl.h:24,
                 from ../deps/v8/third_party/highway/src/hwy/highway.h:722,
                 from ../deps/v8/src/json/json-stringifier.cc:10:
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h: In instantiation of 'struct hwy::N_SCALAR::detail::FixedTagChecker<unsigned char, 16>':
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h:423:7:   required by substitution of 'template<class T, long unsigned int kNumLanes> using hwy::N_SCALAR::FixedTag = typename hwy::N_SCALAR::detail::FixedTagChecker::type [with T = unsigned char; long unsigned int kNumLanes = 16]'
  423 | using FixedTag = typename detail::FixedTagChecker<T, kNumLanes>::type;
      |       ^~~~~~~~
../deps/v8/src/json/json-stringifier.cc:3379:29:   required from 'bool v8::internal::FastJsonStringifier<Char>::AppendStringSIMD(const SrcChar*, size_t, const v8::internal::DisallowGarbageCollection&) [with SrcChar = unsigned char; Char = unsigned char; size_t = long unsigned int; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'
 3379 |   hw::FixedTag<SrcChar, 16> tag;
      |                             ^~~
../deps/v8/src/json/json-stringifier.cc:3316:28:   required from 'bool v8::internal::FastJsonStringifier<Char>::AppendString(const SrcChar*, size_t, const v8::internal::DisallowGarbageCollection&) [with SrcChar = unsigned char; Char = unsigned char; size_t = long unsigned int; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'
 3316 |     return AppendStringSIMD(chars, length, no_gc);
      |            ~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~
../deps/v8/src/json/json-stringifier.cc:2531:17:   required from 'v8::internal::FastJsonStringifierResult v8::internal::FastJsonStringifier<Char>::SerializeString(v8::internal::Tagged<v8::internal::HeapObject>, const v8::internal::DisallowGarbageCollection&) [with StringT = v8::internal::SeqOneByteString; Char = unsigned char; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'
 2531 |     AppendString(chars, length, no_gc);
      |     ~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~
../deps/v8/src/json/json-stringifier.cc:2554:47:   required from 'v8::internal::FastJsonStringifierResult v8::internal::FastJsonStringifier<Char>::TrySerializeSimpleObject(v8::internal::Tagged<v8::internal::Union<v8::internal::Smi, v8::internal::HeapNumber, v8::internal::BigInt, v8::internal::String, v8::internal::Symbol, v8::internal::Boolean, v8::internal::Null, v8::internal::Undefined, v8::internal::JSReceiver> >) [with Char = unsigned char]'
 2554 |       return SerializeString<SeqOneByteString>(obj, no_gc);
      |              ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~
../deps/v8/src/json/json-stringifier.cc:3167:38:   required from 'v8::internal::FastJsonStringifierResult v8::internal::FastJsonStringifier<Char>::SerializeObject(v8::internal::Tagged<v8::internal::Union<v8::internal::Smi, v8::internal::HeapNumber, v8::internal::BigInt, v8::internal::String, v8::internal::Symbol, v8::internal::Boolean, v8::internal::Null, v8::internal::Undefined, v8::internal::JSReceiver> >, const v8::internal::DisallowGarbageCollection&) [with Char = unsigned char; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'
 3167 |   FastJsonStringifierResult result = TrySerializeSimpleObject(object);
      |                                      ^~~~~~~~~~~~~~~~~~~~~~~~
../deps/v8/src/json/json-stringifier.cc:3509:43:   required from here
 3509 |       one_byte_stringifier.SerializeObject(*object, no_gc);
      |       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h:369:27: error: static assertion failed: Too many lanes
  369 |   static_assert(kNumLanes <= HWY_LANES(T), "Too many lanes");
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h:369:27: note: the comparison reduces to '(16 <= 1)'
../deps/v8/third_party/highway/src/hwy/ops/shared-inl.h: In instantiation of 'struct hwy::N_SCALAR::Simd<unsigned char, 16, 0>':
../deps/v8/src/json/json-stringifier.cc:3379:29:   required from 'bool v8::internal::FastJsonStringifier<Char>::AppendStringSIMD(const SrcChar*, size_t, const v8::internal::DisallowGarbageCollection&) [with SrcChar = unsigned char; Char = unsigned char; size_t = long unsigned int; v8::internal::DisallowGarbageCollection = v8::internal::PerThreadAssertScopeEmpty<false, v8::internal::SAFEPOINTS_ASSERT, v8::internal::HEAP_ALLOCATION_ASSERT>]'

This is probably caused by google/highway@f38ee25 ,
which will treat the gcc compiler as broken if the version is less than 16.0, which is not even released yet.
When the gcc compiler is detected as broken, it will fallback to HWY_SCALAR instead of HWY_EMU128 and causing the assertion error.

So one way is to accept that the compiler might do broken stuff and compile anyway with -DHWY_BROKEN_EMU128=0, another way is to use GCC >= 16 or clang.

Link: google/highway#2406 (comment)

Notes to self:

Commit c2843b722ca builds ok with the cross-compiler and is possibly the last one to do so.
The subsequent V8 change in deps: update V8 to 14.2 node#60111 doesn't work (albeit seemingly not with the same error that we now get in HEAD)
HEAD builds with GCC14 cross-compiler seem to build ok if build with CXXFLAGS=-DHWY_BROKEN_EMU128=0 as per the comment above
Raised V8 backport [v24.x][riscv] V8 backport to resolve compile failure node#60989 to resolve the build problems with v24.x

Initial test failure list

Failures on v25.x on TH1520

out/Release/node /home/sxa/node/test/sea/test-single-executable-application-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/es-module/test-esm-detect-ambiguous.mjs
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/es-module/test-esm-loader-entry-url.mjs
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/es-module/test-esm-tla-syntax-errors-not-recognized-as-tla-error.mjs
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-exec-argv.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/es-module/test-typescript-commonjs.mjs
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-snapshot-and-code-cache.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/es-module/test-typescript-eval.mjs
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/es-module/test-typescript-transform.mjs
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/es-module/test-typescript.mjs
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-assert-partial-deep-equal.js
out/Release/node --expose_gc --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-abortsignal-drop-settled-signals.mjs
out/Release/node /home/sxa/node/test/parallel/test-buffer-indexof.js
out/Release/node /home/sxa/node/test/parallel/test-compile-cache-typescript-commonjs.js
out/Release/node /home/sxa/node/test/parallel/test-compile-cache-typescript-strip-miss.js
out/Release/node /home/sxa/node/test/parallel/test-compile-cache-typescript-transform.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-config-file.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-module-strip-types.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-node-output-eval.mjs
out/Release/node /home/sxa/node/test/parallel/test-repl-paste-big-data.js
out/Release/node /home/sxa/node/test/parallel/test-runner-cli.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-runner-global-setup-teardown.mjs
out/Release/node /home/sxa/node/test/parallel/test-snapshot-reproducible.js
out/Release/node /home/sxa/node/test/parallel/test-util-getcallsites.js
out/Release/node /home/sxa/node/test/parallel/test-worker-syntax-error.js
out/Release/node --expose-internals /home/sxa/node/test/sequential/test-worker-heapsnapshot-options.js

Failures on bpi-F3 on Node 24

out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/report/test-report-fatalerror-oomerror-set.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/report/test-report-fatalerror-oomerror-compact.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/report/test-report-fatalerror-oomerror-filename.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-repl-paste-big-data.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-runner-run.mjs
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-snapshot-reproducible.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-strace-openat-openssl.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-stream2-read-sync-stack.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-worker-resource-limits.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/abort/test-abort-fatal-error.js
out/Release/node --expose-internals --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-worker-heapsnapshot-options.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-asset-keys.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-asset-keys-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-assets.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-assets-raw.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-disable-experimental-sea-warning.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-exec-argv.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-exec-argv-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-exec-argv-extension-cli.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-exec-argv-extension-env.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-exec-argv-extension-none.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-inspect.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-inspect-in-sea-flags.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-snapshot.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-snapshot-and-code-cache.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-snapshot-worker.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sea/test-single-executable-application-use-code-cache.js

(Edited to add the following ones after the initial results above)

v22.x on BPiF3

Failed tests:
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-snapshot-reproducible.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-strace-openat-openssl.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-worker-messaging.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-asset-keys.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-asset-keys-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-assets.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-assets-raw.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-disable-experimental-sea-warning.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv-extension-cli.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv-extension-env.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv-extension-none.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-inspect.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-inspect-in-sea-flags.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-snapshot.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-snapshot-and-code-cache.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-snapshot-worker.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-use-code-cache.js
out/Release/node --expose-internals --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-worker-heapsnapshot-options.js
out/Release/node --expose-internals --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-write-heapsnapshot-options.js

v22.x on BPiF3 with backports from the next comment

Note: cherry-picked commits are cdc3ca8a7ed and 22a3eb00e85 and only affect the tests:

[139:10|% 100|+ 4631|-  23]: Done                                             

Failed tests:
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/report/test-report-fatalerror-oomerror-filename.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-child-process-exec-maxbuf.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-snapshot-reproducible.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/parallel/test-worker-messaging.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-asset-keys.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-asset-keys-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-assets.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-assets-raw.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-disable-experimental-sea-warning.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv-empty.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv-extension-cli.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv-extension-env.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-exec-argv-extension-none.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-inspect.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-inspect-in-sea-flags.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-snapshot.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-snapshot-and-code-cache.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-snapshot-worker.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-single-executable-application-use-code-cache.js
out/Release/node --expose-internals --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/sxa/node/test/sequential/test-worker-heapsnapshot-options.js

Noting that while the SEA tests fail consistently, the others are a bit more variable. `parallel/test-net-pingpong.js`, `parallel/test-worker-messaging.js`, `parallel/test-child-process-exec-maxbuf.js`, `sequential/test-get-heapsnapshot-options.js` and `test-snapshot-reproducible.js` and `sequential/test-worker-heapsnapshot-options.js` have all failed at least one one this build.

v25.x-staging (commit c57a40c898c) on bpiF3

[52:44|% 100|+ 4889|-  30]: Done                                              

Failed tests:
out/Release/node /home/sxa/node/test/report/test-report-fatalerror-oomerror-directory.js
out/Release/node /home/sxa/node/test/report/test-report-fatalerror-oomerror-compact.js
out/Release/node /home/sxa/node/test/report/test-report-fatalerror-oomerror-filename.js
out/Release/node /home/sxa/node/test/async-hooks/test-tlswrap.js
out/Release/node /home/sxa/node/test/parallel/test-performance-many-marks.js
out/Release/node /home/sxa/node/test/parallel/test-repl-paste-big-data.js
out/Release/node /home/sxa/node/test/parallel/test-snapshot-reproducible.js
out/Release/node /home/sxa/node/test/parallel/test-stream2-read-sync-stack.js
out/Release/node /home/sxa/node/test/parallel/test-trace-events-fs-sync.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-exec-argv.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-asset-keys-empty.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-assets.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-asset-keys.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-disable-experimental-sea-warning.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-assets-raw.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-empty.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-inspect.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-exec-argv-extension-cli.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-snapshot-and-code-cache.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-inspect-in-sea-flags.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-exec-argv-extension-env.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-snapshot.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-exec-argv-extension-none.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-exec-argv-empty.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-snapshot-worker.js
out/Release/node /home/sxa/node/test/sea/test-single-executable-application-use-code-cache.js
out/Release/node /home/sxa/node/test/node-api/test_sea_addon/test.js
out/Release/node /home/sxa/node/test/abort/test-abort-fatal-error.js
out/Release/node --expose-internals /home/sxa/node/test/sequential/test-worker-heapsnapshot-options.js

(Noting for my own reference that the BPiF3 (SpacemiT-K1) takes about 45 minutes to run the v22 test suite, excluding test-doc. That included timeouts - probably due to a crash - on the single-executable-application tests)

I remember that on SG2042, nodejs 25.2.1 with nodejs/node#60591 and nodejs/node#60588 applied passed all tests except test-snapshot-reproducible.js, which is currently being tracked upstream at https://issues.chromium.org/issues/460332153

@kxxt

Thanks @kxxt I'll take a look at those since I haven't run tests with those included yet and some of the failures were OOMs which could be related to 60591. Was your SG2042 the Milk-V Pioneer or a different board?

Thanks @kxxt I'll take a look at those since I haven't run tests with those included yet and some of the failures were OOMs which could be related to 60591. Was your SG2042 the Milk-V Pioneer or a different board?

Yes. It's a Milk-V Pioneer that I could access remotely(Thanks to felixonmars!). Unfortunately there is a hard-to-debug problem(revyos/revyos#27) specific to that board that could cause node.js to segfault sometimes, which happens sometimes during tests.

There is also bug in old vendor kernel(5.10.x) for TH1520, which could cause weird issues for node.js. That is also mentioned in revyos/revyos#27 (The case of lpi4a).

I had remote access to someone else's Pioneer too at one point but it's not available now. Mine are running a 5.10-based kernel from Ubuntu 24.04 so that link is very useful. It is about time I updated my lpi4a though so I should probably get hold of one of the later kernels as you suggested. I do worry about how reliable the C910/C920 cores are though as we've seen unusual crashes with java on systems with those cores. I'll perhaps need to do a direct compare with the same build on bpiF3 and TH1520 and see if they give the same results, but I've been building different versions on each so far.

@kxxt

I remember that on SG2042, nodejs 25.2.1 with nodejs/node#60591 and nodejs/node#60588 applied passed all tests except test-snapshot-reproducible.js,

@kxxt Did you have to do anything special to get the single executable appliation tests in test/sea to pass? On a TH1520 with 25.x which includes those two commits I'm getting SIGILLs (Could it be trying to run RVV1.0 stuff? Doesn't look like node's v8-options lets you disable those) and elsewhere (e.g. under qemu or on the bpiF3) I'm still getting consistent segmentation faults.
(EDIT: These have passed in some runs in the TH1520, but I'm not yet sure what was special about those!)

I remember that on SG2042, nodejs 25.2.1 with nodejs/node#60591 and nodejs/node#60588 applied passed all tests except test-snapshot-reproducible.js,

@kxxt Did you have to do anything special to get the single executable appliation tests in test/sea to pass? On a TH1520 with 25.x which includes those two commits I'm getting SIGILLs (Could it be trying to run RVV1.0 stuff? Doesn't look like node's v8-options lets you disable those) and elsewhere (e.g. under qemu or on the bpiF3) I'm still getting consistent segmentation faults. (EDIT: These have passed in some runs in the TH1520, but I'm not yet sure what was special about those!)

I don't think I have done anything special for those tests. But the SIGILLs might be related to RVV 1.0 stuff in highway.
It would be a better idea to confirm it in gdb.
Because RVV runtime dispatch is broken in highway(https://github.com/google/highway/blob/4262db0ef703c7a3372f7a921a7b69ffd5db5829/hwy/detect_targets.h#L784 , note the && 0 at the end), IIRC it would unconditionally enable RVV.
So as a distribution targeting rv64gc without V extension, we disable RVV completely in highway with a patch https://github.com/felixonmars/archriscv-packages/blob/master/nodejs-lts-krypton/hwy-broken-rvv.diff
I could open a PR for that patch if you think that's the reasonable default. But I bet folks from ubuntu targeting rva23 profile (with V extension included) would not want it.

Here is a build and test log for 24.11.1: https://archriscv.felixc.at/.status/log.htm?url=logs/nodejs-lts-krypton/nodejs-lts-krypton-24.11.1-1.log (With the patches in https://github.com/felixonmars/archriscv-packages/tree/master/nodejs-lts-krypton)

That's useful information. Given that there are multiple failures on different hardware here I've raised another issue on this and will put excludes in for RISC-V on the SEA tests (which is an experimental feature just now anyway and not looking like it's going to be anything else in the meantime). We can look at doing further debugging on this in the new issue.

we disable RVV completely in highway with a patch https://github.com/felixonmars/archriscv-packages/blob/master/nodejs-lts-krypton/hwy-broken-rvv.diff

Was your passing test suite done with that patch applied on the SG2042?

I'll put some more results in the issue for information but I have successfully been able to run the SEA tests using the build from the TH1520 under qemu on my x64 laptop, but sine they will fail on most people's boards I think it's fair to exclude them for now.

I could open a PR for that patch if you think that's the reasonable default. But I bet folks from ubuntu targeting rva23 profile (with V extension included) would not want it.

I agree. Ideally we'd have this as a runtime switch (or have highway detect it in a better way) so I think I'm ok with leaving the default as-is for now and just disabling the tests.

we disable RVV completely in highway with a patch https://github.com/felixonmars/archriscv-packages/blob/master/nodejs-lts-krypton/hwy-broken-rvv.diff

Was your passing test suite done with that patch applied on the SG2042?

Yes. IIRC when targeting rv64gc with -march=rv64gc and using clang, node.js would throw build errors in highway without that patch.

* This one also appears to affect loongarch64.

fixed: v8/v8@6b1b9bc

In the interests of logging issues for search fodder even if they have solutions I'll note that if you try to build with a version of gcc <14 you'll hit the following error (Spotted in native builds). Not confirmed but I expect it's from the changes in the V8 14.2.231.9 PR which landed back in October:

In file included from ../deps/v8/src/base/cpu.cc:60:
/usr/lib/gcc/riscv64-linux-gnu/13/include/riscv_vector.h:32:2: error: #error "Vector intrinsics require the vector extension."
   32 | #error "Vector intrinsics require the vector extension."
      |  ^~~~~
../deps/v8/src/base/cpu.cc:66:67: warning: target attribute is not supported on this machine [-Wattributes]
   66 | __attribute__((target("arch=+v"))) static unsigned vlen_intrinsic() {
      |                                                                   ^
../deps/v8/src/base/cpu.cc: In function ‘unsigned int vlen_intrinsic()’:
../deps/v8/src/base/cpu.cc:67:32: error: ‘__riscv_vlenb’ was not declared in this scope
   67 |   return static_cast<unsigned>(__riscv_vlenb() * 8);
      |                                ^~~~~~~~~~~~~
make[1]: *** [tools/v8_gypfiles/v8_libbase.target.mk:225: /home/sxa/node/out/Release/obj.target/v8_libbase/deps/v8/src/base/cpu.o] Error 1
make[1]: *** Waiting for unfinished jobs....

Hi, I saw the earlier discussion in #2876 and would like to understand the current status.

What is the main remaining blocker for official linux/riscv64 binaries today? If the community wants to help, what kind of contribution would be most useful?

Thanks.

@sxa

I am not a Node collaborator, so treat this as one contributor's reading of the thread.

The checklist at the top of this issue is the live status. The items that look most load-bearing to me:

RISC-V: "Illegal instruction" crash when vector instructions detected by V8 node#64538, SIGILL on the SpaceMiT K3 when V8 emits vector instructions. @sxa has a native build and test job running and has been narrowing it down. It does not fire on the K1, and a quick attempt at reproducing it under qemu did not work either.
test-code-cache fails on a Linux/riscv64 build with FFI enabled. node#64717, test-code-cache failing when FFI is enabled. Reproducible on a BPI-F3 (K1), but the trail points at the toolchain: it passes with gcc 15, clang-22 on x64 does not show it, and @kxxt could not reproduce it on an SG2042 at all.
[Regression] v8 wasm error when running yarn install with yarn v3+ riscv-forks/electron#3, the wasm trap handler wanting more virtual address space for its guard pages than Sv39 has. That one looks structural, not a bug to squash.
The CITGM run @sxa started in July, to find which modules have no riscv64 prebuilds. Still needs a gap analysis against the other platforms.

The Single Executable Application failures (nodejs/node#61110) are parked rather than fixed, since SEA is experimental anyway.

On what would help: boards and test cycles. Most of what is left is board-specific, and the K3 failure has only ever shown up on that silicon. If you have a riscv64 machine that is neither a K1 nor a K3, running the test suite on it and reporting what breaks would add a data point nobody has today.

@sxa is the one actually doing this work, so he is better placed than me to say where the effort should go.

@gounthar

@gounthar has summarised it well. The only one of the above that I'm currently trying to look at is the illegal instruction but that's slow work building V8 :-) An initial build did not show the crash when running V8's own test suite so I need to identify how to trigger it properly. Also I'm relatively new to the V8 development process at the moment though.

The other failures I'm not actively looking at just now. As per some of the comments earlier in this issue I have run the tests on a TH1520 and JH7110 (VF2) boards but I've not done that for a while so it might be worth re-running to see if there are still issues on those that we're not closely tracking at the moment. Also as Bruno said the gap analysis to identify which of the failures in the CitGM npm module testing are specific to RISC-V would be useful. Which of the issues to pick up really depends on your level of knowledge and interest.

I'll also point out that other than the test failures/crashes the other blocker at the moment for pushing it up to a higher tier would be availability of machines of a suitable performance for inclusion in our regular CI. The K3 that I have is quick enough so I'm hoping that we'll be able to get access to more of them at some point to be able to include them in the regular CI that occurs on each PR.

Latest run (albeit patched to stop the crashes on the RVV issue and including nodejs/node#62251) on the K3's X100 cores:

13:34:50 out/Release/node /home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/parallel/test-snapshot-reproducible.js
13:34:50 out/Release/node /home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/parallel/test-util-styletext.js
13:34:50 out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/parallel/test-util-styletext-hex.js
13:34:50 out/Release/node /home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/node-api/test_sea_addon/test.js
13:34:50 out/Release/node --experimental-wasm-modules /home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/wpt/test-wasm-jsapi.mjs

test-wasm-jsapi.mjs is the only new failure relative to the results in #4099 (comment) but it passed in the next run

Error information from the failing run saved here for reference

13:27:04     [PASS] i64 mutability
13:27:04     [EXPECTED_FAILURE][FAIL] Calling setter without argument
13:27:04     set WebAssembly.Global.value): Argument 0 is required
13:27:04     TypeError: set WebAssembly.Global.value): Argument 0 is required
13:27:04         at Test.<anonymous> (/home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/fixtures/wpt/wasm/jsapi/global/value-get-set.any.js:140:24)
13:27:04         at Test.step (/home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/fixtures/wpt/resources/testharness.js:2869:25)
13:27:04         at test (/home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/fixtures/wpt/resources/testharness.js:633:30)
13:27:04         at /home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/fixtures/wpt/wasm/jsapi/global/value-get-set.any.js:125:1
13:27:04         at Script.runInThisContext (node:vm:138:12)
13:27:04         at runInThisContext (node:vm:318:38)
13:27:04         at Object.<anonymous> (/home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/test/common/wpt/worker.js:102:3)
13:27:04         at Module._compile (node:internal/modules/cjs/loader:1934:14)
13:27:04         at Object..js (node:internal/modules/cjs/loader:2074:10)
13:27:04         at Module.load (node:internal/modules/cjs/loader:1656:32)
13:27:04     Command: /home/iojs/workspace/sxa-node-test-commit-rvlinux/nodes/riscv64/out/Release/node --experimental-wasm-modules undefined 'global/value-get-set.any.js'
13:27:04     
13:27:04     [PASS] Stray argument

node-api SEA test is being excluded under nodejs/node#65569

Latest test results on SG2042 at commit 05447419fbc708fbedaa7bc9740dcd242fab0d08

out/Release/node /home/kxxt/node-riscv/test/parallel/test-snapshot-reproducible.js
out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/kxxt/node-riscv/test/sequential/test-watch-mode.mjs
out/Release/node /home/kxxt/node-riscv/test/wpt/test-encoding.js
out/Release/node /home/kxxt/node-riscv/test/wpt/test-wasm-webapi.js

The last two are flaky. The second one appears to be a regression.

[kxxt@centiskorch node-riscv]$ out/Release/node --test-reporter=./test/common/test-error-reporter.js --test-reporter-destination=stdout /home/kxxt/node-riscv/test/sequential/test-watch-mode.mjs
Test failure: 'should reload env variables when --env-file changes'
Location: test/sequential/test-watch-mode.mjs:227:3
Error: Timed out waiting for restart
    at Timeout.<anonymous> (file:///home/kxxt/node-riscv/test/sequential/test-watch-mode.mjs:96:25)
    at listOnTimeout (node:internal/timers:685:17)
    at process.processTimers (node:internal/timers:618:7)

Test failure: 'should not crash when --env-file-if-exists points to a missing file'
Location: test/sequential/test-watch-mode.mjs:306:3
Error: Timed out waiting for restart
    at Timeout.<anonymous> (file:///home/kxxt/node-riscv/test/sequential/test-watch-mode.mjs:96:25)
    at listOnTimeout (node:internal/timers:685:17)
    at process.processTimers (node:internal/timers:618:7)

Test failure: 'should support multiple --env-file flags'
Location: test/sequential/test-watch-mode.mjs:944:3
Error: Timed out waiting for restart
    at Timeout.<anonymous> (file:///home/kxxt/node-riscv/test/sequential/test-watch-mode.mjs:96:25)
    at listOnTimeout (node:internal/timers:685:17)
    at process.processTimers (node:internal/timers:618:7)

Yeah I have seen failures in test-watch-mode.js too but I wasn't yet certain if it was consistent as I've been focussing on RVV stuff on v26.

Did you only get the failure on that single test and not the other test-watch-mode-* ones?

Yeah I have seen failures in test-watch-mode.js too but I wasn't yet certain if it was consistent as I've been focussing on RVV stuff on v26.

Did you only get the failure on that single test and not the other test-watch-mode-* ones?

Yes. I only get the failure for this single test. I am runni g my CI again to double check.
This should be a recent regression as I didn't get this failure on Jul 25.

I've just done a build from the current main (commit 85e93a165aa21f7796de4f38e2be7c07a232d0eb plus my RVV patch on top) on both the K3 and - at least on the command line - all four of those tests seem to be passing now (tested with 20 iterations) :-)

I do get your timeout error on the slower A100 cores on the K3 though but the exact failure point is not consistent, and also not always at the same line number of the test which does indicate that it's intermittent and the timeouts may need increasing. I've had a quick try with modifying a couple of the timeouts (The 1000 one at line 89) although that gave me another error:

Test failure: 'should not crash when --env-file-if-exists points to a missing file'
Location: test/sequential/test-watch-mode.mjs:306:3
AssertionError [ERR_ASSERTION]: Expected values to be strictly deep-equal:
+ actual - expected

  [
+   "Change detected in '/home/sxa/git/node/test/.tmp.0/34.env'",
+   "Restarting '/home/sxa/git/node/test/.tmp.0/9.js'",
    'ENV: undefined',
    "Completed running '/home/sxa/git/node/test/.tmp.0/9.js'. Waiting for file changes before restarting..."
  ]

    at TestContext.<anonymous> (file:///home/sxa/git/node/test/sequential/test-watch-mode.mjs:318:14)
    at process.processTicksAndRejections (node:internal/process/task_queues:104:5)
    at async Test.run (node:internal/test_runner/test:1409:7)
    at async Promise.all (index 6)
    at async Suite.run (node:internal/test_runner/test:1905:7)
    at async startSubtestAfterBootstrap (node:internal/test_runner/harness:387:3) {
  generatedMessage: true,
  code: 'ERR_ASSERTION',
  actual: [Array],
  expected: [Array],
  operator: 'deepStrictEqual',
  diff: 'simple'
}

Hopefully that won't be too hard to fix. (Although I'll also mention that I'm about to go on holiday!)

EDIT: I also had a quite try with the same commit without my RVV patch and that still passed the other three tests (test-watch-mode.mjs gave an illegal instruction) so it's possible the other problems are now resolved in the codebase, which would be nice, although it looks like they still failed in jenkins.

EDIT2: I tried with the commit that you used (outside jenkins) and got the same failures so they have been fixed in later versions :-)

test: isolate styleText tests from the environment

The V8 15.5 update in nodejs/node#66537 (which went in faster than I expected after 15.2 - nice work!) has a change that breaks my RISC-V builds. The commit causing the problem is v8/v8@16d931d - the compiler has std::hardware_destructive_interference_size defined as 32 where before that commit a value of 64 was hardcoded. It appears that clang-22 on RISC-V sets it to 64 so one option other than reverting/change that commit) is to require that version of clang on RISC-V.

the compiler has std::hardware_destructive_interference_size defined as 32 where before that commit a value of 64 was hardcoded. It appears that clang-22 on RISC-V sets it to 64 so one option other than reverting/change that commit) is to require that version of clang on RISC-V.

For GCC, I proposed to update hardware_destructive_interference_size to 64 for RISC-V in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116662 and the patch for it lands in GCC 15 IIRC.

A quick note on something relevant: There was a refactor of some of the CPU-specific code in V8 15 which landed in Node as part of the V8 15.5 update. This removed src/base/cpu.cc which included the fallback mechanism which meant that when building on glibc prior to 2.39 it would check the CPU fields (effectively /proc/cpuinfo) to determine whether V8 would use things like the vector extensions: https://github.com/v8/v8/blob/0b9dd147a10b9f12751a5ecf1e5f59e88a2a2b7b/src/base/cpu.cc#L1018

In recent weeks that check has been reinstated based on whether the hwprobe syscall succeeds, but it means that the fallback mechanism is not in the current version of Node. I suspect we can backport that at an appropriate time but I'd prefer to get some of the fixes for RVV-related crashes in first.

Navigation Menu

Repository navigation

Tracking issue for problems on RISC-V Linux platform #4099

Build Errors

Bugs

Performance Regressions

Initial test failure list

70 remaining items

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Tracking issue for problems on RISC-V Linux platform #4099

Description

Build Errors

Bugs

Performance Regressions

Activity

sxa commented on Dec 5, 2025

kxxt commented on Dec 6, 2025

sxa commented on Dec 6, 2025

sxa commented on Dec 12, 2025

Initial test failure list

kxxt commented on Dec 13, 2025

sxa commented on Dec 13, 2025

kxxt commented on Dec 13, 2025

sxa commented on Dec 13, 2025

sxa commented on Dec 17, 2025

kxxt commented on Dec 18, 2025

sxa commented on Dec 18, 2025

kxxt commented on Dec 18, 2025

RevySR commented on Feb 4, 2026

sxa commented on Feb 19, 2026

70 remaining items

Polaris-911 commented on Aug 17, 2026

gounthar commented on Aug 17, 2026

sxa commented on Aug 18, 2026

sxa commented on Aug 26, 2026

sxa commented on Aug 26, 2026

kxxt commented on Aug 29, 2026

sxa commented on Sep 1, 2026

kxxt commented on Sep 1, 2026

sxa commented on Sep 3, 2026

sxa commented on Oct 8, 2026

kxxt commented on Oct 8, 2026

sxa commented on Oct 9, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions