On Thu, Oct 26, 2023 at 2:40 PM Liu Yu <[email protected]> wrote: > Thanks for reply ! > > This seems to have no side effects for real-world websites. > And I don't know OCaml, so I will open an issue in v8. >
This will need to be fixed in the spec repository, so please file an issue there: https://github.com/WebAssembly/spec/issues > > 在2023年10月25日星期三 UTC+8 18:32:18<Clemens Backes> 写道： > >> If you look at the JS test file, it contains lots of these lines: >> >> let $1 = instance(...); >> assert...($1, ...); >> let $2 = instance(...); >> assert...($2, ...); >> >> >> So the instances (and also the contained memory objects) never go out of >> scope. >> These files are generated from the spec tests (written in wast format) by >> the reference interpreter. This translation is implemented in the js.ml >> file >> <https://github.com/WebAssembly/spec/blob/main/interpreter/script/js.ml#L599> >> . >> If you know OCaml you could upload a PR to either reset the instances >> when they are not needed any more ($1 = undefined), or to scope the >> instances. >> If you don't know OCaml, you could also open an issue in the spec >> repository and see if someone is willing to fix this. >> >> -Clemens >> >> >> >> On Wed, Oct 25, 2023 at 10:20 AM 'Andreas Haas' via v8-dev < >> [email protected]> wrote: >> >>> Hi, >>> >>> The test allocates 33 wasm memories, 10GB each. It can be that in such a >>> situation garbage collection of wasm memories does not work perfectly, and >>> eventually it cannot find memory anymore to allocate. The tests don't test >>> if you are >>> able to allocate a large number of wasm instances and wasm memories, >>> this is just a side effect.On real-world websites there are typically not >>> so many wasm memories, and they are used more and for longer so that the GC >>> can keep up with it. >>> >>> Cheers, Andreas >>> >>> On Wed, Oct 25, 2023 at 3:53 AM Liu Yu <[email protected]> wrote: >>> >>>> Hi all, >>>> >>>> When porting the code pointer sandbox, I set the sandbox size on >>>> loong64 to 256GB because some loongarch machines are configured with only >>>> 40-bit virtual address space for user space. >>>> >>>> However, after the port, I found that some wasm tests would fail, such >>>> as >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> *=== wasm-spec-tests/memory_copy === >>>> test/wasm-spec-tests/tests/memory_copy.js:74: RangeError: >>>> WebAssembly.Instance(): Out of memory: Cannot allocate Wasm memory for new >>>> instance return new WebAssembly.Instance(module(bytes), imports); >>>> ^RangeError: WebAssembly.Instance(): Out of memory: Cannot allocate Wasm >>>> memory for new instance at instance >>>> (test/wasm-spec-tests/tests/memory_copy.js:74:10) at >>>> test/wasm-spec-tests/tests/memory_copy.js:12772:11Command: >>>> out_dir/loong64.release/d8 --test test/wasm-spec-tests/tests/memory_copy.js >>>> --random-seed=967154330 --nohard-abort --testing-d8-test-runner--- FAILED >>>> ---* >>>> >>>> I changed the sandbox size to 256GB on arm64 linux, and tests will >>>> also go wrong. It seems that the error is caused by insufficient memory in >>>> the sandbox. >>>> >>>> Tests like wasm-spec-tests/memory_copy and wasm-spec-tests/memory_init >>>> include many small wasm casts; Is the number of wasm instance also part of >>>> the test, or does it only need to be correct for each small instance? On >>>> real-world >>>> websites, will there be an error due to insufficient memory because >>>> the sandbox is too small? >>>> >>>> -- >>>> -- >>>> v8-dev mailing list >>>> [email protected] >>>> http://groups.google.com/group/v8-dev >>>> --- >>>> You received this message because you are subscribed to the Google >>>> Groups "v8-dev" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to [email protected]. >>>> To view this discussion on the web visit >>>> https://groups.google.com/d/msgid/v8-dev/e2051aee-c1ee-46a3-97a2-8442984cc384n%40googlegroups.com >>>> <https://groups.google.com/d/msgid/v8-dev/e2051aee-c1ee-46a3-97a2-8442984cc384n%40googlegroups.com?utm_medium=email&utm_source=footer> >>>> . >>>> >>> >>> >>> -- >>> >>> Andreas Haas >>> >>> Software Engineer >>> >>> [email protected] >>> >>> >>> Google Germany GmbH >>> >>> Erika-Mann-Straße 33 >>> >>> 80636 München >>> >>> >>> Geschäftsführer: Paul Manicle, Liana Sebastian >>> >>> Registergericht und -nummer: Hamburg, HRB 86891 >>> >>> Sitz der Gesellschaft: Hamburg >>> >>> >>> Diese E-Mail ist vertraulich. Falls sie diese fälschlicherweise erhalten >>> haben sollten, leiten Sie diese bitte nicht an jemand anderes weiter, >>> löschen Sie alle Kopien und Anhänge davon und lassen Sie mich bitte wissen, >>> dass die E-Mail an die falsche Person gesendet wurde. >>> >>> >>> >>> This e-mail is confidential. If you received this communication by >>> mistake, please don't forward it to anyone else, please erase all copies >>> and attachments, and please let me know that it has gone to the wrong >>> person. >>> >>> -- >>> -- >>> v8-dev mailing list >>> [email protected] >>> http://groups.google.com/group/v8-dev >>> --- >>> You received this message because you are subscribed to the Google >>> Groups "v8-dev" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> >> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/v8-dev/CAELSTve1VR7-51nj2DOF27UOGRB_6G5Dhp4dJhHU6UW8k-3Okw%40mail.gmail.com >>> <https://groups.google.com/d/msgid/v8-dev/CAELSTve1VR7-51nj2DOF27UOGRB_6G5Dhp4dJhHU6UW8k-3Okw%40mail.gmail.com?utm_medium=email&utm_source=footer> >>> . >>> >> >> >> -- >> >> Clemens Backes >> >> Software Engineer >> >> [email protected] >> >> Google Germany GmbH >> >> Erika-Mann-Straße 33 >> >> 80636 München >> >> Geschäftsführer: Paul Manicle, Liana Sebastian >> >> Registergericht und -nummer: Hamburg, HRB 86891 >> >> Sitz der Gesellschaft: Hamburg >> >> Diese E-Mail ist vertraulich. Falls sie diese fälschlicherweise erhalten >> haben sollten, leiten Sie diese bitte nicht an jemand anderes weiter, >> löschen Sie alle Kopien und Anhänge davon und lassen Sie mich bitte wissen, >> dass die E-Mail an die falsche Person gesendet wurde. >> >> >> This e-mail is confidential. If you received this communication by >> mistake, please don't forward it to anyone else, please erase all copies >> and attachments, and please let me know that it has gone to the wrong >> person. >> >> >> -- > -- > v8-dev mailing list > [email protected] > http://groups.google.com/group/v8-dev > --- > You received this message because you are subscribed to the Google Groups > "v8-dev" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/v8-dev/8102bd81-d951-4f3c-a050-dcf690531b7bn%40googlegroups.com > <https://groups.google.com/d/msgid/v8-dev/8102bd81-d951-4f3c-a050-dcf690531b7bn%40googlegroups.com?utm_medium=email&utm_source=footer> > . > -- Clemens Backes Software Engineer [email protected] Google Germany GmbH Erika-Mann-Straße 33 80636 München Geschäftsführer: Paul Manicle, Liana Sebastian Registergericht und -nummer: Hamburg, HRB 86891 Sitz der Gesellschaft: Hamburg Diese E-Mail ist vertraulich. Falls sie diese fälschlicherweise erhalten haben sollten, leiten Sie diese bitte nicht an jemand anderes weiter, löschen Sie alle Kopien und Anhänge davon und lassen Sie mich bitte wissen, dass die E-Mail an die falsche Person gesendet wurde. This e-mail is confidential. If you received this communication by mistake, please don't forward it to anyone else, please erase all copies and attachments, and please let me know that it has gone to the wrong person. -- -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/v8-dev/CAGO%3DqhDp8F3KwA%2Bae_aV4A4BadGzzP%3DsVsC5g_26ZEueLaJJEw%40mail.gmail.com.
