[EH] Add --experimental-new-eh option #6267
Conversation
This adds `--experimental-new-eh` option to `wasm-opt`. The difference between this and `--translate-to-new-eh` is, `--translate-to-new-eh` just runs `TranslateToNewEH` pass, while `--experimental-new-eh` attaches `TranslateToNewEH` pass at the end of the whole optimization pipeline. So if no other passes or optimization options (`-On`) are specified, it is equivalent to `--translate-to-new-eh`. If other optimization passes are specified, it runs them and at the end run the translator to ensure the new EH instructions are emitted. The reason we are doing this this way is that the optimization pipeline as a whole does not support the new EH instruction yet, but we would like to provide an option to emit a reasonably OK code with the new EH instructions. This also means when the optimization level > 3, it will also run the StackIR optimization after the translation. Not sure how to test the output of this option, given that there is not much point in testing the default optimization passes, and it is also not clear how to print the stack IR if the stack ir generation and optimization runs as a part of the pipeline and not the explicit command line options.
tlively
left a comment
tlively
left a comment
There was a problem hiding this comment.
What is the long-term plan for this flag? Will it eventually mean that wasm-opt will run the translation pass at the beginning of the pipeline instead? Will this flag be eventually removed?
Yes, after the optimization pipeline is ready, the translator will run at the beginning of the pipeline to handle the old binaries in the wild.
I think eventually it should. |
| addIfNoDWARFIssues("directize"); | ||
|
|
||
| // Translate to use the new EH instructions if requested. | ||
| if (wasm->features.hasExceptionHandling() && options.experimentalNewEH) { |
There was a problem hiding this comment.
Doing this here means it will run twice if the user does wasm-opt -O3 -O3. Perhaps it would be better to insert this pass in src/tools/wasm-opt.cpp? Then we can add it right at the end.
There was a problem hiding this comment.
A downside to my idea is that we'd need to also add logic there for adding the stackIR passes right after it, copying the 4 lines underneath us. But I think it might still be better.
There was a problem hiding this comment.
Then if a user gives wasm-opt --generate-stack-ir --optimize-stack-ir --experimental-new-eh (or -O3 --experimental-new-eh, which I guess also runs the stack IR generation and optimization), aren't we gonna generate and optimize the stack IR, throw it away, run the translator, and re-generate and re-optimize the stack IR?
There was a problem hiding this comment.
Yes. We also do that with -O3 -O3, as the end of each of them generates stack IR and optimizes it. A little wasteful, but not significant (making the pass pipeline smarter to avoid this overhead likely isn't worth it).
There was a problem hiding this comment.
Is -O3 -O3 common enough? I thought that would only happen if someone carelessly adds it twice, not remembering they have already done it, but I might be wrong.. whereas I thought -O3 --experimental-new-eh would be pretty common, because it's basically just -O3 with the intention of using the new EH, which now runs stack IR generation/optimization twice.
There was a problem hiding this comment.
Chatted w/ @tlively and I learned people run -O3 not only twice but many times... Yeah, it'd be better to run it only once at the end of wasm-opt, even though it means running Stack IR twice.
There was a problem hiding this comment.
Closing this in favor of #6270, which adds the option directly to wasm-opt.cpp.
This adds `--experimental-new-eh` option to `wasm-opt`. The difference between this and `--translate-to-new-eh` is, `--translate-to-new-eh` just runs `TranslateToNewEH` pass, while `--experimental-new-eh` attaches `TranslateToNewEH` pass at the end of the whole optimization pipeline. So if no other passes or optimization options (`-On`) are specified, it is equivalent to `--translate-to-new-eh`. If other optimization passes are specified, it runs them and at the end run the translator to ensure the new EH instructions are emitted. The reason we are doing this this way is that the optimization pipeline as a whole does not support the new EH instruction yet, but we would like to provide an option to emit a reasonably OK code with the new EH instructions. This also means when the optimization level > 3, it will also run the StackIR + local2stack optimization after the translation. Not sure how to test the output of this option, given that there is not much point in testing the default optimization passes, and it is also not clear how to print the stack IR if the stack ir generation and optimization runs as a part of the pipeline and not the explicit command line options. This is created in favor of WebAssembly#6267, which added the option to `optimization-options.h`. This had a problem of running the translator multiple times when `-On` was given multiple times in the command line, which I learned was rather a common usage. This adds the option directly to `wasm-opt.cpp`, which avoids the problem. With this, it is still possible to create and optimize Stack IR unnecessarily, but that feels a better alternative.
This adds `--experimental-new-eh` option to `wasm-opt`. The difference between this and `--translate-to-new-eh` is, `--translate-to-new-eh` just runs `TranslateToNewEH` pass, while `--experimental-new-eh` attaches `TranslateToNewEH` pass at the end of the whole optimization pipeline. So if no other passes or optimization options (`-On`) are specified, it is equivalent to `--translate-to-new-eh`. If other optimization passes are specified, it runs them and at the end run the translator to ensure the new EH instructions are emitted. The reason we are doing this this way is that the optimization pipeline as a whole does not support the new EH instruction yet, but we would like to provide an option to emit a reasonably OK code with the new EH instructions. This also means when the optimization level > 3, it will also run the StackIR + local2stack optimization after the translation. Not sure how to test the output of this option, given that there is not much point in testing the default optimization passes, and it is also not clear how to print the stack IR if the stack ir generation and optimization runs as a part of the pipeline and not the explicit command line options. This is created in favor of #6267, which added the option to `optimization-options.h`. It had a problem of running the translator multiple times when `-On` was given multiple times in the command line, which I learned was rather a common usage. This adds the option directly to `wasm-opt.cpp`, which avoids the problem. With this, it is still possible to create and optimize Stack IR unnecessarily, but that feels a better alternative.
This adds `--experimental-new-eh` option to `wasm-opt`. The difference between this and `--translate-to-new-eh` is, `--translate-to-new-eh` just runs `TranslateToNewEH` pass, while `--experimental-new-eh` attaches `TranslateToNewEH` pass at the end of the whole optimization pipeline. So if no other passes or optimization options (`-On`) are specified, it is equivalent to `--translate-to-new-eh`. If other optimization passes are specified, it runs them and at the end run the translator to ensure the new EH instructions are emitted. The reason we are doing this this way is that the optimization pipeline as a whole does not support the new EH instruction yet, but we would like to provide an option to emit a reasonably OK code with the new EH instructions. This also means when the optimization level > 3, it will also run the StackIR + local2stack optimization after the translation. Not sure how to test the output of this option, given that there is not much point in testing the default optimization passes, and it is also not clear how to print the stack IR if the stack ir generation and optimization runs as a part of the pipeline and not the explicit command line options. This is created in favor of WebAssembly#6267, which added the option to `optimization-options.h`. It had a problem of running the translator multiple times when `-On` was given multiple times in the command line, which I learned was rather a common usage. This adds the option directly to `wasm-opt.cpp`, which avoids the problem. With this, it is still possible to create and optimize Stack IR unnecessarily, but that feels a better alternative.
3 participants
This adds
--experimental-new-ehoption towasm-opt. The difference between this and--translate-to-new-ehis,--translate-to-new-ehjust runsTranslateToNewEHpass, while--experimental-new-ehattachesTranslateToNewEHpass at the end of the whole optimization pipeline. So if no other passes or optimization options (-On) are specified, it is equivalent to--translate-to-new-eh. If other optimization passes are specified, it runs them and at the end run the translator to ensure the new EH instructions are emitted. The reason we are doing this this way is that the optimization pipeline as a whole does not support the new EH instruction yet, but we would like to provide an option to emit a reasonably OK code with the new EH instructions.This also means when the optimization level > 3, it will also run the StackIR optimization after the translation.
Not sure how to test the output of this option, given that there is not much point in testing the default optimization passes, and it is also not clear how to print the stack IR if the stack ir generation and optimization runs as a part of the pipeline and not the explicit command line options.