-
Notifications
You must be signed in to change notification settings - Fork 3.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[compiler-v2] Use livevar analysis to compute copy/move instructions #9361
Conversation
9a21aeb
to
cdc771d
Compare
This connects the live-variable analysis which is already present in the Move prover to the v2 bytecode pipeline. This information is then used in the file-format generator to make better decisions when to move or copy values. Livevar is a standard compiler construction data flow analysis. It computes the set of variables which are alive (being used in subsequent reachable code) before and after each program point. The implementation from the prover uses our dataflow analysis framework and is [here](https://github.com/aptos-labs/aptos-core/blob/206f529c0c9d8488e27d2e50297178f0caf429a5/third_party/move/move-prover/bytecode/src/livevar_analysis.rs#L381). In this PR we create the first processing step in the bytecode pipeline with the new module `pipeline/livevar_analysis_step.rs` which forwards logical work to the existing analysis.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Also: Not sure what to do with this, but the BTreeMap initialization in livevar_analysis.rs (outside of this review) is pretty pessimal, as it is sequential and will rebalance the tree on every insert. Apparently if you create a BTreeMap directly from an iterator it saves the elements in a vector, sorts it (hopefully efficiently if it's already sorted), then creates the BTree more efficiently. I don't have numbers, but maybe be careful about anti-performance patterns like that.
third_party/move/move-compiler-v2/src/file_format_generator/function_generator.rs
Outdated
Show resolved
Hide resolved
third_party/move/move-compiler-v2/src/file_format_generator/function_generator.rs
Outdated
Show resolved
Hide resolved
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, thanks!
while self.stack.len() > top { | ||
let temp = self.stack.pop().unwrap(); | ||
let local = self.temp_to_local(ctx, temp); | ||
self.emit(FF::Bytecode::StLoc(local)); | ||
if ctx.is_alive_before(temp) || ctx.is_alive_after(temp) || self.pinned.contains(&temp) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why are you testing is_alive_before here?
Also, It seems that you only need to save a pinned local if it has been changed on the stack. But abstract_push_result() already flushes a pinned temp.
We may need some targeted test cases that make this situation easier to see.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
When abstract_flush_stack is called, we are interested in the pre state. That is because the flushing will be essentially inserted before the next instruction. This is different to when we use is_alive_after
above: here we are interested in what is alive after this given program point. FWIW this has been discovered by testing; before I would use the after-set.
Regards your 2nd point: There are temps on the stack which may in fact have never been stored into any local (as this is a stack machine -- in the ideal case we need to store as little as possible). Those need to be saved here. You are right we could optimize the case if the temp on the stack is loaded from a non-pinned local. I suggest we keep this optimization for later. Not really need the most optimal code here -- doing the liveness here so we can do meaningful ability and borrow analysis, which require the right placement of copy and move as pre-requisite.
Testing: please note this is quite extensively tested via the transactional tests here, which you do not see as part of this change. In fact I debugged quite a bit based on those tests.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Regards the abstract stack, there is now also a section in the design doc: https://www.notion.so/aptoslabs/Aptos-Move-Compiler-V2-High-Level-Design-6ab5e5ffcd574c0693795e6fdefc1331?pvs=4#fc7f5dc9114a4a6097eeb75cdc2d34ff
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think you should generate tight code based on the information you have available here, rather than being sloppy and assuming it will be fixed later. We have the liveness results, let's use them effectively.
I'm afraid that abstract_flush_stack() is saving temps from the stack into locals in too many cases. It might help to add comments for the pre/post conditions for a few of these shared operations, since the state of the function generator is a bit inconsistent as it is called in a few places at different points in generating code for a given input bytecode.
The function is defined here:
aptos-core/third_party/move/move-compiler-v2/src/file_format_generator/function_generator.rs
Line 672 in 47803cd
fn abstract_flush_stack(&mut self, ctx: &BytecodeContext, top: usize) { |
and the overly-generous test is here:
aptos-core/third_party/move/move-compiler-v2/src/file_format_generator/function_generator.rs
Line 676 in 47803cd
if ctx.is_alive_before(temp) || ctx.is_alive_after(temp) || self.pinned.contains(&temp) |
(1) is_alive_before/after: See function_generator.rs:152-156 (your branch) where we see:
self.gen_bytecode(&bytecode_ctx, &bytecode[i], Some(next_bc));
if !bc.is_branch() && matches!(next_bc, Bytecode::Label(..)) {
// At block boundaries without a preceding branch, need to flush stack
// TODO: to avoid this, we should use the CFG for code generation.
self.abstract_flush_stack(&bytecode_ctx, 0);
So in this particular case, after the bytecode is generated, we carefully store temps including those which are alive before the generated bytecode (even if not live after the bytecode). In other cases (e.g., in abstract_push_args()) the flush is presumably happening before the current bytecode. There's also a call in abstract_push_result() which also seems to happen after the bytecode is generated.
The current code test on line 676 (if ctx.is_alive_before(temp) || ctx.is_alive_after(temp) || self.pinned.contains(&temp)
) is conservative, and safe, in that it saves a superset of the required temps that works for both cases. But it would be better to differentiate the 2 cases (before and after the bytecode) so that spurious temp saves are not done.
(2) is_pinned: The current code seems to ignore the lifetime of a pinned variable and consider it pinned for the rest of the method. Fine, though it would be good to mention it somewhere. Even with this conservative assumption, a pinned temp only needs to be saved if it has not already been copied to its home local.
There are only 2 locations where stack.push() is called. (i) In abstract_push_args(), the temp is saved to a local before being pushed on the stack. This case could presumably happen when pushing args for a Borrow operation, in which case the value would not yet have been saved, and should probably be saved to its home location immediately rather than waiting for abstract_flush_stack() to do it.
(ii) In abstract_push_result(), a temp is pushed, and if it is pinned, then it is saved right away, so there is no need to copy it later when abstract_flush_stack() is called and pops that temp.
So the is_pinned case is redundant in case (ii), and in case (i) it could be saved immediately.
(3) I don't see bytecode output for the transactional tests, so they aren't as useful to understand what is happening.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think you should generate tight code based on the information you have available here, rather than being sloppy and assuming it will be fixed later. We have the liveness results, let's use them effectively.
Thanks for calling my work sloppy :-) But I disagree. You and I have discussed multiple times that the goal is to not implement optimizations at this phase. Its also like this in the engineering plan. Optimizations are in milestone 4. There is a lot more optimizations you can imagine than this single one you are suggesting.
Technically, keeping track of what value on the stack is identical to a value in a local, may require us to do additional data flow analysis. Not 100% sure but perhaps in different paths, this information can be different. This would make this PR significantly larger and require a larger rewrite.
I'm afraid that abstract_flush_stack() is saving temps from the stack into locals in too many cases. It might help to add comments for the pre/post conditions for a few of these shared operations, since the state of the function generator is a bit inconsistent as it is called in a few places at different points in generating code for a given input bytecode.
I adopted this and introduced two different functions: flush_stack_after
and flush_stack_before
.
The function is defined here:
aptos-core/third_party/move/move-compiler-v2/src/file_format_generator/function_generator.rs
Line 672 in 47803cd
fn abstract_flush_stack(&mut self, ctx: &BytecodeContext, top: usize) { .
and the overly-generous test is here:
aptos-core/third_party/move/move-compiler-v2/src/file_format_generator/function_generator.rs
Line 676 in 47803cd
if ctx.is_alive_before(temp) || ctx.is_alive_after(temp) || self.pinned.contains(&temp) (1) is_alive_before/after: See function_generator.rs:152-156 (your branch) where we see:
self.gen_bytecode(&bytecode_ctx, &bytecode[i], Some(next_bc)); if !bc.is_branch() && matches!(next_bc, Bytecode::Label(..)) { // At block boundaries without a preceding branch, need to flush stack // TODO: to avoid this, we should use the CFG for code generation. self.abstract_flush_stack(&bytecode_ctx, 0);
So in this particular case, after the bytecode is generated, we carefully store temps including those which are alive before the generated bytecode (even if not live after the bytecode). In other cases (e.g., in abstract_push_args()) the flush is presumably happening before the current bytecode. There's also a call in abstract_push_result() which also seems to happen after the bytecode is generated.
The current code test on line 676 (
if ctx.is_alive_before(temp) || ctx.is_alive_after(temp) || self.pinned.contains(&temp)
) is conservative, and safe, in that it saves a superset of the required temps that works for both cases. But it would be better to differentiate the 2 cases (before and after the bytecode) so that spurious temp saves are not done.(2) is_pinned: The current code seems to ignore the lifetime of a pinned variable and consider it pinned for the rest of the method. Fine, though it would be good to mention it somewhere. Even with this conservative assumption, a pinned temp only needs to be saved if it has not already been copied to its home local.
One needs borrow analysis to reason about liveness of pinned variables. The borrow analysis maintains a borrow graph which has an edge from the pinned variable to the temp which contains the reference if it is still alive. But this happens in a future analysis which we have not written yet.
There are only 2 locations where stack.push() is called. (i) In abstract_push_args(), the temp is saved to a local before being pushed on the stack. This case could presumably happen when pushing args for a Borrow operation, in which case the value would not yet have been saved, and should probably be saved to its home location immediately rather than waiting for abstract_flush_stack() to do it.
(ii) In abstract_push_result(), a temp is pushed, and if it is pinned, then it is saved right away, so there is no need to copy it later when abstract_flush_stack() is called and pops that temp.
So the is_pinned case is redundant in case (ii), and in case (i) it could be saved immediately.
Can you please check the link to the design doc I provided that we are on the same page what the actual goals are here. For (i) the goal is to not save to local and postpone or avoid the need for that. That's the major motivation for the whole logic here. Regards the (ii) case, this is not redundant because of (i).
(3) I don't see bytecode output for the transactional tests, so they aren't as useful to understand what is happening.
These are transactional tests which compare execution results of the both compilers. There was a PR out with this which multiple folks reviewed. The other tests which show bytecode are also there, and you should see the results.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Btw, since the current code is safe, I'd be ok with accepting this PR if we add an issue to go back and clarify the "stack temp needs to be stored in local" property in a latter PR.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ok.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks ok.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
✅ Forge suite
|
✅ Forge suite
|
✅ Forge suite
|
…aptos-labs#9361) * [compiler-v2] Use livevar analysis to optimize file format generation This connects the live-variable analysis which is already present in the Move prover to the v2 bytecode pipeline. This information is then used in the file-format generator to make better decisions when to move or copy values. Livevar is a standard compiler construction data flow analysis. It computes the set of variables which are alive (being used in subsequent reachable code) before and after each program point. The implementation from the prover uses our dataflow analysis framework and is [here](https://github.com/aptos-labs/aptos-core/blob/206f529c0c9d8488e27d2e50297178f0caf429a5/third_party/move/move-prover/bytecode/src/livevar_analysis.rs#L381). In this PR we create the first processing step in the bytecode pipeline with the new module `pipeline/livevar_analysis_step.rs` which forwards logical work to the existing analysis. * Addressing reviewer comments.
…#9361) * [compiler-v2] Use livevar analysis to optimize file format generation This connects the live-variable analysis which is already present in the Move prover to the v2 bytecode pipeline. This information is then used in the file-format generator to make better decisions when to move or copy values. Livevar is a standard compiler construction data flow analysis. It computes the set of variables which are alive (being used in subsequent reachable code) before and after each program point. The implementation from the prover uses our dataflow analysis framework and is [here](https://github.com/aptos-labs/aptos-core/blob/206f529c0c9d8488e27d2e50297178f0caf429a5/third_party/move/move-prover/bytecode/src/livevar_analysis.rs#L381). In this PR we create the first processing step in the bytecode pipeline with the new module `pipeline/livevar_analysis_step.rs` which forwards logical work to the existing analysis. * Addressing reviewer comments.
* [storage] support sharding in fast_sync handle sharded kv batches handle sharded merkel tree save progress * [TS SDK] Support network and API endpoints on `AptosConfig` (#9549) * support network and api urls on AptosConfig * devnet as default network * Remove tutorial links from descriptions Tutorials are linked on tutorial title and again on the tutorial description just next to the title. Small nitpick, but is slightly confusing when you just arrive to the dev docs, since you expect that second link to be something different. * add profiling crate with cpu and memory profiling fix comments and warnings fix comments resolve comments add crate to cargo.toml add additional functions for different cases of profiling fix cargo.toml * remove unnecessary dependency * fix lint errors * fix lint errors * fix lint error in cargo.toml * [testsuite][pangu][pte] Creating a Pangu Command to Create a Transaction Emitter (#9495) * [testsuite][pangu][pte] Initial work for the transaction emitter. * [testsuite][pangu][pte] Finished the transaction emitter comand. * Update testsuite/test_framework/kubernetes.py good catch! Co-authored-by: Balaji Arun <balaji@aptoslabs.com> * [testsuite][pangu][pte] Addressed Balaji's comments. * [testsuite][pangu][pte] Hotfix for Forge compatibility. * [testsuite][pangu][pte] Another hotfix. * [testsuite][pangu][pte] Added support for dry runs, and fixed a bug in get testnet * [testsuite][pangu][pte] Minor changes. * Update README.md --------- Co-authored-by: Olsen Budanur <olsenbudanur@Olsens-MBP.localdomain> Co-authored-by: Balaji Arun <balaji@aptoslabs.com> * replay-verify: log skipped version when it happens * [dashboards] sync grafana dashboards (#9440) Co-authored-by: rustielin <rustielin@users.noreply.github.com> * [forge stable] Fix HAProxy test - increase latency limits (#9575) * [indexer grpc] reduce the data service streaming channel size (#9590) * [refactoring] Remove unused genesis check from StateView (#9589) `is_genesis()` was defined in `StateView` but never used apart from a single test with `MockVM`. But the test set it to false always... Removing it to have better view interfaces. * [TS SDK] Use `account_transactions` and `account_transaction_aggregate` queries (#9403) * improve indexer client to support tokenv2 sorting and new queries * use account_transactions query * Optional argument on is not positional * [Dev Docs] Add small note about target-version in backup docs. * swich to gcp * add semgrep for github workflows (#9522) add semgrep for GitHub workflows * [GHA] Run smoke and performance tests when appropriate. * [Spec] Ensures of transaction_validation (#9461) * hp_trans_valid * hp_trans_validation * pre-pr * fix trim --------- Co-authored-by: chan-bing <zzywullr@gmail.com> * [GHA] Small tweaks to jobs. * [TS SDK] filter amount > 0 on getTokenOwnersData and include all query fields (#9593) * filter amount > 0 on getTokenOwnersData * export all query fields * update changelog * [Python SDK] Add a token_transfer function to the python SDK (#9422) * Adding transfer_token function to aptos_token_client.py and adding documentation tags for the new your_first_nft tutorial. * Fixed strange formatting * Created a transfer_object function in RestClient class and altered the transfer_token function in the AptosTokenClient to just call transfer_object. Left it in for ease of use * [deps] Update clap dependency to be consistent across crates (#9605) * Fix node compatibility test CI (#9609) * [gas] fix abstract gas cost for storage (#9501) * [forge] Add latency for different workloads (#9415) * remove redundant test (#9613) * [dag] Handle certified node in DAG Driver (#9312) * [dag] Handle certified node in dag driver * [dag] Trigger new round on handler start * [dag] Use only the aptos_time_service crate * [NFT Metadata Crawler] Fix error for defaulting to original raw_image_uri (#9611) * Fix error for defaulting to original raw_image_uri * move db commit outside * [ts-sdk example] Adding an example for rotation offer capability and signer offer capability with signed structs (#9425) * Adding the offer capability example. Uses signed structs in typescript. * remove unnecessary aptosClient from getAccount call Co-authored-by: Maayan <maayan@aptoslabs.com> * Fixing network URL unnecessary code and potential unalignment between network/faucet url * Shortening # hyphens in output string * Moving the chainId to the bottom of the struct list since it's potentially undefined, and explaining in the sign struct function that the proof bytes must be in that specific order. * Cleaning up the code and making it more readable * Formatting * Use a much cleaner and reusable serializable class instead of a struct --------- Co-authored-by: Maayan <maayan@aptoslabs.com> * bump version to 1.18.0 (#9607) * [compiler-v2] Use livevar analysis to optimize file format generation (#9361) * [compiler-v2] Use livevar analysis to optimize file format generation This connects the live-variable analysis which is already present in the Move prover to the v2 bytecode pipeline. This information is then used in the file-format generator to make better decisions when to move or copy values. Livevar is a standard compiler construction data flow analysis. It computes the set of variables which are alive (being used in subsequent reachable code) before and after each program point. The implementation from the prover uses our dataflow analysis framework and is [here](https://github.com/aptos-labs/aptos-core/blob/206f529c0c9d8488e27d2e50297178f0caf429a5/third_party/move/move-prover/bytecode/src/livevar_analysis.rs#L381). In this PR we create the first processing step in the bytecode pipeline with the new module `pipeline/livevar_analysis_step.rs` which forwards logical work to the existing analysis. * Addressing reviewer comments. * [storage] update resource requriement * [TS SDK v2] Add Hex and HexInput types (#9595) Co-authored-by: maayan <maayansavir@gmail.com> * [cleanup] Remove foreign contracts (#9623) * [Spec] Ensures of reconfiguration.move (#9383) * init * hp * hp reconfig * init * fix comment * fix comment --------- Co-authored-by: chan-bing <zzywullr@gmail.com> * Revert "Revert "[NFT Metadata Crawler] Dockerize Parser (#9541)"" (#9622) This reverts commit aa81b95. * Clearer comments in `resource_account.move` (#9555) * Explicitize that the txn_seq_num param on the epilogue is unused With this we can go through all historical txns to make sure. * [dag] Integrate Order Rule with Dag Driver (#9452) * [gas] convert output to InternalGas cost (#9603) * [move-ir-compiler] add Nop token support (#9599) * [dag] Integrate Dag Fetcher with Dag Driver (#9453) * [dag] Integrate Dag Fetcher with Dag Driver * [dag] Handle fetch response in dag handler * [TS SDK v2] Add AccountAddress (#9564) * [CLI] Disallow "rotating" to same private key (#9546) * trivial: fix gas meter mutability * Remove unused deps for CLI (#9531) * typo in functions.md * typo spec-lang.md typo * Update package-upgrades.md space added * Update developer-docs-site/docs/move/book/package-upgrades.md * Update unit-testing.md typo * [GHA] Only run all perf tests on schedule. * [TS SDK example] Add a working example of converting a MultiEd25519 account to a MultiSig account (#9516) * An example of converting a MultiEd25519 account to a MultiSig account Cleaned up the file even more, added assertions and print statements at the end Removing incorrect addresses Formatting and clarifying the steps. Changing network to DEVNET from local Removing confusing comment from prior version and adding comment about making sure secondsTilExpiration is set accordingly Fixing file name Fixing variable name Making header comments less long * Adding the ability to add metadata to the multisig account, since we were using empty arrays before. Asserted that the values on-chain match the input values at the end * Using a much cleaner and reusable serializable class for the signed message * Removing the unnecessary multisig as sender option * Update e2e flow description at top of file * Resolving rebase conflict and simplifying the signature construction * Clarify the process of creating a MultiEd25519 account and that public keys are different from addresses * Improve clarity of comments on token.move * [NFT Metadata Crawler] Fix raw image uri parse error (#9628) * Fix raw_image_uri parse bug * upsert * update compatibility test base * [Dev Docs] Add note about indexer bootstrap support. * Update developer-docs-site/docs/nodes/indexer-fullnode.md * [dashboards] sync grafana dashboards * remove payer from StateValueMetadata * Track slot allocation fee on state metadata * enable state value metadata tracking for devnet * [compiler v2] Fix bugs around compilation of vector code (#9636) * [compiler v2] Fix bugs around compilation of vector code This fixes a few bugs and closes #9629 - Usage of precompiled stdlib in transactional tests: v2 compiler currently (and probably never) supports precompiled modules, working around this - Stack-based bytecode generator created wrong target on generic function calls - Needed to make implicit conversion from `&mut` to `&` explicit in generated bytecode via Freeze operation Added some additional tests while debugging this. * Adding a new option `--print-bytecode` which can be provided to the `//# publish` and `//# run` transactional test runner command. This is the applied (for now) to the `sorter` test case only. Also introduced logic to map well-known vector functions to the associated builtin opcodes. * [TS SDK v2] Reject invalid SHORT form for special addresses in fromString (#9640) * [TS SDK v2] Reject invalid SHORT form for special addresses in fromString * Update account_address.ts * Update account_address.ts * [TS SDK v2] Add equality methods for Hex and AccountAddress (#9641) * Handle empty network address in CLI (#9576) * [CLI][e2e] Tests create-resource-account and derive-resource-account-address (#8757) * CLI e2e for resource account * add account list to test resource account * [prover] lock version dependency for z3 and boogie (#8718) (#9524) Co-authored-by: Aalok Thakkar <aalok@Aaloks-MacBook-Pro.local> * Update workflows (#9650) * Update semgrep.yaml to also run daily * update semgrep rule * fix workflows * Update .github/workflows/semgrep.yaml Co-authored-by: Balaji Arun <balaji@aptoslabs.com> --------- Co-authored-by: Balaji Arun <balaji@aptoslabs.com> * [dag] bootstrap logic (#9455) * add guard field to struct * temp * add executor benchmark profiling * small fixes * lint fix * lint fix * lint fix * lint fix * lint fix * update cargo.lock * [Spec] Ensures of resources.move (#9382) * init * fix md * fix comment * [Spec] Ensures of transaction_fee (#9460) * hp trans_fee * pre-pr * add the condition that proposer == @vm_reserved * fix comment --------- Co-authored-by: chan-bing <zzywullr@gmail.com> * [TS SDK v2] Remove redundant code in AccountAddress.fromString (#9662) * [docs] Rename Aptos Move CLI -> Aptos CLI (#9655) * Thoroughly fix Issue 8875: Error out in compilation when encounter an unknown attribute (#9229) Fix issue-8875 thoroughly: warning for unknown attributes, add --skip-attribute-check flag. Omit warning on aptos_std library files, to avoid warning on some existing code. A few tangential fixes: * [third party/.../test, compiler] Fix Move.toml files to point to local files instead of network to avoid tests using old stdlib. Oddly, this requires fixes to capitalization of std. Fix the resolution warning to make it clear that std needs to be uncapitalized. * [compiler] Hack to move-compiler to avoid warning about unknown attributes in aptos_std=0x1. This avoids surprising warning on currently deployed library code. * [compiler] test cases and stdlib source code updates for attributes checks PR * [aptos stdlib] Now that we have verified that the move-compiler Aptos stdlib hack works, fix stdlib to use #[test_only] on test helper functions instead of #[testonly]. Leave struct attributes since it's a deployment problem to make it disappear. * [Storage Service] Add type support for subscription requests. * [Storage Service] Add subscription stream support. * [Storage Service] Adopt subscription metadata and loop when serving subscriptions. * Make AccountAddress from_str conform to AIP-40, add from_str_strict (#9186) * [TS SDK v2] Remove redundant code in AccountAddress.fromString * Make AccountAddress from_str conform to AIP-40 * Use thiserror for AccountAddressParseError * [Executor-benchmark] Make connected grps of txns fully conflicting grps (#9647) [Executor-benchmark] Make connected grps of txns fully conflicting grps Fully conflicting grps are more meaningful pattern for testing out the sharded executor because they cannot be executed parallely (that is what we intend to benchmark). Cmd: cargo run -p aptos-executor-benchmark -- --block-size 100 --connected-tx-grps 5 --shuffle-connected-txns run-executor --main-signer-accounts 1000 --data-dir /tmp/some-db --checkpoint-dir /tmp/some-checkpoint --blocks 2 For a good random workload we would want 'num_accounts_per_grp' to be much greater than 'num_txns_per_grp'. The command enforces 'num_signer_accounts' > '2 * num_txns_per_grp'. * [NFT Metadata Crawler] Stop parsing completely if token_uri has already been parsed (#9649) * stop parsing completely if token_uri has already been parsed, add logs, add more panics (failed to write to gcs, failed to commit postgres tx) * add comment to clarify * error -> warn * rename * Fixed signer to be formatted properly (#8866) * Fixed signer to be formatted properly - fixed the issue in `string_utils` - replaced `testonly` with `test_only` * Add a feature flag * [Quorum Store] A couple metrics to observe backpressure better (#9239) ### Description A couple metrics to observe backpressure better * [Execution] Do not hold the Arc of committed block in execution. (#9674) * [NFT Metadata Crawler] Increase HTTP request for large files (#9677) * retry time increase * add to constants file * remove unneeded comment * Revert "swich to gcp" This reverts commit 8f28f93. * [NFT Metadata Crawler] Maintain image aspect ration on resize (#9688) * maintain image aspect ration on resize * fix lint * [Spec] update spec for vesting.move (#9115) * new vesting * new vesting * init * fix comment * fix md * add comment * add head * add func total_accumulated_rewards time * rm boogie * fixed timeout * close timeout function's verification * rust lint fix * del head --------- Co-authored-by: chan-bing <zzywullr@gmail.com> * [event_v2] module event extension, attribute and extended check * [eventv2] make v0/v1 to v1/v2 * [event_v2] test * [events v2] Sample of how to check for event type attributes in the extended checker * [eventv2] add module publishing validation * [eventv2] add module event feature * [eventv2] tests and fixes * [eventv2] address comments and forbid script emitting module events * [eventv2] add script event verifier to forbid event emitting * add #[event] to known attributes * trivial: add required feature to dependency to fix tests `e2e-move-tests` when running alone doesn't pass because `pub const GAS_UNIT_PRICE: u64 = 0;` is not in effect * Revert "add #[event] to known attributes" This reverts commit 5af30fb. * Revert "[eventv2] add script event verifier to forbid event emitting" This reverts commit 6a30bcd. * Revert "[eventv2] address comments and forbid script emitting module events" This reverts commit e4a9c2c. * Revert "[eventv2] tests and fixes" This reverts commit 4040b3d. * Revert "[eventv2] add module event feature" This reverts commit 5a12afc. * Revert "[eventv2] add module publishing validation" This reverts commit 7543f6e. * Revert "[events v2] Sample of how to check for event type attributes in the extended checker" This reverts commit 72cb7b4. * Revert "[event_v2] test" This reverts commit c93e61d. * Revert "[eventv2] make v0/v1 to v1/v2" This reverts commit 0ba554c. * Revert "[event_v2] module event extension, attribute and extended check" This reverts commit ed99bf0. * [compiler v2] Move the stackless bytecode crate out of prover (#9698) This moves the `move-stackless-bytecode` crate out of its current location in the `move-prover` into the `move-model`. This is pure relocation. Subsequent PRs will split off prover specific code and move it back to the prover tree. * [gas][docs] example guides (#9633) * [refactoring] Relax trait bounds, reduce dependence on `StateView` (#9644) API only needs move resolver, and not state view. Decoupling `MoveResolverExt` into `AptosMoveResolver + StateView` so that any future changes were not visible on API side. 1. Move API checks for resource groups where it should be. 2. Make trait bounds for config not require state view. Right now state view is used all over the place although it seems this can be avoided. 3. Reduce the usage of `MoveResolverExt`: e.g. session only needs to talk to resolver. 4. State value metadata has its own resolver. This makes storage accesses almost uniform (still need to fix `read_write_set`) which should be implemented in storage adapter. 5. Some minor typo fixes and dead code removal. * [dag] network sender implementation (#9456) * [Storage Service] Add subscription stream support. * [Storage Service] Add tests for subscription streams. * [Storage Service] Add tests for stream looping. * Add get_state_value_u128 method to TStateView (#9097) * [Spec] Ensures for voting (#9462) * new_voting * run lint * little change * fix md * fix schema name * fix error in simplemap --------- Co-authored-by: chan-bing <zzywullr@gmail.com> * [scripts] fixes to arch linux (pacman) * Archlinux / python / pip insist that you use the package manager solutions not python for pip * The `libudev-dev` seems to exist in the default system install and under a wildly different name. * [Block Executor] Refactor view & output traits & mvhashmap (#9659) * [MVHashMap] convenient (shared) view state, aggregator v1 port * [Block executor] new traits for processing the output * [typo*] aptos-token and wallets (#9661) * Update aptos-token.md typo "maxium"/maximum * [typo*]Update wallets.md dot added * [compiler v2] Stackless Bytecode Refactoring (#9713) This is the 2nd step in the refactoring started in #9698. This splits off prover specific parts from the crate `move-model/bytecode` into `move-prover/bytecode-pipeline`. Sharable dataflow analysis and transformation processors, like livevar and reaching definitions, stay. The tests have been split as well, and a common testing driver has been moved in its own test utility crate. Github does not nicely show diffs like this, but this is functionally a no-op which only Moves code around. * [DAG] Adding randomness field in dag node (#9687) * [NFT Metadata Crawler] Fix double slash URI parsing error, add ACKing PubSub message on receive (#9706) * fix double slash uri parsing error, acking on message receive * edit var names * fix test * [api] Add consistent API testing (#9577) * [api] Add consistent API testing (#9146) This commit introduces consistent API testing where we run a consistent load through the network mimicking various user flows. See proposal for more high level information. Changes: Created aptos-api-tester crate which runs 4 user flows: Account creation Coin transfer NFT transfer Module publishing, and outputs results to stdout. Added token-client to the SDK with the methods needed to run Your First NFT. mirroring the typescript SDK and the rust coin client by using existing builders Added PartialEq to aptos_rest_client::Account for testing. Added the crate to the tools docker image to run with GCP Cloud Run. * [api] Add metrics to api tester (#9307) This commit introduces logging tools on top of the consistent API testing introduced here. See proposal for more high level information. Changes: Added metrics pusher to aptos-api-tester with 4 histogram metrics for success, error, fail rates, and latency. Added logs to tests instead of printing to stdout. * [api] Add threads and timestamps to api tests (#9349) This commit builds threads and run ID on top of the consistent API testing introduced here. See proposal for more high level information. Changes: Refactored tests into dedicated file tests.rs. Switched to creating accounts for every tests instead of using the same and created pre test set up routines in testsetups.rs. Added run ID start_time to metrics. * [api] Add support for strictness level for consistency on API tests (#9444) Description This commit builds threads and run ID on top of the consistent API testing introduced here. See proposal for more high level information. One problem we saw in the dashboard is that the problems we observed were regarding eventual consistency. We added support for adjusting the strictness level for such errors. Changes: Refactor of tests into tests module and decomposing into steps (setting up for next PR, individual step timing). Persistent checks for information retrieval. Added token support for Testnet faucet. * [api] Add individual step timing to api tester (#9502) This commit builds individual step timing on top of the consistent API testing introduced here. See proposal for more high level information. This addition allows us to dive deeper on what the issue is should we detect any flow is consistently slower. Changes: Add timer macro in macros.rs. Add helpers for emitting metrics and refactor process_result. Put publish_module in its own thread by creating a tokio runtime. Add sleeps between persistent checks. Reduce API client sleep time to 100ms from 500ms. Make test setups persistent. * [api] Add view function testing to the API tester (#9658) This commit adds a new test which tests a simple view function. * [dag] e2e integration test (#9457) * [event_v2] module event extension, attribute and extended check * [eventv2] make v0/v1 to v1/v2 * [event_v2] test * [events v2] Sample of how to check for event type attributes in the extended checker * [eventv2] add module publishing validation * [eventv2] add module event feature * [eventv2] tests and fixes * [eventv2] address comments and forbid script emitting module events * [eventv2] add script event verifier to forbid event emitting * add #[event] to known attributes * [eventv2] api test * Update staking-pool-operations.md (#8685) * Update staking-pool-operations.md corrected CLI command to create staking_contract previously the command created a direct pool type * Update staking-pool-operations.md add additional 8 0s * Update Docker images (#9699) Co-authored-by: gedigi <gedigi@users.noreply.github.com> * [smart_vector] add destroy function for dropable T (#9434) * [CLI] Add support for setting faucet auth token (#9715) * [CLI] Add support for setting faucet auth token * Use FaucetClient * Unify thread name, and combine all threads in the same thread pool in framegraph produced by cpu_profiler. (#9721) * Feat/pypi ci (#9512) * add python sdk publish ci * remove check version * [framework] Reset the release yaml for upcoming 1.7 release Test Plan: ran yq / generate proposals locally ``` target/performance/aptos-release-builder generate-proposals --release-config aptos-move/aptos-release-builder/data/release.yaml --output-dir output ``` --------- Co-authored-by: Bo Wu <bo@aptoslabs.com> Co-authored-by: Maayan <maayan@aptoslabs.com> Co-authored-by: Álvaro Lillo Igualada <alilloig@gmail.com> Co-authored-by: yunuseozer <yunusozer@berkeley.edu> Co-authored-by: Oğuzhan (Olsen) Budanur <74462406+olsenbudanur@users.noreply.github.com> Co-authored-by: Olsen Budanur <olsenbudanur@Olsens-MBP.localdomain> Co-authored-by: Balaji Arun <balaji@aptoslabs.com> Co-authored-by: aldenhu <msmouse@gmail.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: rustielin <rustielin@users.noreply.github.com> Co-authored-by: igor-aptos <110557261+igor-aptos@users.noreply.github.com> Co-authored-by: larry-aptos <112209412+larry-aptos@users.noreply.github.com> Co-authored-by: George Mitenkov <georgemitenk0v@gmail.com> Co-authored-by: Josh Lind <josh.lind@hotmail.com> Co-authored-by: Gerardo Di Giacomo <gerardo@aptoslabs.com> Co-authored-by: Zorrot Chen <UI_Zorrot@163.com> Co-authored-by: chan-bing <zzywullr@gmail.com> Co-authored-by: Matt <90358481+xbtmatt@users.noreply.github.com> Co-authored-by: Greg Nazario <greg@gnazar.io> Co-authored-by: Daniel Porteous (dport) <daniel@dport.me> Co-authored-by: Victor Gao <10379359+vgao1996@users.noreply.github.com> Co-authored-by: Justin Chang <37165464+just-in-chang@users.noreply.github.com> Co-authored-by: Wolfgang Grieskamp <wg@aptoslabs.com> Co-authored-by: maayan <maayansavir@gmail.com> Co-authored-by: Alin Tomescu <tomescu.alin@gmail.com> Co-authored-by: William Law <williamlaw.wtl@gmail.com> Co-authored-by: Vladislav ~ cryptomolot <88001005+cryptomolot@users.noreply.github.com> Co-authored-by: David Wolinsky <isaac.wolinsky@gmail.com> Co-authored-by: JasperTimm <jaspa@jaspa.codes> Co-authored-by: Jin <128556004+0xjinn@users.noreply.github.com> Co-authored-by: aalok-t <140445856+aalok-t@users.noreply.github.com> Co-authored-by: Aalok Thakkar <aalok@Aaloks-MacBook-Pro.local> Co-authored-by: Brian R. Murphy <132495859+brmataptos@users.noreply.github.com> Co-authored-by: Manu Dhundi <manudhundi@gmail.com> Co-authored-by: Junkil Park <jpark@aptoslabs.com> Co-authored-by: Brian (Sunghoon) Cho <brian@aptoslabs.com> Co-authored-by: Guoteng Rao <3603304+grao1991@users.noreply.github.com> Co-authored-by: Bo Wu <bowu365@gmail.com> Co-authored-by: Aaron Gao <lightmark@gmail.com> Co-authored-by: Andrei Tonkikh <andrei.tonkikh@gmail.com> Co-authored-by: Rati Gelashvili <gelash@users.noreply.github.com> Co-authored-by: Daniel Xiang <66756900+danielxiangzl@users.noreply.github.com> Co-authored-by: Nurullah Giray Kuru <giraykuru@gmail.com> Co-authored-by: michelle-aptos <120680608+michelle-aptos@users.noreply.github.com> Co-authored-by: gedigi <gedigi@users.noreply.github.com> Co-authored-by: Fenix <zhuzhenfeng1993@hotmail.com> Co-authored-by: Perry Randall <perryjrandall@gmail.com>
Live variables is a standard compiler construction data flow analysis. It computes the set of variables which are alive (being used in subsequent reachable code) before and after each program point. The implementation from the prover uses our dataflow analysis framework and is here.
In this PR we create the first processing step in the bytecode pipeline for the V2 compiler with the new module
pipeline/livevar_analysis_step.rs
which forwards logical work to the existing analysis. This is then used primarily to compute copy/move information correctly which is a pre-requisite for ability and borrow safety analysis.