Commit Graph

21 Commits

Author SHA1 Message Date
Patrick Thier
f5a3017122 [test] Disable --always-opt for mjsunit/baseline/batch-compilation.js
Explicitliy add --no-always-opt to prevent Fuzzer failures.

Bug: v8:11853
Change-Id: Idf1ee842fd7e404516e27d1f5f5e6afe87fc5bd1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2951726
Reviewed-by: Victor Gomes <victorgomes@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/master@{#75079}
2021-06-10 14:47:55 +00:00
Patrick Thier
c871945055 [sparkplug][test] Fix NumFuzz (mjsunit/baseline/batch-compilation.js)
--stress-concurrent-inlining has a negative implication for
--lazy-feedback-allocation. We need to explicitly add
--no-stress-concurrent-inlining, since the test relies on lazy feedback
allocation.

Bug: v8:11853
Change-Id: I2bd8f0da05a766dd7282cdb3b70c4a1b478c71cc
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2948647
Commit-Queue: Patrick Thier <pthier@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#75037}
2021-06-09 08:52:28 +00:00
Patrick Thier
03dade3ec2 [sparkplug] Fix NumFuzz issues with batch-compilation test
Add flag --lazy-feedback-allocation to prevent NumFuzz errors.

Bug: v8:11853
Change-Id: I5170ef9db374e168cf248b86dbed3c3e7c87f826
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2944428
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/master@{#74990}
2021-06-07 16:16:14 +00:00
Patrick Thier
6ff1129ca3 [sparkplug] Batch compilation
Instead of compiling a function with baseline immediately when the
interrupt budget is hit, we compile functions in batches to save some
memory protection flips on code pages.

This CL introduces batch compilation behind --baseline-batch-compilation
(enabled on future) and adds a flag
--baseline-batch-compilation-threshold to control the size of batches.

Bug: v8:11790

Change-Id: I3efc360424a14e4b07c6570e48860509ae59e591
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2891656
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/master@{#74913}
2021-06-02 11:41:33 +00:00
Patrick Thier
af59290ffc [test] Force more flags for Sparkplug OSR Test
This test relies on non-concurrent optimizations to precisely test
transitions between the different tiers.

Disable flags that invalidate this test:
--no-always-opt to test transiation from Ignition to Sparkplug
--deopt-every-n-times=0 to test transition from Sparkplug to TF

Bug: v8:11656
Change-Id: I85047015d3f7cfdf00bddad59c9742d5d8f5d223
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2827902
Auto-Submit: Patrick Thier <pthier@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73978}
2021-04-15 16:14:30 +00:00
Patrick Thier
5f49bbdf84 [test][sparkplug] Test Ignition -> Sparkplug OSR
- Add %BaselineOsr to manually trigger OSR to Baseline.
- Add flags to %GetOptimizationStatus to check if the topmost frame is
an Interpreter/Baseline frame.
- Add mjsunit test.

Bug: v8:11420
Change-Id: Id80421ad97ee719a67ef299cc700da9c44f23bae
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2814567
Auto-Submit: Patrick Thier <pthier@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73937}
2021-04-13 14:51:28 +00:00
Patrick Thier
59807ec552 Fix d8.test.verifySourcePositions
Add check, that passed argument is a HeapObject.

Bug: chromium:1196503
Change-Id: I23d951b5581781ad3c6867d81c765d13c329d3a8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2808936
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Auto-Submit: Patrick Thier <pthier@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73820}
2021-04-07 08:59:13 +00:00
Patrick Thier
69d706dc5f Allow only JSFunction/JSBoundFunction in d8.test.verifySourcePositions
Explicitly check for JSFunction or JSBoundFunction and throw if any other
JS type is passed to d8.test.verifySourcePositions.

Bug: chromium:1195717
Change-Id: Id65875526d5d6b3f720850d41d0a8192ec407035
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2807607
Auto-Submit: Patrick Thier <pthier@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73811}
2021-04-06 16:23:42 +00:00
Leszek Swirski
4baf07a769 [sparkplug] Add scratch register for arm64
We didn't have enough scratch registers for a suspend generator
whose field write offsets exceeded the immediate value range.

Bug: v8:11420, chromium:1193493
Change-Id: Iee90db4ef1ec00924bcc4791a8e6ffb9138bb388
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2794424
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73739}
2021-03-31 12:41:04 +00:00
Leszek Swirski
00afef3c7f [sparkplug/ia32] Fix argc clobbering
Fix the InstallBaselineCode path in the InterpreterEntryTrampoline to
restore the clobbered eax (i.e. argc) register.

Bug: v8:11420, chromium:1192459
Change-Id: I97ce5739cf22a08fbb46dbf372ab6276bb802440
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2791567
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Victor Gomes <victorgomes@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Victor Gomes <victorgomes@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73721}
2021-03-30 12:03:01 +00:00
Leszek Swirski
791e085d2f [sparkplug] Add always_sparkplug variant
Add an always_sparkplug testing variant, and fix a couple of issues it
found.

Bug: v8:11420
Change-Id: I7d87a41e3413f40271a0140118531f075d633b23
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2773047
Reviewed-by: Patrick Thier <pthier@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73529}
2021-03-19 09:52:50 +00:00
Leszek Swirski
d7ed4f97c2 [sparkplug] Inline TypeOf checks
In the TestTypeOf handler, use the statically known value of the type-of
literal flag to emit the specific type-of check for that type, rather
than going to the general type-of builtin. These checks are very simple,
so we can get away with hand-writing them in the baseline compiler.

Bug: v8:11420
Change-Id: Id9f51042916abaea62f929a2f95cf2c64dc32ee3
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2772613
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Reviewed-by: Patrick Thier <pthier@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73520}
2021-03-18 17:10:19 +00:00
Patrick Thier
6748e359a2 Accept Proxy in d8.test.verfiySourcePositions
If a Proxy is passed as argument to d8.test.verifySourcePositions,
unwrap the target function.

Bug: chromium:1186491
Change-Id: Iaacbf800f023121a07035a88745919df68f79217
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2748081
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73398}
2021-03-15 09:53:14 +00:00
pthier
2966c8967a Reland "[sparkplug] Change bytecode offset mapping and introduce iterator."
This is a reland of a8b61ef521

The main reason for the revert was not related to this CL and was fixed
with https://crrev.com/c/2739646
In addition debug output in d8.test.verifySourcePositions was removed
due to TSAN complaints.

Original change's description:
> [sparkplug] Change bytecode offset mapping and introduce iterator.
>
> Previously, we recorded pairs of (bytecode offset, sparkplug pc) to
> create a mapping of bytecode offset <-> sparkplug pc.
> These pairs were only recorded after builtin/runtime calls.
> In preparation for deoptimizing to Sparkplug, we need a more precise
> mapping.
> With this CL, we record positions for every bytecode. Instead of storing
> a pair of (bytecode offset, sparkplug pc), we store only the pc,
> calculating the bytecode offset from the index in the mapping table.
> For easier use an iterator to access the mapping is introduced.
>
> Drive-by: Reduce sampling interval in cpu-profiler cctest to get rid of
flaky failures.
>
> Bug: v8:11420, v8:11429
> Change-Id: I36a9171f43a574eb67880cbca6cf9ff7ab291e60
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2720189
> Reviewed-by: Victor Gomes <victorgomes@chromium.org>
> Reviewed-by: Camillo Bruni <cbruni@chromium.org>
> Auto-Submit: Patrick Thier <pthier@chromium.org>
> Commit-Queue: Patrick Thier <pthier@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#73186}
>
> Change-Id: I9ab4cb60da002ef130f8a21ad10ba69e2826a7b6

Change-Id: I9ab4cb60da002ef130f8a21ad10ba69e2826a7b6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2745335
Reviewed-by: Victor Gomes <victorgomes@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73293}
2021-03-09 13:43:05 +00:00
Maya Lekova
6fa780ffdc Revert "[sparkplug] Change bytecode offset mapping and introduce iterator."
This reverts commit a8b61ef521.

Reason for revert: Looks like it breaks GC stress bot - https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Linux64%20GC%20Stress%20-%20custom%20snapshot/35880/overview

Original change's description:
> [sparkplug] Change bytecode offset mapping and introduce iterator.
>
> Previously, we recorded pairs of (bytecode offset, sparkplug pc) to
> create a mapping of bytecode offset <-> sparkplug pc.
> These pairs were only recorded after builtin/runtime calls.
> In preparation for deoptimizing to Sparkplug, we need a more precise
> mapping.
> With this CL, we record positions for every bytecode. Instead of storing
> a pair of (bytecode offset, sparkplug pc), we store only the pc,
> calculating the bytecode offset from the index in the mapping table.
> For easier use an iterator to access the mapping is introduced.
>
> Drive-by: Reduce sampling interval in cpu-profiler cctest to get rid of
> flaky failures.
>
> Bug: v8:11420, v8:11429
> Change-Id: I36a9171f43a574eb67880cbca6cf9ff7ab291e60
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2720189
> Reviewed-by: Victor Gomes <victorgomes@chromium.org>
> Reviewed-by: Camillo Bruni <cbruni@chromium.org>
> Auto-Submit: Patrick Thier <pthier@chromium.org>
> Commit-Queue: Patrick Thier <pthier@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#73186}

Bug: v8:11420
Bug: v8:11429
Change-Id: Ie71e7ce234e7b9ab9a2ec99a983e9900f35baa44
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2735397
Auto-Submit: Maya Lekova <mslekova@chromium.org>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#73187}
2021-03-04 13:01:08 +00:00
pthier
a8b61ef521 [sparkplug] Change bytecode offset mapping and introduce iterator.
Previously, we recorded pairs of (bytecode offset, sparkplug pc) to
create a mapping of bytecode offset <-> sparkplug pc.
These pairs were only recorded after builtin/runtime calls.
In preparation for deoptimizing to Sparkplug, we need a more precise
mapping.
With this CL, we record positions for every bytecode. Instead of storing
a pair of (bytecode offset, sparkplug pc), we store only the pc,
calculating the bytecode offset from the index in the mapping table.
For easier use an iterator to access the mapping is introduced.

Drive-by: Reduce sampling interval in cpu-profiler cctest to get rid of
flaky failures.

Bug: v8:11420, v8:11429
Change-Id: I36a9171f43a574eb67880cbca6cf9ff7ab291e60
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2720189
Reviewed-by: Victor Gomes <victorgomes@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Auto-Submit: Patrick Thier <pthier@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73186}
2021-03-04 12:30:07 +00:00
Leszek Swirski
d0136a5caa [sparkplug] CompileWithBaseline to Compiler::CompileBaseline
Move the CompileWithBaseline interface to the Compiler class, as
CompileBaseline, which will do the additional work of pre-compiling
to bytecode, ensuring there is a feedback vector, and setting the
code on the function closure.

As a drive-by, fix v8_enable_trace_unoptimized to have a blank default
value, so that v8_enable_trace_ignition/v8_enable_trace_baseline_exec
can set it.

Bug: v8:11420, v8:11429
Change-Id: If715161de71f7d9300f3fdcbb50cc678b1fcdfdf
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2697352
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72819}
2021-02-17 16:01:45 +00:00
Leszek Swirski
862c6281f7 [sparkplug] Use NeverOptimize in tests
This prevents conflicts with e.g. --always-opt

Fixed: v8:11462
Change-Id: Ib98378a1bf620f5706bf3eef06f997bd994b1e2c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2700035
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72815}
2021-02-17 15:06:12 +00:00
Victor Gomes
83867ef69e [cleanup] Rename DeoptimizeBaseline to DiscardBaselineCode
- Baseline is to be consider non-optimized code, so for consistency we rename these functions to DiscardBaselineCode.
- Move to debug/, since discarding baseline code is only used by the debugger.
- %DeoptimizeNow and %DeoptimizeFunction are not to be used to tier down from Sparkplug to Ignition

Change-Id: I050607d4d6978907c589e54c57e940979b0a9a15
Bug: v8:11429
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2692699
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Victor Gomes <victorgomes@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72732}
2021-02-15 10:40:19 +00:00
Leszek Swirski
c913ef3a91 [sparkplug] Change Sparkplug to Baseline
Currently we sometimes refer to baseline code or the baseline compiler
by its codename (Sparkplug). The codename is fun, but we should be
consistent and call things by one name or the other. Following the
pattern of Ignition stuff being called "interpreter", we call Sparkplug
"baseline", and leave the codename only in flags and variants.

Bug: v8:11420
Change-Id: I432e5629518be7c7ad38b6acff024c91d4cfd6d3
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2692186
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72696}
2021-02-12 13:51:35 +00:00
Leszek Swirski
c053419e8c [sparkplug] Upstream Sparkplug
Sparkplug is a new baseline, non-optimising second-tier compiler,
designed to fit in the compiler trade-off space between Ignition and
TurboProp/TurboFan.

Design doc:
https://docs.google.com/document/d/13c-xXmFOMcpUQNqo66XWQt3u46TsBjXrHrh4c045l-A/edit?usp=sharing

Bug: v8:11420
Change-Id: Ideb7270db3d6548eedd8337a3f596eb6f8fea6b1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2667514
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72686}
2021-02-12 09:31:06 +00:00