Bug 165519 - [WASM] Add support for truncation operators
Summary: [WASM] Add support for truncation operators
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Keith Miller
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-12-06 23:55 PST by Keith Miller
Modified: 2016-12-07 13:44 PST (History)
5 users (show)

See Also:

Attachments
Patch (41.70 KB, patch)
2016-12-07 00:01 PST, Keith Miller 		ggaren: review+
Details | Formatted Diff | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Keith Miller 2016-12-06 23:55:24 PST 
Add support for truncation operators
Comment 1 Keith Miller 2016-12-07 00:01:43 PST 
Created attachment 296382 [details]
Patch
Comment 2 Geoffrey Garen 2016-12-07 10:39:06 PST 
Comment on attachment 296382 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=296382&action=review

r=me

> Source/JavaScriptCore/assembler/MacroAssemblerX86_64.h:1316
> +        // instructions are the same. Otherwise, we subtract need to int64_t::min() convert then add it back.

Otherwise, we need to: subtract int64_t::min(); truncate double to uint64_t; then add back int64_t::min().

> Source/JavaScriptCore/assembler/MacroAssemblerX86_64.h:1324
> +        m_assembler.movq_i64r(0x8000000000000000, scratchRegister());

Let's use numeric_limits<int64_t>::min() here instead of 0x8.

> Source/JavaScriptCore/assembler/MacroAssemblerX86_64.h:1356
> +        m_assembler.movq_i64r(0x8000000000000000, scratchRegister());

Samesies.

> Source/JavaScriptCore/wasm/WasmB3IRGenerator.cpp:1043
> +        constant = m_currentBlock->appendNew<ConstDoubleValue>(m_proc, Origin(), bitwise_cast<double>(0x43e0000000000000));

static_cast<double>(numeric_limits<int64_t>::min())

> Source/JavaScriptCore/wasm/WasmB3IRGenerator.cpp:1108
> +        constant = m_currentBlock->appendNew<ConstFloatValue>(m_proc, Origin(), bitwise_cast<float>(0x5f000000));

static_cast<float>(numeric_limits<int64_t>::min())
Comment 3 Keith Miller 2016-12-07 13:44:27 PST 
Committed r209476: <http://trac.webkit.org/changeset/209476>