[Webkit-unassigned] [Bug 138315] New: [JSC] Erratum (835769) in Cortex-A53

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=138315

            Bug ID: 138315
           Summary: [JSC] Erratum (835769) in Cortex-A53
    Classification: Unclassified
           Product: WebKit
           Version: 528+ (Nightly build)
          Hardware: Unspecified
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: JavaScriptCore
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: akiss at inf.u-szeged.hu
                CC: dbates at webkit.org, fpizlo at apple.com, loki at webkit.org,
                    msaboff at apple.com

There is an erratum (835769) in some Cortex-A53, which can only be worked around at code generation: It is possible for an A64 64-bit multiply-accumulate instruction to generate an incorrect result when immediately preceded by a memory instruction. In WebKit, that means 4 affected stages, potentially:

1. compilation with gcc or clang
2. processing LLInt asm sources with the offline assembler
3. MacroAssembler
4. LLVM in FTL

1. Fixes for both gcc and clang have already been submitted:
https://gcc.gnu.org/ml/gcc-cvs/2014-10/msg00331.html
https://gcc.gnu.org/ml/gcc-cvs/2014-10/msg00332.html
https://gcc.gnu.org/ml/gcc-cvs/2014-10/msg00333.html
https://gcc.gnu.org/ml/gcc-cvs/2014-10/msg00335.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20141006/239368.html
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20141006/116322.html
So, if building WebKit for Cortex-A53, the build system should detect whether the compiler tool chain has support for the workaround and apply it if available.

2 & 3. Offlineasm & MacroAssembler should be amended with the workaround.

4. LLVM trunk already has the workaround, so as soon as ARM64 (for Linux) has FTL working, WebKit can rely on LLVM generating correct code.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20141103/e50419c4/attachment-0002.html>