harmony-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mikhail Fursov (JIRA)" <j...@apache.org>
Subject [jira] Closed: (HARMONY-2728) [drlvm][jit][opt] InstCodeSelector::tau_arrayLen generates non-serilizable rt-dependent immediate
Date Tue, 17 Jul 2007 11:19:05 GMT

     [ https://issues.apache.org/jira/browse/HARMONY-2728?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Mikhail Fursov closed HARMONY-2728.
-----------------------------------

    Resolution: Won't Fix

Closing this issue as "Won't Fix"
Reason: Yes, most of runtime infos are resolved in "emitter" pass. However there is nothing
wrong to resolve offsets of 'array length' field in LIR selector and allow optimizations like
'cafl' optimize it.

> [drlvm][jit][opt] InstCodeSelector::tau_arrayLen generates non-serilizable rt-dependent
immediate
> -------------------------------------------------------------------------------------------------
>
>                 Key: HARMONY-2728
>                 URL: https://issues.apache.org/jira/browse/HARMONY-2728
>             Project: Harmony
>          Issue Type: Bug
>          Components: DRLVM
>            Reporter: Nikolay Sidelnikov
>            Assignee: Mikhail Fursov
>            Priority: Minor
>
> Function InstCodeSelector::tau_arrayLen (Ia32InstCodeSelector.cpp) contains the
> following code
>   Opnd * offset=irManager.newImmOpnd(typeManager.getInt32Type(),
> arrayType->getArrayLengthOffset());
> which discords with the general Jitrino policy to keep (until code emission)
> all runtime-dependent immediates as operands with 'runtimeInfo' field set to a
> value discribing the immediate in serializable form.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message