You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@trafficserver.apache.org by "Leif Hedstrom (JIRA)" <ji...@apache.org> on 2011/06/26 18:23:47 UTC

[jira] [Commented] (TS-567) Cleanup of memory allocation in libinktomi++

    [ https://issues.apache.org/jira/browse/TS-567?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055110#comment-13055110 ] 

Leif Hedstrom commented on TS-567:
----------------------------------

I think we'll also want some additions done to other allocators, such as the proxy / class allocators (and freelist). E.g. it'd be great to have a high debug level where the class allocators aren't returning memory on the freelist, but instead just use free(). This ought to make it possible to do better memory debugging using tools such as tcmalloc and valgrind (which can keep track of malloc and free's, and warn on bad free's / usage of free'd memory).

> Cleanup of memory allocation in libinktomi++
> --------------------------------------------
>
>                 Key: TS-567
>                 URL: https://issues.apache.org/jira/browse/TS-567
>             Project: Traffic Server
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Leif Hedstrom
>            Assignee: Leif Hedstrom
>            Priority: Critical
>             Fix For: 3.1.0
>
>         Attachments: clear_memory.patch, malloc.txt
>
>
> There's a lot of ifdef's and debugging code paths around various allocation schemes in libinktomi++. We should clean this up, and simplify. It'd also be great to make it easy to enable various other allocators, e.g. jemalloc or tcmalloc, as part of the build system, causing ink_malloc / xmalloc and STL containers to use the new allocator. In particular for STL containers, it means we might need to have some allocator setup / initialization?
> Also, when should we use ink_malloc() vs xmalloc() ? It's quite inconsistently used right now I think. And, there are a few places using malloc() (etc.) directly, which we ought to fix.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira