[prev in list] [next in list] [prev in thread] [next in thread]
List: kde-release-team
Subject: Re: Review Request 111718: Prevent EntityListCache from causing endless loop
From: "Commit Hook" <null () kde ! org>
Date: 2013-07-27 12:04:23
Message-ID: 20130727120423.4847.22843 () vidsolbach ! de
[Download RAW message or body]
[Attachment #2 (multipart/alternative)]
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/111718/
-----------------------------------------------------------
(Updated July 27, 2013, 12:04 p.m.)
Status
------
This change has been marked as submitted.
Review request for KDEPIM-Libraries and Release Team.
Description
-------
Calling ensureCached() can cause the cache and Monitor to end up in an endless loop.
In situation, that there are items 1,2,4,5 and 6 in the cache, calling \
ensureCached(1,2,3,4,5) will call request(3) and return false. request() internally \
calls shrinkCache() which removes all fetched items to make sure the cache does not \
grow indefinitely. After the item 3 is fetched, we end up with cache with only item \
3. Monitor will then again call ensureCached(1,2,3,4,5), which will call \
request(1,2,4,5) which in return calls shrinkCache(), which will remove the item 3 \
and after fetched is finished the cache only contains items 1,2,4 and 5. This way the \
cache ends up in an endless loop because it will never contain all items that Monitor \
requested.
This patch prevents shrinkCache() from removing items that are part of the \
ensureCached() request.
I would like to get this patch to 4.11, because it's not that difficult to trigger \
this bug and in such case you end up with Akonadi resource going nuts (and not \
processing any further requests)
This implementation is still not optimal, because meanwhile the Monitor can call \
ensureCached(7,8,9), which will cause (1,2,4,5) to be removed from cache anyway, so \
it can take several tries for EntityListCache to retrieve the entire set (1,2,3,4,5). \
However now it's guaranteed that at some point the cache will succeed. We can rethink \
(and maybe remove) the cache when server-side recording is implemented.
Diffs
-----
akonadi/entitycache_p.h 5f00917
Diff: http://git.reviewboard.kde.org/r/111718/diff/
Testing
-------
Thanks,
Dan Vrátil
[Attachment #5 (text/html)]
<html>
<body>
<div style="font-family: Verdana, Arial, Helvetica, Sans-Serif;">
<table bgcolor="#f9f3c9" width="100%" cellpadding="8" style="border: 1px #c9c399 \
solid;"> <tr>
<td>
This is an automatically generated e-mail. To reply, visit:
<a href="http://git.reviewboard.kde.org/r/111718/">http://git.reviewboard.kde.org/r/111718/</a>
</td>
</tr>
</table>
<br />
<table bgcolor="#e0e0e0" width="100%" cellpadding="8" style="border: 1px gray \
solid;"> <tr>
<td>
<h1 style="margin-right: 0.2em; padding: 0; font-size: 10pt;">This change has been \
marked as submitted.</h1> </td>
</tr>
</table>
<br />
<table bgcolor="#fefadf" width="100%" cellspacing="0" cellpadding="8" \
style="background-image: \
url('http://git.reviewboard.kde.org/static/rb/images/review_request_box_top_bg.ab6f3b1072c9.png'); \
background-position: left top; background-repeat: repeat-x; border: 1px black \
solid;"> <tr>
<td>
<div>Review request for KDEPIM-Libraries and Release Team.</div>
<div>By Dan Vrátil.</div>
<p style="color: grey;"><i>Updated July 27, 2013, 12:04 p.m.</i></p>
<h1 style="color: #575012; font-size: 10pt; margin-top: 1.5em;">Description </h1>
<table width="100%" bgcolor="#ffffff" cellspacing="0" cellpadding="10" \
style="border: 1px solid #b8b5a0"> <tr>
<td>
<pre style="margin: 0; padding: 0; white-space: pre-wrap; white-space: \
-moz-pre-wrap; white-space: -pre-wrap; white-space: -o-pre-wrap; word-wrap: \
break-word;">Calling ensureCached() can cause the cache and Monitor to end up in an \
endless loop.
In situation, that there are items 1,2,4,5 and 6 in the cache, calling \
ensureCached(1,2,3,4,5) will call request(3) and return false. request() internally \
calls shrinkCache() which removes all fetched items to make sure the cache does not \
grow indefinitely. After the item 3 is fetched, we end up with cache with only item \
3. Monitor will then again call ensureCached(1,2,3,4,5), which will call \
request(1,2,4,5) which in return calls shrinkCache(), which will remove the item 3 \
and after fetched is finished the cache only contains items 1,2,4 and 5. This way the \
cache ends up in an endless loop because it will never contain all items that Monitor \
requested.
This patch prevents shrinkCache() from removing items that are part of the \
ensureCached() request.
I would like to get this patch to 4.11, because it's not that difficult to \
trigger this bug and in such case you end up with Akonadi resource going nuts (and \
not processing any further requests)
This implementation is still not optimal, because meanwhile the Monitor can call \
ensureCached(7,8,9), which will cause (1,2,4,5) to be removed from cache anyway, so \
it can take several tries for EntityListCache to retrieve the entire set (1,2,3,4,5). \
However now it's guaranteed that at some point the cache will succeed. We can \
rethink (and maybe remove) the cache when server-side recording is implemented.</pre> \
</td> </tr>
</table>
<h1 style="color: #575012; font-size: 10pt; margin-top: 1.5em;">Diffs</b> </h1>
<ul style="margin-left: 3em; padding-left: 0;">
<li>akonadi/entitycache_p.h <span style="color: grey">(5f00917)</span></li>
</ul>
<p><a href="http://git.reviewboard.kde.org/r/111718/diff/" style="margin-left: \
3em;">View Diff</a></p>
</td>
</tr>
</table>
</div>
</body>
</html>
_______________________________________________
release-team mailing list
release-team@kde.org
https://mail.kde.org/mailman/listinfo/release-team
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic