CINXE.COM

<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>LKML: "Artem B. Bityuckiy": Re: [RFC] CryptoAPI & Compression</title><link href="/css/message.css" rel="stylesheet" type="text/css" /><link href="/css/wrap.css" rel="alternate stylesheet" type="text/css" title="wrap" /><link href="/css/nowrap.css" rel="stylesheet" type="text/css" title="nowrap" /><link href="/favicon.ico" rel="shortcut icon" /><script src="/js/simple-calendar.js" type="text/javascript"></script><script src="/js/styleswitcher.js" type="text/javascript"></script><link rel="alternate" type="application/rss+xml" title="lkml.org : last 100 messages" href="/rss.php" /><link rel="alternate" type="application/rss+xml" title="lkml.org : last messages by "Artem B. Bityuckiy"" href="/groupie.php?aid=28528" /><script> var _paq = window._paq = window._paq || []; /* tracker methods like "setCustomDimension" should be called before "trackPageView" */ _paq.push(["setDoNotTrack", true]); _paq.push(["disableCookies"]); _paq.push(['trackPageView']); _paq.push(['enableLinkTracking']); (function() { var u="//m.lkml.org/"; _paq.push(['setTrackerUrl', u+'matomo.php']); _paq.push(['setSiteId', '1']); var d=document, g=d.createElement('script'), s=d.getElementsByTagName('script')[0]; g.async=true; g.src=u+'matomo.js'; s.parentNode.insertBefore(g,s); })(); </script></head><body onload="es.jasper.simpleCalendar.init();" itemscope="itemscope" itemtype="http://schema.org/BlogPosting"><table border="0" cellpadding="0" cellspacing="0"><tr><td width="180" align="center"><a href="/"><img style="border:0;width:135px;height:32px" src="/images/toprowlk.gif" alt="lkml.org" /></a></td><td width="32">聽</td><td class="nb"><div><a class="nb" href="/lkml"> [lkml]</a> 聽 <a class="nb" href="/lkml/2005"> [2005]</a> 聽 <a class="nb" href="/lkml/2005/3"> [Mar]</a> 聽 <a class="nb" href="/lkml/2005/3/29"> [29]</a> 聽 <a class="nb" href="/lkml/last100"> [last100]</a> 聽 <a href="/rss.php"><img src="/images/rss-or.gif" border="0" alt="RSS Feed" /></a></div><div>Views: <a href="#" class="nowrap" onclick="setActiveStyleSheet('wrap');return false;">[wrap]</a><a href="#" class="wrap" onclick="setActiveStyleSheet('nowrap');return false;">[no wrap]</a> 聽 <a class="nb" href="/lkml/mheaders/2005/3/29/122" onclick="this.href='/lkml/headers'+'/2005/3/29/122';">[headers]</a>聽 <a href="/lkml/bounce/2005/3/29/122">[forward]</a>聽 </div></td><td width="32">聽</td></tr><tr><td valign="top"><div class="es-jasper-simpleCalendar" baseurl="/lkml/"></div><div class="threadlist">Messages in this thread</div><ul class="threadlist"><li class="root"><a href="/lkml/2005/3/25/104">First message in thread</a></li><li><a href="/lkml/2005/3/28/118">"Artem B. Bityuckiy"</a><ul><li><a href="/lkml/2005/3/29/92">Herbert Xu</a><ul><li class="origin"><a href="/lkml/2005/3/30/313">"Artem B. Bityuckiy"</a><ul><li><a href="/lkml/2005/3/30/313">Herbert Xu</a><ul><li><a href="/lkml/2005/3/31/88">"Artem B. Bityuckiy"</a></li></ul></li></ul></li></ul></li><li><a href="/lkml/2005/3/31/74">Herbert Xu</a><ul><li><a href="/lkml/2005/4/3/58">"Artem B. Bityuckiy"</a><ul><li><a href="/lkml/2005/4/3/60">Herbert Xu</a><ul><li><a href="/lkml/2005/4/3/61">"Artem B. Bityuckiy"</a></li><li><a href="/lkml/2005/4/3/64">"Artem B. Bityuckiy"</a></li></ul></li></ul></li></ul></li></ul></li></ul><div class="threadlist">Patch in this message</div><ul class="threadlist"><li><a href="/lkml/diff/2005/3/29/122/1">Get diff 1</a></li></ul></td><td width="32" rowspan="2" class="c" valign="top"><img src="/images/icornerl.gif" width="32" height="32" alt="/" /></td><td class="c" rowspan="2" valign="top" style="padding-top: 1em"><table><tr><td><table><tr><td class="lp">Date</td><td class="rp" itemprop="datePublished">Tue, 29 Mar 2005 12:55:11 +0100 (BST)</td></tr><tr><td class="lp">From</td><td class="rp" itemprop="author">"Artem B. Bityuckiy" <></td></tr><tr><td class="lp">Subject</td><td class="rp" itemprop="name">Re: [RFC] CryptoAPI & Compression</td></tr></table></td><td></td></tr></table><pre itemprop="articleBody">> Are you sure that 12 bytes is enough for all cases? It would seem<br />> to be safer to use the formula in deflateBound/compressBound from<br />> later versions (> 1.2) of zlib to calculate the reserve.<br />><br />I'm not sure. David Woodhouse (the author) said that this is probably <br />enough in any case but a lot of time has gone since the code was written <br />and he doesn't remember for sure. I have also seen some magic number "12" <br />somewhere in zlib, but I'm not sure.<br /><br />At least my practice shows that 12 is Ok for JFFS2 where we compress fewer <br />then 4K a a time. I'll explore this.<br /><br />> We normally put the operator on the preceding line, i.e.,<br />><br />> while (foo &&<br />> bar) {<br />If this is the the common practice for Linux, then OK. My argument is the <br />GNU Coding style which recommends:<br /><br />----------------------------------------------------------------------<br />When you split an expression into multiple lines, split it before an <br />operator, not after one. Here is the right way:<br /><br /> if (foo_this_is_long && bar > win (x, y, z)<br /> && remaining_condition)<br />----------------------------------------------------------------------<br /><br />while the Linux coding style doesn't mention this AFAIR. And of course, <br />Linux doesn't have to obey that rule. <br /><br />Ok. This is not the final patch but more like RFC and I can re-format and <br />re-send it. :-) Please, feel free to re-format it as you would like <br />yourself.<br /><br />And one more thing I wanted to offer. In the <br />deflate_[compress|uncompress|pcompress] functions we call the <br />zlib_[in|de]flateReset function at the beginning. This is OK. But when we <br />unload the deflate module we don't call zlib_[in|de]flateEnd to free all <br />the zlib internal data. It looks like a bug for me. Please, consider the <br />attached patch.<br /><br />--<br />Best Regards,<br />Artem B. Bityuckiy,<br />St.-Petersburg, Russia.diff -auNrp linux-2.6.11.5/crypto/deflate.c linux-2.6.11.5_changed/crypto/deflate.c<br />--- linux-2.6.11.5/crypto/deflate.c 2005-03-29 15:37:44.000000000 +0400<br />+++ linux-2.6.11.5_changed/crypto/deflate.c 2005-03-29 15:37:38.000000000 +0400<br />@@ -93,11 +93,13 @@ out_free:<br /> <br /> static void deflate_comp_exit(struct deflate_ctx *ctx)<br /> {<br />+ zlib_deflateEnd(&ctx->comp_stream);<br /> vfree(ctx->comp_stream.workspace);<br /> }<br /> <br /> static void deflate_decomp_exit(struct deflate_ctx *ctx)<br /> {<br />+ zlib_inflateEnd(&ctx->decomp_stream);<br /> kfree(ctx->decomp_stream.workspace);<br /> }<br /> </pre></td><td width="32" rowspan="2" class="c" valign="top"><img src="/images/icornerr.gif" width="32" height="32" alt="\" /></td></tr><tr><td align="right" valign="bottom"> 聽 </td></tr><tr><td align="right" valign="bottom">聽</td><td class="c" valign="bottom" style="padding-bottom: 0px"><img src="/images/bcornerl.gif" width="32" height="32" alt="\" /></td><td class="c">聽</td><td class="c" valign="bottom" style="padding-bottom: 0px"><img src="/images/bcornerr.gif" width="32" height="32" alt="/" /></td></tr><tr><td align="right" valign="top" colspan="2"> 聽 </td><td class="lm">Last update: 2005-04-06 13:31 聽聽 [from the cache]<br />漏2003-2020 <a href="http://blog.jasper.es/"><span itemprop="editor">Jasper Spaans</span></a>|hosted at <a href="https://www.digitalocean.com/?refcode=9a8e99d24cf9">Digital Ocean</a> and my Meterkast|<a href="http://blog.jasper.es/categories.html#lkml-ref">Read the blog</a></td><td>聽</td></tr></table><script language="javascript" src="/js/styleswitcher.js" type="text/javascript"></script></body></html>