LKML: Evgeniy Polyakov: Re: [patch 2.6.12-rc1-mm4] fork

<?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: Evgeniy Polyakov: Re: [patch 2.6.12-rc1-mm4] fork_connector: add a fork connector</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 Evgeniy Polyakov" href="/groupie.php?aid=3643" /><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/2005/4"> [Apr]</a> 聽 <a class="nb" href="/lkml/2005/4/8"> [8]</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/2005/4/8/96" onclick="this.href='/lkml/headers'+'/2005/4/8/96';">[headers]</a>聽 <a href="/lkml/bounce/2005/4/8/96">[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/2005/3/31/137">First message in thread</a></li><li><a href="/lkml/2005/4/7/259">Jay Lan</a><ul><li><a href="/lkml/2005/4/7/261">Jay Lan</a><ul><li class="origin"><a href="/lkml/2005/4/8/102">Evgeniy Polyakov</a><ul><li><a href="/lkml/2005/4/8/102">Evgeniy Polyakov</a><ul><li><a href="/lkml/2005/4/8/256">Jay Lan</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/2005/4/8/96/1">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">Subject</td><td class="rp" itemprop="name">Re: [patch 2.6.12-rc1-mm4] fork_connector: add a fork connector</td></tr><tr><td class="lp">From</td><td class="rp" itemprop="author">Evgeniy Polyakov <></td></tr><tr><td class="lp">Date</td><td class="rp" itemprop="datePublished">Fri, 08 Apr 2005 14:24:00 +0400</td></tr></table></td><td></td></tr></table><pre itemprop="articleBody">On Thu, 2005-04-07 at 15:47 -0700, Jay Lan wrote:<br />> BTW, when it happened last time, my program listening to the socket<br />> complained about duplicate messages received.<br />> <br />> Unmatched seq. Rcvd=1824062, expected=1824061 <===<br />> Unmatched seq. Rcvd=1824062, expected=1824063 <===<br />> Unmatched seq. Rcvd=1824348, expected=1824307<br />> <br />> When my program received 1824062 while expecting 1824061<br />> it adjusted itself to expect the next one being 1824063. But instead<br />> the message of 1824062 arrived the second time.<br /><br />Thank you for your report.<br /><br />Could you reproduce a bug with the attached patch?<br /><br />> - jay<br /><br /><br />* looking for johnpol@2ka.mipt.ru-2004/connector--main--0--patch-38 to compare with<br />* comparing to johnpol@2ka.mipt.ru-2004/connector--main--0--patch-38<br />M connector.c<br /><br />* modified files<br /><br />--- orig/drivers/connector/connector.c<br />+++ mod/drivers/connector/connector.c<br />@@ -121,7 +121,7 @@<br /> NETLINK_CB(skb).dst_groups = groups;<br /> <br /> uskb = skb_clone(skb, GFP_ATOMIC);<br />- if (uskb)<br />+ if (uskb && 0)<br /> netlink_unicast(dev->nls, uskb, 0, 0);<br /> <br /> netlink_broadcast(dev->nls, skb, 0, groups, GFP_ATOMIC);<br />@@ -158,7 +158,7 @@<br /> }<br /> spin_unlock_bh(&dev->cbdev->queue_lock);<br /> <br />- return found;<br />+ return (found)?0:-ENODEV;<br /> }<br /> <br /> /*<br />@@ -181,7 +181,6 @@<br /> "requested msg->len=%u[%u], nlh->nlmsg_len=%u, skb->len=%u.\n",<br /> msg->len, NLMSG_SPACE(msg->len + sizeof(*msg)),<br /> nlh->nlmsg_len, skb->len);<br />- kfree_skb(skb);<br /> return -EINVAL;<br /> }<br /> #if 0<br />@@ -215,17 +214,18 @@<br /> skb->len, skb->data_len, skb->truesize, skb->protocol,<br /> skb_cloned(skb), skb_shared(skb));<br /> #endif<br />- while (skb->len >= NLMSG_SPACE(0)) {<br />+ if (skb->len >= NLMSG_SPACE(0)) {<br /> nlh = (struct nlmsghdr *)skb->data;<br />+<br /> if (nlh->nlmsg_len < sizeof(struct cn_msg) ||<br /> skb->len < nlh->nlmsg_len ||<br /> nlh->nlmsg_len > CONNECTOR_MAX_MSG_SIZE) {<br />-#if 0<br />+#if 1<br /> printk(KERN_INFO "nlmsg_len=%u, sizeof(*nlh)=%u\n",<br /> nlh->nlmsg_len, sizeof(*nlh));<br /> #endif<br /> kfree_skb(skb);<br />- break;<br />+ goto out;<br /> }<br /> <br /> len = NLMSG_ALIGN(nlh->nlmsg_len);<br />@@ -233,22 +233,11 @@<br /> len = skb->len;<br /> <br /> err = __cn_rx_skb(skb, nlh);<br />- if (err) {<br />-#if 0<br />- if (err < 0 && (nlh->nlmsg_flags & NLM_F_ACK))<br />- netlink_ack(skb, nlh, -err);<br />-#endif<br />- break;<br />- } else {<br />-#if 0<br />- if (nlh->nlmsg_flags & NLM_F_ACK)<br />- netlink_ack(skb, nlh, 0);<br />-#endif<br />- break;<br />- }<br />- skb_pull(skb, len);<br />+ if (err < 0)<br />+ kfree_skb(skb);<br /> }<br />- <br />+<br />+out:<br /> kfree_skb(__skb);<br /> }<br /> <br /><br /><br /><br /><br />-- <br /> Evgeniy Polyakov<br /><br />Crash is better than data corruption -- Arthur Grabowski<br />[unhandled content-type:application/pgp-signature]</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: 2005-04-08 12:32 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>

CINXE.COM

LKML: Evgeniy Polyakov: Re: [patch 2.6.12-rc1-mm4] fork_connector: add a fork connector