<div dir="ltr"><div>Hi,<br><br>For a chance to get this included in 4.1 (although it'd be off by default) can you test the attached patch with real traffic and report back?<br><br></div><div>Please note this patch has the accept_filter feature enabled so there is no need to do anything besides ensuring it works as expected.<br><br></div><div>Thanks.</div><div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Fri, Sep 4, 2015 at 3:12 PM, Federico Schwindt <span dir="ltr"><<a href="mailto:fgsch@lodoss.net" target="_blank">fgsch@lodoss.net</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote"><span class="">On Fri, Sep 4, 2015 at 2:26 PM, Rafael Zalamena <span dir="ltr"><<a href="mailto:rafaelfz@taghos.com.br" target="_blank">rafaelfz@taghos.com.br</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Em Fri, 4 Sep 2015 13:42:40 +0100<br>
Federico Schwindt <<a href="mailto:fgsch@lodoss.net" target="_blank">fgsch@lodoss.net</a>> escreveu:<br>
<span><br>
> Diff aside looking at the code my impression is that the VTCP_filter_http()<br>
> function is meant to be compiled in always so erroring out if it's not<br>
> supported might be wrong here, or at least not when errno is EOPNOTSUPP/<br>
> multiple times.<br>
<br>
</span>People without it must have a way to find it out. The old way was not<br>
including any code and avoid it at all, but then we had a code that behaves<br>
differently according to the system which seems better than not giving<br>
any clues.<br></blockquote><div><br></div></span><div>The older code will always include VTCP_filter_http() for Linux, it was not calling it though.<br></div><span class=""><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<span><br>
> On 4 Sep 2015 1:35 pm, "Federico Schwindt" <<a href="mailto:fgsch@lodoss.net" target="_blank">fgsch@lodoss.net</a>> wrote:<br>
><br>
> > Did you see my second diff? :)<br>
</span>> --- SNIPPED ---<br>
<br>
Yes, I read your second diff.<br>
<br>
I agree with you that it would be better to kill the accept_filter<br>
param ifdef guards to keep it avaliable in every system, but I don't<br>
think it's a good idea to break the old behavior because people might<br>
be expecting it. That's why I suggested the TCP_DEFER_ACCEPT detection<br>
to make the code more portable instead of just looking for __linux.<br></blockquote><div><br></div></span>My second diff doesn't change any behaviour so nothing will break unless you enable it.<br></div><br></div></div>
</blockquote></div><br></div>