From mboxrd@z Thu Jan 1 00:00:00 1970 From: Walt Holman Subject: Re: Strange network timeouts w/ 2.6.30.5 Date: Fri, 21 Aug 2009 10:45:04 -0500 (CDT) Message-ID: <987943678.21250869504786.JavaMail.root@mail.holmansrus.com> References: <1489093361.01250869377567.JavaMail.root@mail.holmansrus.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Cc: David Miller , linux-kernel@vger.kernel.org, netdev@vger.kernel.org To: Krzysztof Halasa Return-path: Received: from static-76-164-182-57.apid.com ([76.164.182.57]:38666 "EHLO waltsputer.localdomain" rhost-flags-OK-FAIL-OK-FAIL) by vger.kernel.org with ESMTP id S1754425AbZHUPyN (ORCPT ); Fri, 21 Aug 2009 11:54:13 -0400 In-Reply-To: <1489093361.01250869377567.JavaMail.root@mail.holmansrus.com> Sender: netdev-owner@vger.kernel.org List-ID: ----- "Krzysztof Halasa" wrote: > Walt Holman writes: > > > Would something like passing a mem=xx cmdline on x86_64 be > sufficient to test this? > > I think so, though I admit I don't remember using this personally > since > the introduction of e820 RAM mapping support. > Dmesg will show if the memory is limited. For the swiotlb to > effectively > disable no RAM > 0x100000000 may be used. > > But this test isn't IMHO terribly important at this point - the > driver > makes invalid use of the DMA API and it has to change. The test could > only explain _how_ exactly does it fail, we already know _why_ it > does. > -- > Krzysztof Halasa Just got a chance to test this for confirmation. When limiting the RAM to 2GB, the patched driver appears to work OK. -Walt