From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Ball Subject: Re: mmc_test: fix performance tests that go over max_blk_count Date: Sun, 12 Sep 2010 00:47:07 +0100 Message-ID: <20100911234706.GA14042@void.printf.net> References: <4C89ED69.8000301@nokia.com> <20100910203957.GA2966@void.printf.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from void.printf.net ([89.145.121.20]:45615 "EHLO void.printf.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752886Ab0IKXrQ (ORCPT ); Sat, 11 Sep 2010 19:47:16 -0400 Content-Disposition: inline In-Reply-To: <20100910203957.GA2966@void.printf.net> Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Adrian Hunter , Andrew Morton , "linux-mmc@vger.kernel.org" Hi Adrian, On Fri, Sep 10, 2010 at 09:39:57PM +0100, Chris Ball wrote: > My test machine (Dell m6300, Sandisk Extreme III 4GB) no longer BUG()s, > but it's still failing all of the transfer tests: > > mmc0: Test case 22. Multi-block highmem read... > mmc0: Result: OK > mmc0: Test case 23. Best-case read performance... > mmc0: Result: Prepare stage failed! (-12) Tracked this down a little farther. mmc_test_alloc_mem() is bailing out here: if (mem->cnt >= max_segs) { if (page_cnt < min_page_cnt) goto out_free; break; } with vars mem->cnt=1, max_segs=1, page_cnt=0, min_page_cnt=16. Thanks, -- Chris Ball One Laptop Per Child