From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tony Lindgren Subject: Re: DSS2 failure with current linux-omap head Date: Tue, 25 May 2010 15:46:31 -0700 Message-ID: <20100525224630.GK9645@atomide.com> References: <1274429232.4246.65.camel@tubuntu.research.nokia.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mho-01-ewr.mailhop.org ([204.13.248.71]:62375 "EHLO mho-01-ewr.mailhop.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750800Ab0EYWqe (ORCPT ); Tue, 25 May 2010 18:46:34 -0400 Content-Disposition: inline In-Reply-To: <1274429232.4246.65.camel@tubuntu.research.nokia.com> Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: Tomi Valkeinen Cc: ext Steve Sakoman , "linux-omap@vger.kernel.org" * Tomi Valkeinen [100521 01:03]: > On Thu, 2010-05-20 at 00:33 +0200, ext Steve Sakoman wrote: > > I did a quick test build of the current linux-omap head and get a > > failure very early on in the boot process in > > drivers/video/omap2/vram.c code: > > > > Illegal SDRAM size for VRAM > > > > It is generated by the following code: > > > > bdata = NODE_DATA(0)->bdata; > > sdram_start = bdata->node_min_pfn << PAGE_SHIFT; > > sdram_size = (bdata->node_low_pfn << PAGE_SHIFT) - sdram_start; > > > > if (paddr) { > > if ((paddr & ~PAGE_MASK) || paddr < sdram_start || > > paddr + size > sdram_start + sdram_size) { > > pr_err("Illegal SDRAM region for VRAM\n"); > > return; > > } > > > > if (reserve_bootmem(paddr, size, BOOTMEM_EXCLUSIVE) < 0) { > > pr_err("FB: failed to reserve VRAM\n"); > > return; > > } > > } else { > > if (size > sdram_size) { > > pr_err("Illegal SDRAM size for VRAM\n"); > > return; > > } > > > > paddr = virt_to_phys(alloc_bootmem_pages(size)); > > BUG_ON(paddr & ~PAGE_MASK); > > } > > > > > > I modified the error pr_error to display the two values that are > > compared to generate this error: > > > > Illegal SDRAM size for VRAM: size=0xc00000 sdram_size=0x0 > > > > So it appears that the method used to get the sdram size no longer > > works since it returns a size of 0! > > I copied the method from the older omapfb driver, and I have to say that > I've never looked at what it actually does. Perhaps there's a proper way > to get the RAM starts and sizes somewhere... FYI, Russell has several interesting omapfb patches in his lmb branch. Please take a look at that, might make your life easier :) Regards, Tony