From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from webapps.arcom.com ([194.200.159.168]) by canuck.infradead.org with esmtp (Exim 4.52 #1 (Red Hat Linux)) id 1ENAD0-0002Bf-MS for linux-mtd@lists.infradead.org; Wed, 05 Oct 2005 10:25:45 -0400 Message-ID: <4343E24E.8020507@arcom.com> Date: Wed, 05 Oct 2005 15:25:18 +0100 From: David Vrabel MIME-Version: 1.0 To: Jan Pedersen References: <20051004204918.E02AE37646A@rocket.glaze.se> In-Reply-To: <20051004204918.E02AE37646A@rocket.glaze.se> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: linux-mtd@lists.infradead.org Subject: Re: [patch] cfi: remove warning message on expected behaivor List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Jan Pedersen wrote: > When an erase operation is in progress, the DQ5 (data bit 5 / exceeded > timing limit) pin on the flash chips may raise just before operation > complete is detected. This is expected behaivor because when the erase is > complete, DQ5 switches from 'exceeded timing limit' to 'data bit 5' which > therefore might be read as '1' just before operation complete is detected. > This fix is well tested. If you look at the current cfi_cmdset_0002.c (in 2.6) you'll notice that all it does it check for toggling bits. The software timeout handles failed erases and writes rather than checking DQ5. Trying to handle the various error bits was deemed too tricky/complex to get right for interleaved chips. David Vrabel -- David Vrabel, Design Engineer Arcom, Clifton Road Tel: +44 (0)1223 411200 ext. 3233 Cambridge CB1 7EA, UK Web: http://www.arcom.com/