From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [212.234.111.251] (helo=mail.telecom-design.com) by bombadil.infradead.org with esmtp (Exim 4.68 #1 (Red Hat Linux)) id 1KeGus-00025i-BQ for linux-mtd@lists.infradead.org; Fri, 12 Sep 2008 22:15:02 +0000 Received: (from uucp@localhost) by mail.telecom-design.com id m8CM3USi027859 for ; Sat, 13 Sep 2008 00:03:30 +0200 Subject: 8/16/32 bit JEDEC probe From: Michel Stempin To: linux-mtd@lists.infradead.org Content-Type: text/plain Content-Transfer-Encoding: 7bit Date: Sat, 13 Sep 2008 00:03:22 +0200 Message-Id: <1221257002.30016.13.camel@Mikeul> Mime-Version: 1.0 List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Hi list, I have an interesting case to submit. I have a SystemBase ARM9 (AT91SAM9260) Eddy-CPU-V2 that can be populated with either a 32MB SST 39VF3201 (Mfr: 0x00bf, Id: 0x235b) or a 64MB EON EN29LV640B (Mfr: 0x007f, Id: 0x22cb) part. Mine has the SST part. As none of these chips were defined in jedec_probe.c, I defined them there. However, when the probe is done, an SST 49LF080A (Mfr: 0x00bf, Id: 0x005b) is found instead of my SST 39VF3201... What is remarkable is that both chips have the same LSB Id! My conclusion is that the less discriminating 8-bit device detection has precedence over the more discriminating 16-bit device. Shouldn't this be the other way around (ie. first detect 32, then 16, then 8 bit devices)? If I comment the SST 49LF080A definition, everything works as expected. Best regards, -- Michel Stempin