From mboxrd@z Thu Jan 1 00:00:00 1970 From: Seth Forshee Subject: AUTO_EXTVOLT on OMAP2430 Date: Mon, 20 Oct 2008 00:07:26 -0500 Message-ID: <20081020050726.GA12891@ubuntu-workstation> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from wx-out-0506.google.com ([66.249.82.230]:27026 "EHLO wx-out-0506.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751735AbYJTFHt (ORCPT ); Mon, 20 Oct 2008 01:07:49 -0400 Received: by wx-out-0506.google.com with SMTP id h27so744794wxd.4 for ; Sun, 19 Oct 2008 22:07:45 -0700 (PDT) Content-Disposition: inline Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: linux-omap@vger.kernel.org I'm working with what basically amounts to an older version of the linux-omap kernel with substantial modifications. I found that in some situations I was getting about 4 mA more current draw during suspend than at other times. This was ultimately being caused by some of the code that runs in SRAM masking off the PRCM_VOLTCTRL[AUTO_EXTVOLT] bit. I compared to tips in the linux-omap repo for comparison and found the situation is the same -- the bit is set during initialization, but will be cleared by the SRAM code that configures the SDRC. >>From reading the TRM it seems that this bit should only affect the retention and off states, so I don't understand why it needs to be masked during SDRC configuration. Is there any danger in changing the mask to leave this bit unchanged? Or would it be safer to set this bit before suspend and clear it after? Thanks, Seth