From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bjorn Helgaas Subject: Re: [PATCH 1/3] ACPI: Ignore the upper bits of SystemIO addresses Date: Wed, 19 May 2010 10:18:46 -0600 Message-ID: <201005191018.46799.bjorn.helgaas@hp.com> References: <1274283791-3380-1-git-send-email-mjg@redhat.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Return-path: Received: from g4t0014.houston.hp.com ([15.201.24.17]:29444 "EHLO g4t0014.houston.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751522Ab0ESQS6 (ORCPT ); Wed, 19 May 2010 12:18:58 -0400 In-Reply-To: <1274283791-3380-1-git-send-email-mjg@redhat.com> Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: Matthew Garrett Cc: linux-acpi@vger.kernel.org, robert.moore@intel.com, lenb@kernel.org On Wednesday, May 19, 2010 09:43:09 am Matthew Garrett wrote: > Various machines (https://bugzilla.redhat.com/show_bug.cgi?id=585756 for > instance) contain SystemIO spaces with addresses > 16 bits. acpica throws > an error for this, while Windows silently ignores the upper 16 bits and > carries on happily. Provide support for the latter behaviour for bug > compatibility. > > + if (acpi_gbl_ignore_high_ioport_bits) > + /* Windows only uses the lower 16 bits of an address. > + Emulate that */ > + address &= 0xffff; I think this is a good idea. But it makes me a little bit nervous to change addresses supplied by the firmware without any user-visible indication at all. Is it worth doing a WARN_ONCE() sort of thing when we truncate? I know you experimented quite a bit to confirm that Windows does this sort of masking. Do you have any notes about that experimentation that would be useful to add to the bugzilla? Bjorn