From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeff Garzik Subject: Re: [PATCH #upstream, v3] ahci: Implement SATA AHCI FIS-based switching support Date: Thu, 03 Dec 2009 12:42:05 -0500 Message-ID: <4B17F86D.1080805@pobox.com> References: <1252402105.3514.7.camel@zm-desktop> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mail-yx0-f187.google.com ([209.85.210.187]:52725 "EHLO mail-yx0-f187.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751575AbZLCRmA (ORCPT ); Thu, 3 Dec 2009 12:42:00 -0500 Received: by yxe17 with SMTP id 17so1338180yxe.33 for ; Thu, 03 Dec 2009 09:42:06 -0800 (PST) In-Reply-To: <1252402105.3514.7.camel@zm-desktop> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: shane.huang@amd.com Cc: tj@kernel.org, linux-ide@vger.kernel.org On 09/08/2009 05:28 AM, Shane Huang wrote: > Implement SATA AHCI FIS-based switching support. > > Tested on AMD internal reference board. > > Signed-off-by: Shane Huang > Cc: Tejun Heo > --- > Changes from the last version(v2): > * Move fbs_need_dec from struct ahci_port_priv to ahci_error_intr() as > a local variable. > * Add BUG_ON(!pp->fbs_enabled) for ahci_fbs_dec_intr(). > * Refine the retry loop in ahci_fbs_dec_intr(). > * Refine the code to determine failed active link with ap->pmp_link in > ahci_error_intr(). > * WARN_ON() in ahci_port_intr() can storm the machine with consecutive > stack dumps, replace it with WARN_ON_ONCE(). > * Refine ahci_enable_fbs()/ahci_disable_fbs(), and discard the message > for controllers which do not support FBS. > * Add a warning message in ahci_port_start() for those weird controllers > which support CAP.FBSS but do not support PxCMD.FBSCP. > * Several small corrections for better coding style. > > drivers/ata/ahci.c | 225 +++++++++++++++++++++++++++++++++++++++++++++++----- > 1 files changed, 206 insertions(+), 19 deletions(-) OK, I'm ready to get back to this... Would you please resubmit this patch, rediff'd for the current upstream kernel? I will queue it for kernel 2.6.34, because it needs plenty of testing -- both for FBS-enabled AHCI, and to be sure we don't break non-FBS AHCI. Jeff