From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bartlomiej Zolnierkiewicz Subject: Re: [git patches] libata fixes Date: Sun, 24 Feb 2008 18:21:14 +0100 Message-ID: <200802241821.15052.bzolnier@gmail.com> References: <20080224053508.GA17530@havoc.gtf.org> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Return-path: Received: from ug-out-1314.google.com ([66.249.92.174]:50487 "EHLO ug-out-1314.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750924AbYBXRHm convert rfc822-to-8bit (ORCPT ); Sun, 24 Feb 2008 12:07:42 -0500 Received: by ug-out-1314.google.com with SMTP id z38so467459ugc.16 for ; Sun, 24 Feb 2008 09:07:40 -0800 (PST) In-Reply-To: <20080224053508.GA17530@havoc.gtf.org> Content-Disposition: inline Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Jeff Garzik Cc: Andrew Morton , Linus Torvalds , linux-ide@vger.kernel.org, LKML , Alan Cox On Sunday 24 February 2008, Jeff Garzik wrote: [...] > Alan Cox (1): > pata_atiixp: Use 255 sector limit AHCI needs sorting too but this deals with the old interface Signed-off-by: Alan Cox Signed-off-by: Jeff Garzik [...] >> diff --git a/drivers/ata/pata_atiixp.c b/drivers/ata/pata_atiixp.c > index 9623f52..408bdc1 100644 > --- a/drivers/ata/pata_atiixp.c > +++ b/drivers/ata/pata_atiixp.c > @@ -227,7 +227,7 @@ static struct scsi_host_template atiixp_sht = { > .queuecommand = ata_scsi_queuecmd, > .can_queue = ATA_DEF_QUEUE, > .this_id = ATA_SHT_THIS_ID, > - .sg_tablesize = LIBATA_MAX_PRD, > + .sg_tablesize = LIBATA_DUMB_MAX_PRD, > .cmd_per_lun = ATA_SHT_CMD_PER_LUN, > .emulated = ATA_SHT_EMULATED, > .use_clustering = ATA_SHT_USE_CLUSTERING, > @@ -259,7 +259,7 @@ static struct ata_port_operations atiixp_port_ops = { > .bmdma_stop = atiixp_bmdma_stop, > .bmdma_status = ata_bmdma_status, > > - .qc_prep = ata_qc_prep, > + .qc_prep = ata_dumb_qc_prep, > .qc_issue = ata_qc_issue_prot, > > .data_xfer = ata_data_xfer, This patch doesn't enforce "255 sector limit" but instead: - limits max PRD table size from 128 entries to 64 - enables quirk needed for some chipsets to split 64K PRD table entry on two 32K ones (these chipsets choke on size == 0x0000) >>From the patch description it can't be told whether the patch itself is correct and only the patch description is bogus... Alan/Jeff: please elaborate on what the patch is trying to achieve. Thanks, Bart