From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sergei Shtylyov Subject: Re: [PATCH 2/4] hpt366: UltraDMA filter for SATA cards (take 3) Date: Sat, 01 Sep 2007 00:08:05 +0400 Message-ID: <46D87525.6060303@ru.mvista.com> References: <200708252328.03826.sshtylyov@ru.mvista.com> <200708271922.23349.bzolnier@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from h155.mvista.com ([63.81.120.155]:44941 "EHLO imap.sh.mvista.com" rhost-flags-OK-FAIL-OK-FAIL) by vger.kernel.org with ESMTP id S1752091AbXHaUFR (ORCPT ); Fri, 31 Aug 2007 16:05:17 -0400 In-Reply-To: <200708271922.23349.bzolnier@gmail.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Bartlomiej Zolnierkiewicz Cc: linux-ide@vger.kernel.org, linux-kernel@vger.kernel.org Bartlomiej Zolnierkiewicz wrote: >>The Marvell bridge chips used on HighPoint SATA cards do not seem to support >>the UltraDMA modes 1, 2, and 3 as well as any MWDMA modes, so the driver needs >>to account for this in the udma_filter() method. In order to achieve that, do >>the following changes: >>- install the method for all chips, not only HPT36x/370 and impove the code >> formatting by killing the extra tabs while at it; >>- add to the end of the 'switch' statement in the method cases for HPT372[AN] >> and HPT374 chips upon which the known SATA cards are based; >>- use hwif->ultra_mask as a default mask for the ide_dma_filter() method to >> behave correctly; >>- move the HPT370[A] cases below the HPT36x case for consistency... >>Signed-off-by: Sergei Shtylyov >>--- >>Argh! I've managed to put = instead of &= here and there, so please disregard >>the take #2... :-/ > Not to mention that there was already take #2 on Aug 19 2007 > (the version of the patch which is currently in IDE quilt tree)... I've just changed my mind about which series it'd beliong too -- propply need to post [0/n] message before the series.... >>This version doesn't use explicit UltraDMA masks, so converting them to the >>ATA_UDMA* is left for another, global patch. This patch against the current > I have already other patches which are based on the previous version of the > patch and I don't find the idea of re-doing them especially tempting... Too bad. :-) I don't find the idea if the open coded masks depnding on HPT3*_ALLOW_ATA* for a catch all udma_filter() method especially tempting to. :-) >>Linus' tree and unfortunately I was able to only compile test it since that >>tree gives MODPOST warning and dies early on bootup. >>Index: linux-2.6/drivers/ide/pci/hpt366.c >>=================================================================== >>--- linux-2.6.orig/drivers/ide/pci/hpt366.c >>+++ linux-2.6/drivers/ide/pci/hpt366.c >>@@ -1,5 +1,5 @@ >> /* >>- * linux/drivers/ide/pci/hpt366.c Version 1.11 Aug 11, 2007 >>+ * linux/drivers/ide/pci/hpt366.c Version 1.12 Aug 25, 2007 >> * >> * Copyright (C) 1999-2003 Andre Hedrick >> * Portions Copyright (C) 2001 Sun Microsystems, Inc. >>@@ -114,6 +114,7 @@ >> * unify HPT36x/37x timing setup code and the speedproc handlers by joining >> * the register setting lists into the table indexed by the clock selected >> * - set the correct hwif->ultra_mask for each individual chip >>+ * - add UltraDMA mode filtering for the HPT37[24] based SATA cards >> * Sergei Shtylyov, or >> */ >> >>@@ -524,36 +525,38 @@ static int check_in_drive_list(ide_drive > (the real) take #2 also updated hpt3xx_udma_filter() comment I've probably dropped that part in anticipation of adding mdma_filter... I agree -- future has always come. B-) >> static u8 hpt3xx_udma_filter(ide_drive_t *drive) >> { >>- struct hpt_info *info = pci_get_drvdata(HWIF(drive)->pci_dev); >>- u8 mask; >>+ ide_hwif_t *hwif = HWIF(drive); >>+ struct hpt_info *info = pci_get_drvdata(hwif->pci_dev); >>+ u8 mask = hwif->ultra_mask; >> >> switch (info->chip_type) { >>- case HPT370A: >>- if (!HPT370_ALLOW_ATA100_5 || >>- check_in_drive_list(drive, bad_ata100_5)) >>- return 0x1f; >>- else >>- return 0x3f; >>- case HPT370: >>- if (!HPT370_ALLOW_ATA100_5 || >>- check_in_drive_list(drive, bad_ata100_5)) >>- mask = 0x1f; >>- else >>- mask = 0x3f; >>- break; >> case HPT36x: >>- if (!HPT366_ALLOW_ATA66_4 || >>+ if (HPT366_ALLOW_ATA66_4 && >> check_in_drive_list(drive, bad_ata66_4)) >>- mask = 0x0f; >>- else >>- mask = 0x1f; >>+ mask &= ~0x10; >> >>- if (!HPT366_ALLOW_ATA66_3 || >>+ if (HPT366_ALLOW_ATA66_3 && >> check_in_drive_list(drive, bad_ata66_3)) >>- mask = 0x07; >>+ mask &= ~0x08; >> break; >>+ case HPT370 : >>+ case HPT370A: >>+ if (HPT370_ALLOW_ATA100_5 && >>+ check_in_drive_list(drive, bad_ata100_5)) >>+ mask &= ~0x20; >>+ >>+ if (info->chip_type == HPT370A) >>+ return mask; >>+ break; >>+ case HPT372 : >>+ case HPT372A: >>+ case HPT372N: >>+ case HPT374 : >>+ if (ide_dev_is_sata(drive->id)) >>+ mask &= ~0x0e; >>+ /* Fall thru */ >> default: >>- return 0x7f; >>+ return mask; >> } > I really don't see the advantage of "mask &=" over the previous Considered the smaller code? ;-) > code ("mask = ATA_UDMA*") which was just more readable IMO. Like I said, I don't want to be tied by HPT3*_ALLOW_ATA* here and there (BTW, this approach was in revision #0 which I've never published :-). > I'm staying with (the real) take #2 of this patch for now. I'll respin RSN... but not take #2, alas. B-) > Bart MBR, Sergei