linux-ide.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Tejun Heo <htejun@gmail.com>
To: jeff@garzik.org, linux-ide@vger.kernel.org
Cc: Tejun Heo <htejun@gmail.com>, Alan Cox <alan@lxorguk.ukuu.org.uk>
Subject: [PATCH 07/12] libata: fix ata_acpi_gtm_xfermask()
Date: Tue,  6 Nov 2007 14:39:05 +0900	[thread overview]
Message-ID: <1194327551238-git-send-email-htejun@gmail.com> (raw)
In-Reply-To: <1194327550227-git-send-email-htejun@gmail.com>

ata_acpi_gtm_xfermask() as separated out from pacpi_discover_modes()
has various bugs.  Fix them.

* The wrong comparison operator is used when finding for matching
  cycle resulting totally bogus result.

* With the comparion operator fixed, boundary condtion handling is
  clumsy.

* Setting of any DMA mask bit set all bits in PIO mask.

* MWDMA and UDMA blocks are swapped.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
---
 drivers/ata/libata-acpi.c |   42 +++++++++++++++++++++---------------------
 1 files changed, 21 insertions(+), 21 deletions(-)

diff --git a/drivers/ata/libata-acpi.c b/drivers/ata/libata-acpi.c
index 5ffa542..06d9961 100644
--- a/drivers/ata/libata-acpi.c
+++ b/drivers/ata/libata-acpi.c
@@ -456,49 +456,49 @@ EXPORT_SYMBOL_GPL(ata_acpi_udma_cycle);
 unsigned int ata_acpi_gtm_xfermask(struct ata_device *dev,
 				   const struct ata_acpi_gtm *gtm)
 {
+	unsigned int pio_mask = 0, mwdma_mask = 0, udma_mask = 0;
 	int unit, i;
 	u32 t;
-	unsigned long mask = (0x7f << ATA_SHIFT_UDMA) | (0x7 << ATA_SHIFT_MWDMA) | (0x1F << ATA_SHIFT_PIO);
 
 	/* we always use the 0 slot for crap hardware */
 	unit = dev->devno;
 	if (!(gtm->flags & 0x10))
 		unit = 0;
 
+	/* Values larger than the longest cycle results in 0 mask
+	 * while values equal to smaller than the shortest cycle
+	 * results in mask which includes all supported modes.
+	 * Disabled transfer method has the value of 0xffffffff which
+	 * will always result in 0 mask.
+	 */
+
 	/* start by scanning for PIO modes */
-	for (i = 0; i < 7; i++) {
-		t = gtm->drive[unit].pio;
-		if (t <= ata_acpi_pio_cycle[i]) {
-			mask |= (2 << (ATA_SHIFT_PIO + i)) - 1;
+	t = gtm->drive[unit].pio;
+	for (i = 0; i < ARRAY_SIZE(ata_acpi_pio_cycle); i++)
+		if (t > ata_acpi_pio_cycle[i])
 			break;
-		}
-	}
+	pio_mask = (1 << i) - 1;
 
 	/* See if we have MWDMA or UDMA data. We don't bother with
 	 * MWDMA if UDMA is available as this means the BIOS set UDMA
 	 * and our error changedown if it works is UDMA to PIO anyway.
 	 */
-	if (gtm->flags & (1 << (2 * unit))) {
+	t = gtm->drive[unit].dma;
+	if (!(gtm->flags & (1 << (2 * unit)))) {
 		/* MWDMA */
-		for (i = 0; i < 5; i++) {
-			t = gtm->drive[unit].dma;
-			if (t <= ata_acpi_mwdma_cycle[i]) {
-				mask |= (2 << (ATA_SHIFT_MWDMA + i)) - 1;
+		for (i = 0; i < ARRAY_SIZE(ata_acpi_mwdma_cycle); i++)
+			if (t > ata_acpi_mwdma_cycle[i])
 				break;
-			}
-		}
+		mwdma_mask = (1 << i) - 1;
 	} else {
 		/* UDMA */
-		for (i = 0; i < 7; i++) {
-			t = gtm->drive[unit].dma;
-			if (t <= ata_acpi_udma_cycle[i]) {
-				mask |= (2 << (ATA_SHIFT_UDMA + i)) - 1;
+		for (i = 0; i < ARRAY_SIZE(ata_acpi_udma_cycle); i++)
+			if (t > ata_acpi_udma_cycle[i])
 				break;
-			}
-		}
+		udma_mask = (1 << i) - 1;
 	}
 
-	return mask;
+	return ata_pack_xfermask(pio_mask, mwdma_mask, udma_mask);
 }
 EXPORT_SYMBOL_GPL(ata_acpi_gtm_xfermask);
 
-- 
1.5.2.4


  parent reply	other threads:[~2007-11-06  5:39 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-11-06  5:38 [PATCHSET] libata: update timing and fix pata_amd transfer mode selection Tejun Heo
2007-11-06  5:38 ` [PATCH 01/12] ata_generic: unindent loop in generic_set_mode() Tejun Heo
2007-11-06  5:39 ` [PATCH 02/12] libata: export xfermode / PATA timing related functions Tejun Heo
2007-11-06  5:39 ` [PATCH 03/12] libata: clean up xfermode / PATA timing related stuff Tejun Heo
2007-11-06  5:39 ` [PATCH 04/12] libata: kill ata_id_to_dma_mode() Tejun Heo
2007-11-06 10:49   ` Alan Cox
2007-11-06 11:21     ` Tejun Heo
2007-11-24  1:13   ` Jeff Garzik
2007-11-06  5:39 ` [PATCH 05/12] libata: xfer_mask is unsigned int not unsigned long Tejun Heo
2007-11-06 10:51   ` Alan Cox
2007-11-06 10:59     ` Tejun Heo
2007-11-06 17:53       ` Jeff Garzik
2007-11-07  1:12         ` Tejun Heo
2007-11-24  1:13   ` Jeff Garzik
2007-11-24  1:13     ` Jeff Garzik
2007-11-27  8:42       ` Tejun Heo
2007-11-06  5:39 ` [PATCH 06/12] libata: separate out ata_acpi_gtm_xfermask() from pacpi_discover_modes() Tejun Heo
2007-11-06 10:54   ` Alan Cox
2007-11-06 11:00     ` Tejun Heo
2007-11-24  1:14   ` Jeff Garzik
2007-11-06  5:39 ` Tejun Heo [this message]
2007-11-06 10:55   ` [PATCH 07/12] libata: fix ata_acpi_gtm_xfermask() Alan Cox
2007-11-24  1:16   ` Jeff Garzik
2007-11-27  8:40     ` Tejun Heo
2007-11-06  5:39 ` [PATCH 08/12] libata: implement ata_timing_cycle2mode() and use it in libata-acpi and pata_acpi Tejun Heo
2007-11-06 10:59   ` Alan Cox
2007-11-06 11:09     ` Tejun Heo
2007-11-24  1:17   ` Jeff Garzik
2007-11-06  5:39 ` [PATCH 09/12] libata: implement ata_acpi_init_gtm() Tejun Heo
2007-11-06  5:39 ` [PATCH 10/12] libata: reimplement ata_acpi_cbl_80wire() using ata_acpi_gtm_xfermask() Tejun Heo
2007-11-06  5:39 ` [PATCH 11/12] libata: add ATA_CBL_PATA_IGN Tejun Heo
2007-11-06 10:59   ` Alan Cox
2007-11-06 11:02     ` Tejun Heo
2007-11-06 11:25       ` Alan Cox
2007-11-06  5:39 ` [PATCH 12/12] pata_amd: update mode selection for NV PATAs Tejun Heo
2007-11-06 10:59   ` Alan Cox
2007-11-23  1:08 ` [PATCHSET] libata: update timing and fix pata_amd transfer mode selection Tejun Heo
2007-11-24  1:18 ` Jeff Garzik
  -- strict thread matches above, loose matches on Subject: below --
2007-11-27 10:43 [PATCHSET] libata: improve timing code and fix pata_amd transfer mode selection, take #2 Tejun Heo
2007-11-27 10:43 ` [PATCH 07/12] libata: fix ata_acpi_gtm_xfermask() Tejun Heo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1194327551238-git-send-email-htejun@gmail.com \
    --to=htejun@gmail.com \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=jeff@garzik.org \
    --cc=linux-ide@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).