Skip to contentSkip to conversation

pata_atiixp: fix slave device support

PATCH Applied Bartlomiej Zolnierkiewicz 2 participants 3 messages Aug 6, 2009 – Aug 12, 2009 libata

lore
Share & tools
lore

Bartlomiej Zolnierkiewicz

author
#0 lore raw

PATCH 1/1 pata_atiixp: fix slave device support

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH] pata_atiixp: fix slave device support

PIO and MWDMA timings are never programmed for the slave device because timing registers are treated as 16-bit long ones.

The bug is an attixp -> pata_atiixp regression and goes back to:

	commit 669a5db411d85a14f86cd92bc16bf7ab5b8aa235
	Author: Jeff Garzik <[email protected]>
	Date:   Tue Aug 29 18:12:40 2006 -0400

	    [libata] Add a bunch of PATA drivers.

The patch should fix:

	http://bugzilla.kernel.org/show_bug.cgi?id=13921
Reported-by Krystian Juskowiak <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Robert Hancock <[email protected]>
Signed-off-by Bartlomiej Zolnierkiewicz <[email protected]>

---
After this patch pata_atiixp still has two bugs (+ few superfluous oddities) not present in atiixp (although not as severe as this one)..

 drivers/ata/pata_atiixp.c |   19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)
drivers/ata/pata_atiixp.c +10 -9
Index: b/drivers/ata/pata_atiixp.c
===================================================================
@@ -1,6 +1,7 @@
 /*
  * pata_atiixp.c 	- ATI PATA for new ATA layer
  *			  (C) 2005 Red Hat Inc
+ *			  (C) 2009 Bartlomiej Zolnierkiewicz
  *
  * Based on
  *

@@ -61,20 +62,19 @@ static void atiixp_set_pio_timing(struct
 
 	struct pci_dev *pdev = to_pci_dev(ap->host->dev);
 	int dn = 2 * ap->port_no + adev->devno;
-
-	/* Check this is correct - the order is odd in both drivers */
 	int timing_shift = (16 * ap->port_no) + 8 * (adev->devno ^ 1);
-	u16 pio_mode_data, pio_timing_data;
+	u32 pio_timing_data;
+	u16 pio_mode_data;
 
 	pci_read_config_word(pdev, ATIIXP_IDE_PIO_MODE, &pio_mode_data);
 	pio_mode_data &= ~(0x7 << (4 * dn));
 	pio_mode_data |= pio << (4 * dn);
 	pci_write_config_word(pdev, ATIIXP_IDE_PIO_MODE, pio_mode_data);
 
-	pci_read_config_word(pdev, ATIIXP_IDE_PIO_TIMING, &pio_timing_data);
+	pci_read_config_dword(pdev, ATIIXP_IDE_PIO_TIMING, &pio_timing_data);
 	pio_timing_data &= ~(0xFF << timing_shift);
 	pio_timing_data |= (pio_timings[pio] << timing_shift);
-	pci_write_config_word(pdev, ATIIXP_IDE_PIO_TIMING, pio_timing_data);
+	pci_write_config_dword(pdev, ATIIXP_IDE_PIO_TIMING, pio_timing_data);
 }
 
 /**

@@ -119,16 +119,17 @@ static void atiixp_set_dmamode(struct at
 		udma_mode_data |= dma << (4 * dn);
 		pci_write_config_word(pdev, ATIIXP_IDE_UDMA_MODE, udma_mode_data);
 	} else {
-		u16 mwdma_timing_data;
-		/* Check this is correct - the order is odd in both drivers */
 		int timing_shift = (16 * ap->port_no) + 8 * (adev->devno ^ 1);
+		u32 mwdma_timing_data;
 
 		dma -= XFER_MW_DMA_0;
 
-		pci_read_config_word(pdev, ATIIXP_IDE_MWDMA_TIMING, &mwdma_timing_data);
+		pci_read_config_dword(pdev, ATIIXP_IDE_MWDMA_TIMING,
+				      &mwdma_timing_data);
 		mwdma_timing_data &= ~(0xFF << timing_shift);
 		mwdma_timing_data |= (mwdma_timings[dma] << timing_shift);
-		pci_write_config_word(pdev, ATIIXP_IDE_MWDMA_TIMING, mwdma_timing_data);
+		pci_write_config_dword(pdev, ATIIXP_IDE_MWDMA_TIMING,
+				       mwdma_timing_data);
 	}
 	/*
 	 *	We must now look at the PIO mode situation. We may need to
author
#1 lore raw
Quoted 2 lines · On Thursday 06 August 2009 17:40:15 Bartlomiej Zolnierkiewicz wrote:

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH] pata_atiixp: fix slave device support

Argh, I pushed the "Send" button too quick..

Below version with the correct patch description (I doubt that this fix will help with bug #13921)..

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH v2] pata_atiixp: fix second channel support

PIO and MWDMA timings are never programmed for the second channel because timing registers are treated as 16-bit long ones.

The bug is an attixp -> pata_atiixp regression and goes back to:

	commit 669a5db411d85a14f86cd92bc16bf7ab5b8aa235
	Author: Jeff Garzik <[email protected]>
	Date:   Tue Aug 29 18:12:40 2006 -0400

	    [libata] Add a bunch of PATA drivers.
Cc Krystian Juskowiak <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Robert Hancock <[email protected]>
Signed-off-by Bartlomiej Zolnierkiewicz <[email protected]>

---
After this patch pata_atiixp still have two bugs (+ few superfluous oddities) not present in atiixp (although not as severe as this one)..

 drivers/ata/pata_atiixp.c |   19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)
drivers/ata/pata_atiixp.c +10 -9
Index: b/drivers/ata/pata_atiixp.c
===================================================================
@@ -1,6 +1,7 @@
 /*
  * pata_atiixp.c 	- ATI PATA for new ATA layer
  *			  (C) 2005 Red Hat Inc
+ *			  (C) 2009 Bartlomiej Zolnierkiewicz
  *
  * Based on
  *

@@ -61,20 +62,19 @@ static void atiixp_set_pio_timing(struct
 
 	struct pci_dev *pdev = to_pci_dev(ap->host->dev);
 	int dn = 2 * ap->port_no + adev->devno;
-
-	/* Check this is correct - the order is odd in both drivers */
 	int timing_shift = (16 * ap->port_no) + 8 * (adev->devno ^ 1);
-	u16 pio_mode_data, pio_timing_data;
+	u32 pio_timing_data;
+	u16 pio_mode_data;
 
 	pci_read_config_word(pdev, ATIIXP_IDE_PIO_MODE, &pio_mode_data);
 	pio_mode_data &= ~(0x7 << (4 * dn));
 	pio_mode_data |= pio << (4 * dn);
 	pci_write_config_word(pdev, ATIIXP_IDE_PIO_MODE, pio_mode_data);
 
-	pci_read_config_word(pdev, ATIIXP_IDE_PIO_TIMING, &pio_timing_data);
+	pci_read_config_dword(pdev, ATIIXP_IDE_PIO_TIMING, &pio_timing_data);
 	pio_timing_data &= ~(0xFF << timing_shift);
 	pio_timing_data |= (pio_timings[pio] << timing_shift);
-	pci_write_config_word(pdev, ATIIXP_IDE_PIO_TIMING, pio_timing_data);
+	pci_write_config_dword(pdev, ATIIXP_IDE_PIO_TIMING, pio_timing_data);
 }
 
 /**

@@ -119,16 +119,17 @@ static void atiixp_set_dmamode(struct at
 		udma_mode_data |= dma << (4 * dn);
 		pci_write_config_word(pdev, ATIIXP_IDE_UDMA_MODE, udma_mode_data);
 	} else {
-		u16 mwdma_timing_data;
-		/* Check this is correct - the order is odd in both drivers */
 		int timing_shift = (16 * ap->port_no) + 8 * (adev->devno ^ 1);
+		u32 mwdma_timing_data;
 
 		dma -= XFER_MW_DMA_0;
 
-		pci_read_config_word(pdev, ATIIXP_IDE_MWDMA_TIMING, &mwdma_timing_data);
+		pci_read_config_dword(pdev, ATIIXP_IDE_MWDMA_TIMING,
+				      &mwdma_timing_data);
 		mwdma_timing_data &= ~(0xFF << timing_shift);
 		mwdma_timing_data |= (mwdma_timings[dma] << timing_shift);
-		pci_write_config_word(pdev, ATIIXP_IDE_MWDMA_TIMING, mwdma_timing_data);
+		pci_write_config_dword(pdev, ATIIXP_IDE_MWDMA_TIMING,
+				       mwdma_timing_data);
 	}
 	/*
 	 *	We must now look at the PIO mode situation. We may need to
Applied
#2 lore raw
Quoted 28 lines · Bartlomiej Zolnierkiewicz
Quoted 2 lines · On Thursday 06 August 2009 17:40:15 Bartlomiej Zolnierkiewicz wrote:

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH] pata_atiixp: fix slave device support

Argh, I pushed the "Send" button too quick..

Below version with the correct patch description (I doubt that this fix will help with bug #13921)..

From: Bartlomiej Zolnierkiewicz <[email protected]>
Subject: [PATCH v2] pata_atiixp: fix second channel support

PIO and MWDMA timings are never programmed for the second channel because timing registers are treated as 16-bit long ones.

The bug is an attixp -> pata_atiixp regression and goes back to:

	commit 669a5db411d85a14f86cd92bc16bf7ab5b8aa235
	Author: Jeff Garzik <[email protected]>
	Date:   Tue Aug 29 18:12:40 2006 -0400

	    [libata] Add a bunch of PATA drivers.
Cc Krystian Juskowiak <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Robert Hancock <[email protected]>
Signed-off-by Bartlomiej Zolnierkiewicz <[email protected]>
Quoted 6 lines

---
After this patch pata_atiixp still have two bugs (+ few superfluous oddities) not present in atiixp (although not as severe as this one)..

 drivers/ata/pata_atiixp.c |   19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)

applied