Skip to contentSkip to conversation

ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

PATCH New Peng Ma 3 participants 21 messages Mar 27, 2019 – May 24, 2019

Subsystems (2)
lore
Share & tools
lore

Peng Ma

author
#0 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ======================
This board does not use CONFIG_DM_SCSI. Please update
the storage controller to use CONFIG_DM_SCSI before the v2019.07 release. Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

---
depends on:
- https://patchwork.ozlabs.org/project/uboot/list/?series=99168
- https://patchwork.ozlabs.org/project/uboot/list/?series=99167

 drivers/ata/Kconfig    |   10 +
 drivers/ata/Makefile   |    1 +
 drivers/ata/fsl_ahci.c | 1030 ++++++++++++++++++++++++++++++++++++++++
 drivers/ata/fsl_sata.h |    1 +
 4 files changed, 1042 insertions(+)
 create mode 100644 drivers/ata/fsl_ahci.c
drivers/ata/Kconfig +10 -0
index 49a056e941..efac29c709 100644
@@ -59,6 +59,16 @@ config DWC_AHCI
 	  Enable this driver to support Sata devices through
 	  Synopsys DWC AHCI module.
 
+config FSL_AHCI
+	bool "Enable Freescale AHCI driver support"
+	select SCSI_AHCI
+	depends on AHCI
+	depends on DM_SCSI
+	help
+	  Enable this driver to support Sata devices found in
+	  some Freescale PowerPC SoCs.
+
+
 config DWC_AHSATA
 	bool "Enable DWC AHSATA driver support"
 	select LIBATA
drivers/ata/Makefile +1 -0
index 10bed53bb3..93aabf34c7 100644
@@ -4,6 +4,7 @@
 # Wolfgang Denk, DENX Software Engineering, wd at denx.de.
 
 obj-$(CONFIG_DWC_AHCI) += dwc_ahci.o
+obj-$(CONFIG_FSL_AHCI) += fsl_ahci.o
 obj-$(CONFIG_AHCI) += ahci-uclass.o
 obj-$(CONFIG_AHCI_PCI) += ahci-pci.o
 obj-$(CONFIG_SCSI_AHCI) += ahci.o
drivers/ata/fsl_ahci.c new file +1030 -0

1,030 lines not on this page · raw message

drivers/ata/fsl_sata.h +1 -0
index 1e2da10b02..a4ee83d187 100644
@@ -312,6 +312,7 @@ typedef struct fsl_sata {
 	int		wcache;
 	int		flush;
 	int		flush_ext;
+	u32		dma_flag;
 } fsl_sata_t;
 
 #define READ_CMD	0
Signature · 1 line
2.17.1
author
#1 lore raw

PATCH 2/4 ppc: t2080qds: add sata node

This patch is to add sata node for t2080qds

---
depends on:
- https://patchwork.ozlabs.org/project/uboot/list/?series=99168
- https://patchwork.ozlabs.org/project/uboot/list/?series=99167

 arch/powerpc/dts/t2080.dtsi | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
arch/powerpc/dts/t2080.dtsi +16 -0
index 666601ed49..d2bebb08b6 100644
@@ -87,5 +87,21 @@
 			dr_mode = "host";
 			phy_type = "utmi";
 		};
+
+		sata0: sata at 220000 {
+			compatible = "fsl,pq-sata-v2";
+			reg = <0x220000 0x1000>;
+			interrupts = <68 0x2 0 0>;
+			sata-number = <0x0>;
+			sata-fpdma = <0x0>;
+		};
+
+		sata1: sata at 221000 {
+			compatible = "fsl,pq-sata-v2";
+			reg = <0x221000 0x1000>;
+			interrupts = <69 0x2 0 0>;
+			sata-number = <0x0>;
+			sata-fpdma = <0x0>;
+		};
 	};
 };
Signature · 1 line
2.17.1
#2 lore raw

PATCH 2/4 ppc: t2080qds: add sata node

Quoted 12 lines

-----Original Message-----
From: Peng Ma
Sent: Wednesday, March 27, 2019 2:53 PM
To: Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; sr at denx.de; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de; Peng Ma <[email protected]>
Subject: [PATCH 2/4] ppc: t2080qds: add sata node

Quoted 4 lines

This patch is to add sata node for t2080qds

---

This patch has been applied to u-boot-mpc85xx, awaiting upstream.

--pk

author
#3 lore raw

PATCH 3/4 powerpc: mpc85xx: delete FSL_SATA for T2080QDS board.

In reply to Peng Ma

---
depends on:
- https://patchwork.ozlabs.org/project/uboot/list/?series=99168
- https://patchwork.ozlabs.org/project/uboot/list/?series=99167

 arch/powerpc/cpu/mpc85xx/Kconfig | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
arch/powerpc/cpu/mpc85xx/Kconfig +1 -3
index 0057f195b3..aebf168a89 100644
@@ -352,7 +352,6 @@ config TARGET_T2080QDS
 	select PHYS_64BIT
 	select FSL_DDR_FIRST_SLOT_QUAD_CAPABLE
 	select FSL_DDR_INTERACTIVE
-	imply CMD_SATA
 
 config TARGET_T2080RDB
 	bool "Support T2080RDB"
@@ -361,6 +360,7 @@ config TARGET_T2080RDB
 	select SUPPORT_SPL
 	select PHYS_64BIT
 	imply CMD_SATA
+	imply FSL_SATA
 	imply PANIC_HANG
 
 config TARGET_T2081QDS
@@ -1081,10 +1081,8 @@ config ARCH_T2080
 	select SYS_FSL_SEC_COMPAT_4
 	select SYS_PPC64
 	select FSL_IFC
-	imply CMD_SATA
 	imply CMD_NAND
 	imply CMD_REGINFO
-	imply FSL_SATA
 
 config ARCH_T2081
 	bool
Signature · 1 line
2.17.1
#4 lore raw

PATCH 3/4 powerpc: mpc85xx: delete FSL_SATA for T2080QDS board.

Quoted 12 lines

-----Original Message-----
From: Peng Ma
Sent: Wednesday, March 27, 2019 2:54 PM
To: Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; sr at denx.de; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de; Peng Ma <[email protected]>
Subject: [PATCH 3/4] powerpc: mpc85xx: delete FSL_SATA for T2080QDS board.

Quoted 2 lines

---

This patch has been applied to u-boot-mpc85xx, awaiting upstream.

--pk

author
#5 lore raw

PATCH 4/4 configs: enable sata device module in T2080QDS

In reply to Peng Ma

This patch is to enable sata DM for T2080QDS in uboot

---
depends on:
- https://patchwork.ozlabs.org/project/uboot/list/?series=99168
- https://patchwork.ozlabs.org/project/uboot/list/?series=99167

 configs/T2080QDS_NAND_defconfig           | 11 +++++++----
 configs/T2080QDS_SDCARD_defconfig         | 11 +++++++----
 configs/T2080QDS_SECURE_BOOT_defconfig    | 16 +++++++++-------
 configs/T2080QDS_SPIFLASH_defconfig       | 11 +++++++----
 configs/T2080QDS_SRIO_PCIE_BOOT_defconfig | 18 ++++++++++--------
 configs/T2080QDS_defconfig                | 11 +++++++----
 6 files changed, 47 insertions(+), 31 deletions(-)
configs/T2080QDS_NAND_defconfig +7 -4
index bbd2f3e781..e67a50251a 100644
@@ -7,6 +7,7 @@ CONFIG_SPL_DRIVERS_MISC_SUPPORT=y
 CONFIG_SPL=y
 CONFIG_MPC85xx=y
 CONFIG_TARGET_T2080QDS=y
+CONFIG_AHCI=y
 CONFIG_FIT=y
 CONFIG_FIT_VERBOSE=y
 CONFIG_OF_BOARD_SETUP=y
@@ -38,7 +39,10 @@ CONFIG_MTDPARTS_DEFAULT="mtdparts=fe8000000.nor:1m(uboot),5m(kernel),128k(dtb),9
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="t2080qds"
 CONFIG_ENV_IS_IN_NAND=y
+CONFIG_DM=y
+CONFIG_FSL_AHCI=y
 CONFIG_FSL_CAAM=y
+CONFIG_DM_MMC=y
 CONFIG_FSL_ESDHC=y
 CONFIG_MTD_NOR_FLASH=y
 CONFIG_FLASH_CFI_DRIVER=y
@@ -56,12 +60,11 @@ CONFIG_PHYLIB=y
 CONFIG_PHY_AQUANTIA=y
 CONFIG_E1000=y
 CONFIG_MII=y
+CONFIG_SCSI=y
+CONFIG_DM_SCSI=y
 CONFIG_SYS_NS16550=y
 CONFIG_SPI=y
 CONFIG_FSL_ESPI=y
 CONFIG_USB=y
-CONFIG_USB_STORAGE=y
 CONFIG_DM_USB=y
-CONFIG_DM_MMC=y
-CONFIG_BLK=y
-CONFIG_DM=y
+CONFIG_USB_STORAGE=y
configs/T2080QDS_SDCARD_defconfig +7 -4
index df66599475..5f681d6d0f 100644
@@ -8,6 +8,7 @@ CONFIG_SPL_DRIVERS_MISC_SUPPORT=y
 CONFIG_SPL=y
 CONFIG_MPC85xx=y
 CONFIG_TARGET_T2080QDS=y
+CONFIG_AHCI=y
 CONFIG_FIT=y
 CONFIG_FIT_VERBOSE=y
 CONFIG_OF_BOARD_SETUP=y
@@ -38,7 +39,10 @@ CONFIG_MTDPARTS_DEFAULT="mtdparts=fe8000000.nor:1m(uboot),5m(kernel),128k(dtb),9
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="t2080qds"
 CONFIG_ENV_IS_IN_MMC=y
+CONFIG_DM=y
+CONFIG_FSL_AHCI=y
 CONFIG_FSL_CAAM=y
+CONFIG_DM_MMC=y
 CONFIG_FSL_ESDHC=y
 CONFIG_MTD_NOR_FLASH=y
 CONFIG_FLASH_CFI_DRIVER=y
@@ -55,12 +59,11 @@ CONFIG_PHYLIB=y
 CONFIG_PHY_AQUANTIA=y
 CONFIG_E1000=y
 CONFIG_MII=y
+CONFIG_SCSI=y
+CONFIG_DM_SCSI=y
 CONFIG_SYS_NS16550=y
 CONFIG_SPI=y
 CONFIG_FSL_ESPI=y
 CONFIG_USB=y
-CONFIG_USB_STORAGE=y
 CONFIG_DM_USB=y
-CONFIG_DM_MMC=y
-CONFIG_BLK=y
-CONFIG_DM=y
+CONFIG_USB_STORAGE=y
configs/T2080QDS_SECURE_BOOT_defconfig +9 -7
index 766ee7c4ac..f6284fede0 100644
@@ -3,6 +3,8 @@ CONFIG_SYS_TEXT_BASE=0xEFF40000
 CONFIG_SECURE_BOOT=y
 CONFIG_MPC85xx=y
 CONFIG_TARGET_T2080QDS=y
+CONFIG_MPC85XX_HAVE_RESET_VECTOR=y
+CONFIG_AHCI=y
 # CONFIG_SYS_MALLOC_F is not set
 CONFIG_FIT=y
 CONFIG_FIT_VERBOSE=y
@@ -26,7 +28,11 @@ CONFIG_CMD_FAT=y
 CONFIG_CMD_MTDPARTS=y
 CONFIG_MTDIDS_DEFAULT="nor0=fe8000000.nor,nand0=fff800000.flash,spi0=spife110000.0"
 CONFIG_MTDPARTS_DEFAULT="mtdparts=fe8000000.nor:1m(uboot),5m(kernel),128k(dtb),96m(fs),-(user);fff800000.flash:1m(uboot),5m(kernel),128k(dtb),96m(fs),-(user);spife110000.0:1m(uboot),5m(kernel),128k(dtb),-(user)"
+CONFIG_OF_CONTROL=y
+CONFIG_DEFAULT_DEVICE_TREE="t2080qds"
 CONFIG_DM=y
+CONFIG_FSL_AHCI=y
+CONFIG_DM_MMC=y
 CONFIG_FSL_ESDHC=y
 CONFIG_MTD_NOR_FLASH=y
 CONFIG_FLASH_CFI_DRIVER=y
@@ -43,18 +49,14 @@ CONFIG_PHYLIB=y
 CONFIG_PHY_AQUANTIA=y
 CONFIG_E1000=y
 CONFIG_MII=y
+CONFIG_SCSI=y
+CONFIG_DM_SCSI=y
 CONFIG_SYS_NS16550=y
 CONFIG_SPI=y
 CONFIG_FSL_ESPI=y
 CONFIG_USB=y
-CONFIG_USB_STORAGE=y
 CONFIG_DM_USB=y
+CONFIG_USB_STORAGE=y
 CONFIG_RSA=y
 CONFIG_SPL_RSA=y
 CONFIG_RSA_SOFTWARE_EXP=y
-CONFIG_OF_LIBFDT=y
-CONFIG_MPC85XX_HAVE_RESET_VECTOR=y
-CONFIG_DEFAULT_DEVICE_TREE="t2080qds"
-CONFIG_DM_MMC=y
-CONFIG_BLK=y
-CONFIG_OF_CONTROL=y
configs/T2080QDS_SPIFLASH_defconfig +7 -4
index 43c50fb27b..4fe737ddb5 100644
@@ -9,6 +9,7 @@ CONFIG_SPL_SPI_FLASH_SUPPORT=y
 CONFIG_SPL_SPI_SUPPORT=y
 CONFIG_MPC85xx=y
 CONFIG_TARGET_T2080QDS=y
+CONFIG_AHCI=y
 CONFIG_FIT=y
 CONFIG_FIT_VERBOSE=y
 CONFIG_OF_BOARD_SETUP=y
@@ -39,7 +40,10 @@ CONFIG_MTDPARTS_DEFAULT="mtdparts=fe8000000.nor:1m(uboot),5m(kernel),128k(dtb),9
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="t2080qds"
 CONFIG_ENV_IS_IN_SPI_FLASH=y
+CONFIG_DM=y
+CONFIG_FSL_AHCI=y
 CONFIG_FSL_CAAM=y
+CONFIG_DM_MMC=y
 CONFIG_FSL_ESDHC=y
 CONFIG_MTD_NOR_FLASH=y
 CONFIG_FLASH_CFI_DRIVER=y
@@ -56,12 +60,11 @@ CONFIG_PHYLIB=y
 CONFIG_PHY_AQUANTIA=y
 CONFIG_E1000=y
 CONFIG_MII=y
+CONFIG_SCSI=y
+CONFIG_DM_SCSI=y
 CONFIG_SYS_NS16550=y
 CONFIG_SPI=y
 CONFIG_FSL_ESPI=y
 CONFIG_USB=y
-CONFIG_USB_STORAGE=y
 CONFIG_DM_USB=y
-CONFIG_DM_MMC=y
-CONFIG_BLK=y
-CONFIG_DM=y
+CONFIG_USB_STORAGE=y
configs/T2080QDS_SRIO_PCIE_BOOT_defconfig +10 -8
index 869f203160..f18759f9fd 100644
@@ -2,6 +2,8 @@ CONFIG_PPC=y
 CONFIG_SYS_TEXT_BASE=0xFFF40000
 CONFIG_MPC85xx=y
 CONFIG_TARGET_T2080QDS=y
+CONFIG_MPC85XX_HAVE_RESET_VECTOR=y
+CONFIG_AHCI=y
 CONFIG_FIT=y
 CONFIG_FIT_VERBOSE=y
 CONFIG_OF_BOARD_SETUP=y
@@ -22,8 +24,13 @@ CONFIG_CMD_PING=y
 CONFIG_MP=y
 CONFIG_CMD_EXT2=y
 CONFIG_CMD_FAT=y
+CONFIG_OF_CONTROL=y
+CONFIG_DEFAULT_DEVICE_TREE="t2080qds"
 CONFIG_ENV_IS_IN_REMOTE=y
+CONFIG_DM=y
+CONFIG_FSL_AHCI=y
 CONFIG_FSL_CAAM=y
+CONFIG_DM_MMC=y
 CONFIG_FSL_ESDHC=y
 CONFIG_SPI_FLASH=y
 CONFIG_SF_DEFAULT_MODE=0
@@ -35,16 +42,11 @@ CONFIG_PHYLIB=y
 CONFIG_PHY_AQUANTIA=y
 CONFIG_E1000=y
 CONFIG_MII=y
+CONFIG_SCSI=y
+CONFIG_DM_SCSI=y
 CONFIG_SYS_NS16550=y
 CONFIG_SPI=y
 CONFIG_FSL_ESPI=y
 CONFIG_USB=y
-CONFIG_USB_STORAGE=y
 CONFIG_DM_USB=y
-CONFIG_OF_LIBFDT=y
-CONFIG_MPC85XX_HAVE_RESET_VECTOR=y
-CONFIG_OF_CONTROL=y
-CONFIG_DEFAULT_DEVICE_TREE="t2080qds"
-CONFIG_DM_MMC=y
-CONFIG_BLK=y
-CONFIG_DM=y
+CONFIG_USB_STORAGE=y
configs/T2080QDS_defconfig +7 -4
index 28dad83302..8166c13b17 100644
@@ -3,6 +3,7 @@ CONFIG_SYS_TEXT_BASE=0xEFF40000
 CONFIG_MPC85xx=y
 CONFIG_TARGET_T2080QDS=y
 CONFIG_MPC85XX_HAVE_RESET_VECTOR=y
+CONFIG_AHCI=y
 CONFIG_FIT=y
 CONFIG_FIT_VERBOSE=y
 CONFIG_OF_BOARD_SETUP=y
@@ -28,7 +29,10 @@ CONFIG_MTDPARTS_DEFAULT="mtdparts=fe8000000.nor:1m(uboot),5m(kernel),128k(dtb),9
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="t2080qds"
 CONFIG_ENV_IS_IN_FLASH=y
+CONFIG_DM=y
+CONFIG_FSL_AHCI=y
 CONFIG_FSL_CAAM=y
+CONFIG_DM_MMC=y
 CONFIG_FSL_ESDHC=y
 CONFIG_MTD_NOR_FLASH=y
 CONFIG_FLASH_CFI_DRIVER=y
@@ -45,12 +49,11 @@ CONFIG_PHYLIB=y
 CONFIG_PHY_AQUANTIA=y
 CONFIG_E1000=y
 CONFIG_MII=y
+CONFIG_SCSI=y
+CONFIG_DM_SCSI=y
 CONFIG_SYS_NS16550=y
 CONFIG_SPI=y
 CONFIG_FSL_ESPI=y
 CONFIG_USB=y
-CONFIG_USB_STORAGE=y
-CONFIG_DM_MMC=y
-CONFIG_BLK=y
-CONFIG_DM=y
 CONFIG_DM_USB=y
+CONFIG_USB_STORAGE=y
Signature · 1 line
2.17.1
#6 lore raw

PATCH 4/4 configs: enable sata device module in T2080QDS

Quoted 12 lines

-----Original Message-----
From: Peng Ma
Sent: Wednesday, March 27, 2019 2:54 PM
To: Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; sr at denx.de; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de; Peng Ma <[email protected]>
Subject: [PATCH 4/4] configs: enable sata device module in T2080QDS

Quoted 4 lines

This patch is to enable sata DM for T2080QDS in uboot

---

This patch has been applied to u-boot-mpc85xx, awaiting upstream.

--pk

#7 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

In reply to Peng Ma
Quoted 16 lines · Peng Ma

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ======================
This board does not use CONFIG_DM_SCSI. Please update
the storage controller to use CONFIG_DM_SCSI before the v2019.07 release. Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

---
depends on:
- https://patchwork.ozlabs.org/project/uboot/list/?series=99168
- https://patchwork.ozlabs.org/project/uboot/list/?series=99167

Quoted 6 lines
  drivers/ata/Kconfig    |   10 +
  drivers/ata/Makefile   |    1 +
  drivers/ata/fsl_ahci.c | 1030 ++++++++++++++++++++++++++++++++++++++++
  drivers/ata/fsl_sata.h |    1 +
  4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

Thanks,
Stefan

Quoted 1,075 lines
drivers/ata/Kconfig +10 -0
index 49a056e941..efac29c709 100644
@@ -59,6 +59,16 @@ config DWC_AHCI
  	  Enable this driver to support Sata devices through
  	  Synopsys DWC AHCI module.
 
+config FSL_AHCI
+	bool "Enable Freescale AHCI driver support"
+	select SCSI_AHCI
+	depends on AHCI
+	depends on DM_SCSI
+	help
+	  Enable this driver to support Sata devices found in
+	  some Freescale PowerPC SoCs.
+
+
  config DWC_AHSATA
  	bool "Enable DWC AHSATA driver support"
  	select LIBATA
drivers/ata/Makefile +1 -0
index 10bed53bb3..93aabf34c7 100644
@@ -4,6 +4,7 @@
  # Wolfgang Denk, DENX Software Engineering, wd at denx.de.
 
  obj-$(CONFIG_DWC_AHCI) += dwc_ahci.o
+obj-$(CONFIG_FSL_AHCI) += fsl_ahci.o
  obj-$(CONFIG_AHCI) += ahci-uclass.o
  obj-$(CONFIG_AHCI_PCI) += ahci-pci.o
  obj-$(CONFIG_SCSI_AHCI) += ahci.o
drivers/ata/fsl_ahci.c new file +1030 -0
new file mode 100644
index 0000000000..16c6f7a335
@@ -0,0 +1,1030 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * NXP PPC SATA platform driver
+ *
+ * (C) Copyright 2019 NXP, Inc.
+ *
+ */
+#include <common.h>
+#include <asm/fsl_serdes.h>
+#include <dm/lists.h>
+#include <dm.h>
+#include <ahci.h>
+#include <scsi.h>
+#include <libata.h>
+#include <sata.h>
+#include <malloc.h>
+#include <memalign.h>
+#include <fis.h>
+
+#include "fsl_sata.h"
+
+struct fsl_ahci_priv {
+	u32 base;
+	u32 flag;
+	u32 number;
+	fsl_sata_t *fsl_sata;
+};
+
+static int fsl_ahci_bind(struct udevice *dev)
+{
+	return device_bind_driver(dev, "fsl_ahci_scsi", "fsl_ahci_scsi", NULL);
+}
+
+static int fsl_ahci_ofdata_to_platdata(struct udevice *dev)
+{
+	struct fsl_ahci_priv *priv = dev_get_priv(dev);
+
+	priv->number = dev_read_u32_default(dev, "sata-number", -1);
+	priv->flag = dev_read_u32_default(dev, "sata-fpdma", -1);
+
+	priv->base = dev_read_addr(dev);
+	if (priv->base == FDT_ADDR_T_NONE)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int ata_wait_register(unsigned __iomem *addr, u32 mask,
+			     u32 val, u32 timeout_msec)
+{
+	int i;
+
+	for (i = 0; ((in_le32(addr) & mask) != val) && i < timeout_msec; i++)
+		mdelay(1);
+
+	return (i < timeout_msec) ? 0 : -1;
+}
+
+static void fsl_sata_dump_sfis(struct sata_fis_d2h *s)
+{
+	printf("Status FIS dump:\n\r");
+	printf("fis_type:		%02x\n\r", s->fis_type);
+	printf("pm_port_i:		%02x\n\r", s->pm_port_i);
+	printf("status:			%02x\n\r", s->status);
+	printf("error:			%02x\n\r", s->error);
+	printf("lba_low:		%02x\n\r", s->lba_low);
+	printf("lba_mid:		%02x\n\r", s->lba_mid);
+	printf("lba_high:		%02x\n\r", s->lba_high);
+	printf("device:			%02x\n\r", s->device);
+	printf("lba_low_exp:		%02x\n\r", s->lba_low_exp);
+	printf("lba_mid_exp:		%02x\n\r", s->lba_mid_exp);
+	printf("lba_high_exp:		%02x\n\r", s->lba_high_exp);
+	printf("res1:			%02x\n\r", s->res1);
+	printf("sector_count:		%02x\n\r", s->sector_count);
+	printf("sector_count_exp:	%02x\n\r", s->sector_count_exp);
+}
+
+static void fsl_sata_dump_regs(fsl_sata_reg_t __iomem *reg)
+{
+	printf("\n\rSATA:           %08x\n\r", (u32)reg);
+	printf("CQR:            %08x\n\r", in_le32(&reg->cqr));
+	printf("CAR:            %08x\n\r", in_le32(&reg->car));
+	printf("CCR:            %08x\n\r", in_le32(&reg->ccr));
+	printf("CER:            %08x\n\r", in_le32(&reg->cer));
+	printf("CQR:            %08x\n\r", in_le32(&reg->cqr));
+	printf("DER:            %08x\n\r", in_le32(&reg->der));
+	printf("CHBA:           %08x\n\r", in_le32(&reg->chba));
+	printf("HStatus:        %08x\n\r", in_le32(&reg->hstatus));
+	printf("HControl:       %08x\n\r", in_le32(&reg->hcontrol));
+	printf("CQPMP:          %08x\n\r", in_le32(&reg->cqpmp));
+	printf("SIG:            %08x\n\r", in_le32(&reg->sig));
+	printf("ICC:            %08x\n\r", in_le32(&reg->icc));
+	printf("SStatus:        %08x\n\r", in_le32(&reg->sstatus));
+	printf("SError:         %08x\n\r", in_le32(&reg->serror));
+	printf("SControl:       %08x\n\r", in_le32(&reg->scontrol));
+	printf("SNotification:  %08x\n\r", in_le32(&reg->snotification));
+	printf("TransCfg:       %08x\n\r", in_le32(&reg->transcfg));
+	printf("TransStatus:    %08x\n\r", in_le32(&reg->transstatus));
+	printf("LinkCfg:        %08x\n\r", in_le32(&reg->linkcfg));
+	printf("LinkCfg1:       %08x\n\r", in_le32(&reg->linkcfg1));
+	printf("LinkCfg2:       %08x\n\r", in_le32(&reg->linkcfg2));
+	printf("LinkStatus:     %08x\n\r", in_le32(&reg->linkstatus));
+	printf("LinkStatus1:    %08x\n\r", in_le32(&reg->linkstatus1));
+	printf("PhyCtrlCfg:     %08x\n\r", in_le32(&reg->phyctrlcfg));
+	printf("SYSPR:          %08x\n\r", in_be32(&reg->syspr));
+}
+
+static int init_sata(struct fsl_ahci_priv *priv)
+{
+	int i;
+	u32 cda;
+	u32 val32;
+	u32 sig;
+	fsl_sata_t *sata;
+	u32 length, align;
+	cmd_hdr_tbl_t *cmd_hdr;
+	fsl_sata_reg_t __iomem *reg;
+
+	int dev = priv->number;
+
+	if (dev < 0 || dev > (CONFIG_SYS_SATA_MAX_DEVICE - 1)) {
+		printf("the sata index %d is out of ranges\n\r", dev);
+		return -EINVAL;
+	}
+
+#ifdef CONFIG_MPC85xx
+	if (dev == 0 && (!is_serdes_configured(SATA1))) {
+		printf("SATA%d [dev = %d] is not enabled\n", dev + 1, dev);
+		return -EINVAL;
+	}
+	if (dev == 1 && (!is_serdes_configured(SATA2))) {
+		printf("SATA%d [dev = %d] is not enabled\n", dev + 1, dev);
+		return -EINVAL;
+	}
+#endif
+
+	/* Allocate SATA device driver struct */
+	sata = (fsl_sata_t *)malloc(sizeof(fsl_sata_t));
+	if (!sata) {
+		printf("alloc the sata device struct failed\n\r");
+		return -ENOMEM;
+	}
+	/* Zero all of the device driver struct */
+	memset((void *)sata, 0, sizeof(fsl_sata_t));
+
+	sata->dma_flag = priv->flag;
+	snprintf(sata->name, 12, "SATA%d", dev);
+
+	/* Set the controller register base address to device struct */
+	reg = (fsl_sata_reg_t *)priv->base;
+	sata->reg_base = reg;
+
+	/* Allocate the command header table, 4 bytes aligned */
+	length = sizeof(struct cmd_hdr_tbl);
+	align = SATA_HC_CMD_HDR_TBL_ALIGN;
+	sata->cmd_hdr_tbl_offset = (void *)malloc(length + align);
+	if (!sata->cmd_hdr_tbl_offset) {
+		printf("alloc the command header failed\n\r");
+		return -ENOMEM;
+	}
+
+	cmd_hdr = (cmd_hdr_tbl_t *)(((u32)sata->cmd_hdr_tbl_offset + align)
+						& ~(align - 1));
+	sata->cmd_hdr = cmd_hdr;
+
+	/* Zero all of the command header table */
+	memset((void *)sata->cmd_hdr_tbl_offset, 0, length + align);
+
+	/* Allocate command descriptor for all command */
+	length = sizeof(struct cmd_desc) * SATA_HC_MAX_CMD;
+	align = SATA_HC_CMD_DESC_ALIGN;
+	sata->cmd_desc_offset = (void *)malloc(length + align);
+	if (!sata->cmd_desc_offset) {
+		printf("alloc the command descriptor failed\n\r");
+		return -ENOMEM;
+	}
+	sata->cmd_desc = (cmd_desc_t *)(((u32)sata->cmd_desc_offset + align)
+						& ~(align - 1));
+	/* Zero all of command descriptor */
+	memset((void *)sata->cmd_desc_offset, 0, length + align);
+
+	/* Link the command descriptor to command header */
+	for (i = 0; i < SATA_HC_MAX_CMD; i++) {
+		cda = ((u32)sata->cmd_desc + SATA_HC_CMD_DESC_SIZE * i)
+					 & ~(CMD_HDR_CDA_ALIGN - 1);
+		cmd_hdr->cmd_slot[i].cda = cpu_to_le32(cda);
+	}
+
+	/* To have safe state, force the controller offline */
+	val32 = in_le32(&reg->hcontrol);
+	val32 &= ~HCONTROL_ONOFF;
+	val32 |= HCONTROL_FORCE_OFFLINE;
+	out_le32(&reg->hcontrol, val32);
+
+	/* Wait the controller offline */
+	ata_wait_register(&reg->hstatus, HSTATUS_ONOFF, 0, 1000);
+
+	/* Set the command header base address to CHBA register to tell DMA */
+	out_le32(&reg->chba, (u32)cmd_hdr & ~0x3);
+
+	/* Snoop for the command header */
+	val32 = in_le32(&reg->hcontrol);
+	val32 |= HCONTROL_HDR_SNOOP;
+	out_le32(&reg->hcontrol, val32);
+
+	/* Disable all of interrupts */
+	val32 = in_le32(&reg->hcontrol);
+	val32 &= ~HCONTROL_INT_EN_ALL;
+	out_le32(&reg->hcontrol, val32);
+
+	/* Clear all of interrupts */
+	val32 = in_le32(&reg->hstatus);
+	out_le32(&reg->hstatus, val32);
+
+	/* Set the ICC, no interrupt coalescing */
+	out_le32(&reg->icc, 0x01000000);
+
+	/* No PM attatched, the SATA device direct connect */
+	out_le32(&reg->cqpmp, 0);
+
+	/* Clear SError register */
+	val32 = in_le32(&reg->serror);
+	out_le32(&reg->serror, val32);
+
+	/* Clear CER register */
+	val32 = in_le32(&reg->cer);
+	out_le32(&reg->cer, val32);
+
+	/* Clear DER register */
+	val32 = in_le32(&reg->der);
+	out_le32(&reg->der, val32);
+
+	/* No device detection or initialization action requested */
+	out_le32(&reg->scontrol, 0x00000300);
+
+	/* Configure the transport layer, default value */
+	out_le32(&reg->transcfg, 0x08000016);
+
+	/* Configure the link layer, default value */
+	out_le32(&reg->linkcfg, 0x0000ff34);
+
+	/* Bring the controller online */
+	val32 = in_le32(&reg->hcontrol);
+	val32 |= HCONTROL_ONOFF;
+	out_le32(&reg->hcontrol, val32);
+
+	mdelay(100);
+
+	/* print sata device name */
+	printf("%s ", sata->name);
+
+	/* Wait PHY RDY signal changed for 500ms */
+	ata_wait_register(&reg->hstatus, HSTATUS_PHY_RDY,
+			  HSTATUS_PHY_RDY, 500);
+
+	/* Check PHYRDY */
+	val32 = in_le32(&reg->hstatus);
+	if (val32 & HSTATUS_PHY_RDY) {
+		sata->link = 1;
+	} else {
+		sata->link = 0;
+		printf("(No RDY)\n\r");
+		return -EINVAL;
+	}
+
+	/* Wait for signature updated, which is 1st D2H */
+	ata_wait_register(&reg->hstatus, HSTATUS_SIGNATURE,
+			  HSTATUS_SIGNATURE, 10000);
+
+	if (val32 & HSTATUS_SIGNATURE) {
+		sig = in_le32(&reg->sig);
+		debug("Signature updated, the sig =%08x\n\r", sig);
+		sata->ata_device_type = ata_dev_classify(sig);
+	}
+
+	/* Check the speed */
+	val32 = in_le32(&reg->sstatus);
+	if ((val32 & SSTATUS_SPD_MASK) == SSTATUS_SPD_GEN1)
+		printf("(1.5 Gbps)\n\r");
+	else if ((val32 & SSTATUS_SPD_MASK) == SSTATUS_SPD_GEN2)
+		printf("(3 Gbps)\n\r");
+
+	priv->fsl_sata = sata;
+
+	return 0;
+}
+
+static int fsl_ata_exec_ata_cmd(struct fsl_sata *sata,
+				struct sata_fis_h2d *cfis,
+				int is_ncq, int tag,
+				u8 *buffer, u32 len)
+{
+	cmd_hdr_entry_t *cmd_hdr;
+	cmd_desc_t *cmd_desc;
+	sata_fis_h2d_t *h2d;
+	prd_entry_t *prde;
+	u32 ext_c_ddc;
+	u32 prde_count;
+	u32 val32;
+	u32 ttl;
+	u32 der;
+	int i;
+
+	fsl_sata_reg_t *reg = sata->reg_base;
+
+	/* Check xfer length */
+	if (len > SATA_HC_MAX_XFER_LEN) {
+		printf("max transfer length is 64MB\n\r");
+		return 0;
+	}
+
+	/* Setup the command descriptor */
+	cmd_desc = sata->cmd_desc + tag;
+
+	/* Get the pointer cfis of command descriptor */
+	h2d = (sata_fis_h2d_t *)cmd_desc->cfis;
+
+	/* Zero the cfis of command descriptor */
+	memset((void *)h2d, 0, SATA_HC_CMD_DESC_CFIS_SIZE);
+
+	/* Copy the cfis from user to command descriptor */
+	h2d->fis_type = cfis->fis_type;
+	h2d->pm_port_c = cfis->pm_port_c;
+	h2d->command = cfis->command;
+
+	h2d->features = cfis->features;
+	h2d->features_exp = cfis->features_exp;
+
+	h2d->lba_low = cfis->lba_low;
+	h2d->lba_mid = cfis->lba_mid;
+	h2d->lba_high = cfis->lba_high;
+	h2d->lba_low_exp = cfis->lba_low_exp;
+	h2d->lba_mid_exp = cfis->lba_mid_exp;
+	h2d->lba_high_exp = cfis->lba_high_exp;
+
+	if (!is_ncq) {
+		h2d->sector_count = cfis->sector_count;
+		h2d->sector_count_exp = cfis->sector_count_exp;
+	} else { /* NCQ */
+		h2d->sector_count = (u8)(tag << 3);
+	}
+
+	h2d->device = cfis->device;
+	h2d->control = cfis->control;
+
+	/* Setup the PRD table */
+	prde = (prd_entry_t *)cmd_desc->prdt;
+	memset((void *)prde, 0, sizeof(struct prdt));
+
+	prde_count = 0;
+	ttl = len;
+	for (i = 0; i < SATA_HC_MAX_PRD_DIRECT; i++) {
+		if (!len)
+			break;
+		prde->dba = cpu_to_le32((u32)buffer & ~0x3);
+		debug("dba = %08x\n\r", (u32)buffer);
+
+		if (len < PRD_ENTRY_MAX_XFER_SZ) {
+			ext_c_ddc = PRD_ENTRY_DATA_SNOOP | len;
+			debug("ext_c_ddc1 = %08x, len = %08x\n\r",
+			      ext_c_ddc, len);
+			prde->ext_c_ddc = cpu_to_le32(ext_c_ddc);
+			prde_count++;
+			prde++;
+		} else {
+			ext_c_ddc = PRD_ENTRY_DATA_SNOOP; /* 4M bytes */
+			debug("ext_c_ddc2 = %08x, len = %08x\n\r",
+			      ext_c_ddc, len);
+			prde->ext_c_ddc = cpu_to_le32(ext_c_ddc);
+			buffer += PRD_ENTRY_MAX_XFER_SZ;
+			len -= PRD_ENTRY_MAX_XFER_SZ;
+			prde_count++;
+			prde++;
+		}
+	}
+
+	/* Setup the command slot of cmd hdr */
+	cmd_hdr = (cmd_hdr_entry_t *)&sata->cmd_hdr->cmd_slot[tag];
+
+	cmd_hdr->cda = cpu_to_le32((u32)cmd_desc & ~0x3);
+
+	val32 = prde_count << CMD_HDR_PRD_ENTRY_SHIFT;
+	val32 |= sizeof(sata_fis_h2d_t);
+	cmd_hdr->prde_fis_len = cpu_to_le32(val32);
+
+	cmd_hdr->ttl = cpu_to_le32(ttl);
+
+	if (!is_ncq)
+		val32 = CMD_HDR_ATTR_RES | CMD_HDR_ATTR_SNOOP;
+	else
+		val32 = CMD_HDR_ATTR_RES | CMD_HDR_ATTR_SNOOP |
+			CMD_HDR_ATTR_FPDMA;
+
+	tag &= CMD_HDR_ATTR_TAG;
+	val32 |= tag;
+
+	debug("attribute = %08x\n\r", val32);
+	cmd_hdr->attribute = cpu_to_le32(val32);
+
+	/* Make sure cmd desc and cmd slot valid before command issue */
+	sync();
+
+	/* PMP*/
+	val32 = (u32)(h2d->pm_port_c & 0x0f);
+	out_le32(&reg->cqpmp, val32);
+
+	/* Wait no active */
+	if (ata_wait_register(&reg->car, (1 << tag), 0, 10000))
+		printf("Wait no active time out\n\r");
+
+	/* Issue command */
+	if (!(in_le32(&reg->cqr) & (1 << tag))) {
+		val32 = 1 << tag;
+		out_le32(&reg->cqr, val32);
+	}
+
+	/* Wait command completed for 10s */
+	if (ata_wait_register(&reg->ccr, (1 << tag), (1 << tag), 10000)) {
+		if (!is_ncq)
+			printf("Non-NCQ command time out\n\r");
+		else
+			printf("NCQ command time out\n\r");
+	}
+
+	val32 = in_le32(&reg->cer);
+
+	if (val32) {
+		fsl_sata_dump_sfis((struct sata_fis_d2h *)cmd_desc->sfis);
+		printf("CE at device\n\r");
+		fsl_sata_dump_regs(reg);
+		der = in_le32(&reg->der);
+		out_le32(&reg->cer, val32);
+		out_le32(&reg->der, der);
+	}
+
+	/* Clear complete flags */
+	val32 = in_le32(&reg->ccr);
+	out_le32(&reg->ccr, val32);
+
+	return len;
+}
+
+static int fsl_sata_exec_cmd(struct fsl_sata *sata, struct sata_fis_h2d *cfis,
+			     enum cmd_type command_type, int tag, u8 *buffer,
+			     u32 len)
+{
+	int rc;
+
+	if (tag > SATA_HC_MAX_CMD || tag < 0) {
+		printf("tag is out of range, tag=%d\n\r", tag);
+		return -1;
+	}
+
+	switch (command_type) {
+	case CMD_ATA:
+		rc = fsl_ata_exec_ata_cmd(sata, cfis, 0, tag, buffer, len);
+		return rc;
+	case CMD_NCQ:
+		rc = fsl_ata_exec_ata_cmd(sata, cfis, 1, tag, buffer, len);
+		return rc;
+	case CMD_ATAPI:
+	case CMD_VENDOR_BIST:
+	case CMD_BIST:
+		printf("not support now\n\r");
+		return -1;
+	default:
+		break;
+	}
+
+	return -1;
+}
+
+static void fsl_sata_identify(fsl_sata_t *sata, u16 *id)
+{
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = ATA_CMD_ID_ATA;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, (u8 *)id, ATA_ID_WORDS * 2);
+	ata_swap_buf_le16(id, ATA_ID_WORDS);
+}
+
+static void fsl_sata_xfer_mode(fsl_sata_t *sata, u16 *id)
+{
+	sata->pio = id[ATA_ID_PIO_MODES];
+	sata->mwdma = id[ATA_ID_MWDMA_MODES];
+	sata->udma = id[ATA_ID_UDMA_MODES];
+	debug("pio %04x, mwdma %04x, udma %04x\n\r", sata->pio,
+	      sata->mwdma, sata->udma);
+}
+
+static void fsl_sata_init_wcache(fsl_sata_t *sata, u16 *id)
+{
+	if (ata_id_has_wcache(id) && ata_id_wcache_enabled(id))
+		sata->wcache = 1;
+	if (ata_id_has_flush(id))
+		sata->flush = 1;
+	if (ata_id_has_flush_ext(id))
+		sata->flush_ext = 1;
+}
+
+static void fsl_sata_set_features(fsl_sata_t *sata)
+{
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+	u8 udma_cap;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = ATA_CMD_SET_FEATURES;
+	cfis->features = SETFEATURES_XFER;
+
+	/* First check the device capablity */
+	udma_cap = (u8)(sata->udma & 0xff);
+	debug("udma_cap %02x\n\r", udma_cap);
+
+	if (udma_cap == ATA_UDMA6)
+		cfis->sector_count = XFER_UDMA_6;
+	if (udma_cap == ATA_UDMA5)
+		cfis->sector_count = XFER_UDMA_5;
+	if (udma_cap == ATA_UDMA4)
+		cfis->sector_count = XFER_UDMA_4;
+	if (udma_cap == ATA_UDMA3)
+		cfis->sector_count = XFER_UDMA_3;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, NULL, 0);
+}
+
+static u32 fsl_sata_rw_cmd(fsl_sata_t *sata, u32 start, u32 blkcnt,
+			   u8 *buffer, int is_write)
+{
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+	u32 block;
+
+	block = start;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = (is_write) ? ATA_CMD_WRITE : ATA_CMD_READ;
+	cfis->device = ATA_LBA;
+
+	cfis->device |= (block >> 24) & 0xf;
+	cfis->lba_high = (block >> 16) & 0xff;
+	cfis->lba_mid = (block >> 8) & 0xff;
+	cfis->lba_low = block & 0xff;
+	cfis->sector_count = (u8)(blkcnt & 0xff);
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, buffer,
+			  ATA_SECT_SIZE * blkcnt);
+	return blkcnt;
+}
+
+static void fsl_sata_flush_cache(fsl_sata_t *sata)
+{
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = ATA_CMD_FLUSH;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, NULL, 0);
+}
+
+static u32 fsl_sata_rw_cmd_ext(fsl_sata_t *sata, u32 start,
+			       u32 blkcnt, u8 *buffer, int is_write)
+{
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+	u64 block;
+
+	block = (u64)start;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+
+	cfis->command = (is_write) ? ATA_CMD_WRITE_EXT
+				 : ATA_CMD_READ_EXT;
+
+	cfis->lba_high_exp = (block >> 40) & 0xff;
+	cfis->lba_mid_exp = (block >> 32) & 0xff;
+	cfis->lba_low_exp = (block >> 24) & 0xff;
+	cfis->lba_high = (block >> 16) & 0xff;
+	cfis->lba_mid = (block >> 8) & 0xff;
+	cfis->lba_low = block & 0xff;
+	cfis->device = ATA_LBA;
+	cfis->sector_count_exp = (blkcnt >> 8) & 0xff;
+	cfis->sector_count = blkcnt & 0xff;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, buffer,
+			  ATA_SECT_SIZE * blkcnt);
+	return blkcnt;
+}
+
+static u32 fsl_sata_rw_ncq_cmd(fsl_sata_t *sata, u32 start, u32 blkcnt,
+			       u8 *buffer,
+			       int is_write)
+{
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+	int ncq_channel;
+	u64 block;
+
+	if (sata->lba48 != 1) {
+		printf("execute FPDMA command on non-LBA48 hard disk\n\r");
+		return -1;
+	}
+
+	block = (u64)start;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+
+	cfis->command = (is_write) ? ATA_CMD_FPDMA_WRITE
+				 : ATA_CMD_FPDMA_READ;
+
+	cfis->lba_high_exp = (block >> 40) & 0xff;
+	cfis->lba_mid_exp = (block >> 32) & 0xff;
+	cfis->lba_low_exp = (block >> 24) & 0xff;
+	cfis->lba_high = (block >> 16) & 0xff;
+	cfis->lba_mid = (block >> 8) & 0xff;
+	cfis->lba_low = block & 0xff;
+
+	cfis->device = ATA_LBA;
+	cfis->features_exp = (blkcnt >> 8) & 0xff;
+	cfis->features = blkcnt & 0xff;
+
+	if (sata->queue_depth >= SATA_HC_MAX_CMD)
+		ncq_channel = SATA_HC_MAX_CMD - 1;
+	else
+		ncq_channel = sata->queue_depth - 1;
+
+	/* Use the latest queue */
+	fsl_sata_exec_cmd(sata, cfis, CMD_NCQ, ncq_channel, buffer,
+			  ATA_SECT_SIZE * blkcnt);
+	return blkcnt;
+}
+
+static void fsl_sata_flush_cache_ext(fsl_sata_t *sata)
+{
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = ATA_CMD_FLUSH_EXT;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, NULL, 0);
+}
+
+static u32 ata_low_level_rw_lba48(fsl_sata_t *sata, u32 blknr, lbaint_t blkcnt,
+				  const void *buffer, int is_write)
+{
+	u32 start, blks;
+	u8 *addr;
+	int max_blks;
+
+	start = blknr;
+	blks = blkcnt;
+	addr = (u8 *)buffer;
+
+	max_blks = ATA_MAX_SECTORS_LBA48;
+	do {
+		if (blks > max_blks) {
+			if (sata->dma_flag != FLAGS_FPDMA)
+				fsl_sata_rw_cmd_ext(sata, start, max_blks,
+						    addr, is_write);
+			else
+				fsl_sata_rw_ncq_cmd(sata, start, max_blks,
+						    addr, is_write);
+			start += max_blks;
+			blks -= max_blks;
+			addr += ATA_SECT_SIZE * max_blks;
+		} else {
+			if (sata->dma_flag != FLAGS_FPDMA)
+				fsl_sata_rw_cmd_ext(sata, start, blks,
+						    addr, is_write);
+			else
+				fsl_sata_rw_ncq_cmd(sata, start, blks,
+						    addr, is_write);
+			start += blks;
+			blks = 0;
+			addr += ATA_SECT_SIZE * blks;
+		}
+	} while (blks != 0);
+
+	return blks;
+}
+
+static u32 ata_low_level_rw_lba28(fsl_sata_t *sata, u32 blknr, u32 blkcnt,
+				  const void *buffer, int is_write)
+{
+	u32 start, blks;
+	u8 *addr;
+	int max_blks;
+
+	start = blknr;
+	blks = blkcnt;
+	addr = (u8 *)buffer;
+
+	max_blks = ATA_MAX_SECTORS;
+	do {
+		if (blks > max_blks) {
+			fsl_sata_rw_cmd(sata, start, max_blks, addr, is_write);
+			start += max_blks;
+			blks -= max_blks;
+			addr += ATA_SECT_SIZE * max_blks;
+		} else {
+			fsl_sata_rw_cmd(sata, start, blks, addr, is_write);
+			start += blks;
+			blks = 0;
+			addr += ATA_SECT_SIZE * blks;
+		}
+	} while (blks != 0);
+
+	return blks;
+}
+
+/*
+ * SATA interface between low level driver and command layer
+ */
+static int sata_read(fsl_sata_t *sata, ulong blknr, lbaint_t blkcnt,
+		     void *buffer)
+{
+	u32 rc;
+
+	if (sata->lba48)
+		rc = ata_low_level_rw_lba48(sata, blknr, blkcnt, buffer,
+					    READ_CMD);
+	else
+		rc = ata_low_level_rw_lba28(sata, blknr, blkcnt, buffer,
+					    READ_CMD);
+	return rc;
+}
+
+static int sata_write(fsl_sata_t *sata, ulong blknr, lbaint_t blkcnt,
+		      const void *buffer)
+{
+	u32 rc;
+
+	if (sata->lba48) {
+		rc = ata_low_level_rw_lba48(sata, blknr, blkcnt, buffer,
+					    WRITE_CMD);
+		if (sata->wcache && sata->flush_ext)
+			fsl_sata_flush_cache_ext(sata);
+	} else {
+		rc = ata_low_level_rw_lba28(sata, blknr, blkcnt, buffer,
+					    WRITE_CMD);
+		if (sata->wcache && sata->flush)
+			fsl_sata_flush_cache(sata);
+	}
+
+	return rc;
+}
+
+int sata_getinfo(fsl_sata_t *sata, u16 *id)
+{
+	/* if no detected link */
+	if (!sata->link)
+		return -EINVAL;
+
+#ifdef CONFIG_LBA48
+	/* Check if support LBA48 */
+	if (ata_id_has_lba48(id)) {
+		sata->lba48 = 1;
+		debug("Device support LBA48\n\r");
+	} else {
+		debug("Device supports LBA28\n\r");
+	}
+#endif
+
+	/* Get the NCQ queue depth from device */
+	sata->queue_depth = ata_id_queue_depth(id);
+
+	/* Get the xfer mode from device */
+	fsl_sata_xfer_mode(sata, id);
+
+	/* Get the write cache status from device */
+	fsl_sata_init_wcache(sata, id);
+
+	/* Set the xfer mode to highest speed */
+	fsl_sata_set_features(sata);
+
+	return 0;
+}
+
+static int fsl_scsi_exec(fsl_sata_t *sata, struct scsi_cmd *pccb,
+			 bool is_write)
+{
+	int ret;
+	u32 temp;
+	u16 blocks = 0;
+	lbaint_t start = 0;
+	u8 *buffer = pccb->pdata;
+
+	/* Retrieve the base LBA number from the ccb structure. */
+	if (pccb->cmd[0] == SCSI_READ16) {
+		memcpy(&start, pccb->cmd + 2, 8);
+		start = be64_to_cpu(start);
+	} else {
+		memcpy(&temp, pccb->cmd + 2, 4);
+		start = be32_to_cpu(temp);
+	}
+
+	if (pccb->cmd[0] == SCSI_READ16)
+		blocks = (((u16)pccb->cmd[13]) << 8) | ((u16)pccb->cmd[14]);
+	else
+		blocks = (((u16)pccb->cmd[7]) << 8) | ((u16)pccb->cmd[8]);
+
+	debug("scsi_ahci: %s %u blocks starting from lba 0x" LBAFU "\n",
+	      is_write ?  "write" : "read", blocks, start);
+
+	if (is_write)
+		ret = sata_write(sata, start, blocks, buffer);
+	else
+		ret = sata_read(sata, start, blocks, buffer);
+
+	return ret;
+}
+
+static char *fsl_ata_id_strcpy(u16 *target, u16 *src, int len)
+{
+	int i;
+
+	for (i = 0; i < len / 2; i++)
+		target[i] = src[i];
+
+	return (char *)target;
+}
+
+static int fsl_ata_scsiop_inquiry(struct ahci_uc_priv *uc_priv,
+				  struct scsi_cmd *pccb,
+				  fsl_sata_t *sata)
+{
+	u8 port;
+	u16 *idbuf;
+
+	ALLOC_CACHE_ALIGN_BUFFER(u16, tmpid, ATA_ID_WORDS);
+
+	/* Clean ccb data buffer */
+	memset(pccb->pdata, 0, pccb->datalen);
+
+	if (pccb->datalen <= 35)
+		return 0;
+
+	/* Read id from sata */
+	port = pccb->target;
+
+	fsl_sata_identify(sata, (u16 *)tmpid);
+
+	if (!uc_priv->ataid[port]) {
+		uc_priv->ataid[port] = malloc(ATA_ID_WORDS * 2);
+		if (!uc_priv->ataid[port]) {
+			printf("%s: No memory for ataid[port]\n", __func__);
+			return -ENOMEM;
+		}
+	}
+
+	idbuf = uc_priv->ataid[port];
+
+	memcpy(idbuf, tmpid, ATA_ID_WORDS * 2);
+
+	memcpy(&pccb->pdata[8], "ATA     ", 8);
+	fsl_ata_id_strcpy((u16 *)&pccb->pdata[16], &idbuf[ATA_ID_PROD], 16);
+	fsl_ata_id_strcpy((u16 *)&pccb->pdata[32], &idbuf[ATA_ID_FW_REV], 4);
+
+	sata_getinfo(sata, (u16 *)idbuf);
+#ifdef DEBUG
+	ata_dump_id(idbuf);
+#endif
+	return 0;
+}
+
+/*
+ * SCSI READ CAPACITY10 command operation.
+ */
+static int fsl_ata_scsiop_read_capacity10(struct ahci_uc_priv *uc_priv,
+					  struct scsi_cmd *pccb)
+{
+	u32 cap;
+	u64 cap64;
+	u32 block_size;
+
+	if (!uc_priv->ataid[pccb->target]) {
+		printf("scsi_ahci: SCSI READ CAPACITY10 command failure."
+		       "\tNo ATA info!\n"
+		       "\tPlease run SCSI command INQUIRY first!\n");
+		return -EPERM;
+	}
+
+	cap64 = ata_id_n_sectors(uc_priv->ataid[pccb->target]);
+	if (cap64 > 0x100000000ULL)
+		cap64 = 0xffffffff;
+
+	cap = cpu_to_be32(cap64);
+	memcpy(pccb->pdata, &cap, sizeof(cap));
+
+	block_size = cpu_to_be32((u32)512);
+	memcpy(&pccb->pdata[4], &block_size, 4);
+
+	return 0;
+}
+
+/*
+ * SCSI READ CAPACITY16 command operation.
+ */
+static int fsl_ata_scsiop_read_capacity16(struct ahci_uc_priv *uc_priv,
+					  struct scsi_cmd *pccb)
+{
+	u64 cap;
+	u64 block_size;
+
+	if (!uc_priv->ataid[pccb->target]) {
+		printf("scsi_ahci: SCSI READ CAPACITY16 command failure."
+		       "\tNo ATA info!\n"
+		       "\tPlease run SCSI command INQUIRY first!\n");
+		return -EPERM;
+	}
+
+	cap = ata_id_n_sectors(uc_priv->ataid[pccb->target]);
+	cap = cpu_to_be64(cap);
+	memcpy(pccb->pdata, &cap, sizeof(cap));
+
+	block_size = cpu_to_be64((u64)512);
+	memcpy(&pccb->pdata[8], &block_size, 8);
+
+	return 0;
+}
+
+/*
+ * SCSI TEST UNIT READY command operation.
+ */
+static int fsl_ata_scsiop_test_unit_ready(struct ahci_uc_priv *uc_priv,
+					  struct scsi_cmd *pccb)
+{
+	return (uc_priv->ataid[pccb->target]) ? 0 : -EPERM;
+}
+
+static int fsl_ahci_scsi_exec(struct udevice *dev, struct scsi_cmd *pccb)
+{
+	struct ahci_uc_priv *uc_priv = dev_get_uclass_priv(dev->parent);
+	struct fsl_ahci_priv *priv = dev_get_priv(dev->parent);
+	fsl_sata_t *sata = priv->fsl_sata;
+	int ret;
+
+	switch (pccb->cmd[0]) {
+	case SCSI_READ16:
+	case SCSI_READ10:
+		ret = fsl_scsi_exec(sata, pccb, 0);
+		break;
+	case SCSI_WRITE10:
+		ret = fsl_scsi_exec(sata, pccb, 1);
+		break;
+	case SCSI_RD_CAPAC10:
+		ret = fsl_ata_scsiop_read_capacity10(uc_priv, pccb);
+		break;
+	case SCSI_RD_CAPAC16:
+		ret = fsl_ata_scsiop_read_capacity16(uc_priv, pccb);
+		break;
+	case SCSI_TST_U_RDY:
+		ret = fsl_ata_scsiop_test_unit_ready(uc_priv, pccb);
+		break;
+	case SCSI_INQUIRY:
+		ret = fsl_ata_scsiop_inquiry(uc_priv, pccb, sata);
+		break;
+	default:
+		printf("Unsupport SCSI command 0x%02x\n", pccb->cmd[0]);
+		return -ENOTSUPP;
+	}
+
+	if (ret) {
+		debug("SCSI command 0x%02x ret errno %d\n", pccb->cmd[0], ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int fsl_ahci_probe(struct udevice *dev)
+{
+	struct fsl_ahci_priv *priv = dev_get_priv(dev);
+	struct udevice *child_dev;
+	struct scsi_platdata *uc_plat;
+
+	device_find_first_child(dev, &child_dev);
+	if (!child_dev)
+		return -ENODEV;
+	uc_plat = dev_get_uclass_platdata(child_dev);
+	uc_plat->base = priv->base;
+	uc_plat->max_lun = 1;
+	uc_plat->max_id = 1;
+
+	return init_sata(priv);
+}
+
+struct scsi_ops fsl_scsi_ops = {
+	.exec		= fsl_ahci_scsi_exec,
+};
+
+static const struct udevice_id fsl_ahci_ids[] = {
+	{ .compatible = "fsl,pq-sata-v2" },
+	{ }
+};
+
+U_BOOT_DRIVER(fsl_ahci_scsi) = {
+	.name		= "fsl_ahci_scsi",
+	.id		= UCLASS_SCSI,
+	.ops		= &fsl_scsi_ops,
+};
+
+U_BOOT_DRIVER(fsl_ahci) = {
+	.name	= "fsl_ahci",
+	.id	= UCLASS_AHCI,
+	.of_match = fsl_ahci_ids,
+	.bind	= fsl_ahci_bind,
+	.ofdata_to_platdata = fsl_ahci_ofdata_to_platdata,
+	.probe	= fsl_ahci_probe,
+	.priv_auto_alloc_size = sizeof(struct fsl_ahci_priv),
+};
drivers/ata/fsl_sata.h +0 -0
index 1e2da10b02..a4ee83d187 100644

drivers/ata/fsl_sata.h typedef struct fsl_sata { · line 318

Quoted 7 lines
drivers/ata/fsl_sata.h +1 -0
@@ -312,6 +312,7 @@ typedef struct fsl_sata {
  	int		wcache;
  	int		flush;
  	int		flush_ext;
+	u32		dma_flag;
  } fsl_sata_t;
 
  #define READ_CMD	0

Viele Grüße,
Stefan

Signature · 3 lines
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr at denx.de
author
#8 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 54 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:01
To: Peng Ma <[email protected]>; Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 4 lines · Peng Ma

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ====================== This board

does

Quoted 10 lines

not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release.
Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

---
depends on:

	-

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpatchw
ork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99168&amp;data =02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef 3%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072 966&amp;sdata=3Z33Z5raG%2BnbtSUpz2kPCGpefk1byOgy0%2Br3R4DUFU8% 3D&amp;reserved=0

Quoted 3 lines
	-

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpatc

hwork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99167&amp;d ata

Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef 3%7C686e

Quoted 1 line

a1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072966&amp; sdata=sr

Quoted 1 line

%2BCh4UioJw0kuhOiLhc3F6keRhIH8Wflt%2BvMJnHtsY%3D&amp;reserved=

0

Quoted 4 lines
  drivers/ata/Kconfig    |   10 +
  drivers/ata/Makefile   |    1 +
  drivers/ata/fsl_ahci.c | 1030
Quoted 7 lines

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
  drivers/ata/fsl_sata.h |    1 +
  4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

Best Regards,
Peng

Quoted 1,071 lines

Thanks,
Stefan

Quoted 674 lines
drivers/ata/Kconfig index renamed from drivers/ata/Kconfig +0 -0

49a056e941..efac29c709 100644

drivers/ata/Kconfig +10 -0
@@ -59,6 +59,16 @@ config DWC_AHCI
  	  Enable this driver to support Sata devices through
  	  Synopsys DWC AHCI module.
 
+config FSL_AHCI
+	bool "Enable Freescale AHCI driver support"
+	select SCSI_AHCI
+	depends on AHCI
+	depends on DM_SCSI
+	help
+	  Enable this driver to support Sata devices found in
+	  some Freescale PowerPC SoCs.
+
+
  config DWC_AHSATA
  	bool "Enable DWC AHSATA driver support"
  	select LIBATA
drivers/ata/Makefile index renamed from drivers/ata/Makefile +0 -0

10bed53bb3..93aabf34c7 100644

drivers/ata/Makefile +1 -0
@@ -4,6 +4,7 @@
  # Wolfgang Denk, DENX Software Engineering, wd at denx.de.
 
  obj-$(CONFIG_DWC_AHCI) += dwc_ahci.o
+obj-$(CONFIG_FSL_AHCI) += fsl_ahci.o
  obj-$(CONFIG_AHCI) += ahci-uclass.o
  obj-$(CONFIG_AHCI_PCI) += ahci-pci.o
  obj-$(CONFIG_SCSI_AHCI) += ahci.o
drivers/ata/fsl_ahci.c new file renamed from drivers/ata/fsl_ahci.c +0 -0

mode 100644 index 0000000000..16c6f7a335

drivers/ata/fsl_ahci.c new file +636 -0
@@ -0,0 +1,1030 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * NXP PPC SATA platform driver
+ *
+ * (C) Copyright 2019 NXP, Inc.
+ *
+ */
+#include <common.h>
+#include <asm/fsl_serdes.h>
+#include <dm/lists.h>
+#include <dm.h>
+#include <ahci.h>
+#include <scsi.h>
+#include <libata.h>
+#include <sata.h>
+#include <malloc.h>
+#include <memalign.h>
+#include <fis.h>
+
+#include "fsl_sata.h"
+
+struct fsl_ahci_priv {
+	u32 base;
+	u32 flag;
+	u32 number;
+	fsl_sata_t *fsl_sata;
+};
+
+static int fsl_ahci_bind(struct udevice *dev) {
+	return device_bind_driver(dev, "fsl_ahci_scsi", "fsl_ahci_scsi",
+NULL); }
+
+static int fsl_ahci_ofdata_to_platdata(struct udevice *dev) {
+	struct fsl_ahci_priv *priv = dev_get_priv(dev);
+
+	priv->number = dev_read_u32_default(dev, "sata-number", -1);
+	priv->flag = dev_read_u32_default(dev, "sata-fpdma", -1);
+
+	priv->base = dev_read_addr(dev);
+	if (priv->base == FDT_ADDR_T_NONE)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int ata_wait_register(unsigned __iomem *addr, u32 mask,
+			     u32 val, u32 timeout_msec)
+{
+	int i;
+
+	for (i = 0; ((in_le32(addr) & mask) != val) && i < timeout_msec; i++)
+		mdelay(1);
+
+	return (i < timeout_msec) ? 0 : -1;
+}
+
+static void fsl_sata_dump_sfis(struct sata_fis_d2h *s) {
+	printf("Status FIS dump:\n\r");
+	printf("fis_type:		%02x\n\r", s->fis_type);
+	printf("pm_port_i:		%02x\n\r", s->pm_port_i);
+	printf("status:			%02x\n\r", s->status);
+	printf("error:			%02x\n\r", s->error);
+	printf("lba_low:		%02x\n\r", s->lba_low);
+	printf("lba_mid:		%02x\n\r", s->lba_mid);
+	printf("lba_high:		%02x\n\r", s->lba_high);
+	printf("device:			%02x\n\r", s->device);
+	printf("lba_low_exp:		%02x\n\r", s->lba_low_exp);
+	printf("lba_mid_exp:		%02x\n\r", s->lba_mid_exp);
+	printf("lba_high_exp:		%02x\n\r", s->lba_high_exp);
+	printf("res1:			%02x\n\r", s->res1);
+	printf("sector_count:		%02x\n\r", s->sector_count);
+	printf("sector_count_exp:	%02x\n\r", s->sector_count_exp);
+}
+
+static void fsl_sata_dump_regs(fsl_sata_reg_t __iomem *reg) {
+	printf("\n\rSATA:           %08x\n\r", (u32)reg);
+	printf("CQR:            %08x\n\r", in_le32(&reg->cqr));
+	printf("CAR:            %08x\n\r", in_le32(&reg->car));
+	printf("CCR:            %08x\n\r", in_le32(&reg->ccr));
+	printf("CER:            %08x\n\r", in_le32(&reg->cer));
+	printf("CQR:            %08x\n\r", in_le32(&reg->cqr));
+	printf("DER:            %08x\n\r", in_le32(&reg->der));
+	printf("CHBA:           %08x\n\r", in_le32(&reg->chba));
+	printf("HStatus:        %08x\n\r", in_le32(&reg->hstatus));
+	printf("HControl:       %08x\n\r", in_le32(&reg->hcontrol));
+	printf("CQPMP:          %08x\n\r", in_le32(&reg->cqpmp));
+	printf("SIG:            %08x\n\r", in_le32(&reg->sig));
+	printf("ICC:            %08x\n\r", in_le32(&reg->icc));
+	printf("SStatus:        %08x\n\r", in_le32(&reg->sstatus));
+	printf("SError:         %08x\n\r", in_le32(&reg->serror));
+	printf("SControl:       %08x\n\r", in_le32(&reg->scontrol));
+	printf("SNotification:  %08x\n\r", in_le32(&reg->snotification));
+	printf("TransCfg:       %08x\n\r", in_le32(&reg->transcfg));
+	printf("TransStatus:    %08x\n\r", in_le32(&reg->transstatus));
+	printf("LinkCfg:        %08x\n\r", in_le32(&reg->linkcfg));
+	printf("LinkCfg1:       %08x\n\r", in_le32(&reg->linkcfg1));
+	printf("LinkCfg2:       %08x\n\r", in_le32(&reg->linkcfg2));
+	printf("LinkStatus:     %08x\n\r", in_le32(&reg->linkstatus));
+	printf("LinkStatus1:    %08x\n\r", in_le32(&reg->linkstatus1));
+	printf("PhyCtrlCfg:     %08x\n\r", in_le32(&reg->phyctrlcfg));
+	printf("SYSPR:          %08x\n\r", in_be32(&reg->syspr));
+}
+
+static int init_sata(struct fsl_ahci_priv *priv) {
+	int i;
+	u32 cda;
+	u32 val32;
+	u32 sig;
+	fsl_sata_t *sata;
+	u32 length, align;
+	cmd_hdr_tbl_t *cmd_hdr;
+	fsl_sata_reg_t __iomem *reg;
+
+	int dev = priv->number;
+
+	if (dev < 0 || dev > (CONFIG_SYS_SATA_MAX_DEVICE - 1)) {
+		printf("the sata index %d is out of ranges\n\r", dev);
+		return -EINVAL;
+	}
+
+#ifdef CONFIG_MPC85xx
+	if (dev == 0 && (!is_serdes_configured(SATA1))) {
+		printf("SATA%d [dev = %d] is not enabled\n", dev + 1, dev);
+		return -EINVAL;
+	}
+	if (dev == 1 && (!is_serdes_configured(SATA2))) {
+		printf("SATA%d [dev = %d] is not enabled\n", dev + 1, dev);
+		return -EINVAL;
+	}
+#endif
+
+	/* Allocate SATA device driver struct */
+	sata = (fsl_sata_t *)malloc(sizeof(fsl_sata_t));
+	if (!sata) {
+		printf("alloc the sata device struct failed\n\r");
+		return -ENOMEM;
+	}
+	/* Zero all of the device driver struct */
+	memset((void *)sata, 0, sizeof(fsl_sata_t));
+
+	sata->dma_flag = priv->flag;
+	snprintf(sata->name, 12, "SATA%d", dev);
+
+	/* Set the controller register base address to device struct */
+	reg = (fsl_sata_reg_t *)priv->base;
+	sata->reg_base = reg;
+
+	/* Allocate the command header table, 4 bytes aligned */
+	length = sizeof(struct cmd_hdr_tbl);
+	align = SATA_HC_CMD_HDR_TBL_ALIGN;
+	sata->cmd_hdr_tbl_offset = (void *)malloc(length + align);
+	if (!sata->cmd_hdr_tbl_offset) {
+		printf("alloc the command header failed\n\r");
+		return -ENOMEM;
+	}
+
+	cmd_hdr = (cmd_hdr_tbl_t *)(((u32)sata->cmd_hdr_tbl_offset + align)
+						& ~(align - 1));
+	sata->cmd_hdr = cmd_hdr;
+
+	/* Zero all of the command header table */
+	memset((void *)sata->cmd_hdr_tbl_offset, 0, length + align);
+
+	/* Allocate command descriptor for all command */
+	length = sizeof(struct cmd_desc) * SATA_HC_MAX_CMD;
+	align = SATA_HC_CMD_DESC_ALIGN;
+	sata->cmd_desc_offset = (void *)malloc(length + align);
+	if (!sata->cmd_desc_offset) {
+		printf("alloc the command descriptor failed\n\r");
+		return -ENOMEM;
+	}
+	sata->cmd_desc = (cmd_desc_t *)(((u32)sata->cmd_desc_offset + align)
+						& ~(align - 1));
+	/* Zero all of command descriptor */
+	memset((void *)sata->cmd_desc_offset, 0, length + align);
+
+	/* Link the command descriptor to command header */
+	for (i = 0; i < SATA_HC_MAX_CMD; i++) {
+		cda = ((u32)sata->cmd_desc + SATA_HC_CMD_DESC_SIZE * i)
+					 & ~(CMD_HDR_CDA_ALIGN - 1);
+		cmd_hdr->cmd_slot[i].cda = cpu_to_le32(cda);
+	}
+
+	/* To have safe state, force the controller offline */
+	val32 = in_le32(&reg->hcontrol);
+	val32 &= ~HCONTROL_ONOFF;
+	val32 |= HCONTROL_FORCE_OFFLINE;
+	out_le32(&reg->hcontrol, val32);
+
+	/* Wait the controller offline */
+	ata_wait_register(&reg->hstatus, HSTATUS_ONOFF, 0, 1000);
+
+	/* Set the command header base address to CHBA register to tell DMA */
+	out_le32(&reg->chba, (u32)cmd_hdr & ~0x3);
+
+	/* Snoop for the command header */
+	val32 = in_le32(&reg->hcontrol);
+	val32 |= HCONTROL_HDR_SNOOP;
+	out_le32(&reg->hcontrol, val32);
+
+	/* Disable all of interrupts */
+	val32 = in_le32(&reg->hcontrol);
+	val32 &= ~HCONTROL_INT_EN_ALL;
+	out_le32(&reg->hcontrol, val32);
+
+	/* Clear all of interrupts */
+	val32 = in_le32(&reg->hstatus);
+	out_le32(&reg->hstatus, val32);
+
+	/* Set the ICC, no interrupt coalescing */
+	out_le32(&reg->icc, 0x01000000);
+
+	/* No PM attatched, the SATA device direct connect */
+	out_le32(&reg->cqpmp, 0);
+
+	/* Clear SError register */
+	val32 = in_le32(&reg->serror);
+	out_le32(&reg->serror, val32);
+
+	/* Clear CER register */
+	val32 = in_le32(&reg->cer);
+	out_le32(&reg->cer, val32);
+
+	/* Clear DER register */
+	val32 = in_le32(&reg->der);
+	out_le32(&reg->der, val32);
+
+	/* No device detection or initialization action requested */
+	out_le32(&reg->scontrol, 0x00000300);
+
+	/* Configure the transport layer, default value */
+	out_le32(&reg->transcfg, 0x08000016);
+
+	/* Configure the link layer, default value */
+	out_le32(&reg->linkcfg, 0x0000ff34);
+
+	/* Bring the controller online */
+	val32 = in_le32(&reg->hcontrol);
+	val32 |= HCONTROL_ONOFF;
+	out_le32(&reg->hcontrol, val32);
+
+	mdelay(100);
+
+	/* print sata device name */
+	printf("%s ", sata->name);
+
+	/* Wait PHY RDY signal changed for 500ms */
+	ata_wait_register(&reg->hstatus, HSTATUS_PHY_RDY,
+			  HSTATUS_PHY_RDY, 500);
+
+	/* Check PHYRDY */
+	val32 = in_le32(&reg->hstatus);
+	if (val32 & HSTATUS_PHY_RDY) {
+		sata->link = 1;
+	} else {
+		sata->link = 0;
+		printf("(No RDY)\n\r");
+		return -EINVAL;
+	}
+
+	/* Wait for signature updated, which is 1st D2H */
+	ata_wait_register(&reg->hstatus, HSTATUS_SIGNATURE,
+			  HSTATUS_SIGNATURE, 10000);
+
+	if (val32 & HSTATUS_SIGNATURE) {
+		sig = in_le32(&reg->sig);
+		debug("Signature updated, the sig =%08x\n\r", sig);
+		sata->ata_device_type = ata_dev_classify(sig);
+	}
+
+	/* Check the speed */
+	val32 = in_le32(&reg->sstatus);
+	if ((val32 & SSTATUS_SPD_MASK) == SSTATUS_SPD_GEN1)
+		printf("(1.5 Gbps)\n\r");
+	else if ((val32 & SSTATUS_SPD_MASK) == SSTATUS_SPD_GEN2)
+		printf("(3 Gbps)\n\r");
+
+	priv->fsl_sata = sata;
+
+	return 0;
+}
+
+static int fsl_ata_exec_ata_cmd(struct fsl_sata *sata,
+				struct sata_fis_h2d *cfis,
+				int is_ncq, int tag,
+				u8 *buffer, u32 len)
+{
+	cmd_hdr_entry_t *cmd_hdr;
+	cmd_desc_t *cmd_desc;
+	sata_fis_h2d_t *h2d;
+	prd_entry_t *prde;
+	u32 ext_c_ddc;
+	u32 prde_count;
+	u32 val32;
+	u32 ttl;
+	u32 der;
+	int i;
+
+	fsl_sata_reg_t *reg = sata->reg_base;
+
+	/* Check xfer length */
+	if (len > SATA_HC_MAX_XFER_LEN) {
+		printf("max transfer length is 64MB\n\r");
+		return 0;
+	}
+
+	/* Setup the command descriptor */
+	cmd_desc = sata->cmd_desc + tag;
+
+	/* Get the pointer cfis of command descriptor */
+	h2d = (sata_fis_h2d_t *)cmd_desc->cfis;
+
+	/* Zero the cfis of command descriptor */
+	memset((void *)h2d, 0, SATA_HC_CMD_DESC_CFIS_SIZE);
+
+	/* Copy the cfis from user to command descriptor */
+	h2d->fis_type = cfis->fis_type;
+	h2d->pm_port_c = cfis->pm_port_c;
+	h2d->command = cfis->command;
+
+	h2d->features = cfis->features;
+	h2d->features_exp = cfis->features_exp;
+
+	h2d->lba_low = cfis->lba_low;
+	h2d->lba_mid = cfis->lba_mid;
+	h2d->lba_high = cfis->lba_high;
+	h2d->lba_low_exp = cfis->lba_low_exp;
+	h2d->lba_mid_exp = cfis->lba_mid_exp;
+	h2d->lba_high_exp = cfis->lba_high_exp;
+
+	if (!is_ncq) {
+		h2d->sector_count = cfis->sector_count;
+		h2d->sector_count_exp = cfis->sector_count_exp;
+	} else { /* NCQ */
+		h2d->sector_count = (u8)(tag << 3);
+	}
+
+	h2d->device = cfis->device;
+	h2d->control = cfis->control;
+
+	/* Setup the PRD table */
+	prde = (prd_entry_t *)cmd_desc->prdt;
+	memset((void *)prde, 0, sizeof(struct prdt));
+
+	prde_count = 0;
+	ttl = len;
+	for (i = 0; i < SATA_HC_MAX_PRD_DIRECT; i++) {
+		if (!len)
+			break;
+		prde->dba = cpu_to_le32((u32)buffer & ~0x3);
+		debug("dba = %08x\n\r", (u32)buffer);
+
+		if (len < PRD_ENTRY_MAX_XFER_SZ) {
+			ext_c_ddc = PRD_ENTRY_DATA_SNOOP | len;
+			debug("ext_c_ddc1 = %08x, len = %08x\n\r",
+			      ext_c_ddc, len);
+			prde->ext_c_ddc = cpu_to_le32(ext_c_ddc);
+			prde_count++;
+			prde++;
+		} else {
+			ext_c_ddc = PRD_ENTRY_DATA_SNOOP; /* 4M bytes */
+			debug("ext_c_ddc2 = %08x, len = %08x\n\r",
+			      ext_c_ddc, len);
+			prde->ext_c_ddc = cpu_to_le32(ext_c_ddc);
+			buffer += PRD_ENTRY_MAX_XFER_SZ;
+			len -= PRD_ENTRY_MAX_XFER_SZ;
+			prde_count++;
+			prde++;
+		}
+	}
+
+	/* Setup the command slot of cmd hdr */
+	cmd_hdr = (cmd_hdr_entry_t *)&sata->cmd_hdr->cmd_slot[tag];
+
+	cmd_hdr->cda = cpu_to_le32((u32)cmd_desc & ~0x3);
+
+	val32 = prde_count << CMD_HDR_PRD_ENTRY_SHIFT;
+	val32 |= sizeof(sata_fis_h2d_t);
+	cmd_hdr->prde_fis_len = cpu_to_le32(val32);
+
+	cmd_hdr->ttl = cpu_to_le32(ttl);
+
+	if (!is_ncq)
+		val32 = CMD_HDR_ATTR_RES | CMD_HDR_ATTR_SNOOP;
+	else
+		val32 = CMD_HDR_ATTR_RES | CMD_HDR_ATTR_SNOOP |
+			CMD_HDR_ATTR_FPDMA;
+
+	tag &= CMD_HDR_ATTR_TAG;
+	val32 |= tag;
+
+	debug("attribute = %08x\n\r", val32);
+	cmd_hdr->attribute = cpu_to_le32(val32);
+
+	/* Make sure cmd desc and cmd slot valid before command issue */
+	sync();
+
+	/* PMP*/
+	val32 = (u32)(h2d->pm_port_c & 0x0f);
+	out_le32(&reg->cqpmp, val32);
+
+	/* Wait no active */
+	if (ata_wait_register(&reg->car, (1 << tag), 0, 10000))
+		printf("Wait no active time out\n\r");
+
+	/* Issue command */
+	if (!(in_le32(&reg->cqr) & (1 << tag))) {
+		val32 = 1 << tag;
+		out_le32(&reg->cqr, val32);
+	}
+
+	/* Wait command completed for 10s */
+	if (ata_wait_register(&reg->ccr, (1 << tag), (1 << tag), 10000)) {
+		if (!is_ncq)
+			printf("Non-NCQ command time out\n\r");
+		else
+			printf("NCQ command time out\n\r");
+	}
+
+	val32 = in_le32(&reg->cer);
+
+	if (val32) {
+		fsl_sata_dump_sfis((struct sata_fis_d2h *)cmd_desc->sfis);
+		printf("CE at device\n\r");
+		fsl_sata_dump_regs(reg);
+		der = in_le32(&reg->der);
+		out_le32(&reg->cer, val32);
+		out_le32(&reg->der, der);
+	}
+
+	/* Clear complete flags */
+	val32 = in_le32(&reg->ccr);
+	out_le32(&reg->ccr, val32);
+
+	return len;
+}
+
+static int fsl_sata_exec_cmd(struct fsl_sata *sata, struct sata_fis_h2d *cfis,
+			     enum cmd_type command_type, int tag, u8 *buffer,
+			     u32 len)
+{
+	int rc;
+
+	if (tag > SATA_HC_MAX_CMD || tag < 0) {
+		printf("tag is out of range, tag=%d\n\r", tag);
+		return -1;
+	}
+
+	switch (command_type) {
+	case CMD_ATA:
+		rc = fsl_ata_exec_ata_cmd(sata, cfis, 0, tag, buffer, len);
+		return rc;
+	case CMD_NCQ:
+		rc = fsl_ata_exec_ata_cmd(sata, cfis, 1, tag, buffer, len);
+		return rc;
+	case CMD_ATAPI:
+	case CMD_VENDOR_BIST:
+	case CMD_BIST:
+		printf("not support now\n\r");
+		return -1;
+	default:
+		break;
+	}
+
+	return -1;
+}
+
+static void fsl_sata_identify(fsl_sata_t *sata, u16 *id) {
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = ATA_CMD_ID_ATA;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, (u8 *)id, ATA_ID_WORDS * 2);
+	ata_swap_buf_le16(id, ATA_ID_WORDS); }
+
+static void fsl_sata_xfer_mode(fsl_sata_t *sata, u16 *id) {
+	sata->pio = id[ATA_ID_PIO_MODES];
+	sata->mwdma = id[ATA_ID_MWDMA_MODES];
+	sata->udma = id[ATA_ID_UDMA_MODES];
+	debug("pio %04x, mwdma %04x, udma %04x\n\r", sata->pio,
+	      sata->mwdma, sata->udma);
+}
+
+static void fsl_sata_init_wcache(fsl_sata_t *sata, u16 *id) {
+	if (ata_id_has_wcache(id) && ata_id_wcache_enabled(id))
+		sata->wcache = 1;
+	if (ata_id_has_flush(id))
+		sata->flush = 1;
+	if (ata_id_has_flush_ext(id))
+		sata->flush_ext = 1;
+}
+
+static void fsl_sata_set_features(fsl_sata_t *sata) {
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+	u8 udma_cap;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = ATA_CMD_SET_FEATURES;
+	cfis->features = SETFEATURES_XFER;
+
+	/* First check the device capablity */
+	udma_cap = (u8)(sata->udma & 0xff);
+	debug("udma_cap %02x\n\r", udma_cap);
+
+	if (udma_cap == ATA_UDMA6)
+		cfis->sector_count = XFER_UDMA_6;
+	if (udma_cap == ATA_UDMA5)
+		cfis->sector_count = XFER_UDMA_5;
+	if (udma_cap == ATA_UDMA4)
+		cfis->sector_count = XFER_UDMA_4;
+	if (udma_cap == ATA_UDMA3)
+		cfis->sector_count = XFER_UDMA_3;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, NULL, 0); }
+
+static u32 fsl_sata_rw_cmd(fsl_sata_t *sata, u32 start, u32 blkcnt,
+			   u8 *buffer, int is_write)
+{
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+	u32 block;
+
+	block = start;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = (is_write) ? ATA_CMD_WRITE : ATA_CMD_READ;
+	cfis->device = ATA_LBA;
+
+	cfis->device |= (block >> 24) & 0xf;
+	cfis->lba_high = (block >> 16) & 0xff;
+	cfis->lba_mid = (block >> 8) & 0xff;
+	cfis->lba_low = block & 0xff;
+	cfis->sector_count = (u8)(blkcnt & 0xff);
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, buffer,
+			  ATA_SECT_SIZE * blkcnt);
+	return blkcnt;
+}
+
+static void fsl_sata_flush_cache(fsl_sata_t *sata) {
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = ATA_CMD_FLUSH;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, NULL, 0); }
+
+static u32 fsl_sata_rw_cmd_ext(fsl_sata_t *sata, u32 start,
+			       u32 blkcnt, u8 *buffer, int is_write) {
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+	u64 block;
+
+	block = (u64)start;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+
+	cfis->command = (is_write) ? ATA_CMD_WRITE_EXT
+				 : ATA_CMD_READ_EXT;
+
+	cfis->lba_high_exp = (block >> 40) & 0xff;
+	cfis->lba_mid_exp = (block >> 32) & 0xff;
+	cfis->lba_low_exp = (block >> 24) & 0xff;
+	cfis->lba_high = (block >> 16) & 0xff;
+	cfis->lba_mid = (block >> 8) & 0xff;
+	cfis->lba_low = block & 0xff;
+	cfis->device = ATA_LBA;
+	cfis->sector_count_exp = (blkcnt >> 8) & 0xff;
+	cfis->sector_count = blkcnt & 0xff;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, buffer,
+			  ATA_SECT_SIZE * blkcnt);
+	return blkcnt;
+}
+
+static u32 fsl_sata_rw_ncq_cmd(fsl_sata_t *sata, u32 start, u32 blkcnt,
+			       u8 *buffer,
+			       int is_write)
+{
+	struct sata_fis_h2d h2d, *cfis = &h2d;
+	int ncq_channel;
+	u64 block;
+
+	if (sata->lba48 != 1) {
+		printf("execute FPDMA command on non-LBA48 hard disk\n\r");
+		return -1;
+	}
+
+	block = (u64)start;
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+
+	cfis->command = (is_write) ? ATA_CMD_FPDMA_WRITE
+				 : ATA_CMD_FPDMA_READ;
+
+	cfis->lba_high_exp = (block >> 40) & 0xff;
+	cfis->lba_mid_exp = (block >> 32) & 0xff;
+	cfis->lba_low_exp = (block >> 24) & 0xff;
+	cfis->lba_high = (block >> 16) & 0xff;
+	cfis->lba_mid = (block >> 8) & 0xff;
+	cfis->lba_low = block & 0xff;
+
+	cfis->device = ATA_LBA;
+	cfis->features_exp = (blkcnt >> 8) & 0xff;
+	cfis->features = blkcnt & 0xff;
+
+	if (sata->queue_depth >= SATA_HC_MAX_CMD)
+		ncq_channel = SATA_HC_MAX_CMD - 1;
+	else
+		ncq_channel = sata->queue_depth - 1;
+
+	/* Use the latest queue */
+	fsl_sata_exec_cmd(sata, cfis, CMD_NCQ, ncq_channel, buffer,
+			  ATA_SECT_SIZE * blkcnt);
+	return blkcnt;
+}
+
+static void fsl_sata_flush_cache_ext(fsl_sata_t *sata) {
+	struct sata_fis_h2d h2d, *cfis = &h2d;

drivers/ata/fsl_ahci.c line 646

Quoted 10 lines
drivers/ata/fsl_ahci.c new file +10 -0
+
+	memset(cfis, 0, sizeof(struct sata_fis_h2d));
+
+	cfis->fis_type = SATA_FIS_TYPE_REGISTER_H2D;
+	cfis->pm_port_c = 0x80; /* is command */
+	cfis->command = ATA_CMD_FLUSH_EXT;
+
+	fsl_sata_exec_cmd(sata, cfis, CMD_ATA, 0, NULL, 0); }
+
+static u32 ata_low_level_rw_lba48(fsl_sata_t *sata, u32 blknr, lbaint_t

blkcnt,

Quoted 367 lines
drivers/ata/fsl_ahci.c new file +361 -0
+				  const void *buffer, int is_write) {
+	u32 start, blks;
+	u8 *addr;
+	int max_blks;
+
+	start = blknr;
+	blks = blkcnt;
+	addr = (u8 *)buffer;
+
+	max_blks = ATA_MAX_SECTORS_LBA48;
+	do {
+		if (blks > max_blks) {
+			if (sata->dma_flag != FLAGS_FPDMA)
+				fsl_sata_rw_cmd_ext(sata, start, max_blks,
+						    addr, is_write);
+			else
+				fsl_sata_rw_ncq_cmd(sata, start, max_blks,
+						    addr, is_write);
+			start += max_blks;
+			blks -= max_blks;
+			addr += ATA_SECT_SIZE * max_blks;
+		} else {
+			if (sata->dma_flag != FLAGS_FPDMA)
+				fsl_sata_rw_cmd_ext(sata, start, blks,
+						    addr, is_write);
+			else
+				fsl_sata_rw_ncq_cmd(sata, start, blks,
+						    addr, is_write);
+			start += blks;
+			blks = 0;
+			addr += ATA_SECT_SIZE * blks;
+		}
+	} while (blks != 0);
+
+	return blks;
+}
+
+static u32 ata_low_level_rw_lba28(fsl_sata_t *sata, u32 blknr, u32 blkcnt,
+				  const void *buffer, int is_write) {
+	u32 start, blks;
+	u8 *addr;
+	int max_blks;
+
+	start = blknr;
+	blks = blkcnt;
+	addr = (u8 *)buffer;
+
+	max_blks = ATA_MAX_SECTORS;
+	do {
+		if (blks > max_blks) {
+			fsl_sata_rw_cmd(sata, start, max_blks, addr, is_write);
+			start += max_blks;
+			blks -= max_blks;
+			addr += ATA_SECT_SIZE * max_blks;
+		} else {
+			fsl_sata_rw_cmd(sata, start, blks, addr, is_write);
+			start += blks;
+			blks = 0;
+			addr += ATA_SECT_SIZE * blks;
+		}
+	} while (blks != 0);
+
+	return blks;
+}
+
+/*
+ * SATA interface between low level driver and command layer  */
+static int sata_read(fsl_sata_t *sata, ulong blknr, lbaint_t blkcnt,
+		     void *buffer)
+{
+	u32 rc;
+
+	if (sata->lba48)
+		rc = ata_low_level_rw_lba48(sata, blknr, blkcnt, buffer,
+					    READ_CMD);
+	else
+		rc = ata_low_level_rw_lba28(sata, blknr, blkcnt, buffer,
+					    READ_CMD);
+	return rc;
+}
+
+static int sata_write(fsl_sata_t *sata, ulong blknr, lbaint_t blkcnt,
+		      const void *buffer)
+{
+	u32 rc;
+
+	if (sata->lba48) {
+		rc = ata_low_level_rw_lba48(sata, blknr, blkcnt, buffer,
+					    WRITE_CMD);
+		if (sata->wcache && sata->flush_ext)
+			fsl_sata_flush_cache_ext(sata);
+	} else {
+		rc = ata_low_level_rw_lba28(sata, blknr, blkcnt, buffer,
+					    WRITE_CMD);
+		if (sata->wcache && sata->flush)
+			fsl_sata_flush_cache(sata);
+	}
+
+	return rc;
+}
+
+int sata_getinfo(fsl_sata_t *sata, u16 *id) {
+	/* if no detected link */
+	if (!sata->link)
+		return -EINVAL;
+
+#ifdef CONFIG_LBA48
+	/* Check if support LBA48 */
+	if (ata_id_has_lba48(id)) {
+		sata->lba48 = 1;
+		debug("Device support LBA48\n\r");
+	} else {
+		debug("Device supports LBA28\n\r");
+	}
+#endif
+
+	/* Get the NCQ queue depth from device */
+	sata->queue_depth = ata_id_queue_depth(id);
+
+	/* Get the xfer mode from device */
+	fsl_sata_xfer_mode(sata, id);
+
+	/* Get the write cache status from device */
+	fsl_sata_init_wcache(sata, id);
+
+	/* Set the xfer mode to highest speed */
+	fsl_sata_set_features(sata);
+
+	return 0;
+}
+
+static int fsl_scsi_exec(fsl_sata_t *sata, struct scsi_cmd *pccb,
+			 bool is_write)
+{
+	int ret;
+	u32 temp;
+	u16 blocks = 0;
+	lbaint_t start = 0;
+	u8 *buffer = pccb->pdata;
+
+	/* Retrieve the base LBA number from the ccb structure. */
+	if (pccb->cmd[0] == SCSI_READ16) {
+		memcpy(&start, pccb->cmd + 2, 8);
+		start = be64_to_cpu(start);
+	} else {
+		memcpy(&temp, pccb->cmd + 2, 4);
+		start = be32_to_cpu(temp);
+	}
+
+	if (pccb->cmd[0] == SCSI_READ16)
+		blocks = (((u16)pccb->cmd[13]) << 8) | ((u16)pccb->cmd[14]);
+	else
+		blocks = (((u16)pccb->cmd[7]) << 8) | ((u16)pccb->cmd[8]);
+
+	debug("scsi_ahci: %s %u blocks starting from lba 0x" LBAFU "\n",
+	      is_write ?  "write" : "read", blocks, start);
+
+	if (is_write)
+		ret = sata_write(sata, start, blocks, buffer);
+	else
+		ret = sata_read(sata, start, blocks, buffer);
+
+	return ret;
+}
+
+static char *fsl_ata_id_strcpy(u16 *target, u16 *src, int len) {
+	int i;
+
+	for (i = 0; i < len / 2; i++)
+		target[i] = src[i];
+
+	return (char *)target;
+}
+
+static int fsl_ata_scsiop_inquiry(struct ahci_uc_priv *uc_priv,
+				  struct scsi_cmd *pccb,
+				  fsl_sata_t *sata)
+{
+	u8 port;
+	u16 *idbuf;
+
+	ALLOC_CACHE_ALIGN_BUFFER(u16, tmpid, ATA_ID_WORDS);
+
+	/* Clean ccb data buffer */
+	memset(pccb->pdata, 0, pccb->datalen);
+
+	if (pccb->datalen <= 35)
+		return 0;
+
+	/* Read id from sata */
+	port = pccb->target;
+
+	fsl_sata_identify(sata, (u16 *)tmpid);
+
+	if (!uc_priv->ataid[port]) {
+		uc_priv->ataid[port] = malloc(ATA_ID_WORDS * 2);
+		if (!uc_priv->ataid[port]) {
+			printf("%s: No memory for ataid[port]\n", __func__);
+			return -ENOMEM;
+		}
+	}
+
+	idbuf = uc_priv->ataid[port];
+
+	memcpy(idbuf, tmpid, ATA_ID_WORDS * 2);
+
+	memcpy(&pccb->pdata[8], "ATA     ", 8);
+	fsl_ata_id_strcpy((u16 *)&pccb->pdata[16], &idbuf[ATA_ID_PROD], 16);
+	fsl_ata_id_strcpy((u16 *)&pccb->pdata[32], &idbuf[ATA_ID_FW_REV],
+4);
+
+	sata_getinfo(sata, (u16 *)idbuf);
+#ifdef DEBUG
+	ata_dump_id(idbuf);
+#endif
+	return 0;
+}
+
+/*
+ * SCSI READ CAPACITY10 command operation.
+ */
+static int fsl_ata_scsiop_read_capacity10(struct ahci_uc_priv *uc_priv,
+					  struct scsi_cmd *pccb)
+{
+	u32 cap;
+	u64 cap64;
+	u32 block_size;
+
+	if (!uc_priv->ataid[pccb->target]) {
+		printf("scsi_ahci: SCSI READ CAPACITY10 command failure."
+		       "\tNo ATA info!\n"
+		       "\tPlease run SCSI command INQUIRY first!\n");
+		return -EPERM;
+	}
+
+	cap64 = ata_id_n_sectors(uc_priv->ataid[pccb->target]);
+	if (cap64 > 0x100000000ULL)
+		cap64 = 0xffffffff;
+
+	cap = cpu_to_be32(cap64);
+	memcpy(pccb->pdata, &cap, sizeof(cap));
+
+	block_size = cpu_to_be32((u32)512);
+	memcpy(&pccb->pdata[4], &block_size, 4);
+
+	return 0;
+}
+
+/*
+ * SCSI READ CAPACITY16 command operation.
+ */
+static int fsl_ata_scsiop_read_capacity16(struct ahci_uc_priv *uc_priv,
+					  struct scsi_cmd *pccb)
+{
+	u64 cap;
+	u64 block_size;
+
+	if (!uc_priv->ataid[pccb->target]) {
+		printf("scsi_ahci: SCSI READ CAPACITY16 command failure."
+		       "\tNo ATA info!\n"
+		       "\tPlease run SCSI command INQUIRY first!\n");
+		return -EPERM;
+	}
+
+	cap = ata_id_n_sectors(uc_priv->ataid[pccb->target]);
+	cap = cpu_to_be64(cap);
+	memcpy(pccb->pdata, &cap, sizeof(cap));
+
+	block_size = cpu_to_be64((u64)512);
+	memcpy(&pccb->pdata[8], &block_size, 8);
+
+	return 0;
+}
+
+/*
+ * SCSI TEST UNIT READY command operation.
+ */
+static int fsl_ata_scsiop_test_unit_ready(struct ahci_uc_priv *uc_priv,
+					  struct scsi_cmd *pccb)
+{
+	return (uc_priv->ataid[pccb->target]) ? 0 : -EPERM; }
+
+static int fsl_ahci_scsi_exec(struct udevice *dev, struct scsi_cmd
+*pccb) {
+	struct ahci_uc_priv *uc_priv = dev_get_uclass_priv(dev->parent);
+	struct fsl_ahci_priv *priv = dev_get_priv(dev->parent);
+	fsl_sata_t *sata = priv->fsl_sata;
+	int ret;
+
+	switch (pccb->cmd[0]) {
+	case SCSI_READ16:
+	case SCSI_READ10:
+		ret = fsl_scsi_exec(sata, pccb, 0);
+		break;
+	case SCSI_WRITE10:
+		ret = fsl_scsi_exec(sata, pccb, 1);
+		break;
+	case SCSI_RD_CAPAC10:
+		ret = fsl_ata_scsiop_read_capacity10(uc_priv, pccb);
+		break;
+	case SCSI_RD_CAPAC16:
+		ret = fsl_ata_scsiop_read_capacity16(uc_priv, pccb);
+		break;
+	case SCSI_TST_U_RDY:
+		ret = fsl_ata_scsiop_test_unit_ready(uc_priv, pccb);
+		break;
+	case SCSI_INQUIRY:
+		ret = fsl_ata_scsiop_inquiry(uc_priv, pccb, sata);
+		break;
+	default:
+		printf("Unsupport SCSI command 0x%02x\n", pccb->cmd[0]);
+		return -ENOTSUPP;
+	}
+
+	if (ret) {
+		debug("SCSI command 0x%02x ret errno %d\n", pccb->cmd[0], ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int fsl_ahci_probe(struct udevice *dev) {
+	struct fsl_ahci_priv *priv = dev_get_priv(dev);
+	struct udevice *child_dev;
+	struct scsi_platdata *uc_plat;
+
+	device_find_first_child(dev, &child_dev);
+	if (!child_dev)
+		return -ENODEV;
+	uc_plat = dev_get_uclass_platdata(child_dev);
+	uc_plat->base = priv->base;
+	uc_plat->max_lun = 1;
+	uc_plat->max_id = 1;
+
+	return init_sata(priv);
+}
+
+struct scsi_ops fsl_scsi_ops = {
+	.exec		= fsl_ahci_scsi_exec,
+};
+
+static const struct udevice_id fsl_ahci_ids[] = {
+	{ .compatible = "fsl,pq-sata-v2" },
+	{ }
+};
+
+U_BOOT_DRIVER(fsl_ahci_scsi) = {
+	.name		= "fsl_ahci_scsi",
+	.id		= UCLASS_SCSI,
+	.ops		= &fsl_scsi_ops,
+};
+
+U_BOOT_DRIVER(fsl_ahci) = {
+	.name	= "fsl_ahci",
+	.id	= UCLASS_AHCI,
+	.of_match = fsl_ahci_ids,
+	.bind	= fsl_ahci_bind,
+	.ofdata_to_platdata = fsl_ahci_ofdata_to_platdata,
+	.probe	= fsl_ahci_probe,
+	.priv_auto_alloc_size = sizeof(struct fsl_ahci_priv), };
drivers/ata/fsl_sata.h index renamed from drivers/ata/fsl_sata.h +0 -0

1e2da10b02..a4ee83d187 100644

drivers/ata/fsl_sata.h typedef struct fsl_sata { · line 318

Quoted 7 lines
drivers/ata/fsl_sata.h +1 -0
@@ -312,6 +312,7 @@ typedef struct fsl_sata {
  	int		wcache;
  	int		flush;
  	int		flush_ext;
+	u32		dma_flag;
  } fsl_sata_t;
 
  #define READ_CMD	0

Viele Grüße,
Stefan

--
DENX Software Engineering GmbH, Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr at denx.de

#9 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 61 lines · Peng Ma
Quoted 54 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:01
To: Peng Ma <[email protected]>; Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 4 lines · Peng Ma

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ====================== This board

does

Quoted 10 lines

not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release.
Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

---
depends on:

	-

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpatchw
ork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99168&amp;data =02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef 3%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072 966&amp;sdata=3Z33Z5raG%2BnbtSUpz2kPCGpefk1byOgy0%2Br3R4DUFU8% 3D&amp;reserved=0

Quoted 3 lines
	-

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpatc

hwork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99167&amp;d ata

Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef 3%7C686e

Quoted 1 line

a1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072966&amp; sdata=sr

Quoted 1 line

%2BCh4UioJw0kuhOiLhc3F6keRhIH8Wflt%2BvMJnHtsY%3D&amp;reserved=

0

Quoted 4 lines
   drivers/ata/Kconfig    |   10 +
   drivers/ata/Makefile   |    1 +
   drivers/ata/fsl_ahci.c | 1030
Quoted 7 lines

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
   drivers/ata/fsl_sata.h |    1 +
   4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

Quoted 7 lines

[Peng Ma]
Hi Stefan,

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver
When all of our powerpc socs support DM.

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

Thanks,
Stefan

author
#10 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 93 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:48
To: Peng Ma <[email protected]>; Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:01
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 9 lines
Quoted 9 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 4 lines · Peng Ma

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ====================== This

board

Quoted 7 lines
Quoted 7 lines

does

Quoted 6 lines

not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release.
Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

Quoted 7 lines
Quoted 7 lines
Quoted 4 lines

---
depends on:

	-

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpat
chw

ork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99168&amp;data

Quoted 1 line
Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 1 line
Quoted 1 line

3%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072

Quoted 1 line
Quoted 1 line

966&amp;sdata=3Z33Z5raG%2BnbtSUpz2kPCGpefk1byOgy0%2Br3R4DUFU8%

Quoted 6 lines
Quoted 6 lines

3D&amp;reserved=0

Quoted 4 lines
	-

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpa
tc

hwork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99167&amp;d

Quoted 3 lines
Quoted 3 lines

ata

Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 3 lines
Quoted 3 lines

3%7C686e

Quoted 1 line

a1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072966&amp;

Quoted 2 lines
Quoted 2 lines

sdata=sr

Quoted 1 line

%2BCh4UioJw0kuhOiLhc3F6keRhIH8Wflt%2BvMJnHtsY%3D&amp;reserve

d=

Quoted 15 lines
Quoted 5 lines

0

Quoted 4 lines
   drivers/ata/Kconfig    |   10 +
   drivers/ata/Makefile   |    1 +
   drivers/ata/fsl_ahci.c | 1030
Quoted 7 lines

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
   drivers/ata/fsl_sata.h |    1 +
   4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

Quoted 6 lines
Quoted 3 lines

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

[Peng Ma]
Hi Stefan,

I am so sorry to reply late, Other Non dts powerpc board need the old sata driver, so the new Sata driver does not replace the so far.

Best Regards,
Peng

Quoted 3 lines

Thanks,
Stefan

#11 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 97 lines · Peng Ma
Quoted 97 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:48
To: Peng Ma <[email protected]>; Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:01
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 9 lines
Quoted 9 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 4 lines · Peng Ma

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ====================== This

board

Quoted 7 lines
Quoted 7 lines

does

Quoted 6 lines

not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release.
Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

Quoted 7 lines
Quoted 7 lines
Quoted 4 lines

---
depends on:

	-

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpat
chw

ork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99168&amp;data

Quoted 1 line
Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 1 line
Quoted 1 line

3%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072

Quoted 1 line
Quoted 1 line

966&amp;sdata=3Z33Z5raG%2BnbtSUpz2kPCGpefk1byOgy0%2Br3R4DUFU8%

Quoted 6 lines
Quoted 6 lines

3D&amp;reserved=0

Quoted 4 lines
	-

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpa
tc

hwork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99167&amp;d

Quoted 3 lines
Quoted 3 lines

ata

Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 3 lines
Quoted 3 lines

3%7C686e

Quoted 1 line

a1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072966&amp;

Quoted 2 lines
Quoted 2 lines

sdata=sr

Quoted 1 line

%2BCh4UioJw0kuhOiLhc3F6keRhIH8Wflt%2BvMJnHtsY%3D&amp;reserve

d=

Quoted 18 lines
Quoted 5 lines

0

Quoted 4 lines
    drivers/ata/Kconfig    |   10 +
    drivers/ata/Makefile   |    1 +
    drivers/ata/fsl_ahci.c | 1030
Quoted 7 lines

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
    drivers/ata/fsl_sata.h |    1 +
    4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

Quoted 8 lines
Quoted 2 lines

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

[Peng Ma]
Hi Stefan,

I am so sorry to reply late, Other Non dts powerpc board need the old sata driver, so the new
Sata driver does not replace the so far.

Just curious: Which are the "other non dts powerpc boards"? Are there still many? Is there a plan to move them to DT as well? Or should they perhaps be dropped from mainline if not converted to DT? What's stopping the conversion here?

Thanks,
Stefan

author
#12 lore raw
Quoted 132 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年5月15日 17:23
To: Peng Ma <[email protected]>; Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>;
bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: [EXT] Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Caution: EXT Email

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:48
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 39 lines
Quoted 39 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:01
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 9 lines
Quoted 9 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 4 lines · Peng Ma

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ====================== This

board

Quoted 11 lines
Quoted 11 lines

does

Quoted 10 lines

not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release. Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

---
depends on:
-

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fp
at
chw

ork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99168&amp;data

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

3%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

966&amp;sdata=3Z33Z5raG%2BnbtSUpz2kPCGpefk1byOgy0%2Br3R4DUFU8%

Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
Quoted 7 lines

3D&amp;reserved=0

Quoted 5 lines

-
https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2F
pa
tc

hwork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99167&amp;d

Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
Quoted 3 lines

ata

Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
Quoted 3 lines

3%7C686e

Quoted 1 line

a1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072966&amp;

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

sdata=sr

Quoted 1 line

%2BCh4UioJw0kuhOiLhc3F6keRhIH8Wflt%2BvMJnHtsY%3D&amp;reser

ve

Quoted 25 lines
Quoted 16 lines

d=

Quoted 15 lines
Quoted 5 lines

0

Quoted 4 lines
    drivers/ata/Kconfig    |   10 +
    drivers/ata/Makefile   |    1 +
    drivers/ata/fsl_ahci.c | 1030
Quoted 7 lines

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
    drivers/ata/fsl_sata.h |    1 +
    4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

Quoted 6 lines
Quoted 3 lines

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

[Peng Ma]
Hi Stefan,

Quoted 7 lines
Quoted 2 lines

I am so sorry to reply late, Other Non dts powerpc board need the old sata driver, so the new Sata driver does not replace the so far.

Just curious: Which are the "other non dts powerpc boards"? Are there still many? Is there a plan to move them to DT as well? Or should they perhaps be dropped from mainline if not converted to DT? What's stopping the conversion here?

[Peng Ma]
1: There is some platforms still with non dts such as: T1024QDS, T1024RDB, T1040D4RDB etc. 2: I do not know the plan to move them to DT, we only support one platform named T2080QDS so far. 3:Before all of them support DT, we should not delete old sata driver.

Thanks.
Best Regards,
Peng

Quoted 3 lines

Thanks,
Stefan

author
#13 lore raw
In reply to Stefan Roese

Hi Prabhakar,

I see our platforms with powerpc. We only support T2080QDS with DTS. Is there a plan to move them to DT as well? Please let Stefan Know.
Thanks.

Best Regards,
Peng

Quoted 140 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年5月15日 17:23
To: Peng Ma <[email protected]>; Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>;
bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: [EXT] Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Caution: EXT Email

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:48
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 39 lines
Quoted 39 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:01
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 9 lines
Quoted 9 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 4 lines · Peng Ma

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ====================== This

board

Quoted 11 lines
Quoted 11 lines

does

Quoted 10 lines

not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release. Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

---
depends on:
-

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fp
at
chw

ork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99168&amp;data

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

3%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

966&amp;sdata=3Z33Z5raG%2BnbtSUpz2kPCGpefk1byOgy0%2Br3R4DUFU8%

Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
Quoted 7 lines

3D&amp;reserved=0

Quoted 5 lines

-
https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2F
pa
tc

hwork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99167&amp;d

Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
Quoted 3 lines

ata

Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
Quoted 3 lines

3%7C686e

Quoted 1 line

a1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072966&amp;

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

sdata=sr

Quoted 1 line

%2BCh4UioJw0kuhOiLhc3F6keRhIH8Wflt%2BvMJnHtsY%3D&amp;reser

ve

Quoted 20 lines
Quoted 20 lines

d=

Quoted 18 lines
Quoted 5 lines

0

Quoted 4 lines
    drivers/ata/Kconfig    |   10 +
    drivers/ata/Makefile   |    1 +
    drivers/ata/fsl_ahci.c | 1030
Quoted 7 lines

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
    drivers/ata/fsl_sata.h |    1 +
    4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

Quoted 7 lines
Quoted 2 lines

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

[Peng Ma]
Hi Stefan,

I am so sorry to reply late, Other Non dts powerpc board need the old sata driver, so the new Sata driver does not replace the so far.

Just curious: Which are the "other non dts powerpc boards"? Are there still many? Is there a plan to move them to DT as well? Or should they perhaps be dropped from mainline if not converted to DT? What's stopping the conversion here?

Thanks,
Stefan

author
#14 lore raw
In reply to Stefan Roese

Hi Stefan,

Do you have any questions about this patch?
Plese let me know. Thanks.

Best Regars,
Peng

Quoted 140 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年5月15日 17:23
To: Peng Ma <[email protected]>; Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>;
bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: [EXT] Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Caution: EXT Email

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:48
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 39 lines
Quoted 39 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:01
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 9 lines
Quoted 9 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 4 lines · Peng Ma

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ====================== This

board

Quoted 11 lines
Quoted 11 lines

does

Quoted 10 lines

not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release. Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

---
depends on:
-

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fp
at
chw

ork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99168&amp;data

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

3%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

966&amp;sdata=3Z33Z5raG%2BnbtSUpz2kPCGpefk1byOgy0%2Br3R4DUFU8%

Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
Quoted 7 lines

3D&amp;reserved=0

Quoted 5 lines

-
https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2F
pa
tc

hwork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99167&amp;d

Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
Quoted 3 lines

ata

Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
Quoted 3 lines

3%7C686e

Quoted 1 line

a1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072966&amp;

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

sdata=sr

Quoted 1 line

%2BCh4UioJw0kuhOiLhc3F6keRhIH8Wflt%2BvMJnHtsY%3D&amp;reser

ve

Quoted 20 lines
Quoted 20 lines

d=

Quoted 18 lines
Quoted 5 lines

0

Quoted 4 lines
    drivers/ata/Kconfig    |   10 +
    drivers/ata/Makefile   |    1 +
    drivers/ata/fsl_ahci.c | 1030
Quoted 7 lines

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
    drivers/ata/fsl_sata.h |    1 +
    4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

Quoted 7 lines
Quoted 2 lines

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

[Peng Ma]
Hi Stefan,

I am so sorry to reply late, Other Non dts powerpc board need the old sata driver, so the new Sata driver does not replace the so far.

Just curious: Which are the "other non dts powerpc boards"? Are there still many? Is there a plan to move them to DT as well? Or should they perhaps be dropped from mainline if not converted to DT? What's stopping the conversion here?

Thanks,
Stefan

#15 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

In reply to Stefan Roese

Hi Stefan,

Quoted 127 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: Wednesday, May 15, 2019 2:53 PM
To: Peng Ma <[email protected]>; Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:48
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 39 lines
Quoted 39 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 7 lines · Peng Ma
Quoted 7 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年3月27日 18:01
To: Peng Ma <[email protected]>; Prabhakar Kushwaha
<[email protected]>; Shengzhou Liu

<[email protected]>;

Quoted 4 lines
Quoted 4 lines

Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 9 lines
Quoted 9 lines

u-boot at lists.denx.de
Subject: Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Quoted 4 lines · Peng Ma

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ====================== This

board

Quoted 11 lines
Quoted 11 lines

does

Quoted 10 lines

not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release. Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

---
depends on:

	-
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fp
at
chw

ork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99168&amp;data

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

3%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072

Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Quoted 1 line

966&amp;sdata=3Z33Z5raG%2BnbtSUpz2kPCGpefk1byOgy0%2Br3R4DUFU8%

Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
Quoted 7 lines

3D&amp;reserved=0

Quoted 5 lines
	-

https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2F
pa
tc

hwork.ozlabs.org%2Fproject%2Fuboot%2Flist%2F%3Fseries%3D99167&amp;d

Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
Quoted 3 lines

ata

Quoted 1 line

=02%7C01%7Cpeng.ma%40nxp.com%7C71e43bd30bf24799586f08d6b29b1ef

Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
Quoted 3 lines

3%7C686e

Quoted 1 line

a1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636892776652072966&amp;

Quoted 27 lines
Quoted 18 lines
Quoted 2 lines
Quoted 2 lines

sdata=sr

Quoted 1 line

%2BCh4UioJw0kuhOiLhc3F6keRhIH8Wflt%2BvMJnHtsY%3D&amp;reserve

d=

Quoted 15 lines
Quoted 5 lines

0

Quoted 4 lines
    drivers/ata/Kconfig    |   10 +
    drivers/ata/Makefile   |    1 +
    drivers/ata/fsl_ahci.c | 1030
Quoted 7 lines

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
    drivers/ata/fsl_sata.h |    1 +
    4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

Quoted 6 lines
Quoted 3 lines

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

[Peng Ma]
Hi Stefan,

Quoted 8 lines
Quoted 2 lines

I am so sorry to reply late, Other Non dts powerpc board need the old sata driver, so the new Sata driver does not replace the so far.

Just curious: Which are the "other non dts powerpc boards"? Are there still many? Is there a plan to move them to DT as well? Or should they perhaps be dropped from mainline if not converted to DT? What's stopping the conversion here?

There are too many powerpc platforms which are not migrated to DT.

My suggestion will be to have sata driver supporting both. Once all PowerPC platform using this SATA driver migrated, it can be dropped.

--pk.

#16 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

On 23.05.19 06:58, Prabhakar Kushwaha wrote:

<snip>

Quoted 31 lines
Quoted 22 lines
Quoted 22 lines
Quoted 13 lines
Quoted 13 lines
Quoted 10 lines
Quoted 3 lines
     drivers/ata/Kconfig    |   10 +
     drivers/ata/Makefile   |    1 +
     drivers/ata/fsl_ahci.c | 1030

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
     drivers/ata/fsl_sata.h |    1 +
     4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

Quoted 6 lines
Quoted 3 lines

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

[Peng Ma]
Hi Stefan,

Quoted 8 lines
Quoted 2 lines

I am so sorry to reply late, Other Non dts powerpc board need the old sata driver, so the new Sata driver does not replace the so far.

Just curious: Which are the "other non dts powerpc boards"? Are there still many? Is there a plan to move them to DT as well? Or should they perhaps be dropped from mainline if not converted to DT? What's stopping the conversion here?

Quoted 5 lines

There are too many powerpc platforms which are not migrated to DT.

My suggestion will be to have sata driver supporting both. Once all PowerPC platform using this SATA driver migrated, it can be dropped.

I'm fine with this. Please go ahead.

Thanks,
Stefan

author
#17 lore raw

Hi Stefan,

Quoted 50 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年5月23日 13:09
To: Prabhakar Kushwaha <[email protected]>; Peng Ma <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>;
bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: [EXT] Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Caution: EXT Email

On 23.05.19 06:58, Prabhakar Kushwaha wrote:

<snip>

Quoted 31 lines
Quoted 30 lines
Quoted 17 lines
Quoted 17 lines
Quoted 16 lines
Quoted 10 lines
Quoted 3 lines
     drivers/ata/Kconfig    |   10 +
     drivers/ata/Makefile   |    1 +
     drivers/ata/fsl_ahci.c | 1030

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
     drivers/ata/fsl_sata.h |    1 +
     4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

Quoted 7 lines
Quoted 2 lines

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

[Peng Ma]
Hi Stefan,

I am so sorry to reply late, Other Non dts powerpc board need the old sata driver, so the new Sata driver does not replace the so far.

Just curious: Which are the "other non dts powerpc boards"? Are there still many? Is there a plan to move them to DT as well? Or should they perhaps be dropped from mainline if not converted to DT? What's stopping the conversion here?

Quoted 7 lines
Quoted 4 lines

There are too many powerpc platforms which are not migrated to DT.

My suggestion will be to have sata driver supporting both. Once all PowerPC platform using this SATA driver migrated, it can be dropped.

I'm fine with this. Please go ahead.

Ok, thanks for your review.

Best Regards,
Peng

Quoted 2 lines

Thanks,
Stefan

#18 lore raw

Dear Peng,

Quoted 67 lines

-----Original Message-----
From: Peng Ma
Sent: Thursday, May 23, 2019 12:37 PM
To: Stefan Roese <[email protected]>; Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de
Subject: RE: [EXT] Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Hi Stefan,

Quoted 5 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: 2019年5月23日 13:09
To: Prabhakar Kushwaha <[email protected]>; Peng Ma <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika

Gupta

Quoted 4 lines

<[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>; bmeng.cn at gmail.com; make at marvell.com; Andy Tang

<[email protected]>;

Quoted 41 lines

u-boot at lists.denx.de
Subject: [EXT] Re: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

Caution: EXT Email

On 23.05.19 06:58, Prabhakar Kushwaha wrote:

<snip>

Quoted 31 lines
Quoted 30 lines
Quoted 17 lines
Quoted 17 lines
Quoted 16 lines
Quoted 10 lines
Quoted 3 lines
     drivers/ata/Kconfig    |   10 +
     drivers/ata/Makefile   |    1 +
     drivers/ata/fsl_ahci.c | 1030

++++++++++++++++++++++++++++++++++++++++

Quoted 3 lines
     drivers/ata/fsl_sata.h |    1 +
     4 files changed, 1042 insertions(+)

create mode 100644 drivers/ata/fsl_ahci.c

Will this patch series replace the old fsl_sata.c driver? If yes, could you remove this driver as well in this series?

[Peng Ma]
Hi Stefan,

fsl_sata.c used to Non DM sata driver for some powerpc socs. Currently We only have one board that supports dts initialization. I will remove this old driver When all of our powerpc socs support DM.

Quoted 7 lines
Quoted 2 lines

I see, thanks. I just wanted to know, if this new SATA driver is a meant as a replacement for the old non-DM driver.

[Peng Ma]
Hi Stefan,

I am so sorry to reply late, Other Non dts powerpc board need the old sata driver, so the new Sata driver does not replace the so far.

Just curious: Which are the "other non dts powerpc boards"? Are there still many? Is there a plan to move them to DT as well? Or should they perhaps be dropped from mainline if not converted to DT? What's stopping the conversion here?

Quoted 9 lines
Quoted 7 lines
Quoted 4 lines

There are too many powerpc platforms which are not migrated to DT.

My suggestion will be to have sata driver supporting both. Once all PowerPC platform using this SATA driver migrated, it can be dropped.

I'm fine with this. Please go ahead.

Ok, thanks for your review.

I am seeing build errors in drivers/ata/fsl_sata.c with this patch.
Something may have changed from the time the patch has submitted and till I am trying to incorporate.

Can you please fix and provide updated patch on top of the tree as early as possible

--pk

#19 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

In reply to Peng Ma
Quoted 29 lines

-----Original Message-----
From: Peng Ma
Sent: Wednesday, March 27, 2019 2:53 PM
To: Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>;
bmeng.cn at gmail.com; sr at denx.de; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de; Peng Ma <[email protected]>
Subject: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

===================== WARNING ====================== This board does not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release.
Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

---
depends on:
- https://patchwork.ozlabs.org/project/uboot/list/?series=99168
- https://patchwork.ozlabs.org/project/uboot/list/?series=99167

Quoted 7 lines
 drivers/ata/Kconfig    |   10 +
 drivers/ata/Makefile   |    1 +
 drivers/ata/fsl_ahci.c | 1030 ++++++++++++++++++++++++++++++++++++++++
 drivers/ata/fsl_sata.h |    1 +
 4 files changed, 1042 insertions(+)
 create mode 100644 drivers/ata/fsl_ahci.c

New driver is getting added in drivers/ata.

I request driver/ata maintainer to review

--pk

#20 lore raw

PATCH 1/4 ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

In reply to Peng Ma
Quoted 16 lines

-----Original Message-----
From: Peng Ma
Sent: Wednesday, March 27, 2019 2:53 PM
To: Prabhakar Kushwaha <[email protected]>; Shengzhou Liu <[email protected]>; Ruchika Gupta <[email protected]>
Cc: Yinbo Zhu <[email protected]>; sjg at chromium.org; Jagdish Gediya <[email protected]>; York Sun <[email protected]>;
bmeng.cn at gmail.com; sr at denx.de; make at marvell.com; Andy Tang <[email protected]>; u-boot at lists.denx.de; Peng Ma <[email protected]>
Subject: [PATCH 1/4] ata: fsl_ahci: Add sata DM support for Freescale powerpc socs

This patch is to support Freescale sata driver with dts initialized. Also resolved the following problems.

Quoted 8 lines

===================== WARNING ====================== This board does not use CONFIG_DM_SCSI. Please update the storage controller to use CONFIG_DM_SCSI before the v2019.07 release.
Failure to update by the deadline may result in board removal. See doc/driver-model/MIGRATION.txt for more info.
====================================================

This patch has been applied to u-boot-mpc85xx, awaiting upstream.

--pk