iMX23 MMC support and environment addition

PATCH New Otavio Salvador 4 participants 43 messages Jan 22, 2013 – Jan 24, 2013

Subsystems (4)

Arm (32-bit) Arm i.MX and Layerscape SPI MMC, SD and SDIO

lore

Share & tools

lore

Otavio Salvador

author

2013-01-22 01:47 UTC #0 lore raw

cover iMX23 MMC support and environment addition

This patchset adds MMC suuport, in PIO mode for now, to iMX23; it has been boot tested in mx23evk board. I did a brief look at DMA support but it seems not fast to implement so I would prefer to have this patchset merged so other people can start using and playing with it.

Marek Vasut (3):
mx23: Add SSP registers map
mx23: Use XFER_COUNT field in HW_SSP_CTRL0 register in MMC PIO mode
mx23_olinuxino: Add MMC support

Otavio Salvador (6):
mx23: Fix transfer size setting for SPI
mx23evk: Add initial board support
mx23_olinuxino: Skip MMC Card detection
mx23_olinuxino: Add default environment
mxs: Fix MMC reset to avoid write errors and timeout in iMX23
mx23: Use PIO mode support for MMC

 arch/arm/include/asm/arch-mxs/regs-ssp.h     |  28 +++
 board/freescale/mx23evk/Makefile             |  47 +++++
 board/freescale/mx23evk/mx23evk.c            |  81 +++++++++
 board/freescale/mx23evk/spl_boot.c           | 115 ++++++++++++
 board/olimex/mx23_olinuxino/mx23_olinuxino.c |  17 +-
 board/olimex/mx23_olinuxino/spl_boot.c       |  10 ++
 boards.cfg                                   |   1 +
 drivers/mmc/mxsmmc.c                         |  50 ++++--
 drivers/spi/mxs_spi.c                        |   5 +
 include/configs/mx23_olinuxino.h             | 103 ++++++++++-
 include/configs/mx23evk.h                    | 252 +++++++++++++++++++++++++++
 11 files changed, 697 insertions(+), 12 deletions(-)
 create mode 100644 board/freescale/mx23evk/Makefile
 create mode 100644 board/freescale/mx23evk/mx23evk.c
 create mode 100644 board/freescale/mx23evk/spl_boot.c
 create mode 100644 include/configs/mx23evk.h

Signature · 1 line

1.8.1

Otavio Salvador

author

2013-01-22 01:47 UTC #1 lore raw

PATCH 1/9 mx23: Add SSP registers map

From: Marek Vasut <[email protected]>

Signed-off-by Marek Vasut <[email protected]>

Signed-off-by Otavio Salvador <[email protected]>

---
 arch/arm/include/asm/arch-mxs/regs-ssp.h | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

arch/arm/include/asm/arch-mxs/regs-ssp.h +28 -0

index b97ac35..72a59ed 100644
@@ -28,6 +28,28 @@
 #include <asm/arch/regs-common.h>
 
 #ifndef	__ASSEMBLY__
+#if defined(CONFIG_MX23)
+struct mxs_ssp_regs {
+	mxs_reg_32(hw_ssp_ctrl0)
+	mxs_reg_32(hw_ssp_cmd0)
+	mxs_reg_32(hw_ssp_cmd1)
+	mxs_reg_32(hw_ssp_compref)
+	mxs_reg_32(hw_ssp_compmask)
+	mxs_reg_32(hw_ssp_timing)
+	mxs_reg_32(hw_ssp_ctrl1)
+	mxs_reg_32(hw_ssp_data)
+	mxs_reg_32(hw_ssp_sdresp0)
+	mxs_reg_32(hw_ssp_sdresp1)
+	mxs_reg_32(hw_ssp_sdresp2)
+	mxs_reg_32(hw_ssp_sdresp3)
+	mxs_reg_32(hw_ssp_status)
+
+	uint32_t	reserved1[12];
+
+	mxs_reg_32(hw_ssp_debug)
+	mxs_reg_32(hw_ssp_version)
+};
+#elif defined(CONFIG_MX28)
 struct mxs_ssp_regs {
 	mxs_reg_32(hw_ssp_ctrl0)
 	mxs_reg_32(hw_ssp_cmd0)

@@ -50,6 +72,7 @@ struct mxs_ssp_regs {
 	mxs_reg_32(hw_ssp_debug)
 	mxs_reg_32(hw_ssp_version)
 };
+#endif
 
 static inline struct mxs_ssp_regs *mxs_ssp_regs_by_bus(unsigned int port)
 {

@@ -90,6 +113,11 @@ static inline struct mxs_ssp_regs *mxs_ssp_regs_by_bus(unsigned int port)
 #define	SSP_CTRL0_GET_RESP			(1 << 17)
 #define	SSP_CTRL0_ENABLE			(1 << 16)
 
+#ifdef CONFIG_MX23
+#define	SSP_CTRL0_XFER_COUNT_OFFSET		0
+#define	SSP_CTRL0_XFER_COUNT_MASK		0xffff
+#endif
+
 #define	SSP_CMD0_SOFT_TERMINATE			(1 << 26)
 #define	SSP_CMD0_DBL_DATA_RATE_EN		(1 << 25)
 #define	SSP_CMD0_PRIM_BOOT_OP_EN		(1 << 24)

Signature · 1 line

1.8.1

Otavio Salvador

author

2013-01-22 01:47 UTC #2 lore raw

PATCH 2/9 mx23: Use XFER_COUNT field in HW_SSP_CTRL0 register in MMC PIO mode

In reply to Otavio Salvador

From: Marek Vasut <[email protected]>

Signed-off-by Marek Vasut <[email protected]>

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/mmc/mxsmmc.c | 24 +++++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)

drivers/mmc/mxsmmc.c +21 -3

index 0c4cd54..4b178be 100644
@@ -211,14 +211,32 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, struct mmc_data *data)
 		}
 
 		ctrl0 |= SSP_CTRL0_DATA_XFER;
+
+		reg = data->blocksize * data->blocks;
+#if defined(CONFIG_MX23)
+		if (reg & ~SSP_CTRL0_XFER_COUNT_MASK) {
+			printf("MMC%d: Transfer too large (%i bytes)!\n",
+				mmc->block_dev.dev, reg);
+			return -EINVAL;
+		}
+
+		ctrl0 |= reg & SSP_CTRL0_XFER_COUNT_MASK;
+
+		reg = readl(&ssp_regs->hw_ssp_cmd0);
+		reg &= ~0xfff00;
+		reg |= (data->blocks - 1) << 8;
+		reg |= (ffs(data->blocksize) - 1) << 16;
+
+		writel(reg, &ssp_regs->hw_ssp_cmd0);
+#elif defined(CONFIG_MX28)
+		writel(reg, &ssp_regs->hw_ssp_xfer_size);
+
 		reg = ((data->blocks - 1) <<
 			SSP_BLOCK_SIZE_BLOCK_COUNT_OFFSET) |
 			((ffs(data->blocksize) - 1) <<
 			SSP_BLOCK_SIZE_BLOCK_SIZE_OFFSET);
 		writel(reg, &ssp_regs->hw_ssp_block_size);
-
-		reg = data->blocksize * data->blocks;
-		writel(reg, &ssp_regs->hw_ssp_xfer_size);
+#endif
 	}
 
 	/* Kick off the command */

Signature · 1 line

1.8.1

Marek Vasut

2013-01-22 01:45 UTC #3 lore raw

PATCH 2/9 mx23: Use XFER_COUNT field in HW_SSP_CTRL0 register in MMC PIO mode

Dear Otavio Salvador,

Quoted 21 lines

From: Marek Vasut <[email protected]>

Signed-off-by Marek Vasut <[email protected]>

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/mmc/mxsmmc.c | 24 +++++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)

drivers/mmc/mxsmmc.c +5 -0

index 0c4cd54..4b178be 100644
@@ -211,14 +211,32 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd,
 struct mmc_data *data) }
 
 		ctrl0 |= SSP_CTRL0_DATA_XFER;
+
+		reg = data->blocksize * data->blocks;
+#if defined(CONFIG_MX23)
+		if (reg & ~SSP_CTRL0_XFER_COUNT_MASK) {
+			printf("MMC%d: Transfer too large (%i bytes)!\n",

drivers/mmc/mxsmmc.c mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, · line 228

Quoted 10 lines

drivers/mmc/mxsmmc.c +10 -0

+				mmc->block_dev.dev, reg);
+			return -EINVAL;
+		}
+
+		ctrl0 |= reg & SSP_CTRL0_XFER_COUNT_MASK;
+
+		reg = readl(&ssp_regs->hw_ssp_cmd0);
+		reg &= ~0xfff00;
+		reg |= (data->blocks - 1) << 8;
+		reg |= (ffs(data->blocksize) - 1) << 16;

This is using some random hard-coded offsets, I told you numerous times I will NAK this.

Sorry.

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 01:53 UTC #4 lore raw

PATCH 2/9 mx23: Use XFER_COUNT field in HW_SSP_CTRL0 register in MMC PIO mode

Quoted 28 lines · Marek Vasut

Dear Otavio Salvador,

Quoted 26 lines

From: Marek Vasut <[email protected]>

Signed-off-by Marek Vasut <[email protected]>

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/mmc/mxsmmc.c | 24 +++++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)

drivers/mmc/mxsmmc.c +10 -0

index 0c4cd54..4b178be 100644
@@ -211,14 +211,32 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd,
 struct mmc_data *data) }
 
              ctrl0 |= SSP_CTRL0_DATA_XFER;
+
+             reg = data->blocksize * data->blocks;
+#if defined(CONFIG_MX23)
+             if (reg & ~SSP_CTRL0_XFER_COUNT_MASK) {
+                     printf("MMC%d: Transfer too large (%i bytes)!\n",
+                             mmc->block_dev.dev, reg);
+                     return -EINVAL;
+             }
+
+             ctrl0 |= reg & SSP_CTRL0_XFER_COUNT_MASK;

Quoted 8 lines

drivers/mmc/mxsmmc.c mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, struct mmc_data *data) · line 228
Quoted 5 lines
drivers/mmc/mxsmmc.c +5 -0
+
+             reg = readl(&ssp_regs->hw_ssp_cmd0);
+             reg &= ~0xfff00;
+             reg |= (data->blocks - 1) << 8;
+             reg |= (ffs(data->blocksize) - 1) << 16;
This is using some random hard-coded offsets, I told you numerous times I will NAK this.

It is your patch ;-) So ...

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Marek Vasut

2013-01-22 02:01 UTC #5 lore raw

PATCH 2/9 mx23: Use XFER_COUNT field in HW_SSP_CTRL0 register in MMC PIO mode

Dear Otavio Salvador,

Quoted 40 lines · Marek Vasut

Quoted 30 lines

Dear Otavio Salvador,

Quoted 28 lines

From: Marek Vasut <[email protected]>

Signed-off-by Marek Vasut <[email protected]>

Signed-off-by Otavio Salvador <[email protected]>

---

 drivers/mmc/mxsmmc.c | 24 +++++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)

drivers/mmc/mxsmmc.c +10 -0

index 0c4cd54..4b178be 100644
@@ -211,14 +211,32 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd
 *cmd, struct mmc_data *data) }
 
              ctrl0 |= SSP_CTRL0_DATA_XFER;
 
+
+             reg = data->blocksize * data->blocks;
+#if defined(CONFIG_MX23)
+             if (reg & ~SSP_CTRL0_XFER_COUNT_MASK) {
+                     printf("MMC%d: Transfer too large (%i bytes)!\n",
+                             mmc->block_dev.dev, reg);
+                     return -EINVAL;
+             }
+
+             ctrl0 |= reg & SSP_CTRL0_XFER_COUNT_MASK;

Quoted 8 lines

drivers/mmc/mxsmmc.c mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, struct mmc_data *data) · line 228
Quoted 5 lines
drivers/mmc/mxsmmc.c +5 -0
+
+             reg = readl(&ssp_regs->hw_ssp_cmd0);
+             reg &= ~0xfff00;
+             reg |= (data->blocks - 1) << 8;
+             reg |= (ffs(data->blocksize) - 1) << 16;
This is using some random hard-coded offsets, I told you numerous times I will NAK this.

Quoted 1 line

It is your patch ;-) So ...

So I'll NAK it. I have no problem with it. Besides, this is only based on my work, but whatever.

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 01:47 UTC #6 lore raw

PATCH 3/9 mx23: Fix transfer size setting for SPI

In reply to Otavio Salvador

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/spi/mxs_spi.c | 5 +++++
 1 file changed, 5 insertions(+)

drivers/spi/mxs_spi.c +5 -0

index bb865b7..10bf5aa 100644
@@ -167,8 +167,13 @@ static int mxs_spi_xfer_pio(struct mxs_spi_slave *slave,
 		mxs_spi_start_xfer(ssp_regs);
 
 	while (length--) {
+#if defined(CONFIG_MX23)
+		writel(SSP_CTRL0_XFER_COUNT_MASK, &ssp_regs->hw_ssp_ctrl0_clr);
+		writel(1, &ssp_regs->hw_ssp_ctrl0_set);
+#elif defined(CONFIG_MX28)
 		/* We transfer 1 byte */
 		writel(1, &ssp_regs->hw_ssp_xfer_size);
+#endif
 
 		if ((flags & SPI_XFER_END) && !length)
 			mxs_spi_end_xfer(ssp_regs);

Signature · 1 line

1.8.1

Marek Vasut

2013-01-22 01:46 UTC #7 lore raw

PATCH 3/9 mx23: Fix transfer size setting for SPI

Dear Otavio Salvador,

Description missing. Was this ever tested? Does SSP DMA work?

Quoted 13 lines

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/spi/mxs_spi.c | 5 +++++
 1 file changed, 5 insertions(+)

drivers/spi/mxs_spi.c +0 -0

index bb865b7..10bf5aa 100644
@@ -167,8 +167,13 @@ static int mxs_spi_xfer_pio(struct mxs_spi_slave
 *slave, mxs_spi_start_xfer(ssp_regs);
 
 	while (length--) {

drivers/spi/mxs_spi.c static int mxs_spi_xfer_pio(struct mxs_spi_slave · line 179

Quoted 10 lines

drivers/spi/mxs_spi.c +5 -0

+#if defined(CONFIG_MX23)
+		writel(SSP_CTRL0_XFER_COUNT_MASK, &ssp_regs->hw_ssp_ctrl0_clr);
+		writel(1, &ssp_regs->hw_ssp_ctrl0_set);
+#elif defined(CONFIG_MX28)
 		/* We transfer 1 byte */
 		writel(1, &ssp_regs->hw_ssp_xfer_size);
+#endif
 
 		if ((flags & SPI_XFER_END) && !length)
 			mxs_spi_end_xfer(ssp_regs);

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 01:55 UTC #8 lore raw

PATCH 3/9 mx23: Fix transfer size setting for SPI

Quoted 5 lines · Marek Vasut

Dear Otavio Salvador,
Description missing. Was this ever tested? Does SSP DMA work?
Quoted 1 line
Signed-off-by Otavio Salvador <[email protected]>

No; I did not test it as NAND support is not yet done but this was need to get it going. This is exactly same code as done in v3.7 kernel so seems fine.

What you want added in description?

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Marek Vasut

2013-01-22 02:00 UTC #9 lore raw

PATCH 3/9 mx23: Fix transfer size setting for SPI

Dear Otavio Salvador,

Quoted 10 lines · Marek Vasut

Quoted 5 lines
Dear Otavio Salvador,
Description missing. Was this ever tested? Does SSP DMA work?
Quoted 1 line
Signed-off-by Otavio Salvador <[email protected]>
No; I did not test it as NAND support is not yet done but this was need to get it going. This is exactly same code as done in v3.7 kernel so seems fine.

SPI has nothing to do with NAND. Please at least test code or clearly state the code was untested.

Quoted 1 line

What you want added in description?

What this patch does.

Best regards,
Marek Vasut

Marek Vasut

NAK

2013-01-24 18:14 UTC #10 lore raw

PATCH 3/9 mx23: Fix transfer size setting for SPI

In reply to Otavio Salvador

Dear Otavio Salvador,

NAK, this won't work. SSP0 DMA has this +1 offset in it's channel placement (so SSP0 DMA channel is actually channel 1), check the MMC patch.

Quoted 13 lines

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/spi/mxs_spi.c | 5 +++++
 1 file changed, 5 insertions(+)

drivers/spi/mxs_spi.c +0 -0

index bb865b7..10bf5aa 100644
@@ -167,8 +167,13 @@ static int mxs_spi_xfer_pio(struct mxs_spi_slave
 *slave, mxs_spi_start_xfer(ssp_regs);
 
 	while (length--) {

drivers/spi/mxs_spi.c static int mxs_spi_xfer_pio(struct mxs_spi_slave · line 179

Quoted 10 lines

drivers/spi/mxs_spi.c +5 -0

+#if defined(CONFIG_MX23)
+		writel(SSP_CTRL0_XFER_COUNT_MASK, &ssp_regs->hw_ssp_ctrl0_clr);
+		writel(1, &ssp_regs->hw_ssp_ctrl0_set);
+#elif defined(CONFIG_MX28)
 		/* We transfer 1 byte */
 		writel(1, &ssp_regs->hw_ssp_xfer_size);
+#endif
 
 		if ((flags & SPI_XFER_END) && !length)
 			mxs_spi_end_xfer(ssp_regs);

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 01:47 UTC #11 lore raw

PATCH 4/9 mx23evk: Add initial board support

In reply to Otavio Salvador

Signed-off-by Otavio Salvador <[email protected]>

---
 board/freescale/mx23evk/Makefile   |  47 +++++++
 board/freescale/mx23evk/mx23evk.c  |  81 ++++++++++++
 board/freescale/mx23evk/spl_boot.c | 115 +++++++++++++++++
 boards.cfg                         |   1 +
 include/configs/mx23evk.h          | 252 +++++++++++++++++++++++++++++++++++++
 5 files changed, 496 insertions(+)
 create mode 100644 board/freescale/mx23evk/Makefile
 create mode 100644 board/freescale/mx23evk/mx23evk.c
 create mode 100644 board/freescale/mx23evk/spl_boot.c
 create mode 100644 include/configs/mx23evk.h

board/freescale/mx23evk/Makefile new file +47 -0

new file mode 100644
index 0000000..7518d7b
@@ -0,0 +1,47 @@
+#
+# (C) Copyright 2000-2006
+# Wolfgang Denk, DENX Software Engineering, wd at denx.de.
+#
+# See file CREDITS for list of people who contributed to this
+# project.
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU General Public License as
+# published by the Free Software Foundation; either version 2 of
+# the License, or (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 59 Temple Place, Suite 330, Boston,
+# MA 02111-1307 USA
+#
+
+include $(TOPDIR)/config.mk
+
+LIB	= $(obj)lib$(BOARD).o
+
+ifndef	CONFIG_SPL_BUILD
+COBJS	:= mx23evk.o
+else
+COBJS	:= spl_boot.o
+endif
+
+SRCS	:= $(COBJS:.o=.c)
+OBJS	:= $(addprefix $(obj),$(COBJS))
+
+$(LIB):	$(obj).depend $(OBJS)
+	$(call cmd_link_o_target, $(OBJS))
+
+#########################################################################
+
+# defines $(obj).depend target
+include $(SRCTREE)/rules.mk
+
+sinclude $(obj).depend
+
+#########################################################################

board/freescale/mx23evk/mx23evk.c new file +81 -0

81 lines not on this page · raw message

board/freescale/mx23evk/spl_boot.c new file +115 -0

115 lines not on this page · raw message

boards.cfg +1 -0

index 9b31931..7d03620 100644
@@ -193,6 +193,7 @@ imx27lite                    arm         arm926ejs   imx27lite           logicpd
 magnesium                    arm         arm926ejs   imx27lite           logicpd        mx27
 mx23_olinuxino               arm         arm926ejs   mx23_olinuxino      olimex         mxs		mx23_olinuxino
 apx4devkit                   arm         arm926ejs   apx4devkit          bluegiga       mxs		apx4devkit
+mx23evk                      arm         arm926ejs   mx23evk             freescale      mxs		mx23evk
 m28evk                       arm         arm926ejs   m28evk              denx           mxs		m28evk
 mx28evk                      arm         arm926ejs   mx28evk             freescale      mxs		mx28evk
 sc_sps_1                     arm         arm926ejs   sc_sps_1            schulercontrol mxs

include/configs/mx23evk.h new file +252 -0

252 lines not on this page · raw message

Signature · 1 line

1.8.1

Marek Vasut

2013-01-22 01:46 UTC #12 lore raw

PATCH 4/9 mx23evk: Add initial board support

Dear Otavio Salvador,

This should not be a part of the patchset. Description is missing.

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 02:02 UTC #13 lore raw

PATCH 4/9 mx23evk: Add initial board support

Quoted 3 lines · Marek Vasut

Dear Otavio Salvador,
This should not be a part of the patchset. Description is missing.

Well, this is the first version of the board support. I see no better description for it.

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Marek Vasut

2013-01-22 02:16 UTC #14 lore raw

PATCH 4/9 mx23evk: Add initial board support

Dear Otavio Salvador,

Quoted 7 lines · Marek Vasut

Quoted 3 lines
Dear Otavio Salvador,
This should not be a part of the patchset. Description is missing.
Well, this is the first version of the board support. I see no better description for it.

Check other patches adding boards then please. Description of the board is sufficient.

Best regards,
Marek Vasut

Fabio Estevam

reviewer

2013-01-22 02:30 UTC #15 lore raw

PATCH 4/9 mx23evk: Add initial board support

In reply to Otavio Salvador

Quoted 12 lines · Otavio Salvador

Signed-off-by Otavio Salvador <[email protected]>

---
 board/freescale/mx23evk/Makefile   |  47 +++++++
 board/freescale/mx23evk/mx23evk.c  |  81 ++++++++++++
 board/freescale/mx23evk/spl_boot.c | 115 +++++++++++++++++
 boards.cfg                         |   1 +
 include/configs/mx23evk.h          | 252 +++++++++++++++++++++++++++++++++++++
 5 files changed, 496 insertions(+)
 create mode 100644 board/freescale/mx23evk/Makefile
 create mode 100644 board/freescale/mx23evk/mx23evk.c
 create mode 100644 board/freescale/mx23evk/spl_boot.c
 create mode 100644 include/configs/mx23evk.h

An entry to MAINTAINERS file is missing.

include/configs/mx23evk.h line 169

Quoted 10 lines

include/configs/mx23evk.h new file +10 -0

+/*
+ * USB
+ */
+#ifdef CONFIG_CMD_USB
+#define        CONFIG_USB_EHCI
+#define        CONFIG_USB_EHCI_MXS
+#define        CONFIG_EHCI_MXS_PORT 1
+#define        CONFIG_EHCI_IS_TDI
+#define        CONFIG_USB_STORAGE
+#endif

Is USB functional? If not, please remove this.

Marek Vasut

2013-01-22 02:42 UTC #16 lore raw

PATCH 4/9 mx23evk: Add initial board support

Dear Fabio Estevam,

Quoted 20 lines

On Mon, Jan 21, 2013 at 11:47 PM, Otavio Salvador

Quoted 14 lines · <[email protected]> wrote:

Signed-off-by Otavio Salvador <[email protected]>

---

 board/freescale/mx23evk/Makefile   |  47 +++++++
 board/freescale/mx23evk/mx23evk.c  |  81 ++++++++++++
 board/freescale/mx23evk/spl_boot.c | 115 +++++++++++++++++
 boards.cfg                         |   1 +
 include/configs/mx23evk.h          | 252
 +++++++++++++++++++++++++++++++++++++ 5 files changed, 496
 insertions(+)
 create mode 100644 board/freescale/mx23evk/Makefile
 create mode 100644 board/freescale/mx23evk/mx23evk.c
 create mode 100644 board/freescale/mx23evk/spl_boot.c
 create mode 100644 include/configs/mx23evk.h

An entry to MAINTAINERS file is missing.

Quoted 12 lines

include/configs/mx23evk.h line 169

Quoted 10 lines

include/configs/mx23evk.h new file +10 -0

+/*
+ * USB
+ */
+#ifdef CONFIG_CMD_USB
+#define        CONFIG_USB_EHCI
+#define        CONFIG_USB_EHCI_MXS
+#define        CONFIG_EHCI_MXS_PORT 1
+#define        CONFIG_EHCI_IS_TDI
+#define        CONFIG_USB_STORAGE
+#endif

Is USB functional? If not, please remove this.

I don't think so ... besides, MX23 has only one USB port. How can this even compile if USBPHY1 and such values are undefined?

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 10:05 UTC #17 lore raw

PATCH 4/9 mx23evk: Add initial board support

In reply to Fabio Estevam

Quoted 15 lines · Fabio Estevam

Quoted 12 lines · Otavio Salvador

Signed-off-by Otavio Salvador <[email protected]>

---
 board/freescale/mx23evk/Makefile   |  47 +++++++
 board/freescale/mx23evk/mx23evk.c  |  81 ++++++++++++
 board/freescale/mx23evk/spl_boot.c | 115 +++++++++++++++++
 boards.cfg                         |   1 +
 include/configs/mx23evk.h          | 252 +++++++++++++++++++++++++++++++++++++
 5 files changed, 496 insertions(+)
 create mode 100644 board/freescale/mx23evk/Makefile
 create mode 100644 board/freescale/mx23evk/mx23evk.c
 create mode 100644 board/freescale/mx23evk/spl_boot.c
 create mode 100644 include/configs/mx23evk.h

Quoted 1 line

An entry to MAINTAINERS file is missing.

Right.

Quoted 4 lines

Quoted 4 lines
include/configs/mx23evk.h new file +4 -0
+/*
+ * USB
+ */
+#ifdef CONFIG_CMD_USB

Quoted 8 lines

include/configs/mx23evk.h line 169
Quoted 6 lines
include/configs/mx23evk.h new file +6 -0
+#define        CONFIG_USB_EHCI
+#define        CONFIG_USB_EHCI_MXS
+#define        CONFIG_EHCI_MXS_PORT 1
+#define        CONFIG_EHCI_IS_TDI
+#define        CONFIG_USB_STORAGE
+#endif
Is USB functional? If not, please remove this.

It is not; will check if I can drop it.

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Otavio Salvador

author

2013-01-22 01:47 UTC #18 lore raw

PATCH 5/9 mx23_olinuxino: Add MMC support

In reply to Otavio Salvador

From: Marek Vasut <[email protected]>

Signed-off-by Marek Vasut <[email protected]>

---
 board/olimex/mx23_olinuxino/mx23_olinuxino.c | 17 ++++++++++++++++-
 board/olimex/mx23_olinuxino/spl_boot.c       | 10 ++++++++++
 include/configs/mx23_olinuxino.h             | 12 ++++++++++++
 3 files changed, 38 insertions(+), 1 deletion(-)

board/olimex/mx23_olinuxino/mx23_olinuxino.c +16 -1

index 2fcad26..421ed9c 100644
@@ -23,7 +23,10 @@
  */
 
 #include <common.h>
+#include <asm/io.h>
+#include <asm/arch/iomux-mx23.h>
 #include <asm/arch/imx-regs.h>
+#include <asm/arch/clock.h>
 #include <asm/arch/sys_proto.h>
 
 DECLARE_GLOBAL_DATA_PTR;

@@ -33,7 +36,12 @@ DECLARE_GLOBAL_DATA_PTR;
  */
 int board_early_init_f(void)
 {
-	/* SSP clock init will come here soon. */
+	/* IO0 clock at 480MHz */
+	mxs_set_ioclk(MXC_IOCLK0, 480000);
+
+	/* SSP0 clock at 96MHz */
+	mxs_set_sspclk(MXC_SSPCLK0, 96000, 0);
+
 	return 0;
 }

@@ -42,6 +50,13 @@ int dram_init(void)
 	return mxs_dram_init();
 }
 
+#ifdef	CONFIG_CMD_MMC
+int board_mmc_init(bd_t *bis)
+{
+	return mxsmmc_initialize(bis, 0, NULL);
+}
+#endif
+
 int board_init(void)
 {
 	/* Adress of boot parameters */

board/olimex/mx23_olinuxino/spl_boot.c +10 -0

index aaebe6c..3cc40cd 100644
@@ -30,6 +30,7 @@
 #include <asm/arch/sys_proto.h>
 
 #define	MUX_CONFIG_EMI	(MXS_PAD_3V3 | MXS_PAD_12MA | MXS_PAD_NOPULL)
+#define	MUX_CONFIG_SSP	(MXS_PAD_3V3 | MXS_PAD_8MA | MXS_PAD_PULLUP)
 
 const iomux_cfg_t iomux_setup[] = {
 	/* DUART */

@@ -82,6 +83,15 @@ const iomux_cfg_t iomux_setup[] = {
 	MX23_PAD_EMI_CKE__EMI_CKE | MUX_CONFIG_EMI,
 	MX23_PAD_EMI_RASN__EMI_RASN | MUX_CONFIG_EMI,
 	MX23_PAD_EMI_WEN__EMI_WEN | MUX_CONFIG_EMI,
+
+	/* MMC 0 */
+	MX23_PAD_SSP1_CMD__SSP1_CMD | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DETECT__SSP1_DETECT | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DATA0__SSP1_DATA0 | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DATA1__SSP1_DATA1 | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DATA2__SSP1_DATA2 | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DATA3__SSP1_DATA3 | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_SCK__SSP1_SCK | MUX_CONFIG_SSP,
 };
 
 void board_init_ll(void)

include/configs/mx23_olinuxino.h +12 -0

index e5e718c..bc63481 100644
@@ -53,8 +53,10 @@
 #define	CONFIG_DOS_PARTITION
 
 #define	CONFIG_CMD_CACHE
+#define	CONFIG_CMD_EXT2
 #define	CONFIG_CMD_FAT
 #define	CONFIG_CMD_GPIO
+#define	CONFIG_CMD_MMC
 
 /*
  * Memory configurations

@@ -110,6 +112,16 @@
 #define	CONFIG_BAUDRATE			115200	/* Default baud rate */
 
 /*
+ * MMC Driver
+ */
+#ifdef	CONFIG_CMD_MMC
+#define	CONFIG_MMC
+#define	CONFIG_BOUNCE_BUFFER
+#define	CONFIG_GENERIC_MMC
+#define	CONFIG_MXS_MMC
+#endif
+
+/*
  * APBH DMA
  */
 #define CONFIG_APBH_DMA

Signature · 1 line

1.8.1

Marek Vasut

2013-01-22 01:47 UTC #19 lore raw

PATCH 5/9 mx23_olinuxino: Add MMC support

Dear Otavio Salvador,

Description is missing.

Quoted 103 lines

From: Marek Vasut <[email protected]>

Signed-off-by Marek Vasut <[email protected]>

---
 board/olimex/mx23_olinuxino/mx23_olinuxino.c | 17 ++++++++++++++++-
 board/olimex/mx23_olinuxino/spl_boot.c       | 10 ++++++++++
 include/configs/mx23_olinuxino.h             | 12 ++++++++++++
 3 files changed, 38 insertions(+), 1 deletion(-)

(quoted diff) +0 -0

b/board/olimex/mx23_olinuxino/mx23_olinuxino.c index 2fcad26..421ed9c 100644

board/olimex/mx23_olinuxino/mx23_olinuxino.c +16 -1

@@ -23,7 +23,10 @@
  */
 
 #include <common.h>
+#include <asm/io.h>
+#include <asm/arch/iomux-mx23.h>
 #include <asm/arch/imx-regs.h>
+#include <asm/arch/clock.h>
 #include <asm/arch/sys_proto.h>
 
 DECLARE_GLOBAL_DATA_PTR;

@@ -33,7 +36,12 @@ DECLARE_GLOBAL_DATA_PTR;
  */
 int board_early_init_f(void)
 {
-	/* SSP clock init will come here soon. */
+	/* IO0 clock at 480MHz */
+	mxs_set_ioclk(MXC_IOCLK0, 480000);
+
+	/* SSP0 clock at 96MHz */
+	mxs_set_sspclk(MXC_SSPCLK0, 96000, 0);
+
 	return 0;
 }

@@ -42,6 +50,13 @@ int dram_init(void)
 	return mxs_dram_init();
 }
 
+#ifdef	CONFIG_CMD_MMC
+int board_mmc_init(bd_t *bis)
+{
+	return mxsmmc_initialize(bis, 0, NULL);
+}
+#endif
+
 int board_init(void)
 {
 	/* Adress of boot parameters */

(quoted diff) +0 -0

b/board/olimex/mx23_olinuxino/spl_boot.c index aaebe6c..3cc40cd 100644

board/olimex/mx23_olinuxino/spl_boot.c +10 -0

@@ -30,6 +30,7 @@
 #include <asm/arch/sys_proto.h>
 
 #define	MUX_CONFIG_EMI	(MXS_PAD_3V3 | MXS_PAD_12MA | MXS_PAD_NOPULL)
+#define	MUX_CONFIG_SSP	(MXS_PAD_3V3 | MXS_PAD_8MA | MXS_PAD_PULLUP)
 
 const iomux_cfg_t iomux_setup[] = {
 	/* DUART */

@@ -82,6 +83,15 @@ const iomux_cfg_t iomux_setup[] = {
 	MX23_PAD_EMI_CKE__EMI_CKE | MUX_CONFIG_EMI,
 	MX23_PAD_EMI_RASN__EMI_RASN | MUX_CONFIG_EMI,
 	MX23_PAD_EMI_WEN__EMI_WEN | MUX_CONFIG_EMI,
+
+	/* MMC 0 */
+	MX23_PAD_SSP1_CMD__SSP1_CMD | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DETECT__SSP1_DETECT | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DATA0__SSP1_DATA0 | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DATA1__SSP1_DATA1 | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DATA2__SSP1_DATA2 | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_DATA3__SSP1_DATA3 | MUX_CONFIG_SSP,
+	MX23_PAD_SSP1_SCK__SSP1_SCK | MUX_CONFIG_SSP,
 };
 
 void board_init_ll(void)

(quoted diff) +0 -0

b/include/configs/mx23_olinuxino.h index e5e718c..bc63481 100644

include/configs/mx23_olinuxino.h +5 -0

@@ -53,8 +53,10 @@
 #define	CONFIG_DOS_PARTITION
 
 #define	CONFIG_CMD_CACHE
+#define	CONFIG_CMD_EXT2
 #define	CONFIG_CMD_FAT
 #define	CONFIG_CMD_GPIO
+#define	CONFIG_CMD_MMC
 
 /*
  * Memory configurations

@@ -110,6 +112,16 @@
 #define	CONFIG_BAUDRATE			115200	/* Default baud rate */
 
 /*
+ * MMC Driver
+ */
+#ifdef	CONFIG_CMD_MMC

include/configs/mx23_olinuxino.h line 127

Quoted 10 lines

include/configs/mx23_olinuxino.h +7 -0

+#define	CONFIG_MMC
+#define	CONFIG_BOUNCE_BUFFER
+#define	CONFIG_GENERIC_MMC
+#define	CONFIG_MXS_MMC
+#endif
+
+/*
  * APBH DMA
  */
 #define CONFIG_APBH_DMA

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 01:58 UTC #20 lore raw

PATCH 5/9 mx23_olinuxino: Add MMC support

Quoted 7 lines · Marek Vasut

Dear Otavio Salvador,
Description is missing.
Quoted 3 lines
From: Marek Vasut <[email protected]>
Signed-off-by Marek Vasut <[email protected]>

Please provide me the description you wish and I amend it.

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Otavio Salvador

author

2013-01-22 01:47 UTC #21 lore raw

PATCH 6/9 mx23_olinuxino: Skip MMC Card detection

In reply to Otavio Salvador

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/mmc/mxsmmc.c             | 2 ++
 include/configs/mx23_olinuxino.h | 1 +
 2 files changed, 3 insertions(+)

drivers/mmc/mxsmmc.c +2 -0

index 4b178be..618659e 100644
@@ -165,11 +165,13 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, struct mmc_data *data)
 		return TIMEOUT;
 	}
 
+#ifndef CONFIG_MMC_SKIP_CARD_DETECT
 	/* See if card is present */
 	if (readl(&ssp_regs->hw_ssp_status) & SSP_STATUS_CARD_DETECT) {
 		printf("MMC%d: No card detected!\n", mmc->block_dev.dev);
 		return NO_CARD_ERR;
 	}
+#endif
 
 	/* Start building CTRL0 contents */
 	ctrl0 = priv->buswidth;

include/configs/mx23_olinuxino.h +1 -0

index bc63481..a62e3a2 100644
@@ -116,6 +116,7 @@
  */
 #ifdef	CONFIG_CMD_MMC
 #define	CONFIG_MMC
+#define	CONFIG_MMC_SKIP_CARD_DETECT
 #define	CONFIG_BOUNCE_BUFFER
 #define	CONFIG_GENERIC_MMC
 #define	CONFIG_MXS_MMC

Signature · 1 line

1.8.1

Marek Vasut

2013-01-22 01:48 UTC #22 lore raw

PATCH 6/9 mx23_olinuxino: Skip MMC Card detection

Dear Otavio Salvador,

Quoted 11 lines

Signed-off-by Otavio Salvador <[email protected]>
---
 drivers/mmc/mxsmmc.c             | 2 ++
 include/configs/mx23_olinuxino.h | 1 +
 2 files changed, 3 insertions(+)
drivers/mmc/mxsmmc.c +0 -0
index 4b178be..618659e 100644

drivers/mmc/mxsmmc.c mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, · line 168

Quoted 4 lines

drivers/mmc/mxsmmc.c +1 -0

@@ -165,11 +165,13 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd,
 struct mmc_data *data) return TIMEOUT;
 	}
 
+#ifndef CONFIG_MMC_SKIP_CARD_DETECT

Undocumented feature added. I think mxsmmc_register() call can take a param for this.

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 01:58 UTC #23 lore raw

PATCH 6/9 mx23_olinuxino: Skip MMC Card detection

Quoted 16 lines · Marek Vasut

Dear Otavio Salvador,

Quoted 14 lines

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/mmc/mxsmmc.c             | 2 ++
 include/configs/mx23_olinuxino.h | 1 +
 2 files changed, 3 insertions(+)

drivers/mmc/mxsmmc.c +0 -0

index 4b178be..618659e 100644
@@ -165,11 +165,13 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd,
 struct mmc_data *data) return TIMEOUT;
      }

Quoted 4 lines

drivers/mmc/mxsmmc.c mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, struct mmc_data *data) · line 168
Quoted 1 line
drivers/mmc/mxsmmc.c +1 -0
+#ifndef CONFIG_MMC_SKIP_CARD_DETECT
Undocumented feature added. I think mxsmmc_register() call can take a param for this.

Where I should document it? mmc_register cannot.

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Marek Vasut

2013-01-22 01:59 UTC #24 lore raw

PATCH 6/9 mx23_olinuxino: Skip MMC Card detection

Dear Otavio Salvador,

Quoted 17 lines · Marek Vasut

Quoted 16 lines
Dear Otavio Salvador,
Quoted 14 lines
Signed-off-by Otavio Salvador <[email protected]>
---
 drivers/mmc/mxsmmc.c             | 2 ++
 include/configs/mx23_olinuxino.h | 1 +
 2 files changed, 3 insertions(+)
drivers/mmc/mxsmmc.c +0 -0
index 4b178be..618659e 100644
@@ -165,11 +165,13 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd
 *cmd, struct mmc_data *data) return TIMEOUT;
 

Quoted 8 lines

Quoted 6 lines
drivers/mmc/mxsmmc.c mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, struct mmc_data *data) · line 168
Quoted 3 lines
drivers/mmc/mxsmmc.c +1 -0
      }
 
+#ifndef CONFIG_MMC_SKIP_CARD_DETECT
Undocumented feature added. I think mxsmmc_register() call can take a param for this.
Where I should document it? mmc_register cannot.

mxsmmc_initialize() can.

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 01:47 UTC #25 lore raw

PATCH 7/9 mx23_olinuxino: Add default environment

In reply to Otavio Salvador

Signed-off-by Otavio Salvador <[email protected]>

---
 include/configs/mx23_olinuxino.h | 90 +++++++++++++++++++++++++++++++++++++++-
 1 file changed, 89 insertions(+), 1 deletion(-)

include/configs/mx23_olinuxino.h +89 -1

index a62e3a2..7d34168 100644
@@ -140,9 +140,97 @@
 #define	CONFIG_SETUP_MEMORY_TAGS
 #define	CONFIG_BOOTDELAY	3
 #define	CONFIG_BOOTFILE		"uImage"
-#define	CONFIG_BOOTARGS		"console=ttyAMA0,115200n8 "
 #define	CONFIG_LOADADDR		0x42000000
 #define	CONFIG_SYS_LOAD_ADDR	CONFIG_LOADADDR
 #define	CONFIG_OF_LIBFDT
 
+/*
+ * Extra Environments
+ */
+#define CONFIG_EXTRA_ENV_SETTINGS \
+	"update_sd_firmware_filename=u-boot.sd\0" \
+	"update_nand_full_filename=u-boot.nand\0" \
+	"update_nand_firmware_filename=u-boot.sb\0"	\
+	"update_sd_firmware_filename=u-boot.sd\0" \
+	"update_nand_firmware_maxsz=0x100000\0"	\
+	"update_nand_stride=0x40\0"	/* MX28 datasheet ch. 12.12 */ \
+	"update_nand_count=0x4\0"	/* MX28 datasheet ch. 12.12 */ \
+	"update_nand_get_fcb_size="	/* Get size of FCB blocks */ \
+		"nand device 0 ; " \
+		"nand info ; " \
+		"setexpr fcb_sz ${update_nand_stride} * ${update_nand_count};" \
+		"setexpr update_nand_fcb ${fcb_sz} * ${nand_writesize}\0" \
+	"update_nand_full="		    /* Update FCB, DBBT and FW */ \
+		"if tftp ${update_nand_full_filename} ; then " \
+		"run update_nand_get_fcb_size ; " \
+		"nand scrub -y 0x0 ${filesize} ; " \
+		"nand write.raw ${loadaddr} 0x0 ${fcb_sz} ; " \
+		"setexpr update_off ${loadaddr} + ${update_nand_fcb} ; " \
+		"setexpr update_sz ${filesize} - ${update_nand_fcb} ; " \
+		"nand write ${update_off} ${update_nand_fcb} ${update_sz} ; " \
+		"fi\0" \
+	"update_nand_firmware="		/* Update only firmware */ \
+		"if tftp ${update_nand_firmware_filename} ; then " \
+		"run update_nand_get_fcb_size ; " \
+		"setexpr fcb_sz ${update_nand_fcb} * 2 ; " /* FCB + DBBT */ \
+		"setexpr fw_sz ${update_nand_firmware_maxsz} * 2 ; " \
+		"setexpr fw_off ${fcb_sz} + ${update_nand_firmware_maxsz};" \
+		"nand erase ${fcb_sz} ${fw_sz} ; " \
+		"nand write ${loadaddr} ${fcb_sz} ${filesize} ; " \
+		"nand write ${loadaddr} ${fw_off} ${filesize} ; " \
+		"fi\0" \
+	"update_sd_firmware="		/* Update the SD firmware partition */ \
+		"if mmc rescan ; then "	\
+		"if tftp ${update_sd_firmware_filename} ; then " \
+		"setexpr fw_sz ${filesize} / 0x200 ; "	/* SD block size */ \
+		"setexpr fw_sz ${fw_sz} + 1 ; "	\
+		"mmc write ${loadaddr} 0x800 ${fw_sz} ; " \
+		"fi ; "	\
+		"fi\0" \
+	"script=boot.scr\0"	\
+	"uimage=uImage\0" \
+	"console=ttyAMA0\0" \
+	"fdt_file=imx23-evk.dtb\0" \
+	"fdt_addr=0x41000000\0" \
+	"boot_fdt=try\0" \
+	"mmcdev=0\0" \
+	"mmcpart=2\0" \
+	"mmcroot=/dev/mmcblk0p3 rw rootwait\0" \
+	"mmcargs=setenv bootargs console=${console},${baudrate} " \
+		"root=${mmcroot}\0" \
+	"loadbootscript="  \
+		"fatload mmc ${mmcdev}:${mmcpart} ${loadaddr} ${script};\0" \
+	"bootscript=echo Running bootscript from mmc ...; "	\
+		"source\0" \
+	"loaduimage=fatload mmc ${mmcdev}:${mmcpart} ${loadaddr} ${uimage}\0" \
+	"loadfdt=fatload mmc ${mmcdev}:${mmcpart} ${fdt_addr} ${fdt_file}\0" \
+	"mmcboot=echo Booting from mmc ...; " \
+		"run mmcargs; " \
+		"if test ${boot_fdt} = yes || test ${boot_fdt} = try; then " \
+			"if run loadfdt; then " \
+				"bootm ${loadaddr} - ${fdt_addr}; " \
+			"else " \
+				"if test ${boot_fdt} = try; then " \
+					"bootm; " \
+				"else " \
+					"echo WARN: Cannot load the DT; " \
+				"fi; " \
+			"fi; " \
+		"else " \
+			"bootm; " \
+		"fi;\0"
+
+#define CONFIG_BOOTCOMMAND \
+	"mmc dev ${mmcdev}; if mmc rescan; then " \
+		"if run loadbootscript; then " \
+			"run bootscript; " \
+		"else " \
+			"if run loaduimage; then " \
+				"run mmcboot; " \
+			"else " \
+				"echo ERR: Fail to boot from MMC; " \
+			"fi; " \
+		"fi; " \
+	"else exit; fi"
+
 #endif /* __MX23_OLINUXINO_CONFIG_H__ */

Signature · 1 line

1.8.1

Marek Vasut

2013-01-22 01:49 UTC #26 lore raw

PATCH 7/9 mx23_olinuxino: Add default environment

Dear Otavio Salvador,

Description missing.

Quoted 11 lines

Signed-off-by Otavio Salvador <[email protected]>
---
 include/configs/mx23_olinuxino.h | 90
+++++++++++++++++++++++++++++++++++++++- 1 file changed, 89 insertions(+), 1 deletion(-)
(quoted diff) +0 -0
b/include/configs/mx23_olinuxino.h index a62e3a2..7d34168 100644

include/configs/mx23_olinuxino.h line 143

Quoted 4 lines

include/configs/mx23_olinuxino.h +0 -1

@@ -140,9 +140,97 @@
 #define	CONFIG_SETUP_MEMORY_TAGS
 #define	CONFIG_BOOTDELAY	3
 #define	CONFIG_BOOTFILE		"uImage"
-#define	CONFIG_BOOTARGS		"console=ttyAMA0,115200n8 "

No need to remove this.

Quoted 6 lines

include/configs/mx23_olinuxino.h +2 -0

 #define	CONFIG_LOADADDR		0x42000000
 #define	CONFIG_SYS_LOAD_ADDR	CONFIG_LOADADDR
 #define	CONFIG_OF_LIBFDT
 
+/*
+ * Extra Environments

include/configs/mx23_olinuxino.h line 158

Quoted 10 lines

include/configs/mx23_olinuxino.h +10 -0

+ */
+#define CONFIG_EXTRA_ENV_SETTINGS \
+	"update_sd_firmware_filename=u-boot.sd\0" \
+	"update_nand_full_filename=u-boot.nand\0" \
+	"update_nand_firmware_filename=u-boot.sb\0"	\
+	"update_sd_firmware_filename=u-boot.sd\0" \
+	"update_nand_firmware_maxsz=0x100000\0"	\
+	"update_nand_stride=0x40\0"	/* MX28 datasheet ch. 12.12 */ \
+	"update_nand_count=0x4\0"	/* MX28 datasheet ch. 12.12 */ \
+	"update_nand_get_fcb_size="	/* Get size of FCB blocks */ \

This board has no NAND.

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 02:00 UTC #27 lore raw

PATCH 7/9 mx23_olinuxino: Add default environment

Quoted 10 lines · Marek Vasut

Dear Otavio Salvador,
Description missing.
Quoted 6 lines
Signed-off-by Otavio Salvador <[email protected]>
---
 include/configs/mx23_olinuxino.h | 90
+++++++++++++++++++++++++++++++++++++++- 1 file changed, 89 insertions(+), 1 deletion(-)

Quoted 11 lines

include/configs/mx23_olinuxino.h line 143
Quoted 9 lines
(quoted diff) +0 -0
b/include/configs/mx23_olinuxino.h index a62e3a2..7d34168 100644
include/configs/mx23_olinuxino.h +0 -1
@@ -140,9 +140,97 @@
 #define      CONFIG_SETUP_MEMORY_TAGS
 #define      CONFIG_BOOTDELAY        3
 #define      CONFIG_BOOTFILE         "uImage"
-#define      CONFIG_BOOTARGS         "console=ttyAMA0,115200n8 "
No need to remove this.

It has an extended version in the end.

Quoted 4 lines

include/configs/mx23_olinuxino.h +0 -0

 #define      CONFIG_LOADADDR         0x42000000
 #define      CONFIG_SYS_LOAD_ADDR    CONFIG_LOADADDR
 #define      CONFIG_OF_LIBFDT

Quoted 14 lines

include/configs/mx23_olinuxino.h line 158

Quoted 12 lines

include/configs/mx23_olinuxino.h +12 -0

+/*
+ * Extra Environments
+ */
+#define CONFIG_EXTRA_ENV_SETTINGS \
+     "update_sd_firmware_filename=u-boot.sd\0" \
+     "update_nand_full_filename=u-boot.nand\0" \
+     "update_nand_firmware_filename=u-boot.sb\0"     \
+     "update_sd_firmware_filename=u-boot.sd\0" \
+     "update_nand_firmware_maxsz=0x100000\0" \
+     "update_nand_stride=0x40\0"     /* MX28 datasheet ch. 12.12 */ \
+     "update_nand_count=0x4\0"       /* MX28 datasheet ch. 12.12 */ \
+     "update_nand_get_fcb_size="     /* Get size of FCB blocks */ \

This board has no NAND.

Good catch; will drop it for v2 after Stefano and Fabio's comments.

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Fabio Estevam

reviewer

2013-01-22 02:34 UTC #28 lore raw

PATCH 7/9 mx23_olinuxino: Add default environment

In reply to Otavio Salvador

include/configs/mx23_olinuxino.h line 193

Quoted 4 lines · Otavio Salvador

include/configs/mx23_olinuxino.h +4 -0

+       "script=boot.scr\0"     \
+       "uimage=uImage\0" \
+       "console=ttyAMA0\0" \
+       "fdt_file=imx23-evk.dtb\0" \

The dtb file should be imx23-olinuxino.dtb instead.

Otavio Salvador

author

2013-01-22 01:47 UTC #29 lore raw

PATCH 8/9 mxs: Fix MMC reset to avoid write errors and timeout in iMX23

In reply to Otavio Salvador

This does the same reset mask as done in v3.7 Linux kernel code; it fixed the saveenv command in mx23evk.

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/mmc/mxsmmc.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

drivers/mmc/mxsmmc.c +11 -5

index 618659e..0a3f0c4 100644
@@ -354,11 +354,17 @@ static int mxsmmc_init(struct mmc *mmc)
 	/* Reset SSP */
 	mxs_reset_block(&ssp_regs->hw_ssp_ctrl0_reg);
 
-	/* 8 bits word length in MMC mode */
-	clrsetbits_le32(&ssp_regs->hw_ssp_ctrl1,
-		SSP_CTRL1_SSP_MODE_MASK | SSP_CTRL1_WORD_LENGTH_MASK |
-		SSP_CTRL1_DMA_ENABLE,
-		SSP_CTRL1_SSP_MODE_SD_MMC | SSP_CTRL1_WORD_LENGTH_EIGHT_BITS);
+	/* Reset MMC */
+	writel(SSP_CTRL1_SSP_MODE_SD_MMC |
+		SSP_CTRL1_WORD_LENGTH_EIGHT_BITS |
+		SSP_CTRL1_DMA_ENABLE |
+		SSP_CTRL1_POLARITY |
+		SSP_CTRL1_RECV_TIMEOUT_IRQ_EN |
+		SSP_CTRL1_DATA_CRC_IRQ_EN |
+		SSP_CTRL1_DATA_TIMEOUT_IRQ_EN |
+		SSP_CTRL1_RESP_TIMEOUT_IRQ_EN |
+		SSP_CTRL1_RESP_ERR_IRQ_EN,
+		&ssp_regs->hw_ssp_ctrl1_set);
 
 	/* Set initial bit clock 400 KHz */
 	mxs_set_ssp_busclock(priv->id, 400);

Signature · 1 line

1.8.1

Marek Vasut

2013-01-22 01:49 UTC #30 lore raw

PATCH 8/9 mxs: Fix MMC reset to avoid write errors and timeout in iMX23

Dear Otavio Salvador,

Quoted 16 lines

This does the same reset mask as done in v3.7 Linux kernel code; it fixed the saveenv command in mx23evk.
Signed-off-by Otavio Salvador <[email protected]>
---
 drivers/mmc/mxsmmc.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
drivers/mmc/mxsmmc.c +0 -0
index 618659e..0a3f0c4 100644
@@ -354,11 +354,17 @@ static int mxsmmc_init(struct mmc *mmc)
 	/* Reset SSP */
 	mxs_reset_block(&ssp_regs->hw_ssp_ctrl0_reg);
 

drivers/mmc/mxsmmc.c static int mxsmmc_init(struct mmc *mmc) · line 361

Quoted 10 lines

drivers/mmc/mxsmmc.c +5 -5

-	/* 8 bits word length in MMC mode */
-	clrsetbits_le32(&ssp_regs->hw_ssp_ctrl1,
-		SSP_CTRL1_SSP_MODE_MASK | SSP_CTRL1_WORD_LENGTH_MASK |
-		SSP_CTRL1_DMA_ENABLE,
-		SSP_CTRL1_SSP_MODE_SD_MMC | SSP_CTRL1_WORD_LENGTH_EIGHT_BITS);
+	/* Reset MMC */
+	writel(SSP_CTRL1_SSP_MODE_SD_MMC |
+		SSP_CTRL1_WORD_LENGTH_EIGHT_BITS |
+		SSP_CTRL1_DMA_ENABLE |
+		SSP_CTRL1_POLARITY |

Will this addition not break mx28?

drivers/mmc/mxsmmc.c static int mxsmmc_init(struct mmc *mmc) · line 370

Quoted 9 lines

drivers/mmc/mxsmmc.c +6 -0

+		SSP_CTRL1_RECV_TIMEOUT_IRQ_EN |
+		SSP_CTRL1_DATA_CRC_IRQ_EN |
+		SSP_CTRL1_DATA_TIMEOUT_IRQ_EN |
+		SSP_CTRL1_RESP_TIMEOUT_IRQ_EN |
+		SSP_CTRL1_RESP_ERR_IRQ_EN,
+		&ssp_regs->hw_ssp_ctrl1_set);
 
 	/* Set initial bit clock 400 KHz */
 	mxs_set_ssp_busclock(priv->id, 400);

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 02:03 UTC #31 lore raw

PATCH 8/9 mxs: Fix MMC reset to avoid write errors and timeout in iMX23

Quoted 18 lines · Marek Vasut

Dear Otavio Salvador,
Quoted 16 lines
This does the same reset mask as done in v3.7 Linux kernel code; it fixed the saveenv command in mx23evk.
Signed-off-by Otavio Salvador <[email protected]>
---
 drivers/mmc/mxsmmc.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
drivers/mmc/mxsmmc.c +0 -0
index 618659e..0a3f0c4 100644
@@ -354,11 +354,17 @@ static int mxsmmc_init(struct mmc *mmc)
      /* Reset SSP */
      mxs_reset_block(&ssp_regs->hw_ssp_ctrl0_reg);
 

Quoted 12 lines

drivers/mmc/mxsmmc.c static int mxsmmc_init(struct mmc *mmc) · line 361

Quoted 10 lines

drivers/mmc/mxsmmc.c +5 -5

-     /* 8 bits word length in MMC mode */
-     clrsetbits_le32(&ssp_regs->hw_ssp_ctrl1,
-             SSP_CTRL1_SSP_MODE_MASK | SSP_CTRL1_WORD_LENGTH_MASK |
-             SSP_CTRL1_DMA_ENABLE,
-             SSP_CTRL1_SSP_MODE_SD_MMC | SSP_CTRL1_WORD_LENGTH_EIGHT_BITS);
+     /* Reset MMC */
+     writel(SSP_CTRL1_SSP_MODE_SD_MMC |
+             SSP_CTRL1_WORD_LENGTH_EIGHT_BITS |
+             SSP_CTRL1_DMA_ENABLE |
+             SSP_CTRL1_POLARITY |

Will this addition not break mx28?

Not tested. I hope it does not as it has been taken from v3.7 kernel.

Quoted 7 lines

drivers/mmc/mxsmmc.c +6 -0

+             SSP_CTRL1_RECV_TIMEOUT_IRQ_EN |
+             SSP_CTRL1_DATA_CRC_IRQ_EN |
+             SSP_CTRL1_DATA_TIMEOUT_IRQ_EN |
+             SSP_CTRL1_RESP_TIMEOUT_IRQ_EN |
+             SSP_CTRL1_RESP_ERR_IRQ_EN,
+             &ssp_regs->hw_ssp_ctrl1_set);

Quoted 5 lines

drivers/mmc/mxsmmc.c static int mxsmmc_init(struct mmc *mmc) · line 370
Quoted 2 lines
drivers/mmc/mxsmmc.c +0 -0
      /* Set initial bit clock 400 KHz */
      mxs_set_ssp_busclock(priv->id, 400);
Best regards,
Marek Vasut

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Marek Vasut

2013-01-22 02:15 UTC #32 lore raw

PATCH 8/9 mxs: Fix MMC reset to avoid write errors and timeout in iMX23

Dear Otavio Salvador,

Quoted 34 lines · Marek Vasut

Quoted 20 lines
Dear Otavio Salvador,
Quoted 18 lines
This does the same reset mask as done in v3.7 Linux kernel code; it fixed the saveenv command in mx23evk.
Signed-off-by Otavio Salvador <[email protected]>
---
 drivers/mmc/mxsmmc.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
drivers/mmc/mxsmmc.c +0 -0
index 618659e..0a3f0c4 100644
@@ -354,11 +354,17 @@ static int mxsmmc_init(struct mmc *mmc)
 
      /* Reset SSP */
      mxs_reset_block(&ssp_regs->hw_ssp_ctrl0_reg);
 
Quoted 12 lines
drivers/mmc/mxsmmc.c static int mxsmmc_init(struct mmc *mmc) · line 361
Quoted 10 lines
drivers/mmc/mxsmmc.c +0 -5
-     /* 8 bits word length in MMC mode */
-     clrsetbits_le32(&ssp_regs->hw_ssp_ctrl1,
-             SSP_CTRL1_SSP_MODE_MASK | SSP_CTRL1_WORD_LENGTH_MASK |
-             SSP_CTRL1_DMA_ENABLE,
-             SSP_CTRL1_SSP_MODE_SD_MMC |
SSP_CTRL1_WORD_LENGTH_EIGHT_BITS); + /* Reset MMC */
drivers/mmc/mxsmmc.c +4 -0
+     writel(SSP_CTRL1_SSP_MODE_SD_MMC |
+             SSP_CTRL1_WORD_LENGTH_EIGHT_BITS |
+             SSP_CTRL1_DMA_ENABLE |
+             SSP_CTRL1_POLARITY |
Will this addition not break mx28?

Quoted 1 line

Not tested. I hope it does not as it has been taken from v3.7 kernel.

Please, mark all such untested code which can possibly break other platforms in the description.

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 01:47 UTC #33 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

In reply to Otavio Salvador

Signed-off-by Otavio Salvador <[email protected]>

---
 drivers/mmc/mxsmmc.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

drivers/mmc/mxsmmc.c +6 -2

index 0a3f0c4..628d609 100644
@@ -50,6 +50,7 @@ struct mxsmmc_priv {
 	uint32_t		buswidth;
 	int			(*mmc_is_wp)(int);
 	struct mxs_dma_desc	*desc;
+	int force_pio;
 };
 
 #define	MXSMMC_MAX_TIMEOUT	10000

@@ -184,7 +185,7 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, struct mmc_data *data)
 	if (cmd->resp_type & MMC_RSP_136)	/* It's a 136 bits response */
 		ctrl0 |= SSP_CTRL0_LONG_RESP;
 
-	if (data && (data->blocksize * data->blocks < MXSMMC_SMALL_TRANSFER))
+	if (priv->force_pio || (data && (data->blocksize * data->blocks < MXSMMC_SMALL_TRANSFER)))
 		writel(SSP_CTRL1_DMA_ENABLE, &ssp_regs->hw_ssp_ctrl1_clr);
 	else
 		writel(SSP_CTRL1_DMA_ENABLE, &ssp_regs->hw_ssp_ctrl1_set);

@@ -287,7 +288,7 @@ mxsmmc_send_cmd(struct mmc *mmc, struct mmc_cmd *cmd, struct mmc_data *data)
 	if (!data)
 		return 0;
 
-	if (data->blocksize * data->blocks < MXSMMC_SMALL_TRANSFER) {
+	if (priv->force_pio || (data->blocksize * data->blocks < MXSMMC_SMALL_TRANSFER)) {
 		ret = mxsmmc_send_cmd_pio(priv, data);
 		if (ret) {
 			printf("MMC%d: Data timeout with command %d "

@@ -385,9 +386,11 @@ int mxsmmc_initialize(bd_t *bis, int id, int (*wp)(int))
 #if defined(CONFIG_MX23)
 	const unsigned int mxsmmc_max_id = 2;
 	const unsigned int mxsmmc_clk_id = 0;
+	const unsigned int mxsmmc_force_pio = 1;
 #elif defined(CONFIG_MX28)
 	const unsigned int mxsmmc_max_id = 4;
 	const unsigned int mxsmmc_clk_id = id;
+	const unsigned int mxsmmc_force_pio = 0;
 #endif
 
 	if (id >= mxsmmc_max_id)

@@ -417,6 +420,7 @@ int mxsmmc_initialize(bd_t *bis, int id, int (*wp)(int))
 	priv->mmc_is_wp = wp;
 	priv->id = id;
 	priv->regs = mxs_ssp_regs_by_bus(id);
+	priv->force_pio = mxsmmc_force_pio;
 
 	sprintf(mmc->name, "MXS MMC");
 	mmc->send_cmd = mxsmmc_send_cmd;

Signature · 1 line

1.8.1

Marek Vasut

2013-01-22 01:50 UTC #34 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

Dear Otavio Salvador,

NAK. DMA works on MX28 and works in Linux on MX23, therefore the DMA is not broken on either. Thus there's no reason to introduce such "force_pio" stuff.

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-22 01:56 UTC #35 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

Quoted 4 lines · Marek Vasut

Dear Otavio Salvador,
NAK. DMA works on MX28 and works in Linux on MX23, therefore the DMA is not broken on either. Thus there's no reason to introduce such "force_pio" stuff.

Lack of time to work on it is one reason. I won't be able to work on it for some days and I prefer a working solution than skip it for another release.

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Marek Vasut

2013-01-22 01:59 UTC #36 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

Dear Otavio Salvador,

Quoted 10 lines · Marek Vasut

Quoted 5 lines
Dear Otavio Salvador,
NAK. DMA works on MX28 and works in Linux on MX23, therefore the DMA is not broken on either. Thus there's no reason to introduce such "force_pio" stuff.
Lack of time to work on it is one reason. I won't be able to work on it for some days and I prefer a working solution than skip it for another release.

The solution you propose adds undocumented feature and is not working, this is clearly a workaround.

Best regards,
Marek Vasut

Stefano Babic

2013-01-23 09:04 UTC #37 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

In reply to Otavio Salvador

Quoted 6 lines · Otavio Salvador

Quoted 4 lines · Marek Vasut
Dear Otavio Salvador,
NAK. DMA works on MX28 and works in Linux on MX23, therefore the DMA is not broken on either. Thus there's no reason to introduce such "force_pio" stuff.

Quoted 9 lines

Lack of time to work on it is one reason. I won't be able to work on it for some days and I prefer a working solution than skip it for another release.
--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Hi Otavio, hi Marek,

let me say some words about this discussion. I agree that if there is no support for DMA for MX23, we could add support with PIO mode. However, there should be a very good reason, because as far as I understand, added support for DMA is valid for mxs platform. The solution introduced with this patchset lets also the door open for people that can use MX28 in PIO mode. I think (I have already seen in ML Marek's patchset for it) the way to fix DMA for mx23 and then add support for MMC is the rught way to do.

Best regards,
Stefano Babic

Signature · 5 lines

=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: sbabic at denx.de
=====================================================================

Otavio Salvador

author

2013-01-23 11:05 UTC #38 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

Quoted 19 lines · Stefano Babic

Quoted 6 lines · Otavio Salvador
Quoted 4 lines · Marek Vasut
Dear Otavio Salvador,
NAK. DMA works on MX28 and works in Linux on MX23, therefore the DMA is not broken on either. Thus there's no reason to introduce such "force_pio" stuff.
Quoted 9 lines
Lack of time to work on it is one reason. I won't be able to work on it for some days and I prefer a working solution than skip it for another release.
--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br
Hi Otavio, hi Marek,

Quoted 8 lines

let me say some words about this discussion. I agree that if there is no support for DMA for MX23, we could add support with PIO mode. However, there should be a very good reason, because as far as I understand, added support for DMA is valid for mxs platform. The solution introduced with this patchset lets also the door open for people that can use MX28 in PIO mode. I think (I have already seen in ML Marek's patchset for it) the way to fix DMA for mx23 and then add support for MMC is the rught way to do.

So you'd like to have this patch ported to Marek's patchset so the support for PIO mode is kept available? Instead of enabling it for MX23 we can have a config option which allows force PIO mode and disable DMA. This is what you want?

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br

Marek Vasut

2013-01-23 11:26 UTC #39 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

Dear Otavio Salvador,

Quoted 29 lines · Stefano Babic

Quoted 19 lines · Otavio Salvador
Quoted 7 lines · Marek Vasut
Quoted 5 lines
Dear Otavio Salvador,
NAK. DMA works on MX28 and works in Linux on MX23, therefore the DMA is not broken on either. Thus there's no reason to introduce such "force_pio" stuff.
Quoted 8 lines
Lack of time to work on it is one reason. I won't be able to work on it for some days and I prefer a working solution than skip it for another release.
--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br
Hi Otavio, hi Marek,
Quoted 8 lines
let me say some words about this discussion. I agree that if there is no support for DMA for MX23, we could add support with PIO mode. However, there should be a very good reason, because as far as I understand, added support for DMA is valid for mxs platform. The solution introduced with this patchset lets also the door open for people that can use MX28 in PIO mode. I think (I have already seen in ML Marek's patchset for it) the way to fix DMA for mx23 and then add support for MMC is the rught way to do.

Quoted 4 lines

So you'd like to have this patch ported to Marek's patchset so the support for PIO mode is kept available? Instead of enabling it for MX23 we can have a config option which allows force PIO mode and disable DMA. This is what you want?

Forcing PIO is nonsense, why would it be of any use?

Best regards,
Marek Vasut

Stefano Babic

2013-01-23 11:39 UTC #40 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

Quoted 11 lines · Marek Vasut

Quoted 11 lines
Quoted 10 lines
Hi Otavio, hi Marek,
let me say some words about this discussion. I agree that if there is no support for DMA for MX23, we could add support with PIO mode. However, there should be a very good reason, because as far as I understand, added support for DMA is valid for mxs platform. The solution introduced with this patchset lets also the door open for people that can use MX28 in PIO mode. I think (I have already seen in ML Marek's patchset for it) the way to fix DMA for mx23 and then add support for MMC is the rught way to do.

Quoted 6 lines

Quoted 4 lines
So you'd like to have this patch ported to Marek's patchset so the support for PIO mode is kept available? Instead of enabling it for MX23 we can have a config option which allows force PIO mode and disable DMA. This is what you want?
Forcing PIO is nonsense, why would it be of any use?

No, I agree with Marek - if DMA is then working also for MX23, we do not need PIO at all.

Best regards,
Stefano Babic

Signature · 5 lines

=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: sbabic at denx.de
=====================================================================

Marek Vasut

2013-01-23 11:48 UTC #41 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

Dear Stefano Babic,

Quoted 17 lines · Marek Vasut

Quoted 16 lines
Quoted 15 lines
Quoted 10 lines
Hi Otavio, hi Marek,
let me say some words about this discussion. I agree that if there is no support for DMA for MX23, we could add support with PIO mode. However, there should be a very good reason, because as far as I understand, added support for DMA is valid for mxs platform. The solution introduced with this patchset lets also the door open for people that can use MX28 in PIO mode. I think (I have already seen in ML Marek's patchset for it) the way to fix DMA for mx23 and then add support for MMC is the rught way to do.
So you'd like to have this patch ported to Marek's patchset so the support for PIO mode is kept available? Instead of enabling it for MX23 we can have a config option which allows force PIO mode and disable DMA. This is what you want?

Quoted 4 lines

Quoted 1 line
Forcing PIO is nonsense, why would it be of any use?
No, I agree with Marek - if DMA is then working also for MX23, we do not need PIO at all.

Stefano, it's now using mixed mode -- PIO for small transfers, DMA for large transfers. That is the best pick of both worlds ;-)

Best regards,
Marek Vasut

Otavio Salvador

author

2013-01-23 12:12 UTC #42 lore raw

PATCH 9/9 mx23: Use PIO mode support for MMC

Quoted 19 lines · Marek Vasut

Dear Stefano Babic,
Quoted 17 lines · Marek Vasut
Quoted 16 lines
Quoted 15 lines
Quoted 10 lines
Hi Otavio, hi Marek,
let me say some words about this discussion. I agree that if there is no support for DMA for MX23, we could add support with PIO mode. However, there should be a very good reason, because as far as I understand, added support for DMA is valid for mxs platform. The solution introduced with this patchset lets also the door open for people that can use MX28 in PIO mode. I think (I have already seen in ML Marek's patchset for it) the way to fix DMA for mx23 and then add support for MMC is the rught way to do.
So you'd like to have this patch ported to Marek's patchset so the support for PIO mode is kept available? Instead of enabling it for MX23 we can have a config option which allows force PIO mode and disable DMA. This is what you want?

Quoted 7 lines

Quoted 4 lines
Quoted 1 line
Forcing PIO is nonsense, why would it be of any use?
No, I agree with Marek - if DMA is then working also for MX23, we do not need PIO at all.
Stefano, it's now using mixed mode -- PIO for small transfers, DMA for large transfers. That is the best pick of both worlds ;-)

That's fine; I just wanted to check if I had understood it right.

No problem; I will update the environments and send a v2 of my patchset basing on top of Marek's changes.

--
Otavio Salvador O.S. Systems
E-mail: otavio at ossystems.com.br http://www.ossystems.com.br
Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br