[PATCH 0/6] Some fixes and cleanup on imx-dma device registration

The patch set makes some small fixes on imx-dma device registration
code, and then cleans up the code by breaking the concentrated
imx-dma device registration into soc specific setup function.
This is basically what gpio-mxc device registration is doing.
In this way, we can avoid the churn of "#ifdef" and the cpu_is_mx
checking on such a long list of soc family, which makes no sense,
considering more soc supports need to be added and we need to
support single image for multiple socs in the long run.

Shawn Guo (6):
      ARM: mxc: imx-sdma device gets 16K iosize than 4K
      ARM: mxc: sdma on imx25 is a V2 block
      ARM: mxc: change imx-dma default to_version to 1
      ARM: mxc: imx-dma on imx25 has no other TO version but TO1
      ARM: mxc: all three imx51 TOs use TO1 version of sdma script
      ARM: mxc: clean up imx-dma device registration

 arch/arm/mach-imx/mm-imx21.c                    |    3 +
 arch/arm/mach-imx/mm-imx25.c                    |   26 +++
 arch/arm/mach-imx/mm-imx27.c                    |    3 +
 arch/arm/mach-imx/mm-imx31.c                    |   24 +++
 arch/arm/mach-imx/mm-imx35.c                    |   44 +++++
 arch/arm/mach-mx5/mm.c                          |   23 +++
 arch/arm/plat-mxc/devices.c                     |   16 ++-
 arch/arm/plat-mxc/devices/platform-imx-dma.c    |  204 ++---------------------
 arch/arm/plat-mxc/include/mach/devices-common.h |    6 +
 9 files changed, 157 insertions(+), 192 deletions(-)

The sdma on all imx soc gets 16K IO space not 4K.

Signed-off-by: Shawn Guo <redacted>
---
 arch/arm/plat-mxc/devices/platform-imx-dma.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/arch/arm/plat-mxc/devices/platform-imx-dma.c b/arch/arm/plat-mxc/devices/platform-imx-dma.c
index b130f60..222e439 100644
--- a/arch/arm/plat-mxc/devices/platform-imx-dma.c
+++ b/arch/arm/plat-mxc/devices/platform-imx-dma.c

@@ -57,7 +57,7 @@ static struct platform_device __init __maybe_unused *imx_add_imx_sdma(
 	struct resource res[] = {
 		{
 			.start = data->iobase,
-			.end = data->iobase + SZ_4K - 1,
+			.end = data->iobase + SZ_16K - 1,
 			.flags = IORESOURCE_MEM,
 		}, {
 			.start = data->irq,

-- 
1.7.4.1

The sdma on soc imx25 is not a V1 but V2 block.

Signed-off-by: Shawn Guo <redacted>
---
 arch/arm/plat-mxc/devices/platform-imx-dma.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/arch/arm/plat-mxc/devices/platform-imx-dma.c b/arch/arm/plat-mxc/devices/platform-imx-dma.c
index 222e439..1c0ba53 100644
--- a/arch/arm/plat-mxc/devices/platform-imx-dma.c
+++ b/arch/arm/plat-mxc/devices/platform-imx-dma.c

@@ -33,7 +33,7 @@ struct imx_imx_sdma_data {
 
 #ifdef CONFIG_SOC_IMX25
 struct imx_imx_sdma_data imx25_imx_sdma_data __initconst =
-	imx_imx_sdma_data_entry_single(MX25, 1, "imx25", 0);
+	imx_imx_sdma_data_entry_single(MX25, 2, "imx25", 0);
 #endif /* ifdef CONFIG_SOC_IMX25 */
 
 #ifdef CONFIG_SOC_IMX31

-- 
1.7.4.1

The value 0 is not a valid TO version number.  With the current
code, imx-sdma driver will try to load firmware sdma-imx25-to0.bin,
which is obviously not a good name.  Instead, sdma-imx25-to1.bin
makes much more sense.

Signed-off-by: Shawn Guo <redacted>
---
 arch/arm/plat-mxc/devices/platform-imx-dma.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/arch/arm/plat-mxc/devices/platform-imx-dma.c b/arch/arm/plat-mxc/devices/platform-imx-dma.c
index 1c0ba53..c2712b7 100644
--- a/arch/arm/plat-mxc/devices/platform-imx-dma.c
+++ b/arch/arm/plat-mxc/devices/platform-imx-dma.c

@@ -33,22 +33,22 @@ struct imx_imx_sdma_data {
 
 #ifdef CONFIG_SOC_IMX25
 struct imx_imx_sdma_data imx25_imx_sdma_data __initconst =
-	imx_imx_sdma_data_entry_single(MX25, 2, "imx25", 0);
+	imx_imx_sdma_data_entry_single(MX25, 2, "imx25", 1);
 #endif /* ifdef CONFIG_SOC_IMX25 */
 
 #ifdef CONFIG_SOC_IMX31
 struct imx_imx_sdma_data imx31_imx_sdma_data __initdata =
-	imx_imx_sdma_data_entry_single(MX31, 1, "imx31", 0);
+	imx_imx_sdma_data_entry_single(MX31, 1, "imx31", 1);
 #endif /* ifdef CONFIG_SOC_IMX31 */
 
 #ifdef CONFIG_SOC_IMX35
 struct imx_imx_sdma_data imx35_imx_sdma_data __initdata =
-	imx_imx_sdma_data_entry_single(MX35, 2, "imx35", 0);
+	imx_imx_sdma_data_entry_single(MX35, 2, "imx35", 1);
 #endif /* ifdef CONFIG_SOC_IMX35 */
 
 #ifdef CONFIG_SOC_IMX51
 struct imx_imx_sdma_data imx51_imx_sdma_data __initconst =
-	imx_imx_sdma_data_entry_single(MX51, 2, "imx51", 0);
+	imx_imx_sdma_data_entry_single(MX51, 2, "imx51", 1);
 #endif /* ifdef CONFIG_SOC_IMX51 */
 
 static struct platform_device __init __maybe_unused *imx_add_imx_sdma(

-- 
1.7.4.1

The imx25 sdma script only gets TO1 version, so there is no need
to encode "to1" in the variable name.

Signed-off-by: Shawn Guo <redacted>
---
 arch/arm/plat-mxc/devices/platform-imx-dma.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/arm/plat-mxc/devices/platform-imx-dma.c b/arch/arm/plat-mxc/devices/platform-imx-dma.c
index c2712b7..c64f015 100644
--- a/arch/arm/plat-mxc/devices/platform-imx-dma.c
+++ b/arch/arm/plat-mxc/devices/platform-imx-dma.c

@@ -77,7 +77,7 @@ static struct platform_device __init __maybe_unused *imx_add_imx_dma(void)
 }
 
 #ifdef CONFIG_ARCH_MX25
-static struct sdma_script_start_addrs addr_imx25_to1 = {
+static struct sdma_script_start_addrs addr_imx25 = {
 	.ap_2_ap_addr = 729,
 	.uart_2_mcu_addr = 904,
 	.per_2_app_addr = 1255,

@@ -165,7 +165,7 @@ static int __init imxXX_add_imx_dma(void)
 
 #if defined(CONFIG_SOC_IMX25)
 	if (cpu_is_mx25()) {
-		imx25_imx_sdma_data.pdata.script_addrs = &addr_imx25_to1;
+		imx25_imx_sdma_data.pdata.script_addrs = &addr_imx25;
 		ret = imx_add_imx_sdma(&imx25_imx_sdma_data);
 	} else
 #endif

-- 
1.7.4.1

Though there are three TOs of imx51 soc, the sdma script never
changes since TO1, which means all three TOs of imx51 uses TO1
version of sdma script.

The current code passes TO number to imx-sdma driver to load
different firmware for different TO.  That means we have to prepare
3 identical firmwares, sdma-imx25-to1.bin sdma-imx25-to2.bin and
sdma-imx25-to3.bin, to have the kernel capable of running on all
three TOs.  This just makes no sense.

The patch removes the TO number passing and get the default TO1
version of sdma firmware work for all TOs.

Signed-off-by: Shawn Guo <redacted>
---
 arch/arm/plat-mxc/devices/platform-imx-dma.c |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/arch/arm/plat-mxc/devices/platform-imx-dma.c b/arch/arm/plat-mxc/devices/platform-imx-dma.c
index c64f015..2091540 100644
--- a/arch/arm/plat-mxc/devices/platform-imx-dma.c
+++ b/arch/arm/plat-mxc/devices/platform-imx-dma.c

@@ -196,8 +196,6 @@ static int __init imxXX_add_imx_dma(void)
 
 #if defined(CONFIG_SOC_IMX51)
 	if (cpu_is_mx51()) {
-		int to_version = mx51_revision() >> 4;
-		imx51_imx_sdma_data.pdata.to_version = to_version;
 		imx51_imx_sdma_data.pdata.script_addrs = &addr_imx51;
 		ret = imx_add_imx_sdma(&imx51_imx_sdma_data);
 	} else

-- 
1.7.4.1

The patch follows the implementation of gpio-mxc device registration
to break the concentrated imx-dma device registration into soc
specific setup function.  Then we can avoid the churn of "#ifdef"
and the cpu_is_mx checking on such a long list, which makes no sense,
considering more soc supports need to be added and we need to support
single image for multiple socs in the long run.

Signed-off-by: Shawn Guo <redacted>
---
 arch/arm/mach-imx/mm-imx21.c                    |    3 +
 arch/arm/mach-imx/mm-imx25.c                    |   26 +++
 arch/arm/mach-imx/mm-imx27.c                    |    3 +
 arch/arm/mach-imx/mm-imx31.c                    |   24 +++
 arch/arm/mach-imx/mm-imx35.c                    |   44 +++++
 arch/arm/mach-mx5/mm.c                          |   23 +++
 arch/arm/plat-mxc/devices.c                     |   16 ++-
 arch/arm/plat-mxc/devices/platform-imx-dma.c    |  202 ++---------------------
 arch/arm/plat-mxc/include/mach/devices-common.h |    6 +
 9 files changed, 157 insertions(+), 190 deletions(-)

diff --git a/arch/arm/mach-imx/mm-imx21.c b/arch/arm/mach-imx/mm-imx21.c
index f0fb8bc..8d27f47 100644
--- a/arch/arm/mach-imx/mm-imx21.c
+++ b/arch/arm/mach-imx/mm-imx21.c

@@ -22,6 +22,7 @@
 #include <linux/init.h>
 #include <mach/hardware.h>
 #include <mach/common.h>
+#include <mach/devices-common.h>
 #include <asm/pgtable.h>
 #include <asm/mach/map.h>
 #include <mach/irqs.h>

@@ -82,4 +83,6 @@ void __init imx21_soc_init(void)
 	mxc_register_gpio(3, MX21_GPIO4_BASE_ADDR, SZ_4K, MX21_INT_GPIO, 0);
 	mxc_register_gpio(4, MX21_GPIO5_BASE_ADDR, SZ_4K, MX21_INT_GPIO, 0);
 	mxc_register_gpio(5, MX21_GPIO6_BASE_ADDR, SZ_4K, MX21_INT_GPIO, 0);
+
+	imx_add_imx_dma();
 }

diff --git a/arch/arm/mach-imx/mm-imx25.c b/arch/arm/mach-imx/mm-imx25.c
index 1b6d583..0c54520 100644
--- a/arch/arm/mach-imx/mm-imx25.c
+++ b/arch/arm/mach-imx/mm-imx25.c

@@ -24,6 +24,7 @@
 #include <asm/mach/map.h>
 
 #include <mach/common.h>
+#include <mach/devices-common.h>
 #include <mach/hardware.h>
 #include <mach/mx25.h>
 #include <mach/iomux-v3.h>

@@ -61,10 +62,35 @@ void __init mx25_init_irq(void)
 	mxc_init_irq(MX25_IO_ADDRESS(MX25_AVIC_BASE_ADDR));
 }
 
+static struct sdma_script_start_addrs imx25_sdma_script __initdata = {
+	.ap_2_ap_addr = 729,
+	.uart_2_mcu_addr = 904,
+	.per_2_app_addr = 1255,
+	.mcu_2_app_addr = 834,
+	.uartsh_2_mcu_addr = 1120,
+	.per_2_shp_addr = 1329,
+	.mcu_2_shp_addr = 1048,
+	.ata_2_mcu_addr = 1560,
+	.mcu_2_ata_addr = 1479,
+	.app_2_per_addr = 1189,
+	.app_2_mcu_addr = 770,
+	.shp_2_per_addr = 1407,
+	.shp_2_mcu_addr = 979,
+};
+
+static struct sdma_platform_data imx25_sdma_pdata __initdata = {
+	.sdma_version = 2,
+	.cpu_name = "imx25",
+	.to_version = 1,
+	.script_addrs = &imx25_sdma_script,
+};
+
 void __init imx25_soc_init(void)
 {
 	mxc_register_gpio(0, MX25_GPIO1_BASE_ADDR, SZ_16K, MX25_INT_GPIO1, 0);
 	mxc_register_gpio(1, MX25_GPIO2_BASE_ADDR, SZ_16K, MX25_INT_GPIO2, 0);
 	mxc_register_gpio(2, MX25_GPIO3_BASE_ADDR, SZ_16K, MX25_INT_GPIO3, 0);
 	mxc_register_gpio(3, MX25_GPIO4_BASE_ADDR, SZ_16K, MX25_INT_GPIO4, 0);
+
+	imx_add_imx_sdma(MX25_SDMA_BASE_ADDR, MX25_INT_SDMA, &imx25_sdma_pdata);
 }

diff --git a/arch/arm/mach-imx/mm-imx27.c b/arch/arm/mach-imx/mm-imx27.c
index d3700ce..47a6ada 100644
--- a/arch/arm/mach-imx/mm-imx27.c
+++ b/arch/arm/mach-imx/mm-imx27.c

@@ -22,6 +22,7 @@
 #include <linux/init.h>
 #include <mach/hardware.h>
 #include <mach/common.h>
+#include <mach/devices-common.h>
 #include <asm/pgtable.h>
 #include <asm/mach/map.h>
 #include <mach/irqs.h>

@@ -82,4 +83,6 @@ void __init imx27_soc_init(void)
 	mxc_register_gpio(3, MX27_GPIO4_BASE_ADDR, SZ_4K, MX27_INT_GPIO, 0);
 	mxc_register_gpio(4, MX27_GPIO5_BASE_ADDR, SZ_4K, MX27_INT_GPIO, 0);
 	mxc_register_gpio(5, MX27_GPIO6_BASE_ADDR, SZ_4K, MX27_INT_GPIO, 0);
+
+	imx_add_imx_dma();
 }

diff --git a/arch/arm/mach-imx/mm-imx31.c b/arch/arm/mach-imx/mm-imx31.c
index cb16ac6..bc00a2f 100644
--- a/arch/arm/mach-imx/mm-imx31.c
+++ b/arch/arm/mach-imx/mm-imx31.c

@@ -24,6 +24,7 @@
 #include <asm/mach/map.h>
 
 #include <mach/common.h>
+#include <mach/devices-common.h>
 #include <mach/hardware.h>
 #include <mach/iomux-v3.h>
 #include <mach/irqs.h>

@@ -57,9 +58,32 @@ void __init mx31_init_irq(void)
 	mxc_init_irq(MX31_IO_ADDRESS(MX31_AVIC_BASE_ADDR));
 }
 
+static struct sdma_script_start_addrs imx31_to1_sdma_script __initdata = {
+	.per_2_per_addr = 1677,
+};
+
+static struct sdma_script_start_addrs imx31_to2_sdma_script __initdata = {
+	.ap_2_ap_addr = 423,
+	.ap_2_bp_addr = 829,
+	.bp_2_ap_addr = 1029,
+};
+
+static struct sdma_platform_data imx31_sdma_pdata __initdata = {
+	.sdma_version = 1,
+	.cpu_name = "imx31",
+};
+
 void __init imx31_soc_init(void)
 {
+	int to_version = mx31_revision() >> 4;
+
 	mxc_register_gpio(0, MX31_GPIO1_BASE_ADDR, SZ_16K, MX31_INT_GPIO1, 0);
 	mxc_register_gpio(1, MX31_GPIO2_BASE_ADDR, SZ_16K, MX31_INT_GPIO2, 0);
 	mxc_register_gpio(2, MX31_GPIO3_BASE_ADDR, SZ_16K, MX31_INT_GPIO3, 0);
+
+	imx31_sdma_pdata.to_version = to_version;
+	imx31_sdma_pdata.script_addrs =
+		(to_version == 1) ?  &imx31_to1_sdma_script :
+				     &imx31_to2_sdma_script;
+	imx_add_imx_sdma(MX31_SDMA_BASE_ADDR, MX31_INT_SDMA, &imx31_sdma_pdata);
 }

diff --git a/arch/arm/mach-imx/mm-imx35.c b/arch/arm/mach-imx/mm-imx35.c
index 648bfca..b3a474b 100644
--- a/arch/arm/mach-imx/mm-imx35.c
+++ b/arch/arm/mach-imx/mm-imx35.c

@@ -25,6 +25,7 @@
 #include <asm/hardware/cache-l2x0.h>
 
 #include <mach/common.h>
+#include <mach/devices-common.h>
 #include <mach/hardware.h>
 #include <mach/iomux-v3.h>
 #include <mach/irqs.h>

@@ -54,9 +55,52 @@ void __init mx35_init_irq(void)
 	mxc_init_irq(MX35_IO_ADDRESS(MX35_AVIC_BASE_ADDR));
 }
 
+static struct sdma_script_start_addrs imx35_to1_sdma_script __initdata = {
+	.ap_2_ap_addr = 642,
+	.uart_2_mcu_addr = 817,
+	.mcu_2_app_addr = 747,
+	.uartsh_2_mcu_addr = 1183,
+	.per_2_shp_addr = 1033,
+	.mcu_2_shp_addr = 961,
+	.ata_2_mcu_addr = 1333,
+	.mcu_2_ata_addr = 1252,
+	.app_2_mcu_addr = 683,
+	.shp_2_per_addr = 1111,
+	.shp_2_mcu_addr = 892,
+};
+
+static struct sdma_script_start_addrs imx35_to2_sdma_script __initdata = {
+	.ap_2_ap_addr = 729,
+	.uart_2_mcu_addr = 904,
+	.per_2_app_addr = 1597,
+	.mcu_2_app_addr = 834,
+	.uartsh_2_mcu_addr = 1270,
+	.per_2_shp_addr = 1120,
+	.mcu_2_shp_addr = 1048,
+	.ata_2_mcu_addr = 1429,
+	.mcu_2_ata_addr = 1339,
+	.app_2_per_addr = 1531,
+	.app_2_mcu_addr = 770,
+	.shp_2_per_addr = 1198,
+	.shp_2_mcu_addr = 979,
+};
+
+static struct sdma_platform_data imx35_sdma_pdata __initdata = {
+	.sdma_version = 2,
+	.cpu_name = "imx35",
+};
+
 void __init imx35_soc_init(void)
 {
+	int to_version = mx35_revision() >> 4;
+
 	mxc_register_gpio(0, MX35_GPIO1_BASE_ADDR, SZ_16K, MX35_INT_GPIO1, 0);
 	mxc_register_gpio(1, MX35_GPIO2_BASE_ADDR, SZ_16K, MX35_INT_GPIO2, 0);
 	mxc_register_gpio(2, MX35_GPIO3_BASE_ADDR, SZ_16K, MX35_INT_GPIO3, 0);
+
+	imx35_sdma_pdata.to_version = to_version;
+	imx35_sdma_pdata.script_addrs =
+		(to_version == 1) ?  &imx35_to1_sdma_script :
+				     &imx35_to2_sdma_script;
+	imx_add_imx_sdma(MX35_SDMA_BASE_ADDR, MX35_INT_SDMA, &imx35_sdma_pdata);
 }

diff --git a/arch/arm/mach-mx5/mm.c b/arch/arm/mach-mx5/mm.c
index 800bb8b2..5af490a 100644
--- a/arch/arm/mach-mx5/mm.c
+++ b/arch/arm/mach-mx5/mm.c

@@ -18,6 +18,7 @@
 
 #include <mach/hardware.h>
 #include <mach/common.h>
+#include <mach/devices-common.h>
 #include <mach/iomux-v3.h>
 
 /*

@@ -100,12 +101,34 @@ void __init mx53_init_irq(void)
 	tzic_init_irq(tzic_virt);
 }
 
+static struct sdma_script_start_addrs imx51_sdma_script __initdata = {
+	.ap_2_ap_addr = 642,
+	.uart_2_mcu_addr = 817,
+	.mcu_2_app_addr = 747,
+	.mcu_2_shp_addr = 961,
+	.ata_2_mcu_addr = 1473,
+	.mcu_2_ata_addr = 1392,
+	.app_2_per_addr = 1033,
+	.app_2_mcu_addr = 683,
+	.shp_2_per_addr = 1251,
+	.shp_2_mcu_addr = 892,
+};
+
+static struct sdma_platform_data imx51_sdma_pdata __initdata = {
+	.sdma_version = 2,
+	.cpu_name = "imx51",
+	.to_version = 1,
+	.script_addrs = &imx51_sdma_script,
+};
+
 void __init imx51_soc_init(void)
 {
 	mxc_register_gpio(0, MX51_GPIO1_BASE_ADDR, SZ_16K, MX51_MXC_INT_GPIO1_LOW, MX51_MXC_INT_GPIO1_HIGH);
 	mxc_register_gpio(1, MX51_GPIO2_BASE_ADDR, SZ_16K, MX51_MXC_INT_GPIO2_LOW, MX51_MXC_INT_GPIO2_HIGH);
 	mxc_register_gpio(2, MX51_GPIO3_BASE_ADDR, SZ_16K, MX51_MXC_INT_GPIO3_LOW, MX51_MXC_INT_GPIO3_HIGH);
 	mxc_register_gpio(3, MX51_GPIO4_BASE_ADDR, SZ_16K, MX51_MXC_INT_GPIO4_LOW, MX51_MXC_INT_GPIO4_HIGH);
+
+	imx_add_imx_sdma(MX51_SDMA_BASE_ADDR, MX51_INT_SDMA, &imx51_sdma_pdata);
 }
 
 void __init imx53_soc_init(void)

diff --git a/arch/arm/plat-mxc/devices.c b/arch/arm/plat-mxc/devices.c
index fb166b2..0d6ed31 100644
--- a/arch/arm/plat-mxc/devices.c
+++ b/arch/arm/plat-mxc/devices.c

@@ -95,8 +95,22 @@ struct device mxc_aips_bus = {
 	.parent		= &platform_bus,
 };
 
+struct device mxc_ahb_bus = {
+	.init_name	= "mxc_ahb",
+	.parent		= &platform_bus,
+};
+
 static int __init mxc_device_init(void)
 {
-	return device_register(&mxc_aips_bus);
+	int ret;
+
+	ret = device_register(&mxc_aips_bus);
+	if (IS_ERR_VALUE(ret))
+		goto done;
+
+	ret = device_register(&mxc_ahb_bus);
+
+done:
+	return ret;
 }
 core_initcall(mxc_device_init);

diff --git a/arch/arm/plat-mxc/devices/platform-imx-dma.c b/arch/arm/plat-mxc/devices/platform-imx-dma.c
index 2091540..2b0fdb2 100644
--- a/arch/arm/plat-mxc/devices/platform-imx-dma.c
+++ b/arch/arm/plat-mxc/devices/platform-imx-dma.c

@@ -6,205 +6,29 @@
  * the terms of the GNU General Public License version 2 as published by the
  * Free Software Foundation.
  */
-#include <linux/compiler.h>
-#include <linux/err.h>
-#include <linux/init.h>
-
-#include <mach/hardware.h>
 #include <mach/devices-common.h>
-#include <mach/sdma.h>
-
-struct imx_imx_sdma_data {
-	resource_size_t iobase;
-	resource_size_t irq;
-	struct sdma_platform_data pdata;
-};
-
-#define imx_imx_sdma_data_entry_single(soc, _sdma_version, _cpu_name, _to_version)\
-	{								\
-		.iobase = soc ## _SDMA ## _BASE_ADDR,			\
-		.irq = soc ## _INT_SDMA,				\
-		.pdata = {						\
-			.sdma_version = _sdma_version,			\
-			.cpu_name = _cpu_name,				\
-			.to_version = _to_version,			\
-		},							\
-	}
-
-#ifdef CONFIG_SOC_IMX25
-struct imx_imx_sdma_data imx25_imx_sdma_data __initconst =
-	imx_imx_sdma_data_entry_single(MX25, 2, "imx25", 1);
-#endif /* ifdef CONFIG_SOC_IMX25 */
 
-#ifdef CONFIG_SOC_IMX31
-struct imx_imx_sdma_data imx31_imx_sdma_data __initdata =
-	imx_imx_sdma_data_entry_single(MX31, 1, "imx31", 1);
-#endif /* ifdef CONFIG_SOC_IMX31 */
-
-#ifdef CONFIG_SOC_IMX35
-struct imx_imx_sdma_data imx35_imx_sdma_data __initdata =
-	imx_imx_sdma_data_entry_single(MX35, 2, "imx35", 1);
-#endif /* ifdef CONFIG_SOC_IMX35 */
-
-#ifdef CONFIG_SOC_IMX51
-struct imx_imx_sdma_data imx51_imx_sdma_data __initconst =
-	imx_imx_sdma_data_entry_single(MX51, 2, "imx51", 1);
-#endif /* ifdef CONFIG_SOC_IMX51 */
+struct platform_device __init __maybe_unused *imx_add_imx_dma(void)
+{
+	return platform_device_register_resndata(&mxc_ahb_bus,
+			"imx-dma", -1, NULL, 0, NULL, 0);
+}
 
-static struct platform_device __init __maybe_unused *imx_add_imx_sdma(
-		const struct imx_imx_sdma_data *data)
+struct platform_device __init __maybe_unused *imx_add_imx_sdma(
+	resource_size_t iobase, int irq, struct sdma_platform_data *pdata)
 {
 	struct resource res[] = {
 		{
-			.start = data->iobase,
-			.end = data->iobase + SZ_16K - 1,
+			.start = iobase,
+			.end = iobase + SZ_16K - 1,
 			.flags = IORESOURCE_MEM,
 		}, {
-			.start = data->irq,
-			.end = data->irq,
+			.start = irq,
+			.end = irq,
 			.flags = IORESOURCE_IRQ,
 		},
 	};
 
-	return imx_add_platform_device("imx-sdma", -1,
-			res, ARRAY_SIZE(res),
-			&data->pdata, sizeof(data->pdata));
-}
-
-static struct platform_device __init __maybe_unused *imx_add_imx_dma(void)
-{
-	return imx_add_platform_device("imx-dma", -1, NULL, 0, NULL, 0);
-}
-
-#ifdef CONFIG_ARCH_MX25
-static struct sdma_script_start_addrs addr_imx25 = {
-	.ap_2_ap_addr = 729,
-	.uart_2_mcu_addr = 904,
-	.per_2_app_addr = 1255,
-	.mcu_2_app_addr = 834,
-	.uartsh_2_mcu_addr = 1120,
-	.per_2_shp_addr = 1329,
-	.mcu_2_shp_addr = 1048,
-	.ata_2_mcu_addr = 1560,
-	.mcu_2_ata_addr = 1479,
-	.app_2_per_addr = 1189,
-	.app_2_mcu_addr = 770,
-	.shp_2_per_addr = 1407,
-	.shp_2_mcu_addr = 979,
-};
-#endif
-
-#ifdef CONFIG_SOC_IMX31
-static struct sdma_script_start_addrs addr_imx31_to1 = {
-	.per_2_per_addr = 1677,
-};
-
-static struct sdma_script_start_addrs addr_imx31_to2 = {
-	.ap_2_ap_addr = 423,
-	.ap_2_bp_addr = 829,
-	.bp_2_ap_addr = 1029,
-};
-#endif
-
-#ifdef CONFIG_SOC_IMX35
-static struct sdma_script_start_addrs addr_imx35_to1 = {
-	.ap_2_ap_addr = 642,
-	.uart_2_mcu_addr = 817,
-	.mcu_2_app_addr = 747,
-	.uartsh_2_mcu_addr = 1183,
-	.per_2_shp_addr = 1033,
-	.mcu_2_shp_addr = 961,
-	.ata_2_mcu_addr = 1333,
-	.mcu_2_ata_addr = 1252,
-	.app_2_mcu_addr = 683,
-	.shp_2_per_addr = 1111,
-	.shp_2_mcu_addr = 892,
-};
-
-static struct sdma_script_start_addrs addr_imx35_to2 = {
-	.ap_2_ap_addr = 729,
-	.uart_2_mcu_addr = 904,
-	.per_2_app_addr = 1597,
-	.mcu_2_app_addr = 834,
-	.uartsh_2_mcu_addr = 1270,
-	.per_2_shp_addr = 1120,
-	.mcu_2_shp_addr = 1048,
-	.ata_2_mcu_addr = 1429,
-	.mcu_2_ata_addr = 1339,
-	.app_2_per_addr = 1531,
-	.app_2_mcu_addr = 770,
-	.shp_2_per_addr = 1198,
-	.shp_2_mcu_addr = 979,
-};
-#endif
-
-#ifdef CONFIG_SOC_IMX51
-static struct sdma_script_start_addrs addr_imx51 = {
-	.ap_2_ap_addr = 642,
-	.uart_2_mcu_addr = 817,
-	.mcu_2_app_addr = 747,
-	.mcu_2_shp_addr = 961,
-	.ata_2_mcu_addr = 1473,
-	.mcu_2_ata_addr = 1392,
-	.app_2_per_addr = 1033,
-	.app_2_mcu_addr = 683,
-	.shp_2_per_addr = 1251,
-	.shp_2_mcu_addr = 892,
-};
-#endif
-
-static int __init imxXX_add_imx_dma(void)
-{
-	struct platform_device *ret;
-
-#if defined(CONFIG_SOC_IMX21) || defined(CONFIG_SOC_IMX27)
-	if (cpu_is_mx21() || cpu_is_mx27())
-		ret = imx_add_imx_dma();
-	else
-#endif
-
-#if defined(CONFIG_SOC_IMX25)
-	if (cpu_is_mx25()) {
-		imx25_imx_sdma_data.pdata.script_addrs = &addr_imx25;
-		ret = imx_add_imx_sdma(&imx25_imx_sdma_data);
-	} else
-#endif
-
-#if defined(CONFIG_SOC_IMX31)
-	if (cpu_is_mx31()) {
-		int to_version = mx31_revision() >> 4;
-		imx31_imx_sdma_data.pdata.to_version = to_version;
-		if (to_version == 1)
-			imx31_imx_sdma_data.pdata.script_addrs = &addr_imx31_to1;
-		else
-			imx31_imx_sdma_data.pdata.script_addrs = &addr_imx31_to2;
-		ret = imx_add_imx_sdma(&imx31_imx_sdma_data);
-	} else
-#endif
-
-#if defined(CONFIG_SOC_IMX35)
-	if (cpu_is_mx35()) {
-		int to_version = mx35_revision() >> 4;
-		imx35_imx_sdma_data.pdata.to_version = to_version;
-		if (to_version == 1)
-			imx35_imx_sdma_data.pdata.script_addrs = &addr_imx35_to1;
-		else
-			imx35_imx_sdma_data.pdata.script_addrs = &addr_imx35_to2;
-		ret = imx_add_imx_sdma(&imx35_imx_sdma_data);
-	} else
-#endif
-
-#if defined(CONFIG_SOC_IMX51)
-	if (cpu_is_mx51()) {
-		imx51_imx_sdma_data.pdata.script_addrs = &addr_imx51;
-		ret = imx_add_imx_sdma(&imx51_imx_sdma_data);
-	} else
-#endif
-		ret = ERR_PTR(-ENODEV);
-
-	if (IS_ERR(ret))
-		return PTR_ERR(ret);
-
-	return 0;
+	return platform_device_register_resndata(&mxc_ahb_bus, "imx-sdma",
+			-1, res, ARRAY_SIZE(res), pdata, sizeof(*pdata));
 }
-arch_initcall(imxXX_add_imx_dma);

diff --git a/arch/arm/plat-mxc/include/mach/devices-common.h b/arch/arm/plat-mxc/include/mach/devices-common.h
index 03f6266..bf93820 100644
--- a/arch/arm/plat-mxc/include/mach/devices-common.h
+++ b/arch/arm/plat-mxc/include/mach/devices-common.h

@@ -9,8 +9,10 @@
 #include <linux/kernel.h>
 #include <linux/platform_device.h>
 #include <linux/init.h>
+#include <mach/sdma.h>
 
 extern struct device mxc_aips_bus;
+extern struct device mxc_ahb_bus;
 
 struct platform_device *imx_add_platform_device_dmamask(
 		const char *name, int id,

@@ -293,3 +295,7 @@ struct imx_spi_imx_data {
 struct platform_device *__init imx_add_spi_imx(
 		const struct imx_spi_imx_data *data,
 		const struct spi_imx_master *pdata);
+
+struct platform_device *imx_add_imx_dma(void);
+struct platform_device *imx_add_imx_sdma(
+	resource_size_t iobase, int irq, struct sdma_platform_data *pdata);

-- 
1.7.4.1

Hi Shawn,

On Mon, Jun 13, 2011 at 4:37 AM, Shawn Guo [off-list ref] wrote:

Though there are three TOs of imx51 soc, the sdma script never
changes since TO1, which means all three TOs of imx51 uses TO1
version of sdma script.

The current code passes TO number to imx-sdma driver to load
different firmware for different TO. ?That means we have to prepare
3 identical firmwares, sdma-imx25-to1.bin sdma-imx25-to2.bin and
sdma-imx25-to3.bin,

I think you want to mean:
sdma-imx51-to1.bin, sdma-imx51-to2.bin, sdma-imx51-to3.bin

Please fix the commit log.

Regards,

Fabio Estevam

On Mon, Jun 13, 2011 at 09:36:45AM -0300, Fabio Estevam wrote:

Hi Shawn,

On Mon, Jun 13, 2011 at 4:37 AM, Shawn Guo [off-list ref] wrote:

quoted

Though there are three TOs of imx51 soc, the sdma script never
changes since TO1, which means all three TOs of imx51 uses TO1
version of sdma script.

The current code passes TO number to imx-sdma driver to load
different firmware for different TO. ?That means we have to prepare
3 identical firmwares, sdma-imx25-to1.bin sdma-imx25-to2.bin and
sdma-imx25-to3.bin,

I think you want to mean:
sdma-imx51-to1.bin, sdma-imx51-to2.bin, sdma-imx51-to3.bin

Oops, right.

Please fix the commit log.

OK.

-- 
Regards,
Shawn

Though there are three TOs of imx51 soc, the sdma script never
changes since TO1, which means all three TOs of imx51 uses TO1
version of sdma script.

The current code passes TO number to imx-sdma driver to load
different firmware for different TO.  That means we have to prepare
3 identical firmwares, sdma-imx51-to1.bin sdma-imx51-to2.bin and
sdma-imx51-to3.bin, to have the kernel capable of running on all
three TOs.  This just makes no sense.

The patch removes the TO number passing and get the default TO1
version of sdma firmware work for all TOs.

Signed-off-by: Shawn Guo <redacted>
---
Changes since v1:
 * fix typo in imx51 firmware name caught by Fabio Estevam

 arch/arm/plat-mxc/devices/platform-imx-dma.c |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/arch/arm/plat-mxc/devices/platform-imx-dma.c b/arch/arm/plat-mxc/devices/platform-imx-dma.c
index c64f015..2091540 100644
--- a/arch/arm/plat-mxc/devices/platform-imx-dma.c
+++ b/arch/arm/plat-mxc/devices/platform-imx-dma.c

@@ -196,8 +196,6 @@ static int __init imxXX_add_imx_dma(void)
 
 #if defined(CONFIG_SOC_IMX51)
 	if (cpu_is_mx51()) {
-		int to_version = mx51_revision() >> 4;
-		imx51_imx_sdma_data.pdata.to_version = to_version;
 		imx51_imx_sdma_data.pdata.script_addrs = &addr_imx51;
 		ret = imx_add_imx_sdma(&imx51_imx_sdma_data);
 	} else

-- 
1.7.4.1

On Tue, Jun 14, 2011 at 02:20:07PM +0800, Shawn Guo wrote:

Though there are three TOs of imx51 soc, the sdma script never
changes since TO1, which means all three TOs of imx51 uses TO1
version of sdma script.

The current code passes TO number to imx-sdma driver to load
different firmware for different TO.  That means we have to prepare
3 identical firmwares, sdma-imx51-to1.bin sdma-imx51-to2.bin and
sdma-imx51-to3.bin, to have the kernel capable of running on all
three TOs.  This just makes no sense.

The patch removes the TO number passing and get the default TO1
version of sdma firmware work for all TOs.

I don't agree to this approach. The SDMA firmware has been different on
all TO versions of earlier i.MXs. For the linearity of the universe we
should keep this. What about providing a link in userspace?

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

On Tue, Jun 14, 2011 at 10:38:09PM +0200, Sascha Hauer wrote:

On Tue, Jun 14, 2011 at 02:20:07PM +0800, Shawn Guo wrote:

quoted

Though there are three TOs of imx51 soc, the sdma script never
changes since TO1, which means all three TOs of imx51 uses TO1
version of sdma script.

The current code passes TO number to imx-sdma driver to load
different firmware for different TO.  That means we have to prepare
3 identical firmwares, sdma-imx51-to1.bin sdma-imx51-to2.bin and
sdma-imx51-to3.bin, to have the kernel capable of running on all
three TOs.  This just makes no sense.

The patch removes the TO number passing and get the default TO1
version of sdma firmware work for all TOs.

I don't agree to this approach. The SDMA firmware has been different on
all TO versions of earlier i.MXs. For the linearity of the universe we
should keep this. What about providing a link in userspace?

The fact is that mx31 and mx35 are the only two SoCs (relatively old
ones) that have different SDMA firmwares on different TOs.  The
relatively new ones, mx25 (TO1, TO1.1), mx50 (TO1, TO1.1), mx51
(TO1, TO2, TO3) and mx53 (TO1, TO2) all have one version SDMA
firmware for different TOs.  I do not see the mx31/35 case is a
linearity of the universe.  Do we really want to bother ourselves
with userspace link on so many latest SoCs?

-- 
Regards,
Shawn

On Wed, Jun 15, 2011 at 12:29:49PM +0800, Shawn Guo wrote:

On Tue, Jun 14, 2011 at 10:38:09PM +0200, Sascha Hauer wrote:

quoted

On Tue, Jun 14, 2011 at 02:20:07PM +0800, Shawn Guo wrote:

quoted

Though there are three TOs of imx51 soc, the sdma script never
changes since TO1, which means all three TOs of imx51 uses TO1
version of sdma script.

The current code passes TO number to imx-sdma driver to load
different firmware for different TO.  That means we have to prepare
3 identical firmwares, sdma-imx51-to1.bin sdma-imx51-to2.bin and
sdma-imx51-to3.bin, to have the kernel capable of running on all
three TOs.  This just makes no sense.

The patch removes the TO number passing and get the default TO1
version of sdma firmware work for all TOs.

I don't agree to this approach. The SDMA firmware has been different on
all TO versions of earlier i.MXs. For the linearity of the universe we
should keep this. What about providing a link in userspace?

The fact is that mx31 and mx35 are the only two SoCs (relatively old
ones) that have different SDMA firmwares on different TOs.  The
relatively new ones, mx25 (TO1, TO1.1), mx50 (TO1, TO1.1), mx51
(TO1, TO2, TO3) and mx53 (TO1, TO2) all have one version SDMA
firmware for different TOs.  I do not see the mx31/35 case is a
linearity of the universe.  Do we really want to bother ourselves
with userspace link on so many latest SoCs?

Ok, then let's do it differently. Let's go away from "sdma-%s-to%d.bin"
in the sdma driver and pass the string from platform data instead.
Then we can pass the TO version where it's relevant and skip it where
only one firmware exists. Using a sdma-imx51-to1.bin for to2 and to3
is confusing.

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

On Thu, Jun 16, 2011 at 09:30:18AM +0200, Sascha Hauer wrote:

On Wed, Jun 15, 2011 at 12:29:49PM +0800, Shawn Guo wrote:

quoted

On Tue, Jun 14, 2011 at 10:38:09PM +0200, Sascha Hauer wrote:

quoted

On Tue, Jun 14, 2011 at 02:20:07PM +0800, Shawn Guo wrote:

quoted

Though there are three TOs of imx51 soc, the sdma script never
changes since TO1, which means all three TOs of imx51 uses TO1
version of sdma script.

The current code passes TO number to imx-sdma driver to load
different firmware for different TO.  That means we have to prepare
3 identical firmwares, sdma-imx51-to1.bin sdma-imx51-to2.bin and
sdma-imx51-to3.bin, to have the kernel capable of running on all
three TOs.  This just makes no sense.

The patch removes the TO number passing and get the default TO1
version of sdma firmware work for all TOs.

I don't agree to this approach. The SDMA firmware has been different on
all TO versions of earlier i.MXs. For the linearity of the universe we
should keep this. What about providing a link in userspace?

The fact is that mx31 and mx35 are the only two SoCs (relatively old
ones) that have different SDMA firmwares on different TOs.  The
relatively new ones, mx25 (TO1, TO1.1), mx50 (TO1, TO1.1), mx51
(TO1, TO2, TO3) and mx53 (TO1, TO2) all have one version SDMA
firmware for different TOs.  I do not see the mx31/35 case is a
linearity of the universe.  Do we really want to bother ourselves
with userspace link on so many latest SoCs?

Ok, then let's do it differently. Let's go away from "sdma-%s-to%d.bin"
in the sdma driver and pass the string from platform data instead.
Then we can pass the TO version where it's relevant and skip it where
only one firmware exists. Using a sdma-imx51-to1.bin for to2 and to3

That means the sdma firmware dose not change since to1, confusing?
But I'm fine with your idea, so okay, drop the patch from the series,
and let's address this issue with another patch.

Other patches look good to be applied?

is confusing.

-- 
Regards,
Shawn

On Thu, Jun 16, 2011 at 8:51 AM, Shawn Guo [off-list ref] wrote:
...

That means the sdma firmware dose not change since to1, confusing?
But I'm fine with your idea, so okay, drop the patch from the series,
and let's address this issue with another patch.

Other patches look good to be applied?

The other SDMA patches you sent are already in imx-for-2.6.40 now:
http://git.pengutronix.de/?p=imx/linux-2.6.git;a=shortlog;h=refs/heads/imx-for-2.6.40

I think it would be nice if Sascha could reply with a 'applied' email
so we know that the patches we sent were accepted.

Regards,

Fabio Estevam

On Thu, Jun 16, 2011 at 10:59:41PM -0300, Fabio Estevam wrote:

On Thu, Jun 16, 2011 at 8:51 AM, Shawn Guo [off-list ref] wrote:
...

quoted

That means the sdma firmware dose not change since to1, confusing?
But I'm fine with your idea, so okay, drop the patch from the series,
and let's address this issue with another patch.

Other patches look good to be applied?

The other SDMA patches you sent are already in imx-for-2.6.40 now:
http://git.pengutronix.de/?p=imx/linux-2.6.git;a=shortlog;h=refs/heads/imx-for-2.6.40

Ah, thanks.  I have been keeping my eyes on branch imx-for-3.0, so
really did not check imx-for-2.6.40.  But patch #6 which is the main
initiative of the whole patch set is still not applied yet.  I
suppose it will go through imx-for-3.0.

Speaking of the branch name imx-for-3.0, I'm a little confused here.
Shouldn't it be imx-for-3.1, if it targets the next merge window?
I'm wrong?

I think it would be nice if Sascha could reply with a 'applied' email
so we know that the patches we sent were accepted.

+1.  It would be even better with the branch info. 

-- 
Regards,
Shawn

On Fri, Jun 17, 2011 at 10:27:18AM +0800, Shawn Guo wrote:

On Thu, Jun 16, 2011 at 10:59:41PM -0300, Fabio Estevam wrote:

quoted

On Thu, Jun 16, 2011 at 8:51 AM, Shawn Guo [off-list ref] wrote:
...

quoted

That means the sdma firmware dose not change since to1, confusing?
But I'm fine with your idea, so okay, drop the patch from the series,
and let's address this issue with another patch.

Other patches look good to be applied?

The other SDMA patches you sent are already in imx-for-2.6.40 now:
http://git.pengutronix.de/?p=imx/linux-2.6.git;a=shortlog;h=refs/heads/imx-for-2.6.40

Ah, thanks.  I have been keeping my eyes on branch imx-for-3.0, so
really did not check imx-for-2.6.40.  But patch #6 which is the main
initiative of the whole patch set is still not applied yet.  I
suppose it will go through imx-for-3.0.

Speaking of the branch name imx-for-3.0, I'm a little confused here.
Shouldn't it be imx-for-3.1, if it targets the next merge window?
I'm wrong?

quoted

I think it would be nice if Sascha could reply with a 'applied' email
so we know that the patches we sent were accepted.

+1.  It would be even better with the branch info. 

+1. Wait, I am Sascha. Will do in the future.

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Hi Sascha,

Any comment on this one?

Regards,
Shawn

On Mon, Jun 13, 2011 at 03:37:25PM +0800, Shawn Guo wrote:

The patch follows the implementation of gpio-mxc device registration
to break the concentrated imx-dma device registration into soc
specific setup function.  Then we can avoid the churn of "#ifdef"
and the cpu_is_mx checking on such a long list, which makes no sense,
considering more soc supports need to be added and we need to support
single image for multiple socs in the long run.

Signed-off-by: Shawn Guo <redacted>
---
 arch/arm/mach-imx/mm-imx21.c                    |    3 +
 arch/arm/mach-imx/mm-imx25.c                    |   26 +++
 arch/arm/mach-imx/mm-imx27.c                    |    3 +
 arch/arm/mach-imx/mm-imx31.c                    |   24 +++
 arch/arm/mach-imx/mm-imx35.c                    |   44 +++++
 arch/arm/mach-mx5/mm.c                          |   23 +++
 arch/arm/plat-mxc/devices.c                     |   16 ++-
 arch/arm/plat-mxc/devices/platform-imx-dma.c    |  202 ++---------------------
 arch/arm/plat-mxc/include/mach/devices-common.h |    6 +
 9 files changed, 157 insertions(+), 190 deletions(-)

On Mon, Jun 20, 2011 at 10:50:55PM +0800, Shawn Guo wrote:

Hi Sascha,

Any comment on this one?

One more ping: any comment?

The patch has been out there for a while with no comments.  My
understanding is you will need to either give comments or apply
the patch, no?

Regards,
Shawn

On Mon, Jun 13, 2011 at 03:37:25PM +0800, Shawn Guo wrote:

quoted

The patch follows the implementation of gpio-mxc device registration
to break the concentrated imx-dma device registration into soc
specific setup function.  Then we can avoid the churn of "#ifdef"
and the cpu_is_mx checking on such a long list, which makes no sense,
considering more soc supports need to be added and we need to support
single image for multiple socs in the long run.

Signed-off-by: Shawn Guo <redacted>
---
 arch/arm/mach-imx/mm-imx21.c                    |    3 +
 arch/arm/mach-imx/mm-imx25.c                    |   26 +++
 arch/arm/mach-imx/mm-imx27.c                    |    3 +
 arch/arm/mach-imx/mm-imx31.c                    |   24 +++
 arch/arm/mach-imx/mm-imx35.c                    |   44 +++++
 arch/arm/mach-mx5/mm.c                          |   23 +++
 arch/arm/plat-mxc/devices.c                     |   16 ++-
 arch/arm/plat-mxc/devices/platform-imx-dma.c    |  202 ++---------------------
 arch/arm/plat-mxc/include/mach/devices-common.h |    6 +
 9 files changed, 157 insertions(+), 190 deletions(-)

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

-- 
Regards,
Shawn

The patch has been out there for a while with no comments.  My
understanding is you will need to either give comments or apply
the patch, no?

Yes, but there is no timeframe guaranteed. We try our best to respond as fast
as possible, but we (like most others) have a couple of things on our
todo-list.

Regards,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20110621/078457e4/attachment.sig>

On Mon, Jun 20, 2011 at 10:50:55PM +0800, Shawn Guo wrote:

Hi Sascha,

Any comment on this one?

Please respin this series on the for-next branch. Fabio has
added i.MX53 support in the meantime and this does not apply anymore.

Sascha
 

Regards,
Shawn

On Mon, Jun 13, 2011 at 03:37:25PM +0800, Shawn Guo wrote:

quoted

The patch follows the implementation of gpio-mxc device registration
to break the concentrated imx-dma device registration into soc
specific setup function.  Then we can avoid the churn of "#ifdef"
and the cpu_is_mx checking on such a long list, which makes no sense,
considering more soc supports need to be added and we need to support
single image for multiple socs in the long run.

Signed-off-by: Shawn Guo <redacted>
---
 arch/arm/mach-imx/mm-imx21.c                    |    3 +
 arch/arm/mach-imx/mm-imx25.c                    |   26 +++
 arch/arm/mach-imx/mm-imx27.c                    |    3 +
 arch/arm/mach-imx/mm-imx31.c                    |   24 +++
 arch/arm/mach-imx/mm-imx35.c                    |   44 +++++
 arch/arm/mach-mx5/mm.c                          |   23 +++
 arch/arm/plat-mxc/devices.c                     |   16 ++-
 arch/arm/plat-mxc/devices/platform-imx-dma.c    |  202 ++---------------------
 arch/arm/plat-mxc/include/mach/devices-common.h |    6 +
 9 files changed, 157 insertions(+), 190 deletions(-)

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

On Tue, Jun 21, 2011 at 05:13:26PM +0200, Wolfram Sang wrote:

quoted

The patch has been out there for a while with no comments.  My
understanding is you will need to either give comments or apply
the patch, no?

Yes, but there is no timeframe guaranteed. We try our best to respond as fast
as possible, but we (like most others) have a couple of things on our
todo-list.

Totally understandable.  I could have hold this second ping for
another few days.  But I saw Sascha picked up Fabio's mx53 SDMA
patch which grows the stuff I'm trying to clean up.  It actually
is Fabio's patch reminding me to send the patch clean this stuff
up.

So in this particular case, it's not about time frame but picking
up the right patch in the right order.

-- 
Regards,
Shawn

On Wed, Jun 22, 2011 at 12:57:58AM +0200, Sascha Hauer wrote:

On Mon, Jun 20, 2011 at 10:50:55PM +0800, Shawn Guo wrote:

quoted

Hi Sascha,

Any comment on this one?

Please respin this series on the for-next branch. Fabio has
added i.MX53 support in the meantime and this does not apply anymore.

This is exactly the thing I feel strange about.  I sent the patch
trying to stop people adding more churn there.  However, you applied
the patch adding more churn yesterday, and are asking me to rebase
this cleanup patch set today.  IMO, the cleanup patch should be
applied first, and the i.MX53 patch should be asked to respin.

But no problem, I will do.

-- 
Regards,
Shawn

So in this particular case, it's not about time frame but picking
up the right patch in the right order.

Ah, okay. That's a perfectly valid reason, of course. Would be nice to state
the reason in the "ping"-mail next time :)

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20110622/023d23e9/attachment-0001.sig>