Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
"Hemanth V" writes: >> >> A Santosh suggested, you need to break this up into parts that are >> OMAP specific (arch/arm/*) and parts that go via the SPI subsystem >> (drivers/spi/*.) If doing this breaks your series, then fix it > > But then that would cause a problem, we will not be able to > use arch/arm/plat-omap/include/mach/mcspi.h and will need to redefine those > variables in omap2_mcspi.c which obviously is not the preferred way. Anything shared between the driver and the arch code should go into , this way the arch code always compiles. For the driver code, when it's submitted upstream, you state that it has a dependency on the arch code if it is not yet upstream. The point is to separate out the arch code and get it submitted/merged asap so driver code has a smooth path. Kevin -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
- Original Message - From: "Kevin Hilman" To: "Hemanth V" Cc: ; "Tony Lindgren" ; "David Brownell" Sent: Thursday, June 11, 2009 8:17 PM Subject: Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support "Hemanth V" writes: - Original Message - From: "Hemanth V" To: "Tony Lindgren" Cc: Sent: Friday, June 05, 2009 3:28 PM Subject: Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support - Original Message - From: "Tony Lindgren" To: "Hemanth V" Cc: Sent: Tuesday, June 02, 2009 11:36 PM Subject: Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support Hi, Sorry for the delay in replying, few comments below. * Hemanth V [090519 22:57]: This patch adds support for McSPI slave and FIFO. DMA and FIFO could be enabled together for better throughput. Platform config parameters have been added to enable these features on any particular McSPI controller. FIFO can be enabled by defining fifo_depth parameter. fifo_depth needs to be a multiple of buffer size that is used for read/write. These features are useful when you have high throughput devices like WLAN or Modem connected over SPI. Signed-off-by: Hemanth V arch/arm/mach-omap2/devices.c |5 arch/arm/plat-omap/include/mach/mcspi.h | 16 + drivers/spi/omap2_mcspi.c | 343 3 files changed, 325 insertions(+), 39 deletions(-) As this is mostly drivers/spi/omap2_mcspi.c, this patch should get merged via: $ grep -A7 "SPI SUBSYSTEM" MAINTAINERS SPI SUBSYSTEM P: David Brownell M: dbrown...@users.sourceforge.net L: spi-devel-gene...@lists.sourceforge.net S: Maintained F: Documentation/spi/ F: drivers/spi/ F: include/linux/spi/ Please keep linux-omap list Cc'd too so everybody can follow the progress. Tony, is this list active. The archives seem to be flooded with spam mails http://sourceforge.net/mailarchive/forum.php?forum_name=spi-devel-general Kevin, Can u suggest what is to be done in this situation. spi-devel list doesnot seem to be active and Tony is not willing to merge this patch. Should I send this to LKML First, you haven't addressed any of the comments made on the list about your series. Tony isn't merging this patch because most of it should go via the SPI subsystem. A Santosh suggested, you need to break this up into parts that are OMAP specific (arch/arm/*) and parts that go via the SPI subsystem (drivers/spi/*.) If doing this breaks your series, then fix it But then that would cause a problem, we will not be able to use arch/arm/plat-omap/include/mach/mcspi.h and will need to redefine those variables in omap2_mcspi.c which obviously is not the preferred way. because if it breaks compile this way, then upstream maintainers will surely hit the same errors and complain. David Brownell is the SPI maintainer. I suggest you send to him, CC'ing linux-omap. Kevin -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
"Hemanth V" writes: > - Original Message - > From: "Hemanth V" > To: "Tony Lindgren" > Cc: > Sent: Friday, June 05, 2009 3:28 PM > Subject: Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support > > >> - Original Message - >> From: "Tony Lindgren" >> To: "Hemanth V" >> Cc: >> Sent: Tuesday, June 02, 2009 11:36 PM >> Subject: Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support >> >> >>> Hi, >>> >>> Sorry for the delay in replying, few comments below. >>> >>> * Hemanth V [090519 22:57]: >>>> This patch adds support for McSPI slave and FIFO. DMA and FIFO >>>> could be enabled together for better throughput. Platform config >>>> parameters have been added to enable these features on any particular >>>> McSPI controller. >>>> >>>> FIFO can be enabled by defining fifo_depth parameter. fifo_depth needs >>>> to be a multiple of buffer size that is used for read/write. >>>> >>>> These features are useful when you have high throughput devices >>>> like WLAN or Modem connected over SPI. >>>> >>>> Signed-off-by: Hemanth V >>>> arch/arm/mach-omap2/devices.c |5 >>>> arch/arm/plat-omap/include/mach/mcspi.h | 16 + >>>> drivers/spi/omap2_mcspi.c | 343 >>>> >>>> 3 files changed, 325 insertions(+), 39 deletions(-) >>> >>> As this is mostly drivers/spi/omap2_mcspi.c, this patch should get >>> merged via: >>> >>> $ grep -A7 "SPI SUBSYSTEM" MAINTAINERS >>> SPI SUBSYSTEM >>> P: David Brownell >>> M: dbrown...@users.sourceforge.net >>> L: spi-devel-gene...@lists.sourceforge.net >>> S: Maintained >>> F: Documentation/spi/ >>> F: drivers/spi/ >>> F: include/linux/spi/ >>> >>> Please keep linux-omap list Cc'd too so everybody can follow >>> the progress. >> >> Tony, is this list active. The archives seem to be flooded with spam >> mails >> >> http://sourceforge.net/mailarchive/forum.php?forum_name=spi-devel-general >> >>> > > Kevin, Can u suggest what is to be done in this situation. spi-devel > list doesnot seem > to be active and Tony is not willing to merge this patch. Should I > send this to LKML First, you haven't addressed any of the comments made on the list about your series. Tony isn't merging this patch because most of it should go via the SPI subsystem. A Santosh suggested, you need to break this up into parts that are OMAP specific (arch/arm/*) and parts that go via the SPI subsystem (drivers/spi/*.) If doing this breaks your series, then fix it because if it breaks compile this way, then upstream maintainers will surely hit the same errors and complain. David Brownell is the SPI maintainer. I suggest you send to him, CC'ing linux-omap. Kevin -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
- Original Message -
From: "Hemanth V"
To: "Tony Lindgren"
Cc:
Sent: Friday, June 05, 2009 3:28 PM
Subject: Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
- Original Message -
From: "Tony Lindgren"
To: "Hemanth V"
Cc:
Sent: Tuesday, June 02, 2009 11:36 PM
Subject: Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
Hi,
Sorry for the delay in replying, few comments below.
* Hemanth V [090519 22:57]:
This patch adds support for McSPI slave and FIFO. DMA and FIFO
could be enabled together for better throughput. Platform config
parameters have been added to enable these features on any particular
McSPI controller.
FIFO can be enabled by defining fifo_depth parameter. fifo_depth needs
to be a multiple of buffer size that is used for read/write.
These features are useful when you have high throughput devices
like WLAN or Modem connected over SPI.
Signed-off-by: Hemanth V
arch/arm/mach-omap2/devices.c |5
arch/arm/plat-omap/include/mach/mcspi.h | 16 +
drivers/spi/omap2_mcspi.c | 343
3 files changed, 325 insertions(+), 39 deletions(-)
As this is mostly drivers/spi/omap2_mcspi.c, this patch should get
merged via:
$ grep -A7 "SPI SUBSYSTEM" MAINTAINERS
SPI SUBSYSTEM
P: David Brownell
M: dbrown...@users.sourceforge.net
L: spi-devel-gene...@lists.sourceforge.net
S: Maintained
F: Documentation/spi/
F: drivers/spi/
F: include/linux/spi/
Please keep linux-omap list Cc'd too so everybody can follow
the progress.
Tony, is this list active. The archives seem to be flooded with spam
mails
http://sourceforge.net/mailarchive/forum.php?forum_name=spi-devel-general
Kevin, Can u suggest what is to be done in this situation. spi-devel list
doesnot seem
to be active and Tony is not willing to merge this patch. Should I send this
to LKML
Thanks
Hemanth
---
Index: linux-omap-2.6/arch/arm/mach-omap2/devices.c
===
--- linux-omap-2.6.orig/arch/arm/mach-omap2/devices.c 2009-05-19
17:00:21.0 +0530
+++ linux-omap-2.6/arch/arm/mach-omap2/devices.c 2009-05-20
11:02:41.0
+0530
@@ -259,6 +259,7 @@
static struct omap2_mcspi_platform_config omap2_mcspi1_config = {
.num_cs = 4,
+ .force_cs_mode = 1,
};
static struct resource omap2_mcspi1_resources[] = {
@@ -281,6 +282,10 @@
static struct omap2_mcspi_platform_config omap2_mcspi2_config = {
.num_cs = 2,
+ .mode = OMAP2_MCSPI_MASTER,
+ .dma_mode = 1,
+ .force_cs_mode = 0,
+ .fifo_depth = 0,
};
static struct resource omap2_mcspi2_resources[] = {
Index: linux-omap-2.6/arch/arm/plat-omap/include/mach/mcspi.h
===
--- linux-omap-2.6.orig/arch/arm/plat-omap/include/mach/mcspi.h
2009-05-19
17:00:21.0 +0530
+++ linux-omap-2.6/arch/arm/plat-omap/include/mach/mcspi.h 2009-05-20
11:02:41.0 +0530
@@ -1,8 +1,24 @@
#ifndef _OMAP2_MCSPI_H
#define _OMAP2_MCSPI_H
+#define OMAP2_MCSPI_MASTER 0
+#define OMAP2_MCSPI_SLAVE 1
+
struct omap2_mcspi_platform_config {
unsigned short num_cs;
+
+ /* SPI is master or slave */
+ unsigned short mode;
+
+ /* Use only DMA for data transfers */
+ unsigned short dma_mode;
+
+ /* Force chip select mode */
+ unsigned short force_cs_mode;
+
+ /* FIFO depth in bytes, max value 64 */
+ unsigned short fifo_depth;
+
};
struct omap2_mcspi_device_config {
Index: linux-omap-2.6/drivers/spi/omap2_mcspi.c
===
--- linux-omap-2.6.orig/drivers/spi/omap2_mcspi.c 2009-05-19
17:00:21.0
+0530
+++ linux-omap-2.6/drivers/spi/omap2_mcspi.c 2009-05-20
11:02:41.0 +0530
@@ -37,9 +37,11 @@
#include
#include
+#include
#define OMAP2_MCSPI_MAX_FREQ 4800
+#define OMAP2_MCSPI_MAX_FIFODEPTH 64
#define OMAP2_MCSPI_REVISION 0x00
#define OMAP2_MCSPI_SYSCONFIG 0x10
@@ -49,6 +51,7 @@
#define OMAP2_MCSPI_WAKEUPENABLE 0x20
#define OMAP2_MCSPI_SYST 0x24
#define OMAP2_MCSPI_MODULCTRL 0x28
+#define OMAP2_MCSPI_XFERLEVEL 0x7c
/* per-channel banks, 0x14 bytes each, first is: */
#define OMAP2_MCSPI_CHCONF0 0x2c
@@ -85,6 +88,9 @@
#define OMAP2_MCSPI_CHCONF_IS BIT(18)
#define OMAP2_MCSPI_CHCONF_TURBO BIT(19)
#define OMAP2_MCSPI_CHCONF_FORCE BIT(20)
+#define OMAP2_MCSPI_CHCONF_FFER BIT(28)
+#define OMAP2_MCSPI_CHCONF_FFET BIT(27)
+
#define OMAP2_MCSPI_CHSTAT_RXS BIT(0)
#define OMAP2_MCSPI_CHSTAT_TXS BIT(1)
@@ -93,6 +99,7 @@
Please swap BIT(27) to be before BIT(28) to keep them sorted above.
#define OMAP2_MCSPI_CHCTRL_EN BIT(0)
#define OMAP2_MCSPI_WAKEUPENABLE_WKEN BIT(0)
+#define OMAP2_MCSPI_IRQ_EOW BIT(17)
/* We have 2 DMA channels per CS, one for RX and one for TX */
struct omap2_mcspi_dma {
@@ -125,6 +132,10 @@
unsigned long phys;
/* SPI1 has 4 channels, while SPI2 has 2 */
struct omap2_mcspi_dma *dma_
Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
- Original Message -
From: "Tony Lindgren"
To: "Hemanth V"
Cc:
Sent: Tuesday, June 02, 2009 11:36 PM
Subject: Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
Hi,
Sorry for the delay in replying, few comments below.
* Hemanth V [090519 22:57]:
This patch adds support for McSPI slave and FIFO. DMA and FIFO
could be enabled together for better throughput. Platform config
parameters have been added to enable these features on any particular
McSPI controller.
FIFO can be enabled by defining fifo_depth parameter. fifo_depth needs
to be a multiple of buffer size that is used for read/write.
These features are useful when you have high throughput devices
like WLAN or Modem connected over SPI.
Signed-off-by: Hemanth V
arch/arm/mach-omap2/devices.c |5
arch/arm/plat-omap/include/mach/mcspi.h | 16 +
drivers/spi/omap2_mcspi.c | 343
3 files changed, 325 insertions(+), 39 deletions(-)
As this is mostly drivers/spi/omap2_mcspi.c, this patch should get
merged via:
$ grep -A7 "SPI SUBSYSTEM" MAINTAINERS
SPI SUBSYSTEM
P: David Brownell
M: dbrown...@users.sourceforge.net
L: spi-devel-gene...@lists.sourceforge.net
S: Maintained
F: Documentation/spi/
F: drivers/spi/
F: include/linux/spi/
Please keep linux-omap list Cc'd too so everybody can follow
the progress.
Tony, is this list active. The archives seem to be flooded with spam
mails
http://sourceforge.net/mailarchive/forum.php?forum_name=spi-devel-general
---
Index: linux-omap-2.6/arch/arm/mach-omap2/devices.c
===
--- linux-omap-2.6.orig/arch/arm/mach-omap2/devices.c 2009-05-19
17:00:21.0 +0530
+++ linux-omap-2.6/arch/arm/mach-omap2/devices.c 2009-05-20
11:02:41.0
+0530
@@ -259,6 +259,7 @@
static struct omap2_mcspi_platform_config omap2_mcspi1_config = {
.num_cs = 4,
+ .force_cs_mode = 1,
};
static struct resource omap2_mcspi1_resources[] = {
@@ -281,6 +282,10 @@
static struct omap2_mcspi_platform_config omap2_mcspi2_config = {
.num_cs = 2,
+ .mode = OMAP2_MCSPI_MASTER,
+ .dma_mode = 1,
+ .force_cs_mode = 0,
+ .fifo_depth = 0,
};
static struct resource omap2_mcspi2_resources[] = {
Index: linux-omap-2.6/arch/arm/plat-omap/include/mach/mcspi.h
===
--- linux-omap-2.6.orig/arch/arm/plat-omap/include/mach/mcspi.h
2009-05-19
17:00:21.0 +0530
+++ linux-omap-2.6/arch/arm/plat-omap/include/mach/mcspi.h 2009-05-20
11:02:41.0 +0530
@@ -1,8 +1,24 @@
#ifndef _OMAP2_MCSPI_H
#define _OMAP2_MCSPI_H
+#define OMAP2_MCSPI_MASTER 0
+#define OMAP2_MCSPI_SLAVE 1
+
struct omap2_mcspi_platform_config {
unsigned short num_cs;
+
+ /* SPI is master or slave */
+ unsigned short mode;
+
+ /* Use only DMA for data transfers */
+ unsigned short dma_mode;
+
+ /* Force chip select mode */
+ unsigned short force_cs_mode;
+
+ /* FIFO depth in bytes, max value 64 */
+ unsigned short fifo_depth;
+
};
struct omap2_mcspi_device_config {
Index: linux-omap-2.6/drivers/spi/omap2_mcspi.c
===
--- linux-omap-2.6.orig/drivers/spi/omap2_mcspi.c 2009-05-19
17:00:21.0
+0530
+++ linux-omap-2.6/drivers/spi/omap2_mcspi.c 2009-05-20
11:02:41.0 +0530
@@ -37,9 +37,11 @@
#include
#include
+#include
#define OMAP2_MCSPI_MAX_FREQ 4800
+#define OMAP2_MCSPI_MAX_FIFODEPTH 64
#define OMAP2_MCSPI_REVISION 0x00
#define OMAP2_MCSPI_SYSCONFIG 0x10
@@ -49,6 +51,7 @@
#define OMAP2_MCSPI_WAKEUPENABLE 0x20
#define OMAP2_MCSPI_SYST 0x24
#define OMAP2_MCSPI_MODULCTRL 0x28
+#define OMAP2_MCSPI_XFERLEVEL 0x7c
/* per-channel banks, 0x14 bytes each, first is: */
#define OMAP2_MCSPI_CHCONF0 0x2c
@@ -85,6 +88,9 @@
#define OMAP2_MCSPI_CHCONF_IS BIT(18)
#define OMAP2_MCSPI_CHCONF_TURBO BIT(19)
#define OMAP2_MCSPI_CHCONF_FORCE BIT(20)
+#define OMAP2_MCSPI_CHCONF_FFER BIT(28)
+#define OMAP2_MCSPI_CHCONF_FFET BIT(27)
+
#define OMAP2_MCSPI_CHSTAT_RXS BIT(0)
#define OMAP2_MCSPI_CHSTAT_TXS BIT(1)
@@ -93,6 +99,7 @@
Please swap BIT(27) to be before BIT(28) to keep them sorted above.
#define OMAP2_MCSPI_CHCTRL_EN BIT(0)
#define OMAP2_MCSPI_WAKEUPENABLE_WKEN BIT(0)
+#define OMAP2_MCSPI_IRQ_EOW BIT(17)
/* We have 2 DMA channels per CS, one for RX and one for TX */
struct omap2_mcspi_dma {
@@ -125,6 +132,10 @@
unsigned long phys;
/* SPI1 has 4 channels, while SPI2 has 2 */
struct omap2_mcspi_dma *dma_channels;
+ unsigned short mcspi_mode;
+ unsigned short dma_mode;
+ unsigned short force_cs_mode;
+ unsigned short fifo_depth;
};
struct omap2_mcspi_cs {
@@ -133,6 +144,37 @@
int word_len;
};
+#ifdef CONFIG_SPI_DEBUG
+struct reg_type {
+ char name[40];
+ int offset;
+};
+
+static struct reg_type reg_map[] = {
+ {"MCSPI_REV", 0x0},
+ {"MCSPI_SYSCONFIG", 0x10},
Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
Hi,
Sorry for the delay in replying, few comments below.
* Hemanth V [090519 22:57]:
> This patch adds support for McSPI slave and FIFO. DMA and FIFO
> could be enabled together for better throughput. Platform config
> parameters have been added to enable these features on any particular
> McSPI controller.
>
> FIFO can be enabled by defining fifo_depth parameter. fifo_depth needs
> to be a multiple of buffer size that is used for read/write.
>
> These features are useful when you have high throughput devices
> like WLAN or Modem connected over SPI.
>
> Signed-off-by: Hemanth V
> arch/arm/mach-omap2/devices.c |5
> arch/arm/plat-omap/include/mach/mcspi.h | 16 +
> drivers/spi/omap2_mcspi.c | 343
>
> 3 files changed, 325 insertions(+), 39 deletions(-)
As this is mostly drivers/spi/omap2_mcspi.c, this patch should get
merged via:
$ grep -A7 "SPI SUBSYSTEM" MAINTAINERS
SPI SUBSYSTEM
P: David Brownell
M: dbrown...@users.sourceforge.net
L: spi-devel-gene...@lists.sourceforge.net
S: Maintained
F: Documentation/spi/
F: drivers/spi/
F: include/linux/spi/
Please keep linux-omap list Cc'd too so everybody can follow
the progress.
>
> ---
> Index: linux-omap-2.6/arch/arm/mach-omap2/devices.c
> ===
> --- linux-omap-2.6.orig/arch/arm/mach-omap2/devices.c 2009-05-19
> 17:00:21.0 +0530
> +++ linux-omap-2.6/arch/arm/mach-omap2/devices.c 2009-05-20
> 11:02:41.0
> +0530
> @@ -259,6 +259,7 @@
>
> static struct omap2_mcspi_platform_config omap2_mcspi1_config = {
> .num_cs = 4,
> + .force_cs_mode = 1,
> };
>
> static struct resource omap2_mcspi1_resources[] = {
> @@ -281,6 +282,10 @@
>
> static struct omap2_mcspi_platform_config omap2_mcspi2_config = {
> .num_cs = 2,
> + .mode = OMAP2_MCSPI_MASTER,
> + .dma_mode = 1,
> + .force_cs_mode = 0,
> + .fifo_depth = 0,
> };
>
> static struct resource omap2_mcspi2_resources[] = {
> Index: linux-omap-2.6/arch/arm/plat-omap/include/mach/mcspi.h
> ===
> --- linux-omap-2.6.orig/arch/arm/plat-omap/include/mach/mcspi.h
> 2009-05-19
> 17:00:21.0 +0530
> +++ linux-omap-2.6/arch/arm/plat-omap/include/mach/mcspi.h2009-05-20
> 11:02:41.0 +0530
> @@ -1,8 +1,24 @@
> #ifndef _OMAP2_MCSPI_H
> #define _OMAP2_MCSPI_H
>
> +#define OMAP2_MCSPI_MASTER 0
> +#define OMAP2_MCSPI_SLAVE1
> +
> struct omap2_mcspi_platform_config {
> unsigned short num_cs;
> +
> + /* SPI is master or slave */
> + unsigned short mode;
> +
> + /* Use only DMA for data transfers */
> + unsigned short dma_mode;
> +
> + /* Force chip select mode */
> + unsigned short force_cs_mode;
> +
> + /* FIFO depth in bytes, max value 64 */
> + unsigned short fifo_depth;
> +
> };
>
> struct omap2_mcspi_device_config {
> Index: linux-omap-2.6/drivers/spi/omap2_mcspi.c
> ===
> --- linux-omap-2.6.orig/drivers/spi/omap2_mcspi.c 2009-05-19
> 17:00:21.0
> +0530
> +++ linux-omap-2.6/drivers/spi/omap2_mcspi.c 2009-05-20 11:02:41.0
> +0530
> @@ -37,9 +37,11 @@
>
> #include
> #include
> +#include
>
>
> #define OMAP2_MCSPI_MAX_FREQ 4800
> +#define OMAP2_MCSPI_MAX_FIFODEPTH64
>
> #define OMAP2_MCSPI_REVISION 0x00
> #define OMAP2_MCSPI_SYSCONFIG0x10
> @@ -49,6 +51,7 @@
> #define OMAP2_MCSPI_WAKEUPENABLE 0x20
> #define OMAP2_MCSPI_SYST 0x24
> #define OMAP2_MCSPI_MODULCTRL0x28
> +#define OMAP2_MCSPI_XFERLEVEL0x7c
>
> /* per-channel banks, 0x14 bytes each, first is: */
> #define OMAP2_MCSPI_CHCONF0 0x2c
> @@ -85,6 +88,9 @@
> #define OMAP2_MCSPI_CHCONF_ISBIT(18)
> #define OMAP2_MCSPI_CHCONF_TURBO BIT(19)
> #define OMAP2_MCSPI_CHCONF_FORCE BIT(20)
> +#define OMAP2_MCSPI_CHCONF_FFER BIT(28)
> +#define OMAP2_MCSPI_CHCONF_FFET BIT(27)
> +
>
> #define OMAP2_MCSPI_CHSTAT_RXS BIT(0)
> #define OMAP2_MCSPI_CHSTAT_TXS BIT(1)
> @@ -93,6 +99,7 @@
Please swap BIT(27) to be before BIT(28) to keep them sorted above.
> #define OMAP2_MCSPI_CHCTRL_ENBIT(0)
>
> #define OMAP2_MCSPI_WAKEUPENABLE_WKENBIT(0)
> +#define OMAP2_MCSPI_IRQ_EOW BIT(17)
>
> /* We have 2 DMA channels per CS, one for RX and one for TX */
> struct omap2_mcspi_dma {
> @@ -125,6 +132,10 @@
> unsigned long phys;
> /* SPI1 has 4 channels, while SPI2 has 2 */
> struct omap2_mcspi_dma *dma_channels;
> + unsigned short mcspi_mode;
> + unsigned short dma_mode;
> + unsigned short force_cs_mode;
> + unsigned
Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
Tony, Any update on this patch.
Thanks
Hemanth
> This patch adds support for McSPI slave and FIFO. DMA and FIFO
> could be enabled together for better throughput. Platform config
> parameters have been added to enable these features on any particular
> McSPI controller.
>
> FIFO can be enabled by defining fifo_depth parameter. fifo_depth needs
> to be a multiple of buffer size that is used for read/write.
>
> These features are useful when you have high throughput devices
> like WLAN or Modem connected over SPI.
>
> Signed-off-by: Hemanth V
> arch/arm/mach-omap2/devices.c |5
> arch/arm/plat-omap/include/mach/mcspi.h | 16 +
> drivers/spi/omap2_mcspi.c | 343
>
> 3 files changed, 325 insertions(+), 39 deletions(-)
>
> ---
> Index: linux-omap-2.6/arch/arm/mach-omap2/devices.c
> ===
> --- linux-omap-2.6.orig/arch/arm/mach-omap2/devices.c 2009-05-19
> 17:00:21.0 +0530
> +++ linux-omap-2.6/arch/arm/mach-omap2/devices.c 2009-05-20
> 11:02:41.0
> +0530
> @@ -259,6 +259,7 @@
>
> static struct omap2_mcspi_platform_config omap2_mcspi1_config = {
> .num_cs = 4,
> + .force_cs_mode = 1,
> };
>
> static struct resource omap2_mcspi1_resources[] = {
> @@ -281,6 +282,10 @@
>
> static struct omap2_mcspi_platform_config omap2_mcspi2_config = {
> .num_cs = 2,
> + .mode = OMAP2_MCSPI_MASTER,
> + .dma_mode = 1,
> + .force_cs_mode = 0,
> + .fifo_depth = 0,
> };
>
> static struct resource omap2_mcspi2_resources[] = {
> Index: linux-omap-2.6/arch/arm/plat-omap/include/mach/mcspi.h
> ===
> --- linux-omap-2.6.orig/arch/arm/plat-omap/include/mach/mcspi.h
> 2009-05-19
> 17:00:21.0 +0530
> +++ linux-omap-2.6/arch/arm/plat-omap/include/mach/mcspi.h2009-05-20
> 11:02:41.0 +0530
> @@ -1,8 +1,24 @@
> #ifndef _OMAP2_MCSPI_H
> #define _OMAP2_MCSPI_H
>
> +#define OMAP2_MCSPI_MASTER 0
> +#define OMAP2_MCSPI_SLAVE1
> +
> struct omap2_mcspi_platform_config {
> unsigned short num_cs;
> +
> + /* SPI is master or slave */
> + unsigned short mode;
> +
> + /* Use only DMA for data transfers */
> + unsigned short dma_mode;
> +
> + /* Force chip select mode */
> + unsigned short force_cs_mode;
> +
> + /* FIFO depth in bytes, max value 64 */
> + unsigned short fifo_depth;
> +
> };
>
> struct omap2_mcspi_device_config {
> Index: linux-omap-2.6/drivers/spi/omap2_mcspi.c
> ===
> --- linux-omap-2.6.orig/drivers/spi/omap2_mcspi.c 2009-05-19
> 17:00:21.0
> +0530
> +++ linux-omap-2.6/drivers/spi/omap2_mcspi.c 2009-05-20 11:02:41.0
> +0530
> @@ -37,9 +37,11 @@
>
> #include
> #include
> +#include
>
>
> #define OMAP2_MCSPI_MAX_FREQ 4800
> +#define OMAP2_MCSPI_MAX_FIFODEPTH64
>
> #define OMAP2_MCSPI_REVISION 0x00
> #define OMAP2_MCSPI_SYSCONFIG0x10
> @@ -49,6 +51,7 @@
> #define OMAP2_MCSPI_WAKEUPENABLE 0x20
> #define OMAP2_MCSPI_SYST 0x24
> #define OMAP2_MCSPI_MODULCTRL0x28
> +#define OMAP2_MCSPI_XFERLEVEL0x7c
>
> /* per-channel banks, 0x14 bytes each, first is: */
> #define OMAP2_MCSPI_CHCONF0 0x2c
> @@ -85,6 +88,9 @@
> #define OMAP2_MCSPI_CHCONF_ISBIT(18)
> #define OMAP2_MCSPI_CHCONF_TURBO BIT(19)
> #define OMAP2_MCSPI_CHCONF_FORCE BIT(20)
> +#define OMAP2_MCSPI_CHCONF_FFER BIT(28)
> +#define OMAP2_MCSPI_CHCONF_FFET BIT(27)
> +
>
> #define OMAP2_MCSPI_CHSTAT_RXS BIT(0)
> #define OMAP2_MCSPI_CHSTAT_TXS BIT(1)
> @@ -93,6 +99,7 @@
> #define OMAP2_MCSPI_CHCTRL_ENBIT(0)
>
> #define OMAP2_MCSPI_WAKEUPENABLE_WKENBIT(0)
> +#define OMAP2_MCSPI_IRQ_EOW BIT(17)
>
> /* We have 2 DMA channels per CS, one for RX and one for TX */
> struct omap2_mcspi_dma {
> @@ -125,6 +132,10 @@
> unsigned long phys;
> /* SPI1 has 4 channels, while SPI2 has 2 */
> struct omap2_mcspi_dma *dma_channels;
> + unsigned short mcspi_mode;
> + unsigned short dma_mode;
> + unsigned short force_cs_mode;
> + unsigned short fifo_depth;
> };
>
> struct omap2_mcspi_cs {
> @@ -133,6 +144,37 @@
> int word_len;
> };
>
> +#ifdef CONFIG_SPI_DEBUG
> +struct reg_type {
> + char name[40];
> + int offset;
> +};
> +
> +static struct reg_type reg_map[] = {
> + {"MCSPI_REV", 0x0},
> + {"MCSPI_SYSCONFIG", 0x10},
> + {"MCSPI_SYSSTATUS", 0x14},
> + {"MCSPI_IRQSTATUS", 0x18},
> + {"MCSPI_IRQENABLE", 0x1C},
> + {"MCSPI_WAKEUPENABLE", 0x20},
> + {"MCSPI_SYST", 0x24},
> + {"MCSPI_MODULCTRL", 0x28},
>
Re: [PATCH 1/2] McSPI Slave and DMA,FIFO support
Santosh, - Original Message - From: "Shilimkar, Santosh" To: "V, Hemanth" ; Sent: Wednesday, May 20, 2009 11:34 AM Subject: RE: [PATCH 1/2] McSPI Slave and DMA,FIFO support Hemanth, Signed-off-by: Hemanth V arch/arm/mach-omap2/devices.c |5 arch/arm/plat-omap/include/mach/mcspi.h | 16 + drivers/spi/omap2_mcspi.c | 343 drivers/spi/omap2_mcspi.c This should be separate patch since this has touching drivers directory. It would break compililation of the patch -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
RE: [PATCH 1/2] McSPI Slave and DMA,FIFO support
Hemanth, > Signed-off-by: Hemanth V > arch/arm/mach-omap2/devices.c |5 > arch/arm/plat-omap/include/mach/mcspi.h | 16 + > drivers/spi/omap2_mcspi.c | 343 > drivers/spi/omap2_mcspi.c This should be separate patch since this has touching drivers directory. How about arranging this series [1/2]: OMAP platform specific changes ( arch/arm/plat-omap, arch/arm/mach-omap) [2/2]: Drivers specific (drivers/spi) Regards, Santosh -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
