spi: Add PPC4xx SPI driver
PATCH v6 1/1 spi: Add PPC4xx SPI driver
This adds a SPI driver for the SPI controller found in the IBM/AMCC 4xx PowerPC's.
---
Changes in v6:
- Moved comment about high interrupt load to top of file and extended it by explaining that the 4xx SPI controller has no FIFOs.
- Added parameter checking to setup() routine.
- Removed comment about LSB
- Used of_gpio_count() instead creating own static implementation as suggested by Anton.
Changes in v5:
- Don't call setupxfer() from setup() so that the baudrate etc won't get changed while another transfer is active, as suggested by David Brownell.
- module_{init,exit} moved directly to the functions to which they apply.
- Use __func__ instead of __FUNCTION__.
Changes in v4:
- Added fixes suggested by Josh Boyer
- Changed compatible property from "ibm,spi" to "ibm,ppc4xx-spi"
Changes in v3:
- When the device is removed the GPIOs are released. The memory for the GPIO array is freed.
Changes in v2:
- Now the gpios property is correctly decoded and the resulting gpio numbers are used as the devices chip selects.
drivers/spi/Kconfig | 7 + drivers/spi/Makefile | 1 + drivers/spi/spi_ppc4xx.c | 594 ++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 602 insertions(+), 0 deletions(-) create mode 100644 drivers/spi/spi_ppc4xx.c
drivers/spi/Kconfig +7 -0
@@ -155,6 +155,13 @@ config SPI_ORION help This enables using the SPI master controller on the Orion chips. +config SPI_PPC4xx + tristate "PPC4xx SPI Controller" + depends on 4xx && SPI_MASTER + select SPI_BITBANG + help + This selects a driver for the PPC4xx SPI Controller. + config SPI_PXA2XX tristate "PXA2xx SSP SPI master" depends on ARCH_PXA && EXPERIMENTAL
drivers/spi/Makefile +1 -0
@@ -24,6 +24,7 @@ obj-$(CONFIG_SPI_OMAP24XX) += omap2_mcspi.o obj-$(CONFIG_SPI_ORION) += orion_spi.o obj-$(CONFIG_SPI_MPC52xx_PSC) += mpc52xx_psc_spi.o obj-$(CONFIG_SPI_MPC83xx) += spi_mpc83xx.o +obj-$(CONFIG_SPI_PPC4xx) += spi_ppc4xx.o obj-$(CONFIG_SPI_S3C24XX_GPIO) += spi_s3c24xx_gpio.o obj-$(CONFIG_SPI_S3C24XX) += spi_s3c24xx.o obj-$(CONFIG_SPI_TXX9) += spi_txx9.o
drivers/spi/spi_ppc4xx.c new file +594 -0
594 lines not on this page · raw message
Signature · 1 line
1.6.1
A message that is not in the archive (sent off-list or not yet received): 1231411250-25380-1-git-send-email-sr-ynQEQJNshbs@public.gmane.org
Quoted 2 lines · Stefan Roese
This adds a SPI driver for the SPI controller found in the IBM/AMCC 4xx PowerPC's.
Note that given some patches now in the mm tree, this needs something like the appended fixup. Some common code has now moved into the spi core.
- Dave
--- drivers/spi/spi_ppc4xx.c | 19 +++---------------- 1 file changed, 3 insertions(+), 16 deletions(-)
drivers/spi/spi_ppc4xx.c +3 -16
@@ -61,9 +61,6 @@/* RxD ready */ #define SPI_PPC4XX_SR_RBR (0x80 >> 7) -/* the spi->mode bits understood by this driver: */ -#define MODEBITS (SPI_CPHA | SPI_CPOL | SPI_CS_HIGH | SPI_LSB_FIRST) -/* clock settings (SCP and CI) for various SPI modes */ #define SPI_CLK_MODE0 SPI_PPC4XX_MODE_SCP#define SPI_CLK_MODE1 0
@@ -198,9 +195,6 @@ static int spi_ppc4xx_setup(struct spi_dstruct spi_ppc4xx_cs *cs = spi->controller_state;int init =0 ; -if (!spi->bits_per_word) - spi->bits_per_word =8 ; -if (spi->bits_per_word !=8 ) {dev_err (&spi->dev,"invalid bits-per-word (%d)\n" , spi->bits_per_word);
@@ -212,12 +206,6 @@ static int spi_ppc4xx_setup(struct spi_dreturn -EINVAL ; } -if (spi->mode & ~MODEBITS ) { -dev_dbg (&spi->dev,"setup: unsupported mode bits %x\n" , - spi->mode & ~MODEBITS ); -return -EINVAL ; - } -if (cs ==NULL ) { cs =kzalloc (sizeof *cs,GFP_KERNEL );if (!cs)
@@ -268,10 +256,6 @@ static int spi_ppc4xx_setup(struct spi_d } } -dev_dbg (&spi->dev,"%s: mode %d, %u bpw, %d hz\n" , - __func__, spi->mode, spi->bits_per_word, - spi->max_speed_hz); -return 0 ; }
@@ -442,6 +426,9 @@ static int __init spi_ppc4xx_of_probe(st } } +/* the spi->mode bits understood by this driver: */ + master->modebits =SPI_CPHA |SPI_CPOL |SPI_CS_HIGH |SPI_LSB_FIRST ; +/* Setup the state for the bitbang driver */ bbp = &hw->bitbang; bbp->master = hw->master;
------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and around Java (TM) technology - register by April 22, and save $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco. 300 plus technical and hands-on sessions. Register today. Use priority code J9JMT32. http://p.sf.net/sfu/p
Quoted 7 lines · David Brownell
Quoted 2 lines · Stefan Roese
This adds a SPI driver for the SPI controller found in the IBM/AMCC 4xx PowerPC's.
Note that given some patches now in the mm tree, this needs something like the appended fixup. Some common code has now moved into the spi core.
So would you accept the driver for 2.6.31 merge with your suggested change below? If yes, I'll generate a 7th version.
Thanks.
Best regards,
Stefan
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 11 lines · Stefan Roese
Quoted 7 lines · David Brownell
Quoted 2 lines · Stefan Roese
This adds a SPI driver for the SPI controller found in the IBM/AMCC 4xx PowerPC's.
Note that given some patches now in the mm tree, this needs something like the appended fixup. Some common code has now moved into the spi core.
So would you accept the driver for 2.6.31 merge with your suggested change below? If yes, I'll generate a 7th version.
Hold on until I send the rest of my review comments; almost done. Sorry for the delay. Yes, I think a 2.6.31 merge should be doable.
- Dave
Quoted 7 lines
Thanks.
Best regards,
Stefan
------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and around Java (TM) technology - register by April 22, and save $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco. 300 plus technical and hands-on sessions. Register today. Use priority code J9JMT32. http://p.sf.net/sfu/p
Quoted 9 lines · Stefan Roese
This adds a SPI driver for the SPI controller found in the IBM/AMCC 4xx PowerPC's.
+/*
+ * The PPC4xx SPI controller has no FIFO so each sent/received byte will
+ * generate an interrupt to the CPU. This can cause high CPU utilization.
+ * This driver allows platforms to reduce the interrupt load on the CPU
+ * during SPI transfers by settingmax_speed_hzvia the device tree.
Note that an alternate strategy is to use polling instead of IRQs, at least when the data rate is high enough that the IRQ processing is also slowing down the data transfers.
drivers/spi/spi_ppc4xx.c line 46
Quoted 3 lines
drivers/spi/spi_ppc4xx.c new file +3
-0+/* bits in mode register - bit 0 ist MSb */ +/* data latched on leading edge of clock, else trailing edge */ +#define SPI_PPC4XX_MODE_SCP (0x80 >> 3)
Or in short, SCP == CPHA.
drivers/spi/spi_ppc4xx.c line 52
Quoted 6 lines
drivers/spi/spi_ppc4xx.c new file +6
-0+/* port enabled */ +#define SPI_PPC4XX_MODE_SPE (0x80 >> 4) +/* MSB first, else LSB first */ +#define SPI_PPC4XX_MODE_RD (0x80 >> 5) +/* clock invert - idle clock = 1, active clock = 0; else reversed */ +#define SPI_PPC4XX_MODE_CI (0x80 >> 6)
Or in short, CI == CPOL.
Quoted 9 lines
drivers/spi/spi_ppc4xx.c new file +9
-0+/* loopback enable */ +#define SPI_PPC4XX_MODE_IL (0x80 >> 7) +/* bits in control register */ +/* starts a transfer when set */ +#define SPI_PPC4XX_CR_STR (0x80 >> 7) +/* bits in status register */ +/* port is busy with a transfer */ +#define SPI_PPC4XX_SR_BSY (0x80 >> 6) +/* RxD ready */
drivers/spi/spi_ppc4xx.c line 71
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+#define SPI_PPC4XX_SR_RBR (0x80 >> 7) + +/* the spi->mode bits understood by this driver: */ +#define MODEBITS (SPI_CPHA | SPI_CPOL | SPI_CS_HIGH | SPI_LSB_FIRST) + +/* clock settings (SCP and CI) for various SPI modes */ +#define SPI_CLK_MODE0 SPI_PPC4XX_MODE_SCP +#define SPI_CLK_MODE1 0 +#define SPI_CLK_MODE2 SPI_PPC4XX_MODE_CI +#define SPI_CLK_MODE3 (SPI_PPC4XX_MODE_SCP | SPI_PPC4XX_MODE_CI)
Color me puzzled then. Either the definitions of MODE0 and MODE1 are swapped here ... or the comments above are wrong, and SCP should describe "falling" vs "rising" instead of "leading" vs "trailing".
Quoted 5 lines
drivers/spi/spi_ppc4xx.c new file +5
-0+static int spi_ppc4xx_setupxfer (struct spi_device *spi,struct spi_transfer *t) +{ +struct ppc4xx_spi *hw =spi_master_get_devdata (spi->master); +struct spi_ppc4xx_cs *cs = spi->controller_state; +int scr;
drivers/spi/spi_ppc4xx.c line 166
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+u8 cdm =0 ; +u32 speed; +u8 bits_per_word; + + bits_per_word = (t) ? t->bits_per_word : spi->bits_per_word; + speed = (t) ? t->speed_hz : spi->max_speed_hz; + + ... + +if (!speed || (speed > spi->max_speed_hz)) {
This is wrong. Typical case is that t->speed_hz is zero, meaning "use spi->max_speed_hz" ... you treat that as an error.
drivers/spi/spi_ppc4xx.c line 199
Quoted 12 lines
drivers/spi/spi_ppc4xx.c new file +12
-0+dev_err (&spi->dev,"invalid speed_hz (%d)\n" , speed); +return -EINVAL ; + } + + ... +} + +static int spi_ppc4xx_setup (struct spi_device *spi) +{ +int ret; +struct spi_ppc4xx_cs *cs = spi->controller_state; +int init =0 ;
This "init" thing is still wrong.
Consider: board gets set up with one device. Later, *WHILE BUSY TRANSFERRING DATA TO/FROM THAT DEVICE* some other device gets instantiated. Then ...
drivers/spi/spi_ppc4xx.c line 231
Quoted 11 lines
drivers/spi/spi_ppc4xx.c new file +11
-0+if (cs ==NULL ) { + cs =kzalloc (sizeof *cs,GFP_KERNEL ); +if (!cs) +return -ENOMEM ; + spi->controller_state = cs; + +/* +* First time called, so let's init the SPI controller +* at the end of this function +*/ + init =1 ;
"init" becomes true here, although the controller has already been initialized. If it needs to be set up, do it in probe() before registering the spi_master.
drivers/spi/spi_ppc4xx.c line 264
Quoted 11 lines
drivers/spi/spi_ppc4xx.c new file +11
-0+ } + + ... + +/* +* New configuration (mode, speed etc) will be written to the +* controller in spi_ppc4xx_setupxfer(). Only call +* spi_ppc4xx_setupxfer() directly upon first initialization. +*/ +if (init) { + ret =spi_ppc4xx_setupxfer (spi,NULL );
... so blam, you clobber the settings currently being used for the active transfer. So this would be a bug.
If not ... this driver deserves a comment on exactly how unusual this driver is. Specifically, that all spi_device children must be set up *in advance* so that standard calls like spi_new_device() don't work with it; and why.
I don't think I see anything in this driver which would prevent that from working, though. Sure you've got to have a list of chipselect GPIOs in advance, but that's distinct from being unable to use spi_new_device().
drivers/spi/spi_ppc4xx.c line 285
Quoted 11 lines
drivers/spi/spi_ppc4xx.c new file +11
-0+ ... +} + +static void spi_ppc4xx_chipsel (struct spi_device *spi,int value) +{ +struct ppc4xx_spi *hw =spi_master_get_devdata (spi->master); +unsigned int cspol = spi->mode &SPI_CS_HIGH ?1 :0 ; +unsigned int cs = spi->chip_select; + +if (!hw->num_gpios) +return ;
Hmm, num_gpios ... can never be zero, right? You always set it up so that of_gpio_count() GPIOs can all be used as chipselects. Looks like this check is not needed.
drivers/spi/spi_ppc4xx.c line 288
Quoted 3 lines
drivers/spi/spi_ppc4xx.c new file +3
-0+ +if (cs >= hw->num_gpios) +return ;
I guess I don't see why you have num_gpio instead of just using the chipselect count, either. (Assuming that the of_gpio_count routine isn't counting *all* gpios, instead of just ones allocated to this controller, which would be a different issue.)
drivers/spi/spi_ppc4xx.c line 291
Quoted 3 lines
drivers/spi/spi_ppc4xx.c new file +3
-0+ +if (value !=BITBANG_CS_INACTIVE && value !=BITBANG_CS_ACTIVE ) +return ;
Strange, what other values could be passed??
Quoted 12 lines
drivers/spi/spi_ppc4xx.c new file +12
-0+ +if (value ==BITBANG_CS_INACTIVE ) + cspol = !cspol; + +gpio_set_value (hw->gpios[cs], cspol); +} + +static irqreturn_t spi_ppc4xx_int (int irq,void *dev_id) +{ +struct ppc4xx_spi *hw; +u8 status; +u8 data;
drivers/spi/spi_ppc4xx.c line 313
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+unsigned int count; + + hw = (struct ppc4xx_spi *)dev_id; + + status =in_8 (&hw->regs->sr); +if (!status) +return IRQ_NONE ; + +/* should never happen but check anyway */ +if (status &SPI_PPC4XX_SR_BSY ) {
Maybe "if (unlikely(...)) {" then.
drivers/spi/spi_ppc4xx.c line 317
Quoted 4 lines
drivers/spi/spi_ppc4xx.c new file +4
-0+u8 lstatus; +int cnt =0 ; + +dev_dbg (hw->dev,"got interrupt but spi still busy?\n" );
But ... *does* this ever happen? If not, I'd strike this code. And if it does, I'd add a comment explaining what's known about this hardware bug. Without this, the IRQ handler would look pretty tight.
Quoted 11 lines
drivers/spi/spi_ppc4xx.c new file +11
-0+do { +ndelay (10 ); + lstatus =in_8 (&hw->regs->sr); + }while (++cnt <100 && lstatus &SPI_PPC4XX_SR_BSY ); + +if (cnt >=100 ) { +dev_err (hw->dev,"busywait: too many loops!\n" ); +complete (&hw->done); +return IRQ_HANDLED ; + }else { +/* status is always 1 (RBR) here */
drivers/spi/spi_ppc4xx.c line 338
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+ status =in_8 (&hw->regs->sr); +dev_dbg (hw->dev,"loops %d status %x\n" , cnt, status); + } + } + + count = hw->count; + hw->count++; + +if (status &SPI_PPC4XX_SR_RBR ) { +/* Data Ready */
Most SPI hardware I've seen has at most a weak distinction between "tx done" and "rx done" IRQs ... there's little point to separate IRQs, since shifting out the last TX bit is what shifts in the last RX bit.
Unless this is unusual hardware you can probably just assume that there's RX data, and only read it if hw->rx is non-null. Or, if this is *not* set, report the odd behavior...
Quoted 30 lines
drivers/spi/spi_ppc4xx.c new file +29
-0+ data =in_8 (&hw->regs->rxd); +if (hw->rx) + hw->rx[count] = data; + } + + count++; + +if (count < hw->len) { + data = hw->tx ? hw->tx[count] :0 ; +out_8 (&hw->regs->txd, data); +out_8 (&hw->regs->cr,SPI_PPC4XX_CR_STR ); + }else { +complete (&hw->done); + } + +return IRQ_HANDLED ; +} + + ... + +static int __init spi_ppc4xx_of_probe (struct of_device *op, +const struct of_device_id *match) +{ +struct ppc4xx_spi *hw; +struct spi_master *master; +struct spi_bitbang *bbp; +struct resource resource; +struct device_node *np = op->node; +struct device *dev = &op->dev;
drivers/spi/spi_ppc4xx.c line 407
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+struct device_node *opbnp; +int ret; +const unsigned int *clk; + + master =spi_alloc_master (dev,sizeof *hw); +if (master ==NULL ) +return -ENOMEM ; +dev_set_drvdata (dev, master); + hw =spi_master_get_devdata (master); +memset (hw,0 ,sizeof (struct ppc4xx_spi ));
That memset is not required, it comes to you pre-zeroed.
drivers/spi/spi_ppc4xx.c line 414
Quoted 7 lines
drivers/spi/spi_ppc4xx.c new file +7
-0+ + hw->master =spi_master_get (master); + hw->dev = dev; + +init_completion (&hw->done); + + hw->num_gpios =of_gpio_count (np);
Worth a comment what's going on there. I'm assuming this of_gpio_count() thing returns a count of GPIOs somehow associated with this specific device tree node, rather than say all GPIOs known to the device tree...
Quoted 19 lines
drivers/spi/spi_ppc4xx.c new file +19
-0+if (hw->num_gpios) { +int i; + + hw->gpios =kzalloc (sizeof (int ) * hw->num_gpios, +GFP_KERNEL ); +if (!hw->gpios) { + ret = -ENOMEM ; + goto free_master; + } + +for (i =0 ; i < hw->num_gpios; i++) { +int gpio =of_get_gpio (np, i); +if (gpio <0 ) { +dev_err (dev,"Invalid gpio spec %d\n" , i); + ret = gpio; + goto free_gpios; + } + + ret =gpio_request (gpio, np->name);
drivers/spi/spi_ppc4xx.c line 443
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+if (ret <0 ) { +dev_err (dev,"gpio %d already in use\n" , i); + ret = gpio; + goto free_gpios; + } + +gpio_direction_output (gpio,0 ); + hw->gpios[i] = gpio; + } + }
Else ... error, if there are no chipselects?
I've got a patch pending to add a "no chipselect" spi->mode flag. If you expect to support that, please let me know.
drivers/spi/spi_ppc4xx.c line 455
Quoted 12 lines
drivers/spi/spi_ppc4xx.c new file +12
-0+ +/* Setup the state for the bitbang driver */ + bbp = &hw->bitbang; + bbp->master = hw->master; + bbp->setup_transfer = spi_ppc4xx_setupxfer; + bbp->chipselect = spi_ppc4xx_chipsel; + bbp->txrx_bufs = spi_ppc4xx_txrx; + bbp->use_dma =0 ; + bbp->master->setup = spi_ppc4xx_setup; + bbp->master->cleanup = spi_ppc4xx_cleanup; +/* only one SPI controller */ + bbp->master->bus_num =0 ;
Bad assumption. There could be GPIO-based adapters, of course. And aren't these the cores which can be found on higher end Xilinx FPGAs? And, accordingly, integrated with any number of additional controllers?
If so, the restriction is just that this controller be bus #0. If not ... the Kconfig should say more about *which* PPC4xx chips have this controller. (I suspect it's the latter.)
drivers/spi/spi_ppc4xx.c line 456
Quoted 1 line
drivers/spi/spi_ppc4xx.c new file +1
-0+if (bbp->master->num_chipselect ==0 ) {
You didn't set it, so of course it's still zeroed.
drivers/spi/spi_ppc4xx.c line 458
Quoted 2 lines
drivers/spi/spi_ppc4xx.c new file +2
-0+/* this many pins in al GPIO controllers */ + bbp->master->num_chipselect = hw->num_gpios;
So num_chipselect is alays num_gpios, and there's no point to having num_gpios since you could always use num_chipselect.
Quoted 4 lines
drivers/spi/spi_ppc4xx.c new file +4
-0+ } + + ... +}
A message that is not in the archive (sent off-list or not yet received): 200904221300.39747.david-b-yBeKhBN/[email protected]
Hi
i am a newbie and am doing a project on beagle board(running omap3530). i am interfacing an adc(ads7886) to the beagleboard via spi. need to know how the spi works and how a module can access the spi registers of the omap. if someones already made an adc driver can they mail me?
thanks
arnav
Quoted 369 lines · David Brownell
Quoted 9 lines · Stefan Roese
This adds a SPI driver for the SPI controller found in the IBM/AMCC 4xx PowerPC's.
+/*
+ * The PPC4xx SPI controller has no FIFO so each sent/received byte will
+ * generate an interrupt to the CPU. This can cause high CPU utilization.
+ * This driver allows platforms to reduce the interrupt load on the CPU
+ * during SPI transfers by settingmax_speed_hzvia the device tree.Note that an alternate strategy is to use polling instead of IRQs, at least when the data rate is high enough that the IRQ processing is also slowing down the data transfers.
drivers/spi/spi_ppc4xx.c line 46
Quoted 3 lines
drivers/spi/spi_ppc4xx.c new file +3
-0+/* bits in mode register - bit 0 ist MSb */ +/* data latched on leading edge of clock, else trailing edge */ +#define SPI_PPC4XX_MODE_SCP (0x80 >> 3)Or in short, SCP == CPHA.
drivers/spi/spi_ppc4xx.c line 52
Quoted 6 lines
drivers/spi/spi_ppc4xx.c new file +6
-0+/* port enabled */ +#define SPI_PPC4XX_MODE_SPE (0x80 >> 4) +/* MSB first, else LSB first */ +#define SPI_PPC4XX_MODE_RD (0x80 >> 5) +/* clock invert - idle clock = 1, active clock = 0; else reversed */ +#define SPI_PPC4XX_MODE_CI (0x80 >> 6)Or in short, CI == CPOL.
Quoted 9 lines
drivers/spi/spi_ppc4xx.c new file +9
-0+/* loopback enable */ +#define SPI_PPC4XX_MODE_IL (0x80 >> 7) +/* bits in control register */ +/* starts a transfer when set */ +#define SPI_PPC4XX_CR_STR (0x80 >> 7) +/* bits in status register */ +/* port is busy with a transfer */ +#define SPI_PPC4XX_SR_BSY (0x80 >> 6) +/* RxD ready */ drivers/spi/spi_ppc4xx.c line 71
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+#define SPI_PPC4XX_SR_RBR (0x80 >> 7) + +/* the spi->mode bits understood by this driver: */ +#define MODEBITS (SPI_CPHA | SPI_CPOL | SPI_CS_HIGH | SPI_LSB_FIRST) + +/* clock settings (SCP and CI) for various SPI modes */ +#define SPI_CLK_MODE0 SPI_PPC4XX_MODE_SCP +#define SPI_CLK_MODE1 0 +#define SPI_CLK_MODE2 SPI_PPC4XX_MODE_CI +#define SPI_CLK_MODE3 (SPI_PPC4XX_MODE_SCP | SPI_PPC4XX_MODE_CI)Color me puzzled then. Either the definitions of MODE0 and MODE1 are swapped here ... or the comments above are wrong, and SCP should describe "falling" vs "rising" instead of "leading" vs "trailing".
Quoted 6 lines
+static int
spi_ppc4xx_setupxfer(struct spi_device *spi, struct spi_transfer *t)drivers/spi/spi_ppc4xx.c new file +4
-0+{ +struct ppc4xx_spi *hw =spi_master_get_devdata (spi->master); +struct spi_ppc4xx_cs *cs = spi->controller_state; +int scr;drivers/spi/spi_ppc4xx.c line 166
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+u8 cdm =0 ; +u32 speed; +u8 bits_per_word; + + bits_per_word = (t) ? t->bits_per_word : spi->bits_per_word; + speed = (t) ? t->speed_hz : spi->max_speed_hz; + + ... + +if (!speed || (speed > spi->max_speed_hz)) {This is wrong. Typical case is that
t->speed_hzis zero, meaning "usespi->max_speed_hz" ... you treat that as an error.drivers/spi/spi_ppc4xx.c line 199
Quoted 12 lines
drivers/spi/spi_ppc4xx.c new file +12
-0+dev_err (&spi->dev,"invalid speed_hz (%d)\n" , speed); +return -EINVAL ; + } + + ... +} + +static int spi_ppc4xx_setup (struct spi_device *spi) +{ +int ret; +struct spi_ppc4xx_cs *cs = spi->controller_state; +int init =0 ;This "init" thing is still wrong.
Consider: board gets set up with one device. Later, *WHILE BUSY TRANSFERRING DATA TO/FROM THAT DEVICE* some other device gets instantiated. Then ...
drivers/spi/spi_ppc4xx.c line 231
Quoted 11 lines
drivers/spi/spi_ppc4xx.c new file +11
-0+if (cs ==NULL ) { + cs =kzalloc (sizeof *cs,GFP_KERNEL ); +if (!cs) +return -ENOMEM ; + spi->controller_state = cs; + +/* +* First time called, so let's init the SPI controller +* at the end of this function +*/ + init =1 ;"init" becomes true here, although the controller has already been initialized. If it needs to be set up, do it in
probe()before registering thespi_master.drivers/spi/spi_ppc4xx.c line 264
Quoted 11 lines
drivers/spi/spi_ppc4xx.c new file +11
-0+ } + + ... + +/* +* New configuration (mode, speed etc) will be written to the +* controller in spi_ppc4xx_setupxfer(). Only call +* spi_ppc4xx_setupxfer() directly upon first initialization. +*/ +if (init) { + ret =spi_ppc4xx_setupxfer (spi,NULL );... so blam, you clobber the settings currently being used for the active transfer. So this would be a bug.
If not ... this driver deserves a comment on exactly how unusual this driver is. Specifically, that all
spi_devicechildren must be set up *in advance* so that standard calls likespi_new_device()don't work with it; and why.I don't think I see anything in this driver which would prevent that from working, though. Sure you've got to have a list of chipselect GPIOs in advance, but that's distinct from being unable to use
spi_new_device().drivers/spi/spi_ppc4xx.c line 285
Quoted 11 lines
drivers/spi/spi_ppc4xx.c new file +11
-0+ ... +} + +static void spi_ppc4xx_chipsel (struct spi_device *spi,int value) +{ +struct ppc4xx_spi *hw =spi_master_get_devdata (spi->master); +unsigned int cspol = spi->mode &SPI_CS_HIGH ?1 :0 ; +unsigned int cs = spi->chip_select; + +if (!hw->num_gpios) +return ;Hmm,
num_gpios... can never be zero, right? You always set it up so thatof_gpio_count()GPIOs can all be used as chipselects. Looks like this check is not needed.drivers/spi/spi_ppc4xx.c line 288
Quoted 3 lines
drivers/spi/spi_ppc4xx.c new file +3
-0+ +if (cs >= hw->num_gpios) +return ;I guess I don't see why you have
num_gpioinstead of just using the chipselect count, either. (Assuming that theof_gpio_countroutine isn't counting *all* gpios, instead of just ones allocated to this controller, which would be a different issue.)drivers/spi/spi_ppc4xx.c line 291
Quoted 3 lines
drivers/spi/spi_ppc4xx.c new file +3
-0+ +if (value !=BITBANG_CS_INACTIVE && value !=BITBANG_CS_ACTIVE ) +return ;Strange, what other values could be passed??
Quoted 12 lines
drivers/spi/spi_ppc4xx.c new file +12
-0+ +if (value ==BITBANG_CS_INACTIVE ) + cspol = !cspol; + +gpio_set_value (hw->gpios[cs], cspol); +} + +static irqreturn_t spi_ppc4xx_int (int irq,void *dev_id) +{ +struct ppc4xx_spi *hw; +u8 status; +u8 data;drivers/spi/spi_ppc4xx.c line 313
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+unsigned int count; + + hw = (struct ppc4xx_spi *)dev_id; + + status =in_8 (&hw->regs->sr); +if (!status) +return IRQ_NONE ; + +/* should never happen but check anyway */ +if (status &SPI_PPC4XX_SR_BSY ) {Maybe "if (unlikely(...)) {" then.
drivers/spi/spi_ppc4xx.c line 317
Quoted 4 lines
drivers/spi/spi_ppc4xx.c new file +4
-0+u8 lstatus; +int cnt =0 ; + +dev_dbg (hw->dev,"got interrupt but spi still busy?\n" );But ... *does* this ever happen? If not, I'd strike this code. And if it does, I'd add a comment explaining what's known about this hardware bug. Without this, the IRQ handler would look pretty tight.
Quoted 11 lines
drivers/spi/spi_ppc4xx.c new file +11
-0+do { +ndelay (10 ); + lstatus =in_8 (&hw->regs->sr); + }while (++cnt <100 && lstatus &SPI_PPC4XX_SR_BSY ); + +if (cnt >=100 ) { +dev_err (hw->dev,"busywait: too many loops!\n" ); +complete (&hw->done); +return IRQ_HANDLED ; + }else { +/* status is always 1 (RBR) here */ drivers/spi/spi_ppc4xx.c line 338
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+ status =in_8 (&hw->regs->sr); +dev_dbg (hw->dev,"loops %d status %x\n" , cnt, status); + } + } + + count = hw->count; + hw->count++; + +if (status &SPI_PPC4XX_SR_RBR ) { +/* Data Ready */ Most SPI hardware I've seen has at most a weak distinction between "tx done" and "rx done" IRQs ... there's little point to separate IRQs, since shifting out the last TX bit is what shifts in the last RX bit.
Unless this is unusual hardware you can probably just assume that there's RX data, and only read it if
hw->rxis non-null. Or, if this is *not* set, report the odd behavior...Quoted 30 lines
drivers/spi/spi_ppc4xx.c new file +29
-0+ data =in_8 (&hw->regs->rxd); +if (hw->rx) + hw->rx[count] = data; + } + + count++; + +if (count < hw->len) { + data = hw->tx ? hw->tx[count] :0 ; +out_8 (&hw->regs->txd, data); +out_8 (&hw->regs->cr,SPI_PPC4XX_CR_STR ); + }else { +complete (&hw->done); + } + +return IRQ_HANDLED ; +} + + ... + +static int __init spi_ppc4xx_of_probe (struct of_device *op, +const struct of_device_id *match) +{ +struct ppc4xx_spi *hw; +struct spi_master *master; +struct spi_bitbang *bbp; +struct resource resource; +struct device_node *np = op->node; +struct device *dev = &op->dev;drivers/spi/spi_ppc4xx.c line 407
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+struct device_node *opbnp; +int ret; +const unsigned int *clk; + + master =spi_alloc_master (dev,sizeof *hw); +if (master ==NULL ) +return -ENOMEM ; +dev_set_drvdata (dev, master); + hw =spi_master_get_devdata (master); +memset (hw,0 ,sizeof (struct ppc4xx_spi ));That memset is not required, it comes to you pre-zeroed.
drivers/spi/spi_ppc4xx.c line 414
Quoted 7 lines
drivers/spi/spi_ppc4xx.c new file +7
-0+ + hw->master =spi_master_get (master); + hw->dev = dev; + +init_completion (&hw->done); + + hw->num_gpios =of_gpio_count (np);Worth a comment what's going on there. I'm assuming this
of_gpio_count()thing returns a count of GPIOs somehow associated with this specific device tree node, rather than say all GPIOs known to the device tree...Quoted 19 lines
drivers/spi/spi_ppc4xx.c new file +19
-0+if (hw->num_gpios) { +int i; + + hw->gpios =kzalloc (sizeof (int ) * hw->num_gpios, +GFP_KERNEL ); +if (!hw->gpios) { + ret = -ENOMEM ; + goto free_master; + } + +for (i =0 ; i < hw->num_gpios; i++) { +int gpio =of_get_gpio (np, i); +if (gpio <0 ) { +dev_err (dev,"Invalid gpio spec %d\n" , i); + ret = gpio; + goto free_gpios; + } + + ret =gpio_request (gpio, np->name);drivers/spi/spi_ppc4xx.c line 443
Quoted 10 lines
drivers/spi/spi_ppc4xx.c new file +10
-0+if (ret <0 ) { +dev_err (dev,"gpio %d already in use\n" , i); + ret = gpio; + goto free_gpios; + } + +gpio_direction_output (gpio,0 ); + hw->gpios[i] = gpio; + } + }Else ... error, if there are no chipselects?
I've got a patch pending to add a "no chipselect"
spi->modeflag. If you expect to support that, please let me know.drivers/spi/spi_ppc4xx.c line 455
Quoted 12 lines
drivers/spi/spi_ppc4xx.c new file +12
-0+ +/* Setup the state for the bitbang driver */ + bbp = &hw->bitbang; + bbp->master = hw->master; + bbp->setup_transfer = spi_ppc4xx_setupxfer; + bbp->chipselect = spi_ppc4xx_chipsel; + bbp->txrx_bufs = spi_ppc4xx_txrx; + bbp->use_dma =0 ; + bbp->master->setup = spi_ppc4xx_setup; + bbp->master->cleanup = spi_ppc4xx_cleanup; +/* only one SPI controller */ + bbp->master->bus_num =0 ;Bad assumption. There could be GPIO-based adapters, of course. And aren't these the cores which can be found on higher end Xilinx FPGAs? And, accordingly, integrated with any number of additional controllers?
If so, the restriction is just that this controller be bus #0. If not ... the
Kconfigshould say more about *which* PPC4xx chips have this controller. (I suspect it's the latter.)drivers/spi/spi_ppc4xx.c line 456
Quoted 1 line
drivers/spi/spi_ppc4xx.c new file +1
-0+if (bbp->master->num_chipselect ==0 ) {You didn't set it, so of course it's still zeroed.
drivers/spi/spi_ppc4xx.c line 458
Quoted 2 lines
drivers/spi/spi_ppc4xx.c new file +2
-0+/* this many pins in al GPIO controllers */ + bbp->master->num_chipselect = hw->num_gpios;So
num_chipselectis alaysnum_gpios, and there's no point to havingnum_gpiossince you could always usenum_chipselect.
Quoted 16 lines
drivers/spi/spi_ppc4xx.c line 164
Quoted 4 lines
drivers/spi/spi_ppc4xx.c new file +4
-0+ } + + ... +}------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and around Java (TM) technology - register by April 22, and save $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco. 300 plus technical and hands-on sessions. Register today. Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
spi-devel-general mailing list spi-devel-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensign option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment.
http://p.sf.net/sfu/businessobjects
Re: general SPI question (WAS: [PATCH v6] spi: Add PPC4xx SPI driver)
Quoted 1 line · Arnav Das
i am a newbie
Lesson #1: make sure your Subject: lines match the message topic (I did a partial repair) and don't post to the wrong list (e.g. PPC lists for OMAP questions).
Quoted 3 lines
and am doing a project on beagle board(running omap3530). i am interfacing an adc(ads7886) to the beagleboard via spi. need to know how the spi works
Read the ads7886 data sheet, in this case. It uses a subset of SPI ... no input commands, just a 12-bit sample returned each time chipselect pulses.
Quoted 3 lines
and how a module can access the
spi registers of the omap. if someones already made an adc driver can they mail me?
Use drivers/spi/omap2_mcspi.c on OMAP3 boards, at least to start with. If you want streaming conversions you may want to use different modes than are currently supported by that driver.
- Dave