| Message ID | 20240315015708.13948-1-justin.swartz@risingedge.co.za (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | spi: mt7621: allow GPIO chip select lines | expand |
On Fri, Mar 15, 2024 at 03:57:07AM +0200, Justin Swartz wrote: > Add the mt7621_spi_set_cs_gpio() function to control the > logical state of a GPIO chip select line, agnostic of the > electrical line state and activation polarity. The core should handle GPIO chip selects for you?
On 2024-03-15 16:45, Mark Brown wrote: > On Fri, Mar 15, 2024 at 03:57:07AM +0200, Justin Swartz wrote: > >> Add the mt7621_spi_set_cs_gpio() function to control the >> logical state of a GPIO chip select line, agnostic of the >> electrical line state and activation polarity. > > The core should handle GPIO chip selects for you? As far as I can tell, it doesn't - at least as far the state of spi-mt7621.c is concerned prior to the patch, plus kernel configuration choices, device tree definition, and other factors I might not be taking into account. But maybe I'm doing something wrong, or perhaps have a misconfiguration somewhere. So, if you're able to point out something I've done incorrectly, it would be appreciated. To attempt to confirm if the core will handle my desired GPIO chip select lines without explicit state toggling, I tried to set the value of use_gpio_descriptors to true, without any other modifications to spi-mt7621.c as of commit 90d35da658da8cff0d4ecbb5113f5fac9d00eb72: [... Sorry if my tabs decide to be spaces instead ...] ---%--- --- a/drivers/spi/spi-mt7621.c +++ b/drivers/spi/spi-mt7621.c @@ -357,6 +357,7 @@ static int mt7621_spi_probe(struct platform_device *pdev) host->bits_per_word_mask = SPI_BPW_MASK(8); host->dev.of_node = pdev->dev.of_node; host->num_chipselect = 2; + host->use_gpio_descriptors = true; dev_set_drvdata(&pdev->dev, host); ---%--- I use a smallish program to write(2) a few bytes from stdin to an spidev node. ---%--- #include <linux/spi/spidev.h> #include <sys/ioctl.h> #include <fcntl.h> #include <stdio.h> #include <stdint.h> #include <stdlib.h> #include <unistd.h> static int Device = -1; static uint32_t Mode = 0; static uint32_t BitsPerWord = 8; static uint32_t MaxSpeed = 100000; #define BUFFER_SIZE 4096 static uint8_t Buffer[BUFFER_SIZE]; static int openDevice(char *); static void closeDevice(void); static int transmit(void); int main(int argc, char *argv[]) { if (argc != 2) { puts("usage: spiw SPI-DEVICE"); return EXIT_FAILURE; } atexit(closeDevice); if (openDevice(argv[1]) == -1) return EXIT_FAILURE; while (!feof(stdin)) if (transmit() == -1) return EXIT_FAILURE; return EXIT_SUCCESS; } static int openDevice(char *filename) { Device = open(filename, O_RDWR); if (Device == -1) return -1; if (ioctl(Device, SPI_IOC_WR_MODE32, &Mode) == -1) return -1; if (ioctl(Device, SPI_IOC_WR_BITS_PER_WORD, &BitsPerWord) == -1) return -1; if (ioctl(Device, SPI_IOC_WR_MAX_SPEED_HZ, &MaxSpeed) == -1) return -1; return 0; } static void closeDevice(void) { if (Device != -1) close(Device); } static int transmit(void) { size_t length = fread(Buffer, 1, sizeof(Buffer), stdin); if (ferror(stdin)) return -1; return write(Device, Buffer, length); } ---%--- If I send write some data to a device associated with one of the GPIO chip selects while watching the signals on the SPI bus, I can see the expected transitions on SCLK and MOSI but there isn't any change on the expected CS line, nor any others: ~ # printf "\x41" | /tmp/spiw /dev/spidev0.2 A rough diagram, to show that although 'A' was sent, the chip select wasn't activated: ______________________________________________________ CS2 __ __ __ __ __ __ __ __ SCLK ________| |__| |__| |__| |__| |__| |__| |__| | ________ _____ ______ MOSI |__| |_____________________________| : : : : : : : : : : 0 1 0 0 0 0 0 1
On Fri, Mar 15, 2024 at 06:23:09PM +0200, Justin Swartz wrote: > On 2024-03-15 16:45, Mark Brown wrote: > > The core should handle GPIO chip selects for you? > As far as I can tell, it doesn't - at least as far the state > of spi-mt7621.c is concerned prior to the patch, plus kernel > configuration choices, device tree definition, and other > factors I might not be taking into account. > But maybe I'm doing something wrong, or perhaps have a > misconfiguration somewhere. So, if you're able to point out > something I've done incorrectly, it would be appreciated. Look at other drivers that support GPIO chip selects? > To attempt to confirm if the core will handle my desired > GPIO chip select lines without explicit state toggling, > I tried to set the value of use_gpio_descriptors to true, > without any other modifications to spi-mt7621.c as of > commit 90d35da658da8cff0d4ecbb5113f5fac9d00eb72: Please include human readable descriptions of things like commits and issues being discussed in e-mail in your mails, this makes them much easier for humans to read especially when they have no internet access. I do frequently catch up on my mail on flights or while otherwise travelling so this is even more pressing for me than just being about making things a bit easier to read. The core needs to know that the GPIO chip selects are there but once it knows that they're there things like setting the chip select should just work.
On 2024-03-15 19:47, Mark Brown wrote: > On Fri, Mar 15, 2024 at 06:23:09PM +0200, Justin Swartz wrote: >> On 2024-03-15 16:45, Mark Brown wrote: > >> > The core should handle GPIO chip selects for you? > >> As far as I can tell, it doesn't - at least as far the state >> of spi-mt7621.c is concerned prior to the patch, plus kernel >> configuration choices, device tree definition, and other >> factors I might not be taking into account. > >> But maybe I'm doing something wrong, or perhaps have a >> misconfiguration somewhere. So, if you're able to point out >> something I've done incorrectly, it would be appreciated. > > Look at other drivers that support GPIO chip selects? Of the 43 drivers (of drivers/spi/*.c) that setup the spi_controller's use_gpio_descriptors as true: 39 drivers use the transfer_one hook, and 4 drivers use the transfer_one_message hook. Drivers that use the transfer_one hook benefit from the core taking care of chip selection on their behalf. Drivers that use the transfer_one_message hook handle chip selection on their own, within the function they've pointed the hook at. There's comment prior to the declaration of the spi_controller structure (in include/linux/spi/spi.h) that says the following about the transfer_one* hooks (beginning at line 493): * Note: transfer_one and transfer_one_message are mutually * exclusive; when both are set, the generic subsystem does * not call your transfer_one callback. Considering spi-mt7621.c was implemented using the transfer_one_message() hook, I'd assumed that it made more sense to take the approach of determining whether a chip select was native or GPIO, and then calling a function that is responsible only for the control over a single type of chip select to ensure that I was not influencing the previous native chip select logic in any drastic manner. To me that seems less intrusive, and allows less room for potential breakage for existing users of this driver (who are native chip select users only), than the outright refactoring of the mt7621_spi_transfer_one_message() function into an mt7621_spi_transfer_one() function instead. Based on reading (some of) drivers/spi/spi.c and looking at cflow-generated callgraph of drivers/spi/spi.c, to determine where spi_set_cs() and any gpiod_* functions are called, I believe that only the transfer_one() hook approach leads to SPI core control of the GPIO chip select lines - via the core's own spi_transfer_one_message(). >> To attempt to confirm if the core will handle my desired >> GPIO chip select lines without explicit state toggling, >> I tried to set the value of use_gpio_descriptors to true, >> without any other modifications to spi-mt7621.c as of >> commit 90d35da658da8cff0d4ecbb5113f5fac9d00eb72: > > Please include human readable descriptions of things like commits and > issues being discussed in e-mail in your mails, this makes them much > easier for humans to read especially when they have no internet access. > I do frequently catch up on my mail on flights or while otherwise > travelling so this is even more pressing for me than just being about > making things a bit easier to read. I understand that. The unlabelled commit was: $ git log | grep -A5 90d35da658da8cff0d4ecbb5113f5fac9d00eb72 commit 90d35da658da8cff0d4ecbb5113f5fac9d00eb72 Author: Linus Torvalds <torvalds@linux-foundation.org> Date: Sun Mar 3 13:02:52 2024 -0800 Linux 6.8-rc7 > The core needs to know that the GPIO chip selects are there but once it > knows that they're there things like setting the chip select should > just > work. This seems to be true for transfer_one() hookers only.
On Fri, Mar 15, 2024 at 10:21:53PM +0200, Justin Swartz wrote: > On 2024-03-15 19:47, Mark Brown wrote: > > Look at other drivers that support GPIO chip selects? > Of the 43 drivers (of drivers/spi/*.c) that setup the > spi_controller's use_gpio_descriptors as true: > 39 drivers use the transfer_one hook, and > 4 drivers use the transfer_one_message hook. > Drivers that use the transfer_one hook benefit from the core > taking care of chip selection on their behalf. > Drivers that use the transfer_one_message hook handle chip > selection on their own, within the function they've pointed > the hook at. Oh, this is an old school driver. Glancing at the code I can't see any particular reason why it's not using transfer_one(), you should just convert the driver to that which will reduce the open coding and just generally improve functionality. You could add a callback to flush the write FIFO or add that into the write function, I'm not sure if there's a meaningful performance benefit there. > Considering spi-mt7621.c was implemented using the > transfer_one_message() hook, I'd assumed that it made more I think it's just old and based on having gone through staging likely based on even older BSP code.
diff --git a/drivers/spi/spi-mt7621.c b/drivers/spi/spi-mt7621.c index 4e9053d03..87e164c86 100644 --- a/drivers/spi/spi-mt7621.c +++ b/drivers/spi/spi-mt7621.c @@ -52,6 +52,8 @@ #define MT7621_CPOL BIT(4) #define MT7621_LSB_FIRST BIT(3) +#define MT7621_NATIVE_CS_COUNT 2 + struct mt7621_spi { struct spi_controller *host; void __iomem *base; @@ -75,7 +77,19 @@ static inline void mt7621_spi_write(struct mt7621_spi *rs, u32 reg, u32 val) iowrite32(val, rs->base + reg); } -static void mt7621_spi_set_cs(struct spi_device *spi, int enable) +static int mt7621_spi_set_cs_gpio(struct spi_device *spi, int enable) +{ + struct gpio_desc *gpiod = spi_get_csgpiod(spi, 0); + int cs = spi_get_chipselect(spi, 0); + int status = gpiod_direction_output(gpiod, enable); + + if (status) + dev_err(&spi->dev, "set_gpio: failed to set CS%d", cs); + + return status; +} + +static void mt7621_spi_set_native_cs(struct spi_device *spi, int enable) { struct mt7621_spi *rs = spidev_to_mt7621_spi(spi); int cs = spi_get_chipselect(spi, 0); @@ -99,6 +113,15 @@ static void mt7621_spi_set_cs(struct spi_device *spi, int enable) mt7621_spi_write(rs, MT7621_SPI_POLAR, polar); } +static int mt7621_spi_set_cs(struct spi_device *spi, int enable) +{ + if (spi_is_csgpiod(spi)) + return mt7621_spi_set_cs_gpio(spi, enable); + + mt7621_spi_set_native_cs(spi,enable); + return 0; +} + static int mt7621_spi_prepare(struct spi_device *spi, unsigned int speed) { struct mt7621_spi *rs = spidev_to_mt7621_spi(spi); @@ -266,7 +289,9 @@ static int mt7621_spi_transfer_one_message(struct spi_controller *host, } /* Assert CS */ - mt7621_spi_set_cs(spi, 1); + status = mt7621_spi_set_cs(spi, 1); + if (status) + goto msg_done; m->actual_length = 0; list_for_each_entry(t, &m->transfers, transfer_list) { @@ -290,7 +315,7 @@ static int mt7621_spi_transfer_one_message(struct spi_controller *host, /* Flush data and deassert CS */ mt7621_spi_flush(rs); - mt7621_spi_set_cs(spi, 0); + status = mt7621_spi_set_cs(spi, 0); msg_done: m->status = status; @@ -313,9 +338,18 @@ static int mt7621_spi_setup(struct spi_device *spi) return -EINVAL; } + if (spi_is_csgpiod(spi)) + return mt7621_spi_set_cs_gpio(spi, 0); + return 0; } +static void mt7621_spi_cleanup(struct spi_device *spi) +{ + if (spi_is_csgpiod(spi)) + mt7621_spi_set_cs_gpio(spi, 0); +} + static const struct of_device_id mt7621_spi_match[] = { { .compatible = "ralink,mt7621-spi" }, {}, @@ -353,10 +387,14 @@ static int mt7621_spi_probe(struct platform_device *pdev) host->mode_bits = SPI_LSB_FIRST; host->flags = SPI_CONTROLLER_HALF_DUPLEX; host->setup = mt7621_spi_setup; + host->cleanup = mt7621_spi_cleanup; host->transfer_one_message = mt7621_spi_transfer_one_message; host->bits_per_word_mask = SPI_BPW_MASK(8); host->dev.of_node = pdev->dev.of_node; - host->num_chipselect = 2; + + host->max_native_cs = MT7621_NATIVE_CS_COUNT; + host->num_chipselect = host->max_native_cs; + host->use_gpio_descriptors = true; dev_set_drvdata(&pdev->dev, host);
Extract a magic number, from mt7621_spi_probe(), used to declare the number of chip select lines (which co-incides with the native chip select count of 2) to a macro. Use the newly defined MT7621_NATIVE_CS_COUNT macro to instead populate both the spi_controller's max_native_cs and num_chipselect members. Declare that the spi_controller should use_gpio_descriptors if present in the device properties (such as those declared in the cs-gpio property of a "ralink,mt7621-spi" compatible device-tree node) so that the SPI core will recalulcate num_chipselect to account for the GPIO descriptors that it should have populated in the cs_gpiod array member. Add the mt7621_spi_set_cs_gpio() function to control the logical state of a GPIO chip select line, agnostic of the electrical line state and activation polarity. Add the mt7621_spi_cleanup() function to ensure that every GPIO chip select will be deactivated when no longer in use. Extend mt7621_spi_setup() so that if an SPI device is associated with a GPIO chip select, its chip select line will be deactivated before use. Rename mt7621_spi_set_cs() to mt7621_spi_set_native_cs(), and redefine mt7621_spi_set_cs() to determine whether: to call mt7621_spi_set_cs_gpio(), in the case of the passed SPI device being associated with a GPIO chip select line, or to call mt7621_spi_set_set_native_cs() instead. Modify mt7621_transfer_one_message() to take into account that mt7621_spi_set_cs() now returns an int and should use the returned value for spi_message status indication if a failure related to GPIO access has occured. Signed-off-by: Justin Swartz <justin.swartz@risingedge.co.za> --- See Documentation/devicetree/bindings/spi/spi-controller.yaml for information about cs-gpios semantics. Example: &spi0 { cs-gpios = <0>, <0>, <&gpio 18 GPIO_ACTIVE_LOW>, /* WDT_RST_N */ <&gpio 19 GPIO_ACTIVE_HIGH>; /* PERST_N */ status = "ok"; ... spidev@2 { compatible = "defective,by-design"; reg = <2>; spi-max-frequency = <16000000>; }; spidev@3 { compatible = "defective,by-design"; reg = <3>; spi-cs-high; spi-max-frequency = <16000000>; }; }; drivers/spi/spi-mt7621.c | 46 ++++++++++++++++++++++++++++++++++++---- 1 file changed, 42 insertions(+), 4 deletions(-) --