| Message ID | cover.1615038553.git.syednwaris@gmail.com |
|---|---|
| Headers | show |
| Series | Introduce the for_each_set_nbits macro | expand |
On 3/6/21 3:06 PM, Syed Nayyar Waris wrote: > This patch reimplements the xgpio_set_multiple() function in > drivers/gpio/gpio-xilinx.c to use the new generic functions: > bitmap_get_value() and bitmap_set_value(). The code is now simpler > to read and understand. Moreover, instead of looping for each bit > in xgpio_set_multiple() function, now we can check each channel at > a time and save cycles. > > Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com> > Cc: Michal Simek <michal.simek@xilinx.com> > Signed-off-by: Syed Nayyar Waris <syednwaris@gmail.com> > Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com> > --- > drivers/gpio/gpio-xilinx.c | 63 +++++++++++++++++++------------------- > 1 file changed, 32 insertions(+), 31 deletions(-) > > diff --git a/drivers/gpio/gpio-xilinx.c b/drivers/gpio/gpio-xilinx.c > index be539381fd82..8445e69cf37b 100644 > --- a/drivers/gpio/gpio-xilinx.c > +++ b/drivers/gpio/gpio-xilinx.c > @@ -15,6 +15,7 @@ > #include <linux/of_device.h> > #include <linux/of_platform.h> > #include <linux/slab.h> > +#include "gpiolib.h" > > /* Register Offset Definitions */ > #define XGPIO_DATA_OFFSET (0x0) /* Data register */ > @@ -141,37 +142,37 @@ static void xgpio_set_multiple(struct gpio_chip *gc, unsigned long *mask, > { > unsigned long flags; > struct xgpio_instance *chip = gpiochip_get_data(gc); > - int index = xgpio_index(chip, 0); > - int offset, i; > - > - spin_lock_irqsave(&chip->gpio_lock[index], flags); > - > - /* Write to GPIO signals */ > - for (i = 0; i < gc->ngpio; i++) { > - if (*mask == 0) > - break; > - /* Once finished with an index write it out to the register */ > - if (index != xgpio_index(chip, i)) { > - xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > - index * XGPIO_CHANNEL_OFFSET, > - chip->gpio_state[index]); > - spin_unlock_irqrestore(&chip->gpio_lock[index], flags); > - index = xgpio_index(chip, i); > - spin_lock_irqsave(&chip->gpio_lock[index], flags); > - } > - if (__test_and_clear_bit(i, mask)) { > - offset = xgpio_offset(chip, i); > - if (test_bit(i, bits)) > - chip->gpio_state[index] |= BIT(offset); > - else > - chip->gpio_state[index] &= ~BIT(offset); > - } > - } > - > - xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > - index * XGPIO_CHANNEL_OFFSET, chip->gpio_state[index]); > - > - spin_unlock_irqrestore(&chip->gpio_lock[index], flags); > + u32 *const state = chip->gpio_state; > + unsigned int *const width = chip->gpio_width; > + > + DECLARE_BITMAP(old, 64); > + DECLARE_BITMAP(new, 64); > + DECLARE_BITMAP(changed, 64); > + > + spin_lock_irqsave(&chip->gpio_lock[0], flags); > + spin_lock(&chip->gpio_lock[1]); > + > + bitmap_set_value(old, 64, state[0], width[0], 0); > + bitmap_set_value(old, 64, state[1], width[1], width[0]); > + bitmap_replace(new, old, bits, mask, gc->ngpio); > + > + bitmap_set_value(old, 64, state[0], 32, 0); > + bitmap_set_value(old, 64, state[1], 32, 32); > + state[0] = bitmap_get_value(new, 0, width[0]); > + state[1] = bitmap_get_value(new, width[0], width[1]); > + bitmap_set_value(new, 64, state[0], 32, 0); > + bitmap_set_value(new, 64, state[1], 32, 32); > + bitmap_xor(changed, old, new, 64); > + > + if (((u32 *)changed)[0]) > + xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET, > + state[0]); > + if (((u32 *)changed)[1]) > + xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > + XGPIO_CHANNEL_OFFSET, state[1]); > + > + spin_unlock(&chip->gpio_lock[1]); > + spin_unlock_irqrestore(&chip->gpio_lock[0], flags); > } > > /** > Srinivas N: Can you please test this code? Thanks, Michal
On Mon, Mar 8, 2021 at 8:13 AM Michal Simek <michal.simek@xilinx.com> wrote: > > > > On 3/6/21 3:06 PM, Syed Nayyar Waris wrote: > > This patch reimplements the xgpio_set_multiple() function in > > drivers/gpio/gpio-xilinx.c to use the new generic functions: > > bitmap_get_value() and bitmap_set_value(). The code is now simpler > > to read and understand. Moreover, instead of looping for each bit > > in xgpio_set_multiple() function, now we can check each channel at > > a time and save cycles. > > > > Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com> > > Cc: Michal Simek <michal.simek@xilinx.com> > > Signed-off-by: Syed Nayyar Waris <syednwaris@gmail.com> > > Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com> > > --- > > drivers/gpio/gpio-xilinx.c | 63 +++++++++++++++++++------------------- > > 1 file changed, 32 insertions(+), 31 deletions(-) > > > > diff --git a/drivers/gpio/gpio-xilinx.c b/drivers/gpio/gpio-xilinx.c > > index be539381fd82..8445e69cf37b 100644 > > --- a/drivers/gpio/gpio-xilinx.c > > +++ b/drivers/gpio/gpio-xilinx.c > > @@ -15,6 +15,7 @@ > > #include <linux/of_device.h> > > #include <linux/of_platform.h> > > #include <linux/slab.h> > > +#include "gpiolib.h" > > > > /* Register Offset Definitions */ > > #define XGPIO_DATA_OFFSET (0x0) /* Data register */ > > @@ -141,37 +142,37 @@ static void xgpio_set_multiple(struct gpio_chip *gc, unsigned long *mask, > > { > > unsigned long flags; > > struct xgpio_instance *chip = gpiochip_get_data(gc); > > - int index = xgpio_index(chip, 0); > > - int offset, i; > > - > > - spin_lock_irqsave(&chip->gpio_lock[index], flags); > > - > > - /* Write to GPIO signals */ > > - for (i = 0; i < gc->ngpio; i++) { > > - if (*mask == 0) > > - break; > > - /* Once finished with an index write it out to the register */ > > - if (index != xgpio_index(chip, i)) { > > - xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > > - index * XGPIO_CHANNEL_OFFSET, > > - chip->gpio_state[index]); > > - spin_unlock_irqrestore(&chip->gpio_lock[index], flags); > > - index = xgpio_index(chip, i); > > - spin_lock_irqsave(&chip->gpio_lock[index], flags); > > - } > > - if (__test_and_clear_bit(i, mask)) { > > - offset = xgpio_offset(chip, i); > > - if (test_bit(i, bits)) > > - chip->gpio_state[index] |= BIT(offset); > > - else > > - chip->gpio_state[index] &= ~BIT(offset); > > - } > > - } > > - > > - xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > > - index * XGPIO_CHANNEL_OFFSET, chip->gpio_state[index]); > > - > > - spin_unlock_irqrestore(&chip->gpio_lock[index], flags); > > + u32 *const state = chip->gpio_state; > > + unsigned int *const width = chip->gpio_width; > > + > > + DECLARE_BITMAP(old, 64); > > + DECLARE_BITMAP(new, 64); > > + DECLARE_BITMAP(changed, 64); > > + > > + spin_lock_irqsave(&chip->gpio_lock[0], flags); > > + spin_lock(&chip->gpio_lock[1]); > > + > > + bitmap_set_value(old, 64, state[0], width[0], 0); > > + bitmap_set_value(old, 64, state[1], width[1], width[0]); > > + bitmap_replace(new, old, bits, mask, gc->ngpio); > > + > > + bitmap_set_value(old, 64, state[0], 32, 0); > > + bitmap_set_value(old, 64, state[1], 32, 32); > > + state[0] = bitmap_get_value(new, 0, width[0]); > > + state[1] = bitmap_get_value(new, width[0], width[1]); > > + bitmap_set_value(new, 64, state[0], 32, 0); > > + bitmap_set_value(new, 64, state[1], 32, 32); > > + bitmap_xor(changed, old, new, 64); > > + > > + if (((u32 *)changed)[0]) > > + xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET, > > + state[0]); > > + if (((u32 *)changed)[1]) > > + xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > > + XGPIO_CHANNEL_OFFSET, state[1]); > > + > > + spin_unlock(&chip->gpio_lock[1]); > > + spin_unlock_irqrestore(&chip->gpio_lock[0], flags); > > } > > > > /** > > > > Srinivas N: Can you please test this code? > > Thanks, > Michal Hey, any chance of getting that Tested-by? Bart
On Sat, Mar 6, 2021 at 4:08 PM Syed Nayyar Waris <syednwaris@gmail.com> wrote: > > This patch reimplements the xgpio_set_multiple() function in > drivers/gpio/gpio-xilinx.c to use the new generic functions: > bitmap_get_value() and bitmap_set_value(). The code is now simpler > to read and understand. Moreover, instead of looping for each bit > in xgpio_set_multiple() function, now we can check each channel at > a time and save cycles. ... > + u32 *const state = chip->gpio_state; > + unsigned int *const width = chip->gpio_width; > + Extra blank line. > + DECLARE_BITMAP(old, 64); > + DECLARE_BITMAP(new, 64); > + DECLARE_BITMAP(changed, 64); > + spin_lock_irqsave(&chip->gpio_lock[0], flags); > + spin_lock(&chip->gpio_lock[1]); I understand why this is done at the top of the function in the original code. I do not understand why you put some operations under spin lock. Have you checked what each of these spin locks protects? Please check and try to lock as minimum as possible. > + bitmap_set_value(old, 64, state[0], width[0], 0); > + bitmap_set_value(old, 64, state[1], width[1], width[0]); > + bitmap_replace(new, old, bits, mask, gc->ngpio); > + > + bitmap_set_value(old, 64, state[0], 32, 0); > + bitmap_set_value(old, 64, state[1], 32, 32); > + state[0] = bitmap_get_value(new, 0, width[0]); > + state[1] = bitmap_get_value(new, width[0], width[1]); > + bitmap_set_value(new, 64, state[0], 32, 0); > + bitmap_set_value(new, 64, state[1], 32, 32); > + bitmap_xor(changed, old, new, 64); Original code and this is cryptic. Can you add a few comments explaining what is going on here? > + spin_unlock(&chip->gpio_lock[1]); > + spin_unlock_irqrestore(&chip->gpio_lock[0], flags); -- With Best Regards, Andy Shevchenko
On Sat, Mar 6, 2021 at 4:08 PM Syed Nayyar Waris <syednwaris@gmail.com> wrote: > + bitmap_set_value(old, 64, state[0], 32, 0); > + bitmap_set_value(old, 64, state[1], 32, 32); Isn't it effectively bitnap_from_arr32() ? > + bitmap_set_value(new, 64, state[0], 32, 0); > + bitmap_set_value(new, 64, state[1], 32, 32); Ditto. -- With Best Regards, Andy Shevchenko
On Fri, Mar 26, 2021 at 11:32 PM Andy Shevchenko <andy.shevchenko@gmail.com> wrote: > > On Sat, Mar 6, 2021 at 4:08 PM Syed Nayyar Waris <syednwaris@gmail.com> wrote: > > > + bitmap_set_value(old, 64, state[0], 32, 0); > > + bitmap_set_value(old, 64, state[1], 32, 32); > > Isn't it effectively bitnap_from_arr32() ? > > > + bitmap_set_value(new, 64, state[0], 32, 0); > > + bitmap_set_value(new, 64, state[1], 32, 32); > > Ditto. > > -- > With Best Regards, > Andy Shevchenko Hi Andy, With bitmap_set_value() we are also specifying the offset (or start) position too. so that the remainder of the array remains unaffected. I think it would not be feasible to use bitmap_from/to_arr32() here. Regards Syed Nayyar Waris
On Sat, Mar 27, 2021 at 09:29:26AM +0200, Andy Shevchenko wrote: > On Saturday, March 27, 2021, Syed Nayyar Waris <syednwaris@gmail.com> wrote: > > > On Fri, Mar 26, 2021 at 11:32 PM Andy Shevchenko > > <andy.shevchenko@gmail.com> wrote: > > > > > > On Sat, Mar 6, 2021 at 4:08 PM Syed Nayyar Waris <syednwaris@gmail.com> > > wrote: > > > > > > > + bitmap_set_value(old, 64, state[0], 32, 0); > > > > + bitmap_set_value(old, 64, state[1], 32, 32); > > > > > > Isn't it effectively bitnap_from_arr32() ? > > > > > > > + bitmap_set_value(new, 64, state[0], 32, 0); > > > > + bitmap_set_value(new, 64, state[1], 32, 32); > > > > > > Ditto. > > > > > > -- > > > With Best Regards, > > > Andy Shevchenko > > > > Hi Andy, > > > > With bitmap_set_value() we are also specifying the offset (or start) > > position too. so that the remainder of the array remains unaffected. I > > think it would not be feasible to use bitmap_from/to_arr32() here. > > > You have hard coded start and nbits parameters to 32. How is it not the > same? Would these four lines become something like this: bitmap_from_arr32(old, state, 64); ... bitmap_from_arr32(new, state, 64); Sincerely, William Breathitt Gray
On Sat, Mar 27, 2021 at 2:02 PM William Breathitt Gray <vilhelm.gray@gmail.com> wrote: > On Sat, Mar 27, 2021 at 09:29:26AM +0200, Andy Shevchenko wrote: > > On Saturday, March 27, 2021, Syed Nayyar Waris <syednwaris@gmail.com> wrote: > > > On Fri, Mar 26, 2021 at 11:32 PM Andy Shevchenko > > > <andy.shevchenko@gmail.com> wrote: > > > > On Sat, Mar 6, 2021 at 4:08 PM Syed Nayyar Waris <syednwaris@gmail.com> > > > wrote: > > > > > > > > > + bitmap_set_value(old, 64, state[0], 32, 0); > > > > > + bitmap_set_value(old, 64, state[1], 32, 32); > > > > > > > > Isn't it effectively bitnap_from_arr32() ? > > > > > > > > > + bitmap_set_value(new, 64, state[0], 32, 0); > > > > > + bitmap_set_value(new, 64, state[1], 32, 32); > > > > > > > > Ditto. > > > With bitmap_set_value() we are also specifying the offset (or start) > > > position too. so that the remainder of the array remains unaffected. I > > > think it would not be feasible to use bitmap_from/to_arr32() here. > > > > > > You have hard coded start and nbits parameters to 32. How is it not the > > same? > > Would these four lines become something like this: > > bitmap_from_arr32(old, state, 64); > ... > bitmap_from_arr32(new, state, 64); This is my understanding, but I might miss something. I mean driver specifics that make my proposal incorrect. -- With Best Regards, Andy Shevchenko
On Wed, Mar 31, 2021 at 8:56 PM Srinivas Neeli <sneeli@xilinx.com> wrote: > > Hi, > > > -----Original Message----- > > From: Bartosz Golaszewski <bgolaszewski@baylibre.com> > > Sent: Friday, March 26, 2021 10:58 PM > > To: Michal Simek <michals@xilinx.com> > > Cc: Syed Nayyar Waris <syednwaris@gmail.com>; Srinivas Neeli > > <sneeli@xilinx.com>; Andy Shevchenko > > <andriy.shevchenko@linux.intel.com>; William Breathitt Gray > > <vilhelm.gray@gmail.com>; Arnd Bergmann <arnd@arndb.de>; Robert > > Richter <rrichter@marvell.com>; Linus Walleij <linus.walleij@linaro.org>; > > Masahiro Yamada <yamada.masahiro@socionext.com>; Andrew Morton > > <akpm@linux-foundation.org>; Zhang Rui <rui.zhang@intel.com>; Daniel > > Lezcano <daniel.lezcano@linaro.org>; Amit Kucheria > > <amit.kucheria@verdurent.com>; Linux-Arch <linux-arch@vger.kernel.org>; > > linux-gpio <linux-gpio@vger.kernel.org>; LKML <linux- > > kernel@vger.kernel.org>; arm-soc <linux-arm-kernel@lists.infradead.org>; > > linux-pm <linux-pm@vger.kernel.org>; Srinivas Goud <sgoud@xilinx.com> > > Subject: Re: [PATCH v3 3/3] gpio: xilinx: Utilize generic bitmap_get_value and > > _set_value > > > > On Mon, Mar 8, 2021 at 8:13 AM Michal Simek <michal.simek@xilinx.com> > > wrote: > > > > > > > > > > > > On 3/6/21 3:06 PM, Syed Nayyar Waris wrote: > > > > This patch reimplements the xgpio_set_multiple() function in > > > > drivers/gpio/gpio-xilinx.c to use the new generic functions: > > > > bitmap_get_value() and bitmap_set_value(). The code is now simpler > > > > to read and understand. Moreover, instead of looping for each bit in > > > > xgpio_set_multiple() function, now we can check each channel at a > > > > time and save cycles. > > > > > > > > Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com> > > > > Cc: Michal Simek <michal.simek@xilinx.com> > > > > Signed-off-by: Syed Nayyar Waris <syednwaris@gmail.com> > > > > Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com> > > > > --- > > > > drivers/gpio/gpio-xilinx.c | 63 > > > > +++++++++++++++++++------------------- > > > > 1 file changed, 32 insertions(+), 31 deletions(-) > > > > > > > > diff --git a/drivers/gpio/gpio-xilinx.c b/drivers/gpio/gpio-xilinx.c > > > > index be539381fd82..8445e69cf37b 100644 > > > > --- a/drivers/gpio/gpio-xilinx.c > > > > +++ b/drivers/gpio/gpio-xilinx.c > > > > @@ -15,6 +15,7 @@ > > > > #include <linux/of_device.h> > > > > #include <linux/of_platform.h> > > > > #include <linux/slab.h> > > > > +#include "gpiolib.h" > > > > > > > > /* Register Offset Definitions */ > > > > #define XGPIO_DATA_OFFSET (0x0) /* Data register */ > > > > @@ -141,37 +142,37 @@ static void xgpio_set_multiple(struct > > > > gpio_chip *gc, unsigned long *mask, { > > > > unsigned long flags; > > > > struct xgpio_instance *chip = gpiochip_get_data(gc); > > > > - int index = xgpio_index(chip, 0); > > > > - int offset, i; > > > > - > > > > - spin_lock_irqsave(&chip->gpio_lock[index], flags); > > > > - > > > > - /* Write to GPIO signals */ > > > > - for (i = 0; i < gc->ngpio; i++) { > > > > - if (*mask == 0) > > > > - break; > > > > - /* Once finished with an index write it out to the register */ > > > > - if (index != xgpio_index(chip, i)) { > > > > - xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > > > > - index * XGPIO_CHANNEL_OFFSET, > > > > - chip->gpio_state[index]); > > > > - spin_unlock_irqrestore(&chip->gpio_lock[index], flags); > > > > - index = xgpio_index(chip, i); > > > > - spin_lock_irqsave(&chip->gpio_lock[index], flags); > > > > - } > > > > - if (__test_and_clear_bit(i, mask)) { > > > > - offset = xgpio_offset(chip, i); > > > > - if (test_bit(i, bits)) > > > > - chip->gpio_state[index] |= BIT(offset); > > > > - else > > > > - chip->gpio_state[index] &= ~BIT(offset); > > > > - } > > > > - } > > > > - > > > > - xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > > > > - index * XGPIO_CHANNEL_OFFSET, chip->gpio_state[index]); > > > > - > > > > - spin_unlock_irqrestore(&chip->gpio_lock[index], flags); > > > > + u32 *const state = chip->gpio_state; > > > > + unsigned int *const width = chip->gpio_width; > > > > + > > > > + DECLARE_BITMAP(old, 64); > > > > + DECLARE_BITMAP(new, 64); > > > > + DECLARE_BITMAP(changed, 64); > > > > + > > > > + spin_lock_irqsave(&chip->gpio_lock[0], flags); > > > > + spin_lock(&chip->gpio_lock[1]); > > > > + > > > > + bitmap_set_value(old, 64, state[0], width[0], 0); > > > > + bitmap_set_value(old, 64, state[1], width[1], width[0]); > > > > + bitmap_replace(new, old, bits, mask, gc->ngpio); > > > > + > > > > + bitmap_set_value(old, 64, state[0], 32, 0); > > > > + bitmap_set_value(old, 64, state[1], 32, 32); > > > > + state[0] = bitmap_get_value(new, 0, width[0]); > > > > + state[1] = bitmap_get_value(new, width[0], width[1]); > > > > + bitmap_set_value(new, 64, state[0], 32, 0); > > > > + bitmap_set_value(new, 64, state[1], 32, 32); > > > > + bitmap_xor(changed, old, new, 64); > > > > + > > > > + if (((u32 *)changed)[0]) > > > > + xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET, > > > > + state[0]); > > > > + if (((u32 *)changed)[1]) > > > > + xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > > > > + XGPIO_CHANNEL_OFFSET, state[1]); > > > > + > > > > + spin_unlock(&chip->gpio_lock[1]); > > > > + spin_unlock_irqrestore(&chip->gpio_lock[0], flags); > > > > } > > > > > > > > /** > > > > > > > > > > Srinivas N: Can you please test this code? > > > > > > Thanks, > > > Michal > > > > Hey, any chance of getting that Tested-by? > I tested patches with few modifications in code (spin_lock handling and merge conflict). > functionality wise it's working fine. > > > > > Bart Hi Bartosz, May I please know the URL of the tree that you are using. I had been using the tree below for submitting this patchset on GPIO to you. https://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio.git I think I am using the wrong tree. On which tree should I base my patches on for my next (v4) submission? Should I use the tree below? : https://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux.git Regards Syed Nayyar Waris
On Thu, Apr 1, 2021 at 1:16 PM Syed Nayyar Waris <syednwaris@gmail.com> wrote: > > On Wed, Mar 31, 2021 at 8:56 PM Srinivas Neeli <sneeli@xilinx.com> wrote: > > > > Hi, > > > > > -----Original Message----- > > > From: Bartosz Golaszewski <bgolaszewski@baylibre.com> > > > Sent: Friday, March 26, 2021 10:58 PM > > > To: Michal Simek <michals@xilinx.com> > > > Cc: Syed Nayyar Waris <syednwaris@gmail.com>; Srinivas Neeli > > > <sneeli@xilinx.com>; Andy Shevchenko > > > <andriy.shevchenko@linux.intel.com>; William Breathitt Gray > > > <vilhelm.gray@gmail.com>; Arnd Bergmann <arnd@arndb.de>; Robert > > > Richter <rrichter@marvell.com>; Linus Walleij <linus.walleij@linaro.org>; > > > Masahiro Yamada <yamada.masahiro@socionext.com>; Andrew Morton > > > <akpm@linux-foundation.org>; Zhang Rui <rui.zhang@intel.com>; Daniel > > > Lezcano <daniel.lezcano@linaro.org>; Amit Kucheria > > > <amit.kucheria@verdurent.com>; Linux-Arch <linux-arch@vger.kernel.org>; > > > linux-gpio <linux-gpio@vger.kernel.org>; LKML <linux- > > > kernel@vger.kernel.org>; arm-soc <linux-arm-kernel@lists.infradead.org>; > > > linux-pm <linux-pm@vger.kernel.org>; Srinivas Goud <sgoud@xilinx.com> > > > Subject: Re: [PATCH v3 3/3] gpio: xilinx: Utilize generic bitmap_get_value and > > > _set_value > > > > > > On Mon, Mar 8, 2021 at 8:13 AM Michal Simek <michal.simek@xilinx.com> > > > wrote: > > > > > > > > > > > > > > > > On 3/6/21 3:06 PM, Syed Nayyar Waris wrote: > > > > > This patch reimplements the xgpio_set_multiple() function in > > > > > drivers/gpio/gpio-xilinx.c to use the new generic functions: > > > > > bitmap_get_value() and bitmap_set_value(). The code is now simpler > > > > > to read and understand. Moreover, instead of looping for each bit in > > > > > xgpio_set_multiple() function, now we can check each channel at a > > > > > time and save cycles. > > > > > > > > > > Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com> > > > > > Cc: Michal Simek <michal.simek@xilinx.com> > > > > > Signed-off-by: Syed Nayyar Waris <syednwaris@gmail.com> > > > > > Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com> > > > > > --- > > > > > drivers/gpio/gpio-xilinx.c | 63 > > > > > +++++++++++++++++++------------------- > > > > > 1 file changed, 32 insertions(+), 31 deletions(-) > > > > > > > > > > diff --git a/drivers/gpio/gpio-xilinx.c b/drivers/gpio/gpio-xilinx.c > > > > > index be539381fd82..8445e69cf37b 100644 > > > > > --- a/drivers/gpio/gpio-xilinx.c > > > > > +++ b/drivers/gpio/gpio-xilinx.c > > > > > @@ -15,6 +15,7 @@ > > > > > #include <linux/of_device.h> > > > > > #include <linux/of_platform.h> > > > > > #include <linux/slab.h> > > > > > +#include "gpiolib.h" > > > > > > > > > > /* Register Offset Definitions */ > > > > > #define XGPIO_DATA_OFFSET (0x0) /* Data register */ > > > > > @@ -141,37 +142,37 @@ static void xgpio_set_multiple(struct > > > > > gpio_chip *gc, unsigned long *mask, { > > > > > unsigned long flags; > > > > > struct xgpio_instance *chip = gpiochip_get_data(gc); > > > > > - int index = xgpio_index(chip, 0); > > > > > - int offset, i; > > > > > - > > > > > - spin_lock_irqsave(&chip->gpio_lock[index], flags); > > > > > - > > > > > - /* Write to GPIO signals */ > > > > > - for (i = 0; i < gc->ngpio; i++) { > > > > > - if (*mask == 0) > > > > > - break; > > > > > - /* Once finished with an index write it out to the register */ > > > > > - if (index != xgpio_index(chip, i)) { > > > > > - xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > > > > > - index * XGPIO_CHANNEL_OFFSET, > > > > > - chip->gpio_state[index]); > > > > > - spin_unlock_irqrestore(&chip->gpio_lock[index], flags); > > > > > - index = xgpio_index(chip, i); > > > > > - spin_lock_irqsave(&chip->gpio_lock[index], flags); > > > > > - } > > > > > - if (__test_and_clear_bit(i, mask)) { > > > > > - offset = xgpio_offset(chip, i); > > > > > - if (test_bit(i, bits)) > > > > > - chip->gpio_state[index] |= BIT(offset); > > > > > - else > > > > > - chip->gpio_state[index] &= ~BIT(offset); > > > > > - } > > > > > - } > > > > > - > > > > > - xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > > > > > - index * XGPIO_CHANNEL_OFFSET, chip->gpio_state[index]); > > > > > - > > > > > - spin_unlock_irqrestore(&chip->gpio_lock[index], flags); > > > > > + u32 *const state = chip->gpio_state; > > > > > + unsigned int *const width = chip->gpio_width; > > > > > + > > > > > + DECLARE_BITMAP(old, 64); > > > > > + DECLARE_BITMAP(new, 64); > > > > > + DECLARE_BITMAP(changed, 64); > > > > > + > > > > > + spin_lock_irqsave(&chip->gpio_lock[0], flags); > > > > > + spin_lock(&chip->gpio_lock[1]); > > > > > + > > > > > + bitmap_set_value(old, 64, state[0], width[0], 0); > > > > > + bitmap_set_value(old, 64, state[1], width[1], width[0]); > > > > > + bitmap_replace(new, old, bits, mask, gc->ngpio); > > > > > + > > > > > + bitmap_set_value(old, 64, state[0], 32, 0); > > > > > + bitmap_set_value(old, 64, state[1], 32, 32); > > > > > + state[0] = bitmap_get_value(new, 0, width[0]); > > > > > + state[1] = bitmap_get_value(new, width[0], width[1]); > > > > > + bitmap_set_value(new, 64, state[0], 32, 0); > > > > > + bitmap_set_value(new, 64, state[1], 32, 32); > > > > > + bitmap_xor(changed, old, new, 64); > > > > > + > > > > > + if (((u32 *)changed)[0]) > > > > > + xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET, > > > > > + state[0]); > > > > > + if (((u32 *)changed)[1]) > > > > > + xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET + > > > > > + XGPIO_CHANNEL_OFFSET, state[1]); > > > > > + > > > > > + spin_unlock(&chip->gpio_lock[1]); > > > > > + spin_unlock_irqrestore(&chip->gpio_lock[0], flags); > > > > > } > > > > > > > > > > /** > > > > > > > > > > > > > Srinivas N: Can you please test this code? > > > > > > > > Thanks, > > > > Michal > > > > > > Hey, any chance of getting that Tested-by? > > I tested patches with few modifications in code (spin_lock handling and merge conflict). > > functionality wise it's working fine. > > > > > > > > Bart > > Hi Bartosz, > > May I please know the URL of the tree that you are using. I had been > using the tree below for submitting this patchset on GPIO to you. > https://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio.git > > I think I am using the wrong tree. On which tree should I base my > patches on for my next (v4) submission? Should I use the tree below? > : > https://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux.git > > Regards > Syed Nayyar Waris Yes this is the one. Please address new issues raised by reviewers. Bart
On Mon, Mar 29, 2021 at 8:54 PM Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote: > > On Sat, Mar 06, 2021 at 07:36:30PM +0530, Syed Nayyar Waris wrote: > > This patch reimplements the xgpio_set_multiple() function in > > drivers/gpio/gpio-xilinx.c to use the new generic functions: > > bitmap_get_value() and bitmap_set_value(). The code is now simpler > > to read and understand. Moreover, instead of looping for each bit > > in xgpio_set_multiple() function, now we can check each channel at > > a time and save cycles. > > ... > > > + u32 *const state = chip->gpio_state; > > Looking at this... What's the point of the const here? > > Am I right that this tells: pointer is a const, while the data underneath > can be modified? > > > + unsigned int *const width = chip->gpio_width; > > Ditto. > > Putting const:s here and there for sake of the const is not good practice. > It makes code harder to read. > > -- > With Best Regards, > Andy Shevchenko > Okay. I will incorporate your comments in my next submission. Thank You. Regards Syed Nayyar Waris
On Sat, Mar 27, 2021 at 10:05 PM Andy Shevchenko <andy.shevchenko@gmail.com> wrote: > > On Sat, Mar 27, 2021 at 2:02 PM William Breathitt Gray > <vilhelm.gray@gmail.com> wrote: > > On Sat, Mar 27, 2021 at 09:29:26AM +0200, Andy Shevchenko wrote: > > > On Saturday, March 27, 2021, Syed Nayyar Waris <syednwaris@gmail.com> wrote: > > > > On Fri, Mar 26, 2021 at 11:32 PM Andy Shevchenko > > > > <andy.shevchenko@gmail.com> wrote: > > > > > On Sat, Mar 6, 2021 at 4:08 PM Syed Nayyar Waris <syednwaris@gmail.com> > > > > wrote: > > > > > > > > > > > + bitmap_set_value(old, 64, state[0], 32, 0); > > > > > > + bitmap_set_value(old, 64, state[1], 32, 32); > > > > > > > > > > Isn't it effectively bitnap_from_arr32() ? > > > > > > > > > > > + bitmap_set_value(new, 64, state[0], 32, 0); > > > > > > + bitmap_set_value(new, 64, state[1], 32, 32); > > > > > > > > > > Ditto. > > > > > With bitmap_set_value() we are also specifying the offset (or start) > > > > position too. so that the remainder of the array remains unaffected. I > > > > think it would not be feasible to use bitmap_from/to_arr32() here. > > > > > > > > > You have hard coded start and nbits parameters to 32. How is it not the > > > same? > > > > Would these four lines become something like this: > > > > bitmap_from_arr32(old, state, 64); > > ... > > bitmap_from_arr32(new, state, 64); > > This is my understanding, but I might miss something. I mean driver > specifics that make my proposal incorrect. > > -- > With Best Regards, > Andy Shevchenko I initially (incorrectly) thought that all of the bitmap_set_value() statements have to be replaced. But now I realised, only those specific bitmap_set_value() calls containing 32 bits width have to replaced. I will incorporate the above review comments in my next v4 submission. Regards Syed Nayyar Waris
On Fri, Mar 26, 2021 at 11:27 PM Andy Shevchenko <andy.shevchenko@gmail.com> wrote: > > On Sat, Mar 6, 2021 at 4:08 PM Syed Nayyar Waris <syednwaris@gmail.com> wrote: > > > > This patch reimplements the xgpio_set_multiple() function in > > drivers/gpio/gpio-xilinx.c to use the new generic functions: > > bitmap_get_value() and bitmap_set_value(). The code is now simpler > > to read and understand. Moreover, instead of looping for each bit > > in xgpio_set_multiple() function, now we can check each channel at > > a time and save cycles. > > ... > > > + u32 *const state = chip->gpio_state; > > + unsigned int *const width = chip->gpio_width; > > > + > > Extra blank line. > > > + DECLARE_BITMAP(old, 64); > > + DECLARE_BITMAP(new, 64); > > + DECLARE_BITMAP(changed, 64); > > > + spin_lock_irqsave(&chip->gpio_lock[0], flags); > > + spin_lock(&chip->gpio_lock[1]); > > I understand why this is done at the top of the function in the original code. > I do not understand why you put some operations under spin lock. > > Have you checked what each of these spin locks protects? > Please check and try to lock as minimum as possible. > > > + bitmap_set_value(old, 64, state[0], width[0], 0); > > + bitmap_set_value(old, 64, state[1], width[1], width[0]); > > + bitmap_replace(new, old, bits, mask, gc->ngpio); > > + > > + bitmap_set_value(old, 64, state[0], 32, 0); > > + bitmap_set_value(old, 64, state[1], 32, 32); > > + state[0] = bitmap_get_value(new, 0, width[0]); > > + state[1] = bitmap_get_value(new, width[0], width[1]); > > + bitmap_set_value(new, 64, state[0], 32, 0); > > + bitmap_set_value(new, 64, state[1], 32, 32); > > + bitmap_xor(changed, old, new, 64); > > Original code and this is cryptic. Can you add a few comments > explaining what is going on here? > > > + spin_unlock(&chip->gpio_lock[1]); > > + spin_unlock_irqrestore(&chip->gpio_lock[0], flags); > > -- > With Best Regards, > Andy Shevchenko Have removed the extra line and added comments. Regarding locking - I see that now there is just a single lock available instead of 2 locks. Have made necessary changes. Thanks https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/drivers/gpio/gpio-xilinx.c?id=37ef334680800263b32bb96a5156a4b47f0244a2 Regards Syed Nayyar Waris
On Mon, Mar 29, 2021 at 8:54 PM Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote: > > On Sat, Mar 06, 2021 at 07:36:30PM +0530, Syed Nayyar Waris wrote: > > This patch reimplements the xgpio_set_multiple() function in > > drivers/gpio/gpio-xilinx.c to use the new generic functions: > > bitmap_get_value() and bitmap_set_value(). The code is now simpler > > to read and understand. Moreover, instead of looping for each bit > > in xgpio_set_multiple() function, now we can check each channel at > > a time and save cycles. > > ... > > > + u32 *const state = chip->gpio_state; > > Looking at this... What's the point of the const here? > > Am I right that this tells: pointer is a const, while the data underneath > can be modified? Yes you are right and the data underneath can be modified. I have removed the 'const' in v4 > > > + unsigned int *const width = chip->gpio_width; > > Ditto. > > Putting const:s here and there for sake of the const is not good practice. > It makes code harder to read. Okay. > > -- > With Best Regards, > Andy Shevchenko > >
Hello Bartosz, Since this patchset primarily affects GPIO drivers, would you like to pick it up through your GPIO tree? This patchset introduces a new generic version of for_each_set_nbits. The previous version of for_each_set_clump8 used a fixed size 8-bit clump, but the new generic version can work with clump of any size but less than or equal to BITS_PER_LONG. The patchset utilizes the new macro in several GPIO drivers. The earlier 8-bit for_each_set_clump8 facilitated a for-loop syntax that iterates over a memory region entire groups of set bits at a time. For example, suppose you would like to iterate over a 32-bit integer 8 bits at a time, skipping over 8-bit groups with no set bit, where XXXXXXXX represents the current 8-bit group: Example: 10111110 00000000 11111111 00110011 First loop: 10111110 00000000 11111111 XXXXXXXX Second loop: 10111110 00000000 XXXXXXXX 00110011 Third loop: XXXXXXXX 00000000 11111111 00110011 Each iteration of the loop returns the next 8-bit group that has at least one set bit. But with the new for_each_set_nbits the clump size can be different from 8 bits. Moreover, the clump can be split at word boundary in situations where word size is not multiple of clump size. Following are examples showing the working of new macro for clump sizes of 24 bits and 6 bits. Example 1: clump size: 24 bits, Number of clumps (or ports): 10 bitmap stores the bit information from where successive clumps are retrieved. /* bitmap memory region */ 0x00aa0000ff000000; /* Most significant bits */ 0xaaaaaa0000ff0000; 0x000000aa000000aa; 0xbbbbabcdeffedcba; /* Least significant bits */ Different iterations of for_each_set_nbits:- 'offset' is the bit position and 'clump' is the 24 bit clump from the above bitmap. Iteration first: offset: 0 clump: 0xfedcba Iteration second: offset: 24 clump: 0xabcdef Iteration third: offset: 48 clump: 0xaabbbb Iteration fourth: offset: 96 clump: 0xaa Iteration fifth: offset: 144 clump: 0xff Iteration sixth: offset: 168 clump: 0xaaaaaa Iteration seventh: offset: 216 clump: 0xff Loop breaks because in the end the remaining bits (0x00aa) size was less than clump size of 24 bits. In above example it can be seen that in iteration third, the 24 bit clump that was retrieved was split between bitmap[0] and bitmap[1]. This example also shows that 24 bit zeroes if present in between, were skipped (preserving the previous for_each_set_macro8 behaviour). Example 2: clump size = 6 bits, Number of clumps (or ports) = 3. /* bitmap memory region */ 0x00aa0000ff000000; /* Most significant bits */ 0xaaaaaa0000ff0000; 0x0f00000000000000; 0x0000000000000ac0; /* Least significant bits */ Different iterations of for_each_set_nbits: 'offset' is the bit position and 'clump' is the 6 bit clump from the above bitmap. Iteration first: offset: 6 clump: 0x2b Loop breaks because 6 * 3 = 18 bits traversed in bitmap. Here 6 * 3 is clump size * no. of clumps. Changes in v3: - [Patch 1/3]: Rename for_each_set_clump to for_each_set_nbits. - [Patch 1/3]: Shift function definitions outside 'ifdef CONFIG_DEBUG_FS' macro guard to resolve build (linking) error in xilinx Patch[3/3]. - [Patch 2/3]: Rename for_each_set_clump to for_each_set_nbits. Changes in v2: - [Patch 1/3]: Shift the macros and related functions to gpiolib inside gpio/. Reduce the visibilty of 'for_each_set_clump' to gpio. - [Patch 1/3]: Remove __builtin_unreachable and simply use return statement. - Remove tests from lib/test_bitmap.c as 'for_each_set_clump' is now localised inside gpio/ only. Syed Nayyar Waris (3): gpiolib: Introduce the for_each_set_nbits macro gpio: thunderx: Utilize for_each_set_nbits macro gpio: xilinx: Utilize generic bitmap_get_value and _set_value drivers/gpio/gpio-thunderx.c | 13 ++++-- drivers/gpio/gpio-xilinx.c | 63 ++++++++++++------------- drivers/gpio/gpiolib.c | 90 ++++++++++++++++++++++++++++++++++++ drivers/gpio/gpiolib.h | 28 +++++++++++ 4 files changed, 158 insertions(+), 36 deletions(-) base-commit: e71ba9452f0b5b2e8dc8aa5445198cd9214a6a62