From: kernel-TqfNSX0MhmxHKSADF0wUEw@public.gmane.org To: Mark Brown <broonie-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>, Eric Anholt <eric-WhKQ6XTQaPysTnJN9+BGXg@public.gmane.org>, Stefan Wahren <stefan.wahren-eS4NqCHxEME@public.gmane.org>, linux-spi-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-rpi-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org Subject: [PATCH 1/5] spi: core: allow defining time that cs is deasserted Date: Sat, 23 Feb 2019 08:49:48 +0000 [thread overview] Message-ID: <20190223084952.14758-2-kernel@martin.sperl.org> (raw) In-Reply-To: <20190223084952.14758-1-kernel-TqfNSX0MhmxHKSADF0wUEw@public.gmane.org> From: Martin Sperl <kernel-TqfNSX0MhmxHKSADF0wUEw@public.gmane.org> For some SPI devices that support speed_hz > 1MHz the default 10 us delay when cs_change = 1 is typically way to long and may result in poor spi bus utilization. This patch makes it possible to control the delay at micro or nano second resolution on a per spi_transfer basis. It even allows an "as fast as possible" mode with: xfer.cs_change_delay_unit = SPI_DELAY_UNIT_NSECS; xfer.cs_change_delay = 0; The delay code is shared between delay_usecs and cs_change_delay for consistency and reuse, so in the future this change_delay_unit could also apply to delay_usec as well. Note that on slower SOCs/CPU actually reaching ns deasserts on cs is not realistic as the gpio overhead alone (without any delays added ) may already leave cs deasserted for more than 1us - at least on a raspberry pi. But at the very least this way we can keep it as short as possible. Signed-off-by: Martin Sperl <kernel-TqfNSX0MhmxHKSADF0wUEw@public.gmane.org> --- drivers/spi/spi.c | 59 ++++++++++++++++++++++++++++++++++++++++--------- include/linux/spi/spi.h | 7 ++++++ 2 files changed, 56 insertions(+), 10 deletions(-) diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c index 9a7def7c3237..90659eb29660 100644 --- a/drivers/spi/spi.c +++ b/drivers/spi/spi.c @@ -1071,6 +1071,52 @@ static int spi_transfer_wait(struct spi_controller *ctlr, return 0; } +static void _spi_transfer_delay_ns(u32 ns) +{ + if (!ns) + return; + if (ns <= 1000) { + ndelay(ns); + } else { + u32 us = DIV_ROUND_UP(ns, 1000); + + if (us <= 10) + udelay(us); + else + usleep_range(us, us + DIV_ROUND_UP(us, 10)); + } +} + +static void _spi_transfer_cs_change_delay(struct spi_message *msg, + struct spi_transfer *xfer) +{ + u32 delay = xfer->cs_change_delay; + u32 unit = xfer->cs_change_delay_unit; + + /* return early on "fast" mode - for everything but USECS */ + if (!delay && unit != SPI_DELAY_UNIT_USECS) + return; + + switch (unit) { + case SPI_DELAY_UNIT_USECS: + /* for compatibility use default of 10us */ + if (!delay) + delay = 10000; + else + delay *= 1000; + break; + case SPI_DELAY_UNIT_NSECS: /* nothing to do here */ + break; + default: + dev_err_once(&msg->spi->dev, + "Use of unsupported delay unit %i, using default of 10us\n", + xfer->cs_change_delay_unit); + delay = 10000; + } + /* now sleep for the requested amount of time */ + _spi_transfer_delay_ns(delay); +} + /* * spi_transfer_one_message - Default implementation of transfer_one_message() * @@ -1129,14 +1175,8 @@ static int spi_transfer_one_message(struct spi_controller *ctlr, if (msg->status != -EINPROGRESS) goto out; - if (xfer->delay_usecs) { - u16 us = xfer->delay_usecs; - - if (us <= 10) - udelay(us); - else - usleep_range(us, us + DIV_ROUND_UP(us, 10)); - } + if (xfer->delay_usecs) + _spi_transfer_delay_ns(xfer->delay_usecs * 1000); if (xfer->cs_change) { if (list_is_last(&xfer->transfer_list, @@ -1144,7 +1184,7 @@ static int spi_transfer_one_message(struct spi_controller *ctlr, keep_cs = true; } else { spi_set_cs(msg->spi, false); - udelay(10); + _spi_transfer_cs_change_delay(msg, xfer); spi_set_cs(msg->spi, true); } } @@ -3642,4 +3682,3 @@ static int __init spi_init(void) * include needing to have boardinfo data structures be much more public. */ postcore_initcall(spi_init); - diff --git a/include/linux/spi/spi.h b/include/linux/spi/spi.h index 314d922ca607..f2ce1fb403ef 100644 --- a/include/linux/spi/spi.h +++ b/include/linux/spi/spi.h @@ -703,6 +703,9 @@ extern void spi_res_release(struct spi_controller *ctlr, * @bits_per_word: select a bits_per_word other than the device default * for this transfer. If 0 the default (from @spi_device) is used. * @cs_change: affects chipselect after this transfer completes + * @cs_change_delay: delay between cs deassert and assert when + * @cs_change is set and @spi_transfer is not the last in @spi_message + * @cs_change_delay_unit: unit of cs_change_delay * @delay_usecs: microseconds to delay after this transfer before * (optionally) changing the chipselect status, then starting * the next transfer or completing this @spi_message. @@ -789,6 +792,10 @@ struct spi_transfer { #define SPI_NBITS_QUAD 0x04 /* 4bits transfer */ u8 bits_per_word; u16 delay_usecs; + u16 cs_change_delay; + u8 cs_change_delay_unit; +#define SPI_DELAY_UNIT_USECS 0 +#define SPI_DELAY_UNIT_NSECS 1 u32 speed_hz; u16 word_delay; -- 2.11.0
WARNING: multiple messages have this Message-ID (diff)
From: kernel@martin.sperl.org To: Mark Brown <broonie@kernel.org>, Eric Anholt <eric@anholt.net>, Stefan Wahren <stefan.wahren@i2se.com>, linux-spi@vger.kernel.org, linux-rpi-kernel@lists.infradead.org, linux-arm-kernel@lists.infradead.org Cc: Martin Sperl <kernel@martin.sperl.org> Subject: [PATCH 1/5] spi: core: allow defining time that cs is deasserted Date: Sat, 23 Feb 2019 08:49:48 +0000 [thread overview] Message-ID: <20190223084952.14758-2-kernel@martin.sperl.org> (raw) In-Reply-To: <20190223084952.14758-1-kernel@martin.sperl.org> From: Martin Sperl <kernel@martin.sperl.org> For some SPI devices that support speed_hz > 1MHz the default 10 us delay when cs_change = 1 is typically way to long and may result in poor spi bus utilization. This patch makes it possible to control the delay at micro or nano second resolution on a per spi_transfer basis. It even allows an "as fast as possible" mode with: xfer.cs_change_delay_unit = SPI_DELAY_UNIT_NSECS; xfer.cs_change_delay = 0; The delay code is shared between delay_usecs and cs_change_delay for consistency and reuse, so in the future this change_delay_unit could also apply to delay_usec as well. Note that on slower SOCs/CPU actually reaching ns deasserts on cs is not realistic as the gpio overhead alone (without any delays added ) may already leave cs deasserted for more than 1us - at least on a raspberry pi. But at the very least this way we can keep it as short as possible. Signed-off-by: Martin Sperl <kernel@martin.sperl.org> --- drivers/spi/spi.c | 59 ++++++++++++++++++++++++++++++++++++++++--------- include/linux/spi/spi.h | 7 ++++++ 2 files changed, 56 insertions(+), 10 deletions(-) diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c index 9a7def7c3237..90659eb29660 100644 --- a/drivers/spi/spi.c +++ b/drivers/spi/spi.c @@ -1071,6 +1071,52 @@ static int spi_transfer_wait(struct spi_controller *ctlr, return 0; } +static void _spi_transfer_delay_ns(u32 ns) +{ + if (!ns) + return; + if (ns <= 1000) { + ndelay(ns); + } else { + u32 us = DIV_ROUND_UP(ns, 1000); + + if (us <= 10) + udelay(us); + else + usleep_range(us, us + DIV_ROUND_UP(us, 10)); + } +} + +static void _spi_transfer_cs_change_delay(struct spi_message *msg, + struct spi_transfer *xfer) +{ + u32 delay = xfer->cs_change_delay; + u32 unit = xfer->cs_change_delay_unit; + + /* return early on "fast" mode - for everything but USECS */ + if (!delay && unit != SPI_DELAY_UNIT_USECS) + return; + + switch (unit) { + case SPI_DELAY_UNIT_USECS: + /* for compatibility use default of 10us */ + if (!delay) + delay = 10000; + else + delay *= 1000; + break; + case SPI_DELAY_UNIT_NSECS: /* nothing to do here */ + break; + default: + dev_err_once(&msg->spi->dev, + "Use of unsupported delay unit %i, using default of 10us\n", + xfer->cs_change_delay_unit); + delay = 10000; + } + /* now sleep for the requested amount of time */ + _spi_transfer_delay_ns(delay); +} + /* * spi_transfer_one_message - Default implementation of transfer_one_message() * @@ -1129,14 +1175,8 @@ static int spi_transfer_one_message(struct spi_controller *ctlr, if (msg->status != -EINPROGRESS) goto out; - if (xfer->delay_usecs) { - u16 us = xfer->delay_usecs; - - if (us <= 10) - udelay(us); - else - usleep_range(us, us + DIV_ROUND_UP(us, 10)); - } + if (xfer->delay_usecs) + _spi_transfer_delay_ns(xfer->delay_usecs * 1000); if (xfer->cs_change) { if (list_is_last(&xfer->transfer_list, @@ -1144,7 +1184,7 @@ static int spi_transfer_one_message(struct spi_controller *ctlr, keep_cs = true; } else { spi_set_cs(msg->spi, false); - udelay(10); + _spi_transfer_cs_change_delay(msg, xfer); spi_set_cs(msg->spi, true); } } @@ -3642,4 +3682,3 @@ static int __init spi_init(void) * include needing to have boardinfo data structures be much more public. */ postcore_initcall(spi_init); - diff --git a/include/linux/spi/spi.h b/include/linux/spi/spi.h index 314d922ca607..f2ce1fb403ef 100644 --- a/include/linux/spi/spi.h +++ b/include/linux/spi/spi.h @@ -703,6 +703,9 @@ extern void spi_res_release(struct spi_controller *ctlr, * @bits_per_word: select a bits_per_word other than the device default * for this transfer. If 0 the default (from @spi_device) is used. * @cs_change: affects chipselect after this transfer completes + * @cs_change_delay: delay between cs deassert and assert when + * @cs_change is set and @spi_transfer is not the last in @spi_message + * @cs_change_delay_unit: unit of cs_change_delay * @delay_usecs: microseconds to delay after this transfer before * (optionally) changing the chipselect status, then starting * the next transfer or completing this @spi_message. @@ -789,6 +792,10 @@ struct spi_transfer { #define SPI_NBITS_QUAD 0x04 /* 4bits transfer */ u8 bits_per_word; u16 delay_usecs; + u16 cs_change_delay; + u8 cs_change_delay_unit; +#define SPI_DELAY_UNIT_USECS 0 +#define SPI_DELAY_UNIT_NSECS 1 u32 speed_hz; u16 word_delay; -- 2.11.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
next prev parent reply other threads:[~2019-02-23 8:49 UTC|newest] Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-02-23 8:49 [PATCH 0/5] allow to define cs deassert times in us, ns and SCK-len kernel 2019-02-23 8:49 ` kernel 2019-02-23 8:49 ` [PATCH 2/5] spi: core: allow reporting the effectivly used speed_hz for a transfer kernel 2019-02-23 8:49 ` kernel 2019-02-23 8:49 ` [PATCH 3/5] spi: core: allow defining time that cs is deasserted as a multiple of SCK kernel 2019-02-23 8:49 ` kernel [not found] ` <20190223124010.y7lsncknnxoblvgz@wunner.de> [not found] ` <20190223124010.y7lsncknnxoblvgz-JFq808J9C/izQB+pC5nmwQ@public.gmane.org> 2019-02-23 13:15 ` kernel-TqfNSX0MhmxHKSADF0wUEw 2019-02-23 13:15 ` kernel [not found] ` <20190224103913.bjw7g6ievr75iawz@wunner.de> [not found] ` <20190224103913.bjw7g6ievr75iawz-JFq808J9C/izQB+pC5nmwQ@public.gmane.org> 2019-02-24 11:03 ` kernel-TqfNSX0MhmxHKSADF0wUEw 2019-02-24 11:03 ` kernel 2019-02-26 11:37 ` Mark Brown 2019-05-07 10:07 ` kernel 2019-05-07 10:07 ` kernel [not found] ` <20190223084952.14758-1-kernel-TqfNSX0MhmxHKSADF0wUEw@public.gmane.org> 2019-02-23 8:49 ` kernel-TqfNSX0MhmxHKSADF0wUEw [this message] 2019-02-23 8:49 ` [PATCH 1/5] spi: core: allow defining time that cs is deasserted kernel [not found] ` <20190223084952.14758-2-kernel-TqfNSX0MhmxHKSADF0wUEw@public.gmane.org> 2019-05-08 9:34 ` Applied "spi: core: allow defining time that cs is deasserted" to the spi tree Mark Brown 2019-05-08 9:34 ` Mark Brown 2019-02-23 8:49 ` [PATCH 4/5] spi: bcm2835: support effective_speed_hz kernel-TqfNSX0MhmxHKSADF0wUEw 2019-02-23 8:49 ` kernel 2019-05-13 15:14 ` Mark Brown
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20190223084952.14758-2-kernel@martin.sperl.org \ --to=kernel-tqfnsx0mhmxhksadf0wuew@public.gmane.org \ --cc=broonie-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org \ --cc=eric-WhKQ6XTQaPysTnJN9+BGXg@public.gmane.org \ --cc=linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org \ --cc=linux-rpi-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org \ --cc=linux-spi-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \ --cc=stefan.wahren-eS4NqCHxEME@public.gmane.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.