Message-ID: <20081019204858.GA3143@sortiz.org>
Date:	Sun, 19 Oct 2008 22:48:58 +0200
From:	Samuel Ortiz <samuel@...tiz.org>
To:	Bryan Wu <cooloney@...nel.org>
Cc:	jgarzik@...ox.com, irda-users@...ts.sourceforge.net,
	netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
	Graf Yang <graf.yang@...log.com>
Subject: Re: [PATCH 1/1] [net/irda]: new Blackfin on-chip SIR IrDA driver

Hi Bryan,

On Mon, Oct 13, 2008 at 06:04:54PM +0800, Bryan Wu wrote:
> From: Graf Yang <graf.yang@...log.com>
> 
> Signed-off-by: Graf Yang <graf.yang@...log.com>
> Signed-off-by: Bryan Wu <cooloney@...nel.org>
> ---
>  drivers/net/irda/Kconfig    |   52 +++
>  drivers/net/irda/Makefile   |    1 +
>  drivers/net/irda/bfin_sir.c |  824 +++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 877 insertions(+), 0 deletions(-)
>  create mode 100644 drivers/net/irda/bfin_sir.c
> 
> diff --git a/drivers/net/irda/Kconfig b/drivers/net/irda/Kconfig
> index e631755..4e14d03 100644
> --- a/drivers/net/irda/Kconfig
> +++ b/drivers/net/irda/Kconfig
> @@ -17,6 +17,58 @@ config IRTTY_SIR
>  
>  	  If unsure, say Y.
>  
> +config BFIN_SIR
> +	tristate "Blackfin SIR on UART"
> +	depends on BLACKFIN && IRDA && (!SERIAL_BFIN_UART0 || !SERIAL_BFIN_UART1 || !SERIAL_BFIN_UART2 || !SERIAL_BFIN_UART3 )

Do you really need to have those compiled out to have IrDA running ?

> +	help
> +	  Say Y here if your want to enable SIR function on Blackfin UART 
> +	  devices.
> +
> +	  To activate this driver you can start irattach like:
> +	  "irattach irda0 -s"
> +
> +	  Saying M, it will be built as a module named bfin_sir.
> +
> +	  Note that you need to turn off one of the serial drivers for SIR
> +	  to use that UART.
> +
> +config BFIN_SIR0
> +	bool "Blackfin SIR on UART0"
> +	depends on BFIN_SIR && !SERIAL_BFIN_UART0
> +
> +config BFIN_SIR1
> +	bool "Blackfin SIR on UART1"
> +	depends on BFIN_SIR && !SERIAL_BFIN_UART1 && (!BF531 && !BF532 && !BF533 && !BF561)
> +
> +config BFIN_SIR2
> +	bool "Blackfin SIR on UART2"
> +	depends on BFIN_SIR && !SERIAL_BFIN_UART2 && (BF548 || BF549)
> +
> +config BFIN_SIR3
> +	bool "Blackfin SIR on UART3"
> +	depends on BFIN_SIR && !SERIAL_BFIN_UART3 && (BF542 || BF544 || BF548 || BF549)
> +
> +choice
> +	prompt "SIR Mode"
> +	depends on BFIN_SIR
> +	default SIR_BFIN_DMA
> +	help
> +	  This driver supports the built-in serial ports of the Blackfin family
> +	  of CPUs
This help is bogus.


> +config SIR_BFIN_DMA
> +	bool "DMA mode"
> +	depends on !DMA_UNCACHED_NONE
> +	help
> +	  This driver works under DMA mode.
> +
> +config SIR_BFIN_PIO
> +	bool "PIO mode"
> +	help
> +	  This driver works under PIO mode.
I would prefer this to be a module parameter rather than a Kconfig option.


> +endchoice
> +
>  comment "Dongle support"
>  
>  config DONGLE
> diff --git a/drivers/net/irda/Makefile b/drivers/net/irda/Makefile
> index 5d20fde..d82e1e3 100644
> --- a/drivers/net/irda/Makefile
> +++ b/drivers/net/irda/Makefile
> @@ -21,6 +21,7 @@ obj-$(CONFIG_MCS_FIR)	        += mcs7780.o
>  obj-$(CONFIG_AU1000_FIR)	+= au1k_ir.o
>  # SIR drivers
>  obj-$(CONFIG_IRTTY_SIR)		+= irtty-sir.o	sir-dev.o
> +obj-$(CONFIG_BFIN_SIR)		+= bfin_sir.o
>  # dongle drivers for SIR drivers
>  obj-$(CONFIG_ESI_DONGLE)	+= esi-sir.o
>  obj-$(CONFIG_TEKRAM_DONGLE)	+= tekram-sir.o
> diff --git a/drivers/net/irda/bfin_sir.c b/drivers/net/irda/bfin_sir.c
> new file mode 100644
> index 0000000..55d2d58
> --- /dev/null
> +++ b/drivers/net/irda/bfin_sir.c
> @@ -0,0 +1,824 @@
> +/*
> + * Blackfin Infra-red Driver
> + *
> + * Copyright 2006-2008 Analog Devices Inc.
> + *
> + * Enter bugs at http://blackfin.uclinux.org/
> + *
> + * Licensed under the GPL-2 or later.
> + *
> + */
> +
> +#include <linux/module.h>
> +#include <linux/netdevice.h>
> +#include <linux/interrupt.h>
> +#include <linux/delay.h>
> +#include <linux/platform_device.h>
> +#include <linux/dma-mapping.h>
> +
> +#include <net/irda/irda.h>
> +#include <net/irda/wrapper.h>
> +#include <net/irda/irda_device.h>
> +
> +#include <asm/irq.h>
> +#include <asm/dma.h>
> +#include <asm/cacheflush.h>
> +
> +#include <mach/bfin_sir.h>
Those ones should be part of this patch. There are many symbols in this .c
file that are not part of the patch, which makes it a bit more difficult to
review.


> +#ifdef CONFIG_SIR_BFIN_DMA
> +#define DMA_SIR_RX_XCNT        10
> +#define DMA_SIR_RX_YCNT        (PAGE_SIZE / DMA_SIR_RX_XCNT)
> +#define DMA_SIR_RX_FLUSH_JIFS  (HZ * 4 / 250)
> +#endif
> +
> +static int max_rate = 115200;
> +
> +static void turnaround_delay(unsigned long last_jif, int mtt)
> +{
> +	long ticks;
> +
> +	mtt = mtt < 10000 ? 10000 : mtt;
> +	ticks = 1 + mtt / (USEC_PER_SEC / HZ);
> +	schedule_timeout_interruptible(ticks);
> +}
> +static void __init bfin_sir_init_ports(int i)
> +{
> +	sir_ports[i].membase   = (void __iomem *)bfin_sir_port_resource[i].base_addr;
> +	sir_ports[i].irq = bfin_sir_port_resource[i].irq;
> +	sir_ports[i].clk = get_sclk();
> +#ifdef CONFIG_SIR_BFIN_DMA
> +	sir_ports[i].tx_done        = 1;
> +	init_timer(&(sir_ports[i].rx_dma_timer));
> +#endif
> +	sir_ports[i].rx_dma_channel = bfin_sir_port_resource[i].rx_dma_channel;
> +	sir_ports[i].tx_dma_channel = bfin_sir_port_resource[i].tx_dma_channel;
> +}
> +
> +static void bfin_sir_stop_tx(struct bfin_sir_port *port)
> +{
> +#ifndef CONFIG_BF54x
> +	unsigned short ier;
> +#endif
> +#ifdef CONFIG_SIR_BFIN_DMA
> +	disable_dma(port->tx_dma_channel);
> +#endif
> +
> +	while (!(SIR_UART_GET_LSR(port) & THRE))
> +		continue;
> +#ifdef CONFIG_BF54x
> +	SIR_UART_PUT_LSR(port, TFI);
> +	SIR_UART_CLEAR_IER(port, ETBEI);
> +#else
> +	ier = SIR_UART_GET_IER(port);
> +	ier &= ~ETBEI;
> +	SIR_UART_PUT_IER(port, ier);
> +#endif
> +}
> +
> +static void bfin_sir_enable_tx(struct bfin_sir_port *port)
> +{
> +#ifndef CONFIG_BF54x
> +	unsigned short ier;
> +#endif
> +#ifdef CONFIG_BF54x
> +	SIR_UART_SET_IER(port, ETBEI);
> +#else
> +	ier = SIR_UART_GET_IER(port);
> +	ier |= ETBEI;
> +	SIR_UART_PUT_IER(port, ier);
> +#endif
> +}
> +
> +static void bfin_sir_stop_rx(struct bfin_sir_port *port)
> +{
> +#ifdef CONFIG_BF54x
> +	SIR_UART_CLEAR_IER(port, ERBFI);
> +#else
> +	unsigned short ier;
> +
> +	ier = SIR_UART_GET_IER(port);
> +	ier &= ~ERBFI;
> +	SIR_UART_PUT_IER(port, ier);
> +#endif
> +}
> +
> +static void bfin_sir_enable_rx(struct bfin_sir_port *port)
> +{
> +#ifdef CONFIG_BF54x
> +	SIR_UART_SET_IER(port, ERBFI);
> +#else
> +	SIR_UART_PUT_IER(port, SIR_UART_GET_IER(port) | ERBFI);
> +#endif
> +}
All those ifdefs are quite ugly, imo.
Why don't you have e.g. one sir_uart_set_ier() (please note the lowercase
usage as I'd expect SIR_UART_SET_IER() to be a macro...), and add the needed
ifdefs there ?
Also, all your bfin_sir.h files have a lot of duplicated code, I would really
like to see those one cleaned up.

> +static int bfin_sir_set_speed(struct bfin_sir_port *port, int speed)
> +{
> +	int ret = -EINVAL;
> +	unsigned int quot;
> +	unsigned short val, lsr, lcr = 0;
> +
> +	lcr = WLS(8);
> +
> +	switch (speed) {
> +	case 9600:
> +	case 19200:
> +	case 38400:
> +	case 57600:
> +	case 115200:
> +
> +		quot = (port->clk + (8 * speed)) / (16 * speed);
> +
> +		do {
> +			lsr = SIR_UART_GET_LSR(port);
> +		} while (!(lsr & TEMT));
> +
> +		/* Clear UCEN bit to reset the UART state machine
> +		 * and control registers
> +		 */
> +		val = SIR_UART_GET_GCTL(port);
> +		val &= ~UCEN;
> +		SIR_UART_PUT_GCTL(port, val);
> +
> +#ifndef CONFIG_BF54x
> +		/* Set DLAB in LCR to Access DLL and DLH */
> +		val = SIR_UART_GET_LCR(port);
> +		val |= DLAB;
> +		SIR_UART_PUT_LCR(port, val);
> +		SSYNC();
> +#endif
> +
> +		SIR_UART_PUT_DLL(port, quot & 0xFF);
> +		SSYNC();
> +		SIR_UART_PUT_DLH(port, (quot >> 8) & 0xFF);
> +		SSYNC();
> +
> +#ifndef CONFIG_BF54x
> +		/* Clear DLAB in LCR to Access THR RBR IER */
> +		val = SIR_UART_GET_LCR(port);
> +		val &= ~DLAB;
> +		SIR_UART_PUT_LCR(port, val);
> +		SSYNC();
> +#endif
Here you could have a routine that would NOP if you're not running on a BF54x
machine.
Also, I'm surprised you dont have runtime machine id detection. I know that's
none of my business, but are you guys working towards having one single kernel
supporting all blackfin CPUs ?


> +		SIR_UART_PUT_LCR(port, lcr);
> +
> +		val = SIR_UART_GET_GCTL(port);
> +		val |= UCEN;
> +		SIR_UART_PUT_GCTL(port, val);
> +
> +		ret = 0;
> +		/*printk(KERN_DEBUG "bfin_sir: Set new speed %d\n", speed);*/
> +		break;
> +	default:
> +		printk(KERN_WARNING "bfin_sir: Invalid speed %d\n", speed);
> +		break;
> +	}
> +
> +	val = SIR_UART_GET_GCTL(port);
> +	/* If not add the 'RPOLC', we can't catch the receive interrupt.
> +	 * It's related with the HW layout and the IR transiver.
> +	 */
> +	val |= IREN | RPOLC;
> +	SIR_UART_PUT_GCTL(port, val);
> +	return ret;
> +}
> +
> +static int bfin_sir_is_receiving(struct net_device *dev)
> +{
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +
> +	if (!(SIR_UART_GET_IER(port) & ERBFI))
> +		return 0;
> +	return self->rx_buff.state != OUTSIDE_FRAME;
> +}
> +
> +#ifdef CONFIG_SIR_BFIN_PIO
> +static void bfin_sir_tx_chars(struct net_device *dev)
> +{
> +	unsigned int chr;
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +
> +	if (self->tx_buff.len != 0) {
> +		chr = *(self->tx_buff.data);
> +		SIR_UART_PUT_CHAR(port, chr);
> +		self->tx_buff.data++;
> +		self->tx_buff.len--;
> +	} else {
> +		self->stats.tx_packets++;
> +		self->stats.tx_bytes += self->tx_buff.data - self->tx_buff.head;
> +		if (self->newspeed) {
> +			bfin_sir_set_speed(port, self->newspeed);
> +			self->speed = self->newspeed;
> +			self->newspeed = 0;
> +		}
> +		bfin_sir_stop_tx(port);
> +		bfin_sir_enable_rx(port);
> +		/* I'm hungry! */
> +		netif_wake_queue(dev);
> +	}
> +}
> +
> +static void bfin_sir_rx_chars(struct net_device *dev)
> +{
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +	unsigned char ch;
> +
> +	SIR_UART_CLEAR_LSR(port);
> +	ch = SIR_UART_GET_CHAR(port);
> +	async_unwrap_char(dev, &self->stats, &self->rx_buff, ch);
> +	dev->last_rx = jiffies;
> +}
> +
> +static irqreturn_t bfin_sir_rx_int(int irq, void *dev_id)
> +{
> +	struct net_device *dev = dev_id;
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +	while ((SIR_UART_GET_LSR(port) & DR))
> +		bfin_sir_rx_chars(dev);
Shouldnt this be handled from a bottom half. You're busy looping in an
interrupt hanlder here, that doesnt look too good.


> +
> +	return IRQ_HANDLED;
> +}
> +
> +static irqreturn_t bfin_sir_tx_int(int irq, void *dev_id)
> +{
> +	struct net_device *dev = dev_id;
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +
> +	if (SIR_UART_GET_LSR(port) & THRE)
> +		bfin_sir_tx_chars(dev);
> +	return IRQ_HANDLED;
> +}
> +#endif /* CONFIG_SIR_BFIN_PIO */
> +
> +#ifdef CONFIG_SIR_BFIN_DMA
> +static void bfin_sir_dma_tx_chars(struct net_device *dev)
> +{
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +
> +	if (!port->tx_done)
> +		return;
> +	port->tx_done = 0;
> +
> +	if (self->tx_buff.len == 0) {
> +		self->stats.tx_packets++;
> +		if (self->newspeed) {
> +			bfin_sir_set_speed(port, self->newspeed);
> +			self->speed = self->newspeed;
> +			self->newspeed = 0;
> +		}
> +		bfin_sir_enable_rx(port);
> +		port->tx_done = 1;
> +		netif_wake_queue(dev);
> +		return;
> +	}
> +
> +	blackfin_dcache_flush_range((unsigned long)(self->tx_buff.data),
> +		(unsigned long)(self->tx_buff.data+self->tx_buff.len));
> +	set_dma_config(port->tx_dma_channel,
> +		set_bfin_dma_config(DIR_READ, DMA_FLOW_STOP,
> +			INTR_ON_BUF, DIMENSION_LINEAR, DATA_SIZE_8,
> +			DMA_SYNC_RESTART));
> +	set_dma_start_addr(port->tx_dma_channel,
> +		(unsigned long)(self->tx_buff.data));
> +	set_dma_x_count(port->tx_dma_channel, self->tx_buff.len);
> +	set_dma_x_modify(port->tx_dma_channel, 1);
> +	enable_dma(port->tx_dma_channel);
> +	bfin_sir_enable_tx(port);
> +}
> +
> +static irqreturn_t bfin_sir_dma_tx_int(int irq, void *dev_id)
> +{
> +	struct net_device *dev = dev_id;
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +	spin_lock(&self->lock);
> +	if (!(get_dma_curr_irqstat(port->tx_dma_channel)&DMA_RUN)) {
> +		clear_dma_irqstat(port->tx_dma_channel);
> +		bfin_sir_stop_tx(port);
> +
> +		self->stats.tx_packets++;
> +		self->stats.tx_bytes += self->tx_buff.len;
> +		self->tx_buff.len = 0;
> +		if (self->newspeed) {
> +			bfin_sir_set_speed(port, self->newspeed);
> +			self->speed = self->newspeed;
> +			self->newspeed = 0;
> +		}
> +		bfin_sir_enable_rx(port);
> +		/* I'm hungry! */
> +		netif_wake_queue(dev);
> +		port->tx_done = 1;
> +	}
> +	spin_unlock(&self->lock);
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static void bfin_sir_dma_rx_chars(struct net_device *dev)
> +{
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +	int i;
> +
> +	SIR_UART_CLEAR_LSR(port);
> +
> +	for (i = port->rx_dma_buf.head; i < port->rx_dma_buf.tail; i++)
> +		async_unwrap_char(dev, &self->stats, &self->rx_buff, port->rx_dma_buf.buf[i]);
> +}
> +
> +void bfin_sir_rx_dma_timeout(struct net_device *dev)
> +{
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +	int x_pos, pos;
> +	int flags = 0;
> +
> +	spin_lock_irqsave(&self->lock, flags);
> +	x_pos = DMA_SIR_RX_XCNT - get_dma_curr_xcount(port->rx_dma_channel);
> +	if (x_pos == DMA_SIR_RX_XCNT)
> +		x_pos = 0;
> +
> +	pos = port->rx_dma_nrows * DMA_SIR_RX_XCNT + x_pos;
> +
> +	if (pos > port->rx_dma_buf.tail) {
> +		port->rx_dma_buf.tail = pos;
> +		bfin_sir_dma_rx_chars(dev);
> +		port->rx_dma_buf.head = port->rx_dma_buf.tail;
> +	}
> +	spin_unlock_irqrestore(&self->lock, flags);
> +}
> +
> +static irqreturn_t bfin_sir_dma_rx_int(int irq, void *dev_id)
> +{
> +	struct net_device *dev = dev_id;
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +	unsigned short irqstat;
> +
> +	spin_lock(&self->lock);
> +
> +	port->rx_dma_nrows++;
> +	port->rx_dma_buf.tail = DMA_SIR_RX_XCNT * port->rx_dma_nrows;
> +	bfin_sir_dma_rx_chars(dev);
> +	if (port->rx_dma_nrows >= DMA_SIR_RX_YCNT) {
> +		port->rx_dma_nrows = 0;
> +		port->rx_dma_buf.tail = 0;
> +	}
> +	port->rx_dma_buf.head = port->rx_dma_buf.tail;
> +
> +	irqstat = get_dma_curr_irqstat(port->rx_dma_channel);
> +	clear_dma_irqstat(port->rx_dma_channel);
> +	spin_unlock(&self->lock);
> +
> +	mod_timer(&(port->rx_dma_timer), jiffies + DMA_SIR_RX_FLUSH_JIFS);
> +	return IRQ_HANDLED;
> +}
> +#endif /* CONFIG_SIR_BFIN_DMA */
> +
> +static int bfin_sir_startup(struct bfin_sir_port *port, struct net_device *dev)
> +{
> +#ifdef CONFIG_SIR_BFIN_DMA
> +	dma_addr_t dma_handle;
> +#endif /* CONFIG_SIR_BFIN_DMA */
> +
> +	if (request_dma(port->rx_dma_channel, "BFIN_UART_RX") < 0) {
> +		printk(KERN_WARNING "bfin_sir: Unable to attach SIR RX DMA channel\n");
> +		return -EBUSY;
> +	}
> +
> +	if (request_dma(port->tx_dma_channel, "BFIN_UART_TX") < 0) {
> +		printk(KERN_WARNING "bfin_sir: Unable to attach SIR TX DMA channel\n");
> +		free_dma(port->rx_dma_channel);
> +		return -EBUSY;
> +	}
Why arent the request_dma() not ifdef CONFIG_SIR_BFIN_DMA ?
Here as well, I'd prefer to have 2 sub routines, one for DMA init and another
one for the PIO case. That combined with having dma selection as a module
parameter would make the whole thing better.



> +#ifdef CONFIG_SIR_BFIN_DMA
> +
> +	set_dma_callback(port->rx_dma_channel, bfin_sir_dma_rx_int, dev);
> +	set_dma_callback(port->tx_dma_channel, bfin_sir_dma_tx_int, dev);
> +
> +	port->rx_dma_buf.buf = (unsigned char *)dma_alloc_coherent(NULL, PAGE_SIZE, &dma_handle, GFP_DMA);
> +	port->rx_dma_buf.head = 0;
> +	port->rx_dma_buf.tail = 0;
> +	port->rx_dma_nrows = 0;
> +
> +	set_dma_config(port->rx_dma_channel,
> +				set_bfin_dma_config(DIR_WRITE, DMA_FLOW_AUTO,
> +									INTR_ON_ROW, DIMENSION_2D,
> +									DATA_SIZE_8, DMA_SYNC_RESTART));
> +	set_dma_x_count(port->rx_dma_channel, DMA_SIR_RX_XCNT);
> +	set_dma_x_modify(port->rx_dma_channel, 1);
> +	set_dma_y_count(port->rx_dma_channel, DMA_SIR_RX_YCNT);
> +	set_dma_y_modify(port->rx_dma_channel, 1);
> +	set_dma_start_addr(port->rx_dma_channel, (unsigned long)port->rx_dma_buf.buf);
> +	enable_dma(port->rx_dma_channel);
> +
> +	port->rx_dma_timer.data = (unsigned long)(dev);
> +	port->rx_dma_timer.function = (void *)bfin_sir_rx_dma_timeout;
> +
> +#else
> +
> +	if (request_irq(port->irq, bfin_sir_rx_int, IRQF_DISABLED, "BFIN_SIR_RX", dev)) {
> +		printk(KERN_WARNING "bfin_sir: Unable to attach SIR RX interrupt\n");
> +		return -EBUSY;
> +	}
> +
> +	if (request_irq(port->irq+1, bfin_sir_tx_int, IRQF_DISABLED, "BFIN_SIR_TX", dev)) {
> +		printk(KERN_WARNING "bfin_sir: Unable to attach SIR TX interrupt\n");
> +		free_irq(port->irq, dev);
> +		return -EBUSY;
> +	}
> +#endif
> +
> +	return 0;
> +}
> +
> +static void bfin_sir_shutdown(struct bfin_sir_port *port, struct net_device *dev)
> +{
> +	unsigned short val;
> +
> +	bfin_sir_stop_rx(port);
> +#ifdef CONFIG_BF54x
> +	SIR_UART_CLEAR_IER(port, 0xF);
> +#else
> +	SIR_UART_PUT_IER(port, 0);
> +#endif
> +	val = SIR_UART_GET_GCTL(port);
> +	val &= ~(UCEN | IREN | RPOLC);
> +	SIR_UART_PUT_GCTL(port, val);
> +
> +#ifdef CONFIG_SIR_BFIN_DMA
> +	disable_dma(port->tx_dma_channel);
> +	disable_dma(port->rx_dma_channel);
> +	del_timer(&(port->rx_dma_timer));
> +	dma_free_coherent(NULL, PAGE_SIZE, port->rx_dma_buf.buf, 0);
> +#else
> +	free_irq(port->irq+1, dev);
> +	free_irq(port->irq, dev);
> +#endif
> +	free_dma(port->tx_dma_channel);
> +	free_dma(port->rx_dma_channel);
> +}
> +
> +#ifdef CONFIG_PM
> +static int bfin_sir_suspend(struct platform_device *pdev, pm_message_t state)
> +{
> +	struct bfin_sir_port *psir_ports;
> +	struct net_device *dev;
> +	struct bfin_sir_self *self;
> +	int i;
> +
> +	psir_ports = platform_get_drvdata(pdev);
> +	if (!psir_ports)
> +		return 0;
> +	for (i = nr_sirs-1; i >= 0; i--) {
> +		dev = (psir_ports+i)->dev;
> +		self = dev->priv;
> +		if (self->open) {
> +			flush_scheduled_work();
> +			bfin_sir_shutdown(self->sir_port, dev);
> +			netif_device_detach(dev);
> +		}
> +	}
> +
> +	return 0;
> +}
> +static int bfin_sir_resume(struct platform_device *pdev)
> +{
> +	struct bfin_sir_port *psir_ports;
> +	struct net_device *dev;
> +	struct bfin_sir_self *self;
> +	struct bfin_sir_port *port;
> +	int i;
> +
> +	psir_ports = platform_get_drvdata(pdev);
> +	if (!psir_ports) return 0;
> +	for (i = nr_sirs-1; i >= 0; i--) {
> +		dev = (psir_ports+i)->dev;
> +		self = dev->priv;
> +		port = self->sir_port;
> +		if (self->open) {
> +			if (self->newspeed) {
> +				self->speed = self->newspeed;
> +				self->newspeed = 0;
> +			}
> +			bfin_sir_startup(port, dev);
> +			bfin_sir_set_speed(port, 9600);
> +			bfin_sir_enable_rx(port);
> +			netif_device_attach(dev);
> +		}
> +	}
> +	return 0;
> +}
> +#else
> +#define bfin_sir_suspend   NULL
> +#define bfin_sir_resume    NULL
> +#endif
> +
> +static void bfin_sir_send_work(struct work_struct *work)
> +{
> +	struct bfin_sir_self  *self = container_of(work, struct bfin_sir_self, work);
> +	struct net_device *dev = self->sir_port->dev;
> +	struct bfin_sir_port *port = self->sir_port;
> +	unsigned short val;
> +	int tx_cnt = 10;
> +
> +	while (bfin_sir_is_receiving(dev) && --tx_cnt)
> +		turnaround_delay(dev->last_rx, self->mtt);
> +
> +	bfin_sir_stop_rx(port);
> +
> +	/* To avoid losting RX interrupt, we reset IR function before
> +	 * sending data. We also can set the speed, which will
> +	 * reset all the UART.
> +	 */
> +	val = SIR_UART_GET_GCTL(port);
> +	val &= ~(IREN | RPOLC);
> +	SIR_UART_PUT_GCTL(port, val);
> +	SSYNC();
> +	val |= IREN | RPOLC;
> +	SIR_UART_PUT_GCTL(port, val);
> +	SSYNC();
> +	/* bfin_sir_set_speed(port, self->speed); */
> +
> +#ifdef CONFIG_SIR_BFIN_DMA
> +	bfin_sir_dma_tx_chars(dev);
> +#else
> +	bfin_sir_enable_tx(port);
> +#endif
> +	dev->trans_start = jiffies;
> +}
> +
> +static int bfin_sir_hard_xmit(struct sk_buff *skb, struct net_device *dev)
> +{
> +	struct bfin_sir_self *self = dev->priv;
> +	int speed = irda_get_next_speed(skb);
> +
> +	netif_stop_queue(dev);
> +
> +	self->mtt = irda_get_mtt(skb);
> +
> +	if (speed != self->speed && speed != -1)
> +		self->newspeed = speed;
> +
> +	self->tx_buff.data = self->tx_buff.head;
> +	if (skb->len == 0)
> +		self->tx_buff.len = 0;
> +	else
> +		self->tx_buff.len = async_wrap_skb(skb, self->tx_buff.data, self->tx_buff.truesize);
> +
> +	schedule_work(&self->work);
> +	dev_kfree_skb(skb);
> +
> +	return 0;
> +}
> +
> +static int bfin_sir_ioctl(struct net_device *dev, struct ifreq *ifreq, int cmd)
> +{
> +	struct if_irda_req *rq = (struct if_irda_req *)ifreq;
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +	int ret = 0;
> +
> +	switch (cmd) {
> +	case SIOCSBANDWIDTH:
> +		if (capable(CAP_NET_ADMIN)) {
> +			if (self->open) {
> +				ret = bfin_sir_set_speed(port, rq->ifr_baudrate);
> +				bfin_sir_enable_rx(port);
> +			} else {
> +				printk(KERN_WARNING "bfin_sir: SIOCSBANDWIDTH: !netif_running\n");
> +				ret = 0;
> +			}
> +		}
> +		break;
> +
> +	case SIOCSMEDIABUSY:
> +		ret = -EPERM;
> +		if (capable(CAP_NET_ADMIN)) {
> +			irda_device_set_media_busy(dev, TRUE);
> +			ret = 0;
> +		}
> +		break;
> +
> +	case SIOCGRECEIVING:
> +		rq->ifr_receiving = bfin_sir_is_receiving(dev);
> +		break;
> +
> +	default:
> +		ret = -EOPNOTSUPP;
> +		break;
> +	}
> +
> +	return ret;
> +}
> +
> +static struct net_device_stats *bfin_sir_stats(struct net_device *dev)
> +{
> +	struct bfin_sir_self *self = dev->priv;
> +	return &self->stats;
> +}
> +
> +static int bfin_sir_open(struct net_device *dev)
> +{
> +	struct bfin_sir_self *self = dev->priv;
> +	struct bfin_sir_port *port = self->sir_port;
> +	int err = -ENOMEM;
> +
> +	self->newspeed = 0;
> +	self->speed = 9600;
> +
> +	spin_lock_init(&self->lock);
> +
> +	err = bfin_sir_startup(port, dev);
> +	if (err)
> +		goto err_startup;
> +
> +	bfin_sir_set_speed(port, 9600);
> +
> +	self->irlap = irlap_open(dev, &self->qos, DRIVER_NAME);
> +	if (!self->irlap)
> +		goto err_irlap;
> +
> +	INIT_WORK(&self->work, bfin_sir_send_work);
> +
> +	/*
> +	 * Now enable the interrupt then start the queue
> +	 */
> +	self->open = 1;
> +	bfin_sir_enable_rx(port);
> +
> +	netif_start_queue(dev);
> +
> +	return 0;
> +
> +err_irlap:
> +	self->open = 0;
> +	bfin_sir_shutdown(port, dev);
> +err_startup:
> +	return err;
> +}
> +
> +static int bfin_sir_stop(struct net_device *dev)
> +{
> +	struct bfin_sir_self *self = dev->priv;
> +
> +	flush_scheduled_work();
> +	bfin_sir_shutdown(self->sir_port, dev);
> +
> +	if (self->rxskb) {
> +		dev_kfree_skb(self->rxskb);
> +		self->rxskb = NULL;
> +	}
> +
> +	/* Stop IrLAP */
> +	if (self->irlap) {
> +		irlap_close(self->irlap);
> +		self->irlap = NULL;
> +	}
> +
> +	netif_stop_queue(dev);
> +	self->open = 0;
> +
> +	return 0;
> +}
> +
> +static int bfin_sir_init_iobuf(iobuff_t *io, int size)
> +{
> +	io->head = kmalloc(size, GFP_KERNEL | GFP_DMA);
> +	if (io->head != NULL) {
I think the typical kernel code would do something like:

if (!io->head)
	return -ENOMEM;

and then go on with your initialisation code.


> +		io->truesize = size;
> +		io->in_frame = FALSE;
> +		io->state    = OUTSIDE_FRAME;
> +		io->data     = io->head;
> +	}
> +	return io->head ? 0 : -ENOMEM;
> +}
> +
> +static int __devinit bfin_sir_probe(struct platform_device *pdev)
> +{
> +	struct net_device *dev;
> +	struct bfin_sir_self *self;
> +	unsigned int baudrate_mask;
> +	int i, err = 0;
> +
> +	err = bfin_sir_hw_init();
> +
> +	for (i = 0; i < nr_sirs; i++) {
I dont like that neither.
Just for the record, nr_sirs is defined, e.g. as :

struct bfin_sir_port_res bfin_sir_port_resource[] = {
#ifdef CONFIG_BFIN_SIR0
	{
	0xFFC00400,
	IRQ_UART0_RX,
	CH_UART0_RX,
	CH_UART0_TX,
	},
#endif
#ifdef CONFIG_BFIN_SIR1
	{
	0xFFC02000,
	IRQ_UART1_RX,
	CH_UART1_RX,
	CH_UART1_TX,
	},
#endif
};

int nr_sirs = ARRAY_SIZE(bfin_sir_port_resource);

I think the uart(s) that your board is going to use for IrDA should be passed
as through some platform data structure. That would allow you to get rid of
all those CONFIG_BFIN_SIR*, and would make the probe routine nicer.
You may want to pass your UART base address, irq, and channels as platform
resources, although that's less important.

> +		bfin_sir_init_ports(i);
> +		dev = alloc_irdadev(sizeof(struct bfin_sir_self));
> +		if (!dev)
> +			goto err_mem_1;
> +
> +		self = dev->priv;
> +		self->dev = &pdev->dev;
> +		self->sir_port = &sir_ports[i];
> +		sir_ports[i].dev = dev;
> +
> +		err = bfin_sir_init_iobuf(&self->rx_buff, IRDA_SKB_MAX_MTU);
> +		if (err)
> +			goto err_mem_2;
> +		err = bfin_sir_init_iobuf(&self->tx_buff, IRDA_SIR_MAX_FRAME);
> +		if (err)
> +			goto err_mem_3;
> +
> +		dev->hard_start_xmit = bfin_sir_hard_xmit;
> +		dev->open            = bfin_sir_open;
> +		dev->stop            = bfin_sir_stop;
> +		dev->do_ioctl        = bfin_sir_ioctl;
> +		dev->get_stats       = bfin_sir_stats;
> +		dev->irq             = sir_ports[i].irq;
> +
> +		irda_init_max_qos_capabilies(&self->qos);
> +
> +		baudrate_mask = IR_9600;
> +
> +		switch (max_rate) {
> +		case 115200:		baudrate_mask |= IR_115200;
> +		case 57600:		baudrate_mask |= IR_57600;
> +		case 38400:		baudrate_mask |= IR_38400;
> +		case 19200:		baudrate_mask |= IR_19200;
> +		}
> +
> +		self->qos.baud_rate.bits &= baudrate_mask;
> +
> +		self->qos.min_turn_time.bits = 1; /* 10 ms or more */
> +
> +		irda_qos_bits_to_value(&self->qos);
> +
> +		err = register_netdev(dev);
> +
> +		if (err) {
> +err_mem_3:
> +			kfree(self->tx_buff.head);
> +err_mem_2:
> +			kfree(self->rx_buff.head);
> +err_mem_1:
> +			free_netdev(dev);
> +		}
> +	}
> +	if (err == 0)
> +		platform_set_drvdata(pdev, sir_ports);
> +	return err;
> +}
> +
> +static int __devexit bfin_sir_remove(struct platform_device *pdev)
> +{
> +	int i;
> +	struct bfin_sir_port *psir_ports;
> +	struct net_device *dev = NULL;
> +	struct bfin_sir_self *self;
> +
> +
> +	psir_ports = platform_get_drvdata(pdev);
> +	if (!psir_ports)
> +		return 0;
> +	for (i = nr_sirs-1; i >= 0; i--) {
> +		dev = (psir_ports+i)->dev;
> +		self = dev->priv;
> +		unregister_netdev(dev);
> +		kfree(self->tx_buff.head);
> +		kfree(self->rx_buff.head);
> +		free_netdev(dev);
> +	}
> +	platform_set_drvdata(pdev, NULL);
> +
> +	return 0;
> +}
> +
> +static struct platform_driver bfin_ir_driver = {
> +	.probe   = bfin_sir_probe,
> +	.remove  = __devexit_p(bfin_sir_remove),
> +	.suspend = bfin_sir_suspend,
> +	.resume  = bfin_sir_resume,
> +	.driver  = {
> +		.name = DRIVER_NAME,
> +	},
> +};
> +
> +static int __init bfin_sir_init(void)
> +{
> +	return platform_driver_register(&bfin_ir_driver);
> +}
> +
> +static void __exit bfin_sir_exit(void)
> +{
> +	platform_driver_unregister(&bfin_ir_driver);
> +}
> +
> +module_init(bfin_sir_init);
> +module_exit(bfin_sir_exit);
> +
> +module_param(max_rate, int, 0);
> +MODULE_PARM_DESC(max_rate, "Maximum baud rate (115200, 57600, 38400, 19200, 9600)");
> +
> +MODULE_AUTHOR("Graf.Yang <graf.yang@...log.com>");
> +MODULE_DESCRIPTION("Blackfin IrDA driver");
> +MODULE_LICENSE("GPL");
> -- 
> 1.5.6

Cheers,
Samuel.

-- 
Intel Open Source Technology Centre
http://oss.intel.com/

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Powered by blists - more mailing lists