* [PATCH] OMAP: Add support for IrDA driver
@ 2006-08-07 9:12 Komal Shah
2006-08-11 5:47 ` Samuel Ortiz
0 siblings, 1 reply; 3+ messages in thread
From: Komal Shah @ 2006-08-07 9:12 UTC (permalink / raw)
To: samuel, irda-users; +Cc: , tony, r-woodruff2, akpm
[-- Attachment #1: Type: text/plain, Size: 476 bytes --]
Samuel/Tony,
I have attached the OMAP IrDA patch for OMAP1610/1710 and OMAP242x
for review. See that for IrDA driver to work on H3 and H4, we need
gpio-expander functions to go in mainline, but it can work on H2.
FYI...we are in process of submitting OMAP I2C driver to mainline,
and right now I am working on the 2nd round of review comments
from Jean.
---Komal Shah
http://komalshah.blogspot.com
--
http://www.fastmail.fm - Choose from over 50 domains or use your own
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-OMAP-Add-support-for-IrDA-driver.patch --]
[-- Type: text/x-patch; name="0001-OMAP-Add-support-for-IrDA-driver.patch", Size: 24186 bytes --]
From nobody Mon Sep 17 00:00:00 2001
From: Komal Shah <komal_shah802003@yahoo.com>
Date: Mon, 7 Aug 2006 18:38:45 +0530
Subject: [PATCH] OMAP: Add support for IrDA driver
Patch adds IrDA driver support for Texas Instruments OMAP
(http://www.ti.com/omap) based processors like OMAP1610/1710 and OMAP242x.
Signed-off-by: Komal Shah <komal_shah802003@yahoo.com>
---
drivers/net/irda/Kconfig | 10
drivers/net/irda/Makefile | 1
drivers/net/irda/omap-ir.c | 903 ++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 914 insertions(+), 0 deletions(-)
create mode 100644 drivers/net/irda/omap-ir.c
e725e5afaf6b7f9903ee8b48faf1faeff8604ac7
diff --git a/drivers/net/irda/Kconfig b/drivers/net/irda/Kconfig
index e9e6d99..8d29a0a 100644
--- a/drivers/net/irda/Kconfig
+++ b/drivers/net/irda/Kconfig
@@ -432,5 +432,15 @@ config MCS_FIR
To compile it as a module, choose M here: the module will be called
mcs7780.
+config OMAP_IR
+ tristate "OMAP IrDA(SIR/MIR/FIR)"
+ depends on IRDA && ARCH_OMAP
+ select GPIOEXPANDER_OMAP if (MACH_OMAP_H3 || MACH_OMAP_H4)
+ help
+ Say Y here if you want to build support for the Teaxas Instruments
+ OMAP IrDA device driver, which supports SIR/MIR/FIR. This driver
+ relies on platform specific helper routines so available capabilities
+ may vary from one OMAP target to another.
+
endmenu
diff --git a/drivers/net/irda/Makefile b/drivers/net/irda/Makefile
index 5be09f1..89c6389 100644
--- a/drivers/net/irda/Makefile
+++ b/drivers/net/irda/Makefile
@@ -20,6 +20,7 @@ obj-$(CONFIG_VLSI_FIR) += vlsi_ir.o
obj-$(CONFIG_VIA_FIR) += via-ircc.o
obj-$(CONFIG_PXA_FICP) += pxaficp_ir.o
obj-$(CONFIG_MCS_FIR) += mcs7780.o
+obj-$(CONFIG_OMAP_IR) += omap-ir.o
# Old dongle drivers for old SIR drivers
obj-$(CONFIG_ESI_DONGLE_OLD) += esi.o
obj-$(CONFIG_TEKRAM_DONGLE_OLD) += tekram.o
diff --git a/drivers/net/irda/omap-ir.c b/drivers/net/irda/omap-ir.c
new file mode 100644
index 0000000..c51fb48
--- /dev/null
+++ b/drivers/net/irda/omap-ir.c
@@ -0,0 +1,903 @@
+/*
+ * BRIEF MODULE DESCRIPTION
+ *
+ * Infra-red driver for the OMAP1610-H2 and OMAP1710-H3 and H4 Platforms
+ * (SIR/MIR/FIR modes)
+ * (based on omap-sir.c)
+ *
+ * Copyright 2003 MontaVista Software Inc.
+ * Author: MontaVista Software, Inc.
+ * source@mvista.com
+ *
+ * Copyright 2004 Texas Instruments.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; either version 2 of the License, or (at your
+ * option) any later version.
+ *
+ * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
+ * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN
+ * NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
+ * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ Modifications:
+ Feb 2004, Texas Instruments
+ - Ported to 2.6 kernel (Feb 2004).
+ *
+ Apr 2004, Texas Instruments
+ - Added support for H3 (Apr 2004).
+ Nov 2004, Texas Instruments
+ - Added support for Power Management.
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/netdevice.h>
+#include <linux/slab.h>
+#include <linux/rtnetlink.h>
+#include <linux/interrupt.h>
+#include <linux/delay.h>
+#include <linux/ioport.h>
+#include <linux/dma-mapping.h>
+#include <linux/platform_device.h>
+#include <linux/i2c.h>
+
+#include <net/irda/irda.h>
+#include <net/irda/irmod.h>
+#include <net/irda/wrapper.h>
+#include <net/irda/irda_device.h>
+
+#include <asm/irq.h>
+#include <asm/io.h>
+#include <asm/hardware.h>
+#include <asm/serial.h>
+#include <asm/mach-types.h>
+#include <asm/dma.h>
+#include <asm/arch/mux.h>
+#include <asm/arch/gpio.h>
+#include <asm/arch/irda.h>
+
+#define UART3_EFR_EN (1 << 4)
+#define UART3_MCR_EN_TCR_TLR (1 << 6)
+
+#define UART3_LCR_WL_8 (3 << 0)
+#define UART3_LCR_SP2 (1 << 2)
+#define UART3_LCR_DIVEN (1 << 7)
+
+#define UART3_FCR_FIFO_EN (1 << 0)
+#define UART3_FCR_FIFO_RX (1 << 1)
+#define UART3_FCR_FIFO_TX (1 << 2)
+#define UART3_FCR_FIFO_DMA1 (1 << 3)
+#define UART3_FCR_FIFO_TX_TRIG16 (1 << 4)
+#define UART3_FCR_FIFO_RX_TRIG16 (1 << 6)
+#define UART3_FCR_CONFIG (\
+ UART3_FCR_FIFO_EN | UART3_FCR_FIFO_RX |\
+ UART3_FCR_FIFO_TX | UART3_FCR_FIFO_DMA1 |\
+ UART3_FCR_FIFO_TX_TRIG16 |\
+ UART3_FCR_FIFO_RX_TRIG16)
+
+#define UART3_SCR_TX_TRIG1 (1 << 6)
+#define UART3_SCR_RX_TRIG1 (1 << 7)
+
+#define UART3_MDR1_RESET (0x07)
+#define UART3_MDR1_SIR (1 << 0)
+#define UART3_MDR1_MIR (4 << 0)
+#define UART3_MDR1_FIR (5 << 0)
+#define UART3_MDR1_SIP_AUTO (1 << 6)
+
+#define UART3_MDR2_TRIG1 (0 << 1)
+#define UART3_MDR2_IRTX_UNDERRUN (1 << 0)
+
+#define UART3_ACERG_TX_UNDERRUN_DIS (1 << 4)
+#define UART3_ACERG_SD_MODE_LOW (1 << 6)
+#define UART3_ACERG_DIS_IR_RX (1 << 5)
+
+#define UART3_IER_EOF (1 << 5)
+#define UART3_IER_CTS (1 << 7)
+
+#define UART3_IIR_TX_STATUS (1 << 5)
+#define UART3_IIR_EOF (0x80)
+
+#define IS_FIR(si) ((si)->speed >= 4000000)
+#define IRDA_FRAME_SIZE_LIMIT 4096
+
+static int rx_state = 0; /* RX state for IOCTL */
+
+struct omap_irda {
+ unsigned char open;
+ int speed; /* Current IrDA speed */
+ int newspeed;
+
+ struct net_device_stats stats;
+ struct irlap_cb *irlap;
+ struct qos_info qos;
+
+ int rx_dma_channel;
+ int tx_dma_channel;
+
+ dma_addr_t rx_buf_dma_phys; /* Physical address of RX DMA buffer */
+ dma_addr_t tx_buf_dma_phys; /* Physical address of TX DMA buffer */
+
+ void *rx_buf_dma_virt; /* Virtual address of RX DMA buffer */
+ void *tx_buf_dma_virt; /* Virtual address of TX DMA buffer */
+
+ struct device *dev;
+ struct omap_irda_config *pdata;
+};
+
+static void inline uart_reg_out(int idx, u8 val)
+{
+ omap_writeb(val, idx);
+}
+
+static u8 inline uart_reg_in(int idx)
+{
+ u8 b = omap_readb(idx);
+ return b;
+}
+
+/* forward declarations */
+extern void irda_device_setup(struct net_device *dev);
+extern void omap_stop_dma(int lch);
+static int omap_irda_set_speed(struct net_device *dev, int speed);
+
+static void omap_irda_start_rx_dma(struct omap_irda *si)
+{
+ /* Configure DMA */
+ omap_set_dma_src_params(si->rx_dma_channel, 0x3, 0x0,
+ si->pdata->src_start,
+ 0, 0);
+
+ omap_enable_dma_irq(si->rx_dma_channel, 0x01);
+
+ omap_set_dma_dest_params(si->rx_dma_channel, 0x0, 0x1,
+ si->rx_buf_dma_phys,
+ 0, 0);
+
+ omap_set_dma_transfer_params(si->rx_dma_channel, 0x0,
+ IRDA_FRAME_SIZE_LIMIT, 0x1,
+ 0x0, si->pdata->rx_trigger, 0);
+
+ omap_start_dma(si->rx_dma_channel);
+}
+
+static void omap_start_tx_dma(struct omap_irda *si, int size)
+{
+ /* Configure DMA */
+ omap_set_dma_dest_params(si->tx_dma_channel, 0x03, 0x0,
+ si->pdata->dest_start, 0, 0);
+
+ omap_enable_dma_irq(si->tx_dma_channel, 0x01);
+
+ omap_set_dma_src_params(si->tx_dma_channel, 0x0, 0x1,
+ si->tx_buf_dma_phys,
+ 0, 0);
+
+ omap_set_dma_transfer_params(si->tx_dma_channel, 0x0, size, 0x1,
+ 0x0, si->pdata->tx_trigger, 0);
+
+ /* Start DMA */
+ omap_start_dma(si->tx_dma_channel);
+}
+
+/* DMA RX callback - normally, we should not go here,
+ * it calls only if something is going wrong
+ */
+static void omap_irda_rx_dma_callback(int lch, u16 ch_status, void *data)
+{
+ struct net_device *dev = data;
+ struct omap_irda *si = dev->priv;
+
+ printk(KERN_ERR "RX Transfer error or very big frame\n");
+
+ /* Clear interrupts */
+ uart_reg_in(UART3_IIR);
+
+ si->stats.rx_frame_errors++;
+
+ uart_reg_in(UART3_RESUME);
+
+ /* Re-init RX DMA */
+ omap_irda_start_rx_dma(si);
+}
+
+/* DMA TX callback - calling when frame transfer has been finished */
+static void omap_irda_tx_dma_callback(int lch, u16 ch_status, void *data)
+{
+ struct net_device *dev = data;
+ struct omap_irda *si = dev->priv;
+
+ /*Stop DMA controller */
+ omap_stop_dma(si->tx_dma_channel);
+}
+
+/*
+ * Set the IrDA communications speed.
+ * Interrupt have to be disabled here.
+ */
+static int omap_irda_startup(struct net_device *dev)
+{
+ struct omap_irda *si = dev->priv;
+
+
+ /* FIXME: use clk_* apis for UART3 clock*/
+ /* Enable UART3 clock and set UART3 to IrDA mode */
+ if (machine_is_omap_h2() || machine_is_omap_h3())
+ omap_writel(omap_readl(MOD_CONF_CTRL_0) | (1 << 31) | (1 << 15),
+ MOD_CONF_CTRL_0);
+
+ /* Only for H2?
+ */
+ if (si->pdata->transceiver_mode && machine_is_omap_h2()) {
+ /* Is it select_irda on H2 ? */
+ omap_writel(omap_readl(FUNC_MUX_CTRL_A) | 7,
+ FUNC_MUX_CTRL_A);
+ si->pdata->transceiver_mode(si->dev, IR_SIRMODE);
+ }
+
+ uart_reg_out(UART3_MDR1, UART3_MDR1_RESET); /* Reset mode */
+
+ /* Clear DLH and DLL */
+ uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
+
+ uart_reg_out(UART3_DLL, 0);
+ uart_reg_out(UART3_DLH, 0);
+ uart_reg_out(UART3_LCR, 0xbf); /* FIXME: Add #define */
+
+ uart_reg_out(UART3_EFR, UART3_EFR_EN);
+ uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
+
+ /* Enable access to UART3_TLR and UART3_TCR registers */
+ uart_reg_out(UART3_MCR, UART3_MCR_EN_TCR_TLR);
+
+ uart_reg_out(UART3_SCR, 0);
+ /* Set Rx trigger to 1 and Tx trigger to 1 */
+ uart_reg_out(UART3_TLR, 0);
+
+ /* Set LCR to 8 bits and 1 stop bit */
+ uart_reg_out(UART3_LCR, 0x03);
+
+ /* Clear RX and TX FIFO and enable FIFO */
+ /* Use DMA Req for transfers */
+ uart_reg_out(UART3_FCR, UART3_FCR_CONFIG);
+
+ uart_reg_out(UART3_MCR, 0);
+
+ uart_reg_out(UART3_SCR, UART3_SCR_TX_TRIG1 |
+ UART3_SCR_RX_TRIG1);
+
+ /* Enable UART3 SIR Mode,(Frame-length method to end frames) */
+ uart_reg_out(UART3_MDR1, UART3_MDR1_SIR);
+
+ /* Set Status FIFO trig to 1 */
+ uart_reg_out(UART3_MDR2, 0);
+
+ /* Enables RXIR input */
+ /* and disable TX underrun */
+ /* SEND_SIP pulse */
+ uart_reg_out(UART3_ACREG, UART3_ACERG_SD_MODE_LOW |
+ UART3_ACERG_TX_UNDERRUN_DIS);
+
+ /* Enable EOF Interrupt only */
+ uart_reg_out(UART3_IER, UART3_IER_CTS | UART3_IER_EOF);
+
+ /* Set Maximum Received Frame size to 2048 bytes */
+ uart_reg_out(UART3_RXFLL, 0x00);
+ uart_reg_out(UART3_RXFLH, 0x08);
+
+ uart_reg_in(UART3_RESUME);
+
+ return 0;
+}
+
+static int omap_irda_shutdown(struct omap_irda *si)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+
+ /* Disable all UART3 Interrupts */
+ uart_reg_out(UART3_IER, 0);
+
+ /* Disable UART3 and disable baud rate generator */
+ uart_reg_out(UART3_MDR1, UART3_MDR1_RESET);
+
+ /* set SD_MODE pin to high and Disable RX IR */
+ uart_reg_out(UART3_ACREG, (UART3_ACERG_DIS_IR_RX |
+ ~(UART3_ACERG_SD_MODE_LOW)));
+
+ /* Clear DLH and DLL */
+ uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
+ uart_reg_out(UART3_DLL, 0);
+ uart_reg_out(UART3_DLH, 0);
+
+ local_irq_restore(flags);
+
+ return 0;
+}
+
+static irqreturn_t
+omap_irda_irq(int irq, void *dev_id, struct pt_regs *hw_regs)
+{
+ struct net_device *dev = dev_id;
+ struct omap_irda *si = dev->priv;
+ struct sk_buff *skb;
+
+ u8 status;
+ int w = 0;
+
+ /* Clear EOF interrupt */
+ status = uart_reg_in(UART3_IIR);
+
+ if (status & UART3_IIR_TX_STATUS) {
+ u8 mdr2 = uart_reg_in(UART3_MDR2);
+ if (mdr2 & UART3_MDR2_IRTX_UNDERRUN)
+ printk(KERN_ERR "IrDA Buffer underrun error\n");
+
+ si->stats.tx_packets++;
+
+ if (si->newspeed) {
+ omap_irda_set_speed(dev, si->newspeed);
+ si->newspeed = 0;
+ }
+
+ netif_wake_queue(dev);
+ if (!(status & UART3_IIR_EOF))
+ return IRQ_HANDLED;
+ }
+
+ /* Stop DMA and if there are no errors, send frame to upper layer */
+ omap_stop_dma(si->rx_dma_channel);
+
+ status = uart_reg_in(UART3_SFLSR); /* Take a frame status */
+
+ if (status != 0) { /* Bad frame? */
+ si->stats.rx_frame_errors++;
+ uart_reg_in(UART3_RESUME);
+ } else {
+ /* We got a frame! */
+ skb = alloc_skb(IRDA_FRAME_SIZE_LIMIT, GFP_ATOMIC);
+
+ if (!skb) {
+ printk(KERN_ERR "omap_sir: out of memory for RX SKB\n");
+ return IRQ_HANDLED;
+ }
+ /*
+ * Align any IP headers that may be contained
+ * within the frame.
+ */
+
+ skb_reserve(skb, 1);
+
+ w = OMAP_DMA_CDAC_REG(si->rx_dma_channel);
+
+ if (cpu_is_omap16xx())
+ w -= OMAP1_DMA_CDSA_L_REG(si->rx_dma_channel);
+ if (cpu_is_omap24xx())
+ w -= OMAP2_DMA_CDSA_REG(si->rx_dma_channel);
+
+ if (!IS_FIR(si)) {
+ /* Copy DMA buffer to skb */
+ memcpy(skb_put(skb, w - 2), si->rx_buf_dma_virt, w - 2);
+ } else {
+ /* Copy DMA buffer to skb */
+ memcpy(skb_put(skb, w - 4), si->rx_buf_dma_virt, w - 4);
+ }
+
+ skb->dev = dev;
+ skb->mac.raw = skb->data;
+ skb->protocol = htons(ETH_P_IRDA);
+ si->stats.rx_packets++;
+ si->stats.rx_bytes += skb->len;
+ netif_receive_skb(skb); /* Send data to upper level */
+ }
+
+ /* Re-init RX DMA */
+ omap_irda_start_rx_dma(si);
+
+ dev->last_rx = jiffies;
+
+ return IRQ_HANDLED;
+}
+
+static int omap_irda_hard_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+ struct omap_irda *si = dev->priv;
+ int speed = irda_get_next_speed(skb);
+ int mtt = irda_get_mtt(skb);
+ int xbofs = irda_get_next_xbofs(skb);
+
+
+ /*
+ * Does this packet contain a request to change the interface
+ * speed? If so, remember it until we complete the transmission
+ * of this frame.
+ */
+ if (speed != si->speed && speed != -1)
+ si->newspeed = speed;
+
+ if (xbofs) /* Set number of addtional BOFS */
+ uart_reg_out(UART3_EBLR, xbofs + 1);
+
+ /*
+ * If this is an empty frame, we can bypass a lot.
+ */
+ if (skb->len == 0) {
+ if (si->newspeed) {
+ si->newspeed = 0;
+ omap_irda_set_speed(dev, speed);
+ }
+ dev_kfree_skb(skb);
+ return 0;
+ }
+
+ netif_stop_queue(dev);
+
+ /* Copy skb data to DMA buffer */
+ memcpy(si->tx_buf_dma_virt, skb->data, skb->len);
+
+ /* Copy skb data to DMA buffer */
+ si->stats.tx_bytes += skb->len;
+
+ /* Set frame length */
+ uart_reg_out(UART3_TXFLL, (skb->len & 0xff));
+ uart_reg_out(UART3_TXFLH, (skb->len >> 8));
+
+ if (mtt > 1000)
+ mdelay(mtt / 1000);
+ else
+ udelay(mtt);
+
+ /* Start TX DMA transfer */
+ omap_start_tx_dma(si, skb->len);
+
+ /* We can free skb now because it's already in DMA buffer */
+ dev_kfree_skb(skb);
+
+ dev->trans_start = jiffies;
+
+ return 0;
+}
+
+static int
+omap_irda_ioctl(struct net_device *dev, struct ifreq *ifreq, int cmd)
+{
+ struct if_irda_req *rq = (struct if_irda_req *)ifreq;
+ struct omap_irda *si = dev->priv;
+ int ret = -EOPNOTSUPP;
+
+
+ switch (cmd) {
+ case SIOCSBANDWIDTH:
+ if (capable(CAP_NET_ADMIN)) {
+ /*
+ * We are unable to set the speed if the
+ * device is not running.
+ */
+ if (si->open)
+ ret = omap_irda_set_speed(dev,
+ rq->ifr_baudrate);
+ else {
+ printk(KERN_ERR "omap_irda_ioctl: 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 = rx_state;
+ break;
+
+ default:
+ break;
+ }
+
+ return ret;
+}
+
+static struct net_device_stats *omap_irda_stats(struct net_device *dev)
+{
+ struct omap_irda *si = dev->priv;
+ return &si->stats;
+}
+
+static int omap_irda_start(struct net_device *dev)
+{
+ struct omap_irda *si = dev->priv;
+ int err;
+
+ si->speed = 9600;
+
+ err = request_irq(dev->irq, omap_irda_irq, 0, dev->name, dev);
+ if (err)
+ goto err_irq;
+
+ /*
+ * The interrupt must remain disabled for now.
+ */
+ disable_irq(dev->irq);
+
+ /* Request DMA channels for IrDA hardware */
+ if (omap_request_dma(si->pdata->rx_channel, "IrDA Rx DMA",
+ (void *)omap_irda_rx_dma_callback,
+ dev, &(si->rx_dma_channel))) {
+ printk(KERN_ERR "Failed to request IrDA Rx DMA\n");
+ goto err_irq;
+ }
+
+ if (omap_request_dma(si->pdata->tx_channel, "IrDA Tx DMA",
+ (void *)omap_irda_tx_dma_callback,
+ dev, &(si->tx_dma_channel))) {
+ printk(KERN_ERR "Failed to request IrDA Tx DMA\n");
+ goto err_irq;
+ }
+
+ /* Allocate TX and RX buffers for DMA channels */
+ si->rx_buf_dma_virt =
+ dma_alloc_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
+ &(si->rx_buf_dma_phys),
+ GFP_KERNEL);
+
+ if (!si->rx_buf_dma_virt) {
+ printk(KERN_ERR "Unable to allocate memory for rx_buf_dma\n");
+ goto err_irq;
+ }
+
+ si->tx_buf_dma_virt =
+ dma_alloc_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
+ &(si->tx_buf_dma_phys),
+ GFP_KERNEL);
+
+ if (!si->tx_buf_dma_virt) {
+ printk(KERN_ERR "Unable to allocate memory for tx_buf_dma\n");
+ goto err_mem1;
+ }
+
+ /*
+ * Setup the serial port for the specified config.
+ */
+ if (si->pdata->select_irda)
+ si->pdata->select_irda(si->dev, IR_SEL);
+
+ err = omap_irda_startup(dev);
+
+ if (err)
+ goto err_startup;
+
+ omap_irda_set_speed(dev, si->speed = 9600);
+
+ /*
+ * Open a new IrLAP layer instance.
+ */
+ si->irlap = irlap_open(dev, &si->qos, "omap_sir");
+
+ err = -ENOMEM;
+ if (!si->irlap)
+ goto err_irlap;
+
+ /* Now enable the interrupt and start the queue */
+ si->open = 1;
+
+ /* Start RX DMA */
+ omap_irda_start_rx_dma(si);
+
+ enable_irq(dev->irq);
+ netif_start_queue(dev);
+
+ return 0;
+
+err_irlap:
+ si->open = 0;
+ omap_irda_shutdown(si);
+err_startup:
+ dma_free_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
+ si->tx_buf_dma_virt, si->tx_buf_dma_phys);
+err_mem1:
+ dma_free_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
+ si->rx_buf_dma_virt, si->rx_buf_dma_phys);
+err_irq:
+ free_irq(dev->irq, dev);
+ return err;
+}
+
+static int omap_irda_stop(struct net_device *dev)
+{
+ struct omap_irda *si = dev->priv;
+
+ disable_irq(dev->irq);
+
+ netif_stop_queue(dev);
+
+ omap_free_dma(si->rx_dma_channel);
+ omap_free_dma(si->tx_dma_channel);
+
+ if (si->rx_buf_dma_virt)
+ dma_free_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
+ si->rx_buf_dma_virt, si->rx_buf_dma_phys);
+ if (si->tx_buf_dma_virt)
+ dma_free_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
+ si->tx_buf_dma_virt, si->tx_buf_dma_phys);
+
+ omap_irda_shutdown(si);
+
+ /* Stop IrLAP */
+ if (si->irlap) {
+ irlap_close(si->irlap);
+ si->irlap = NULL;
+ }
+
+ si->open = 0;
+
+ /*
+ * Free resources
+ */
+ free_irq(dev->irq, dev);
+
+ return 0;
+}
+
+static int omap_irda_set_speed(struct net_device *dev, int speed)
+{
+ struct omap_irda *si = dev->priv;
+ int divisor;
+ unsigned long flags;
+
+ /* Set IrDA speed */
+ if (speed <= 115200) {
+
+ local_irq_save(flags);
+
+ /* SIR mode */
+ if (si->pdata->transceiver_mode)
+ si->pdata->transceiver_mode(si->dev, IR_SIRMODE);
+
+ /* Set SIR mode */
+ uart_reg_out(UART3_MDR1, 1);
+ uart_reg_out(UART3_EBLR, 1);
+
+ divisor = 48000000 / (16 * speed); /* Base clock 48 MHz */
+
+ uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
+ uart_reg_out(UART3_DLL, (divisor & 0xff));
+ uart_reg_out(UART3_DLH, (divisor >> 8));
+ uart_reg_out(UART3_LCR, 0x03);
+
+ uart_reg_out(UART3_MCR, 0);
+
+ local_irq_restore(flags);
+ } else if (speed <= 1152000) {
+
+ local_irq_save(flags);
+
+ /* Set MIR mode, auto SIP */
+ uart_reg_out(UART3_MDR1, UART3_MDR1_MIR |
+ UART3_MDR1_SIP_AUTO);
+
+ uart_reg_out(UART3_EBLR, 2);
+
+ divisor = 48000000 / (41 * speed); /* Base clock 48 MHz */
+
+ uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
+ uart_reg_out(UART3_DLL, (divisor & 0xff));
+ uart_reg_out(UART3_DLH, (divisor >> 8));
+ uart_reg_out(UART3_LCR, 0x03);
+
+ if (si->pdata->transceiver_mode)
+ si->pdata->transceiver_mode(si->dev, IR_MIRMODE);
+
+ local_irq_restore(flags);
+ } else {
+ local_irq_save(flags);
+
+ /* FIR mode */
+ uart_reg_out(UART3_MDR1, UART3_MDR1_FIR |
+ UART3_MDR1_SIP_AUTO);
+
+ if (si->pdata->transceiver_mode)
+ si->pdata->transceiver_mode(si->dev, IR_FIRMODE);
+
+ local_irq_restore(flags);
+ }
+
+ si->speed = speed;
+
+ return 0;
+}
+
+#ifdef CONFIG_PM
+/*
+ * Suspend the IrDA interface.
+ */
+static int omap_irda_suspend(struct platform_device *pdev, pm_message_t state)
+{
+ struct net_device *dev = platform_get_drvdata(pdev);
+ struct omap_irda *si = dev->priv;
+
+ if (!dev)
+ return 0;
+
+ if (si->open) {
+ /*
+ * Stop the transmit queue
+ */
+ netif_device_detach(dev);
+ disable_irq(dev->irq);
+ omap_irda_shutdown(si);
+ }
+ return 0;
+}
+
+/*
+ * Resume the IrDA interface.
+ */
+static int omap_irda_resume(struct platform_device *pdev)
+{
+ struct net_device *dev = platform_get_drvdata(pdev);
+ struct omap_irda *si= dev->priv;
+
+ if (!dev)
+ return 0;
+
+ if (si->open) {
+ /*
+ * If we missed a speed change, initialise at the new speed
+ * directly. It is debatable whether this is actually
+ * required, but in the interests of continuing from where
+ * we left off it is desireable. The converse argument is
+ * that we should re-negotiate at 9600 baud again.
+ */
+ if (si->newspeed) {
+ si->speed = si->newspeed;
+ si->newspeed = 0;
+ }
+
+ omap_irda_startup(dev);
+ omap_irda_set_speed(dev, si->speed);
+ enable_irq(dev->irq);
+
+ /*
+ * This automatically wakes up the queue
+ */
+ netif_device_attach(dev);
+ }
+
+ return 0;
+}
+#else
+#define omap_irda_suspend NULL
+#define omap_irda_resume NULL
+#endif
+
+static int omap_irda_probe(struct platform_device *pdev)
+{
+ struct net_device *dev;
+ struct omap_irda *si;
+ struct omap_irda_config *pdata = pdev->dev.platform_data;
+ unsigned int baudrate_mask;
+ int err = 0;
+ int irq = NO_IRQ;
+
+ if (!pdata) {
+ printk(KERN_ERR "IrDA Platform data not supplied\n");
+ return -ENOENT;
+ }
+
+ if (!pdata->rx_channel || !pdata->tx_channel) {
+ printk(KERN_ERR "IrDA invalid rx/tx channel value\n");
+ return -ENOENT;
+ }
+
+ irq = platform_get_irq(pdev, 0);
+ if (irq <= 0) {
+ printk(KERN_WARNING "no irq for IrDA\n");
+ return -ENOENT;
+ }
+
+ dev = alloc_irdadev(sizeof(struct omap_irda));
+ if (!dev)
+ goto err_mem_1;
+
+
+ si = dev->priv;
+ si->dev = &pdev->dev;
+ si->pdata = pdata;
+
+ dev->hard_start_xmit = omap_irda_hard_xmit;
+ dev->open = omap_irda_start;
+ dev->stop = omap_irda_stop;
+ dev->do_ioctl = omap_irda_ioctl;
+ dev->get_stats = omap_irda_stats;
+ dev->irq = irq;
+
+ irda_init_max_qos_capabilies(&si->qos);
+
+ baudrate_mask = 0;
+ if (si->pdata->transceiver_cap & IR_SIRMODE)
+ baudrate_mask |= IR_9600|IR_19200|IR_38400|IR_57600|IR_115200;
+ if (si->pdata->transceiver_cap & IR_MIRMODE)
+ baudrate_mask |= IR_57600 | IR_1152000;
+ if (si->pdata->transceiver_cap & IR_FIRMODE)
+ baudrate_mask |= IR_4000000 << 8;
+
+ si->qos.baud_rate.bits &= baudrate_mask;
+ si->qos.min_turn_time.bits = 7;
+
+ irda_qos_bits_to_value(&si->qos);
+
+ /* Any better way to avoid this? No. */
+ if (machine_is_omap_h3() || machine_is_omap_h4())
+ INIT_WORK(&si->pdata->gpio_expa, NULL, NULL);
+
+ err = register_netdev(dev);
+ if (!err)
+ platform_set_drvdata(pdev, dev);
+ else
+ free_netdev(dev);
+
+err_mem_1:
+ return err;
+}
+
+static int omap_irda_remove(struct platform_device *pdev)
+{
+ struct net_device *dev = platform_get_drvdata(pdev);
+
+ if (pdev) {
+ unregister_netdev(dev);
+ free_netdev(dev);
+ }
+ return 0;
+}
+
+static struct platform_driver omapir_driver = {
+ .probe = omap_irda_probe,
+ .remove = omap_irda_remove,
+ .suspend = omap_irda_suspend,
+ .resume = omap_irda_resume,
+ .driver = {
+ .name = "omapirda",
+ },
+};
+
+static char __initdata banner[] = KERN_INFO "OMAP IrDA driver initializing\n";
+
+static int __init omap_irda_init(void)
+{
+ printk(banner);
+ return platform_driver_register(&omapir_driver);
+}
+
+static void __exit omap_irda_exit(void)
+{
+ platform_driver_unregister(&omapir_driver);
+}
+
+module_init(omap_irda_init);
+module_exit(omap_irda_exit);
+
+MODULE_AUTHOR("MontaVista");
+MODULE_DESCRIPTION("OMAP IrDA Driver");
+MODULE_LICENSE("GPL");
+
--
1.3.3
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] OMAP: Add support for IrDA driver
2006-08-07 9:12 [PATCH] OMAP: Add support for IrDA driver Komal Shah
@ 2006-08-11 5:47 ` Samuel Ortiz
2006-08-11 9:20 ` [irda-users] " Komal Shah
0 siblings, 1 reply; 3+ messages in thread
From: Samuel Ortiz @ 2006-08-11 5:47 UTC (permalink / raw)
To: Komal Shah; +Cc: samuel, irda-users, tony, akpm, r-woodruff2, linux-kernel
Hi Komal,
On Mon, Aug 07, 2006 at 02:42:33PM +0530, Komal Shah wrote:
> Samuel/Tony,
>
> I have attached the OMAP IrDA patch for OMAP1610/1710 and OMAP242x
> for review. See that for IrDA driver to work on H3 and H4, we need
> gpio-expander functions to go in mainline, but it can work on H2.
Overall the patch looks good, I have some comments though. See below:
> +++ b/drivers/net/irda/Kconfig
> @@ -432,5 +432,15 @@ config MCS_FIR
> To compile it as a module, choose M here: the module will be called
> mcs7780.
>
> +config OMAP_IR
> + tristate "OMAP IrDA(SIR/MIR/FIR)"
> + depends on IRDA && ARCH_OMAP
> + select GPIOEXPANDER_OMAP if (MACH_OMAP_H3 || MACH_OMAP_H4)
> + help
> + Say Y here if you want to build support for the Teaxas Instruments
s/Teaxas/Texas/ ;-)
> diff --git a/drivers/net/irda/Makefile b/drivers/net/irda/Makefile
> index 5be09f1..89c6389 100644
> --- a/drivers/net/irda/Makefile
> +++ b/drivers/net/irda/Makefile
> @@ -20,6 +20,7 @@ obj-$(CONFIG_VLSI_FIR) += vlsi_ir.o
> obj-$(CONFIG_VIA_FIR) += via-ircc.o
> obj-$(CONFIG_PXA_FICP) += pxaficp_ir.o
> obj-$(CONFIG_MCS_FIR) += mcs7780.o
> +obj-$(CONFIG_OMAP_IR) += omap-ir.o
> # Old dongle drivers for old SIR drivers
> obj-$(CONFIG_ESI_DONGLE_OLD) += esi.o
> obj-$(CONFIG_TEKRAM_DONGLE_OLD) += tekram.o
> diff --git a/drivers/net/irda/omap-ir.c b/drivers/net/irda/omap-ir.c
> new file mode 100644
> index 0000000..c51fb48
> --- /dev/null
> +++ b/drivers/net/irda/omap-ir.c
> @@ -0,0 +1,903 @@
> +/*
> + * BRIEF MODULE DESCRIPTION
> + *
> + * Infra-red driver for the OMAP1610-H2 and OMAP1710-H3 and H4 Platforms
> + * (SIR/MIR/FIR modes)
> + * (based on omap-sir.c)
> + *
> + * Copyright 2003 MontaVista Software Inc.
> + * Author: MontaVista Software, Inc.
> + * source@mvista.com
> + *
> + * Copyright 2004 Texas Instruments.
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms of the GNU General Public License as published by the
> + * Free Software Foundation; either version 2 of the License, or (at your
> + * option) any later version.
> + *
> + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
> + * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
> + * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN
> + * NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
> + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
> + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
> + * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
> + * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
> + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
> + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
> + *
> + * You should have received a copy of the GNU General Public License along
> + * with this program; if not, write to the Free Software Foundation, Inc.,
> + * 675 Mass Ave, Cambridge, MA 02139, USA.
> + *
> + Modifications:
> + Feb 2004, Texas Instruments
> + - Ported to 2.6 kernel (Feb 2004).
> + *
> + Apr 2004, Texas Instruments
> + - Added support for H3 (Apr 2004).
> + Nov 2004, Texas Instruments
> + - Added support for Power Management.
> + */
> +
> +#include <linux/config.h>
> +#include <linux/module.h>
> +#include <linux/types.h>
> +#include <linux/init.h>
> +#include <linux/errno.h>
> +#include <linux/netdevice.h>
> +#include <linux/slab.h>
> +#include <linux/rtnetlink.h>
> +#include <linux/interrupt.h>
> +#include <linux/delay.h>
> +#include <linux/ioport.h>
> +#include <linux/dma-mapping.h>
> +#include <linux/platform_device.h>
> +#include <linux/i2c.h>
> +
> +#include <net/irda/irda.h>
> +#include <net/irda/irmod.h>
> +#include <net/irda/wrapper.h>
> +#include <net/irda/irda_device.h>
> +
> +#include <asm/irq.h>
> +#include <asm/io.h>
> +#include <asm/hardware.h>
> +#include <asm/serial.h>
> +#include <asm/mach-types.h>
> +#include <asm/dma.h>
> +#include <asm/arch/mux.h>
> +#include <asm/arch/gpio.h>
> +#include <asm/arch/irda.h>
> +
> +#define UART3_EFR_EN (1 << 4)
> +#define UART3_MCR_EN_TCR_TLR (1 << 6)
> +
> +#define UART3_LCR_WL_8 (3 << 0)
> +#define UART3_LCR_SP2 (1 << 2)
> +#define UART3_LCR_DIVEN (1 << 7)
> +
> +#define UART3_FCR_FIFO_EN (1 << 0)
> +#define UART3_FCR_FIFO_RX (1 << 1)
> +#define UART3_FCR_FIFO_TX (1 << 2)
> +#define UART3_FCR_FIFO_DMA1 (1 << 3)
> +#define UART3_FCR_FIFO_TX_TRIG16 (1 << 4)
> +#define UART3_FCR_FIFO_RX_TRIG16 (1 << 6)
> +#define UART3_FCR_CONFIG (\
> + UART3_FCR_FIFO_EN | UART3_FCR_FIFO_RX |\
> + UART3_FCR_FIFO_TX | UART3_FCR_FIFO_DMA1 |\
> + UART3_FCR_FIFO_TX_TRIG16 |\
> + UART3_FCR_FIFO_RX_TRIG16)
> +
> +#define UART3_SCR_TX_TRIG1 (1 << 6)
> +#define UART3_SCR_RX_TRIG1 (1 << 7)
> +
> +#define UART3_MDR1_RESET (0x07)
> +#define UART3_MDR1_SIR (1 << 0)
> +#define UART3_MDR1_MIR (4 << 0)
> +#define UART3_MDR1_FIR (5 << 0)
> +#define UART3_MDR1_SIP_AUTO (1 << 6)
> +
> +#define UART3_MDR2_TRIG1 (0 << 1)
> +#define UART3_MDR2_IRTX_UNDERRUN (1 << 0)
> +
> +#define UART3_ACERG_TX_UNDERRUN_DIS (1 << 4)
> +#define UART3_ACERG_SD_MODE_LOW (1 << 6)
> +#define UART3_ACERG_DIS_IR_RX (1 << 5)
> +
> +#define UART3_IER_EOF (1 << 5)
> +#define UART3_IER_CTS (1 << 7)
> +
> +#define UART3_IIR_TX_STATUS (1 << 5)
> +#define UART3_IIR_EOF (0x80)
> +
> +#define IS_FIR(si) ((si)->speed >= 4000000)
> +#define IRDA_FRAME_SIZE_LIMIT 4096
You should use IRDA_SKB_MAX_MTU and IRDA_SIR_MAX_FRAME for respectively your
max Rx and Tx sizes.
> +static int rx_state = 0; /* RX state for IOCTL */
rx_state could probably be part of omap_irda.
> +struct omap_irda {
> + unsigned char open;
> + int speed; /* Current IrDA speed */
> + int newspeed;
> +
> + struct net_device_stats stats;
> + struct irlap_cb *irlap;
> + struct qos_info qos;
> +
> + int rx_dma_channel;
> + int tx_dma_channel;
> +
> + dma_addr_t rx_buf_dma_phys; /* Physical address of RX DMA buffer */
> + dma_addr_t tx_buf_dma_phys; /* Physical address of TX DMA buffer */
> +
> + void *rx_buf_dma_virt; /* Virtual address of RX DMA buffer */
> + void *tx_buf_dma_virt; /* Virtual address of TX DMA buffer */
> +
> + struct device *dev;
> + struct omap_irda_config *pdata;
You forgot to define omap_irda_config.
Same thing for IR_SEL, IR_SIRMODE, IR_FIRMODE and IR_MIRMODE.
> +};
> +
> +static void inline uart_reg_out(int idx, u8 val)
> +{
> + omap_writeb(val, idx);
> +}
> +
> +static u8 inline uart_reg_in(int idx)
> +{
> + u8 b = omap_readb(idx);
> + return b;
> +}
> +
> +/* forward declarations */
> +extern void irda_device_setup(struct net_device *dev);
This is useless as you don't use this function anywhere in the code below.
> +extern void omap_stop_dma(int lch);
> +static int omap_irda_set_speed(struct net_device *dev, int speed);
> +
> +static void omap_irda_start_rx_dma(struct omap_irda *si)
> +{
> + /* Configure DMA */
> + omap_set_dma_src_params(si->rx_dma_channel, 0x3, 0x0,
> + si->pdata->src_start,
> + 0, 0);
> +
> + omap_enable_dma_irq(si->rx_dma_channel, 0x01);
> +
> + omap_set_dma_dest_params(si->rx_dma_channel, 0x0, 0x1,
> + si->rx_buf_dma_phys,
> + 0, 0);
> +
> + omap_set_dma_transfer_params(si->rx_dma_channel, 0x0,
> + IRDA_FRAME_SIZE_LIMIT, 0x1,
> + 0x0, si->pdata->rx_trigger, 0);
> + omap_start_dma(si->rx_dma_channel);
> +}
> +
> +static void omap_start_tx_dma(struct omap_irda *si, int size)
> +{
> + /* Configure DMA */
> + omap_set_dma_dest_params(si->tx_dma_channel, 0x03, 0x0,
> + si->pdata->dest_start, 0, 0);
> +
> + omap_enable_dma_irq(si->tx_dma_channel, 0x01);
> +
> + omap_set_dma_src_params(si->tx_dma_channel, 0x0, 0x1,
> + si->tx_buf_dma_phys,
> + 0, 0);
> +
> + omap_set_dma_transfer_params(si->tx_dma_channel, 0x0, size, 0x1,
> + 0x0, si->pdata->tx_trigger, 0);
> +
> + /* Start DMA */
> + omap_start_dma(si->tx_dma_channel);
> +}
> +
> +/* DMA RX callback - normally, we should not go here,
> + * it calls only if something is going wrong
> + */
> +static void omap_irda_rx_dma_callback(int lch, u16 ch_status, void *data)
> +{
> + struct net_device *dev = data;
> + struct omap_irda *si = dev->priv;
A detail here, but I think that "si" for a variable name is a bit meaningless.
> + printk(KERN_ERR "RX Transfer error or very big frame\n");
> +
> + /* Clear interrupts */
> + uart_reg_in(UART3_IIR);
> +
> + si->stats.rx_frame_errors++;
> +
> + uart_reg_in(UART3_RESUME);
> +
> + /* Re-init RX DMA */
> + omap_irda_start_rx_dma(si);
> +}
> +
> +/* DMA TX callback - calling when frame transfer has been finished */
> +static void omap_irda_tx_dma_callback(int lch, u16 ch_status, void *data)
> +{
> + struct net_device *dev = data;
> + struct omap_irda *si = dev->priv;
> +
> + /*Stop DMA controller */
> + omap_stop_dma(si->tx_dma_channel);
> +}
> +
> +/*
> + * Set the IrDA communications speed.
> + * Interrupt have to be disabled here.
> + */
> +static int omap_irda_startup(struct net_device *dev)
> +{
> + struct omap_irda *si = dev->priv;
> +
> +
> + /* FIXME: use clk_* apis for UART3 clock*/
> + /* Enable UART3 clock and set UART3 to IrDA mode */
> + if (machine_is_omap_h2() || machine_is_omap_h3())
> + omap_writel(omap_readl(MOD_CONF_CTRL_0) | (1 << 31) | (1 << 15),
> + MOD_CONF_CTRL_0);
> +
> + /* Only for H2?
> + */
> + if (si->pdata->transceiver_mode && machine_is_omap_h2()) {
> + /* Is it select_irda on H2 ? */
> + omap_writel(omap_readl(FUNC_MUX_CTRL_A) | 7,
> + FUNC_MUX_CTRL_A);
> + si->pdata->transceiver_mode(si->dev, IR_SIRMODE);
> + }
> +
> + uart_reg_out(UART3_MDR1, UART3_MDR1_RESET); /* Reset mode */
> +
> + /* Clear DLH and DLL */
> + uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
> +
> + uart_reg_out(UART3_DLL, 0);
> + uart_reg_out(UART3_DLH, 0);
> + uart_reg_out(UART3_LCR, 0xbf); /* FIXME: Add #define */
> +
> + uart_reg_out(UART3_EFR, UART3_EFR_EN);
> + uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
> +
> + /* Enable access to UART3_TLR and UART3_TCR registers */
> + uart_reg_out(UART3_MCR, UART3_MCR_EN_TCR_TLR);
> +
> + uart_reg_out(UART3_SCR, 0);
> + /* Set Rx trigger to 1 and Tx trigger to 1 */
> + uart_reg_out(UART3_TLR, 0);
> +
> + /* Set LCR to 8 bits and 1 stop bit */
> + uart_reg_out(UART3_LCR, 0x03);
> +
> + /* Clear RX and TX FIFO and enable FIFO */
> + /* Use DMA Req for transfers */
> + uart_reg_out(UART3_FCR, UART3_FCR_CONFIG);
> +
> + uart_reg_out(UART3_MCR, 0);
> +
> + uart_reg_out(UART3_SCR, UART3_SCR_TX_TRIG1 |
> + UART3_SCR_RX_TRIG1);
> +
> + /* Enable UART3 SIR Mode,(Frame-length method to end frames) */
> + uart_reg_out(UART3_MDR1, UART3_MDR1_SIR);
> +
> + /* Set Status FIFO trig to 1 */
> + uart_reg_out(UART3_MDR2, 0);
> +
> + /* Enables RXIR input */
> + /* and disable TX underrun */
> + /* SEND_SIP pulse */
> + uart_reg_out(UART3_ACREG, UART3_ACERG_SD_MODE_LOW |
> + UART3_ACERG_TX_UNDERRUN_DIS);
> +
> + /* Enable EOF Interrupt only */
> + uart_reg_out(UART3_IER, UART3_IER_CTS | UART3_IER_EOF);
> +
> + /* Set Maximum Received Frame size to 2048 bytes */
> + uart_reg_out(UART3_RXFLL, 0x00);
> + uart_reg_out(UART3_RXFLH, 0x08);
> +
> + uart_reg_in(UART3_RESUME);
> +
> + return 0;
> +}
> +
> +static int omap_irda_shutdown(struct omap_irda *si)
> +{
> + unsigned long flags;
> +
> + local_irq_save(flags);
> +
> + /* Disable all UART3 Interrupts */
> + uart_reg_out(UART3_IER, 0);
> +
> + /* Disable UART3 and disable baud rate generator */
> + uart_reg_out(UART3_MDR1, UART3_MDR1_RESET);
> +
> + /* set SD_MODE pin to high and Disable RX IR */
> + uart_reg_out(UART3_ACREG, (UART3_ACERG_DIS_IR_RX |
> + ~(UART3_ACERG_SD_MODE_LOW)));
> +
> + /* Clear DLH and DLL */
> + uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
> + uart_reg_out(UART3_DLL, 0);
> + uart_reg_out(UART3_DLH, 0);
> +
> + local_irq_restore(flags);
> +
> + return 0;
> +}
> +
> +static irqreturn_t
> +omap_irda_irq(int irq, void *dev_id, struct pt_regs *hw_regs)
> +{
> + struct net_device *dev = dev_id;
> + struct omap_irda *si = dev->priv;
> + struct sk_buff *skb;
> +
> + u8 status;
> + int w = 0;
> +
> + /* Clear EOF interrupt */
> + status = uart_reg_in(UART3_IIR);
> +
> + if (status & UART3_IIR_TX_STATUS) {
> + u8 mdr2 = uart_reg_in(UART3_MDR2);
> + if (mdr2 & UART3_MDR2_IRTX_UNDERRUN)
> + printk(KERN_ERR "IrDA Buffer underrun error\n");
> +
> + si->stats.tx_packets++;
> +
> + if (si->newspeed) {
> + omap_irda_set_speed(dev, si->newspeed);
> + si->newspeed = 0;
> + }
> +
> + netif_wake_queue(dev);
> + if (!(status & UART3_IIR_EOF))
> + return IRQ_HANDLED;
> + }
> +
> + /* Stop DMA and if there are no errors, send frame to upper layer */
> + omap_stop_dma(si->rx_dma_channel);
> +
> + status = uart_reg_in(UART3_SFLSR); /* Take a frame status */
> +
> + if (status != 0) { /* Bad frame? */
> + si->stats.rx_frame_errors++;
> + uart_reg_in(UART3_RESUME);
> + } else {
> + /* We got a frame! */
> + skb = alloc_skb(IRDA_FRAME_SIZE_LIMIT, GFP_ATOMIC);
Use dev_alloc_skb() for allocating RX packets.
> +
> + if (!skb) {
> + printk(KERN_ERR "omap_sir: out of memory for RX SKB\n");
> + return IRQ_HANDLED;
> + }
> + /*
> + * Align any IP headers that may be contained
> + * within the frame.
> + */
> +
> + skb_reserve(skb, 1);
> +
> + w = OMAP_DMA_CDAC_REG(si->rx_dma_channel);
> +
> + if (cpu_is_omap16xx())
> + w -= OMAP1_DMA_CDSA_L_REG(si->rx_dma_channel);
> + if (cpu_is_omap24xx())
> + w -= OMAP2_DMA_CDSA_REG(si->rx_dma_channel);
> +
> + if (!IS_FIR(si)) {
> + /* Copy DMA buffer to skb */
> + memcpy(skb_put(skb, w - 2), si->rx_buf_dma_virt, w - 2);
> + } else {
> + /* Copy DMA buffer to skb */
> + memcpy(skb_put(skb, w - 4), si->rx_buf_dma_virt, w - 4);
> + }
> +
> + skb->dev = dev;
> + skb->mac.raw = skb->data;
> + skb->protocol = htons(ETH_P_IRDA);
> + si->stats.rx_packets++;
> + si->stats.rx_bytes += skb->len;
> + netif_receive_skb(skb); /* Send data to upper level */
> + }
> +
> + /* Re-init RX DMA */
> + omap_irda_start_rx_dma(si);
> +
> + dev->last_rx = jiffies;
> +
> + return IRQ_HANDLED;
> +}
> +
> +static int omap_irda_hard_xmit(struct sk_buff *skb, struct net_device *dev)
> +{
> + struct omap_irda *si = dev->priv;
> + int speed = irda_get_next_speed(skb);
> + int mtt = irda_get_mtt(skb);
> + int xbofs = irda_get_next_xbofs(skb);
> +
> +
> + /*
> + * Does this packet contain a request to change the interface
> + * speed? If so, remember it until we complete the transmission
> + * of this frame.
> + */
> + if (speed != si->speed && speed != -1)
> + si->newspeed = speed;
> +
> + if (xbofs) /* Set number of addtional BOFS */
> + uart_reg_out(UART3_EBLR, xbofs + 1);
> +
> + /*
> + * If this is an empty frame, we can bypass a lot.
> + */
> + if (skb->len == 0) {
> + if (si->newspeed) {
> + si->newspeed = 0;
> + omap_irda_set_speed(dev, speed);
> + }
> + dev_kfree_skb(skb);
> + return 0;
> + }
> +
> + netif_stop_queue(dev);
> +
> + /* Copy skb data to DMA buffer */
> + memcpy(si->tx_buf_dma_virt, skb->data, skb->len);
> +
> + /* Copy skb data to DMA buffer */
> + si->stats.tx_bytes += skb->len;
> +
> + /* Set frame length */
> + uart_reg_out(UART3_TXFLL, (skb->len & 0xff));
> + uart_reg_out(UART3_TXFLH, (skb->len >> 8));
> +
> + if (mtt > 1000)
> + mdelay(mtt / 1000);
> + else
> + udelay(mtt);
> +
> + /* Start TX DMA transfer */
> + omap_start_tx_dma(si, skb->len);
> +
> + /* We can free skb now because it's already in DMA buffer */
> + dev_kfree_skb(skb);
> +
> + dev->trans_start = jiffies;
> +
> + return 0;
> +}
> +
> +static int
> +omap_irda_ioctl(struct net_device *dev, struct ifreq *ifreq, int cmd)
> +{
> + struct if_irda_req *rq = (struct if_irda_req *)ifreq;
> + struct omap_irda *si = dev->priv;
> + int ret = -EOPNOTSUPP;
> +
> +
> + switch (cmd) {
> + case SIOCSBANDWIDTH:
> + if (capable(CAP_NET_ADMIN)) {
> + /*
> + * We are unable to set the speed if the
> + * device is not running.
> + */
> + if (si->open)
> + ret = omap_irda_set_speed(dev,
> + rq->ifr_baudrate);
> + else {
> + printk(KERN_ERR "omap_irda_ioctl: 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 = rx_state;
> + break;
rx_state is always set to 0, so this looks a bit useless.
> + default:
> + break;
> + }
> +
> + return ret;
> +}
> +
> +static struct net_device_stats *omap_irda_stats(struct net_device *dev)
> +{
> + struct omap_irda *si = dev->priv;
> + return &si->stats;
> +}
> +
> +static int omap_irda_start(struct net_device *dev)
> +{
> + struct omap_irda *si = dev->priv;
> + int err;
> +
> + si->speed = 9600;
> +
> + err = request_irq(dev->irq, omap_irda_irq, 0, dev->name, dev);
> + if (err)
> + goto err_irq;
> +
> + /*
> + * The interrupt must remain disabled for now.
> + */
> + disable_irq(dev->irq);
> +
> + /* Request DMA channels for IrDA hardware */
> + if (omap_request_dma(si->pdata->rx_channel, "IrDA Rx DMA",
> + (void *)omap_irda_rx_dma_callback,
> + dev, &(si->rx_dma_channel))) {
> + printk(KERN_ERR "Failed to request IrDA Rx DMA\n");
> + goto err_irq;
> + }
> +
> + if (omap_request_dma(si->pdata->tx_channel, "IrDA Tx DMA",
> + (void *)omap_irda_tx_dma_callback,
> + dev, &(si->tx_dma_channel))) {
> + printk(KERN_ERR "Failed to request IrDA Tx DMA\n");
> + goto err_irq;
> + }
> +
> + /* Allocate TX and RX buffers for DMA channels */
> + si->rx_buf_dma_virt =
> + dma_alloc_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
> + &(si->rx_buf_dma_phys),
> + GFP_KERNEL);
> +
> + if (!si->rx_buf_dma_virt) {
> + printk(KERN_ERR "Unable to allocate memory for rx_buf_dma\n");
> + goto err_irq;
> + }
> +
> + si->tx_buf_dma_virt =
> + dma_alloc_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
> + &(si->tx_buf_dma_phys),
> + GFP_KERNEL);
> +
> + if (!si->tx_buf_dma_virt) {
> + printk(KERN_ERR "Unable to allocate memory for tx_buf_dma\n");
> + goto err_mem1;
> + }
> +
> + /*
> + * Setup the serial port for the specified config.
> + */
> + if (si->pdata->select_irda)
> + si->pdata->select_irda(si->dev, IR_SEL);
> +
> + err = omap_irda_startup(dev);
> +
> + if (err)
> + goto err_startup;
> +
> + omap_irda_set_speed(dev, si->speed = 9600);
> +
> + /*
> + * Open a new IrLAP layer instance.
> + */
> + si->irlap = irlap_open(dev, &si->qos, "omap_sir");
> +
> + err = -ENOMEM;
> + if (!si->irlap)
> + goto err_irlap;
> +
> + /* Now enable the interrupt and start the queue */
> + si->open = 1;
> +
> + /* Start RX DMA */
> + omap_irda_start_rx_dma(si);
> +
> + enable_irq(dev->irq);
> + netif_start_queue(dev);
> +
> + return 0;
> +
> +err_irlap:
> + si->open = 0;
> + omap_irda_shutdown(si);
> +err_startup:
> + dma_free_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
> + si->tx_buf_dma_virt, si->tx_buf_dma_phys);
> +err_mem1:
> + dma_free_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
> + si->rx_buf_dma_virt, si->rx_buf_dma_phys);
> +err_irq:
> + free_irq(dev->irq, dev);
> + return err;
> +}
> +
> +static int omap_irda_stop(struct net_device *dev)
> +{
> + struct omap_irda *si = dev->priv;
> +
> + disable_irq(dev->irq);
> +
> + netif_stop_queue(dev);
> +
> + omap_free_dma(si->rx_dma_channel);
> + omap_free_dma(si->tx_dma_channel);
> +
> + if (si->rx_buf_dma_virt)
> + dma_free_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
> + si->rx_buf_dma_virt, si->rx_buf_dma_phys);
> + if (si->tx_buf_dma_virt)
> + dma_free_coherent(NULL, IRDA_FRAME_SIZE_LIMIT,
> + si->tx_buf_dma_virt, si->tx_buf_dma_phys);
> +
> + omap_irda_shutdown(si);
> +
> + /* Stop IrLAP */
> + if (si->irlap) {
> + irlap_close(si->irlap);
> + si->irlap = NULL;
> + }
> +
> + si->open = 0;
> +
> + /*
> + * Free resources
> + */
> + free_irq(dev->irq, dev);
> +
> + return 0;
> +}
> +
> +static int omap_irda_set_speed(struct net_device *dev, int speed)
> +{
> + struct omap_irda *si = dev->priv;
> + int divisor;
> + unsigned long flags;
> +
> + /* Set IrDA speed */
> + if (speed <= 115200) {
> +
> + local_irq_save(flags);
> +
> + /* SIR mode */
> + if (si->pdata->transceiver_mode)
> + si->pdata->transceiver_mode(si->dev, IR_SIRMODE);
> +
> + /* Set SIR mode */
> + uart_reg_out(UART3_MDR1, 1);
> + uart_reg_out(UART3_EBLR, 1);
> +
> + divisor = 48000000 / (16 * speed); /* Base clock 48 MHz */
> +
> + uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
> + uart_reg_out(UART3_DLL, (divisor & 0xff));
> + uart_reg_out(UART3_DLH, (divisor >> 8));
> + uart_reg_out(UART3_LCR, 0x03);
> +
> + uart_reg_out(UART3_MCR, 0);
> +
> + local_irq_restore(flags);
> + } else if (speed <= 1152000) {
> +
> + local_irq_save(flags);
> +
> + /* Set MIR mode, auto SIP */
> + uart_reg_out(UART3_MDR1, UART3_MDR1_MIR |
> + UART3_MDR1_SIP_AUTO);
> +
> + uart_reg_out(UART3_EBLR, 2);
> +
> + divisor = 48000000 / (41 * speed); /* Base clock 48 MHz */
> +
> + uart_reg_out(UART3_LCR, UART3_LCR_DIVEN);
> + uart_reg_out(UART3_DLL, (divisor & 0xff));
> + uart_reg_out(UART3_DLH, (divisor >> 8));
> + uart_reg_out(UART3_LCR, 0x03);
> +
> + if (si->pdata->transceiver_mode)
> + si->pdata->transceiver_mode(si->dev, IR_MIRMODE);
> +
> + local_irq_restore(flags);
> + } else {
> + local_irq_save(flags);
> +
> + /* FIR mode */
> + uart_reg_out(UART3_MDR1, UART3_MDR1_FIR |
> + UART3_MDR1_SIP_AUTO);
> +
> + if (si->pdata->transceiver_mode)
> + si->pdata->transceiver_mode(si->dev, IR_FIRMODE);
> +
> + local_irq_restore(flags);
> + }
> +
> + si->speed = speed;
> +
> + return 0;
> +}
> +
> +#ifdef CONFIG_PM
> +/*
> + * Suspend the IrDA interface.
> + */
> +static int omap_irda_suspend(struct platform_device *pdev, pm_message_t state)
> +{
> + struct net_device *dev = platform_get_drvdata(pdev);
> + struct omap_irda *si = dev->priv;
> +
> + if (!dev)
> + return 0;
> +
> + if (si->open) {
> + /*
> + * Stop the transmit queue
> + */
> + netif_device_detach(dev);
> + disable_irq(dev->irq);
> + omap_irda_shutdown(si);
> + }
> + return 0;
> +}
> +
> +/*
> + * Resume the IrDA interface.
> + */
> +static int omap_irda_resume(struct platform_device *pdev)
> +{
> + struct net_device *dev = platform_get_drvdata(pdev);
> + struct omap_irda *si= dev->priv;
> +
> + if (!dev)
> + return 0;
> +
> + if (si->open) {
> + /*
> + * If we missed a speed change, initialise at the new speed
> + * directly. It is debatable whether this is actually
> + * required, but in the interests of continuing from where
> + * we left off it is desireable. The converse argument is
> + * that we should re-negotiate at 9600 baud again.
> + */
> + if (si->newspeed) {
> + si->speed = si->newspeed;
> + si->newspeed = 0;
> + }
> +
> + omap_irda_startup(dev);
> + omap_irda_set_speed(dev, si->speed);
> + enable_irq(dev->irq);
> +
> + /*
> + * This automatically wakes up the queue
> + */
> + netif_device_attach(dev);
> + }
> +
> + return 0;
> +}
> +#else
> +#define omap_irda_suspend NULL
> +#define omap_irda_resume NULL
> +#endif
> +
> +static int omap_irda_probe(struct platform_device *pdev)
> +{
> + struct net_device *dev;
> + struct omap_irda *si;
> + struct omap_irda_config *pdata = pdev->dev.platform_data;
> + unsigned int baudrate_mask;
> + int err = 0;
> + int irq = NO_IRQ;
> +
> + if (!pdata) {
> + printk(KERN_ERR "IrDA Platform data not supplied\n");
> + return -ENOENT;
> + }
> +
> + if (!pdata->rx_channel || !pdata->tx_channel) {
> + printk(KERN_ERR "IrDA invalid rx/tx channel value\n");
> + return -ENOENT;
> + }
I see that rx_channel and tx_channel are part of your omap_irda_config, which
seems to be a platform specific structure. However, DMA channels are
architecture specific and could be removed from this structure, isn't it ?
> +
> + irq = platform_get_irq(pdev, 0);
> + if (irq <= 0) {
> + printk(KERN_WARNING "no irq for IrDA\n");
> + return -ENOENT;
> + }
> +
> + dev = alloc_irdadev(sizeof(struct omap_irda));
> + if (!dev)
> + goto err_mem_1;
> +
> +
> + si = dev->priv;
> + si->dev = &pdev->dev;
> + si->pdata = pdata;
> +
> + dev->hard_start_xmit = omap_irda_hard_xmit;
> + dev->open = omap_irda_start;
> + dev->stop = omap_irda_stop;
> + dev->do_ioctl = omap_irda_ioctl;
> + dev->get_stats = omap_irda_stats;
> + dev->irq = irq;
> +
> + irda_init_max_qos_capabilies(&si->qos);
> +
> + baudrate_mask = 0;
> + if (si->pdata->transceiver_cap & IR_SIRMODE)
> + baudrate_mask |= IR_9600|IR_19200|IR_38400|IR_57600|IR_115200;
> + if (si->pdata->transceiver_cap & IR_MIRMODE)
> + baudrate_mask |= IR_57600 | IR_1152000;
> + if (si->pdata->transceiver_cap & IR_FIRMODE)
> + baudrate_mask |= IR_4000000 << 8;
> +
> + si->qos.baud_rate.bits &= baudrate_mask;
> + si->qos.min_turn_time.bits = 7;
> +
> + irda_qos_bits_to_value(&si->qos);
> +
> + /* Any better way to avoid this? No. */
> + if (machine_is_omap_h3() || machine_is_omap_h4())
> + INIT_WORK(&si->pdata->gpio_expa, NULL, NULL);
What is this for ?
> + err = register_netdev(dev);
> + if (!err)
> + platform_set_drvdata(pdev, dev);
> + else
> + free_netdev(dev);
> +
> +err_mem_1:
> + return err;
> +}
> +
> +static int omap_irda_remove(struct platform_device *pdev)
> +{
> + struct net_device *dev = platform_get_drvdata(pdev);
> +
> + if (pdev) {
> + unregister_netdev(dev);
> + free_netdev(dev);
> + }
> + return 0;
> +}
> +
> +static struct platform_driver omapir_driver = {
> + .probe = omap_irda_probe,
> + .remove = omap_irda_remove,
> + .suspend = omap_irda_suspend,
> + .resume = omap_irda_resume,
> + .driver = {
> + .name = "omapirda",
> + },
> +};
> +
> +static char __initdata banner[] = KERN_INFO "OMAP IrDA driver initializing\n";
> +
> +static int __init omap_irda_init(void)
> +{
> + printk(banner);
> + return platform_driver_register(&omapir_driver);
> +}
> +
> +static void __exit omap_irda_exit(void)
> +{
> + platform_driver_unregister(&omapir_driver);
> +}
> +
> +module_init(omap_irda_init);
> +module_exit(omap_irda_exit);
> +
> +MODULE_AUTHOR("MontaVista");
> +MODULE_DESCRIPTION("OMAP IrDA Driver");
> +MODULE_LICENSE("GPL");
> +
> --
> 1.3.3
Cheers,
Samuel.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [irda-users] [PATCH] OMAP: Add support for IrDA driver
2006-08-11 5:47 ` Samuel Ortiz
@ 2006-08-11 9:20 ` Komal Shah
0 siblings, 0 replies; 3+ messages in thread
From: Komal Shah @ 2006-08-11 9:20 UTC (permalink / raw)
To: Samuel Ortiz; +Cc: akpm, samuel, tony, linux-kernel, irda-users, r-woodruff2
--- Samuel Ortiz <samuel@sortiz.org> wrote:
> Hi Komal,
>
> On Mon, Aug 07, 2006 at 02:42:33PM +0530, Komal Shah wrote:
> > Samuel/Tony,
> >
> > I have attached the OMAP IrDA patch for OMAP1610/1710 and OMAP242x
> > for review. See that for IrDA driver to work on H3 and H4, we need
> > gpio-expander functions to go in mainline, but it can work on H2.
> Overall the patch looks good, I have some comments though. See below:
Thanx for the review.
> > +
> > +#define UART3_IIR_TX_STATUS (1 << 5)
> > +#define UART3_IIR_EOF (0x80)
> > +
> > +#define IS_FIR(si) ((si)->speed >= 4000000)
> > +#define IRDA_FRAME_SIZE_LIMIT 4096
> You should use IRDA_SKB_MAX_MTU and IRDA_SIR_MAX_FRAME for
> respectively your
> max Rx and Tx sizes.
I will update this and to all the places where FRAME_SIZE_LIMIT was
used as default ro rx and tx sizes.
>
>
> > +static int rx_state = 0; /* RX state for IOCTL */
> rx_state could probably be part of omap_irda.
I think this can be removed all together, as it remains zero only.
>
>
> > +struct omap_irda {
> > + unsigned char open;
> > + int speed; /* Current IrDA speed */
> > + int newspeed;
> > +
> > + struct net_device_stats stats;
> > + struct irlap_cb *irlap;
> > + struct qos_info qos;
> > +
> > + int rx_dma_channel;
> > + int tx_dma_channel;
> > +
> > + dma_addr_t rx_buf_dma_phys; /* Physical address of RX DMA buffer
> */
> > + dma_addr_t tx_buf_dma_phys; /* Physical address of TX DMA buffer
> */
> > +
> > + void *rx_buf_dma_virt; /* Virtual address of RX DMA buffer */
> > + void *tx_buf_dma_virt; /* Virtual address of TX DMA buffer */
> > +
> > + struct device *dev;
> > + struct omap_irda_config *pdata;
> You forgot to define omap_irda_config.
> Same thing for IR_SEL, IR_SIRMODE, IR_FIRMODE and IR_MIRMODE.
Check include/asm-arm/arch-omap/irda.h
> > +static void omap_irda_rx_dma_callback(int lch, u16 ch_status, void
> *data)
> > +{
> > + struct net_device *dev = data;
> > + struct omap_irda *si = dev->priv;
> A detail here, but I think that "si" for a variable name is a bit
> meaningless.
Ok. I will update this.
> > +
> > + status = uart_reg_in(UART3_SFLSR); /* Take a frame status */
> > +
> > + if (status != 0) { /* Bad frame? */
> > + si->stats.rx_frame_errors++;
> > + uart_reg_in(UART3_RESUME);
> > + } else {
> > + /* We got a frame! */
> > + skb = alloc_skb(IRDA_FRAME_SIZE_LIMIT, GFP_ATOMIC);
> Use dev_alloc_skb() for allocating RX packets.
Ok. Done
> > +
> > + case SIOCGRECEIVING:
> > + rq->ifr_receiving = rx_state;
> > + break;
> rx_state is always set to 0, so this looks a bit useless.
I can remove the "case SIOCGRECEIVING" altogether then.
> > +
> > + if (!pdata->rx_channel || !pdata->tx_channel) {
> > + printk(KERN_ERR "IrDA invalid rx/tx channel value\n");
> > + return -ENOENT;
> > + }
> I see that rx_channel and tx_channel are part of your
> omap_irda_config, which
> seems to be a platform specific structure. However, DMA channels are
> architecture specific and could be removed from this structure, isn't
> it ?
Correct. Then I have to think passing those information using either
device resources or revert back to #ifdefers in omap-ir.c. Any
suggestions?
> > +
> > + irda_qos_bits_to_value(&si->qos);
> > +
> > + /* Any better way to avoid this? No. */
> > + if (machine_is_omap_h3() || machine_is_omap_h4())
> > + INIT_WORK(&si->pdata->gpio_expa, NULL, NULL);
> What is this for ?
As H3 and H4 platforms uses gpio_expander (through I2C) for selecting
between IrDA and GPS Module, and that code is being called from IRQ
context, as you know that gpip_expander code ultimately leades to I2C
code and which can sleep.
See these links:
http://linux.omap.com/pipermail/linux-omap-open-source/2005-December/thread.html
http://linux.omap.com/pipermail/linux-omap-open-source/2005-December/005996.html
http://linux.omap.com/pipermail/linux-omap-open-source/2005-December/006019.html
---Komal Shah
http://komalshah.blogspot.com/
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2006-08-11 9:20 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-08-07 9:12 [PATCH] OMAP: Add support for IrDA driver Komal Shah
2006-08-11 5:47 ` Samuel Ortiz
2006-08-11 9:20 ` [irda-users] " Komal Shah
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox