OpenE2K
/
linux
6
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
linux/drivers/serial/sunzilog.c

1601 lines
39 KiB
C
Raw Normal View History

[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
/* sunzilog.c: Zilog serial driver for Sparc systems.
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
*
* Driver for Zilog serial chips found on Sun workstations and
* servers. This driver could actually be made more generic.
*
* This is based on the old drivers/sbus/char/zs.c code. A lot
* of code has been simply moved over directly from there but
* much has been rewritten. Credits therefore go out to Eddie
* C. Dost, Pete Zaitcev, Ted Ts'o and Alex Buell for their
* work there.
*
[SERIAL]: Fix console write locking in sparc drivers. Mirror the logic in 8250 for proper console write locking when SYSRQ is triggered or an OOPS is in progress. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-07-16 08:53:32 +02:00
* Copyright (C) 2002, 2006, 2007 David S. Miller (davem@davemloft.net)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
*/
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/delay.h>
#include <linux/tty.h>
#include <linux/tty_flip.h>
#include <linux/major.h>
#include <linux/string.h>
#include <linux/ptrace.h>
#include <linux/ioport.h>
#include <linux/slab.h>
#include <linux/circ_buf.h>
#include <linux/serial.h>
#include <linux/sysrq.h>
#include <linux/console.h>
#include <linux/spinlock.h>
#ifdef CONFIG_SERIO
#include <linux/serio.h>
#endif
#include <linux/init.h>
#include <asm/io.h>
#include <asm/irq.h>
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
#include <asm/prom.h>
#include <asm/of_device.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
#if defined(CONFIG_SERIAL_SUNZILOG_CONSOLE) && defined(CONFIG_MAGIC_SYSRQ)
#define SUPPORT_SYSRQ
#endif
#include <linux/serial_core.h>
#include "suncore.h"
#include "sunzilog.h"
/* On 32-bit sparcs we need to delay after register accesses
* to accommodate sun4 systems, but we do not need to flush writes.
* On 64-bit sparc we only need to flush single writes to ensure
* completion.
*/
#ifndef CONFIG_SPARC64
#define ZSDELAY() udelay(5)
#define ZSDELAY_LONG() udelay(20)
#define ZS_WSYNC(channel) do { } while (0)
#else
#define ZSDELAY()
#define ZSDELAY_LONG()
#define ZS_WSYNC(__channel) \
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
readb(&((__channel)->control))
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
#endif
#define ZS_CLOCK 4915200 /* Zilog input clock rate. */
#define ZS_CLOCK_DIVISOR 16 /* Divisor this driver uses. */
/*
* We wrap our port structure around the generic uart_port.
*/
struct uart_sunzilog_port {
struct uart_port port;
/* IRQ servicing chain. */
struct uart_sunzilog_port *next;
/* Current values of Zilog write registers. */
unsigned char curregs[NUM_ZSREGS];
unsigned int flags;
#define SUNZILOG_FLAG_CONS_KEYB 0x00000001
#define SUNZILOG_FLAG_CONS_MOUSE 0x00000002
#define SUNZILOG_FLAG_IS_CONS 0x00000004
#define SUNZILOG_FLAG_IS_KGDB 0x00000008
#define SUNZILOG_FLAG_MODEM_STATUS 0x00000010
#define SUNZILOG_FLAG_IS_CHANNEL_A 0x00000020
#define SUNZILOG_FLAG_REGS_HELD 0x00000040
#define SUNZILOG_FLAG_TX_STOPPED 0x00000080
#define SUNZILOG_FLAG_TX_ACTIVE 0x00000100
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
#define SUNZILOG_FLAG_ESCC 0x00000200
#define SUNZILOG_FLAG_ISR_HANDLER 0x00000400
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
unsigned int cflag;
unsigned char parity_mask;
unsigned char prev_status;
#ifdef CONFIG_SERIO
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
struct serio serio;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
int serio_open;
#endif
};
#define ZILOG_CHANNEL_FROM_PORT(PORT) ((struct zilog_channel __iomem *)((PORT)->membase))
#define UART_ZILOG(PORT) ((struct uart_sunzilog_port *)(PORT))
#define ZS_IS_KEYB(UP) ((UP)->flags & SUNZILOG_FLAG_CONS_KEYB)
#define ZS_IS_MOUSE(UP) ((UP)->flags & SUNZILOG_FLAG_CONS_MOUSE)
#define ZS_IS_CONS(UP) ((UP)->flags & SUNZILOG_FLAG_IS_CONS)
#define ZS_IS_KGDB(UP) ((UP)->flags & SUNZILOG_FLAG_IS_KGDB)
#define ZS_WANTS_MODEM_STATUS(UP) ((UP)->flags & SUNZILOG_FLAG_MODEM_STATUS)
#define ZS_IS_CHANNEL_A(UP) ((UP)->flags & SUNZILOG_FLAG_IS_CHANNEL_A)
#define ZS_REGS_HELD(UP) ((UP)->flags & SUNZILOG_FLAG_REGS_HELD)
#define ZS_TX_STOPPED(UP) ((UP)->flags & SUNZILOG_FLAG_TX_STOPPED)
#define ZS_TX_ACTIVE(UP) ((UP)->flags & SUNZILOG_FLAG_TX_ACTIVE)
/* Reading and writing Zilog8530 registers. The delays are to make this
* driver work on the Sun4 which needs a settling delay after each chip
* register access, other machines handle this in hardware via auxiliary
* flip-flops which implement the settle time we do in software.
*
* The port lock must be held and local IRQs must be disabled
* when {read,write}_zsreg is invoked.
*/
static unsigned char read_zsreg(struct zilog_channel __iomem *channel,
unsigned char reg)
{
unsigned char retval;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(reg, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
retval = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
return retval;
}
static void write_zsreg(struct zilog_channel __iomem *channel,
unsigned char reg, unsigned char value)
{
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(reg, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(value, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
}
static void sunzilog_clear_fifo(struct zilog_channel __iomem *channel)
{
int i;
for (i = 0; i < 32; i++) {
unsigned char regval;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
regval = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
if (regval & Rx_CH_AV)
break;
regval = read_zsreg(channel, R1);
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
readb(&channel->data);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
if (regval & (PAR_ERR | Rx_OVR | CRC_ERR)) {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(ERR_RES, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
}
}
}
/* This function must only be called when the TX is not busy. The UART
* port lock must be held and local interrupts disabled.
*/
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
static int __load_zsregs(struct zilog_channel __iomem *channel, unsigned char *regs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
int i;
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
int escc;
unsigned char r15;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
/* Let pending transmits finish. */
for (i = 0; i < 1000; i++) {
unsigned char stat = read_zsreg(channel, R1);
if (stat & ALL_SNT)
break;
udelay(100);
}
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(ERR_RES, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
sunzilog_clear_fifo(channel);
/* Disable all interrupts. */
write_zsreg(channel, R1,
regs[R1] & ~(RxINT_MASK | TxINT_ENAB | EXT_INT_ENAB));
/* Set parity, sync config, stop bits, and clock divisor. */
write_zsreg(channel, R4, regs[R4]);
/* Set misc. TX/RX control bits. */
write_zsreg(channel, R10, regs[R10]);
/* Set TX/RX controls sans the enable bits. */
write_zsreg(channel, R3, regs[R3] & ~RxENAB);
write_zsreg(channel, R5, regs[R5] & ~TxENAB);
/* Synchronous mode config. */
write_zsreg(channel, R6, regs[R6]);
write_zsreg(channel, R7, regs[R7]);
/* Don't mess with the interrupt vector (R2, unused by us) and
* master interrupt control (R9). We make sure this is setup
* properly at probe time then never touch it again.
*/
/* Disable baud generator. */
write_zsreg(channel, R14, regs[R14] & ~BRENAB);
/* Clock mode control. */
write_zsreg(channel, R11, regs[R11]);
/* Lower and upper byte of baud rate generator divisor. */
write_zsreg(channel, R12, regs[R12]);
write_zsreg(channel, R13, regs[R13]);
/* Now rewrite R14, with BRENAB (if set). */
write_zsreg(channel, R14, regs[R14]);
/* External status interrupt control. */
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
write_zsreg(channel, R15, (regs[R15] | WR7pEN) & ~FIFOEN);
/* ESCC Extension Register */
r15 = read_zsreg(channel, R15);
if (r15 & 0x01) {
write_zsreg(channel, R7, regs[R7p]);
/* External status interrupt and FIFO control. */
write_zsreg(channel, R15, regs[R15] & ~WR7pEN);
escc = 1;
} else {
/* Clear FIFO bit case it is an issue */
regs[R15] &= ~FIFOEN;
escc = 0;
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
/* Reset external status interrupts. */
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
write_zsreg(channel, R0, RES_EXT_INT); /* First Latch */
write_zsreg(channel, R0, RES_EXT_INT); /* Second Latch */
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
/* Rewrite R3/R5, this time without enables masked. */
write_zsreg(channel, R3, regs[R3]);
write_zsreg(channel, R5, regs[R5]);
/* Rewrite R1, this time without IRQ enabled masked. */
write_zsreg(channel, R1, regs[R1]);
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
return escc;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
/* Reprogram the Zilog channel HW registers with the copies found in the
* software state struct. If the transmitter is busy, we defer this update
* until the next TX complete interrupt. Else, we do it right now.
*
* The UART port lock must be held and local interrupts disabled.
*/
static void sunzilog_maybe_update_regs(struct uart_sunzilog_port *up,
struct zilog_channel __iomem *channel)
{
if (!ZS_REGS_HELD(up)) {
if (ZS_TX_ACTIVE(up)) {
up->flags |= SUNZILOG_FLAG_REGS_HELD;
} else {
__load_zsregs(channel, up->curregs);
}
}
}
static void sunzilog_change_mouse_baud(struct uart_sunzilog_port *up)
{
unsigned int cur_cflag = up->cflag;
int brg, new_baud;
up->cflag &= ~CBAUD;
up->cflag |= suncore_mouse_baud_cflag_next(cur_cflag, &new_baud);
brg = BPS_TO_BRG(new_baud, ZS_CLOCK / ZS_CLOCK_DIVISOR);
up->curregs[R12] = (brg & 0xff);
up->curregs[R13] = (brg >> 8) & 0xff;
sunzilog_maybe_update_regs(up, ZILOG_CHANNEL_FROM_PORT(&up->port));
}
static void sunzilog_kbdms_receive_chars(struct uart_sunzilog_port *up,
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
unsigned char ch, int is_break)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
if (ZS_IS_KEYB(up)) {
/* Stop-A is handled by drivers/char/keyboard.c now. */
#ifdef CONFIG_SERIO
if (up->serio_open)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
serio_interrupt(&up->serio, ch, 0);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
#endif
} else if (ZS_IS_MOUSE(up)) {
int ret = suncore_mouse_baud_detection(ch, is_break);
switch (ret) {
case 2:
sunzilog_change_mouse_baud(up);
/* fallthru */
case 1:
break;
case 0:
#ifdef CONFIG_SERIO
if (up->serio_open)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
serio_interrupt(&up->serio, ch, 0);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
#endif
break;
};
}
}
static struct tty_struct *
sunzilog_receive_chars(struct uart_sunzilog_port *up,
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
struct zilog_channel __iomem *channel)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
struct tty_struct *tty;
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 05:54:13 +01:00
unsigned char ch, r1, flag;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
tty = NULL;
if (up->port.info != NULL && /* Unopened serial console */
up->port.info->tty != NULL) /* Keyboard || mouse */
tty = up->port.info->tty;
for (;;) {
r1 = read_zsreg(channel, R1);
if (r1 & (PAR_ERR | Rx_OVR | CRC_ERR)) {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(ERR_RES, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
}
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
ch = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
/* This funny hack depends upon BRK_ABRT not interfering
* with the other bits we care about in R1.
*/
if (ch & BRK_ABRT)
r1 |= BRK_ABRT;
if (!(ch & Rx_CH_AV))
break;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
ch = readb(&channel->data);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ch &= up->parity_mask;
if (unlikely(ZS_IS_KEYB(up)) || unlikely(ZS_IS_MOUSE(up))) {
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
sunzilog_kbdms_receive_chars(up, ch, 0);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
continue;
}
if (tty == NULL) {
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
uart_handle_sysrq_char(&up->port, ch);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
continue;
}
/* A real serial line, record the character and status. */
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 05:54:13 +01:00
flag = TTY_NORMAL;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
up->port.icount.rx++;
if (r1 & (BRK_ABRT | PAR_ERR | Rx_OVR | CRC_ERR)) {
if (r1 & BRK_ABRT) {
r1 &= ~(PAR_ERR | CRC_ERR);
up->port.icount.brk++;
if (uart_handle_break(&up->port))
continue;
}
else if (r1 & PAR_ERR)
up->port.icount.parity++;
else if (r1 & CRC_ERR)
up->port.icount.frame++;
if (r1 & Rx_OVR)
up->port.icount.overrun++;
r1 &= up->port.read_status_mask;
if (r1 & BRK_ABRT)
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 05:54:13 +01:00
flag = TTY_BREAK;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
else if (r1 & PAR_ERR)
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 05:54:13 +01:00
flag = TTY_PARITY;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
else if (r1 & CRC_ERR)
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 05:54:13 +01:00
flag = TTY_FRAME;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
if (uart_handle_sysrq_char(&up->port, ch))
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
continue;
if (up->port.ignore_status_mask == 0xff ||
(r1 & up->port.ignore_status_mask) == 0) {
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 05:54:13 +01:00
tty_insert_flip_char(tty, ch, flag);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 05:54:13 +01:00
if (r1 & Rx_OVR)
tty_insert_flip_char(tty, 0, TTY_OVERRUN);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
return tty;
}
static void sunzilog_status_handle(struct uart_sunzilog_port *up,
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
struct zilog_channel __iomem *channel)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
unsigned char status;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
status = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(RES_EXT_INT, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
if (status & BRK_ABRT) {
if (ZS_IS_MOUSE(up))
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
sunzilog_kbdms_receive_chars(up, 0, 1);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (ZS_IS_CONS(up)) {
/* Wait for BREAK to deassert to avoid potentially
* confusing the PROM.
*/
while (1) {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
status = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
if (!(status & BRK_ABRT))
break;
}
sun_do_break();
return;
}
}
if (ZS_WANTS_MODEM_STATUS(up)) {
if (status & SYNC)
up->port.icount.dsr++;
/* The Zilog just gives us an interrupt when DCD/CTS/etc. change.
* But it does not tell us which bit has changed, we have to keep
* track of this ourselves.
*/
if ((status ^ up->prev_status) ^ DCD)
uart_handle_dcd_change(&up->port,
(status & DCD));
if ((status ^ up->prev_status) ^ CTS)
uart_handle_cts_change(&up->port,
(status & CTS));
wake_up_interruptible(&up->port.info->delta_msr_wait);
}
up->prev_status = status;
}
static void sunzilog_transmit_chars(struct uart_sunzilog_port *up,
struct zilog_channel __iomem *channel)
{
struct circ_buf *xmit;
if (ZS_IS_CONS(up)) {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
unsigned char status = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
/* TX still busy? Just wait for the next TX done interrupt.
*
* It can occur because of how we do serial console writes. It would
* be nice to transmit console writes just like we normally would for
* a TTY line. (ie. buffered and TX interrupt driven). That is not
* easy because console writes cannot sleep. One solution might be
* to poll on enough port->xmit space becomming free. -DaveM
*/
if (!(status & Tx_BUF_EMP))
return;
}
up->flags &= ~SUNZILOG_FLAG_TX_ACTIVE;
if (ZS_REGS_HELD(up)) {
__load_zsregs(channel, up->curregs);
up->flags &= ~SUNZILOG_FLAG_REGS_HELD;
}
if (ZS_TX_STOPPED(up)) {
up->flags &= ~SUNZILOG_FLAG_TX_STOPPED;
goto ack_tx_int;
}
if (up->port.x_char) {
up->flags |= SUNZILOG_FLAG_TX_ACTIVE;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(up->port.x_char, &channel->data);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
up->port.icount.tx++;
up->port.x_char = 0;
return;
}
if (up->port.info == NULL)
goto ack_tx_int;
xmit = &up->port.info->xmit;
[SPARC64]: Fix oops on runlevel change with serial console. Incorrect uart_write_wakeup() calls cause reference to a NULL tty pointer in sunsab and sunzilog serial drivers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-10-11 05:43:22 +02:00
if (uart_circ_empty(xmit))
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
goto ack_tx_int;
[SPARC64]: Fix oops on runlevel change with serial console. Incorrect uart_write_wakeup() calls cause reference to a NULL tty pointer in sunsab and sunzilog serial drivers. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-10-11 05:43:22 +02:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (uart_tx_stopped(&up->port))
goto ack_tx_int;
up->flags |= SUNZILOG_FLAG_TX_ACTIVE;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(xmit->buf[xmit->tail], &channel->data);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE - 1);
up->port.icount.tx++;
if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
uart_write_wakeup(&up->port);
return;
ack_tx_int:
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(RES_Tx_P, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
}
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
static irqreturn_t sunzilog_interrupt(int irq, void *dev_id)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
struct uart_sunzilog_port *up = dev_id;
while (up) {
struct zilog_channel __iomem *channel
= ZILOG_CHANNEL_FROM_PORT(&up->port);
struct tty_struct *tty;
unsigned char r3;
spin_lock(&up->port.lock);
r3 = read_zsreg(channel, R3);
/* Channel A */
tty = NULL;
if (r3 & (CHAEXT | CHATxIP | CHARxIP)) {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(RES_H_IUS, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
if (r3 & CHARxIP)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
tty = sunzilog_receive_chars(up, channel);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (r3 & CHAEXT)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
sunzilog_status_handle(up, channel);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (r3 & CHATxIP)
sunzilog_transmit_chars(up, channel);
}
spin_unlock(&up->port.lock);
if (tty)
tty_flip_buffer_push(tty);
/* Channel B */
up = up->next;
channel = ZILOG_CHANNEL_FROM_PORT(&up->port);
spin_lock(&up->port.lock);
tty = NULL;
if (r3 & (CHBEXT | CHBTxIP | CHBRxIP)) {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(RES_H_IUS, &channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
if (r3 & CHBRxIP)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
tty = sunzilog_receive_chars(up, channel);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (r3 & CHBEXT)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 15:55:46 +02:00
sunzilog_status_handle(up, channel);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (r3 & CHBTxIP)
sunzilog_transmit_chars(up, channel);
}
spin_unlock(&up->port.lock);
if (tty)
tty_flip_buffer_push(tty);
up = up->next;
}
return IRQ_HANDLED;
}
/* A convenient way to quickly get R0 status. The caller must _not_ hold the
* port lock, it is acquired here.
*/
static __inline__ unsigned char sunzilog_read_channel_status(struct uart_port *port)
{
struct zilog_channel __iomem *channel;
unsigned char status;
channel = ZILOG_CHANNEL_FROM_PORT(port);
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
status = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
return status;
}
/* The port lock is not held. */
static unsigned int sunzilog_tx_empty(struct uart_port *port)
{
[PATCH] Serial: Adjust serial locking This patch changes the way serial ports are locked when getting modem status. This change is necessary because we will need to atomically read the modem status and take action depending on the CTS status. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-06-29 10:42:38 +02:00
unsigned long flags;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
unsigned char status;
unsigned int ret;
[PATCH] Serial: Adjust serial locking This patch changes the way serial ports are locked when getting modem status. This change is necessary because we will need to atomically read the modem status and take action depending on the CTS status. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-06-29 10:42:38 +02:00
spin_lock_irqsave(&port->lock, flags);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
status = sunzilog_read_channel_status(port);
[PATCH] Serial: Adjust serial locking This patch changes the way serial ports are locked when getting modem status. This change is necessary because we will need to atomically read the modem status and take action depending on the CTS status. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-06-29 10:42:38 +02:00
spin_unlock_irqrestore(&port->lock, flags);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (status & Tx_BUF_EMP)
ret = TIOCSER_TEMT;
else
ret = 0;
return ret;
}
[PATCH] Serial: Adjust serial locking This patch changes the way serial ports are locked when getting modem status. This change is necessary because we will need to atomically read the modem status and take action depending on the CTS status. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-06-29 10:42:38 +02:00
/* The port lock is held and interrupts are disabled. */
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
static unsigned int sunzilog_get_mctrl(struct uart_port *port)
{
unsigned char status;
unsigned int ret;
status = sunzilog_read_channel_status(port);
ret = 0;
if (status & DCD)
ret |= TIOCM_CAR;
if (status & SYNC)
ret |= TIOCM_DSR;
if (status & CTS)
ret |= TIOCM_CTS;
return ret;
}
/* The port lock is held and interrupts are disabled. */
static void sunzilog_set_mctrl(struct uart_port *port, unsigned int mctrl)
{
struct uart_sunzilog_port *up = (struct uart_sunzilog_port *) port;
struct zilog_channel __iomem *channel = ZILOG_CHANNEL_FROM_PORT(port);
unsigned char set_bits, clear_bits;
set_bits = clear_bits = 0;
if (mctrl & TIOCM_RTS)
set_bits |= RTS;
else
clear_bits |= RTS;
if (mctrl & TIOCM_DTR)
set_bits |= DTR;
else
clear_bits |= DTR;
/* NOTE: Not subject to 'transmitter active' rule. */
up->curregs[R5] |= set_bits;
up->curregs[R5] &= ~clear_bits;
write_zsreg(channel, R5, up->curregs[R5]);
}
/* The port lock is held and interrupts are disabled. */
[SERIAL] Clean up and fix tty transmission start/stoping The start_tx and stop_tx methods were passed a flag to indicate whether the start/stop was from the tty start/stop callbacks, and some drivers used this flag to decide whether to ask the UART to immediately stop transmission (where the UART supports such a feature.) There are other cases when we wish this to occur - when CTS is lowered, or if we change from soft to hard flow control and CTS is inactive. In these cases, this flag was false, and we would allow the transmitter to drain before stopping. There is really only one case where we want to let the transmitter drain before disabling, and that's when we run out of characters to send. Hence, re-jig the start_tx and stop_tx methods to eliminate this flag, and introduce new functions for the special "disable and allow transmitter to drain" case. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-08-31 11:12:14 +02:00
static void sunzilog_stop_tx(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
struct uart_sunzilog_port *up = (struct uart_sunzilog_port *) port;
up->flags |= SUNZILOG_FLAG_TX_STOPPED;
}
/* The port lock is held and interrupts are disabled. */
[SERIAL] Clean up and fix tty transmission start/stoping The start_tx and stop_tx methods were passed a flag to indicate whether the start/stop was from the tty start/stop callbacks, and some drivers used this flag to decide whether to ask the UART to immediately stop transmission (where the UART supports such a feature.) There are other cases when we wish this to occur - when CTS is lowered, or if we change from soft to hard flow control and CTS is inactive. In these cases, this flag was false, and we would allow the transmitter to drain before stopping. There is really only one case where we want to let the transmitter drain before disabling, and that's when we run out of characters to send. Hence, re-jig the start_tx and stop_tx methods to eliminate this flag, and introduce new functions for the special "disable and allow transmitter to drain" case. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-08-31 11:12:14 +02:00
static void sunzilog_start_tx(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
struct uart_sunzilog_port *up = (struct uart_sunzilog_port *) port;
struct zilog_channel __iomem *channel = ZILOG_CHANNEL_FROM_PORT(port);
unsigned char status;
up->flags |= SUNZILOG_FLAG_TX_ACTIVE;
up->flags &= ~SUNZILOG_FLAG_TX_STOPPED;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
status = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
/* TX busy? Just wait for the TX done interrupt. */
if (!(status & Tx_BUF_EMP))
return;
/* Send the first character to jump-start the TX done
* IRQ sending engine.
*/
if (port->x_char) {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(port->x_char, &channel->data);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
port->icount.tx++;
port->x_char = 0;
} else {
struct circ_buf *xmit = &port->info->xmit;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(xmit->buf[xmit->tail], &channel->data);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE - 1);
port->icount.tx++;
if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
uart_write_wakeup(&up->port);
}
}
/* The port lock is held. */
static void sunzilog_stop_rx(struct uart_port *port)
{
struct uart_sunzilog_port *up = UART_ZILOG(port);
struct zilog_channel __iomem *channel;
if (ZS_IS_CONS(up))
return;
channel = ZILOG_CHANNEL_FROM_PORT(port);
/* Disable all RX interrupts. */
up->curregs[R1] &= ~RxINT_MASK;
sunzilog_maybe_update_regs(up, channel);
}
/* The port lock is held. */
static void sunzilog_enable_ms(struct uart_port *port)
{
struct uart_sunzilog_port *up = (struct uart_sunzilog_port *) port;
struct zilog_channel __iomem *channel = ZILOG_CHANNEL_FROM_PORT(port);
unsigned char new_reg;
new_reg = up->curregs[R15] | (DCDIE | SYNCIE | CTSIE);
if (new_reg != up->curregs[R15]) {
up->curregs[R15] = new_reg;
/* NOTE: Not subject to 'transmitter active' rule. */
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
write_zsreg(channel, R15, up->curregs[R15] & ~WR7pEN);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
}
/* The port lock is not held. */
static void sunzilog_break_ctl(struct uart_port *port, int break_state)
{
struct uart_sunzilog_port *up = (struct uart_sunzilog_port *) port;
struct zilog_channel __iomem *channel = ZILOG_CHANNEL_FROM_PORT(port);
unsigned char set_bits, clear_bits, new_reg;
unsigned long flags;
set_bits = clear_bits = 0;
if (break_state)
set_bits |= SND_BRK;
else
clear_bits |= SND_BRK;
spin_lock_irqsave(&port->lock, flags);
new_reg = (up->curregs[R5] | set_bits) & ~clear_bits;
if (new_reg != up->curregs[R5]) {
up->curregs[R5] = new_reg;
/* NOTE: Not subject to 'transmitter active' rule. */
write_zsreg(channel, R5, up->curregs[R5]);
}
spin_unlock_irqrestore(&port->lock, flags);
}
static void __sunzilog_startup(struct uart_sunzilog_port *up)
{
struct zilog_channel __iomem *channel;
channel = ZILOG_CHANNEL_FROM_PORT(&up->port);
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
up->prev_status = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
/* Enable receiver and transmitter. */
up->curregs[R3] |= RxENAB;
up->curregs[R5] |= TxENAB;
up->curregs[R1] |= EXT_INT_ENAB | INT_ALL_Rx | TxINT_ENAB;
sunzilog_maybe_update_regs(up, channel);
}
static int sunzilog_startup(struct uart_port *port)
{
struct uart_sunzilog_port *up = UART_ZILOG(port);
unsigned long flags;
if (ZS_IS_CONS(up))
return 0;
spin_lock_irqsave(&port->lock, flags);
__sunzilog_startup(up);
spin_unlock_irqrestore(&port->lock, flags);
return 0;
}
/*
* The test for ZS_IS_CONS is explained by the following e-mail:
*****
* From: Russell King <rmk@arm.linux.org.uk>
* Date: Sun, 8 Dec 2002 10:18:38 +0000
*
* On Sun, Dec 08, 2002 at 02:43:36AM -0500, Pete Zaitcev wrote:
* > I boot my 2.5 boxes using "console=ttyS0,9600" argument,
* > and I noticed that something is not right with reference
* > counting in this case. It seems that when the console
* > is open by kernel initially, this is not accounted
* > as an open, and uart_startup is not called.
*
* That is correct. We are unable to call uart_startup when the serial
* console is initialised because it may need to allocate memory (as
* request_irq does) and the memory allocators may not have been
* initialised.
*
* 1. initialise the port into a state where it can send characters in the
* console write method.
*
* 2. don't do the actual hardware shutdown in your shutdown() method (but
* do the normal software shutdown - ie, free irqs etc)
*****
*/
static void sunzilog_shutdown(struct uart_port *port)
{
struct uart_sunzilog_port *up = UART_ZILOG(port);
struct zilog_channel __iomem *channel;
unsigned long flags;
if (ZS_IS_CONS(up))
return;
spin_lock_irqsave(&port->lock, flags);
channel = ZILOG_CHANNEL_FROM_PORT(port);
/* Disable receiver and transmitter. */
up->curregs[R3] &= ~RxENAB;
up->curregs[R5] &= ~TxENAB;
/* Disable all interrupts and BRK assertion. */
up->curregs[R1] &= ~(EXT_INT_ENAB | TxINT_ENAB | RxINT_MASK);
up->curregs[R5] &= ~SND_BRK;
sunzilog_maybe_update_regs(up, channel);
spin_unlock_irqrestore(&port->lock, flags);
}
/* Shared by TTY driver and serial console setup. The port lock is held
* and local interrupts are disabled.
*/
static void
sunzilog_convert_to_zs(struct uart_sunzilog_port *up, unsigned int cflag,
unsigned int iflag, int brg)
{
up->curregs[R10] = NRZ;
up->curregs[R11] = TCBR | RCBR;
/* Program BAUD and clock source. */
up->curregs[R4] &= ~XCLK_MASK;
up->curregs[R4] |= X16CLK;
up->curregs[R12] = brg & 0xff;
up->curregs[R13] = (brg >> 8) & 0xff;
up->curregs[R14] = BRSRC | BRENAB;
/* Character size, stop bits, and parity. */
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R3] &= ~RxN_MASK;
up->curregs[R5] &= ~TxN_MASK;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
switch (cflag & CSIZE) {
case CS5:
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R3] |= Rx5;
up->curregs[R5] |= Tx5;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
up->parity_mask = 0x1f;
break;
case CS6:
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R3] |= Rx6;
up->curregs[R5] |= Tx6;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
up->parity_mask = 0x3f;
break;
case CS7:
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R3] |= Rx7;
up->curregs[R5] |= Tx7;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
up->parity_mask = 0x7f;
break;
case CS8:
default:
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R3] |= Rx8;
up->curregs[R5] |= Tx8;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
up->parity_mask = 0xff;
break;
};
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R4] &= ~0x0c;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (cflag & CSTOPB)
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R4] |= SB2;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
else
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R4] |= SB1;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (cflag & PARENB)
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R4] |= PAR_ENAB;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
else
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R4] &= ~PAR_ENAB;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (!(cflag & PARODD))
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R4] |= PAR_EVEN;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
else
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R4] &= ~PAR_EVEN;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
up->port.read_status_mask = Rx_OVR;
if (iflag & INPCK)
up->port.read_status_mask |= CRC_ERR | PAR_ERR;
if (iflag & (BRKINT | PARMRK))
up->port.read_status_mask |= BRK_ABRT;
up->port.ignore_status_mask = 0;
if (iflag & IGNPAR)
up->port.ignore_status_mask |= CRC_ERR | PAR_ERR;
if (iflag & IGNBRK) {
up->port.ignore_status_mask |= BRK_ABRT;
if (iflag & IGNPAR)
up->port.ignore_status_mask |= Rx_OVR;
}
if ((cflag & CREAD) == 0)
up->port.ignore_status_mask = 0xff;
}
/* The port lock is not held. */
static void
[PATCH] tty: switch to ktermios This is the grungy swap all the occurrences in the right places patch that goes with the updates. At this point we have the same functionality as before (except that sgttyb() returns speeds not zero) and are ready to begin turning new stuff on providing nobody reports lots of bugs If you are a tty driver author converting an out of tree driver the only impact should be termios->ktermios name changes for the speed/property setting functions from your upper layers. If you are implementing your own TCGETS function before then your driver was broken already and its about to get a whole lot more painful for you so please fix it 8) Also fill in c_ispeed/ospeed on init for most devices, although the current code will do this for you anyway but I'd like eventually to lose that extra paranoia [akpm@osdl.org: bluetooth fix] [mp3@de.ibm.com: sclp fix] [mp3@de.ibm.com: warning fix for tty3270] [hugh@veritas.com: fix tty_ioctl powerpc build] [jdike@addtoit.com: uml: fix ->set_termios declaration] Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Martin Peschke <mp3@de.ibm.com> Acked-by: Peter Oberparleiter <oberpar@de.ibm.com> Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:38:45 +01:00
sunzilog_set_termios(struct uart_port *port, struct ktermios *termios,
struct ktermios *old)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
struct uart_sunzilog_port *up = (struct uart_sunzilog_port *) port;
unsigned long flags;
int baud, brg;
baud = uart_get_baud_rate(port, termios, old, 1200, 76800);
spin_lock_irqsave(&up->port.lock, flags);
brg = BPS_TO_BRG(baud, ZS_CLOCK / ZS_CLOCK_DIVISOR);
sunzilog_convert_to_zs(up, termios->c_cflag, termios->c_iflag, brg);
if (UART_ENABLE_MS(&up->port, termios->c_cflag))
up->flags |= SUNZILOG_FLAG_MODEM_STATUS;
else
up->flags &= ~SUNZILOG_FLAG_MODEM_STATUS;
up->cflag = termios->c_cflag;
sunzilog_maybe_update_regs(up, ZILOG_CHANNEL_FROM_PORT(port));
uart_update_timeout(port, termios->c_cflag, baud);
spin_unlock_irqrestore(&up->port.lock, flags);
}
static const char *sunzilog_type(struct uart_port *port)
{
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
struct uart_sunzilog_port *up = UART_ZILOG(port);
return (up->flags & SUNZILOG_FLAG_ESCC) ? "zs (ESCC)" : "zs";
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
/* We do not request/release mappings of the registers here, this
* happens at early serial probe time.
*/
static void sunzilog_release_port(struct uart_port *port)
{
}
static int sunzilog_request_port(struct uart_port *port)
{
return 0;
}
/* These do not need to do anything interesting either. */
static void sunzilog_config_port(struct uart_port *port, int flags)
{
}
/* We do not support letting the user mess with the divisor, IRQ, etc. */
static int sunzilog_verify_port(struct uart_port *port, struct serial_struct *ser)
{
return -EINVAL;
}
static struct uart_ops sunzilog_pops = {
.tx_empty = sunzilog_tx_empty,
.set_mctrl = sunzilog_set_mctrl,
.get_mctrl = sunzilog_get_mctrl,
.stop_tx = sunzilog_stop_tx,
.start_tx = sunzilog_start_tx,
.stop_rx = sunzilog_stop_rx,
.enable_ms = sunzilog_enable_ms,
.break_ctl = sunzilog_break_ctl,
.startup = sunzilog_startup,
.shutdown = sunzilog_shutdown,
.set_termios = sunzilog_set_termios,
.type = sunzilog_type,
.release_port = sunzilog_release_port,
.request_port = sunzilog_request_port,
.config_port = sunzilog_config_port,
.verify_port = sunzilog_verify_port,
};
sparc: sunzilog uart order I have a sparcstation 20 clone with a lot of on board serial ports. The serial core code assumes that uarts are assigned contiguously and that may not be the case when there are multiple zs devices present. This patch insures that uart chips are placed in front of keyboard/mouse chips in the port table. ffd37420: ttyS0 at MMIO 0xf1100000 (irq = 44) is a zs (ESCC) Console: ttyS0 (SunZilog zs0) console [ttyS0] enabled ffd37420: ttyS1 at MMIO 0xf1100004 (irq = 44) is a zs (ESCC) ffd37500: Keyboard at MMIO 0xf1000000 (irq = 44) is a zs ffd37500: Mouse at MMIO 0xf1000004 (irq = 44) is a zs ffd3c5c0: ttyS2 at MMIO 0xf1100008 (irq = 44) is a zs (ESCC) ffd3c5c0: ttyS3 at MMIO 0xf110000c (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS4 at MMIO 0xf1100010 (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS5 at MMIO 0xf1100014 (irq = 44) is a zs (ESCC) ffd3c780: ttyS6 at MMIO 0xf1100018 (irq = 44) is a zs (ESCC) ffd3c780: ttyS7 at MMIO 0xf110001c (irq = 44) is a zs (ESCC) Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 12:37:51 +02:00
static int uart_chip_count;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
static struct uart_sunzilog_port *sunzilog_port_table;
static struct zilog_layout __iomem **sunzilog_chip_regs;
static struct uart_sunzilog_port *sunzilog_irq_chain;
static struct uart_driver sunzilog_reg = {
.owner = THIS_MODULE,
.driver_name = "ttyS",
.dev_name = "ttyS",
.major = TTY_MAJOR,
};
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
static int __init sunzilog_alloc_tables(int num_sunzilog)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
struct uart_sunzilog_port *up;
unsigned long size;
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
int num_channels = num_sunzilog * 2;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
int i;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
size = num_channels * sizeof(struct uart_sunzilog_port);
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
sunzilog_port_table = kzalloc(size, GFP_KERNEL);
if (!sunzilog_port_table)
return -ENOMEM;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
for (i = 0; i < num_channels; i++) {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
up = &sunzilog_port_table[i];
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
spin_lock_init(&up->port.lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
if (i == 0)
sunzilog_irq_chain = up;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
if (i < num_channels - 1)
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
up->next = up + 1;
else
up->next = NULL;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
size = num_sunzilog * sizeof(struct zilog_layout __iomem *);
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
sunzilog_chip_regs = kzalloc(size, GFP_KERNEL);
if (!sunzilog_chip_regs) {
kfree(sunzilog_port_table);
sunzilog_irq_chain = NULL;
return -ENOMEM;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
return 0;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
static void sunzilog_free_tables(void)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
kfree(sunzilog_port_table);
sunzilog_irq_chain = NULL;
kfree(sunzilog_chip_regs);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
#define ZS_PUT_CHAR_MAX_DELAY 2000 /* 10 ms */
[SERIAL] kernel console should send CRLF not LFCR Glen Turner reported that writing LFCR rather than the more traditional CRLF causes issues with some terminals. Since this aflicts many serial drivers, extract the common code to a library function (uart_console_write) and arrange for each driver to supply a "putchar" function. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-03-20 21:00:09 +01:00
static void sunzilog_putchar(struct uart_port *port, int ch)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
[PATCH] __iomem annotations in sunzilog Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-10-10 23:44:47 +02:00
struct zilog_channel __iomem *channel = ZILOG_CHANNEL_FROM_PORT(port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
int loops = ZS_PUT_CHAR_MAX_DELAY;
/* This is a timed polling loop so do not switch the explicit
* udelay with ZSDELAY as that is a NOP on some platforms. -DaveM
*/
do {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
unsigned char val = readb(&channel->control);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
if (val & Tx_BUF_EMP) {
ZSDELAY();
break;
}
udelay(5);
} while (--loops);
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
writeb(ch, &channel->data);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
ZSDELAY();
ZS_WSYNC(channel);
}
#ifdef CONFIG_SERIO
static DEFINE_SPINLOCK(sunzilog_serio_lock);
static int sunzilog_serio_write(struct serio *serio, unsigned char ch)
{
struct uart_sunzilog_port *up = serio->port_data;
unsigned long flags;
spin_lock_irqsave(&sunzilog_serio_lock, flags);
[SERIAL] kernel console should send CRLF not LFCR Glen Turner reported that writing LFCR rather than the more traditional CRLF causes issues with some terminals. Since this aflicts many serial drivers, extract the common code to a library function (uart_console_write) and arrange for each driver to supply a "putchar" function. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-03-20 21:00:09 +01:00
sunzilog_putchar(&up->port, ch);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
spin_unlock_irqrestore(&sunzilog_serio_lock, flags);
return 0;
}
static int sunzilog_serio_open(struct serio *serio)
{
struct uart_sunzilog_port *up = serio->port_data;
unsigned long flags;
int ret;
spin_lock_irqsave(&sunzilog_serio_lock, flags);
if (!up->serio_open) {
up->serio_open = 1;
ret = 0;
} else
ret = -EBUSY;
spin_unlock_irqrestore(&sunzilog_serio_lock, flags);
return ret;
}
static void sunzilog_serio_close(struct serio *serio)
{
struct uart_sunzilog_port *up = serio->port_data;
unsigned long flags;
spin_lock_irqsave(&sunzilog_serio_lock, flags);
up->serio_open = 0;
spin_unlock_irqrestore(&sunzilog_serio_lock, flags);
}
#endif /* CONFIG_SERIO */
#ifdef CONFIG_SERIAL_SUNZILOG_CONSOLE
static void
sunzilog_console_write(struct console *con, const char *s, unsigned int count)
{
struct uart_sunzilog_port *up = &sunzilog_port_table[con->index];
unsigned long flags;
[SERIAL]: Fix console write locking in sparc drivers. Mirror the logic in 8250 for proper console write locking when SYSRQ is triggered or an OOPS is in progress. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-07-16 08:53:32 +02:00
int locked = 1;
local_irq_save(flags);
if (up->port.sysrq) {
locked = 0;
} else if (oops_in_progress) {
locked = spin_trylock(&up->port.lock);
} else
spin_lock(&up->port.lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] kernel console should send CRLF not LFCR Glen Turner reported that writing LFCR rather than the more traditional CRLF causes issues with some terminals. Since this aflicts many serial drivers, extract the common code to a library function (uart_console_write) and arrange for each driver to supply a "putchar" function. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-03-20 21:00:09 +01:00
uart_console_write(&up->port, s, count, sunzilog_putchar);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
udelay(2);
[SERIAL]: Fix console write locking in sparc drivers. Mirror the logic in 8250 for proper console write locking when SYSRQ is triggered or an OOPS is in progress. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-07-16 08:53:32 +02:00
if (locked)
spin_unlock(&up->port.lock);
local_irq_restore(flags);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
static int __init sunzilog_console_setup(struct console *con, char *options)
{
struct uart_sunzilog_port *up = &sunzilog_port_table[con->index];
unsigned long flags;
int baud, brg;
[SERIAL] sunzilog: Mirror the sunsab serial setup bug fix. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-08-24 00:53:39 +02:00
if (up->port.type != PORT_SUNZILOG)
return -1;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
printk(KERN_INFO "Console: ttyS%d (SunZilog zs%d)\n",
(sunzilog_reg.minor - 64) + con->index, con->index);
/* Get firmware console settings. */
sunserial_console_termios(con);
/* Firmware console speed is limited to 150-->38400 baud so
* this hackish cflag thing is OK.
*/
switch (con->cflag & CBAUD) {
case B150: baud = 150; break;
case B300: baud = 300; break;
case B600: baud = 600; break;
case B1200: baud = 1200; break;
case B2400: baud = 2400; break;
case B4800: baud = 4800; break;
default: case B9600: baud = 9600; break;
case B19200: baud = 19200; break;
case B38400: baud = 38400; break;
};
brg = BPS_TO_BRG(baud, ZS_CLOCK / ZS_CLOCK_DIVISOR);
spin_lock_irqsave(&up->port.lock, flags);
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R15] |= BRKIE;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
sunzilog_convert_to_zs(up, con->cflag, 0, brg);
sunzilog_set_mctrl(&up->port, TIOCM_DTR | TIOCM_RTS);
__sunzilog_startup(up);
spin_unlock_irqrestore(&up->port.lock, flags);
return 0;
}
[SPARC32]: Fix sparc32 modpost warnings with sunzilog Fix this 2.6.19-rc1 build warnings from modpost: WARNING: vmlinux - Section mismatch: reference to .init.text:sunzilog_console_setup from .data between 'sunzilog_console' (at offset 0x8394) and 'devices_subsys' Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-10-10 23:44:01 +02:00
static struct console sunzilog_console_ops = {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
.name = "ttyS",
.write = sunzilog_console_write,
.device = uart_console_device,
.setup = sunzilog_console_setup,
.flags = CON_PRINTBUFFER,
.index = -1,
.data = &sunzilog_reg,
};
[SPARC64]: Prevent registering wrong serial console. If the console is not for a particular Sun serial controller, set the drv->cons to NULL. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-14 05:09:10 +01:00
static inline struct console *SUNZILOG_CONSOLE(void)
{
[SPARC32]: Fix sparc32 modpost warnings with sunzilog Fix this 2.6.19-rc1 build warnings from modpost: WARNING: vmlinux - Section mismatch: reference to .init.text:sunzilog_console_setup from .data between 'sunzilog_console' (at offset 0x8394) and 'devices_subsys' Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-10-10 23:44:01 +02:00
return &sunzilog_console_ops;
[SPARC64]: Prevent registering wrong serial console. If the console is not for a particular Sun serial controller, set the drv->cons to NULL. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-14 05:09:10 +01:00
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
#else
[SPARC64]: Prevent registering wrong serial console. If the console is not for a particular Sun serial controller, set the drv->cons to NULL. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-14 05:09:10 +01:00
#define SUNZILOG_CONSOLE() (NULL)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
#endif
sparc: sunzilog.c remove unused argument Remove unused argument in function call. Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-27 08:10:19 +02:00
static void __devinit sunzilog_init_kbdms(struct uart_sunzilog_port *up)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
int baud, brg;
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
if (up->flags & SUNZILOG_FLAG_CONS_KEYB) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
up->cflag = B1200 | CS8 | CLOCAL | CREAD;
baud = 1200;
} else {
up->cflag = B4800 | CS8 | CLOCAL | CREAD;
baud = 4800;
}
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R15] |= BRKIE;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
brg = BPS_TO_BRG(baud, ZS_CLOCK / ZS_CLOCK_DIVISOR);
sunzilog_convert_to_zs(up, up->cflag, 0, brg);
sunzilog_set_mctrl(&up->port, TIOCM_DTR | TIOCM_RTS);
__sunzilog_startup(up);
}
#ifdef CONFIG_SERIO
[SERIAL] sunzilog: section mismatch fix This patch fixes section mismatch warnings in the sunzilog driver. Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-19 21:06:02 +02:00
static void __devinit sunzilog_register_serio(struct uart_sunzilog_port *up)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
struct serio *serio = &up->serio;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
serio->port_data = up;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
serio->id.type = SERIO_RS232;
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
if (up->flags & SUNZILOG_FLAG_CONS_KEYB) {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
serio->id.proto = SERIO_SUNKBD;
strlcpy(serio->name, "zskbd", sizeof(serio->name));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
} else {
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
serio->id.proto = SERIO_SUN;
serio->id.extra = 1;
strlcpy(serio->name, "zsms", sizeof(serio->name));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
strlcpy(serio->phys,
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
((up->flags & SUNZILOG_FLAG_CONS_KEYB) ?
"zs/serio0" : "zs/serio1"),
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
sizeof(serio->phys));
serio->write = sunzilog_serio_write;
serio->open = sunzilog_serio_open;
serio->close = sunzilog_serio_close;
serio->dev.parent = up->port.dev;
serio_register_port(serio);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
#endif
[SERIAL] sunzilog: Mark sunzilog_init_hw as __devinit. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 04:52:35 +02:00
static void __devinit sunzilog_init_hw(struct uart_sunzilog_port *up)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
struct zilog_channel __iomem *channel;
unsigned long flags;
int baud, brg;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
channel = ZILOG_CHANNEL_FROM_PORT(&up->port);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
spin_lock_irqsave(&up->port.lock, flags);
if (ZS_IS_CHANNEL_A(up)) {
write_zsreg(channel, R9, FHWRES);
ZSDELAY_LONG();
(void) read_zsreg(channel, R0);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
if (up->flags & (SUNZILOG_FLAG_CONS_KEYB |
SUNZILOG_FLAG_CONS_MOUSE)) {
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R1] = EXT_INT_ENAB | INT_ALL_Rx | TxINT_ENAB;
up->curregs[R4] = PAR_EVEN | X16CLK | SB1;
up->curregs[R3] = RxENAB | Rx8;
up->curregs[R5] = TxENAB | Tx8;
up->curregs[R6] = 0x00; /* SDLC Address */
up->curregs[R7] = 0x7E; /* SDLC Flag */
up->curregs[R9] = NV;
up->curregs[R7p] = 0x00;
sparc: sunzilog.c remove unused argument Remove unused argument in function call. Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-27 08:10:19 +02:00
sunzilog_init_kbdms(up);
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
/* Only enable interrupts if an ISR handler available */
if (up->flags & SUNZILOG_FLAG_ISR_HANDLER)
up->curregs[R9] |= MIE;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
write_zsreg(channel, R9, up->curregs[R9]);
} else {
/* Normal serial TTY. */
up->parity_mask = 0xff;
up->curregs[R1] = EXT_INT_ENAB | INT_ALL_Rx | TxINT_ENAB;
up->curregs[R4] = PAR_EVEN | X16CLK | SB1;
up->curregs[R3] = RxENAB | Rx8;
up->curregs[R5] = TxENAB | Tx8;
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R6] = 0x00; /* SDLC Address */
up->curregs[R7] = 0x7E; /* SDLC Flag */
up->curregs[R9] = NV;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
up->curregs[R10] = NRZ;
up->curregs[R11] = TCBR | RCBR;
baud = 9600;
brg = BPS_TO_BRG(baud, ZS_CLOCK / ZS_CLOCK_DIVISOR);
up->curregs[R12] = (brg & 0xff);
up->curregs[R13] = (brg >> 8) & 0xff;
up->curregs[R14] = BRSRC | BRENAB;
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
up->curregs[R15] = FIFOEN; /* Use FIFO if on ESCC */
up->curregs[R7p] = TxFIFO_LVL | RxFIFO_LVL;
if (__load_zsregs(channel, up->curregs)) {
up->flags |= SUNZILOG_FLAG_ESCC;
}
/* Only enable interrupts if an ISR handler available */
if (up->flags & SUNZILOG_FLAG_ISR_HANDLER)
up->curregs[R9] |= MIE;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
write_zsreg(channel, R9, up->curregs[R9]);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
spin_unlock_irqrestore(&up->port.lock, flags);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
#ifdef CONFIG_SERIO
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
if (up->flags & (SUNZILOG_FLAG_CONS_KEYB |
SUNZILOG_FLAG_CONS_MOUSE))
sunzilog_register_serio(up);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
#endif
}
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
static int zilog_irq = -1;
[SERIAL] sunzilog: Register IRQ after all devices have been probed. Otherwise we will deref half-initialized channel pointers and crash in the interrupt handler. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-18 06:07:17 +02:00
static int __devinit zs_probe(struct of_device *op, const struct of_device_id *match)
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
{
sparc: sunzilog uart order I have a sparcstation 20 clone with a lot of on board serial ports. The serial core code assumes that uarts are assigned contiguously and that may not be the case when there are multiple zs devices present. This patch insures that uart chips are placed in front of keyboard/mouse chips in the port table. ffd37420: ttyS0 at MMIO 0xf1100000 (irq = 44) is a zs (ESCC) Console: ttyS0 (SunZilog zs0) console [ttyS0] enabled ffd37420: ttyS1 at MMIO 0xf1100004 (irq = 44) is a zs (ESCC) ffd37500: Keyboard at MMIO 0xf1000000 (irq = 44) is a zs ffd37500: Mouse at MMIO 0xf1000004 (irq = 44) is a zs ffd3c5c0: ttyS2 at MMIO 0xf1100008 (irq = 44) is a zs (ESCC) ffd3c5c0: ttyS3 at MMIO 0xf110000c (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS4 at MMIO 0xf1100010 (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS5 at MMIO 0xf1100014 (irq = 44) is a zs (ESCC) ffd3c780: ttyS6 at MMIO 0xf1100018 (irq = 44) is a zs (ESCC) ffd3c780: ttyS7 at MMIO 0xf110001c (irq = 44) is a zs (ESCC) Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 12:37:51 +02:00
static int kbm_inst, uart_inst;
int inst;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
struct uart_sunzilog_port *up;
struct zilog_layout __iomem *rp;
sparc: sunzilog uart order I have a sparcstation 20 clone with a lot of on board serial ports. The serial core code assumes that uarts are assigned contiguously and that may not be the case when there are multiple zs devices present. This patch insures that uart chips are placed in front of keyboard/mouse chips in the port table. ffd37420: ttyS0 at MMIO 0xf1100000 (irq = 44) is a zs (ESCC) Console: ttyS0 (SunZilog zs0) console [ttyS0] enabled ffd37420: ttyS1 at MMIO 0xf1100004 (irq = 44) is a zs (ESCC) ffd37500: Keyboard at MMIO 0xf1000000 (irq = 44) is a zs ffd37500: Mouse at MMIO 0xf1000004 (irq = 44) is a zs ffd3c5c0: ttyS2 at MMIO 0xf1100008 (irq = 44) is a zs (ESCC) ffd3c5c0: ttyS3 at MMIO 0xf110000c (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS4 at MMIO 0xf1100010 (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS5 at MMIO 0xf1100014 (irq = 44) is a zs (ESCC) ffd3c780: ttyS6 at MMIO 0xf1100018 (irq = 44) is a zs (ESCC) ffd3c780: ttyS7 at MMIO 0xf110001c (irq = 44) is a zs (ESCC) Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 12:37:51 +02:00
int keyboard_mouse = 0;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
int err;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
if (of_find_property(op->node, "keyboard", NULL))
keyboard_mouse = 1;
sparc: sunzilog uart order I have a sparcstation 20 clone with a lot of on board serial ports. The serial core code assumes that uarts are assigned contiguously and that may not be the case when there are multiple zs devices present. This patch insures that uart chips are placed in front of keyboard/mouse chips in the port table. ffd37420: ttyS0 at MMIO 0xf1100000 (irq = 44) is a zs (ESCC) Console: ttyS0 (SunZilog zs0) console [ttyS0] enabled ffd37420: ttyS1 at MMIO 0xf1100004 (irq = 44) is a zs (ESCC) ffd37500: Keyboard at MMIO 0xf1000000 (irq = 44) is a zs ffd37500: Mouse at MMIO 0xf1000004 (irq = 44) is a zs ffd3c5c0: ttyS2 at MMIO 0xf1100008 (irq = 44) is a zs (ESCC) ffd3c5c0: ttyS3 at MMIO 0xf110000c (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS4 at MMIO 0xf1100010 (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS5 at MMIO 0xf1100014 (irq = 44) is a zs (ESCC) ffd3c780: ttyS6 at MMIO 0xf1100018 (irq = 44) is a zs (ESCC) ffd3c780: ttyS7 at MMIO 0xf110001c (irq = 44) is a zs (ESCC) Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 12:37:51 +02:00
/* uarts must come before keyboards/mice */
if (keyboard_mouse)
inst = uart_chip_count + kbm_inst;
else
inst = uart_inst;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
sunzilog_chip_regs[inst] = of_ioremap(&op->resource[0], 0,
sizeof(struct zilog_layout),
"zs");
if (!sunzilog_chip_regs[inst])
return -ENOMEM;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
rp = sunzilog_chip_regs[inst];
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Remove duplicate IRQ registry in zs_probe(). We do it now in sunzilog_init() after all devices have been probed. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 06:04:04 +02:00
if (zilog_irq == -1)
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
zilog_irq = op->irqs[0];
[SPARC64]: Prevent registering wrong serial console. If the console is not for a particular Sun serial controller, set the drv->cons to NULL. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-14 05:09:10 +01:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
up = &sunzilog_port_table[inst * 2];
/* Channel A */
up[0].port.mapbase = op->resource[0].start + 0x00;
up[0].port.membase = (void __iomem *) &rp->channelA;
up[0].port.iotype = UPIO_MEM;
up[0].port.irq = op->irqs[0];
up[0].port.uartclk = ZS_CLOCK;
up[0].port.fifosize = 1;
up[0].port.ops = &sunzilog_pops;
up[0].port.type = PORT_SUNZILOG;
up[0].port.flags = 0;
up[0].port.line = (inst * 2) + 0;
up[0].port.dev = &op->dev;
up[0].flags |= SUNZILOG_FLAG_IS_CHANNEL_A;
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
if (keyboard_mouse)
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
up[0].flags |= SUNZILOG_FLAG_CONS_KEYB;
sunzilog_init_hw(&up[0]);
/* Channel B */
up[1].port.mapbase = op->resource[0].start + 0x04;
up[1].port.membase = (void __iomem *) &rp->channelB;
up[1].port.iotype = UPIO_MEM;
up[1].port.irq = op->irqs[0];
up[1].port.uartclk = ZS_CLOCK;
up[1].port.fifosize = 1;
up[1].port.ops = &sunzilog_pops;
up[1].port.type = PORT_SUNZILOG;
up[1].port.flags = 0;
up[1].port.line = (inst * 2) + 1;
up[1].port.dev = &op->dev;
up[1].flags |= 0;
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
if (keyboard_mouse)
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
up[1].flags |= SUNZILOG_FLAG_CONS_MOUSE;
sunzilog_init_hw(&up[1]);
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
if (!keyboard_mouse) {
[SPARC]: Fix serial console device detection. The current scheme works on static interpretation of text names, which is wrong. The output-device setting, for example, must be resolved via an alias or similar to a full path name to the console device. Paths also contain an optional set of 'options', which starts with a colon at the end of the path. The option area is used to specify which of two serial ports ('a' or 'b') the path refers to when a device node drives multiple ports. 'a' is assumed if the option specification is missing. This was caught by the UltraSPARC-T1 simulator. The 'output-device' property was set to 'ttya' and we didn't pick upon the fact that this is an OBP alias set to '/virtual-devices/console'. Instead we saw it as the first serial console device, instead of the hypervisor console. The infrastructure is now there to take advantage of this to resolve the console correctly even in multi-head situations in fbcon too. Thanks to Greg Onufer for the bug report. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-07-21 01:59:26 +02:00
if (sunserial_console_match(SUNZILOG_CONSOLE(), op->node,
&sunzilog_reg, up[0].port.line))
up->flags |= SUNZILOG_FLAG_IS_CONS;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
err = uart_add_one_port(&sunzilog_reg, &up[0].port);
if (err) {
[SPARC64]: Fix of_iounmap() region release. We need to pass in the resource otherwise we cannot release the region properly. We must know whether it is an I/O or MEM resource. Spotted by Eric Brower. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-12-29 06:01:32 +01:00
of_iounmap(&op->resource[0],
rp, sizeof(struct zilog_layout));
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
return err;
}
[SPARC]: Fix serial console device detection. The current scheme works on static interpretation of text names, which is wrong. The output-device setting, for example, must be resolved via an alias or similar to a full path name to the console device. Paths also contain an optional set of 'options', which starts with a colon at the end of the path. The option area is used to specify which of two serial ports ('a' or 'b') the path refers to when a device node drives multiple ports. 'a' is assumed if the option specification is missing. This was caught by the UltraSPARC-T1 simulator. The 'output-device' property was set to 'ttya' and we didn't pick upon the fact that this is an OBP alias set to '/virtual-devices/console'. Instead we saw it as the first serial console device, instead of the hypervisor console. The infrastructure is now there to take advantage of this to resolve the console correctly even in multi-head situations in fbcon too. Thanks to Greg Onufer for the bug report. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-07-21 01:59:26 +02:00
if (sunserial_console_match(SUNZILOG_CONSOLE(), op->node,
&sunzilog_reg, up[1].port.line))
up->flags |= SUNZILOG_FLAG_IS_CONS;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
err = uart_add_one_port(&sunzilog_reg, &up[1].port);
if (err) {
uart_remove_one_port(&sunzilog_reg, &up[0].port);
[SPARC64]: Fix of_iounmap() region release. We need to pass in the resource otherwise we cannot release the region properly. We must know whether it is an I/O or MEM resource. Spotted by Eric Brower. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-12-29 06:01:32 +01:00
of_iounmap(&op->resource[0],
rp, sizeof(struct zilog_layout));
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
return err;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
sparc: sunzilog uart order I have a sparcstation 20 clone with a lot of on board serial ports. The serial core code assumes that uarts are assigned contiguously and that may not be the case when there are multiple zs devices present. This patch insures that uart chips are placed in front of keyboard/mouse chips in the port table. ffd37420: ttyS0 at MMIO 0xf1100000 (irq = 44) is a zs (ESCC) Console: ttyS0 (SunZilog zs0) console [ttyS0] enabled ffd37420: ttyS1 at MMIO 0xf1100004 (irq = 44) is a zs (ESCC) ffd37500: Keyboard at MMIO 0xf1000000 (irq = 44) is a zs ffd37500: Mouse at MMIO 0xf1000004 (irq = 44) is a zs ffd3c5c0: ttyS2 at MMIO 0xf1100008 (irq = 44) is a zs (ESCC) ffd3c5c0: ttyS3 at MMIO 0xf110000c (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS4 at MMIO 0xf1100010 (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS5 at MMIO 0xf1100014 (irq = 44) is a zs (ESCC) ffd3c780: ttyS6 at MMIO 0xf1100018 (irq = 44) is a zs (ESCC) ffd3c780: ttyS7 at MMIO 0xf110001c (irq = 44) is a zs (ESCC) Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 12:37:51 +02:00
uart_inst++;
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
} else {
[SERIAL]: Fix 32-bit warnings in sunzilog.c and sunsu.c resource_size_t can be either a u64 or a u32, and we can't really know for sure, so when printing such a value out always use long-long printf formatting and cast the argument to that type. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-08-26 00:17:31 +02:00
printk(KERN_INFO "%s: Keyboard at MMIO 0x%llx (irq = %d) "
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
"is a %s\n",
[SERIAL]: Fix 32-bit warnings in sunzilog.c and sunsu.c resource_size_t can be either a u64 or a u32, and we can't really know for sure, so when printing such a value out always use long-long printf formatting and cast the argument to that type. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-08-26 00:17:31 +02:00
op->dev.bus_id,
(unsigned long long) up[0].port.mapbase,
op->irqs[0], sunzilog_type(&up[0].port));
printk(KERN_INFO "%s: Mouse at MMIO 0x%llx (irq = %d) "
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
"is a %s\n",
[SERIAL]: Fix 32-bit warnings in sunzilog.c and sunsu.c resource_size_t can be either a u64 or a u32, and we can't really know for sure, so when printing such a value out always use long-long printf formatting and cast the argument to that type. Signed-off-by: David S. Miller <davem@davemloft.net>
2007-08-26 00:17:31 +02:00
op->dev.bus_id,
(unsigned long long) up[1].port.mapbase,
op->irqs[0], sunzilog_type(&up[1].port));
sparc: sunzilog uart order I have a sparcstation 20 clone with a lot of on board serial ports. The serial core code assumes that uarts are assigned contiguously and that may not be the case when there are multiple zs devices present. This patch insures that uart chips are placed in front of keyboard/mouse chips in the port table. ffd37420: ttyS0 at MMIO 0xf1100000 (irq = 44) is a zs (ESCC) Console: ttyS0 (SunZilog zs0) console [ttyS0] enabled ffd37420: ttyS1 at MMIO 0xf1100004 (irq = 44) is a zs (ESCC) ffd37500: Keyboard at MMIO 0xf1000000 (irq = 44) is a zs ffd37500: Mouse at MMIO 0xf1000004 (irq = 44) is a zs ffd3c5c0: ttyS2 at MMIO 0xf1100008 (irq = 44) is a zs (ESCC) ffd3c5c0: ttyS3 at MMIO 0xf110000c (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS4 at MMIO 0xf1100010 (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS5 at MMIO 0xf1100014 (irq = 44) is a zs (ESCC) ffd3c780: ttyS6 at MMIO 0xf1100018 (irq = 44) is a zs (ESCC) ffd3c780: ttyS7 at MMIO 0xf110001c (irq = 44) is a zs (ESCC) Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 12:37:51 +02:00
kbm_inst++;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
[SERIAL] sunzilog: Register IRQ after all devices have been probed. Otherwise we will deref half-initialized channel pointers and crash in the interrupt handler. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-18 06:07:17 +02:00
dev_set_drvdata(&op->dev, &up[0]);
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
return 0;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
static void __devexit zs_remove_one(struct uart_sunzilog_port *up)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
{
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
if (ZS_IS_KEYB(up) || ZS_IS_MOUSE(up)) {
#ifdef CONFIG_SERIO
serio_unregister_port(&up->serio);
#endif
} else
uart_remove_one_port(&sunzilog_reg, &up->port);
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SPARC64]: Fix of_iounmap() region release. We need to pass in the resource otherwise we cannot release the region properly. We must know whether it is an I/O or MEM resource. Spotted by Eric Brower. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-12-29 06:01:32 +01:00
static int __devexit zs_remove(struct of_device *op)
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
{
[SPARC64]: Fix of_iounmap() region release. We need to pass in the resource otherwise we cannot release the region properly. We must know whether it is an I/O or MEM resource. Spotted by Eric Brower. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-12-29 06:01:32 +01:00
struct uart_sunzilog_port *up = dev_get_drvdata(&op->dev);
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
struct zilog_layout __iomem *regs;
zs_remove_one(&up[0]);
zs_remove_one(&up[1]);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
regs = sunzilog_chip_regs[up[0].port.line / 2];
[SPARC64]: Fix of_iounmap() region release. We need to pass in the resource otherwise we cannot release the region properly. We must know whether it is an I/O or MEM resource. Spotted by Eric Brower. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-12-29 06:01:32 +01:00
of_iounmap(&op->resource[0], regs, sizeof(struct zilog_layout));
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
[SPARC64]: Fix of_iounmap() region release. We need to pass in the resource otherwise we cannot release the region properly. We must know whether it is an I/O or MEM resource. Spotted by Eric Brower. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-12-29 06:01:32 +01:00
dev_set_drvdata(&op->dev, NULL);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
return 0;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
static struct of_device_id zs_match[] = {
{
.name = "zs",
},
{},
};
MODULE_DEVICE_TABLE(of, zs_match);
static struct of_platform_driver zs_driver = {
.name = "zs",
.match_table = zs_match,
.probe = zs_probe,
.remove = __devexit_p(zs_remove),
};
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
static int __init sunzilog_init(void)
{
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
struct device_node *dp;
sparc: sunzilog uart order I have a sparcstation 20 clone with a lot of on board serial ports. The serial core code assumes that uarts are assigned contiguously and that may not be the case when there are multiple zs devices present. This patch insures that uart chips are placed in front of keyboard/mouse chips in the port table. ffd37420: ttyS0 at MMIO 0xf1100000 (irq = 44) is a zs (ESCC) Console: ttyS0 (SunZilog zs0) console [ttyS0] enabled ffd37420: ttyS1 at MMIO 0xf1100004 (irq = 44) is a zs (ESCC) ffd37500: Keyboard at MMIO 0xf1000000 (irq = 44) is a zs ffd37500: Mouse at MMIO 0xf1000004 (irq = 44) is a zs ffd3c5c0: ttyS2 at MMIO 0xf1100008 (irq = 44) is a zs (ESCC) ffd3c5c0: ttyS3 at MMIO 0xf110000c (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS4 at MMIO 0xf1100010 (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS5 at MMIO 0xf1100014 (irq = 44) is a zs (ESCC) ffd3c780: ttyS6 at MMIO 0xf1100018 (irq = 44) is a zs (ESCC) ffd3c780: ttyS7 at MMIO 0xf110001c (irq = 44) is a zs (ESCC) Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 12:37:51 +02:00
int err;
int num_keybms = 0;
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
int num_sunzilog = 0;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
for_each_node_by_name(dp, "zs") {
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
num_sunzilog++;
[SERIAL] sunzilog: Fix instance enumeration. Just do a linear enumeration so that we handle sun4d systems correctly. As a consequence, eliminate the hard coded keyboard and mouse channel line values, use the CONS_{KEYB,MS} flags instead. Also, report the keyboard/mouse Zilog channels just like the uart ones do. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-20 07:55:08 +02:00
if (of_find_property(dp, "keyboard", NULL))
num_keybms++;
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
if (num_sunzilog) {
err = sunzilog_alloc_tables(num_sunzilog);
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
if (err)
[SERIAL] sunzilog: Register IRQ after all devices have been probed. Otherwise we will deref half-initialized channel pointers and crash in the interrupt handler. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-18 06:07:17 +02:00
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
sparc: sunzilog uart order I have a sparcstation 20 clone with a lot of on board serial ports. The serial core code assumes that uarts are assigned contiguously and that may not be the case when there are multiple zs devices present. This patch insures that uart chips are placed in front of keyboard/mouse chips in the port table. ffd37420: ttyS0 at MMIO 0xf1100000 (irq = 44) is a zs (ESCC) Console: ttyS0 (SunZilog zs0) console [ttyS0] enabled ffd37420: ttyS1 at MMIO 0xf1100004 (irq = 44) is a zs (ESCC) ffd37500: Keyboard at MMIO 0xf1000000 (irq = 44) is a zs ffd37500: Mouse at MMIO 0xf1000004 (irq = 44) is a zs ffd3c5c0: ttyS2 at MMIO 0xf1100008 (irq = 44) is a zs (ESCC) ffd3c5c0: ttyS3 at MMIO 0xf110000c (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS4 at MMIO 0xf1100010 (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS5 at MMIO 0xf1100014 (irq = 44) is a zs (ESCC) ffd3c780: ttyS6 at MMIO 0xf1100018 (irq = 44) is a zs (ESCC) ffd3c780: ttyS7 at MMIO 0xf110001c (irq = 44) is a zs (ESCC) Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 12:37:51 +02:00
uart_chip_count = num_sunzilog - num_keybms;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
sparc: sunzilog uart order I have a sparcstation 20 clone with a lot of on board serial ports. The serial core code assumes that uarts are assigned contiguously and that may not be the case when there are multiple zs devices present. This patch insures that uart chips are placed in front of keyboard/mouse chips in the port table. ffd37420: ttyS0 at MMIO 0xf1100000 (irq = 44) is a zs (ESCC) Console: ttyS0 (SunZilog zs0) console [ttyS0] enabled ffd37420: ttyS1 at MMIO 0xf1100004 (irq = 44) is a zs (ESCC) ffd37500: Keyboard at MMIO 0xf1000000 (irq = 44) is a zs ffd37500: Mouse at MMIO 0xf1000004 (irq = 44) is a zs ffd3c5c0: ttyS2 at MMIO 0xf1100008 (irq = 44) is a zs (ESCC) ffd3c5c0: ttyS3 at MMIO 0xf110000c (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS4 at MMIO 0xf1100010 (irq = 44) is a zs (ESCC) ffd3c6a0: ttyS5 at MMIO 0xf1100014 (irq = 44) is a zs (ESCC) ffd3c780: ttyS6 at MMIO 0xf1100018 (irq = 44) is a zs (ESCC) ffd3c780: ttyS7 at MMIO 0xf110001c (irq = 44) is a zs (ESCC) Signed-off-by: Robert Reif <reif@earthlink.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-04-24 12:37:51 +02:00
err = sunserial_register_minors(&sunzilog_reg,
uart_chip_count * 2);
[SERIAL] sunzilog: Register IRQ after all devices have been probed. Otherwise we will deref half-initialized channel pointers and crash in the interrupt handler. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-18 06:07:17 +02:00
if (err)
goto out_free_tables;
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
}
[SERIAL] sunzilog: Register IRQ after all devices have been probed. Otherwise we will deref half-initialized channel pointers and crash in the interrupt handler. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-18 06:07:17 +02:00
err = of_register_driver(&zs_driver, &of_bus_type);
if (err)
goto out_unregister_uart;
if (zilog_irq != -1) {
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
struct uart_sunzilog_port *up = sunzilog_irq_chain;
[SERIAL] sunzilog: Register IRQ after all devices have been probed. Otherwise we will deref half-initialized channel pointers and crash in the interrupt handler. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-18 06:07:17 +02:00
err = request_irq(zilog_irq, sunzilog_interrupt, IRQF_SHARED,
"zs", sunzilog_irq_chain);
if (err)
goto out_unregister_driver;
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
/* Enable Interrupts */
while (up) {
struct zilog_channel __iomem *channel;
/* printk (KERN_INFO "Enable IRQ for ZILOG Hardware %p\n", up); */
channel = ZILOG_CHANNEL_FROM_PORT(&up->port);
up->flags |= SUNZILOG_FLAG_ISR_HANDLER;
up->curregs[R9] |= MIE;
write_zsreg(channel, R9, up->curregs[R9]);
up = up->next;
}
[SERIAL] sunzilog: Register IRQ after all devices have been probed. Otherwise we will deref half-initialized channel pointers and crash in the interrupt handler. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-18 06:07:17 +02:00
}
out:
return err;
out_unregister_driver:
of_unregister_driver(&zs_driver);
out_unregister_uart:
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
if (num_sunzilog) {
sunserial_unregister_minors(&sunzilog_reg, num_sunzilog);
[SERIAL] sunzilog: Register IRQ after all devices have been probed. Otherwise we will deref half-initialized channel pointers and crash in the interrupt handler. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-18 06:07:17 +02:00
sunzilog_reg.cons = NULL;
}
out_free_tables:
sunzilog_free_tables();
goto out;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
static void __exit sunzilog_exit(void)
{
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
of_unregister_driver(&zs_driver);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
if (zilog_irq != -1) {
[SERIAL] sunzilog: Interrupt enable before ISR handler installed This patch changes the interrupt enable sequence for the sunzilog driver so that interrupts are not enabled untill after the interrupt handler has been installed. If this is not done, some SS1 and SS2 sun4c systems panic on un-handled interrupt before the handler gets installed preventing boot. It also adds in support for the ESCC version of the zilog chips. The changes mean that the FIFO will be enabled for ESCC versions of the SCC UART. My interpretation of the SCC manual and the existing interrupt handler code is that it sould be able to make good use of the FIFO without issues. Signed-off-by: Mark Fortescue <mark@mtfhpc.demon.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-05-09 22:49:04 +02:00
struct uart_sunzilog_port *up = sunzilog_irq_chain;
/* Disable Interrupts */
while (up) {
struct zilog_channel __iomem *channel;
/* printk (KERN_INFO "Disable IRQ for ZILOG Hardware %p\n", up); */
channel = ZILOG_CHANNEL_FROM_PORT(&up->port);
up->flags &= ~SUNZILOG_FLAG_ISR_HANDLER;
up->curregs[R9] &= ~MIE;
write_zsreg(channel, R9, up->curregs[R9]);
up = up->next;
}
[SERIAL] sunzilog: Fix bugs in device deregristration. 1) Need to unregister 2 ports per of_device. 2) Need to of_iounmap() 1 mapping per of_device. 3) Need to free up the IRQ only after all devices have been unregistered. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:40 +02:00
free_irq(zilog_irq, sunzilog_irq_chain);
zilog_irq = -1;
}
[SERIAL] sparc: Infrastructure to fix section mismatch bugs. This patch against 2.6.23 sparc-2.6.git contains a number of minor cleanups of the sparc serial drivers. Initially I fixed this build warning: WARNING: vmlinux.o(.text+0x107a2c): Section mismatch: reference to .init.text:add_preferred_console (between 'sunserial_console_match' and 'sunserial_console_termios') which is done by declaring sunserial_console_match() as __init. This resulted in build warnings on sunserial_current_minor. To resolve these the variable was changed so it is no longer global, and to hide operations on it inside 2 new functions. These functions handle the UART minor handling code that is common to all sparc serial drivers. These changes allowed to clean up the uart counters in all the sparc serial drivers, and the administration of minor device numbers. Lastly, sunserial_console_termios() does not need to be exported since it is only called from non-modular code. Sadly, the following build warning still exists: WARNING: vmlinux.o(__ksymtab+0x2910): Section mismatch: reference to .init.text:sunserial_console_match (between '__ksymtab_sunserial_console_match' and '__ksymtab_sunserial_unregister_minors') This could be resolved by not exporting sunserial_console_match(), but this is not possible at the moment because it is being called from modular code. On the other hand, this is a bogus warning since it comes from a ksymtab section. Signed-off-by: Martin Habets <errandir_news@mph.eclipse.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2007-12-11 12:37:04 +01:00
if (sunzilog_reg.nr) {
sunserial_unregister_minors(&sunzilog_reg, sunzilog_reg.nr);
[SERIAL] sunzilog: Convert to of_driver. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:13:17 +02:00
sunzilog_free_tables();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
}
}
module_init(sunzilog_init);
module_exit(sunzilog_exit);
MODULE_AUTHOR("David S. Miller");
MODULE_DESCRIPTION("Sun Zilog serial port driver");
[SERIAL] sun{su,zilog}: Add missing MODULE_*() niceties. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-30 00:14:17 +02:00
MODULE_VERSION("2.0");
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 00:20:36 +02:00
MODULE_LICENSE("GPL");