ide-io.c

linux 内核源代码

/*
 *	IDE I/O functions
 *
 *	Basic PIO and command management functionality.
 *
 * This code was split off from ide.c. See ide.c for history and original
 * copyrights.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of the GNU General Public License as published by the
 * Free Software Foundation; either version 2, or (at your option) any
 * later version.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * For the avoidance of doubt the "preferred form" of this code is one which
 * is in an open non patent encumbered format. Where cryptographic key signing
 * forms part of the process of creating an executable the information
 * including keys needed to generate an equivalently functional executable
 * are deemed to be part of the source code.
 */
 
 
#include <linux/module.h>
#include <linux/types.h>
#include <linux/string.h>
#include <linux/kernel.h>
#include <linux/timer.h>
#include <linux/mm.h>
#include <linux/interrupt.h>
#include <linux/major.h>
#include <linux/errno.h>
#include <linux/genhd.h>
#include <linux/blkpg.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/pci.h>
#include <linux/delay.h>
#include <linux/ide.h>
#include <linux/completion.h>
#include <linux/reboot.h>
#include <linux/cdrom.h>
#include <linux/seq_file.h>
#include <linux/device.h>
#include <linux/kmod.h>
#include <linux/scatterlist.h>
#include <linux/bitops.h>

#include <asm/byteorder.h>
#include <asm/irq.h>
#include <asm/uaccess.h>
#include <asm/io.h>

static int __ide_end_request(ide_drive_t *drive, struct request *rq,
			     int uptodate, unsigned int nr_bytes, int dequeue)
{
	int ret = 1;

	/*
	 * if failfast is set on a request, override number of sectors and
	 * complete the whole request right now
	 */
	if (blk_noretry_request(rq) && end_io_error(uptodate))
		nr_bytes = rq->hard_nr_sectors << 9;

	if (!blk_fs_request(rq) && end_io_error(uptodate) && !rq->errors)
		rq->errors = -EIO;

	/*
	 * decide whether to reenable DMA -- 3 is a random magic for now,
	 * if we DMA timeout more than 3 times, just stay in PIO
	 */
	if (drive->state == DMA_PIO_RETRY && drive->retry_pio <= 3) {
		drive->state = 0;
		HWGROUP(drive)->hwif->ide_dma_on(drive);
	}

	if (!end_that_request_chunk(rq, uptodate, nr_bytes)) {
		add_disk_randomness(rq->rq_disk);
		if (dequeue) {
			if (!list_empty(&rq->queuelist))
				blkdev_dequeue_request(rq);
			HWGROUP(drive)->rq = NULL;
		}
		end_that_request_last(rq, uptodate);
		ret = 0;
	}

	return ret;
}

/**
 *	ide_end_request		-	complete an IDE I/O
 *	@drive: IDE device for the I/O
 *	@uptodate:
 *	@nr_sectors: number of sectors completed
 *
 *	This is our end_request wrapper function. We complete the I/O
 *	update random number input and dequeue the request, which if
 *	it was tagged may be out of order.
 */

int ide_end_request (ide_drive_t *drive, int uptodate, int nr_sectors)
{
	unsigned int nr_bytes = nr_sectors << 9;
	struct request *rq;
	unsigned long flags;
	int ret = 1;

	/*
	 * room for locking improvements here, the calls below don't
	 * need the queue lock held at all
	 */
	spin_lock_irqsave(&ide_lock, flags);
	rq = HWGROUP(drive)->rq;

	if (!nr_bytes) {
		if (blk_pc_request(rq))
			nr_bytes = rq->data_len;
		else
			nr_bytes = rq->hard_cur_sectors << 9;
	}

	ret = __ide_end_request(drive, rq, uptodate, nr_bytes, 1);

	spin_unlock_irqrestore(&ide_lock, flags);
	return ret;
}
EXPORT_SYMBOL(ide_end_request);

/*
 * Power Management state machine. This one is rather trivial for now,
 * we should probably add more, like switching back to PIO on suspend
 * to help some BIOSes, re-do the door locking on resume, etc...
 */

enum {
	ide_pm_flush_cache	= ide_pm_state_start_suspend,
	idedisk_pm_standby,

	idedisk_pm_restore_pio	= ide_pm_state_start_resume,
	idedisk_pm_idle,
	ide_pm_restore_dma,
};

static void ide_complete_power_step(ide_drive_t *drive, struct request *rq, u8 stat, u8 error)
{
	struct request_pm_state *pm = rq->data;

	if (drive->media != ide_disk)
		return;

	switch (pm->pm_step) {
	case ide_pm_flush_cache:	/* Suspend step 1 (flush cache) complete */
		if (pm->pm_state == PM_EVENT_FREEZE)
			pm->pm_step = ide_pm_state_completed;
		else
			pm->pm_step = idedisk_pm_standby;
		break;
	case idedisk_pm_standby:	/* Suspend step 2 (standby) complete */
		pm->pm_step = ide_pm_state_completed;
		break;
	case idedisk_pm_restore_pio:	/* Resume step 1 complete */
		pm->pm_step = idedisk_pm_idle;
		break;
	case idedisk_pm_idle:		/* Resume step 2 (idle) complete */
		pm->pm_step = ide_pm_restore_dma;
		break;
	}
}

static ide_startstop_t ide_start_power_step(ide_drive_t *drive, struct request *rq)
{
	struct request_pm_state *pm = rq->data;
	ide_task_t *args = rq->special;

	memset(args, 0, sizeof(*args));

	switch (pm->pm_step) {
	case ide_pm_flush_cache:	/* Suspend step 1 (flush cache) */
		if (drive->media != ide_disk)
			break;
		/* Not supported? Switch to next step now. */
		if (!drive->wcache || !ide_id_has_flush_cache(drive->id)) {
			ide_complete_power_step(drive, rq, 0, 0);
			return ide_stopped;
		}
		if (ide_id_has_flush_cache_ext(drive->id))
			args->tfRegister[IDE_COMMAND_OFFSET] = WIN_FLUSH_CACHE_EXT;
		else
			args->tfRegister[IDE_COMMAND_OFFSET] = WIN_FLUSH_CACHE;
		args->command_type = IDE_DRIVE_TASK_NO_DATA;
		args->handler	   = &task_no_data_intr;
		return do_rw_taskfile(drive, args);

	case idedisk_pm_standby:	/* Suspend step 2 (standby) */
		args->tfRegister[IDE_COMMAND_OFFSET] = WIN_STANDBYNOW1;
		args->command_type = IDE_DRIVE_TASK_NO_DATA;
		args->handler	   = &task_no_data_intr;
		return do_rw_taskfile(drive, args);

	case idedisk_pm_restore_pio:	/* Resume step 1 (restore PIO) */
		ide_set_max_pio(drive);
		/*
		 * skip idedisk_pm_idle for ATAPI devices
		 */
		if (drive->media != ide_disk)
			pm->pm_step = ide_pm_restore_dma;
		else
			ide_complete_power_step(drive, rq, 0, 0);
		return ide_stopped;

	case idedisk_pm_idle:		/* Resume step 2 (idle) */
		args->tfRegister[IDE_COMMAND_OFFSET] = WIN_IDLEIMMEDIATE;
		args->command_type = IDE_DRIVE_TASK_NO_DATA;
		args->handler = task_no_data_intr;
		return do_rw_taskfile(drive, args);

	case ide_pm_restore_dma:	/* Resume step 3 (restore DMA) */
		/*
		 * Right now, all we do is call ide_set_dma(drive),
		 * we could be smarter and check for current xfer_speed
		 * in struct drive etc...
		 */
		if (drive->hwif->ide_dma_on == NULL)
			break;
		drive->hwif->dma_off_quietly(drive);
		/*
		 * TODO: respect ->using_dma setting
		 */
		ide_set_dma(drive);
		break;
	}
	pm->pm_step = ide_pm_state_completed;
	return ide_stopped;
}

/**
 *	ide_end_dequeued_request	-	complete an IDE I/O
 *	@drive: IDE device for the I/O
 *	@uptodate:
 *	@nr_sectors: number of sectors completed
 *
 *	Complete an I/O that is no longer on the request queue. This
 *	typically occurs when we pull the request and issue a REQUEST_SENSE.
 *	We must still finish the old request but we must not tamper with the
 *	queue in the meantime.
 *
 *	NOTE: This path does not handle barrier, but barrier is not supported
 *	on ide-cd anyway.
 */

int ide_end_dequeued_request(ide_drive_t *drive, struct request *rq,
			     int uptodate, int nr_sectors)
{
	unsigned long flags;
	int ret;

	spin_lock_irqsave(&ide_lock, flags);
	BUG_ON(!blk_rq_started(rq));
	ret = __ide_end_request(drive, rq, uptodate, nr_sectors << 9, 0);
	spin_unlock_irqrestore(&ide_lock, flags);

	return ret;
}
EXPORT_SYMBOL_GPL(ide_end_dequeued_request);


/**
 *	ide_complete_pm_request - end the current Power Management request
 *	@drive: target drive
 *	@rq: request
 *
 *	This function cleans up the current PM request and stops the queue
 *	if necessary.
 */
static void ide_complete_pm_request (ide_drive_t *drive, struct request *rq)
{
	unsigned long flags;

#ifdef DEBUG_PM
	printk("%s: completing PM request, %s\n", drive->name,
	       blk_pm_suspend_request(rq) ? "suspend" : "resume");
#endif
	spin_lock_irqsave(&ide_lock, flags);
	if (blk_pm_suspend_request(rq)) {
		blk_stop_queue(drive->queue);
	} else {
		drive->blocked = 0;
		blk_start_queue(drive->queue);
	}
	blkdev_dequeue_request(rq);
	HWGROUP(drive)->rq = NULL;
	end_that_request_last(rq, 1);
	spin_unlock_irqrestore(&ide_lock, flags);
}

/**
 *	ide_end_drive_cmd	-	end an explicit drive command
 *	@drive: command 
 *	@stat: status bits
 *	@err: error bits
 *
 *	Clean up after success/failure of an explicit drive command.
 *	These get thrown onto the queue so they are synchronized with
 *	real I/O operations on the drive.
 *
 *	In LBA48 mode we have to read the register set twice to get
 *	all the extra information out.
 */
 
void ide_end_drive_cmd (ide_drive_t *drive, u8 stat, u8 err)
{
	ide_hwif_t *hwif = HWIF(drive);
	unsigned long flags;
	struct request *rq;

	spin_lock_irqsave(&ide_lock, flags);
	rq = HWGROUP(drive)->rq;
	spin_unlock_irqrestore(&ide_lock, flags);

	if (rq->cmd_type == REQ_TYPE_ATA_CMD) {
		u8 *args = (u8 *) rq->buffer;
		if (rq->errors == 0)
			rq->errors = !OK_STAT(stat,READY_STAT,BAD_STAT);

		if (args) {
			args[0] = stat;
			args[1] = err;
			args[2] = hwif->INB(IDE_NSECTOR_REG);
		}
	} else if (rq->cmd_type == REQ_TYPE_ATA_TASK) {
		u8 *args = (u8 *) rq->buffer;
		if (rq->errors == 0)
			rq->errors = !OK_STAT(stat,READY_STAT,BAD_STAT);

		if (args) {
			args[0] = stat;
			args[1] = err;
			/* be sure we're looking at the low order bits */
			hwif->OUTB(drive->ctl & ~0x80, IDE_CONTROL_REG);
			args[2] = hwif->INB(IDE_NSECTOR_REG);
			args[3] = hwif->INB(IDE_SECTOR_REG);
			args[4] = hwif->INB(IDE_LCYL_REG);
			args[5] = hwif->INB(IDE_HCYL_REG);
			args[6] = hwif->INB(IDE_SELECT_REG);
		}
	} else if (rq->cmd_type == REQ_TYPE_ATA_TASKFILE) {
		ide_task_t *args = (ide_task_t *) rq->special;
		if (rq->errors == 0)
			rq->errors = !OK_STAT(stat,READY_STAT,BAD_STAT);
			
		if (args) {
			if (args->tf_in_flags.b.data) {
				u16 data				= hwif->INW(IDE_DATA_REG);
				args->tfRegister[IDE_DATA_OFFSET]	= (data) & 0xFF;
				args->hobRegister[IDE_DATA_OFFSET]	= (data >> 8) & 0xFF;
			}
			args->tfRegister[IDE_ERROR_OFFSET]   = err;
			/* be sure we're looking at the low order bits */
			hwif->OUTB(drive->ctl & ~0x80, IDE_CONTROL_REG);
			args->tfRegister[IDE_NSECTOR_OFFSET] = hwif->INB(IDE_NSECTOR_REG);
			args->tfRegister[IDE_SECTOR_OFFSET]  = hwif->INB(IDE_SECTOR_REG);
			args->tfRegister[IDE_LCYL_OFFSET]    = hwif->INB(IDE_LCYL_REG);
			args->tfRegister[IDE_HCYL_OFFSET]    = hwif->INB(IDE_HCYL_REG);
			args->tfRegister[IDE_SELECT_OFFSET]  = hwif->INB(IDE_SELECT_REG);
			args->tfRegister[IDE_STATUS_OFFSET]  = stat;

			if (drive->addressing == 1) {
				hwif->OUTB(drive->ctl|0x80, IDE_CONTROL_REG);
				args->hobRegister[IDE_FEATURE_OFFSET]	= hwif->INB(IDE_FEATURE_REG);
				args->hobRegister[IDE_NSECTOR_OFFSET]	= hwif->INB(IDE_NSECTOR_REG);
				args->hobRegister[IDE_SECTOR_OFFSET]	= hwif->INB(IDE_SECTOR_REG);
				args->hobRegister[IDE_LCYL_OFFSET]	= hwif->INB(IDE_LCYL_REG);
				args->hobRegister[IDE_HCYL_OFFSET]	= hwif->INB(IDE_HCYL_REG);
			}
		}
	} else if (blk_pm_request(rq)) {
		struct request_pm_state *pm = rq->data;
#ifdef DEBUG_PM
		printk("%s: complete_power_step(step: %d, stat: %x, err: %x)\n",
			drive->name, rq->pm->pm_step, stat, err);
#endif
		ide_complete_power_step(drive, rq, stat, err);
		if (pm->pm_step == ide_pm_state_completed)
			ide_complete_pm_request(drive, rq);
		return;
	}

	spin_lock_irqsave(&ide_lock, flags);
	blkdev_dequeue_request(rq);
	HWGROUP(drive)->rq = NULL;
	rq->errors = err;
	end_that_request_last(rq, !rq->errors);
	spin_unlock_irqrestore(&ide_lock, flags);
}

EXPORT_SYMBOL(ide_end_drive_cmd);

/**
 *	try_to_flush_leftover_data	-	flush junk
 *	@drive: drive to flush
 *
 *	try_to_flush_leftover_data() is invoked in response to a drive
 *	unexpectedly having its DRQ_STAT bit set.  As an alternative to
 *	resetting the drive, this routine tries to clear the condition
 *	by read a sector's worth of data from the drive.  Of course,
 *	this may not help if the drive is *waiting* for data from *us*.
 */
static void try_to_flush_leftover_data (ide_drive_t *drive)
{
	int i = (drive->mult_count ? drive->mult_count : 1) * SECTOR_WORDS;

	if (drive->media != ide_disk)
		return;
	while (i > 0) {
		u32 buffer[16];
		u32 wcount = (i > 16) ? 16 : i;

		i -= wcount;
		HWIF(drive)->ata_input_data(drive, buffer, wcount);
	}
}

static void ide_kill_rq(ide_drive_t *drive, struct request *rq)
{
	if (rq->rq_disk) {
		ide_driver_t *drv;

		drv = *(ide_driver_t **)rq->rq_disk->private_data;
		drv->end_request(drive, 0, 0);
	} else
		ide_end_request(drive, 0, 0);
}

static ide_startstop_t ide_ata_error(ide_drive_t *drive, struct request *rq, u8 stat, u8 err)
{
	ide_hwif_t *hwif = drive->hwif;

	if (stat & BUSY_STAT || ((stat & WRERR_STAT) && !drive->nowerr)) {
		/* other bits are useless when BUSY */
		rq->errors |= ERROR_RESET;
	} else if (stat & ERR_STAT) {
		/* err has different meaning on cdrom and tape */