ide-disk.c

linux设备驱动开发详解（书代码）

/*
 *  linux/drivers/ide/ide-disk.c	Version 1.18	Mar 05, 2003
 *
 *  Copyright (C) 1994-1998  Linus Torvalds & authors (see below)
 *  Copyright (C) 1998-2002  Linux ATA Development
 *				Andre Hedrick <[e-mail]>
 *  Copyright (C) 2003	     Red Hat <[e-mail]>
 */

/*
 *  Mostly written by Mark Lord <[e-mail]>
 *                and Gadi Oxman <[e-mail]>
 *                and Andre Hedrick <[e-mail]>
 *
 * This is the IDE/ATA disk driver, as evolved from hd.c and ide.c.
 *
 * Version 1.00		move disk only code from ide.c to ide-disk.c
 *			support optional byte-swapping of all data
 * Version 1.01		fix previous byte-swapping code
 * Version 1.02		remove ", LBA" from drive identification msgs
 * Version 1.03		fix display of id->buf_size for big-endian
 * Version 1.04		add /proc configurable settings and S.M.A.R.T support
 * Version 1.05		add capacity support for ATA3 >= 8GB
 * Version 1.06		get boot-up messages to show full cyl count
 * Version 1.07		disable door-locking if it fails
 * Version 1.08		fixed CHS/LBA translations for ATA4 > 8GB,
 *			process of adding new ATA4 compliance.
 *			fixed problems in allowing fdisk to see
 *			the entire disk.
 * Version 1.09		added increment of rq->sector in ide_multwrite
 *			added UDMA 3/4 reporting
 * Version 1.10		request queue changes, Ultra DMA 100
 * Version 1.11		added 48-bit lba
 * Version 1.12		adding taskfile io access method
 * Version 1.13		added standby and flush-cache for notifier
 * Version 1.14		added acoustic-wcache
 * Version 1.15		convert all calls to ide_raw_taskfile
 *				since args will return register content.
 * Version 1.16		added suspend-resume-checkpower
 * Version 1.17		do flush on standy, do flush on ATA < ATA6
 *			fix wcache setup.
 */

#define IDEDISK_VERSION	"1.18"

#undef REALLY_SLOW_IO		/* most systems can safely undef this */

//#define DEBUG

#include <linux/config.h>
#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/slab.h>
#include <linux/delay.h>

#define _IDE_DISK

#include <linux/ide.h>

#include <asm/byteorder.h>
#include <asm/irq.h>
#include <asm/uaccess.h>
#include <asm/io.h>
#include <asm/div64.h>

struct ide_disk_obj {
	ide_drive_t	*drive;
	ide_driver_t	*driver;
	struct gendisk	*disk;
	struct kref	kref;
};

static DECLARE_MUTEX(idedisk_ref_sem);

#define to_ide_disk(obj) container_of(obj, struct ide_disk_obj, kref)

#define ide_disk_g(disk) \
	container_of((disk)->private_data, struct ide_disk_obj, driver)

static struct ide_disk_obj *ide_disk_get(struct gendisk *disk)
{
	struct ide_disk_obj *idkp = NULL;

	down(&idedisk_ref_sem);
	idkp = ide_disk_g(disk);
	if (idkp)
		kref_get(&idkp->kref);
	up(&idedisk_ref_sem);
	return idkp;
}

static void ide_disk_release(struct kref *);

static void ide_disk_put(struct ide_disk_obj *idkp)
{
	down(&idedisk_ref_sem);
	kref_put(&idkp->kref, ide_disk_release);
	up(&idedisk_ref_sem);
}

/*
 * lba_capacity_is_ok() performs a sanity check on the claimed "lba_capacity"
 * value for this drive (from its reported identification information).
 *
 * Returns:	1 if lba_capacity looks sensible
 *		0 otherwise
 *
 * It is called only once for each drive.
 */
static int lba_capacity_is_ok (struct hd_driveid *id)
{
	unsigned long lba_sects, chs_sects, head, tail;

	/* No non-LBA info .. so valid! */
	if (id->cyls == 0)
		return 1;

	/*
	 * The ATA spec tells large drives to return
	 * C/H/S = 16383/16/63 independent of their size.
	 * Some drives can be jumpered to use 15 heads instead of 16.
	 * Some drives can be jumpered to use 4092 cyls instead of 16383.
	 */
	if ((id->cyls == 16383
	     || (id->cyls == 4092 && id->cur_cyls == 16383)) &&
	    id->sectors == 63 &&
	    (id->heads == 15 || id->heads == 16) &&
	    (id->lba_capacity >= 16383*63*id->heads))
		return 1;

	lba_sects   = id->lba_capacity;
	chs_sects   = id->cyls * id->heads * id->sectors;

	/* perform a rough sanity check on lba_sects:  within 10% is OK */
	if ((lba_sects - chs_sects) < chs_sects/10)
		return 1;

	/* some drives have the word order reversed */
	head = ((lba_sects >> 16) & 0xffff);
	tail = (lba_sects & 0xffff);
	lba_sects = (head | (tail << 16));
	if ((lba_sects - chs_sects) < chs_sects/10) {
		id->lba_capacity = lba_sects;
		return 1;	/* lba_capacity is (now) good */
	}

	return 0;	/* lba_capacity value may be bad */
}

/*
 * __ide_do_rw_disk() issues READ and WRITE commands to a disk,
 * using LBA if supported, or CHS otherwise, to address sectors.
 */
static ide_startstop_t __ide_do_rw_disk(ide_drive_t *drive, struct request *rq, sector_t block)
{
	ide_hwif_t *hwif	= HWIF(drive);
	unsigned int dma	= drive->using_dma;
	u8 lba48		= (drive->addressing == 1) ? 1 : 0;
	task_ioreg_t command	= WIN_NOP;
	ata_nsector_t		nsectors;

	nsectors.all		= (u16) rq->nr_sectors;

	if (hwif->no_lba48_dma && lba48 && dma) {
		if (block + rq->nr_sectors > 1ULL << 28)
			dma = 0;
		else
			lba48 = 0;
	}

	if (!dma) {
		ide_init_sg_cmd(drive, rq);
		ide_map_sg(drive, rq);
	}

	if (IDE_CONTROL_REG)
		hwif->OUTB(drive->ctl, IDE_CONTROL_REG);

	/* FIXME: SELECT_MASK(drive, 0) ? */

	if (drive->select.b.lba) {
		if (lba48) {
			task_ioreg_t tasklets[10];

			pr_debug("%s: LBA=0x%012llx\n", drive->name,
					(unsigned long long)block);

			tasklets[0] = 0;
			tasklets[1] = 0;
			tasklets[2] = nsectors.b.low;
			tasklets[3] = nsectors.b.high;
			tasklets[4] = (task_ioreg_t) block;
			tasklets[5] = (task_ioreg_t) (block>>8);
			tasklets[6] = (task_ioreg_t) (block>>16);
			tasklets[7] = (task_ioreg_t) (block>>24);
			if (sizeof(block) == 4) {
				tasklets[8] = (task_ioreg_t) 0;
				tasklets[9] = (task_ioreg_t) 0;
			} else {
				tasklets[8] = (task_ioreg_t)((u64)block >> 32);
				tasklets[9] = (task_ioreg_t)((u64)block >> 40);
			}
#ifdef DEBUG
			printk("%s: 0x%02x%02x 0x%02x%02x%02x%02x%02x%02x\n",
				drive->name, tasklets[3], tasklets[2],
				tasklets[9], tasklets[8], tasklets[7],
				tasklets[6], tasklets[5], tasklets[4]);
#endif
			hwif->OUTB(tasklets[1], IDE_FEATURE_REG);
			hwif->OUTB(tasklets[3], IDE_NSECTOR_REG);
			hwif->OUTB(tasklets[7], IDE_SECTOR_REG);
			hwif->OUTB(tasklets[8], IDE_LCYL_REG);
			hwif->OUTB(tasklets[9], IDE_HCYL_REG);

			hwif->OUTB(tasklets[0], IDE_FEATURE_REG);
			hwif->OUTB(tasklets[2], IDE_NSECTOR_REG);
			hwif->OUTB(tasklets[4], IDE_SECTOR_REG);
			hwif->OUTB(tasklets[5], IDE_LCYL_REG);
			hwif->OUTB(tasklets[6], IDE_HCYL_REG);
			hwif->OUTB(0x00|drive->select.all,IDE_SELECT_REG);
		} else {
			hwif->OUTB(0x00, IDE_FEATURE_REG);
			hwif->OUTB(nsectors.b.low, IDE_NSECTOR_REG);
			hwif->OUTB(block, IDE_SECTOR_REG);
			hwif->OUTB(block>>=8, IDE_LCYL_REG);
			hwif->OUTB(block>>=8, IDE_HCYL_REG);
			hwif->OUTB(((block>>8)&0x0f)|drive->select.all,IDE_SELECT_REG);
		}
	} else {
		unsigned int sect,head,cyl,track;
		track = (int)block / drive->sect;
		sect  = (int)block % drive->sect + 1;
		hwif->OUTB(sect, IDE_SECTOR_REG);
		head  = track % drive->head;
		cyl   = track / drive->head;

		pr_debug("%s: CHS=%u/%u/%u\n", drive->name, cyl, head, sect);

		hwif->OUTB(0x00, IDE_FEATURE_REG);
		hwif->OUTB(nsectors.b.low, IDE_NSECTOR_REG);
		hwif->OUTB(cyl, IDE_LCYL_REG);
		hwif->OUTB(cyl>>8, IDE_HCYL_REG);
		hwif->OUTB(head|drive->select.all,IDE_SELECT_REG);
	}

	if (dma) {
		if (!hwif->dma_setup(drive)) {
			if (rq_data_dir(rq)) {
				command = lba48 ? WIN_WRITEDMA_EXT : WIN_WRITEDMA;
				if (drive->vdma)
					command = lba48 ? WIN_WRITE_EXT: WIN_WRITE;
			} else {
				command = lba48 ? WIN_READDMA_EXT : WIN_READDMA;
				if (drive->vdma)
					command = lba48 ? WIN_READ_EXT: WIN_READ;
			}
			hwif->dma_exec_cmd(drive, command);
			hwif->dma_start(drive);
			return ide_started;
		}
		/* fallback to PIO */
		ide_init_sg_cmd(drive, rq);
	}

	if (rq_data_dir(rq) == READ) {

		if (drive->mult_count) {
			hwif->data_phase = TASKFILE_MULTI_IN;
			command = lba48 ? WIN_MULTREAD_EXT : WIN_MULTREAD;
		} else {
			hwif->data_phase = TASKFILE_IN;
			command = lba48 ? WIN_READ_EXT : WIN_READ;
		}

		ide_execute_command(drive, command, &task_in_intr, WAIT_CMD, NULL);
		return ide_started;
	} else {
		if (drive->mult_count) {
			hwif->data_phase = TASKFILE_MULTI_OUT;
			command = lba48 ? WIN_MULTWRITE_EXT : WIN_MULTWRITE;
		} else {
			hwif->data_phase = TASKFILE_OUT;
			command = lba48 ? WIN_WRITE_EXT : WIN_WRITE;
		}

		/* FIXME: ->OUTBSYNC ? */
		hwif->OUTB(command, IDE_COMMAND_REG);

		return pre_task_out_intr(drive, rq);
	}
}

/*
 * 268435455  == 137439 MB or 28bit limit
 * 320173056  == 163929 MB or 48bit addressing
 * 1073741822 == 549756 MB or 48bit addressing fake drive
 */

static ide_startstop_t ide_do_rw_disk (ide_drive_t *drive, struct request *rq, sector_t block)
{
	ide_hwif_t *hwif = HWIF(drive);

	BUG_ON(drive->blocked);

	if (!blk_fs_request(rq)) {
		blk_dump_rq_flags(rq, "ide_do_rw_disk - bad command");
		ide_end_request(drive, 0, 0);
		return ide_stopped;
	}

	pr_debug("%s: %sing: block=%llu, sectors=%lu, buffer=0x%08lx\n",
		 drive->name, rq_data_dir(rq) == READ ? "read" : "writ",
		 (unsigned long long)block, rq->nr_sectors,
		 (unsigned long)rq->buffer);

	if (hwif->rw_disk)
		hwif->rw_disk(drive, rq);

	return __ide_do_rw_disk(drive, rq, block);
}

/*
 * Queries for true maximum capacity of the drive.
 * Returns maximum LBA address (> 0) of the drive, 0 if failed.
 */
static unsigned long idedisk_read_native_max_address(ide_drive_t *drive)
{
	ide_task_t args;
	unsigned long addr = 0;

	/* Create IDE/ATA command request structure */
	memset(&args, 0, sizeof(ide_task_t));
	args.tfRegister[IDE_SELECT_OFFSET]	= 0x40;
	args.tfRegister[IDE_COMMAND_OFFSET]	= WIN_READ_NATIVE_MAX;
	args.command_type			= IDE_DRIVE_TASK_NO_DATA;
	args.handler				= &task_no_data_intr;
	/* submit command request */
	ide_raw_taskfile(drive, &args, NULL);

	/* if OK, compute maximum address value */
	if ((args.tfRegister[IDE_STATUS_OFFSET] & 0x01) == 0) {
		addr = ((args.tfRegister[IDE_SELECT_OFFSET] & 0x0f) << 24)
		     | ((args.tfRegister[  IDE_HCYL_OFFSET]       ) << 16)
		     | ((args.tfRegister[  IDE_LCYL_OFFSET]       ) <<  8)
		     | ((args.tfRegister[IDE_SECTOR_OFFSET]       ));
		addr++;	/* since the return value is (maxlba - 1), we add 1 */
	}
	return addr;
}

static unsigned long long idedisk_read_native_max_address_ext(ide_drive_t *drive)
{
	ide_task_t args;
	unsigned long long addr = 0;

	/* Create IDE/ATA command request structure */
	memset(&args, 0, sizeof(ide_task_t));

	args.tfRegister[IDE_SELECT_OFFSET]	= 0x40;
	args.tfRegister[IDE_COMMAND_OFFSET]	= WIN_READ_NATIVE_MAX_EXT;
	args.command_type			= IDE_DRIVE_TASK_NO_DATA;
	args.handler				= &task_no_data_intr;
        /* submit command request */
        ide_raw_taskfile(drive, &args, NULL);

	/* if OK, compute maximum address value */
	if ((args.tfRegister[IDE_STATUS_OFFSET] & 0x01) == 0) {
		u32 high = (args.hobRegister[IDE_HCYL_OFFSET] << 16) |
			   (args.hobRegister[IDE_LCYL_OFFSET] <<  8) |
			    args.hobRegister[IDE_SECTOR_OFFSET];
		u32 low  = ((args.tfRegister[IDE_HCYL_OFFSET])<<16) |
			   ((args.tfRegister[IDE_LCYL_OFFSET])<<8) |
			    (args.tfRegister[IDE_SECTOR_OFFSET]);
		addr = ((__u64)high << 24) | low;
		addr++;	/* since the return value is (maxlba - 1), we add 1 */
	}
	return addr;
}

/*
 * Sets maximum virtual LBA address of the drive.
 * Returns new maximum virtual LBA address (> 0) or 0 on failure.
 */
static unsigned long idedisk_set_max_address(ide_drive_t *drive, unsigned long addr_req)
{
	ide_task_t args;
	unsigned long addr_set = 0;
	
	addr_req--;
	/* Create IDE/ATA command request structure */
	memset(&args, 0, sizeof(ide_task_t));
	args.tfRegister[IDE_SECTOR_OFFSET]	= ((addr_req >>  0) & 0xff);
	args.tfRegister[IDE_LCYL_OFFSET]	= ((addr_req >>  8) & 0xff);
	args.tfRegister[IDE_HCYL_OFFSET]	= ((addr_req >> 16) & 0xff);
	args.tfRegister[IDE_SELECT_OFFSET]	= ((addr_req >> 24) & 0x0f) | 0x40;
	args.tfRegister[IDE_COMMAND_OFFSET]	= WIN_SET_MAX;
	args.command_type			= IDE_DRIVE_TASK_NO_DATA;
	args.handler				= &task_no_data_intr;
	/* submit command request */
	ide_raw_taskfile(drive, &args, NULL);
	/* if OK, read new maximum address value */
	if ((args.tfRegister[IDE_STATUS_OFFSET] & 0x01) == 0) {
		addr_set = ((args.tfRegister[IDE_SELECT_OFFSET] & 0x0f) << 24)
			 | ((args.tfRegister[  IDE_HCYL_OFFSET]       ) << 16)
			 | ((args.tfRegister[  IDE_LCYL_OFFSET]       ) <<  8)
			 | ((args.tfRegister[IDE_SECTOR_OFFSET]       ));
		addr_set++;
	}
	return addr_set;