OpenCloudOS-Kernel/drivers/ata/pata_legacy.c

1306 lines
34 KiB
C
Raw Normal View History

/*
* pata-legacy.c - Legacy port PATA/SATA controller driver.
* Copyright 2005/2006 Red Hat, all rights reserved.
*
* 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.
*
* You should have received a copy of the GNU General Public License
* along with this program; see the file COPYING. If not, write to
* the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
*
* An ATA driver for the legacy ATA ports.
*
* Data Sources:
* Opti 82C465/82C611 support: Data sheets at opti-inc.com
* HT6560 series:
* Promise 20230/20620:
* http://www.ryston.cz/petr/vlb/pdc20230b.html
* http://www.ryston.cz/petr/vlb/pdc20230c.html
* http://www.ryston.cz/petr/vlb/pdc20630.html
*
* Unsupported but docs exist:
* Appian/Adaptec AIC25VL01/Cirrus Logic PD7220
*
* This driver handles legacy (that is "ISA/VLB side") IDE ports found
* on PC class systems. There are three hybrid devices that are exceptions
* The Cyrix 5510/5520 where a pre SFF ATA device is on the bridge and
* the MPIIX where the tuning is PCI side but the IDE is "ISA side".
*
* Specific support is included for the ht6560a/ht6560b/opti82c611a/
* opti82c465mv/promise 20230c/20630/winbond83759A
*
* Use the autospeed and pio_mask options with:
* Appian ADI/2 aka CLPD7220 or AIC25VL01.
* Use the jumpers, autospeed and set pio_mask to the mode on the jumpers with
* Goldstar GM82C711, PIC-1288A-125, UMC 82C871F, Winbond W83759,
* Winbond W83759A, Promise PDC20230-B
*
* For now use autospeed and pio_mask as above with the W83759A. This may
* change.
*
*/
pata_legacy: wait for async probing The basic problem here that pata_legacy attaches the host, sees if it found any devices and detaches it if none were found. With async probing, it's not waiting until discovery is finished before deciding it has no devices and trying the detach leading to this warning: ata1: PATA max PIO4 cmd 0x1f0 ctl 0x3f6 irq 14 ------------[ cut here ]------------ WARNING: at drivers/ata/libata-core.c:6222 ata_host_detach+0x75/0x90() Modules linked in: Pid: 1, comm: swapper Not tainted 2.6.30-rc7 #1 Call Trace: [<c01fbb05>] ? ata_host_detach+0x75/0x90 [<c01fbb05>] ? ata_host_detach+0x75/0x90 [<c01139b5>] ? warn_slowpath_common+0x45/0x80 [<c01139fa>] ? warn_slowpath_null+0xa/0x10 [<c01fbb05>] ? ata_host_detach+0x75/0x90 [<c02f40e0>] ? legacy_init+0x44e/0x87f [<c02f3c92>] ? legacy_init+0x0/0x87f [<c0101021>] ? _stext+0x21/0x140 [<c01890ff>] ? proc_register+0x2f/0x190 [<c018938c>] ? create_proc_entry+0x5c/0xc0 [<c0135ebe>] ? register_irq_proc+0x6e/0x90 [<c02e6484>] ? kernel_init+0x6e/0xbf [<c02e6416>] ? kernel_init+0x0/0xbf [<c01031d7>] ? kernel_thread_helper+0x7/0x10 ---[ end trace ef1ee36e873ae3a0 ]--- Because it detaches before the probe is complete. One way to fix it would be to put an async_synchronize_full() before looking for devices, which this patch does. A better way might be to separate libata into its own domain and only wait for that. Reported-by: Mikael Pettersson <mikpe@it.uu.se> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2009-06-05 22:41:39 +08:00
#include <linux/async.h>
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/pci.h>
#include <linux/init.h>
#include <linux/blkdev.h>
#include <linux/delay.h>
#include <scsi/scsi_host.h>
#include <linux/ata.h>
#include <linux/libata.h>
#include <linux/platform_device.h>
#define DRV_NAME "pata_legacy"
#define DRV_VERSION "0.6.5"
#define NR_HOST 6
static int all;
module_param(all, int, 0444);
MODULE_PARM_DESC(all, "Grab all legacy port devices, even if PCI(0=off, 1=on)");
struct legacy_data {
unsigned long timing;
u8 clock[2];
u8 last;
int fast;
struct platform_device *platform_dev;
};
enum controller {
BIOS = 0,
SNOOP = 1,
PDC20230 = 2,
HT6560A = 3,
HT6560B = 4,
OPTI611A = 5,
OPTI46X = 6,
QDI6500 = 7,
QDI6580 = 8,
QDI6580DP = 9, /* Dual channel mode is different */
W83759A = 10,
UNKNOWN = -1
};
struct legacy_probe {
unsigned char *name;
unsigned long port;
unsigned int irq;
unsigned int slot;
enum controller type;
unsigned long private;
};
struct legacy_controller {
const char *name;
struct ata_port_operations *ops;
unsigned int pio_mask;
unsigned int flags;
unsigned int pflags;
int (*setup)(struct platform_device *, struct legacy_probe *probe,
struct legacy_data *data);
};
static int legacy_port[NR_HOST] = { 0x1f0, 0x170, 0x1e8, 0x168, 0x1e0, 0x160 };
static struct legacy_probe probe_list[NR_HOST];
static struct legacy_data legacy_data[NR_HOST];
static struct ata_host *legacy_host[NR_HOST];
static int nr_legacy_host;
static int probe_all; /* Set to check all ISA port ranges */
static int ht6560a; /* HT 6560A on primary 1, second 2, both 3 */
static int ht6560b; /* HT 6560A on primary 1, second 2, both 3 */
static int opti82c611a; /* Opti82c611A on primary 1, sec 2, both 3 */
static int opti82c46x; /* Opti 82c465MV present(pri/sec autodetect) */
static int qdi; /* Set to probe QDI controllers */
static int winbond; /* Set to probe Winbond controllers,
give I/O port if non standard */
static int autospeed; /* Chip present which snoops speed changes */
static int pio_mask = ATA_PIO4; /* PIO range for autospeed devices */
static int iordy_mask = 0xFFFFFFFF; /* Use iordy if available */
/**
* legacy_probe_add - Add interface to probe list
* @port: Controller port
* @irq: IRQ number
* @type: Controller type
* @private: Controller specific info
*
* Add an entry into the probe list for ATA controllers. This is used
* to add the default ISA slots and then to build up the table
* further according to other ISA/VLB/Weird device scans
*
* An I/O port list is used to keep ordering stable and sane, as we
* don't have any good way to talk about ordering otherwise
*/
static int legacy_probe_add(unsigned long port, unsigned int irq,
enum controller type, unsigned long private)
{
struct legacy_probe *lp = &probe_list[0];
int i;
struct legacy_probe *free = NULL;
for (i = 0; i < NR_HOST; i++) {
if (lp->port == 0 && free == NULL)
free = lp;
/* Matching port, or the correct slot for ordering */
if (lp->port == port || legacy_port[i] == port) {
free = lp;
break;
}
lp++;
}
if (free == NULL) {
printk(KERN_ERR "pata_legacy: Too many interfaces.\n");
return -1;
}
/* Fill in the entry for later probing */
free->port = port;
free->irq = irq;
free->type = type;
free->private = private;
return 0;
}
/**
* legacy_set_mode - mode setting
* @link: IDE link
* @unused: Device that failed when error is returned
*
* Use a non standard set_mode function. We don't want to be tuned.
*
* The BIOS configured everything. Our job is not to fiddle. Just use
* whatever PIO the hardware is using and leave it at that. When we
* get some kind of nice user driven API for control then we can
* expand on this as per hdparm in the base kernel.
*/
static int legacy_set_mode(struct ata_link *link, struct ata_device **unused)
{
struct ata_device *dev;
ata_for_each_dev(dev, link, ENABLED) {
ata_dev_printk(dev, KERN_INFO, "configured for PIO\n");
dev->pio_mode = XFER_PIO_0;
dev->xfer_mode = XFER_PIO_0;
dev->xfer_shift = ATA_SHIFT_PIO;
dev->flags |= ATA_DFLAG_PIO;
}
return 0;
}
static struct scsi_host_template legacy_sht = {
ATA_PIO_SHT(DRV_NAME),
};
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
static const struct ata_port_operations legacy_base_port_ops = {
.inherits = &ata_sff_port_ops,
.cable_detect = ata_cable_40wire,
};
/*
* These ops are used if the user indicates the hardware
* snoops the commands to decide on the mode and handles the
* mode selection "magically" itself. Several legacy controllers
* do this. The mode range can be set if it is not 0x1F by setting
* pio_mask as well.
*/
static struct ata_port_operations simple_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.sff_data_xfer = ata_sff_data_xfer_noirq,
};
static struct ata_port_operations legacy_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.sff_data_xfer = ata_sff_data_xfer_noirq,
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.set_mode = legacy_set_mode,
};
/*
* Promise 20230C and 20620 support
*
* This controller supports PIO0 to PIO2. We set PIO timings
* conservatively to allow for 50MHz Vesa Local Bus. The 20620 DMA
* support is weird being DMA to controller and PIO'd to the host
* and not supported.
*/
static void pdc20230_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
int tries = 5;
int pio = adev->pio_mode - XFER_PIO_0;
u8 rt;
unsigned long flags;
/* Safe as UP only. Force I/Os to occur together */
local_irq_save(flags);
/* Unlock the control interface */
do {
inb(0x1F5);
outb(inb(0x1F2) | 0x80, 0x1F2);
inb(0x1F2);
inb(0x3F6);
inb(0x3F6);
inb(0x1F2);
inb(0x1F2);
}
while ((inb(0x1F2) & 0x80) && --tries);
local_irq_restore(flags);
outb(inb(0x1F4) & 0x07, 0x1F4);
rt = inb(0x1F3);
rt &= 0x07 << (3 * adev->devno);
if (pio)
rt |= (1 + 3 * pio) << (3 * adev->devno);
udelay(100);
outb(inb(0x1F2) | 0x01, 0x1F2);
udelay(100);
inb(0x1F5);
}
static unsigned int pdc_data_xfer_vlb(struct ata_device *dev,
unsigned char *buf, unsigned int buflen, int rw)
{
int slop = buflen & 3;
struct ata_port *ap = dev->link->ap;
/* 32bit I/O capable *and* we need to write a whole number of dwords */
if (ata_id_has_dword_io(dev->id) && (slop == 0 || slop == 3)
&& (ap->pflags & ATA_PFLAG_PIO32)) {
unsigned long flags;
local_irq_save(flags);
/* Perform the 32bit I/O synchronization sequence */
ioread8(ap->ioaddr.nsect_addr);
ioread8(ap->ioaddr.nsect_addr);
ioread8(ap->ioaddr.nsect_addr);
/* Now the data */
if (rw == READ)
ioread32_rep(ap->ioaddr.data_addr, buf, buflen >> 2);
else
iowrite32_rep(ap->ioaddr.data_addr, buf, buflen >> 2);
if (unlikely(slop)) {
__le32 pad;
if (rw == READ) {
pad = cpu_to_le32(ioread32(ap->ioaddr.data_addr));
memcpy(buf + buflen - slop, &pad, slop);
} else {
memcpy(&pad, buf + buflen - slop, slop);
iowrite32(le32_to_cpu(pad), ap->ioaddr.data_addr);
}
buflen += 4 - slop;
}
local_irq_restore(flags);
} else
buflen = ata_sff_data_xfer_noirq(dev, buf, buflen, rw);
return buflen;
}
static struct ata_port_operations pdc20230_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.set_piomode = pdc20230_set_piomode,
.sff_data_xfer = pdc_data_xfer_vlb,
};
/*
* Holtek 6560A support
*
* This controller supports PIO0 to PIO2 (no IORDY even though higher
* timings can be loaded).
*/
static void ht6560a_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
u8 active, recover;
struct ata_timing t;
/* Get the timing data in cycles. For now play safe at 50Mhz */
ata_timing_compute(adev, adev->pio_mode, &t, 20000, 1000);
active = clamp_val(t.active, 2, 15);
recover = clamp_val(t.recover, 4, 15);
inb(0x3E6);
inb(0x3E6);
inb(0x3E6);
inb(0x3E6);
iowrite8(recover << 4 | active, ap->ioaddr.device_addr);
ioread8(ap->ioaddr.status_addr);
}
static struct ata_port_operations ht6560a_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.set_piomode = ht6560a_set_piomode,
};
/*
* Holtek 6560B support
*
* This controller supports PIO0 to PIO4. We honour the BIOS/jumper FIFO
* setting unless we see an ATAPI device in which case we force it off.
*
* FIXME: need to implement 2nd channel support.
*/
static void ht6560b_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
u8 active, recover;
struct ata_timing t;
/* Get the timing data in cycles. For now play safe at 50Mhz */
ata_timing_compute(adev, adev->pio_mode, &t, 20000, 1000);
active = clamp_val(t.active, 2, 15);
recover = clamp_val(t.recover, 2, 16);
recover &= 0x15;
inb(0x3E6);
inb(0x3E6);
inb(0x3E6);
inb(0x3E6);
iowrite8(recover << 4 | active, ap->ioaddr.device_addr);
if (adev->class != ATA_DEV_ATA) {
u8 rconf = inb(0x3E6);
if (rconf & 0x24) {
rconf &= ~0x24;
outb(rconf, 0x3E6);
}
}
ioread8(ap->ioaddr.status_addr);
}
static struct ata_port_operations ht6560b_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.set_piomode = ht6560b_set_piomode,
};
/*
* Opti core chipset helpers
*/
/**
* opti_syscfg - read OPTI chipset configuration
* @reg: Configuration register to read
*
* Returns the value of an OPTI system board configuration register.
*/
static u8 opti_syscfg(u8 reg)
{
unsigned long flags;
u8 r;
/* Uniprocessor chipset and must force cycles adjancent */
local_irq_save(flags);
outb(reg, 0x22);
r = inb(0x24);
local_irq_restore(flags);
return r;
}
/*
* Opti 82C611A
*
* This controller supports PIO0 to PIO3.
*/
static void opti82c611a_set_piomode(struct ata_port *ap,
struct ata_device *adev)
{
u8 active, recover, setup;
struct ata_timing t;
struct ata_device *pair = ata_dev_pair(adev);
int clock;
int khz[4] = { 50000, 40000, 33000, 25000 };
u8 rc;
/* Enter configuration mode */
ioread16(ap->ioaddr.error_addr);
ioread16(ap->ioaddr.error_addr);
iowrite8(3, ap->ioaddr.nsect_addr);
/* Read VLB clock strapping */
clock = 1000000000 / khz[ioread8(ap->ioaddr.lbah_addr) & 0x03];
/* Get the timing data in cycles */
ata_timing_compute(adev, adev->pio_mode, &t, clock, 1000);
/* Setup timing is shared */
if (pair) {
struct ata_timing tp;
ata_timing_compute(pair, pair->pio_mode, &tp, clock, 1000);
ata_timing_merge(&t, &tp, &t, ATA_TIMING_SETUP);
}
active = clamp_val(t.active, 2, 17) - 2;
recover = clamp_val(t.recover, 1, 16) - 1;
setup = clamp_val(t.setup, 1, 4) - 1;
/* Select the right timing bank for write timing */
rc = ioread8(ap->ioaddr.lbal_addr);
rc &= 0x7F;
rc |= (adev->devno << 7);
iowrite8(rc, ap->ioaddr.lbal_addr);
/* Write the timings */
iowrite8(active << 4 | recover, ap->ioaddr.error_addr);
/* Select the right bank for read timings, also
load the shared timings for address */
rc = ioread8(ap->ioaddr.device_addr);
rc &= 0xC0;
rc |= adev->devno; /* Index select */
rc |= (setup << 4) | 0x04;
iowrite8(rc, ap->ioaddr.device_addr);
/* Load the read timings */
iowrite8(active << 4 | recover, ap->ioaddr.data_addr);
/* Ensure the timing register mode is right */
rc = ioread8(ap->ioaddr.lbal_addr);
rc &= 0x73;
rc |= 0x84;
iowrite8(rc, ap->ioaddr.lbal_addr);
/* Exit command mode */
iowrite8(0x83, ap->ioaddr.nsect_addr);
}
static struct ata_port_operations opti82c611a_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.set_piomode = opti82c611a_set_piomode,
};
/*
* Opti 82C465MV
*
* This controller supports PIO0 to PIO3. Unlike the 611A the MVB
* version is dual channel but doesn't have a lot of unique registers.
*/
static void opti82c46x_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
u8 active, recover, setup;
struct ata_timing t;
struct ata_device *pair = ata_dev_pair(adev);
int clock;
int khz[4] = { 50000, 40000, 33000, 25000 };
u8 rc;
u8 sysclk;
/* Get the clock */
sysclk = opti_syscfg(0xAC) & 0xC0; /* BIOS set */
/* Enter configuration mode */
ioread16(ap->ioaddr.error_addr);
ioread16(ap->ioaddr.error_addr);
iowrite8(3, ap->ioaddr.nsect_addr);
/* Read VLB clock strapping */
clock = 1000000000 / khz[sysclk];
/* Get the timing data in cycles */
ata_timing_compute(adev, adev->pio_mode, &t, clock, 1000);
/* Setup timing is shared */
if (pair) {
struct ata_timing tp;
ata_timing_compute(pair, pair->pio_mode, &tp, clock, 1000);
ata_timing_merge(&t, &tp, &t, ATA_TIMING_SETUP);
}
active = clamp_val(t.active, 2, 17) - 2;
recover = clamp_val(t.recover, 1, 16) - 1;
setup = clamp_val(t.setup, 1, 4) - 1;
/* Select the right timing bank for write timing */
rc = ioread8(ap->ioaddr.lbal_addr);
rc &= 0x7F;
rc |= (adev->devno << 7);
iowrite8(rc, ap->ioaddr.lbal_addr);
/* Write the timings */
iowrite8(active << 4 | recover, ap->ioaddr.error_addr);
/* Select the right bank for read timings, also
load the shared timings for address */
rc = ioread8(ap->ioaddr.device_addr);
rc &= 0xC0;
rc |= adev->devno; /* Index select */
rc |= (setup << 4) | 0x04;
iowrite8(rc, ap->ioaddr.device_addr);
/* Load the read timings */
iowrite8(active << 4 | recover, ap->ioaddr.data_addr);
/* Ensure the timing register mode is right */
rc = ioread8(ap->ioaddr.lbal_addr);
rc &= 0x73;
rc |= 0x84;
iowrite8(rc, ap->ioaddr.lbal_addr);
/* Exit command mode */
iowrite8(0x83, ap->ioaddr.nsect_addr);
/* We need to know this for quad device on the MVB */
ap->host->private_data = ap;
}
/**
* opt82c465mv_qc_issue - command issue
* @qc: command pending
*
* Called when the libata layer is about to issue a command. We wrap
* this interface so that we can load the correct ATA timings. The
* MVB has a single set of timing registers and these are shared
* across channels. As there are two registers we really ought to
* track the last two used values as a sort of register window. For
* now we just reload on a channel switch. On the single channel
* setup this condition never fires so we do nothing extra.
*
* FIXME: dual channel needs ->serialize support
*/
static unsigned int opti82c46x_qc_issue(struct ata_queued_cmd *qc)
{
struct ata_port *ap = qc->ap;
struct ata_device *adev = qc->dev;
/* If timings are set and for the wrong channel (2nd test is
due to a libata shortcoming and will eventually go I hope) */
if (ap->host->private_data != ap->host
&& ap->host->private_data != NULL)
opti82c46x_set_piomode(ap, adev);
return ata_sff_qc_issue(qc);
}
static struct ata_port_operations opti82c46x_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.set_piomode = opti82c46x_set_piomode,
.qc_issue = opti82c46x_qc_issue,
};
static void qdi6500_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
struct ata_timing t;
struct legacy_data *ld_qdi = ap->host->private_data;
int active, recovery;
u8 timing;
/* Get the timing data in cycles */
ata_timing_compute(adev, adev->pio_mode, &t, 30303, 1000);
if (ld_qdi->fast) {
active = 8 - clamp_val(t.active, 1, 8);
recovery = 18 - clamp_val(t.recover, 3, 18);
} else {
active = 9 - clamp_val(t.active, 2, 9);
recovery = 15 - clamp_val(t.recover, 0, 15);
}
timing = (recovery << 4) | active | 0x08;
ld_qdi->clock[adev->devno] = timing;
outb(timing, ld_qdi->timing);
}
/**
* qdi6580dp_set_piomode - PIO setup for dual channel
* @ap: Port
* @adev: Device
*
* In dual channel mode the 6580 has one clock per channel and we have
* to software clockswitch in qc_issue.
*/
static void qdi6580dp_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
struct ata_timing t;
struct legacy_data *ld_qdi = ap->host->private_data;
int active, recovery;
u8 timing;
/* Get the timing data in cycles */
ata_timing_compute(adev, adev->pio_mode, &t, 30303, 1000);
if (ld_qdi->fast) {
active = 8 - clamp_val(t.active, 1, 8);
recovery = 18 - clamp_val(t.recover, 3, 18);
} else {
active = 9 - clamp_val(t.active, 2, 9);
recovery = 15 - clamp_val(t.recover, 0, 15);
}
timing = (recovery << 4) | active | 0x08;
ld_qdi->clock[adev->devno] = timing;
outb(timing, ld_qdi->timing + 2 * ap->port_no);
/* Clear the FIFO */
if (adev->class != ATA_DEV_ATA)
outb(0x5F, ld_qdi->timing + 3);
}
/**
* qdi6580_set_piomode - PIO setup for single channel
* @ap: Port
* @adev: Device
*
* In single channel mode the 6580 has one clock per device and we can
* avoid the requirement to clock switch. We also have to load the timing
* into the right clock according to whether we are master or slave.
*/
static void qdi6580_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
struct ata_timing t;
struct legacy_data *ld_qdi = ap->host->private_data;
int active, recovery;
u8 timing;
/* Get the timing data in cycles */
ata_timing_compute(adev, adev->pio_mode, &t, 30303, 1000);
if (ld_qdi->fast) {
active = 8 - clamp_val(t.active, 1, 8);
recovery = 18 - clamp_val(t.recover, 3, 18);
} else {
active = 9 - clamp_val(t.active, 2, 9);
recovery = 15 - clamp_val(t.recover, 0, 15);
}
timing = (recovery << 4) | active | 0x08;
ld_qdi->clock[adev->devno] = timing;
outb(timing, ld_qdi->timing + 2 * adev->devno);
/* Clear the FIFO */
if (adev->class != ATA_DEV_ATA)
outb(0x5F, ld_qdi->timing + 3);
}
/**
* qdi_qc_issue - command issue
* @qc: command pending
*
* Called when the libata layer is about to issue a command. We wrap
* this interface so that we can load the correct ATA timings.
*/
static unsigned int qdi_qc_issue(struct ata_queued_cmd *qc)
{
struct ata_port *ap = qc->ap;
struct ata_device *adev = qc->dev;
struct legacy_data *ld_qdi = ap->host->private_data;
if (ld_qdi->clock[adev->devno] != ld_qdi->last) {
if (adev->pio_mode) {
ld_qdi->last = ld_qdi->clock[adev->devno];
outb(ld_qdi->clock[adev->devno], ld_qdi->timing +
2 * ap->port_no);
}
}
return ata_sff_qc_issue(qc);
}
static unsigned int vlb32_data_xfer(struct ata_device *adev, unsigned char *buf,
unsigned int buflen, int rw)
{
struct ata_port *ap = adev->link->ap;
int slop = buflen & 3;
if (ata_id_has_dword_io(adev->id) && (slop == 0 || slop == 3)
&& (ap->pflags & ATA_PFLAG_PIO32)) {
if (rw == WRITE)
iowrite32_rep(ap->ioaddr.data_addr, buf, buflen >> 2);
else
ioread32_rep(ap->ioaddr.data_addr, buf, buflen >> 2);
if (unlikely(slop)) {
__le32 pad;
if (rw == WRITE) {
memcpy(&pad, buf + buflen - slop, slop);
iowrite32(le32_to_cpu(pad), ap->ioaddr.data_addr);
} else {
pad = cpu_to_le32(ioread32(ap->ioaddr.data_addr));
memcpy(buf + buflen - slop, &pad, slop);
}
}
return (buflen + 3) & ~3;
} else
return ata_sff_data_xfer(adev, buf, buflen, rw);
}
static int qdi_port(struct platform_device *dev,
struct legacy_probe *lp, struct legacy_data *ld)
{
if (devm_request_region(&dev->dev, lp->private, 4, "qdi") == NULL)
return -EBUSY;
ld->timing = lp->private;
return 0;
}
static struct ata_port_operations qdi6500_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.set_piomode = qdi6500_set_piomode,
.qc_issue = qdi_qc_issue,
.sff_data_xfer = vlb32_data_xfer,
};
static struct ata_port_operations qdi6580_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.set_piomode = qdi6580_set_piomode,
.sff_data_xfer = vlb32_data_xfer,
};
static struct ata_port_operations qdi6580dp_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.set_piomode = qdi6580dp_set_piomode,
.qc_issue = qdi_qc_issue,
.sff_data_xfer = vlb32_data_xfer,
};
static DEFINE_SPINLOCK(winbond_lock);
static void winbond_writecfg(unsigned long port, u8 reg, u8 val)
{
unsigned long flags;
spin_lock_irqsave(&winbond_lock, flags);
outb(reg, port + 0x01);
outb(val, port + 0x02);
spin_unlock_irqrestore(&winbond_lock, flags);
}
static u8 winbond_readcfg(unsigned long port, u8 reg)
{
u8 val;
unsigned long flags;
spin_lock_irqsave(&winbond_lock, flags);
outb(reg, port + 0x01);
val = inb(port + 0x02);
spin_unlock_irqrestore(&winbond_lock, flags);
return val;
}
static void winbond_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
struct ata_timing t;
struct legacy_data *ld_winbond = ap->host->private_data;
int active, recovery;
u8 reg;
int timing = 0x88 + (ap->port_no * 4) + (adev->devno * 2);
reg = winbond_readcfg(ld_winbond->timing, 0x81);
/* Get the timing data in cycles */
if (reg & 0x40) /* Fast VLB bus, assume 50MHz */
ata_timing_compute(adev, adev->pio_mode, &t, 20000, 1000);
else
ata_timing_compute(adev, adev->pio_mode, &t, 30303, 1000);
active = (clamp_val(t.active, 3, 17) - 1) & 0x0F;
recovery = (clamp_val(t.recover, 1, 15) + 1) & 0x0F;
timing = (active << 4) | recovery;
winbond_writecfg(ld_winbond->timing, timing, reg);
/* Load the setup timing */
reg = 0x35;
if (adev->class != ATA_DEV_ATA)
reg |= 0x08; /* FIFO off */
if (!ata_pio_need_iordy(adev))
reg |= 0x02; /* IORDY off */
reg |= (clamp_val(t.setup, 0, 3) << 6);
winbond_writecfg(ld_winbond->timing, timing + 1, reg);
}
static int winbond_port(struct platform_device *dev,
struct legacy_probe *lp, struct legacy_data *ld)
{
if (devm_request_region(&dev->dev, lp->private, 4, "winbond") == NULL)
return -EBUSY;
ld->timing = lp->private;
return 0;
}
static struct ata_port_operations winbond_port_ops = {
libata: implement and use ops inheritance libata lets low level drivers build ata_port_operations table and register it with libata core layer. This allows low level drivers high level of flexibility but also burdens them with lots of boilerplate entries. This becomes worse for drivers which support related similar controllers which differ slightly. They share most of the operations except for a few. However, the driver still needs to list all operations for each variant. This results in large number of duplicate entries, which is not only inefficient but also error-prone as it becomes very difficult to tell what the actual differences are. This duplicate boilerplates all over the low level drivers also make updating the core layer exteremely difficult and error-prone. When compounded with multi-branched development model, it ends up accumulating inconsistencies over time. Some of those inconsistencies cause immediate problems and fixed. Others just remain there dormant making maintenance increasingly difficult. To rectify the problem, this patch implements ata_port_operations inheritance. To allow LLDs to easily re-use their own ops tables overriding only specific methods, this patch implements poor man's class inheritance. An ops table has ->inherits field which can be set to any ops table as long as it doesn't create a loop. When the host is started, the inheritance chain is followed and any operation which isn't specified is taken from the nearest ancestor which has it specified. This operation is called finalization and done only once per an ops table and the LLD doesn't have to do anything special about it other than making the ops table non-const such that libata can update it. libata provides four base ops tables lower drivers can inherit from - base, sata, pmp, sff and bmdma. To avoid overriding these ops accidentaly, these ops are declared const and LLDs should always inherit these instead of using them directly. After finalization, all the ops table are identical before and after the patch except for setting .irq_handler to ata_interrupt in drivers which didn't use to. The .irq_handler doesn't have any actual effect and the field will soon be removed by later patch. * sata_sx4 is still using old style EH and currently doesn't take advantage of ops inheritance. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-03-25 11:22:49 +08:00
.inherits = &legacy_base_port_ops,
.set_piomode = winbond_set_piomode,
.sff_data_xfer = vlb32_data_xfer,
};
static struct legacy_controller controllers[] = {
{"BIOS", &legacy_port_ops, 0x1F,
ATA_FLAG_NO_IORDY, 0, NULL },
{"Snooping", &simple_port_ops, 0x1F,
0, 0, NULL },
{"PDC20230", &pdc20230_port_ops, 0x7,
ATA_FLAG_NO_IORDY,
ATA_PFLAG_PIO32 | ATA_PFLAG_PIO32CHANGE, NULL },
{"HT6560A", &ht6560a_port_ops, 0x07,
ATA_FLAG_NO_IORDY, 0, NULL },
{"HT6560B", &ht6560b_port_ops, 0x1F,
ATA_FLAG_NO_IORDY, 0, NULL },
{"OPTI82C611A", &opti82c611a_port_ops, 0x0F,
0, 0, NULL },
{"OPTI82C46X", &opti82c46x_port_ops, 0x0F,
0, 0, NULL },
{"QDI6500", &qdi6500_port_ops, 0x07,
ATA_FLAG_NO_IORDY,
ATA_PFLAG_PIO32 | ATA_PFLAG_PIO32CHANGE, qdi_port },
{"QDI6580", &qdi6580_port_ops, 0x1F,
0, ATA_PFLAG_PIO32 | ATA_PFLAG_PIO32CHANGE, qdi_port },
{"QDI6580DP", &qdi6580dp_port_ops, 0x1F,
0, ATA_PFLAG_PIO32 | ATA_PFLAG_PIO32CHANGE, qdi_port },
{"W83759A", &winbond_port_ops, 0x1F,
0, ATA_PFLAG_PIO32 | ATA_PFLAG_PIO32CHANGE,
winbond_port }
};
/**
* probe_chip_type - Discover controller
* @probe: Probe entry to check
*
* Probe an ATA port and identify the type of controller. We don't
* check if the controller appears to be driveless at this point.
*/
static __init int probe_chip_type(struct legacy_probe *probe)
{
int mask = 1 << probe->slot;
if (winbond && (probe->port == 0x1F0 || probe->port == 0x170)) {
u8 reg = winbond_readcfg(winbond, 0x81);
reg |= 0x80; /* jumpered mode off */
winbond_writecfg(winbond, 0x81, reg);
reg = winbond_readcfg(winbond, 0x83);
reg |= 0xF0; /* local control */
winbond_writecfg(winbond, 0x83, reg);
reg = winbond_readcfg(winbond, 0x85);
reg |= 0xF0; /* programmable timing */
winbond_writecfg(winbond, 0x85, reg);
reg = winbond_readcfg(winbond, 0x81);
if (reg & mask)
return W83759A;
}
if (probe->port == 0x1F0) {
unsigned long flags;
local_irq_save(flags);
/* Probes */
outb(inb(0x1F2) | 0x80, 0x1F2);
inb(0x1F5);
inb(0x1F2);
inb(0x3F6);
inb(0x3F6);
inb(0x1F2);
inb(0x1F2);
if ((inb(0x1F2) & 0x80) == 0) {
/* PDC20230c or 20630 ? */
printk(KERN_INFO "PDC20230-C/20630 VLB ATA controller"
" detected.\n");
udelay(100);
inb(0x1F5);
local_irq_restore(flags);
return PDC20230;
} else {
outb(0x55, 0x1F2);
inb(0x1F2);
inb(0x1F2);
if (inb(0x1F2) == 0x00)
printk(KERN_INFO "PDC20230-B VLB ATA "
"controller detected.\n");
local_irq_restore(flags);
return BIOS;
}
local_irq_restore(flags);
}
if (ht6560a & mask)
return HT6560A;
if (ht6560b & mask)
return HT6560B;
if (opti82c611a & mask)
return OPTI611A;
if (opti82c46x & mask)
return OPTI46X;
if (autospeed & mask)
return SNOOP;
return BIOS;
}
/**
* legacy_init_one - attach a legacy interface
* @pl: probe record
*
* Register an ISA bus IDE interface. Such interfaces are PIO and we
* assume do not support IRQ sharing.
*/
static __init int legacy_init_one(struct legacy_probe *probe)
{
struct legacy_controller *controller = &controllers[probe->type];
int pio_modes = controller->pio_mask;
unsigned long io = probe->port;
u32 mask = (1 << probe->slot);
struct ata_port_operations *ops = controller->ops;
struct legacy_data *ld = &legacy_data[probe->slot];
struct ata_host *host = NULL;
struct ata_port *ap;
struct platform_device *pdev;
struct ata_device *dev;
void __iomem *io_addr, *ctrl_addr;
u32 iordy = (iordy_mask & mask) ? 0: ATA_FLAG_NO_IORDY;
int ret;
iordy |= controller->flags;
pdev = platform_device_register_simple(DRV_NAME, probe->slot, NULL, 0);
if (IS_ERR(pdev))
return PTR_ERR(pdev);
ret = -EBUSY;
if (devm_request_region(&pdev->dev, io, 8, "pata_legacy") == NULL ||
devm_request_region(&pdev->dev, io + 0x0206, 1,
"pata_legacy") == NULL)
goto fail;
ret = -ENOMEM;
io_addr = devm_ioport_map(&pdev->dev, io, 8);
ctrl_addr = devm_ioport_map(&pdev->dev, io + 0x0206, 1);
if (!io_addr || !ctrl_addr)
goto fail;
if (controller->setup)
if (controller->setup(pdev, probe, ld) < 0)
goto fail;
host = ata_host_alloc(&pdev->dev, 1);
if (!host)
goto fail;
ap = host->ports[0];
ap->ops = ops;
ap->pio_mask = pio_modes;
ap->flags |= ATA_FLAG_SLAVE_POSS | iordy;
ap->pflags |= controller->pflags;
ap->ioaddr.cmd_addr = io_addr;
ap->ioaddr.altstatus_addr = ctrl_addr;
ap->ioaddr.ctl_addr = ctrl_addr;
ata_sff_std_ports(&ap->ioaddr);
ap->host->private_data = ld;
ata_port_desc(ap, "cmd 0x%lx ctl 0x%lx", io, io + 0x0206);
ret = ata_host_activate(host, probe->irq, ata_sff_interrupt, 0,
&legacy_sht);
if (ret)
goto fail;
pata_legacy: wait for async probing The basic problem here that pata_legacy attaches the host, sees if it found any devices and detaches it if none were found. With async probing, it's not waiting until discovery is finished before deciding it has no devices and trying the detach leading to this warning: ata1: PATA max PIO4 cmd 0x1f0 ctl 0x3f6 irq 14 ------------[ cut here ]------------ WARNING: at drivers/ata/libata-core.c:6222 ata_host_detach+0x75/0x90() Modules linked in: Pid: 1, comm: swapper Not tainted 2.6.30-rc7 #1 Call Trace: [<c01fbb05>] ? ata_host_detach+0x75/0x90 [<c01fbb05>] ? ata_host_detach+0x75/0x90 [<c01139b5>] ? warn_slowpath_common+0x45/0x80 [<c01139fa>] ? warn_slowpath_null+0xa/0x10 [<c01fbb05>] ? ata_host_detach+0x75/0x90 [<c02f40e0>] ? legacy_init+0x44e/0x87f [<c02f3c92>] ? legacy_init+0x0/0x87f [<c0101021>] ? _stext+0x21/0x140 [<c01890ff>] ? proc_register+0x2f/0x190 [<c018938c>] ? create_proc_entry+0x5c/0xc0 [<c0135ebe>] ? register_irq_proc+0x6e/0x90 [<c02e6484>] ? kernel_init+0x6e/0xbf [<c02e6416>] ? kernel_init+0x0/0xbf [<c01031d7>] ? kernel_thread_helper+0x7/0x10 ---[ end trace ef1ee36e873ae3a0 ]--- Because it detaches before the probe is complete. One way to fix it would be to put an async_synchronize_full() before looking for devices, which this patch does. A better way might be to separate libata into its own domain and only wait for that. Reported-by: Mikael Pettersson <mikpe@it.uu.se> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2009-06-05 22:41:39 +08:00
async_synchronize_full();
ld->platform_dev = pdev;
/* Nothing found means we drop the port as its probably not there */
ret = -ENODEV;
ata_for_each_dev(dev, &ap->link, ALL) {
if (!ata_dev_absent(dev)) {
legacy_host[probe->slot] = host;
ld->platform_dev = pdev;
return 0;
}
}
ata_host_detach(host);
fail:
platform_device_unregister(pdev);
return ret;
}
/**
* legacy_check_special_cases - ATA special cases
* @p: PCI device to check
* @master: set this if we find an ATA master
* @master: set this if we find an ATA secondary
*
* A small number of vendors implemented early PCI ATA interfaces
* on bridge logic without the ATA interface being PCI visible.
* Where we have a matching PCI driver we must skip the relevant
* device here. If we don't know about it then the legacy driver
* is the right driver anyway.
*/
static void __init legacy_check_special_cases(struct pci_dev *p, int *primary,
int *secondary)
{
/* Cyrix CS5510 pre SFF MWDMA ATA on the bridge */
if (p->vendor == 0x1078 && p->device == 0x0000) {
*primary = *secondary = 1;
return;
}
/* Cyrix CS5520 pre SFF MWDMA ATA on the bridge */
if (p->vendor == 0x1078 && p->device == 0x0002) {
*primary = *secondary = 1;
return;
}
/* Intel MPIIX - PIO ATA on non PCI side of bridge */
if (p->vendor == 0x8086 && p->device == 0x1234) {
u16 r;
pci_read_config_word(p, 0x6C, &r);
if (r & 0x8000) {
/* ATA port enabled */
if (r & 0x4000)
*secondary = 1;
else
*primary = 1;
}
return;
}
}
static __init void probe_opti_vlb(void)
{
/* If an OPTI 82C46X is present find out where the channels are */
static const char *optis[4] = {
"3/463MV", "5MV",
"5MVA", "5MVB"
};
u8 chans = 1;
u8 ctrl = (opti_syscfg(0x30) & 0xC0) >> 6;
opti82c46x = 3; /* Assume master and slave first */
printk(KERN_INFO DRV_NAME ": Opti 82C46%s chipset support.\n",
optis[ctrl]);
if (ctrl == 3)
chans = (opti_syscfg(0x3F) & 0x20) ? 2 : 1;
ctrl = opti_syscfg(0xAC);
/* Check enabled and this port is the 465MV port. On the
MVB we may have two channels */
if (ctrl & 8) {
if (chans == 2) {
legacy_probe_add(0x1F0, 14, OPTI46X, 0);
legacy_probe_add(0x170, 15, OPTI46X, 0);
}
if (ctrl & 4)
legacy_probe_add(0x170, 15, OPTI46X, 0);
else
legacy_probe_add(0x1F0, 14, OPTI46X, 0);
} else
legacy_probe_add(0x1F0, 14, OPTI46X, 0);
}
static __init void qdi65_identify_port(u8 r, u8 res, unsigned long port)
{
static const unsigned long ide_port[2] = { 0x170, 0x1F0 };
/* Check card type */
if ((r & 0xF0) == 0xC0) {
/* QD6500: single channel */
if (r & 8)
/* Disabled ? */
return;
legacy_probe_add(ide_port[r & 0x01], 14 + (r & 0x01),
QDI6500, port);
}
if (((r & 0xF0) == 0xA0) || (r & 0xF0) == 0x50) {
/* QD6580: dual channel */
if (!request_region(port + 2 , 2, "pata_qdi")) {
release_region(port, 2);
return;
}
res = inb(port + 3);
/* Single channel mode ? */
if (res & 1)
legacy_probe_add(ide_port[r & 0x01], 14 + (r & 0x01),
QDI6580, port);
else { /* Dual channel mode */
legacy_probe_add(0x1F0, 14, QDI6580DP, port);
/* port + 0x02, r & 0x04 */
legacy_probe_add(0x170, 15, QDI6580DP, port + 2);
}
release_region(port + 2, 2);
}
}
static __init void probe_qdi_vlb(void)
{
unsigned long flags;
static const unsigned long qd_port[2] = { 0x30, 0xB0 };
int i;
/*
* Check each possible QD65xx base address
*/
for (i = 0; i < 2; i++) {
unsigned long port = qd_port[i];
u8 r, res;
if (request_region(port, 2, "pata_qdi")) {
/* Check for a card */
local_irq_save(flags);
/* I have no h/w that needs this delay but it
is present in the historic code */
r = inb(port);
udelay(1);
outb(0x19, port);
udelay(1);
res = inb(port);
udelay(1);
outb(r, port);
udelay(1);
local_irq_restore(flags);
/* Fail */
if (res == 0x19) {
release_region(port, 2);
continue;
}
/* Passes the presence test */
r = inb(port + 1);
udelay(1);
/* Check port agrees with port set */
if ((r & 2) >> 1 == i)
qdi65_identify_port(r, res, port);
release_region(port, 2);
}
}
}
/**
* legacy_init - attach legacy interfaces
*
* Attach legacy IDE interfaces by scanning the usual IRQ/port suspects.
* Right now we do not scan the ide0 and ide1 address but should do so
* for non PCI systems or systems with no PCI IDE legacy mode devices.
* If you fix that note there are special cases to consider like VLB
* drivers and CS5510/20.
*/
static __init int legacy_init(void)
{
int i;
int ct = 0;
int primary = 0;
int secondary = 0;
int pci_present = 0;
struct legacy_probe *pl = &probe_list[0];
int slot = 0;
struct pci_dev *p = NULL;
for_each_pci_dev(p) {
int r;
/* Check for any overlap of the system ATA mappings. Native
mode controllers stuck on these addresses or some devices
in 'raid' mode won't be found by the storage class test */
for (r = 0; r < 6; r++) {
if (pci_resource_start(p, r) == 0x1f0)
primary = 1;
if (pci_resource_start(p, r) == 0x170)
secondary = 1;
}
/* Check for special cases */
legacy_check_special_cases(p, &primary, &secondary);
/* If PCI bus is present then don't probe for tertiary
legacy ports */
pci_present = 1;
}
if (winbond == 1)
winbond = 0x130; /* Default port, alt is 1B0 */
if (primary == 0 || all)
legacy_probe_add(0x1F0, 14, UNKNOWN, 0);
if (secondary == 0 || all)
legacy_probe_add(0x170, 15, UNKNOWN, 0);
if (probe_all || !pci_present) {
/* ISA/VLB extra ports */
legacy_probe_add(0x1E8, 11, UNKNOWN, 0);
legacy_probe_add(0x168, 10, UNKNOWN, 0);
legacy_probe_add(0x1E0, 8, UNKNOWN, 0);
legacy_probe_add(0x160, 12, UNKNOWN, 0);
}
if (opti82c46x)
probe_opti_vlb();
if (qdi)
probe_qdi_vlb();
for (i = 0; i < NR_HOST; i++, pl++) {
if (pl->port == 0)
continue;
if (pl->type == UNKNOWN)
pl->type = probe_chip_type(pl);
pl->slot = slot++;
if (legacy_init_one(pl) == 0)
ct++;
}
if (ct != 0)
return 0;
return -ENODEV;
}
static __exit void legacy_exit(void)
{
int i;
for (i = 0; i < nr_legacy_host; i++) {
struct legacy_data *ld = &legacy_data[i];
ata_host_detach(legacy_host[i]);
platform_device_unregister(ld->platform_dev);
}
}
MODULE_AUTHOR("Alan Cox");
MODULE_DESCRIPTION("low-level driver for legacy ATA");
MODULE_LICENSE("GPL");
MODULE_VERSION(DRV_VERSION);
module_param(probe_all, int, 0);
module_param(autospeed, int, 0);
module_param(ht6560a, int, 0);
module_param(ht6560b, int, 0);
module_param(opti82c611a, int, 0);
module_param(opti82c46x, int, 0);
module_param(qdi, int, 0);
module_param(pio_mask, int, 0);
module_param(iordy_mask, int, 0);
module_init(legacy_init);
module_exit(legacy_exit);