linux/drivers/ata/pata_ns87415.c

423 lines
11 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* pata_ns87415.c - NS87415 (and PARISC SUPERIO 87560) PATA
*
* (C) 2005 Red Hat <alan@lxorguk.ukuu.org.uk>
*
* This is a fairly generic MWDMA controller. It has some limitations
* as it requires timing reloads on PIO/DMA transitions but it is otherwise
* fairly well designed.
*
* This driver assumes the firmware has left the chip in a valid ST506
* compliant state, either legacy IRQ 14/15 or native INTA shared. You
* may need to add platform code if your system fails to do this.
*
* The same cell appears in the 87560 controller used by some PARISC
* systems. This has its own special mountain of errata.
*
* TODO:
* Get someone to test on SPARC
* Implement lazy pio/dma switching for better performance
* 8bit shared timing.
* See if we need to kill the FIFO for ATAPI
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/pci.h>
#include <linux/blkdev.h>
#include <linux/delay.h>
#include <linux/device.h>
#include <scsi/scsi_host.h>
#include <linux/libata.h>
#include <linux/ata.h>
#define DRV_NAME "pata_ns87415"
#define DRV_VERSION "0.0.1"
/**
* ns87415_set_mode - Initialize host controller mode timings
* @ap: Port whose timings we are configuring
* @adev: Device whose timings we are configuring
* @mode: Mode to set
*
* Program the mode registers for this controller, channel and
* device. Because the chip is quite an old design we have to do this
* for PIO/DMA switches.
*
* LOCKING:
* None (inherited from caller).
*/
static void ns87415_set_mode(struct ata_port *ap, struct ata_device *adev, u8 mode)
{
struct pci_dev *dev = to_pci_dev(ap->host->dev);
int unit = 2 * ap->port_no + adev->devno;
int timing = 0x44 + 2 * unit;
unsigned long T = 1000000000 / 33333; /* PCI clocks */
struct ata_timing t;
u16 clocking;
u8 iordy;
u8 status;
/* Timing register format is 17 - low nybble read timing with
the high nybble being 16 - x for recovery time in PCI clocks */
ata_timing_compute(adev, adev->pio_mode, &t, T, 0);
clocking = 17 - clamp_val(t.active, 2, 17);
clocking |= (16 - clamp_val(t.recover, 1, 16)) << 4;
/* Use the same timing for read and write bytes */
clocking |= (clocking << 8);
pci_write_config_word(dev, timing, clocking);
/* Set the IORDY enable versus DMA enable on or off properly */
pci_read_config_byte(dev, 0x42, &iordy);
iordy &= ~(1 << (4 + unit));
if (mode >= XFER_MW_DMA_0 || !ata_pio_need_iordy(adev))
iordy |= (1 << (4 + unit));
/* Paranoia: We shouldn't ever get here with busy write buffers
but if so wait */
pci_read_config_byte(dev, 0x43, &status);
while (status & 0x03) {
udelay(1);
pci_read_config_byte(dev, 0x43, &status);
}
/* Flip the IORDY/DMA bits now we are sure the write buffers are
clear */
pci_write_config_byte(dev, 0x42, iordy);
/* TODO: Set byte 54 command timing to the best 8bit
mode shared by all four devices */
}
/**
* ns87415_set_piomode - Initialize host controller PATA PIO timings
* @ap: Port whose timings we are configuring
* @adev: Device to program
*
* Set PIO mode for device, in host controller PCI config space.
*
* LOCKING:
* None (inherited from caller).
*/
static void ns87415_set_piomode(struct ata_port *ap, struct ata_device *adev)
{
ns87415_set_mode(ap, adev, adev->pio_mode);
}
/**
* ns87415_bmdma_setup - Set up DMA
* @qc: Command block
*
* Set up for bus mastering DMA. We have to do this ourselves
* rather than use the helper due to a chip erratum
*/
static void ns87415_bmdma_setup(struct ata_queued_cmd *qc)
{
struct ata_port *ap = qc->ap;
unsigned int rw = (qc->tf.flags & ATA_TFLAG_WRITE);
u8 dmactl;
/* load PRD table addr. */
mb(); /* make sure PRD table writes are visible to controller */
iowrite32(ap->bmdma_prd_dma, ap->ioaddr.bmdma_addr + ATA_DMA_TABLE_OFS);
/* specify data direction, triple-check start bit is clear */
dmactl = ioread8(ap->ioaddr.bmdma_addr + ATA_DMA_CMD);
dmactl &= ~(ATA_DMA_WR | ATA_DMA_START);
/* Due to an erratum we need to write these bits to the wrong
place - which does save us an I/O bizarrely */
dmactl |= ATA_DMA_INTR | ATA_DMA_ERR;
if (!rw)
dmactl |= ATA_DMA_WR;
iowrite8(dmactl, ap->ioaddr.bmdma_addr + ATA_DMA_CMD);
/* issue r/w command */
ap->ops->sff_exec_command(ap, &qc->tf);
}
/**
* ns87415_bmdma_start - Begin DMA transfer
* @qc: Command block
*
* Switch the timings for the chip and set up for a DMA transfer
* before the DMA burst begins.
*
* FIXME: We should do lazy switching on bmdma_start versus
* ata_pio_data_xfer for better performance.
*/
static void ns87415_bmdma_start(struct ata_queued_cmd *qc)
{
ns87415_set_mode(qc->ap, qc->dev, qc->dev->dma_mode);
ata_bmdma_start(qc);
}
/**
* ns87415_bmdma_stop - End DMA transfer
* @qc: Command block
*
* End DMA mode and switch the controller back into PIO mode
*/
static void ns87415_bmdma_stop(struct ata_queued_cmd *qc)
{
ata_bmdma_stop(qc);
ns87415_set_mode(qc->ap, qc->dev, qc->dev->pio_mode);
}
/**
* ns87415_irq_clear - Clear interrupt
* @ap: Channel to clear
*
* Erratum: Due to a chip bug registers 02 and 0A bit 1 and 2 (the
* error bits) are reset by writing to register 00 or 08.
*/
static void ns87415_irq_clear(struct ata_port *ap)
{
void __iomem *mmio = ap->ioaddr.bmdma_addr;
if (!mmio)
return;
iowrite8((ioread8(mmio + ATA_DMA_CMD) | ATA_DMA_INTR | ATA_DMA_ERR),
mmio + ATA_DMA_CMD);
}
/**
* ns87415_check_atapi_dma - ATAPI DMA filter
* @qc: Command block
*
* Disable ATAPI DMA (for now). We may be able to do DMA if we
* kill the prefetching. This isn't clear.
*/
static int ns87415_check_atapi_dma(struct ata_queued_cmd *qc)
{
return -EOPNOTSUPP;
}
#if defined(CONFIG_SUPERIO)
/* SUPERIO 87560 is a PoS chip that NatSem denies exists.
* Unfortunately, it's built-in on all Astro-based PA-RISC workstations
* which use the integrated NS87514 cell for CD-ROM support.
* i.e we have to support for CD-ROM installs.
* See drivers/parisc/superio.c for more gory details.
*
* Workarounds taken from drivers/ide/pci/ns87415.c
*/
#include <asm/superio.h>
#define SUPERIO_IDE_MAX_RETRIES 25
/**
* ns87560_read_buggy - workaround buggy Super I/O chip
* @port: Port to read
*
* Work around chipset problems in the 87560 SuperIO chip
*/
static u8 ns87560_read_buggy(void __iomem *port)
{
u8 tmp;
int retries = SUPERIO_IDE_MAX_RETRIES;
do {
tmp = ioread8(port);
if (tmp != 0)
return tmp;
udelay(50);
} while(retries-- > 0);
return tmp;
}
/**
* ns87560_check_status
* @ap: channel to check
*
* Return the status of the channel working around the
* 87560 flaws.
*/
static u8 ns87560_check_status(struct ata_port *ap)
{
return ns87560_read_buggy(ap->ioaddr.status_addr);
}
/**
* ns87560_tf_read - input device's ATA taskfile shadow registers
* @ap: Port from which input is read
* @tf: ATA taskfile register set for storing input
*
* Reads ATA taskfile registers for currently-selected device
* into @tf. Work around the 87560 bugs.
*
* LOCKING:
* Inherited from caller.
*/
void ns87560_tf_read(struct ata_port *ap, struct ata_taskfile *tf)
{
struct ata_ioports *ioaddr = &ap->ioaddr;
tf->status = ns87560_check_status(ap);
tf->error = ioread8(ioaddr->error_addr);
tf->nsect = ioread8(ioaddr->nsect_addr);
tf->lbal = ioread8(ioaddr->lbal_addr);
tf->lbam = ioread8(ioaddr->lbam_addr);
tf->lbah = ioread8(ioaddr->lbah_addr);
tf->device = ns87560_read_buggy(ioaddr->device_addr);
if (tf->flags & ATA_TFLAG_LBA48) {
iowrite8(tf->ctl | ATA_HOB, ioaddr->ctl_addr);
tf->hob_feature = ioread8(ioaddr->error_addr);
tf->hob_nsect = ioread8(ioaddr->nsect_addr);
tf->hob_lbal = ioread8(ioaddr->lbal_addr);
tf->hob_lbam = ioread8(ioaddr->lbam_addr);
tf->hob_lbah = ioread8(ioaddr->lbah_addr);
iowrite8(tf->ctl, ioaddr->ctl_addr);
ap->last_ctl = tf->ctl;
}
}
/**
* ns87560_bmdma_status
* @ap: channel to check
*
* Return the DMA status of the channel working around the
* 87560 flaws.
*/
static u8 ns87560_bmdma_status(struct ata_port *ap)
{
return ns87560_read_buggy(ap->ioaddr.bmdma_addr + ATA_DMA_STATUS);
}
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
#endif /* 87560 SuperIO Support */
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 struct ata_port_operations ns87415_pata_ops = {
.inherits = &ata_bmdma_port_ops,
.check_atapi_dma = ns87415_check_atapi_dma,
.bmdma_setup = ns87415_bmdma_setup,
.bmdma_start = ns87415_bmdma_start,
.bmdma_stop = ns87415_bmdma_stop,
.sff_irq_clear = ns87415_irq_clear,
.cable_detect = ata_cable_40wire,
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_piomode = ns87415_set_piomode,
};
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
#if defined(CONFIG_SUPERIO)
static struct ata_port_operations ns87560_pata_ops = {
.inherits = &ns87415_pata_ops,
.sff_tf_read = ns87560_tf_read,
.sff_check_status = ns87560_check_status,
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
.bmdma_status = ns87560_bmdma_status,
};
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
#endif
static struct scsi_host_template ns87415_sht = {
ATA_BMDMA_SHT(DRV_NAME),
};
static void ns87415_fixup(struct pci_dev *pdev)
{
/* Select 512 byte sectors */
pci_write_config_byte(pdev, 0x55, 0xEE);
/* Select PIO0 8bit clocking */
pci_write_config_byte(pdev, 0x54, 0xB7);
}
/**
* ns87415_init_one - Register 87415 ATA PCI device with kernel services
* @pdev: PCI device to register
* @ent: Entry in ns87415_pci_tbl matching with @pdev
*
* Called from kernel PCI layer. We probe for combined mode (sigh),
* and then hand over control to libata, for it to do the rest.
*
* LOCKING:
* Inherited from PCI layer (may sleep).
*
* RETURNS:
* Zero on success, or -ERRNO value.
*/
static int ns87415_init_one (struct pci_dev *pdev, const struct pci_device_id *ent)
{
static const struct ata_port_info info = {
.flags = ATA_FLAG_SLAVE_POSS,
.pio_mask = ATA_PIO4,
.mwdma_mask = ATA_MWDMA2,
.port_ops = &ns87415_pata_ops,
};
const struct ata_port_info *ppi[] = { &info, NULL };
int rc;
#if defined(CONFIG_SUPERIO)
static const struct ata_port_info info87560 = {
.flags = ATA_FLAG_SLAVE_POSS,
.pio_mask = ATA_PIO4,
.mwdma_mask = ATA_MWDMA2,
.port_ops = &ns87560_pata_ops,
};
if (PCI_SLOT(pdev->devfn) == 0x0E)
ppi[0] = &info87560;
#endif
ata_print_version_once(&pdev->dev, DRV_VERSION);
rc = pcim_enable_device(pdev);
if (rc)
return rc;
ns87415_fixup(pdev);
return ata_pci_bmdma_init_one(pdev, ppi, &ns87415_sht, NULL, 0);
}
static const struct pci_device_id ns87415_pci_tbl[] = {
{ PCI_VDEVICE(NS, PCI_DEVICE_ID_NS_87415), },
{ } /* terminate list */
};
#ifdef CONFIG_PM_SLEEP
static int ns87415_reinit_one(struct pci_dev *pdev)
{
struct ata_host *host = pci_get_drvdata(pdev);
int rc;
rc = ata_pci_device_do_resume(pdev);
if (rc)
return rc;
ns87415_fixup(pdev);
ata_host_resume(host);
return 0;
}
#endif
static struct pci_driver ns87415_pci_driver = {
.name = DRV_NAME,
.id_table = ns87415_pci_tbl,
.probe = ns87415_init_one,
.remove = ata_pci_remove_one,
#ifdef CONFIG_PM_SLEEP
.suspend = ata_pci_device_suspend,
.resume = ns87415_reinit_one,
#endif
};
module_pci_driver(ns87415_pci_driver);
MODULE_AUTHOR("Alan Cox");
MODULE_DESCRIPTION("ATA low-level driver for NS87415 controllers");
MODULE_LICENSE("GPL");
MODULE_DEVICE_TABLE(pci, ns87415_pci_tbl);
MODULE_VERSION(DRV_VERSION);