Skip to contentSkip to conversation

devicefs & sleep support for IDE

PATCH Pavel Machek 3 participants 5 messages Sep 22, 2002 – Sep 26, 2002

lore
Share & tools
lore
#0 lore raw

Hi!

New patch, rediffed against 2.5.36.

More patches will be needed to support IDE properly (like DVD burners you mentioned), but this is known to fix data corruption. It has zero impact on actual I/O. It affects initialization and suspend only. Please apply this time.

Pavel

drivers/ide/ide-disk.c +70 -2
@@ -544,6 +544,7 @@
  */
 static ide_startstop_t do_rw_disk (ide_drive_t *drive, struct request *rq, unsigned long block)
 {
+	BUG_ON(drive->blocked);
 	if (!blk_fs_request(rq)) {
 		blk_dump_rq_flags(rq, "do_rw_disk - bad command");
 		ide_end_request(drive, 0, 0);
@@ -1495,8 +1496,63 @@
  	ide_add_setting(drive,	"max_failures",		SETTING_RW,					-1,			-1,			TYPE_INT,	0,	65535,				1,	1,	&drive->max_failures,		NULL);
 }
 
+static int idedisk_suspend(struct device *dev, u32 state, u32 level)
+{
+	ide_drive_t *drive = dev->driver_data;
+
+	/* I hope that every freeze operations from the upper levels have
+	 * already been done...
+	 */
+
+	BUG_ON(in_interrupt());
+
+	if (level != SUSPEND_SAVE_STATE)
+		return 0;
+
+	/* wait until all commands are finished */
+	/* FIXME: waiting for spinlocks should be done instead. */
+	while (HWGROUP(drive)->handler)
+		yield();
+
+	/* set the drive to standby */
+	printk(KERN_INFO "suspending: %s ", drive->name);
+	if (drive->driver) {
+		if (drive->driver->standby)
+			drive->driver->standby(drive);
+	}
+	drive->blocked = 1;
+
+	while (HWGROUP(drive)->handler)
+		yield();
+
+	return 0;
+}
+
+static int idedisk_resume(struct device *dev, u32 level)
+{
+	ide_drive_t *drive = dev->driver_data;
+
+	if (level != RESUME_RESTORE_STATE)
+		return 0;
+	if (!drive->blocked)
+		panic("ide: Resume but not suspended?\n");
+
+	drive->blocked = 0;
+	return 0;
+}
+
+
+/* This is just a hook for the overall driver tree.
+ */
+
+static struct device_driver idedisk_devdrv = {
+	.lock = RW_LOCK_UNLOCKED,
+	.suspend = idedisk_suspend,
+	.resume = idedisk_resume,
+};
+
 static int idedisk_ioctl (ide_drive_t *drive, struct inode *inode,
-		struct file *file, unsigned int cmd, unsigned long arg)
+	struct file *file, unsigned int cmd, unsigned long arg)
 {
 #if 0
 HDIO_GET_ADDRESS
@@ -1517,12 +1573,13 @@
 {
 	struct hd_driveid *id = drive->id;
 	unsigned long capacity;
+	int myid = -1;
 
 #if 0
 	if (IS_PDC4030_DRIVE)
 		DRIVER(drive)->do_request = promise_rw_disk;
 #endif
-	
+
 	idedisk_add_settings(drive);
 
 	if (id == NULL)
@@ -1540,6 +1597,15 @@
 			drive->doorlocking = 1;
 		}
 	}
+	{
+		ide_hwif_t *hwif = HWIF(drive);
+		sprintf(drive->device.bus_id, "%d", myid);
+		sprintf(drive->device.name, "ide-disk");
+		drive->device.driver = &idedisk_devdrv;
+		drive->device.parent = &hwif->device;
+		drive->device.driver_data = drive;
+		device_register(&drive->device);
+	}
 
 #if 1
 	(void) probe_lba_addressing(drive, 1);
@@ -1623,6 +1689,8 @@
 static int idedisk_cleanup (ide_drive_t *drive)
 {
 	struct gendisk *g = drive->disk;
+
+	put_device(&drive->device);
 	if ((drive->id->cfs_enable_2 & 0x3000) && drive->wcache)
 		if (do_idedisk_flushcache(drive))
 			printk (KERN_INFO "%s: Write Cache FAILED Flushing!\n",
drivers/ide/ide-pnp.c +3 -1
@@ -52,6 +52,7 @@
 static int __init pnpide_generic_init(struct pci_dev *dev, int enable)
 {
 	hw_regs_t hw;
+	ide_hwif_t *hwif;
 	int index;
 
 	if (!enable)
@@ -67,10 +68,11 @@
 //			generic_pnp_ide_iops,
 			DEV_IRQ(dev, 0));
 
-	index = ide_register_hw(&hw, NULL);
+	index = ide_register_hw(&hw, &hwif);
 
 	if (index != -1) {
 	    	printk(KERN_INFO "ide%d: %s IDE interface\n", index, DEV_NAME(dev));
+		hwif->pci_dev = dev;
 		return 0;
 	}
 
drivers/ide/ide-probe.c +10 -0
@@ -46,6 +46,7 @@
 #include <linux/delay.h>
 #include <linux/ide.h>
 #include <linux/spinlock.h>
+#include <linux/pci.h>
 
 #include <asm/byteorder.h>
 #include <asm/irq.h>
@@ -563,6 +564,15 @@
 {
 	u32 i = 0;
 
+	sprintf(hwif->device.bus_id, "%04x", hwif->io_ports[IDE_DATA_OFFSET]);
+	sprintf(hwif->device.name, "ide");
+	hwif->device.driver_data = hwif;
+	if (hwif->pci_dev)
+		hwif->device.parent = &hwif->pci_dev->dev;
+	else
+		hwif->device.parent = NULL; /* Would like to do = &device_legacy */
+	device_register(&hwif->device);
+
 	if (hwif->mmio == 2)
 		return;
 	if (hwif->io_ports[IDE_CONTROL_OFFSET])
drivers/ide/ide.c +3 -1
@@ -152,6 +152,8 @@
 #include <linux/reboot.h>
 #include <linux/cdrom.h>
 #include <linux/seq_file.h>
+#include <linux/device.h>
+#include <linux/kmod.h>
 
 #include <asm/byteorder.h>
 #include <asm/irq.h>
@@ -161,7 +163,6 @@
 
 #include "ide_modes.h"
 
-#include <linux/kmod.h>
 
 /* default maximum number of failures */
 #define IDE_DEFAULT_MAX_FAILURES 	1
@@ -1755,6 +1756,7 @@
 	hwif = &ide_hwifs[index];
 	if (!hwif->present)
 		goto abort;
+	put_device(&hwif->device);
 	for (unit = 0; unit < MAX_DRIVES; ++unit) {
 		drive = &hwif->drives[unit];
 		if (!drive->present)
include/linux/ide.h +4 -0
@@ -17,6 +17,7 @@
 #include <linux/interrupt.h>
 #include <linux/bitops.h>
 #include <linux/bio.h>
+#include <linux/device.h>
 #include <linux/pci.h>
 #include <asm/byteorder.h>
 #include <asm/system.h>
@@ -789,6 +790,7 @@
 	unsigned autotune	: 2;	/* 1=autotune, 2=noautotune, 0=default */
 	unsigned remap_0_to_1	: 2;	/* 0=remap if ezdrive, 1=remap, 2=noremap */
 	unsigned ata_flash	: 1;	/* 1=present, 0=default */
+	unsigned blocked        : 1;	/* 1=powermanagment told us not to do anything, so sleep nicely */
 	unsigned addressing;		/*      : 3;
 					 *  0=28-bit
 					 *  1=48-bit
@@ -835,6 +837,7 @@
 	int		crc_count;	/* crc counter to reduce drive speed */
 	struct list_head list;
 	struct gendisk *disk;
+	struct device	device;		/* for driverfs */
 } ide_drive_t;
 
 typedef struct ide_pio_ops_s {
@@ -1073,6 +1076,7 @@
 	unsigned	no_dsc     : 1;	/* 0 default, 1 dsc_overlap disabled */
 
 	void		*hwif_data;	/* extra hwif data */
+	struct device	device;		/* for devicefs */
 } ide_hwif_t;
 
 /*
Signature · 2 lines
Worst form of spam? Adding advertisment signatures ala sourceforge.net.
What goes next? Inserting advertisment *into* email?
#1 lore raw
Quoted 7 lines

New patch, rediffed against 2.5.36.

More patches will be needed to support IDE properly (like DVD burners you mentioned), but this is known to fix data corruption. It has zero impact on actual I/O. It affects initialization and suspend only. Please apply this time.

Basic driver model support for IDE is in 2.5.38. This just involves creating an IDE bus type, and registering drives as devices. I.e. there is no driver set for any of the drives.

I do have a couple of comments though.

drivers/ide/ide-disk.c line 1552

Quoted 5 lines
drivers/ide/ide-disk.c +5 -0
+static struct device_driver idedisk_devdrv = {
+	.lock = RW_LOCK_UNLOCKED,
+	.suspend = idedisk_suspend,
+	.resume = idedisk_resume,
+};

You don't need to initialize .lock. But, you do need a .name and .bus. The driver won't even be registered unless .bus is set.

include/linux/ide.h line 841

Quoted 6 lines
include/linux/ide.h +1 -0
@@ -835,6 +837,7 @@
 	int		crc_count;	/* crc counter to reduce drive speed */
 	struct list_head list;
 	struct gendisk *disk;
+	struct device	device;		/* for driverfs */
 } ide_drive_t;

There is a struct device in struct gendisk; that should suffice. But note that you may have to do an extra conversion in order to access it in the driver callbacks.

include/linux/ide.h line 1079

Quoted 1 line
include/linux/ide.h +1 -0
+	struct device	device;		/* for devicefs */

Please: it's driver model support, not driverfs. And devicefs does not even exist. :)

Thanks,

-pat

#2 lore raw

Hi!

Quoted 10 lines
Quoted 6 lines

New patch, rediffed against 2.5.36.

More patches will be needed to support IDE properly (like DVD burners you mentioned), but this is known to fix data corruption. It has zero impact on actual I/O. It affects initialization and suspend only. Please apply this time.

Basic driver model support for IDE is in 2.5.38. This just involves creating an IDE bus type, and registering drives as devices. I.e. there is no driver set for any of the drives.

Yep, I saw the support and it confused me a lot.

Questions: is it possible that in hwif_register you don't need to initialize parent?

Where is device_put of hwif->gendev? I miss it.

Quoted 11 lines

I do have a couple of comments though.

drivers/ide/ide-disk.c line 1552

Quoted 5 lines
drivers/ide/ide-disk.c +5 -0
+static struct device_driver idedisk_devdrv = {
+	.lock = RW_LOCK_UNLOCKED,
+	.suspend = idedisk_suspend,
+	.resume = idedisk_resume,
+};

You don't need to initialize .lock. But, you do need a .name and .bus. The driver won't even be registered unless .bus is set.

Quoted 10 lines

include/linux/ide.h line 841

Quoted 6 lines
include/linux/ide.h +1 -0
@@ -835,6 +837,7 @@
 	int		crc_count;	/* crc counter to reduce drive speed */
 	struct list_head list;
 	struct gendisk *disk;
+	struct device	device;		/* for driverfs */
 } ide_drive_t;

There is a struct device in struct gendisk; that should suffice. But note that you may have to do an extra conversion in order to access it in the driver callbacks.

Thanx for info.

Ouch. There are actually two devices in struct gendisk. I choosed disk_dev. Was it right?

Quoted 4 lines

include/linux/ide.h line 1079

Quoted 1 line
include/linux/ide.h +1 -0
+	struct device	device;		/* for devicefs */

Please: it's driver model support, not driverfs. And devicefs does not even exist. :)

Okay, okay ;-).
Pavel

Signature · 2 lines
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.
#3 lore raw
Quoted 3 lines

Questions: is it possible that in hwif_register you don't need to initialize parent?

Actually, looking at it again, the struct device in hwif_t should probably go away. We should initialize the parent device to the struct device in the struct pci_dev of the controller; at least for PCI controllers.

For non-PCI controllers, is there anything else that describes the controller besides hwif_t ?

Quoted 1 line

Where is device_put of hwif->gendev? I miss it.

There is none (yet), as there is no driver attached to it. If we remove it in favor of the PCI device, then we get it in the pci driver for the controller; and we'd need one for non-PCI controllers.

Quoted 2 lines

Ouch. There are actually two devices in struct gendisk. I choosed disk_dev. Was it right?

Yes. The other one:

	struct device *driverfs_dev;

was from the SCSI people. I think they had good intentions, but I'm not sure what they were doing in several places (including here).

-pat

#4 lore raw
Quoted 5 lines · Patrick Mochel

Actually, looking at it again, the struct device in hwif_t should probably go away. We should initialize the parent device to the struct device in the struct pci_dev of the controller; at least for PCI controllers.

From a 2.4 point of view and a making IDE work point of view leave it

alone for now.

Quoted 2 lines

For non-PCI controllers, is there anything else that describes the controller besides hwif_t ?

Undefined and basically to the IDE core "none of your business"

Alan