From ec2cdedf798385a9397ac50dd0405dd658f8529c Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:30:15 +0900 Subject: block: allow deleting zero length partition delete_partition() was noop for zero length partition. As the addition code allows creating zero lenght partition and deletion is assumed to always succeed, this causes memory leak for zero length partitions. Allow zero length partitions to end their meaningless lives. While at it, allow deleting zero lenght partition via BLKPG_DEL_PARTITION ioctl too. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- fs/partitions/check.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs/partitions') diff --git a/fs/partitions/check.c b/fs/partitions/check.c index ecc3330972e..68f3e41ae66 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -325,8 +325,6 @@ void delete_partition(struct gendisk *disk, int part) if (!p) return; - if (!p->nr_sects) - return; disk->part[part-1] = NULL; p->start_sect = 0; p->nr_sects = 0; -- cgit v1.2.3-70-g09d2 From 88e341261ca4d39eec21b212961c77eff51105f7 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:30:16 +0900 Subject: block: update add_partition() error handling d805dda4 tried to fix error case handling in add_partition() but had a few problems. * disk->part[] entry is set early and left dangling if operation fails. * Once device initialized, the last put_device() is responsible for freeing all the resources. The failure path freed part_stats and p regardless of put_device() causing double free. * holders subdir holds reference to the disk device, so failure path should remove it to release resources properly which was missing. This patch fixes the above problems and while at it move partition slot busy check into add_partition() for completeness and inlines holders subdirectory creation. Using separate function for it just obfuscates the code. Signed-off-by: Tejun Heo Cc: Abdel Benamrouche Signed-off-by: Jens Axboe --- block/ioctl.c | 7 ++----- fs/partitions/check.c | 40 +++++++++++++++++++++------------------- 2 files changed, 23 insertions(+), 24 deletions(-) (limited to 'fs/partitions') diff --git a/block/ioctl.c b/block/ioctl.c index c722de0ef2e..eb046aeede8 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -43,12 +43,9 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user || pstart < 0 || plength < 0) return -EINVAL; } - /* partition number in use? */ + mutex_lock(&bdev->bd_mutex); - if (disk->part[part - 1]) { - mutex_unlock(&bdev->bd_mutex); - return -EBUSY; - } + /* overlap? */ for (i = 0; i < disk->minors - 1; i++) { struct hd_struct *s = disk->part[i]; diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 68f3e41ae66..16f98d82460 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -300,15 +300,6 @@ struct device_type part_type = { .release = part_release, }; -static inline void partition_sysfs_add_subdir(struct hd_struct *p) -{ - struct kobject *k; - - k = kobject_get(&p->dev.kobj); - p->holder_dir = kobject_create_and_add("holders", k); - kobject_put(k); -} - static inline void disk_sysfs_add_subdirs(struct gendisk *disk) { struct kobject *k; @@ -347,13 +338,16 @@ int add_partition(struct gendisk *disk, int part, sector_t start, sector_t len, struct hd_struct *p; int err; + if (disk->part[part - 1]) + return -EBUSY; + p = kzalloc(sizeof(*p), GFP_KERNEL); if (!p) return -ENOMEM; if (!init_part_stats(p)) { err = -ENOMEM; - goto out0; + goto out_free; } p->start_sect = start; p->nr_sects = len; @@ -372,34 +366,42 @@ int add_partition(struct gendisk *disk, int part, sector_t start, sector_t len, p->dev.class = &block_class; p->dev.type = &part_type; p->dev.parent = &disk->dev; - disk->part[part-1] = p; /* delay uevent until 'holders' subdir is created */ p->dev.uevent_suppress = 1; err = device_add(&p->dev); if (err) - goto out1; - partition_sysfs_add_subdir(p); + goto out_put; + + err = -ENOMEM; + p->holder_dir = kobject_create_and_add("holders", &p->dev.kobj); + if (!p->holder_dir) + goto out_del; + p->dev.uevent_suppress = 0; if (flags & ADDPART_FLAG_WHOLEDISK) { err = device_create_file(&p->dev, &dev_attr_whole_disk); if (err) - goto out2; + goto out_del; } + /* everything is up and running, commence */ + disk->part[part - 1] = p; + /* suppress uevent if the disk supresses it */ if (!disk->dev.uevent_suppress) kobject_uevent(&p->dev.kobj, KOBJ_ADD); return 0; -out2: +out_free: + kfree(p); + return err; +out_del: + kobject_put(p->holder_dir); device_del(&p->dev); -out1: +out_put: put_device(&p->dev); - free_part_stats(p); -out0: - kfree(p); return err; } -- cgit v1.2.3-70-g09d2 From cf771cb5a7b716f3f9e532fd42a1e3a0a75adec5 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 3 Sep 2008 09:01:09 +0200 Subject: block: make variable and argument names more consistent In hd_struct, @partno is used to denote partition number and a number of other places use @part to denote hd_struct. Functions use @part and @index instead. This causes confusion and makes it difficult to use consistent variable names for hd_struct. Always use @partno if a variable represents partition number. Also, print out functions use @f or @part for seq_file argument. Use @seqf uniformly instead. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/genhd.c | 54 ++++++++++++++++++++++++--------------------------- block/ioctl.c | 15 +++++++------- fs/block_dev.c | 8 ++++---- fs/partitions/check.c | 33 ++++++++++++++++--------------- include/linux/genhd.h | 12 ++++++------ 5 files changed, 60 insertions(+), 62 deletions(-) (limited to 'fs/partitions') diff --git a/block/genhd.c b/block/genhd.c index 11038fbc75e..dc9ad4c171e 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -43,14 +43,14 @@ static inline int major_to_index(int major) } #ifdef CONFIG_PROC_FS -void blkdev_show(struct seq_file *f, off_t offset) +void blkdev_show(struct seq_file *seqf, off_t offset) { struct blk_major_name *dp; if (offset < BLKDEV_MAJOR_HASH_SIZE) { mutex_lock(&block_class_lock); for (dp = major_names[offset]; dp; dp = dp->next) - seq_printf(f, "%3d %s\n", dp->major, dp->name); + seq_printf(seqf, "%3d %s\n", dp->major, dp->name); mutex_unlock(&block_class_lock); } } @@ -157,7 +157,7 @@ void blk_unregister_region(dev_t devt, unsigned long range) EXPORT_SYMBOL(blk_unregister_region); -static struct kobject *exact_match(dev_t devt, int *part, void *data) +static struct kobject *exact_match(dev_t devt, int *partno, void *data) { struct gendisk *p = data; @@ -217,9 +217,9 @@ void unlink_gendisk(struct gendisk *disk) * This function gets the structure containing partitioning * information for the given device @devt. */ -struct gendisk *get_gendisk(dev_t devt, int *part) +struct gendisk *get_gendisk(dev_t devt, int *partno) { - struct kobject *kobj = kobj_lookup(bdev_map, devt, part); + struct kobject *kobj = kobj_lookup(bdev_map, devt, partno); struct device *dev = kobj_to_dev(kobj); return kobj ? dev_to_disk(dev) : NULL; @@ -336,23 +336,12 @@ static void *show_partition_start(struct seq_file *seqf, loff_t *pos) return p; } -static int show_partition(struct seq_file *part, void *v) +static int show_partition(struct seq_file *seqf, void *v) { struct gendisk *sgp = v; int n; char buf[BDEVNAME_SIZE]; - /* - * Print header if start told us to do. This is to preserve - * the original behavior of not printing header if no - * partition exists. This hackery will be removed later with - * class iteration clean up. - */ - if (part->private) { - seq_puts(part, "major minor #blocks name\n\n"); - part->private = NULL; - } - /* Don't show non-partitionable removeable devices or empty devices */ if (!get_capacity(sgp) || (sgp->minors == 1 && (sgp->flags & GENHD_FL_REMOVABLE))) @@ -361,7 +350,7 @@ static int show_partition(struct seq_file *part, void *v) return 0; /* show the full disk and all non-0 size partitions of it */ - seq_printf(part, "%4d %4d %10llu %s\n", + seq_printf(seqf, "%4d %4d %10llu %s\n", sgp->major, sgp->first_minor, (unsigned long long)get_capacity(sgp) >> 1, disk_name(sgp, 0, buf)); @@ -370,7 +359,7 @@ static int show_partition(struct seq_file *part, void *v) continue; if (sgp->part[n]->nr_sects == 0) continue; - seq_printf(part, "%4d %4d %10llu %s\n", + seq_printf(seqf, "%4d %4d %10llu %s\n", sgp->major, n + 1 + sgp->first_minor, (unsigned long long)sgp->part[n]->nr_sects >> 1 , disk_name(sgp, n + 1, buf)); @@ -388,7 +377,7 @@ const struct seq_operations partitions_op = { #endif -static struct kobject *base_probe(dev_t devt, int *part, void *data) +static struct kobject *base_probe(dev_t devt, int *partno, void *data) { if (request_module("block-major-%d-%d", MAJOR(devt), MINOR(devt)) > 0) /* Make old-style 2.4 aliases work */ @@ -564,7 +553,14 @@ static struct device_type disk_type = { }; #ifdef CONFIG_PROC_FS -static int diskstats_show(struct seq_file *s, void *v) +/* + * aggregate disk stat collector. Uses the same stats that the sysfs + * entries do, above, but makes them available through one seq_file. + * + * The output looks suspiciously like /proc/partitions with a bunch of + * extra fields. + */ +static int diskstats_show(struct seq_file *seqf, void *v) { struct gendisk *gp = v; char buf[BDEVNAME_SIZE]; @@ -572,7 +568,7 @@ static int diskstats_show(struct seq_file *s, void *v) /* if (&gp->dev.kobj.entry == block_class.devices.next) - seq_puts(s, "major minor name" + seq_puts(seqf, "major minor name" " rio rmerge rsect ruse wio wmerge " "wsect wuse running use aveq" "\n\n"); @@ -581,7 +577,7 @@ static int diskstats_show(struct seq_file *s, void *v) preempt_disable(); disk_round_stats(gp); preempt_enable(); - seq_printf(s, "%4d %4d %s %lu %lu %llu %u %lu %lu %llu %u %u %u %u\n", + seq_printf(seqf, "%4d %4d %s %lu %lu %llu %u %lu %lu %llu %u %u %u %u\n", gp->major, gp->first_minor, disk_name(gp, 0, buf), disk_stat_read(gp, ios[0]), disk_stat_read(gp, merges[0]), (unsigned long long)disk_stat_read(gp, sectors[0]), @@ -603,7 +599,7 @@ static int diskstats_show(struct seq_file *s, void *v) preempt_disable(); part_round_stats(hd); preempt_enable(); - seq_printf(s, "%4d %4d %s %lu %lu %llu " + seq_printf(seqf, "%4d %4d %s %lu %lu %llu " "%u %lu %lu %llu %u %u %u %u\n", gp->major, n + gp->first_minor + 1, disk_name(gp, n + 1, buf), @@ -655,7 +651,7 @@ void genhd_media_change_notify(struct gendisk *disk) EXPORT_SYMBOL_GPL(genhd_media_change_notify); #endif /* 0 */ -dev_t blk_lookup_devt(const char *name, int part) +dev_t blk_lookup_devt(const char *name, int partno) { dev_t devt = MKDEV(0, 0); struct class_dev_iter iter; @@ -665,9 +661,9 @@ dev_t blk_lookup_devt(const char *name, int part) while ((dev = class_dev_iter_next(&iter))) { struct gendisk *disk = dev_to_disk(dev); - if (!strcmp(dev->bus_id, name) && part < disk->minors) { + if (!strcmp(dev->bus_id, name) && partno < disk->minors) { devt = MKDEV(MAJOR(dev->devt), - MINOR(dev->devt) + part); + MINOR(dev->devt) + partno); break; } } @@ -777,10 +773,10 @@ int bdev_read_only(struct block_device *bdev) EXPORT_SYMBOL(bdev_read_only); -int invalidate_partition(struct gendisk *disk, int index) +int invalidate_partition(struct gendisk *disk, int partno) { int res = 0; - struct block_device *bdev = bdget_disk(disk, index); + struct block_device *bdev = bdget_disk(disk, partno); if (bdev) { fsync_bdev(bdev); res = __invalidate_device(bdev); diff --git a/block/ioctl.c b/block/ioctl.c index eb046aeede8..d77f5e280a6 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -15,7 +15,7 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user struct blkpg_ioctl_arg a; struct blkpg_partition p; long long start, length; - int part; + int partno; int i; int err; @@ -28,8 +28,8 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user disk = bdev->bd_disk; if (bdev != bdev->bd_contains) return -EINVAL; - part = p.pno; - if (part <= 0 || part >= disk->minors) + partno = p.pno; + if (partno <= 0 || partno >= disk->minors) return -EINVAL; switch (a.op) { case BLKPG_ADD_PARTITION: @@ -59,13 +59,14 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user } } /* all seems OK */ - err = add_partition(disk, part, start, length, ADDPART_FLAG_NONE); + err = add_partition(disk, partno, start, length, + ADDPART_FLAG_NONE); mutex_unlock(&bdev->bd_mutex); return err; case BLKPG_DEL_PARTITION: - if (!disk->part[part-1]) + if (!disk->part[partno - 1]) return -ENXIO; - bdevp = bdget_disk(disk, part); + bdevp = bdget_disk(disk, partno); if (!bdevp) return -ENOMEM; mutex_lock(&bdevp->bd_mutex); @@ -79,7 +80,7 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user invalidate_bdev(bdevp); mutex_lock_nested(&bdev->bd_mutex, 1); - delete_partition(disk, part); + delete_partition(disk, partno); mutex_unlock(&bdev->bd_mutex); mutex_unlock(&bdevp->bd_mutex); bdput(bdevp); diff --git a/fs/block_dev.c b/fs/block_dev.c index aff54219e04..de0776cd721 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -930,7 +930,7 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) struct module *owner = NULL; struct gendisk *disk; int ret; - int part; + int partno; int perm = 0; if (file->f_mode & FMODE_READ) @@ -949,7 +949,7 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) ret = -ENXIO; file->f_mapping = bdev->bd_inode->i_mapping; lock_kernel(); - disk = get_gendisk(bdev->bd_dev, &part); + disk = get_gendisk(bdev->bd_dev, &partno); if (!disk) { unlock_kernel(); bdput(bdev); @@ -961,7 +961,7 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) if (!bdev->bd_openers) { bdev->bd_disk = disk; bdev->bd_contains = bdev; - if (!part) { + if (!partno) { struct backing_dev_info *bdi; if (disk->fops->open) { ret = disk->fops->open(bdev->bd_inode, file); @@ -989,7 +989,7 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) if (ret) goto out_first; bdev->bd_contains = whole; - p = disk->part[part - 1]; + p = disk->part[partno - 1]; bdev->bd_inode->i_data.backing_dev_info = whole->bd_inode->i_data.backing_dev_info; if (!(disk->flags & GENHD_FL_UP) || !p || !p->nr_sects) { diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 16f98d82460..b86aab1b0df 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -120,22 +120,22 @@ static int (*check_part[])(struct parsed_partitions *, struct block_device *) = * a pointer to that same buffer (for convenience). */ -char *disk_name(struct gendisk *hd, int part, char *buf) +char *disk_name(struct gendisk *hd, int partno, char *buf) { - if (!part) + if (!partno) snprintf(buf, BDEVNAME_SIZE, "%s", hd->disk_name); else if (isdigit(hd->disk_name[strlen(hd->disk_name)-1])) - snprintf(buf, BDEVNAME_SIZE, "%sp%d", hd->disk_name, part); + snprintf(buf, BDEVNAME_SIZE, "%sp%d", hd->disk_name, partno); else - snprintf(buf, BDEVNAME_SIZE, "%s%d", hd->disk_name, part); + snprintf(buf, BDEVNAME_SIZE, "%s%d", hd->disk_name, partno); return buf; } const char *bdevname(struct block_device *bdev, char *buf) { - int part = MINOR(bdev->bd_dev) - bdev->bd_disk->first_minor; - return disk_name(bdev->bd_disk, part, buf); + int partno = MINOR(bdev->bd_dev) - bdev->bd_disk->first_minor; + return disk_name(bdev->bd_disk, partno, buf); } EXPORT_SYMBOL(bdevname); @@ -310,13 +310,13 @@ static inline void disk_sysfs_add_subdirs(struct gendisk *disk) kobject_put(k); } -void delete_partition(struct gendisk *disk, int part) +void delete_partition(struct gendisk *disk, int partno) { - struct hd_struct *p = disk->part[part-1]; + struct hd_struct *p = disk->part[partno - 1]; if (!p) return; - disk->part[part-1] = NULL; + disk->part[partno - 1] = NULL; p->start_sect = 0; p->nr_sects = 0; part_stat_set_all(p, 0); @@ -333,12 +333,13 @@ static ssize_t whole_disk_show(struct device *dev, static DEVICE_ATTR(whole_disk, S_IRUSR | S_IRGRP | S_IROTH, whole_disk_show, NULL); -int add_partition(struct gendisk *disk, int part, sector_t start, sector_t len, int flags) +int add_partition(struct gendisk *disk, int partno, + sector_t start, sector_t len, int flags) { struct hd_struct *p; int err; - if (disk->part[part - 1]) + if (disk->part[partno - 1]) return -EBUSY; p = kzalloc(sizeof(*p), GFP_KERNEL); @@ -351,18 +352,18 @@ int add_partition(struct gendisk *disk, int part, sector_t start, sector_t len, } p->start_sect = start; p->nr_sects = len; - p->partno = part; + p->partno = partno; p->policy = disk->policy; if (isdigit(disk->dev.bus_id[strlen(disk->dev.bus_id)-1])) snprintf(p->dev.bus_id, BUS_ID_SIZE, - "%sp%d", disk->dev.bus_id, part); + "%sp%d", disk->dev.bus_id, partno); else snprintf(p->dev.bus_id, BUS_ID_SIZE, - "%s%d", disk->dev.bus_id, part); + "%s%d", disk->dev.bus_id, partno); device_initialize(&p->dev); - p->dev.devt = MKDEV(disk->major, disk->first_minor + part); + p->dev.devt = MKDEV(disk->major, disk->first_minor + partno); p->dev.class = &block_class; p->dev.type = &part_type; p->dev.parent = &disk->dev; @@ -386,7 +387,7 @@ int add_partition(struct gendisk *disk, int part, sector_t start, sector_t len, } /* everything is up and running, commence */ - disk->part[part - 1] = p; + disk->part[partno - 1] = p; /* suppress uevent if the disk supresses it */ if (!disk->dev.uevent_suppress) diff --git a/include/linux/genhd.h b/include/linux/genhd.h index c64e659c984..d1723c0a860 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -365,7 +365,7 @@ extern int get_blkdev_list(char *, int); extern void add_disk(struct gendisk *disk); extern void del_gendisk(struct gendisk *gp); extern void unlink_gendisk(struct gendisk *gp); -extern struct gendisk *get_gendisk(dev_t dev, int *part); +extern struct gendisk *get_gendisk(dev_t dev, int *partno); extern void set_device_ro(struct block_device *bdev, int flag); extern void set_disk_ro(struct gendisk *disk, int flag); @@ -534,8 +534,8 @@ struct unixware_disklabel { #define ADDPART_FLAG_RAID 1 #define ADDPART_FLAG_WHOLEDISK 2 -extern dev_t blk_lookup_devt(const char *name, int part); -extern char *disk_name (struct gendisk *hd, int part, char *buf); +extern dev_t blk_lookup_devt(const char *name, int partno); +extern char *disk_name (struct gendisk *hd, int partno, char *buf); extern int rescan_partitions(struct gendisk *disk, struct block_device *bdev); extern int __must_check add_partition(struct gendisk *, int, sector_t, sector_t, int); @@ -553,16 +553,16 @@ extern void blk_register_region(dev_t devt, unsigned long range, void *data); extern void blk_unregister_region(dev_t devt, unsigned long range); -static inline struct block_device *bdget_disk(struct gendisk *disk, int index) +static inline struct block_device *bdget_disk(struct gendisk *disk, int partno) { - return bdget(MKDEV(disk->major, disk->first_minor) + index); + return bdget(MKDEV(disk->major, disk->first_minor) + partno); } #else /* CONFIG_BLOCK */ static inline void printk_all_partitions(void) { } -static inline dev_t blk_lookup_devt(const char *name, int part) +static inline dev_t blk_lookup_devt(const char *name, int partno) { dev_t devt = MKDEV(0, 0); return devt; -- cgit v1.2.3-70-g09d2 From f331c0296f2a9fee0d396a70598b954062603015 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 3 Sep 2008 09:01:48 +0200 Subject: block: don't depend on consecutive minor space * Implement disk_devt() and part_devt() and use them to directly access devt instead of computing it from ->major and ->first_minor. Note that all references to ->major and ->first_minor outside of block layer is used to determine devt of the disk (the part0) and as ->major and ->first_minor will continue to represent devt for the disk, converting these users aren't strictly necessary. However, convert them for consistency. * Implement disk_max_parts() to avoid directly deferencing genhd->minors. * Update bdget_disk() such that it doesn't assume consecutive minor space. * Move devt computation from register_disk() to add_disk() and make it the only one (all other usages use the initially determined value). These changes clean up the code and will help disk->part dereference fix and extended block device numbers. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/genhd.c | 107 +++++++++++++++++++++++++----------- block/ioctl.c | 6 +- drivers/block/pktcdvd.c | 2 +- drivers/block/ps3disk.c | 2 +- drivers/char/random.c | 6 +- drivers/md/dm-ioctl.c | 4 +- drivers/md/dm-stripe.c | 4 +- drivers/md/dm.c | 7 ++- drivers/memstick/core/mspro_block.c | 2 +- drivers/mmc/card/block.c | 2 +- drivers/s390/block/dasd_proc.c | 3 +- drivers/s390/block/dcssblk.c | 4 +- drivers/scsi/sr.c | 2 +- fs/block_dev.c | 2 +- fs/partitions/check.c | 19 ++++--- include/linux/genhd.h | 27 +++++++-- 16 files changed, 132 insertions(+), 67 deletions(-) (limited to 'fs/partitions') diff --git a/block/genhd.c b/block/genhd.c index dc9ad4c171e..fa32d09fda2 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -186,13 +186,14 @@ void add_disk(struct gendisk *disk) int retval; disk->flags |= GENHD_FL_UP; - blk_register_region(MKDEV(disk->major, disk->first_minor), - disk->minors, NULL, exact_match, exact_lock, disk); + disk->dev.devt = MKDEV(disk->major, disk->first_minor); + blk_register_region(disk_devt(disk), disk->minors, NULL, + exact_match, exact_lock, disk); register_disk(disk); blk_register_queue(disk); bdi = &disk->queue->backing_dev_info; - bdi_register_dev(bdi, MKDEV(disk->major, disk->first_minor)); + bdi_register_dev(bdi, disk_devt(disk)); retval = sysfs_create_link(&disk->dev.kobj, &bdi->dev->kobj, "bdi"); WARN_ON(retval); } @@ -205,8 +206,7 @@ void unlink_gendisk(struct gendisk *disk) sysfs_remove_link(&disk->dev.kobj, "bdi"); bdi_unregister(&disk->queue->backing_dev_info); blk_unregister_queue(disk); - blk_unregister_region(MKDEV(disk->major, disk->first_minor), - disk->minors); + blk_unregister_region(disk_devt(disk), disk->minors); } /** @@ -225,6 +225,38 @@ struct gendisk *get_gendisk(dev_t devt, int *partno) return kobj ? dev_to_disk(dev) : NULL; } +/** + * bdget_disk - do bdget() by gendisk and partition number + * @disk: gendisk of interest + * @partno: partition number + * + * Find partition @partno from @disk, do bdget() on it. + * + * CONTEXT: + * Don't care. + * + * RETURNS: + * Resulting block_device on success, NULL on failure. + */ +extern struct block_device *bdget_disk(struct gendisk *disk, int partno) +{ + dev_t devt = MKDEV(0, 0); + + if (partno == 0) + devt = disk_devt(disk); + else { + struct hd_struct *part = disk->part[partno - 1]; + + if (part && part->nr_sects) + devt = part_devt(part); + } + + if (likely(devt != MKDEV(0, 0))) + return bdget(devt); + return NULL; +} +EXPORT_SYMBOL(bdget_disk); + /* * print a full list of all partitions - intended for places where the root * filesystem can't be mounted and thus to give the victim some idea of what @@ -255,7 +287,7 @@ void __init printk_all_partitions(void) * option takes. */ printk("%02x%02x %10llu %s", - disk->major, disk->first_minor, + MAJOR(disk_devt(disk)), MINOR(disk_devt(disk)), (unsigned long long)get_capacity(disk) >> 1, disk_name(disk, 0, buf)); if (disk->driverfs_dev != NULL && @@ -266,15 +298,15 @@ void __init printk_all_partitions(void) printk(" (driver?)\n"); /* now show the partitions */ - for (n = 0; n < disk->minors - 1; ++n) { - if (disk->part[n] == NULL) - continue; - if (disk->part[n]->nr_sects == 0) + for (n = 0; n < disk_max_parts(disk); ++n) { + struct hd_struct *part = disk->part[n]; + + if (!part || !part->nr_sects) continue; printk(" %02x%02x %10llu %s\n", - disk->major, n + 1 + disk->first_minor, - (unsigned long long)disk->part[n]->nr_sects >> 1, - disk_name(disk, n + 1, buf)); + MAJOR(part_devt(part)), MINOR(part_devt(part)), + (unsigned long long)part->nr_sects >> 1, + disk_name(disk, part->partno, buf)); } } class_dev_iter_exit(&iter); @@ -343,26 +375,27 @@ static int show_partition(struct seq_file *seqf, void *v) char buf[BDEVNAME_SIZE]; /* Don't show non-partitionable removeable devices or empty devices */ - if (!get_capacity(sgp) || - (sgp->minors == 1 && (sgp->flags & GENHD_FL_REMOVABLE))) + if (!get_capacity(sgp) || (!disk_max_parts(sgp) && + (sgp->flags & GENHD_FL_REMOVABLE))) return 0; if (sgp->flags & GENHD_FL_SUPPRESS_PARTITION_INFO) return 0; /* show the full disk and all non-0 size partitions of it */ seq_printf(seqf, "%4d %4d %10llu %s\n", - sgp->major, sgp->first_minor, + MAJOR(disk_devt(sgp)), MINOR(disk_devt(sgp)), (unsigned long long)get_capacity(sgp) >> 1, disk_name(sgp, 0, buf)); - for (n = 0; n < sgp->minors - 1; n++) { - if (!sgp->part[n]) + for (n = 0; n < disk_max_parts(sgp); n++) { + struct hd_struct *part = sgp->part[n]; + if (!part) continue; - if (sgp->part[n]->nr_sects == 0) + if (part->nr_sects == 0) continue; seq_printf(seqf, "%4d %4d %10llu %s\n", - sgp->major, n + 1 + sgp->first_minor, - (unsigned long long)sgp->part[n]->nr_sects >> 1 , - disk_name(sgp, n + 1, buf)); + MAJOR(part_devt(part)), MINOR(part_devt(part)), + (unsigned long long)part->nr_sects >> 1, + disk_name(sgp, part->partno, buf)); } return 0; @@ -578,7 +611,8 @@ static int diskstats_show(struct seq_file *seqf, void *v) disk_round_stats(gp); preempt_enable(); seq_printf(seqf, "%4d %4d %s %lu %lu %llu %u %lu %lu %llu %u %u %u %u\n", - gp->major, gp->first_minor, disk_name(gp, 0, buf), + MAJOR(disk_devt(gp)), MINOR(disk_devt(gp)), + disk_name(gp, 0, buf), disk_stat_read(gp, ios[0]), disk_stat_read(gp, merges[0]), (unsigned long long)disk_stat_read(gp, sectors[0]), jiffies_to_msecs(disk_stat_read(gp, ticks[0])), @@ -590,7 +624,7 @@ static int diskstats_show(struct seq_file *seqf, void *v) jiffies_to_msecs(disk_stat_read(gp, time_in_queue))); /* now show all non-0 size partitions of it */ - for (n = 0; n < gp->minors - 1; n++) { + for (n = 0; n < disk_max_parts(gp); n++) { struct hd_struct *hd = gp->part[n]; if (!hd || !hd->nr_sects) @@ -601,8 +635,8 @@ static int diskstats_show(struct seq_file *seqf, void *v) preempt_enable(); seq_printf(seqf, "%4d %4d %s %lu %lu %llu " "%u %lu %lu %llu %u %u %u %u\n", - gp->major, n + gp->first_minor + 1, - disk_name(gp, n + 1, buf), + MAJOR(part_devt(hd)), MINOR(part_devt(hd)), + disk_name(gp, hd->partno, buf), part_stat_read(hd, ios[0]), part_stat_read(hd, merges[0]), (unsigned long long)part_stat_read(hd, sectors[0]), @@ -661,11 +695,22 @@ dev_t blk_lookup_devt(const char *name, int partno) while ((dev = class_dev_iter_next(&iter))) { struct gendisk *disk = dev_to_disk(dev); - if (!strcmp(dev->bus_id, name) && partno < disk->minors) { - devt = MKDEV(MAJOR(dev->devt), - MINOR(dev->devt) + partno); - break; + if (strcmp(dev->bus_id, name)) + continue; + if (partno < 0 || partno > disk_max_parts(disk)) + continue; + + if (partno == 0) + devt = disk_devt(disk); + else { + struct hd_struct *part = disk->part[partno - 1]; + + if (!part || !part->nr_sects) + continue; + + devt = part_devt(part); } + break; } class_dev_iter_exit(&iter); return devt; @@ -755,7 +800,7 @@ void set_disk_ro(struct gendisk *disk, int flag) { int i; disk->policy = flag; - for (i = 0; i < disk->minors - 1; i++) + for (i = 0; i < disk_max_parts(disk); i++) if (disk->part[i]) disk->part[i]->policy = flag; } diff --git a/block/ioctl.c b/block/ioctl.c index d77f5e280a6..403f7d7e0c2 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -29,7 +29,7 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user if (bdev != bdev->bd_contains) return -EINVAL; partno = p.pno; - if (partno <= 0 || partno >= disk->minors) + if (partno <= 0 || partno > disk_max_parts(disk)) return -EINVAL; switch (a.op) { case BLKPG_ADD_PARTITION: @@ -47,7 +47,7 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user mutex_lock(&bdev->bd_mutex); /* overlap? */ - for (i = 0; i < disk->minors - 1; i++) { + for (i = 0; i < disk_max_parts(disk); i++) { struct hd_struct *s = disk->part[i]; if (!s) @@ -96,7 +96,7 @@ static int blkdev_reread_part(struct block_device *bdev) struct gendisk *disk = bdev->bd_disk; int res; - if (disk->minors == 1 || bdev != bdev->bd_contains) + if (!disk_max_parts(disk) || bdev != bdev->bd_contains) return -EINVAL; if (!capable(CAP_SYS_ADMIN)) return -EACCES; diff --git a/drivers/block/pktcdvd.c b/drivers/block/pktcdvd.c index 29b7a648cc6..e1a90bbb474 100644 --- a/drivers/block/pktcdvd.c +++ b/drivers/block/pktcdvd.c @@ -2911,7 +2911,7 @@ static int pkt_setup_dev(dev_t dev, dev_t* pkt_dev) if (!disk->queue) goto out_mem2; - pd->pkt_dev = MKDEV(disk->major, disk->first_minor); + pd->pkt_dev = MKDEV(pktdev_major, idx); ret = pkt_new_dev(pd, dev); if (ret) goto out_new_dev; diff --git a/drivers/block/ps3disk.c b/drivers/block/ps3disk.c index 4b0d6c7f4c6..936466f62af 100644 --- a/drivers/block/ps3disk.c +++ b/drivers/block/ps3disk.c @@ -541,7 +541,7 @@ static int ps3disk_remove(struct ps3_system_bus_device *_dev) struct ps3disk_private *priv = dev->sbd.core.driver_data; mutex_lock(&ps3disk_mask_mutex); - __clear_bit(priv->gendisk->first_minor / PS3DISK_MINORS, + __clear_bit(MINOR(disk_devt(priv->gendisk)) / PS3DISK_MINORS, &ps3disk_mask); mutex_unlock(&ps3disk_mask_mutex); del_gendisk(priv->gendisk); diff --git a/drivers/char/random.c b/drivers/char/random.c index 7ce1ac4baa6..6af435b8986 100644 --- a/drivers/char/random.c +++ b/drivers/char/random.c @@ -661,10 +661,10 @@ void add_disk_randomness(struct gendisk *disk) if (!disk || !disk->random) return; /* first major is 1, so we get >= 0x200 here */ - DEBUG_ENT("disk event %d:%d\n", disk->major, disk->first_minor); + DEBUG_ENT("disk event %d:%d\n", + MAJOR(disk_devt(disk)), MINOR(disk_devt(disk))); - add_timer_randomness(disk->random, - 0x100 + MKDEV(disk->major, disk->first_minor)); + add_timer_randomness(disk->random, 0x100 + disk_devt(disk)); } #endif diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c index b262c0042de..c3de311117a 100644 --- a/drivers/md/dm-ioctl.c +++ b/drivers/md/dm-ioctl.c @@ -426,7 +426,7 @@ static int list_devices(struct dm_ioctl *param, size_t param_size) old_nl->next = (uint32_t) ((void *) nl - (void *) old_nl); disk = dm_disk(hc->md); - nl->dev = huge_encode_dev(MKDEV(disk->major, disk->first_minor)); + nl->dev = huge_encode_dev(disk_devt(disk)); nl->next = 0; strcpy(nl->name, hc->name); @@ -539,7 +539,7 @@ static int __dev_status(struct mapped_device *md, struct dm_ioctl *param) if (dm_suspended(md)) param->flags |= DM_SUSPEND_FLAG; - param->dev = huge_encode_dev(MKDEV(disk->major, disk->first_minor)); + param->dev = huge_encode_dev(disk_devt(disk)); /* * Yes, this will be out of date by the time it gets back diff --git a/drivers/md/dm-stripe.c b/drivers/md/dm-stripe.c index 4de90ab3968..b745d8ac625 100644 --- a/drivers/md/dm-stripe.c +++ b/drivers/md/dm-stripe.c @@ -284,8 +284,8 @@ static int stripe_end_io(struct dm_target *ti, struct bio *bio, memset(major_minor, 0, sizeof(major_minor)); sprintf(major_minor, "%d:%d", - bio->bi_bdev->bd_disk->major, - bio->bi_bdev->bd_disk->first_minor); + MAJOR(disk_devt(bio->bi_bdev->bd_disk)), + MINOR(disk_devt(bio->bi_bdev->bd_disk))); /* * Test to see which stripe drive triggered the event diff --git a/drivers/md/dm.c b/drivers/md/dm.c index ace998ce59f..a78caad2999 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -1146,7 +1146,7 @@ static void unlock_fs(struct mapped_device *md); static void free_dev(struct mapped_device *md) { - int minor = md->disk->first_minor; + int minor = MINOR(disk_devt(md->disk)); if (md->suspended_bdev) { unlock_fs(md); @@ -1267,7 +1267,7 @@ static struct mapped_device *dm_find_md(dev_t dev) md = idr_find(&_minor_idr, minor); if (md && (md == MINOR_ALLOCED || - (dm_disk(md)->first_minor != minor) || + (MINOR(disk_devt(dm_disk(md))) != minor) || test_bit(DMF_FREEING, &md->flags))) { md = NULL; goto out; @@ -1318,7 +1318,8 @@ void dm_put(struct mapped_device *md) if (atomic_dec_and_lock(&md->holders, &_minor_lock)) { map = dm_get_table(md); - idr_replace(&_minor_idr, MINOR_ALLOCED, dm_disk(md)->first_minor); + idr_replace(&_minor_idr, MINOR_ALLOCED, + MINOR(disk_devt(dm_disk(md)))); set_bit(DMF_FREEING, &md->flags); spin_unlock(&_minor_lock); if (!dm_suspended(md)) { diff --git a/drivers/memstick/core/mspro_block.c b/drivers/memstick/core/mspro_block.c index d2d2318dafa..82bf649ef13 100644 --- a/drivers/memstick/core/mspro_block.c +++ b/drivers/memstick/core/mspro_block.c @@ -197,7 +197,7 @@ static int mspro_block_bd_open(struct inode *inode, struct file *filp) static int mspro_block_disk_release(struct gendisk *disk) { struct mspro_block_data *msb = disk->private_data; - int disk_id = disk->first_minor >> MSPRO_BLOCK_PART_SHIFT; + int disk_id = MINOR(disk_devt(disk)) >> MSPRO_BLOCK_PART_SHIFT; mutex_lock(&mspro_block_disk_lock); diff --git a/drivers/mmc/card/block.c b/drivers/mmc/card/block.c index ebc8b9d7761..97156b689e8 100644 --- a/drivers/mmc/card/block.c +++ b/drivers/mmc/card/block.c @@ -83,7 +83,7 @@ static void mmc_blk_put(struct mmc_blk_data *md) mutex_lock(&open_lock); md->usage--; if (md->usage == 0) { - int devidx = md->disk->first_minor >> MMC_SHIFT; + int devidx = MINOR(disk_devt(md->disk)) >> MMC_SHIFT; __clear_bit(devidx, dev_use); put_disk(md->disk); diff --git a/drivers/s390/block/dasd_proc.c b/drivers/s390/block/dasd_proc.c index 03c0e40a92f..e3b5c4d3036 100644 --- a/drivers/s390/block/dasd_proc.c +++ b/drivers/s390/block/dasd_proc.c @@ -76,7 +76,8 @@ dasd_devices_show(struct seq_file *m, void *v) /* Print kdev. */ if (block->gdp) seq_printf(m, " at (%3d:%6d)", - block->gdp->major, block->gdp->first_minor); + MAJOR(disk_devt(block->gdp)), + MINOR(disk_devt(block->gdp))); else seq_printf(m, " at (???:??????)"); /* Print device name. */ diff --git a/drivers/s390/block/dcssblk.c b/drivers/s390/block/dcssblk.c index 711b3004b3e..9481e4a3f76 100644 --- a/drivers/s390/block/dcssblk.c +++ b/drivers/s390/block/dcssblk.c @@ -114,7 +114,7 @@ dcssblk_assign_free_minor(struct dcssblk_dev_info *dev_info) found = 0; // test if minor available list_for_each_entry(entry, &dcssblk_devices, lh) - if (minor == entry->gd->first_minor) + if (minor == MINOR(disk_devt(entry->gd))) found++; if (!found) break; // got unused minor } @@ -397,7 +397,7 @@ dcssblk_add_store(struct device *dev, struct device_attribute *attr, const char goto unload_seg; } sprintf(dev_info->gd->disk_name, "dcssblk%d", - dev_info->gd->first_minor); + MINOR(disk_devt(dev_info->gd))); list_add_tail(&dev_info->lh, &dcssblk_devices); if (!try_module_get(THIS_MODULE)) { diff --git a/drivers/scsi/sr.c b/drivers/scsi/sr.c index 27f5bfd1def..8dbe3798d5f 100644 --- a/drivers/scsi/sr.c +++ b/drivers/scsi/sr.c @@ -878,7 +878,7 @@ static void sr_kref_release(struct kref *kref) struct gendisk *disk = cd->disk; spin_lock(&sr_index_lock); - clear_bit(disk->first_minor, sr_index_bits); + clear_bit(MINOR(disk_devt(disk)), sr_index_bits); spin_unlock(&sr_index_lock); unregister_cdrom(&cd->cdi); diff --git a/fs/block_dev.c b/fs/block_dev.c index de0776cd721..72e0a2887cb 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -892,7 +892,7 @@ int check_disk_change(struct block_device *bdev) if (bdops->revalidate_disk) bdops->revalidate_disk(bdev->bd_disk); - if (bdev->bd_disk->minors > 1) + if (disk_max_parts(bdev->bd_disk)) bdev->bd_invalidated = 1; return 1; } diff --git a/fs/partitions/check.c b/fs/partitions/check.c index b86aab1b0df..e77fa144a07 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -134,7 +134,11 @@ char *disk_name(struct gendisk *hd, int partno, char *buf) const char *bdevname(struct block_device *bdev, char *buf) { - int partno = MINOR(bdev->bd_dev) - bdev->bd_disk->first_minor; + int partno = 0; + + if (bdev->bd_part) + partno = bdev->bd_part->partno; + return disk_name(bdev->bd_disk, partno, buf); } @@ -169,7 +173,7 @@ check_partition(struct gendisk *hd, struct block_device *bdev) if (isdigit(state->name[strlen(state->name)-1])) sprintf(state->name, "p"); - state->limit = hd->minors; + state->limit = disk_max_parts(hd) + 1; i = res = err = 0; while (!res && check_part[i]) { memset(&state->parts, 0, sizeof(state->parts)); @@ -416,7 +420,6 @@ void register_disk(struct gendisk *disk) int err; disk->dev.parent = disk->driverfs_dev; - disk->dev.devt = MKDEV(disk->major, disk->first_minor); strlcpy(disk->dev.bus_id, disk->disk_name, BUS_ID_SIZE); /* ewww... some of these buggers have / in the name... */ @@ -440,7 +443,7 @@ void register_disk(struct gendisk *disk) disk_sysfs_add_subdirs(disk); /* No minors to use for partitions */ - if (disk->minors == 1) + if (!disk_max_parts(disk)) goto exit; /* No such device (e.g., media were just removed) */ @@ -463,8 +466,8 @@ exit: kobject_uevent(&disk->dev.kobj, KOBJ_ADD); /* announce possible partitions */ - for (i = 1; i < disk->minors; i++) { - p = disk->part[i-1]; + for (i = 0; i < disk_max_parts(disk); i++) { + p = disk->part[i]; if (!p || !p->nr_sects) continue; kobject_uevent(&p->dev.kobj, KOBJ_ADD); @@ -482,7 +485,7 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) if (res) return res; bdev->bd_invalidated = 0; - for (p = 1; p < disk->minors; p++) + for (p = 1; p <= disk_max_parts(disk); p++) delete_partition(disk, p); if (disk->fops->revalidate_disk) disk->fops->revalidate_disk(disk); @@ -545,7 +548,7 @@ void del_gendisk(struct gendisk *disk) int p; /* invalidate stuff */ - for (p = disk->minors - 1; p > 0; p--) { + for (p = disk_max_parts(disk); p > 0; p--) { invalidate_partition(disk, p); delete_partition(disk, p); } diff --git a/include/linux/genhd.h b/include/linux/genhd.h index d1723c0a860..0ff75329199 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -111,10 +111,14 @@ struct hd_struct { #define GENHD_FL_FAIL 64 struct gendisk { + /* major, first_minor and minors are input parameters only, + * don't use directly. Use disk_devt() and disk_max_parts(). + */ int major; /* major number of driver */ int first_minor; int minors; /* maximum number of minors, =1 for * disks that can't be partitioned. */ + char disk_name[32]; /* name of major driver */ struct hd_struct **part; /* [indexed by minor - 1] */ struct block_device_operations *fops; @@ -152,6 +156,21 @@ static inline struct gendisk *part_to_disk(struct hd_struct *part) return NULL; } +static inline int disk_max_parts(struct gendisk *disk) +{ + return disk->minors - 1; +} + +static inline dev_t disk_devt(struct gendisk *disk) +{ + return disk->dev.devt; +} + +static inline dev_t part_devt(struct hd_struct *part) +{ + return part->dev.devt; +} + /* * Macros to operate on percpu disk statistics: * @@ -163,7 +182,7 @@ static inline struct hd_struct *disk_map_sector(struct gendisk *gendiskp, { struct hd_struct *part; int i; - for (i = 0; i < gendiskp->minors - 1; i++) { + for (i = 0; i < disk_max_parts(gendiskp); i++) { part = gendiskp->part[i]; if (part && part->start_sect <= sector && sector < part->start_sect + part->nr_sects) @@ -366,6 +385,7 @@ extern void add_disk(struct gendisk *disk); extern void del_gendisk(struct gendisk *gp); extern void unlink_gendisk(struct gendisk *gp); extern struct gendisk *get_gendisk(dev_t dev, int *partno); +extern struct block_device *bdget_disk(struct gendisk *disk, int partno); extern void set_device_ro(struct block_device *bdev, int flag); extern void set_disk_ro(struct gendisk *disk, int flag); @@ -553,11 +573,6 @@ extern void blk_register_region(dev_t devt, unsigned long range, void *data); extern void blk_unregister_region(dev_t devt, unsigned long range); -static inline struct block_device *bdget_disk(struct gendisk *disk, int partno) -{ - return bdget(MKDEV(disk->major, disk->first_minor) + partno); -} - #else /* CONFIG_BLOCK */ static inline void printk_all_partitions(void) { } -- cgit v1.2.3-70-g09d2 From e71bf0d0ee89e51b92776391c5634938236977d5 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 3 Sep 2008 09:03:02 +0200 Subject: block: fix disk->part[] dereferencing race disk->part[] is protected by its matching bdev's lock. However, non-critical accesses like collecting stats and printing out sysfs and proc information used to be performed without any locking. As partitions can come and go dynamically, partitions can go away underneath those non-critical accesses. As some of those accesses are writes, this theoretically can lead to silent corruption. This patch fixes the race by using RCU for the partition array and dev reference counter to hold partitions. * Rename disk->part[] to disk->__part[] to make sure no one outside genhd layer proper accesses it directly. * Use RCU for disk->__part[] dereferencing. * Implement disk_{get|put}_part() which can be used to get and put partitions from gendisk respectively. * Iterators are implemented to help iterate through all partitions safely. * Functions which require RCU readlock are marked with _rcu suffix. * Use disk_put_part() in __blkdev_put() instead of directly putting the contained kobject. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/blk-core.c | 20 ++++- block/blk-merge.c | 9 +- block/genhd.c | 218 ++++++++++++++++++++++++++++++++++++++------- block/ioctl.c | 26 +++--- drivers/block/aoe/aoecmd.c | 6 +- fs/block_dev.c | 15 ++-- fs/partitions/check.c | 70 +++++++++------ include/linux/genhd.h | 53 ++++++++--- 8 files changed, 323 insertions(+), 94 deletions(-) (limited to 'fs/partitions') diff --git a/block/blk-core.c b/block/blk-core.c index a0dc2e72fcb..d6128d9ad60 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -60,7 +60,9 @@ static void drive_stat_acct(struct request *rq, int new_io) if (!blk_fs_request(rq) || !rq->rq_disk) return; - part = disk_map_sector(rq->rq_disk, rq->sector); + rcu_read_lock(); + + part = disk_map_sector_rcu(rq->rq_disk, rq->sector); if (!new_io) __all_stat_inc(rq->rq_disk, part, merges[rw], rq->sector); else { @@ -71,6 +73,8 @@ static void drive_stat_acct(struct request *rq, int new_io) part->in_flight++; } } + + rcu_read_unlock(); } void blk_queue_congestion_threshold(struct request_queue *q) @@ -1557,12 +1561,14 @@ static int __end_that_request_first(struct request *req, int error, } if (blk_fs_request(req) && req->rq_disk) { - struct hd_struct *part = - disk_map_sector(req->rq_disk, req->sector); const int rw = rq_data_dir(req); + struct hd_struct *part; + rcu_read_lock(); + part = disk_map_sector_rcu(req->rq_disk, req->sector); all_stat_add(req->rq_disk, part, sectors[rw], nr_bytes >> 9, req->sector); + rcu_read_unlock(); } total_bytes = bio_nbytes = 0; @@ -1746,7 +1752,11 @@ static void end_that_request_last(struct request *req, int error) if (disk && blk_fs_request(req) && req != &req->q->bar_rq) { unsigned long duration = jiffies - req->start_time; const int rw = rq_data_dir(req); - struct hd_struct *part = disk_map_sector(disk, req->sector); + struct hd_struct *part; + + rcu_read_lock(); + + part = disk_map_sector_rcu(disk, req->sector); __all_stat_inc(disk, part, ios[rw], req->sector); __all_stat_add(disk, part, ticks[rw], duration, req->sector); @@ -1756,6 +1766,8 @@ static void end_that_request_last(struct request *req, int error) part_round_stats(part); part->in_flight--; } + + rcu_read_unlock(); } if (req->end_io) diff --git a/block/blk-merge.c b/block/blk-merge.c index 9b17da698d7..eb2a3ca5830 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -387,14 +387,19 @@ static int attempt_merge(struct request_queue *q, struct request *req, elv_merge_requests(q, req, next); if (req->rq_disk) { - struct hd_struct *part = - disk_map_sector(req->rq_disk, req->sector); + struct hd_struct *part; + + rcu_read_lock(); + + part = disk_map_sector_rcu(req->rq_disk, req->sector); disk_round_stats(req->rq_disk); req->rq_disk->in_flight--; if (part) { part_round_stats(part); part->in_flight--; } + + rcu_read_unlock(); } req->ioprio = ioprio_best(req->ioprio, next->ioprio); diff --git a/block/genhd.c b/block/genhd.c index fa32d09fda2..b431d654394 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -26,6 +26,158 @@ struct kobject *block_depr; static struct device_type disk_type; +/** + * disk_get_part - get partition + * @disk: disk to look partition from + * @partno: partition number + * + * Look for partition @partno from @disk. If found, increment + * reference count and return it. + * + * CONTEXT: + * Don't care. + * + * RETURNS: + * Pointer to the found partition on success, NULL if not found. + */ +struct hd_struct *disk_get_part(struct gendisk *disk, int partno) +{ + struct hd_struct *part; + + if (unlikely(partno < 1 || partno > disk_max_parts(disk))) + return NULL; + rcu_read_lock(); + part = rcu_dereference(disk->__part[partno - 1]); + if (part) + get_device(&part->dev); + rcu_read_unlock(); + + return part; +} +EXPORT_SYMBOL_GPL(disk_get_part); + +/** + * disk_part_iter_init - initialize partition iterator + * @piter: iterator to initialize + * @disk: disk to iterate over + * @flags: DISK_PITER_* flags + * + * Initialize @piter so that it iterates over partitions of @disk. + * + * CONTEXT: + * Don't care. + */ +void disk_part_iter_init(struct disk_part_iter *piter, struct gendisk *disk, + unsigned int flags) +{ + piter->disk = disk; + piter->part = NULL; + + if (flags & DISK_PITER_REVERSE) + piter->idx = disk_max_parts(piter->disk) - 1; + else + piter->idx = 0; + + piter->flags = flags; +} +EXPORT_SYMBOL_GPL(disk_part_iter_init); + +/** + * disk_part_iter_next - proceed iterator to the next partition and return it + * @piter: iterator of interest + * + * Proceed @piter to the next partition and return it. + * + * CONTEXT: + * Don't care. + */ +struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter) +{ + int inc, end; + + /* put the last partition */ + disk_put_part(piter->part); + piter->part = NULL; + + rcu_read_lock(); + + /* determine iteration parameters */ + if (piter->flags & DISK_PITER_REVERSE) { + inc = -1; + end = -1; + } else { + inc = 1; + end = disk_max_parts(piter->disk); + } + + /* iterate to the next partition */ + for (; piter->idx != end; piter->idx += inc) { + struct hd_struct *part; + + part = rcu_dereference(piter->disk->__part[piter->idx]); + if (!part) + continue; + if (!(piter->flags & DISK_PITER_INCL_EMPTY) && !part->nr_sects) + continue; + + get_device(&part->dev); + piter->part = part; + piter->idx += inc; + break; + } + + rcu_read_unlock(); + + return piter->part; +} +EXPORT_SYMBOL_GPL(disk_part_iter_next); + +/** + * disk_part_iter_exit - finish up partition iteration + * @piter: iter of interest + * + * Called when iteration is over. Cleans up @piter. + * + * CONTEXT: + * Don't care. + */ +void disk_part_iter_exit(struct disk_part_iter *piter) +{ + disk_put_part(piter->part); + piter->part = NULL; +} +EXPORT_SYMBOL_GPL(disk_part_iter_exit); + +/** + * disk_map_sector_rcu - map sector to partition + * @disk: gendisk of interest + * @sector: sector to map + * + * Find out which partition @sector maps to on @disk. This is + * primarily used for stats accounting. + * + * CONTEXT: + * RCU read locked. The returned partition pointer is valid only + * while preemption is disabled. + * + * RETURNS: + * Found partition on success, NULL if there's no matching partition. + */ +struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, sector_t sector) +{ + int i; + + for (i = 0; i < disk_max_parts(disk); i++) { + struct hd_struct *part = rcu_dereference(disk->__part[i]); + + if (part && part->start_sect <= sector && + sector < part->start_sect + part->nr_sects) + return part; + } + return NULL; +} +EXPORT_SYMBOL_GPL(disk_map_sector_rcu); + /* * Can be deleted altogether. Later. * @@ -245,10 +397,12 @@ extern struct block_device *bdget_disk(struct gendisk *disk, int partno) if (partno == 0) devt = disk_devt(disk); else { - struct hd_struct *part = disk->part[partno - 1]; + struct hd_struct *part; + part = disk_get_part(disk, partno); if (part && part->nr_sects) devt = part_devt(part); + disk_put_part(part); } if (likely(devt != MKDEV(0, 0))) @@ -270,8 +424,9 @@ void __init printk_all_partitions(void) class_dev_iter_init(&iter, &block_class, NULL, &disk_type); while ((dev = class_dev_iter_next(&iter))) { struct gendisk *disk = dev_to_disk(dev); + struct disk_part_iter piter; + struct hd_struct *part; char buf[BDEVNAME_SIZE]; - int n; /* * Don't show empty devices or things that have been @@ -298,16 +453,13 @@ void __init printk_all_partitions(void) printk(" (driver?)\n"); /* now show the partitions */ - for (n = 0; n < disk_max_parts(disk); ++n) { - struct hd_struct *part = disk->part[n]; - - if (!part || !part->nr_sects) - continue; + disk_part_iter_init(&piter, disk, 0); + while ((part = disk_part_iter_next(&piter))) printk(" %02x%02x %10llu %s\n", MAJOR(part_devt(part)), MINOR(part_devt(part)), (unsigned long long)part->nr_sects >> 1, disk_name(disk, part->partno, buf)); - } + disk_part_iter_exit(&piter); } class_dev_iter_exit(&iter); } @@ -371,7 +523,8 @@ static void *show_partition_start(struct seq_file *seqf, loff_t *pos) static int show_partition(struct seq_file *seqf, void *v) { struct gendisk *sgp = v; - int n; + struct disk_part_iter piter; + struct hd_struct *part; char buf[BDEVNAME_SIZE]; /* Don't show non-partitionable removeable devices or empty devices */ @@ -386,17 +539,14 @@ static int show_partition(struct seq_file *seqf, void *v) MAJOR(disk_devt(sgp)), MINOR(disk_devt(sgp)), (unsigned long long)get_capacity(sgp) >> 1, disk_name(sgp, 0, buf)); - for (n = 0; n < disk_max_parts(sgp); n++) { - struct hd_struct *part = sgp->part[n]; - if (!part) - continue; - if (part->nr_sects == 0) - continue; + + disk_part_iter_init(&piter, sgp, 0); + while ((part = disk_part_iter_next(&piter))) seq_printf(seqf, "%4d %4d %10llu %s\n", MAJOR(part_devt(part)), MINOR(part_devt(part)), (unsigned long long)part->nr_sects >> 1, disk_name(sgp, part->partno, buf)); - } + disk_part_iter_exit(&piter); return 0; } @@ -571,7 +721,7 @@ static void disk_release(struct device *dev) struct gendisk *disk = dev_to_disk(dev); kfree(disk->random); - kfree(disk->part); + kfree(disk->__part); free_disk_stats(disk); kfree(disk); } @@ -596,8 +746,9 @@ static struct device_type disk_type = { static int diskstats_show(struct seq_file *seqf, void *v) { struct gendisk *gp = v; + struct disk_part_iter piter; + struct hd_struct *hd; char buf[BDEVNAME_SIZE]; - int n; /* if (&gp->dev.kobj.entry == block_class.devices.next) @@ -624,12 +775,8 @@ static int diskstats_show(struct seq_file *seqf, void *v) jiffies_to_msecs(disk_stat_read(gp, time_in_queue))); /* now show all non-0 size partitions of it */ - for (n = 0; n < disk_max_parts(gp); n++) { - struct hd_struct *hd = gp->part[n]; - - if (!hd || !hd->nr_sects) - continue; - + disk_part_iter_init(&piter, gp, 0); + while ((hd = disk_part_iter_next(&piter))) { preempt_disable(); part_round_stats(hd); preempt_enable(); @@ -650,6 +797,7 @@ static int diskstats_show(struct seq_file *seqf, void *v) jiffies_to_msecs(part_stat_read(hd, time_in_queue)) ); } + disk_part_iter_exit(&piter); return 0; } @@ -703,12 +851,16 @@ dev_t blk_lookup_devt(const char *name, int partno) if (partno == 0) devt = disk_devt(disk); else { - struct hd_struct *part = disk->part[partno - 1]; + struct hd_struct *part; - if (!part || !part->nr_sects) + part = disk_get_part(disk, partno); + if (!part || !part->nr_sects) { + disk_put_part(part); continue; + } devt = part_devt(part); + disk_put_part(part); } break; } @@ -735,9 +887,9 @@ struct gendisk *alloc_disk_node(int minors, int node_id) } if (minors > 1) { int size = (minors - 1) * sizeof(struct hd_struct *); - disk->part = kmalloc_node(size, + disk->__part = kmalloc_node(size, GFP_KERNEL | __GFP_ZERO, node_id); - if (!disk->part) { + if (!disk->__part) { free_disk_stats(disk); kfree(disk); return NULL; @@ -798,10 +950,14 @@ EXPORT_SYMBOL(set_device_ro); void set_disk_ro(struct gendisk *disk, int flag) { - int i; + struct disk_part_iter piter; + struct hd_struct *part; + disk->policy = flag; - for (i = 0; i < disk_max_parts(disk); i++) - if (disk->part[i]) disk->part[i]->policy = flag; + disk_part_iter_init(&piter, disk, DISK_PITER_INCL_EMPTY); + while ((part = disk_part_iter_next(&piter))) + part->policy = flag; + disk_part_iter_exit(&piter); } EXPORT_SYMBOL(set_disk_ro); diff --git a/block/ioctl.c b/block/ioctl.c index 403f7d7e0c2..a5f672ad55f 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -12,11 +12,12 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user { struct block_device *bdevp; struct gendisk *disk; + struct hd_struct *part; struct blkpg_ioctl_arg a; struct blkpg_partition p; + struct disk_part_iter piter; long long start, length; int partno; - int i; int err; if (!capable(CAP_SYS_ADMIN)) @@ -47,28 +48,33 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user mutex_lock(&bdev->bd_mutex); /* overlap? */ - for (i = 0; i < disk_max_parts(disk); i++) { - struct hd_struct *s = disk->part[i]; - - if (!s) - continue; - if (!(start+length <= s->start_sect || - start >= s->start_sect + s->nr_sects)) { + disk_part_iter_init(&piter, disk, + DISK_PITER_INCL_EMPTY); + while ((part = disk_part_iter_next(&piter))) { + if (!(start + length <= part->start_sect || + start >= part->start_sect + part->nr_sects)) { + disk_part_iter_exit(&piter); mutex_unlock(&bdev->bd_mutex); return -EBUSY; } } + disk_part_iter_exit(&piter); + /* all seems OK */ err = add_partition(disk, partno, start, length, ADDPART_FLAG_NONE); mutex_unlock(&bdev->bd_mutex); return err; case BLKPG_DEL_PARTITION: - if (!disk->part[partno - 1]) + part = disk_get_part(disk, partno); + if (!part) return -ENXIO; - bdevp = bdget_disk(disk, partno); + + bdevp = bdget(part_devt(part)); + disk_put_part(part); if (!bdevp) return -ENOMEM; + mutex_lock(&bdevp->bd_mutex); if (bdevp->bd_openers) { mutex_unlock(&bdevp->bd_mutex); diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c index 885d1409521..84c03d65dcc 100644 --- a/drivers/block/aoe/aoecmd.c +++ b/drivers/block/aoe/aoecmd.c @@ -757,11 +757,15 @@ diskstats(struct gendisk *disk, struct bio *bio, ulong duration, sector_t sector const int rw = bio_data_dir(bio); struct hd_struct *part; - part = disk_map_sector(disk, sector); + rcu_read_lock(); + + part = disk_map_sector_rcu(disk, sector); all_stat_inc(disk, part, ios[rw], sector); all_stat_add(disk, part, ticks[rw], duration, sector); all_stat_add(disk, part, sectors[rw], n_sect, sector); all_stat_add(disk, part, io_ticks, duration, sector); + + rcu_read_unlock(); } void diff --git a/fs/block_dev.c b/fs/block_dev.c index 72e0a2887cb..2f2873b9a04 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -929,6 +929,7 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) { struct module *owner = NULL; struct gendisk *disk; + struct hd_struct *part = NULL; int ret; int partno; int perm = 0; @@ -978,7 +979,6 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) if (bdev->bd_invalidated) rescan_partitions(disk, bdev); } else { - struct hd_struct *p; struct block_device *whole; whole = bdget_disk(disk, 0); ret = -ENOMEM; @@ -989,16 +989,16 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) if (ret) goto out_first; bdev->bd_contains = whole; - p = disk->part[partno - 1]; + part = disk_get_part(disk, partno); bdev->bd_inode->i_data.backing_dev_info = whole->bd_inode->i_data.backing_dev_info; - if (!(disk->flags & GENHD_FL_UP) || !p || !p->nr_sects) { + if (!(disk->flags & GENHD_FL_UP) || + !part || !part->nr_sects) { ret = -ENXIO; goto out_first; } - kobject_get(&p->dev.kobj); - bdev->bd_part = p; - bd_set_size(bdev, (loff_t) p->nr_sects << 9); + bdev->bd_part = part; + bd_set_size(bdev, (loff_t)part->nr_sects << 9); } } else { put_disk(disk); @@ -1027,6 +1027,7 @@ out_first: __blkdev_put(bdev->bd_contains, 1); bdev->bd_contains = NULL; put_disk(disk); + disk_put_part(part); module_put(owner); out: mutex_unlock(&bdev->bd_mutex); @@ -1119,7 +1120,7 @@ static int __blkdev_put(struct block_device *bdev, int for_part) module_put(owner); if (bdev->bd_contains != bdev) { - kobject_put(&bdev->bd_part->dev.kobj); + disk_put_part(bdev->bd_part); bdev->bd_part = NULL; } bdev->bd_disk = NULL; diff --git a/fs/partitions/check.c b/fs/partitions/check.c index e77fa144a07..96c8bf41e45 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -314,19 +314,29 @@ static inline void disk_sysfs_add_subdirs(struct gendisk *disk) kobject_put(k); } +static void delete_partition_rcu_cb(struct rcu_head *head) +{ + struct hd_struct *part = container_of(head, struct hd_struct, rcu_head); + + part->start_sect = 0; + part->nr_sects = 0; + part_stat_set_all(part, 0); + put_device(&part->dev); +} + void delete_partition(struct gendisk *disk, int partno) { - struct hd_struct *p = disk->part[partno - 1]; + struct hd_struct *part; - if (!p) + part = disk->__part[partno-1]; + if (!part) return; - disk->part[partno - 1] = NULL; - p->start_sect = 0; - p->nr_sects = 0; - part_stat_set_all(p, 0); - kobject_put(p->holder_dir); - device_del(&p->dev); - put_device(&p->dev); + + rcu_assign_pointer(disk->__part[partno-1], NULL); + kobject_put(part->holder_dir); + device_del(&part->dev); + + call_rcu(&part->rcu_head, delete_partition_rcu_cb); } static ssize_t whole_disk_show(struct device *dev, @@ -343,7 +353,7 @@ int add_partition(struct gendisk *disk, int partno, struct hd_struct *p; int err; - if (disk->part[partno - 1]) + if (disk->__part[partno - 1]) return -EBUSY; p = kzalloc(sizeof(*p), GFP_KERNEL); @@ -391,7 +401,8 @@ int add_partition(struct gendisk *disk, int partno, } /* everything is up and running, commence */ - disk->part[partno - 1] = p; + INIT_RCU_HEAD(&p->rcu_head); + rcu_assign_pointer(disk->__part[partno - 1], p); /* suppress uevent if the disk supresses it */ if (!disk->dev.uevent_suppress) @@ -414,9 +425,9 @@ out_put: void register_disk(struct gendisk *disk) { struct block_device *bdev; + struct disk_part_iter piter; + struct hd_struct *part; char *s; - int i; - struct hd_struct *p; int err; disk->dev.parent = disk->driverfs_dev; @@ -466,16 +477,16 @@ exit: kobject_uevent(&disk->dev.kobj, KOBJ_ADD); /* announce possible partitions */ - for (i = 0; i < disk_max_parts(disk); i++) { - p = disk->part[i]; - if (!p || !p->nr_sects) - continue; - kobject_uevent(&p->dev.kobj, KOBJ_ADD); - } + disk_part_iter_init(&piter, disk, 0); + while ((part = disk_part_iter_next(&piter))) + kobject_uevent(&part->dev.kobj, KOBJ_ADD); + disk_part_iter_exit(&piter); } int rescan_partitions(struct gendisk *disk, struct block_device *bdev) { + struct disk_part_iter piter; + struct hd_struct *part; struct parsed_partitions *state; int p, res; @@ -485,8 +496,12 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) if (res) return res; bdev->bd_invalidated = 0; - for (p = 1; p <= disk_max_parts(disk); p++) - delete_partition(disk, p); + + disk_part_iter_init(&piter, disk, DISK_PITER_INCL_EMPTY); + while ((part = disk_part_iter_next(&piter))) + delete_partition(disk, part->partno); + disk_part_iter_exit(&piter); + if (disk->fops->revalidate_disk) disk->fops->revalidate_disk(disk); if (!get_capacity(disk) || !(state = check_partition(disk, bdev))) @@ -545,13 +560,18 @@ EXPORT_SYMBOL(read_dev_sector); void del_gendisk(struct gendisk *disk) { - int p; + struct disk_part_iter piter; + struct hd_struct *part; /* invalidate stuff */ - for (p = disk_max_parts(disk); p > 0; p--) { - invalidate_partition(disk, p); - delete_partition(disk, p); + disk_part_iter_init(&piter, disk, + DISK_PITER_INCL_EMPTY | DISK_PITER_REVERSE); + while ((part = disk_part_iter_next(&piter))) { + invalidate_partition(disk, part->partno); + delete_partition(disk, part->partno); } + disk_part_iter_exit(&piter); + invalidate_partition(disk, 0); disk->capacity = 0; disk->flags &= ~GENHD_FL_UP; diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 0ff75329199..7fbba19e076 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -11,6 +11,7 @@ #include #include +#include #ifdef CONFIG_BLOCK @@ -100,6 +101,7 @@ struct hd_struct { #else struct disk_stats dkstats; #endif + struct rcu_head rcu_head; }; #define GENHD_FL_REMOVABLE 1 @@ -120,7 +122,14 @@ struct gendisk { * disks that can't be partitioned. */ char disk_name[32]; /* name of major driver */ - struct hd_struct **part; /* [indexed by minor - 1] */ + + /* Array of pointers to partitions indexed by partno - 1. + * Protected with matching bdev lock but stat and other + * non-critical accesses use RCU. Always access through + * helpers. + */ + struct hd_struct **__part; + struct block_device_operations *fops; struct request_queue *queue; void *private_data; @@ -171,25 +180,41 @@ static inline dev_t part_devt(struct hd_struct *part) return part->dev.devt; } +extern struct hd_struct *disk_get_part(struct gendisk *disk, int partno); + +static inline void disk_put_part(struct hd_struct *part) +{ + if (likely(part)) + put_device(&part->dev); +} + +/* + * Smarter partition iterator without context limits. + */ +#define DISK_PITER_REVERSE (1 << 0) /* iterate in the reverse direction */ +#define DISK_PITER_INCL_EMPTY (1 << 1) /* include 0-sized parts */ + +struct disk_part_iter { + struct gendisk *disk; + struct hd_struct *part; + int idx; + unsigned int flags; +}; + +extern void disk_part_iter_init(struct disk_part_iter *piter, + struct gendisk *disk, unsigned int flags); +extern struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter); +extern void disk_part_iter_exit(struct disk_part_iter *piter); + +extern struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, + sector_t sector); + /* * Macros to operate on percpu disk statistics: * * The __ variants should only be called in critical sections. The full * variants disable/enable preemption. */ -static inline struct hd_struct *disk_map_sector(struct gendisk *gendiskp, - sector_t sector) -{ - struct hd_struct *part; - int i; - for (i = 0; i < disk_max_parts(gendiskp); i++) { - part = gendiskp->part[i]; - if (part && part->start_sect <= sector - && sector < part->start_sect + part->nr_sects) - return part; - } - return NULL; -} #ifdef CONFIG_SMP #define __disk_stat_add(gendiskp, field, addnd) \ -- cgit v1.2.3-70-g09d2 From c9959059161ddd7bf4670cf47367033d6b2f79c4 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:47:21 +0900 Subject: block: fix diskstats access There are two variants of stat functions - ones prefixed with double underbars which don't care about preemption and ones without which disable preemption before manipulating per-cpu counters. It's unclear whether the underbarred ones assume that preemtion is disabled on entry as some callers don't do that. This patch unifies diskstats access by implementing disk_stat_lock() and disk_stat_unlock() which take care of both RCU (for partition access) and preemption (for per-cpu counter access). diskstats access should always be enclosed between the two functions. As such, there's no need for the versions which disables preemption. They're removed and double underbars ones are renamed to drop the underbars. As an extra argument is added, there's no danger of using the old version unconverted. disk_stat_lock() uses get_cpu() and returns the cpu index and all diskstat functions which access per-cpu counters now has @cpu argument to help RT. This change adds RCU or preemption operations at some places but also collapses several preemption ops into one at others. Overall, the performance difference should be negligible as all involved ops are very lightweight per-cpu ones. Signed-off-by: Tejun Heo Cc: Peter Zijlstra Signed-off-by: Jens Axboe --- block/blk-core.c | 52 +++++++++-------- block/blk-merge.c | 11 ++-- block/genhd.c | 20 ++++--- drivers/block/aoe/aoecmd.c | 15 ++--- drivers/md/dm.c | 26 +++++---- drivers/md/linear.c | 7 ++- drivers/md/multipath.c | 7 ++- drivers/md/raid0.c | 7 ++- drivers/md/raid1.c | 8 ++- drivers/md/raid10.c | 7 ++- drivers/md/raid5.c | 8 ++- fs/partitions/check.c | 7 ++- include/linux/genhd.h | 139 +++++++++++++++++++-------------------------- 13 files changed, 158 insertions(+), 156 deletions(-) (limited to 'fs/partitions') diff --git a/block/blk-core.c b/block/blk-core.c index d6128d9ad60..e0a5ee36849 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -56,25 +56,26 @@ static void drive_stat_acct(struct request *rq, int new_io) { struct hd_struct *part; int rw = rq_data_dir(rq); + int cpu; if (!blk_fs_request(rq) || !rq->rq_disk) return; - rcu_read_lock(); - + cpu = disk_stat_lock(); part = disk_map_sector_rcu(rq->rq_disk, rq->sector); + if (!new_io) - __all_stat_inc(rq->rq_disk, part, merges[rw], rq->sector); + all_stat_inc(cpu, rq->rq_disk, part, merges[rw], rq->sector); else { - disk_round_stats(rq->rq_disk); + disk_round_stats(cpu, rq->rq_disk); rq->rq_disk->in_flight++; if (part) { - part_round_stats(part); + part_round_stats(cpu, part); part->in_flight++; } } - rcu_read_unlock(); + disk_stat_unlock(); } void blk_queue_congestion_threshold(struct request_queue *q) @@ -997,7 +998,7 @@ static inline void add_request(struct request_queue *q, struct request *req) * /proc/diskstats. This accounts immediately for all queue usage up to * the current jiffies and restarts the counters again. */ -void disk_round_stats(struct gendisk *disk) +void disk_round_stats(int cpu, struct gendisk *disk) { unsigned long now = jiffies; @@ -1005,15 +1006,15 @@ void disk_round_stats(struct gendisk *disk) return; if (disk->in_flight) { - __disk_stat_add(disk, time_in_queue, - disk->in_flight * (now - disk->stamp)); - __disk_stat_add(disk, io_ticks, (now - disk->stamp)); + disk_stat_add(cpu, disk, time_in_queue, + disk->in_flight * (now - disk->stamp)); + disk_stat_add(cpu, disk, io_ticks, (now - disk->stamp)); } disk->stamp = now; } EXPORT_SYMBOL_GPL(disk_round_stats); -void part_round_stats(struct hd_struct *part) +void part_round_stats(int cpu, struct hd_struct *part) { unsigned long now = jiffies; @@ -1021,9 +1022,9 @@ void part_round_stats(struct hd_struct *part) return; if (part->in_flight) { - __part_stat_add(part, time_in_queue, - part->in_flight * (now - part->stamp)); - __part_stat_add(part, io_ticks, (now - part->stamp)); + part_stat_add(cpu, part, time_in_queue, + part->in_flight * (now - part->stamp)); + part_stat_add(cpu, part, io_ticks, (now - part->stamp)); } part->stamp = now; } @@ -1563,12 +1564,13 @@ static int __end_that_request_first(struct request *req, int error, if (blk_fs_request(req) && req->rq_disk) { const int rw = rq_data_dir(req); struct hd_struct *part; + int cpu; - rcu_read_lock(); + cpu = disk_stat_lock(); part = disk_map_sector_rcu(req->rq_disk, req->sector); - all_stat_add(req->rq_disk, part, sectors[rw], - nr_bytes >> 9, req->sector); - rcu_read_unlock(); + all_stat_add(cpu, req->rq_disk, part, sectors[rw], + nr_bytes >> 9, req->sector); + disk_stat_unlock(); } total_bytes = bio_nbytes = 0; @@ -1753,21 +1755,21 @@ static void end_that_request_last(struct request *req, int error) unsigned long duration = jiffies - req->start_time; const int rw = rq_data_dir(req); struct hd_struct *part; + int cpu; - rcu_read_lock(); - + cpu = disk_stat_lock(); part = disk_map_sector_rcu(disk, req->sector); - __all_stat_inc(disk, part, ios[rw], req->sector); - __all_stat_add(disk, part, ticks[rw], duration, req->sector); - disk_round_stats(disk); + all_stat_inc(cpu, disk, part, ios[rw], req->sector); + all_stat_add(cpu, disk, part, ticks[rw], duration, req->sector); + disk_round_stats(cpu, disk); disk->in_flight--; if (part) { - part_round_stats(part); + part_round_stats(cpu, part); part->in_flight--; } - rcu_read_unlock(); + disk_stat_unlock(); } if (req->end_io) diff --git a/block/blk-merge.c b/block/blk-merge.c index eb2a3ca5830..d926a24bf1f 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -388,18 +388,19 @@ static int attempt_merge(struct request_queue *q, struct request *req, if (req->rq_disk) { struct hd_struct *part; + int cpu; - rcu_read_lock(); - + cpu = disk_stat_lock(); part = disk_map_sector_rcu(req->rq_disk, req->sector); - disk_round_stats(req->rq_disk); + + disk_round_stats(cpu, req->rq_disk); req->rq_disk->in_flight--; if (part) { - part_round_stats(part); + part_round_stats(cpu, part); part->in_flight--; } - rcu_read_unlock(); + disk_stat_unlock(); } req->ioprio = ioprio_best(req->ioprio, next->ioprio); diff --git a/block/genhd.c b/block/genhd.c index b431d654394..430626e440f 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -633,10 +633,11 @@ static ssize_t disk_stat_show(struct device *dev, struct device_attribute *attr, char *buf) { struct gendisk *disk = dev_to_disk(dev); + int cpu; - preempt_disable(); - disk_round_stats(disk); - preempt_enable(); + cpu = disk_stat_lock(); + disk_round_stats(cpu, disk); + disk_stat_unlock(); return sprintf(buf, "%8lu %8lu %8llu %8u " "%8lu %8lu %8llu %8u " @@ -749,6 +750,7 @@ static int diskstats_show(struct seq_file *seqf, void *v) struct disk_part_iter piter; struct hd_struct *hd; char buf[BDEVNAME_SIZE]; + int cpu; /* if (&gp->dev.kobj.entry == block_class.devices.next) @@ -758,9 +760,9 @@ static int diskstats_show(struct seq_file *seqf, void *v) "\n\n"); */ - preempt_disable(); - disk_round_stats(gp); - preempt_enable(); + cpu = disk_stat_lock(); + disk_round_stats(cpu, gp); + disk_stat_unlock(); seq_printf(seqf, "%4d %4d %s %lu %lu %llu %u %lu %lu %llu %u %u %u %u\n", MAJOR(disk_devt(gp)), MINOR(disk_devt(gp)), disk_name(gp, 0, buf), @@ -777,9 +779,9 @@ static int diskstats_show(struct seq_file *seqf, void *v) /* now show all non-0 size partitions of it */ disk_part_iter_init(&piter, gp, 0); while ((hd = disk_part_iter_next(&piter))) { - preempt_disable(); - part_round_stats(hd); - preempt_enable(); + cpu = disk_stat_lock(); + part_round_stats(cpu, hd); + disk_stat_unlock(); seq_printf(seqf, "%4d %4d %s %lu %lu %llu " "%u %lu %lu %llu %u %u %u %u\n", MAJOR(part_devt(hd)), MINOR(part_devt(hd)), diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c index 84c03d65dcc..17eed8c025d 100644 --- a/drivers/block/aoe/aoecmd.c +++ b/drivers/block/aoe/aoecmd.c @@ -756,16 +756,17 @@ diskstats(struct gendisk *disk, struct bio *bio, ulong duration, sector_t sector unsigned long n_sect = bio->bi_size >> 9; const int rw = bio_data_dir(bio); struct hd_struct *part; + int cpu; - rcu_read_lock(); - + cpu = disk_stat_lock(); part = disk_map_sector_rcu(disk, sector); - all_stat_inc(disk, part, ios[rw], sector); - all_stat_add(disk, part, ticks[rw], duration, sector); - all_stat_add(disk, part, sectors[rw], n_sect, sector); - all_stat_add(disk, part, io_ticks, duration, sector); - rcu_read_unlock(); + all_stat_inc(cpu, disk, part, ios[rw], sector); + all_stat_add(cpu, disk, part, ticks[rw], duration, sector); + all_stat_add(cpu, disk, part, sectors[rw], n_sect, sector); + all_stat_add(cpu, disk, part, io_ticks, duration, sector); + + disk_stat_unlock(); } void diff --git a/drivers/md/dm.c b/drivers/md/dm.c index a78caad2999..653624792ea 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -377,12 +377,13 @@ static void free_tio(struct mapped_device *md, struct dm_target_io *tio) static void start_io_acct(struct dm_io *io) { struct mapped_device *md = io->md; + int cpu; io->start_time = jiffies; - preempt_disable(); - disk_round_stats(dm_disk(md)); - preempt_enable(); + cpu = disk_stat_lock(); + disk_round_stats(cpu, dm_disk(md)); + disk_stat_unlock(); dm_disk(md)->in_flight = atomic_inc_return(&md->pending); } @@ -391,15 +392,15 @@ static int end_io_acct(struct dm_io *io) struct mapped_device *md = io->md; struct bio *bio = io->bio; unsigned long duration = jiffies - io->start_time; - int pending; + int pending, cpu; int rw = bio_data_dir(bio); - preempt_disable(); - disk_round_stats(dm_disk(md)); - preempt_enable(); - dm_disk(md)->in_flight = pending = atomic_dec_return(&md->pending); + cpu = disk_stat_lock(); + disk_round_stats(cpu, dm_disk(md)); + disk_stat_add(cpu, dm_disk(md), ticks[rw], duration); + disk_stat_unlock(); - disk_stat_add(dm_disk(md), ticks[rw], duration); + dm_disk(md)->in_flight = pending = atomic_dec_return(&md->pending); return !pending; } @@ -885,6 +886,7 @@ static int dm_request(struct request_queue *q, struct bio *bio) int r = -EIO; int rw = bio_data_dir(bio); struct mapped_device *md = q->queuedata; + int cpu; /* * There is no use in forwarding any barrier request since we can't @@ -897,8 +899,10 @@ static int dm_request(struct request_queue *q, struct bio *bio) down_read(&md->io_lock); - disk_stat_inc(dm_disk(md), ios[rw]); - disk_stat_add(dm_disk(md), sectors[rw], bio_sectors(bio)); + cpu = disk_stat_lock(); + disk_stat_inc(cpu, dm_disk(md), ios[rw]); + disk_stat_add(cpu, dm_disk(md), sectors[rw], bio_sectors(bio)); + disk_stat_unlock(); /* * If we're suspended we have to queue diff --git a/drivers/md/linear.c b/drivers/md/linear.c index b1eebf88c20..00cbc8e4729 100644 --- a/drivers/md/linear.c +++ b/drivers/md/linear.c @@ -318,14 +318,17 @@ static int linear_make_request (struct request_queue *q, struct bio *bio) mddev_t *mddev = q->queuedata; dev_info_t *tmp_dev; sector_t block; + int cpu; if (unlikely(bio_barrier(bio))) { bio_endio(bio, -EOPNOTSUPP); return 0; } - disk_stat_inc(mddev->gendisk, ios[rw]); - disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio)); + cpu = disk_stat_lock(); + disk_stat_inc(cpu, mddev->gendisk, ios[rw]); + disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); + disk_stat_unlock(); tmp_dev = which_dev(mddev, bio->bi_sector); block = bio->bi_sector >> 1; diff --git a/drivers/md/multipath.c b/drivers/md/multipath.c index c4779ccba1c..182f5a94cdc 100644 --- a/drivers/md/multipath.c +++ b/drivers/md/multipath.c @@ -147,6 +147,7 @@ static int multipath_make_request (struct request_queue *q, struct bio * bio) struct multipath_bh * mp_bh; struct multipath_info *multipath; const int rw = bio_data_dir(bio); + int cpu; if (unlikely(bio_barrier(bio))) { bio_endio(bio, -EOPNOTSUPP); @@ -158,8 +159,10 @@ static int multipath_make_request (struct request_queue *q, struct bio * bio) mp_bh->master_bio = bio; mp_bh->mddev = mddev; - disk_stat_inc(mddev->gendisk, ios[rw]); - disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio)); + cpu = disk_stat_lock(); + disk_stat_inc(cpu, mddev->gendisk, ios[rw]); + disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); + disk_stat_unlock(); mp_bh->path = multipath_map(conf); if (mp_bh->path < 0) { diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c index 18361063566..e26030fa59a 100644 --- a/drivers/md/raid0.c +++ b/drivers/md/raid0.c @@ -399,14 +399,17 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio) sector_t chunk; sector_t block, rsect; const int rw = bio_data_dir(bio); + int cpu; if (unlikely(bio_barrier(bio))) { bio_endio(bio, -EOPNOTSUPP); return 0; } - disk_stat_inc(mddev->gendisk, ios[rw]); - disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio)); + cpu = disk_stat_lock(); + disk_stat_inc(cpu, mddev->gendisk, ios[rw]); + disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); + disk_stat_unlock(); chunk_size = mddev->chunk_size >> 10; chunk_sects = mddev->chunk_size >> 9; diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index 0b82030c265..babb13036f9 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -779,7 +779,7 @@ static int make_request(struct request_queue *q, struct bio * bio) struct page **behind_pages = NULL; const int rw = bio_data_dir(bio); const int do_sync = bio_sync(bio); - int do_barriers; + int cpu, do_barriers; mdk_rdev_t *blocked_rdev; /* @@ -804,8 +804,10 @@ static int make_request(struct request_queue *q, struct bio * bio) bitmap = mddev->bitmap; - disk_stat_inc(mddev->gendisk, ios[rw]); - disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio)); + cpu = disk_stat_lock(); + disk_stat_inc(cpu, mddev->gendisk, ios[rw]); + disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); + disk_stat_unlock(); /* * make_request() can abort the operation when READA is being diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index d3b9aa09628..5ec80da0a9d 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -789,6 +789,7 @@ static int make_request(struct request_queue *q, struct bio * bio) mirror_info_t *mirror; r10bio_t *r10_bio; struct bio *read_bio; + int cpu; int i; int chunk_sects = conf->chunk_mask + 1; const int rw = bio_data_dir(bio); @@ -843,8 +844,10 @@ static int make_request(struct request_queue *q, struct bio * bio) */ wait_barrier(conf); - disk_stat_inc(mddev->gendisk, ios[rw]); - disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio)); + cpu = disk_stat_lock(); + disk_stat_inc(cpu, mddev->gendisk, ios[rw]); + disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); + disk_stat_unlock(); r10_bio = mempool_alloc(conf->r10bio_pool, GFP_NOIO); diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 37e546528f9..5899f211515 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -3387,7 +3387,7 @@ static int make_request(struct request_queue *q, struct bio * bi) sector_t logical_sector, last_sector; struct stripe_head *sh; const int rw = bio_data_dir(bi); - int remaining; + int cpu, remaining; if (unlikely(bio_barrier(bi))) { bio_endio(bi, -EOPNOTSUPP); @@ -3396,8 +3396,10 @@ static int make_request(struct request_queue *q, struct bio * bi) md_write_start(mddev, bi); - disk_stat_inc(mddev->gendisk, ios[rw]); - disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bi)); + cpu = disk_stat_lock(); + disk_stat_inc(cpu, mddev->gendisk, ios[rw]); + disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bi)); + disk_stat_unlock(); if (rw == READ && mddev->reshape_position == MaxSector && diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 96c8bf41e45..c442f0aadac 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -219,10 +219,11 @@ static ssize_t part_stat_show(struct device *dev, struct device_attribute *attr, char *buf) { struct hd_struct *p = dev_to_part(dev); + int cpu; - preempt_disable(); - part_round_stats(p); - preempt_enable(); + cpu = disk_stat_lock(); + part_round_stats(cpu, p); + disk_stat_unlock(); return sprintf(buf, "%8lu %8lu %8llu %8u " "%8lu %8lu %8llu %8u " diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 7fbba19e076..ac8a901f200 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -209,16 +209,24 @@ extern void disk_part_iter_exit(struct disk_part_iter *piter); extern struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, sector_t sector); -/* +/* * Macros to operate on percpu disk statistics: * - * The __ variants should only be called in critical sections. The full - * variants disable/enable preemption. + * {disk|part|all}_stat_{add|sub|inc|dec}() modify the stat counters + * and should be called between disk_stat_lock() and + * disk_stat_unlock(). + * + * part_stat_read() can be called at any time. + * + * part_stat_{add|set_all}() and {init|free}_part_stats are for + * internal use only. */ - #ifdef CONFIG_SMP -#define __disk_stat_add(gendiskp, field, addnd) \ - (per_cpu_ptr(gendiskp->dkstats, smp_processor_id())->field += addnd) +#define disk_stat_lock() ({ rcu_read_lock(); get_cpu(); }) +#define disk_stat_unlock() do { put_cpu(); rcu_read_unlock(); } while (0) + +#define disk_stat_add(cpu, gendiskp, field, addnd) \ + (per_cpu_ptr(gendiskp->dkstats, cpu)->field += addnd) #define disk_stat_read(gendiskp, field) \ ({ \ @@ -229,7 +237,8 @@ extern struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, res; \ }) -static inline void disk_stat_set_all(struct gendisk *gendiskp, int value) { +static inline void disk_stat_set_all(struct gendisk *gendiskp, int value) +{ int i; for_each_possible_cpu(i) @@ -237,14 +246,14 @@ static inline void disk_stat_set_all(struct gendisk *gendiskp, int value) { sizeof(struct disk_stats)); } -#define __part_stat_add(part, field, addnd) \ - (per_cpu_ptr(part->dkstats, smp_processor_id())->field += addnd) +#define part_stat_add(cpu, part, field, addnd) \ + (per_cpu_ptr(part->dkstats, cpu)->field += addnd) -#define __all_stat_add(gendiskp, part, field, addnd, sector) \ -({ \ - if (part) \ - __part_stat_add(part, field, addnd); \ - __disk_stat_add(gendiskp, field, addnd); \ +#define all_stat_add(cpu, gendiskp, part, field, addnd, sector) \ +({ \ + if (part) \ + part_stat_add(cpu, part, field, addnd); \ + disk_stat_add(cpu, gendiskp, field, addnd); \ }) #define part_stat_read(part, field) \ @@ -264,10 +273,13 @@ static inline void part_stat_set_all(struct hd_struct *part, int value) memset(per_cpu_ptr(part->dkstats, i), value, sizeof(struct disk_stats)); } - + #else /* !CONFIG_SMP */ -#define __disk_stat_add(gendiskp, field, addnd) \ - (gendiskp->dkstats.field += addnd) +#define disk_stat_lock() ({ rcu_read_lock(); 0; }) +#define disk_stat_unlock() rcu_read_unlock() + +#define disk_stat_add(cpu, gendiskp, field, addnd) \ + (gendiskp->dkstats.field += addnd) #define disk_stat_read(gendiskp, field) (gendiskp->dkstats.field) static inline void disk_stat_set_all(struct gendisk *gendiskp, int value) @@ -275,14 +287,14 @@ static inline void disk_stat_set_all(struct gendisk *gendiskp, int value) memset(&gendiskp->dkstats, value, sizeof (struct disk_stats)); } -#define __part_stat_add(part, field, addnd) \ +#define part_stat_add(cpu, part, field, addnd) \ (part->dkstats.field += addnd) -#define __all_stat_add(gendiskp, part, field, addnd, sector) \ -({ \ - if (part) \ - part->dkstats.field += addnd; \ - __disk_stat_add(gendiskp, field, addnd); \ +#define all_stat_add(cpu, gendiskp, part, field, addnd, sector) \ +({ \ + if (part) \ + part_stat_add(cpu, part, field, addnd); \ + disk_stat_add(cpu, gendiskp, field, addnd); \ }) #define part_stat_read(part, field) (part->dkstats.field) @@ -294,63 +306,26 @@ static inline void part_stat_set_all(struct hd_struct *part, int value) #endif /* CONFIG_SMP */ -#define disk_stat_add(gendiskp, field, addnd) \ - do { \ - preempt_disable(); \ - __disk_stat_add(gendiskp, field, addnd); \ - preempt_enable(); \ - } while (0) - -#define __disk_stat_dec(gendiskp, field) __disk_stat_add(gendiskp, field, -1) -#define disk_stat_dec(gendiskp, field) disk_stat_add(gendiskp, field, -1) - -#define __disk_stat_inc(gendiskp, field) __disk_stat_add(gendiskp, field, 1) -#define disk_stat_inc(gendiskp, field) disk_stat_add(gendiskp, field, 1) - -#define __disk_stat_sub(gendiskp, field, subnd) \ - __disk_stat_add(gendiskp, field, -subnd) -#define disk_stat_sub(gendiskp, field, subnd) \ - disk_stat_add(gendiskp, field, -subnd) - -#define part_stat_add(gendiskp, field, addnd) \ - do { \ - preempt_disable(); \ - __part_stat_add(gendiskp, field, addnd);\ - preempt_enable(); \ - } while (0) - -#define __part_stat_dec(gendiskp, field) __part_stat_add(gendiskp, field, -1) -#define part_stat_dec(gendiskp, field) part_stat_add(gendiskp, field, -1) - -#define __part_stat_inc(gendiskp, field) __part_stat_add(gendiskp, field, 1) -#define part_stat_inc(gendiskp, field) part_stat_add(gendiskp, field, 1) - -#define __part_stat_sub(gendiskp, field, subnd) \ - __part_stat_add(gendiskp, field, -subnd) -#define part_stat_sub(gendiskp, field, subnd) \ - part_stat_add(gendiskp, field, -subnd) - -#define all_stat_add(gendiskp, part, field, addnd, sector) \ - do { \ - preempt_disable(); \ - __all_stat_add(gendiskp, part, field, addnd, sector); \ - preempt_enable(); \ - } while (0) - -#define __all_stat_dec(gendiskp, field, sector) \ - __all_stat_add(gendiskp, field, -1, sector) -#define all_stat_dec(gendiskp, field, sector) \ - all_stat_add(gendiskp, field, -1, sector) - -#define __all_stat_inc(gendiskp, part, field, sector) \ - __all_stat_add(gendiskp, part, field, 1, sector) -#define all_stat_inc(gendiskp, part, field, sector) \ - all_stat_add(gendiskp, part, field, 1, sector) - -#define __all_stat_sub(gendiskp, part, field, subnd, sector) \ - __all_stat_add(gendiskp, part, field, -subnd, sector) -#define all_stat_sub(gendiskp, part, field, subnd, sector) \ - all_stat_add(gendiskp, part, field, -subnd, sector) +#define disk_stat_dec(cpu, gendiskp, field) \ + disk_stat_add(cpu, gendiskp, field, -1) +#define disk_stat_inc(cpu, gendiskp, field) \ + disk_stat_add(cpu, gendiskp, field, 1) +#define disk_stat_sub(cpu, gendiskp, field, subnd) \ + disk_stat_add(cpu, gendiskp, field, -subnd) + +#define part_stat_dec(cpu, gendiskp, field) \ + part_stat_add(cpu, gendiskp, field, -1) +#define part_stat_inc(cpu, gendiskp, field) \ + part_stat_add(cpu, gendiskp, field, 1) +#define part_stat_sub(cpu, gendiskp, field, subnd) \ + part_stat_add(cpu, gendiskp, field, -subnd) + +#define all_stat_dec(cpu, gendiskp, field, sector) \ + all_stat_add(cpu, gendiskp, field, -1, sector) +#define all_stat_inc(cpu, gendiskp, part, field, sector) \ + all_stat_add(cpu, gendiskp, part, field, 1, sector) +#define all_stat_sub(cpu, gendiskp, part, field, subnd, sector) \ + all_stat_add(cpu, gendiskp, part, field, -subnd, sector) /* Inlines to alloc and free disk stats in struct gendisk */ #ifdef CONFIG_SMP @@ -401,8 +376,8 @@ static inline void free_part_stats(struct hd_struct *part) #endif /* CONFIG_SMP */ /* drivers/block/ll_rw_blk.c */ -extern void disk_round_stats(struct gendisk *disk); -extern void part_round_stats(struct hd_struct *part); +extern void disk_round_stats(int cpu, struct gendisk *disk); +extern void part_round_stats(int cpu, struct hd_struct *part); /* drivers/block/genhd.c */ extern int get_blkdev_list(char *, int); -- cgit v1.2.3-70-g09d2 From bcce3de1be61e424deef35d1e86e86a35c4b6e65 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:47:22 +0900 Subject: block: implement extended dev numbers Implement extended device numbers. A block driver can tell block layer that it wants to use extended device numbers. After the usual minor space is used up, block layer automatically allocates devt's from EXT_BLOCK_MAJOR. Currently only one major number is allocated for this but as the allocation is strictly on-demand, ~1mil minor space under it should suffice unless the system actually has more than ~1mil partitions and if that ever happens adding more majors to the extended devt area is easy. Due to internal implementation issues, the first partition can't be allocated on the extended area. In other words, genhd->minors should at least be 1. This limitation will be lifted by later changes. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/genhd.c | 120 +++++++++++++++++++++++++++++++++++++++++++++++--- fs/partitions/check.c | 9 +++- include/linux/genhd.h | 13 ++++-- include/linux/major.h | 2 + 4 files changed, 135 insertions(+), 9 deletions(-) (limited to 'fs/partitions') diff --git a/block/genhd.c b/block/genhd.c index 430626e440f..7bbfed05cec 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -16,6 +16,7 @@ #include #include #include +#include #include "blk.h" @@ -24,6 +25,15 @@ static DEFINE_MUTEX(block_class_lock); struct kobject *block_depr; #endif +/* for extended dynamic devt allocation, currently only one major is used */ +#define MAX_EXT_DEVT (1 << MINORBITS) + +/* For extended devt allocation. ext_devt_mutex prevents look up + * results from going away underneath its user. + */ +static DEFINE_MUTEX(ext_devt_mutex); +static DEFINE_IDR(ext_devt_idr); + static struct device_type disk_type; /** @@ -288,6 +298,74 @@ EXPORT_SYMBOL(unregister_blkdev); static struct kobj_map *bdev_map; +/** + * blk_alloc_devt - allocate a dev_t for a partition + * @part: partition to allocate dev_t for + * @gfp_mask: memory allocation flag + * @devt: out parameter for resulting dev_t + * + * Allocate a dev_t for block device. + * + * RETURNS: + * 0 on success, allocated dev_t is returned in *@devt. -errno on + * failure. + * + * CONTEXT: + * Might sleep. + */ +int blk_alloc_devt(struct hd_struct *part, dev_t *devt) +{ + struct gendisk *disk = part_to_disk(part); + int idx, rc; + + /* in consecutive minor range? */ + if (part->partno < disk->minors) { + *devt = MKDEV(disk->major, disk->first_minor + part->partno); + return 0; + } + + /* allocate ext devt */ + do { + if (!idr_pre_get(&ext_devt_idr, GFP_KERNEL)) + return -ENOMEM; + rc = idr_get_new(&ext_devt_idr, part, &idx); + } while (rc == -EAGAIN); + + if (rc) + return rc; + + if (idx > MAX_EXT_DEVT) { + idr_remove(&ext_devt_idr, idx); + return -EBUSY; + } + + *devt = MKDEV(BLOCK_EXT_MAJOR, idx); + return 0; +} + +/** + * blk_free_devt - free a dev_t + * @devt: dev_t to free + * + * Free @devt which was allocated using blk_alloc_devt(). + * + * CONTEXT: + * Might sleep. + */ +void blk_free_devt(dev_t devt) +{ + might_sleep(); + + if (devt == MKDEV(0, 0)) + return; + + if (MAJOR(devt) == BLOCK_EXT_MAJOR) { + mutex_lock(&ext_devt_mutex); + idr_remove(&ext_devt_idr, MINOR(devt)); + mutex_unlock(&ext_devt_mutex); + } +} + /* * Register device numbers dev..(dev+range-1) * range must be nonzero @@ -371,10 +449,27 @@ void unlink_gendisk(struct gendisk *disk) */ struct gendisk *get_gendisk(dev_t devt, int *partno) { - struct kobject *kobj = kobj_lookup(bdev_map, devt, partno); - struct device *dev = kobj_to_dev(kobj); + struct gendisk *disk = NULL; + + if (MAJOR(devt) != BLOCK_EXT_MAJOR) { + struct kobject *kobj; + + kobj = kobj_lookup(bdev_map, devt, partno); + if (kobj) + disk = dev_to_disk(kobj_to_dev(kobj)); + } else { + struct hd_struct *part; + + mutex_lock(&ext_devt_mutex); + part = idr_find(&ext_devt_idr, MINOR(devt)); + if (part && get_disk(part_to_disk(part))) { + *partno = part->partno; + disk = part_to_disk(part); + } + mutex_unlock(&ext_devt_mutex); + } - return kobj ? dev_to_disk(dev) : NULL; + return disk; } /** @@ -877,18 +972,30 @@ struct gendisk *alloc_disk(int minors) } struct gendisk *alloc_disk_node(int minors, int node_id) +{ + return alloc_disk_ext_node(minors, 0, node_id); +} + +struct gendisk *alloc_disk_ext(int minors, int ext_minors) +{ + return alloc_disk_ext_node(minors, ext_minors, -1); +} + +struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) { struct gendisk *disk; disk = kmalloc_node(sizeof(struct gendisk), GFP_KERNEL | __GFP_ZERO, node_id); if (disk) { + int tot_minors = minors + ext_minors; + if (!init_disk_stats(disk)) { kfree(disk); return NULL; } - if (minors > 1) { - int size = (minors - 1) * sizeof(struct hd_struct *); + if (tot_minors > 1) { + int size = (tot_minors - 1) * sizeof(struct hd_struct *); disk->__part = kmalloc_node(size, GFP_KERNEL | __GFP_ZERO, node_id); if (!disk->__part) { @@ -898,6 +1005,7 @@ struct gendisk *alloc_disk_node(int minors, int node_id) } } disk->minors = minors; + disk->ext_minors = ext_minors; rand_initialize_disk(disk); disk->dev.class = &block_class; disk->dev.type = &disk_type; @@ -910,6 +1018,8 @@ struct gendisk *alloc_disk_node(int minors, int node_id) EXPORT_SYMBOL(alloc_disk); EXPORT_SYMBOL(alloc_disk_node); +EXPORT_SYMBOL(alloc_disk_ext); +EXPORT_SYMBOL(alloc_disk_ext_node); struct kobject *get_disk(struct gendisk *disk) { diff --git a/fs/partitions/check.c b/fs/partitions/check.c index c442f0aadac..0d4b7f28f13 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -333,6 +333,7 @@ void delete_partition(struct gendisk *disk, int partno) if (!part) return; + blk_free_devt(part_devt(part)); rcu_assign_pointer(disk->__part[partno-1], NULL); kobject_put(part->holder_dir); device_del(&part->dev); @@ -352,6 +353,7 @@ int add_partition(struct gendisk *disk, int partno, sector_t start, sector_t len, int flags) { struct hd_struct *p; + dev_t devt = MKDEV(0, 0); int err; if (disk->__part[partno - 1]) @@ -378,11 +380,15 @@ int add_partition(struct gendisk *disk, int partno, "%s%d", disk->dev.bus_id, partno); device_initialize(&p->dev); - p->dev.devt = MKDEV(disk->major, disk->first_minor + partno); p->dev.class = &block_class; p->dev.type = &part_type; p->dev.parent = &disk->dev; + err = blk_alloc_devt(p, &devt); + if (err) + goto out_put; + p->dev.devt = devt; + /* delay uevent until 'holders' subdir is created */ p->dev.uevent_suppress = 1; err = device_add(&p->dev); @@ -419,6 +425,7 @@ out_del: device_del(&p->dev); out_put: put_device(&p->dev); + blk_free_devt(devt); return err; } diff --git a/include/linux/genhd.h b/include/linux/genhd.h index ac8a901f200..6fc53242406 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -113,13 +113,15 @@ struct hd_struct { #define GENHD_FL_FAIL 64 struct gendisk { - /* major, first_minor and minors are input parameters only, - * don't use directly. Use disk_devt() and disk_max_parts(). + /* major, first_minor, minors and ext_minors are input + * parameters only, don't use directly. Use disk_devt() and + * disk_max_parts(). */ int major; /* major number of driver */ int first_minor; int minors; /* maximum number of minors, =1 for * disks that can't be partitioned. */ + int ext_minors; /* number of extended dynamic minors */ char disk_name[32]; /* name of major driver */ @@ -167,7 +169,7 @@ static inline struct gendisk *part_to_disk(struct hd_struct *part) static inline int disk_max_parts(struct gendisk *disk) { - return disk->minors - 1; + return disk->minors + disk->ext_minors - 1; } static inline dev_t disk_devt(struct gendisk *disk) @@ -554,6 +556,8 @@ struct unixware_disklabel { #define ADDPART_FLAG_RAID 1 #define ADDPART_FLAG_WHOLEDISK 2 +extern int blk_alloc_devt(struct hd_struct *part, dev_t *devt); +extern void blk_free_devt(dev_t devt); extern dev_t blk_lookup_devt(const char *name, int partno); extern char *disk_name (struct gendisk *hd, int partno, char *buf); @@ -564,6 +568,9 @@ extern void printk_all_partitions(void); extern struct gendisk *alloc_disk_node(int minors, int node_id); extern struct gendisk *alloc_disk(int minors); +extern struct gendisk *alloc_disk_ext_node(int minors, int ext_minrs, + int node_id); +extern struct gendisk *alloc_disk_ext(int minors, int ext_minors); extern struct kobject *get_disk(struct gendisk *disk); extern void put_disk(struct gendisk *disk); extern void blk_register_region(dev_t devt, unsigned long range, diff --git a/include/linux/major.h b/include/linux/major.h index 53d5fafd85c..88249452b93 100644 --- a/include/linux/major.h +++ b/include/linux/major.h @@ -170,4 +170,6 @@ #define VIOTAPE_MAJOR 230 +#define BLOCK_EXT_MAJOR 259 + #endif -- cgit v1.2.3-70-g09d2 From ed9e1982347b36573cd622ee5f4e2a7ccd79b3fd Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:05 +0900 Subject: block: implement and use {disk|part}_to_dev() Implement {disk|part}_to_dev() and use them to access generic device instead of directly dereferencing {disk|part}->dev. To make sure no user is left behind, rename generic devices fields to __dev. This is in preparation of unifying partition 0 handling with other partitions. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/blk-integrity.c | 5 +-- block/blk-sysfs.c | 4 +-- block/genhd.c | 27 ++++++++-------- drivers/block/aoe/aoeblk.c | 4 +-- drivers/block/nbd.c | 4 +-- drivers/ide/ide-probe.c | 2 +- drivers/md/dm.c | 4 +-- drivers/md/md.c | 10 +++--- fs/block_dev.c | 4 +-- fs/partitions/check.c | 79 ++++++++++++++++++++++++---------------------- include/linux/genhd.h | 20 ++++++------ 11 files changed, 86 insertions(+), 77 deletions(-) (limited to 'fs/partitions') diff --git a/block/blk-integrity.c b/block/blk-integrity.c index d87606eaca1..69023da6315 100644 --- a/block/blk-integrity.c +++ b/block/blk-integrity.c @@ -331,7 +331,8 @@ int blk_integrity_register(struct gendisk *disk, struct blk_integrity *template) return -1; if (kobject_init_and_add(&bi->kobj, &integrity_ktype, - &disk->dev.kobj, "%s", "integrity")) { + &disk_to_dev(disk)->kobj, + "%s", "integrity")) { kmem_cache_free(integrity_cachep, bi); return -1; } @@ -375,7 +376,7 @@ void blk_integrity_unregister(struct gendisk *disk) kobject_uevent(&bi->kobj, KOBJ_REMOVE); kobject_del(&bi->kobj); - kobject_put(&disk->dev.kobj); + kobject_put(&disk_to_dev(disk)->kobj); kmem_cache_free(integrity_cachep, bi); } EXPORT_SYMBOL(blk_integrity_unregister); diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c index 304ec73ab82..b9a6ed16664 100644 --- a/block/blk-sysfs.c +++ b/block/blk-sysfs.c @@ -310,7 +310,7 @@ int blk_register_queue(struct gendisk *disk) if (!q->request_fn) return 0; - ret = kobject_add(&q->kobj, kobject_get(&disk->dev.kobj), + ret = kobject_add(&q->kobj, kobject_get(&disk_to_dev(disk)->kobj), "%s", "queue"); if (ret < 0) return ret; @@ -339,6 +339,6 @@ void blk_unregister_queue(struct gendisk *disk) kobject_uevent(&q->kobj, KOBJ_REMOVE); kobject_del(&q->kobj); - kobject_put(&disk->dev.kobj); + kobject_put(&disk_to_dev(disk)->kobj); } } diff --git a/block/genhd.c b/block/genhd.c index 67e5a59ced2..0a2f16bd54b 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -59,7 +59,7 @@ struct hd_struct *disk_get_part(struct gendisk *disk, int partno) rcu_read_lock(); part = rcu_dereference(disk->__part[partno - 1]); if (part) - get_device(&part->dev); + get_device(part_to_dev(part)); rcu_read_unlock(); return part; @@ -130,7 +130,7 @@ struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter) if (!(piter->flags & DISK_PITER_INCL_EMPTY) && !part->nr_sects) continue; - get_device(&part->dev); + get_device(part_to_dev(part)); piter->part = part; piter->idx += inc; break; @@ -435,7 +435,7 @@ static struct kobject *exact_match(dev_t devt, int *partno, void *data) { struct gendisk *p = data; - return &p->dev.kobj; + return &disk_to_dev(p)->kobj; } static int exact_lock(dev_t devt, void *data) @@ -460,7 +460,7 @@ void add_disk(struct gendisk *disk) int retval; disk->flags |= GENHD_FL_UP; - disk->dev.devt = MKDEV(disk->major, disk->first_minor); + disk_to_dev(disk)->devt = MKDEV(disk->major, disk->first_minor); blk_register_region(disk_devt(disk), disk->minors, NULL, exact_match, exact_lock, disk); register_disk(disk); @@ -468,7 +468,8 @@ void add_disk(struct gendisk *disk) bdi = &disk->queue->backing_dev_info; bdi_register_dev(bdi, disk_devt(disk)); - retval = sysfs_create_link(&disk->dev.kobj, &bdi->dev->kobj, "bdi"); + retval = sysfs_create_link(&disk_to_dev(disk)->kobj, &bdi->dev->kobj, + "bdi"); WARN_ON(retval); } @@ -477,7 +478,7 @@ EXPORT_SYMBOL(del_gendisk); /* in partitions/check.c */ void unlink_gendisk(struct gendisk *disk) { - sysfs_remove_link(&disk->dev.kobj, "bdi"); + sysfs_remove_link(&disk_to_dev(disk)->kobj, "bdi"); bdi_unregister(&disk->queue->backing_dev_info); blk_unregister_queue(disk); blk_unregister_region(disk_devt(disk), disk->minors); @@ -903,7 +904,7 @@ static int diskstats_show(struct seq_file *seqf, void *v) int cpu; /* - if (&gp->dev.kobj.entry == block_class.devices.next) + if (&disk_to_dev(gp)->kobj.entry == block_class.devices.next) seq_puts(seqf, "major minor name" " rio rmerge rsect ruse wio wmerge " "wsect wuse running use aveq" @@ -972,7 +973,7 @@ static void media_change_notify_thread(struct work_struct *work) * set enviroment vars to indicate which event this is for * so that user space will know to go check the media status. */ - kobject_uevent_env(&gd->dev.kobj, KOBJ_CHANGE, envp); + kobject_uevent_env(&disk_to_dev(gd)->kobj, KOBJ_CHANGE, envp); put_device(gd->driverfs_dev); } @@ -1062,9 +1063,9 @@ struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) disk->minors = minors; disk->ext_minors = ext_minors; rand_initialize_disk(disk); - disk->dev.class = &block_class; - disk->dev.type = &disk_type; - device_initialize(&disk->dev); + disk_to_dev(disk)->class = &block_class; + disk_to_dev(disk)->type = &disk_type; + device_initialize(disk_to_dev(disk)); INIT_WORK(&disk->async_notify, media_change_notify_thread); } @@ -1086,7 +1087,7 @@ struct kobject *get_disk(struct gendisk *disk) owner = disk->fops->owner; if (owner && !try_module_get(owner)) return NULL; - kobj = kobject_get(&disk->dev.kobj); + kobj = kobject_get(&disk_to_dev(disk)->kobj); if (kobj == NULL) { module_put(owner); return NULL; @@ -1100,7 +1101,7 @@ EXPORT_SYMBOL(get_disk); void put_disk(struct gendisk *disk) { if (disk) - kobject_put(&disk->dev.kobj); + kobject_put(&disk_to_dev(disk)->kobj); } EXPORT_SYMBOL(put_disk); diff --git a/drivers/block/aoe/aoeblk.c b/drivers/block/aoe/aoeblk.c index 0c39782b266..3edb6cb7d68 100644 --- a/drivers/block/aoe/aoeblk.c +++ b/drivers/block/aoe/aoeblk.c @@ -109,12 +109,12 @@ static const struct attribute_group attr_group = { static int aoedisk_add_sysfs(struct aoedev *d) { - return sysfs_create_group(&d->gd->dev.kobj, &attr_group); + return sysfs_create_group(&disk_to_dev(d->gd)->kobj, &attr_group); } void aoedisk_rm_sysfs(struct aoedev *d) { - sysfs_remove_group(&d->gd->dev.kobj, &attr_group); + sysfs_remove_group(&disk_to_dev(d->gd)->kobj, &attr_group); } static int diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c index 1778e4a2c67..7b3351260d5 100644 --- a/drivers/block/nbd.c +++ b/drivers/block/nbd.c @@ -403,7 +403,7 @@ static int nbd_do_it(struct nbd_device *lo) BUG_ON(lo->magic != LO_MAGIC); lo->pid = current->pid; - ret = sysfs_create_file(&lo->disk->dev.kobj, &pid_attr.attr); + ret = sysfs_create_file(&disk_to_dev(lo->disk)->kobj, &pid_attr.attr); if (ret) { printk(KERN_ERR "nbd: sysfs_create_file failed!"); return ret; @@ -412,7 +412,7 @@ static int nbd_do_it(struct nbd_device *lo) while ((req = nbd_read_stat(lo)) != NULL) nbd_end_request(req); - sysfs_remove_file(&lo->disk->dev.kobj, &pid_attr.attr); + sysfs_remove_file(&disk_to_dev(lo->disk)->kobj, &pid_attr.attr); return 0; } diff --git a/drivers/ide/ide-probe.c b/drivers/ide/ide-probe.c index a51a30e9eab..70aa86c8807 100644 --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -1188,7 +1188,7 @@ static struct kobject *exact_match(dev_t dev, int *part, void *data) { struct gendisk *p = data; *part &= (1 << PARTN_BITS) - 1; - return &p->dev.kobj; + return &disk_to_dev(p)->kobj; } static int exact_lock(dev_t dev, void *data) diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 653624792ea..637806695bb 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -1186,7 +1186,7 @@ static void event_callback(void *context) list_splice_init(&md->uevent_list, &uevents); spin_unlock_irqrestore(&md->uevent_lock, flags); - dm_send_uevents(&uevents, &md->disk->dev.kobj); + dm_send_uevents(&uevents, &disk_to_dev(md->disk)->kobj); atomic_inc(&md->event_nr); wake_up(&md->eventq); @@ -1643,7 +1643,7 @@ out: *---------------------------------------------------------------*/ void dm_kobject_uevent(struct mapped_device *md) { - kobject_uevent(&md->disk->dev.kobj, KOBJ_CHANGE); + kobject_uevent(&disk_to_dev(md->disk)->kobj, KOBJ_CHANGE); } uint32_t dm_next_uevent_seq(struct mapped_device *md) diff --git a/drivers/md/md.c b/drivers/md/md.c index deeac4b4417..96e9fccd2ea 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -1465,9 +1465,9 @@ static int bind_rdev_to_array(mdk_rdev_t * rdev, mddev_t * mddev) goto fail; if (rdev->bdev->bd_part) - ko = &rdev->bdev->bd_part->dev.kobj; + ko = &part_to_dev(rdev->bdev->bd_part)->kobj; else - ko = &rdev->bdev->bd_disk->dev.kobj; + ko = &disk_to_dev(rdev->bdev->bd_disk)->kobj; if ((err = sysfs_create_link(&rdev->kobj, ko, "block"))) { kobject_del(&rdev->kobj); goto fail; @@ -3470,8 +3470,8 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data) disk->queue = mddev->queue; add_disk(disk); mddev->gendisk = disk; - error = kobject_init_and_add(&mddev->kobj, &md_ktype, &disk->dev.kobj, - "%s", "md"); + error = kobject_init_and_add(&mddev->kobj, &md_ktype, + &disk_to_dev(disk)->kobj, "%s", "md"); mutex_unlock(&disks_mutex); if (error) printk(KERN_WARNING "md: cannot register %s/md - name in use\n", @@ -3761,7 +3761,7 @@ static int do_md_run(mddev_t * mddev) sysfs_notify(&mddev->kobj, NULL, "array_state"); sysfs_notify(&mddev->kobj, NULL, "sync_action"); sysfs_notify(&mddev->kobj, NULL, "degraded"); - kobject_uevent(&mddev->gendisk->dev.kobj, KOBJ_CHANGE); + kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE); return 0; } diff --git a/fs/block_dev.c b/fs/block_dev.c index 2f2873b9a04..a02df22f37c 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -543,9 +543,9 @@ EXPORT_SYMBOL(bd_release); static struct kobject *bdev_get_kobj(struct block_device *bdev) { if (bdev->bd_contains != bdev) - return kobject_get(&bdev->bd_part->dev.kobj); + return kobject_get(&part_to_dev(bdev->bd_part)->kobj); else - return kobject_get(&bdev->bd_disk->dev.kobj); + return kobject_get(&disk_to_dev(bdev->bd_disk)->kobj); } static struct kobject *bdev_get_holder(struct block_device *bdev) diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 0d4b7f28f13..ac0df3acdcd 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -309,7 +309,7 @@ static inline void disk_sysfs_add_subdirs(struct gendisk *disk) { struct kobject *k; - k = kobject_get(&disk->dev.kobj); + k = kobject_get(&disk_to_dev(disk)->kobj); disk->holder_dir = kobject_create_and_add("holders", k); disk->slave_dir = kobject_create_and_add("slaves", k); kobject_put(k); @@ -322,7 +322,7 @@ static void delete_partition_rcu_cb(struct rcu_head *head) part->start_sect = 0; part->nr_sects = 0; part_stat_set_all(part, 0); - put_device(&part->dev); + put_device(part_to_dev(part)); } void delete_partition(struct gendisk *disk, int partno) @@ -336,7 +336,7 @@ void delete_partition(struct gendisk *disk, int partno) blk_free_devt(part_devt(part)); rcu_assign_pointer(disk->__part[partno-1], NULL); kobject_put(part->holder_dir); - device_del(&part->dev); + device_del(part_to_dev(part)); call_rcu(&part->rcu_head, delete_partition_rcu_cb); } @@ -354,6 +354,9 @@ int add_partition(struct gendisk *disk, int partno, { struct hd_struct *p; dev_t devt = MKDEV(0, 0); + struct device *ddev = disk_to_dev(disk); + struct device *pdev; + const char *dname; int err; if (disk->__part[partno - 1]) @@ -367,42 +370,43 @@ int add_partition(struct gendisk *disk, int partno, err = -ENOMEM; goto out_free; } + pdev = part_to_dev(p); + p->start_sect = start; p->nr_sects = len; p->partno = partno; p->policy = disk->policy; - if (isdigit(disk->dev.bus_id[strlen(disk->dev.bus_id)-1])) - snprintf(p->dev.bus_id, BUS_ID_SIZE, - "%sp%d", disk->dev.bus_id, partno); + dname = dev_name(ddev); + if (isdigit(dname[strlen(dname) - 1])) + snprintf(pdev->bus_id, BUS_ID_SIZE, "%sp%d", dname, partno); else - snprintf(p->dev.bus_id, BUS_ID_SIZE, - "%s%d", disk->dev.bus_id, partno); + snprintf(pdev->bus_id, BUS_ID_SIZE, "%s%d", dname, partno); - device_initialize(&p->dev); - p->dev.class = &block_class; - p->dev.type = &part_type; - p->dev.parent = &disk->dev; + device_initialize(pdev); + pdev->class = &block_class; + pdev->type = &part_type; + pdev->parent = ddev; err = blk_alloc_devt(p, &devt); if (err) - goto out_put; - p->dev.devt = devt; + goto out_free; + pdev->devt = devt; /* delay uevent until 'holders' subdir is created */ - p->dev.uevent_suppress = 1; - err = device_add(&p->dev); + pdev->uevent_suppress = 1; + err = device_add(pdev); if (err) goto out_put; err = -ENOMEM; - p->holder_dir = kobject_create_and_add("holders", &p->dev.kobj); + p->holder_dir = kobject_create_and_add("holders", &pdev->kobj); if (!p->holder_dir) goto out_del; - p->dev.uevent_suppress = 0; + pdev->uevent_suppress = 0; if (flags & ADDPART_FLAG_WHOLEDISK) { - err = device_create_file(&p->dev, &dev_attr_whole_disk); + err = device_create_file(pdev, &dev_attr_whole_disk); if (err) goto out_del; } @@ -412,8 +416,8 @@ int add_partition(struct gendisk *disk, int partno, rcu_assign_pointer(disk->__part[partno - 1], p); /* suppress uevent if the disk supresses it */ - if (!disk->dev.uevent_suppress) - kobject_uevent(&p->dev.kobj, KOBJ_ADD); + if (!ddev->uevent_suppress) + kobject_uevent(&pdev->kobj, KOBJ_ADD); return 0; @@ -422,9 +426,9 @@ out_free: return err; out_del: kobject_put(p->holder_dir); - device_del(&p->dev); + device_del(pdev); out_put: - put_device(&p->dev); + put_device(pdev); blk_free_devt(devt); return err; } @@ -432,30 +436,31 @@ out_put: /* Not exported, helper to add_disk(). */ void register_disk(struct gendisk *disk) { + struct device *ddev = disk_to_dev(disk); struct block_device *bdev; struct disk_part_iter piter; struct hd_struct *part; char *s; int err; - disk->dev.parent = disk->driverfs_dev; + ddev->parent = disk->driverfs_dev; - strlcpy(disk->dev.bus_id, disk->disk_name, BUS_ID_SIZE); + strlcpy(ddev->bus_id, disk->disk_name, BUS_ID_SIZE); /* ewww... some of these buggers have / in the name... */ - s = strchr(disk->dev.bus_id, '/'); + s = strchr(ddev->bus_id, '/'); if (s) *s = '!'; /* delay uevents, until we scanned partition table */ - disk->dev.uevent_suppress = 1; + ddev->uevent_suppress = 1; - if (device_add(&disk->dev)) + if (device_add(ddev)) return; #ifndef CONFIG_SYSFS_DEPRECATED - err = sysfs_create_link(block_depr, &disk->dev.kobj, - kobject_name(&disk->dev.kobj)); + err = sysfs_create_link(block_depr, &ddev->kobj, + kobject_name(&ddev->kobj)); if (err) { - device_del(&disk->dev); + device_del(ddev); return; } #endif @@ -481,13 +486,13 @@ void register_disk(struct gendisk *disk) exit: /* announce disk after possible partitions are created */ - disk->dev.uevent_suppress = 0; - kobject_uevent(&disk->dev.kobj, KOBJ_ADD); + ddev->uevent_suppress = 0; + kobject_uevent(&ddev->kobj, KOBJ_ADD); /* announce possible partitions */ disk_part_iter_init(&piter, disk, 0); while ((part = disk_part_iter_next(&piter))) - kobject_uevent(&part->dev.kobj, KOBJ_ADD); + kobject_uevent(&part_to_dev(part)->kobj, KOBJ_ADD); disk_part_iter_exit(&piter); } @@ -518,7 +523,7 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) return -EIO; /* tell userspace that the media / partition table may have changed */ - kobject_uevent(&disk->dev.kobj, KOBJ_CHANGE); + kobject_uevent(&disk_to_dev(disk)->kobj, KOBJ_CHANGE); for (p = 1; p < state->limit; p++) { sector_t size = state->parts[p].size; @@ -591,7 +596,7 @@ void del_gendisk(struct gendisk *disk) kobject_put(disk->slave_dir); disk->driverfs_dev = NULL; #ifndef CONFIG_SYSFS_DEPRECATED - sysfs_remove_link(block_depr, disk->dev.bus_id); + sysfs_remove_link(block_depr, dev_name(disk_to_dev(disk))); #endif - device_del(&disk->dev); + device_del(disk_to_dev(disk)); } diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 6fc53242406..e4e18c509ac 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -15,9 +15,11 @@ #ifdef CONFIG_BLOCK -#define kobj_to_dev(k) container_of(k, struct device, kobj) -#define dev_to_disk(device) container_of(device, struct gendisk, dev) -#define dev_to_part(device) container_of(device, struct hd_struct, dev) +#define kobj_to_dev(k) container_of((k), struct device, kobj) +#define dev_to_disk(device) container_of((device), struct gendisk, __dev) +#define dev_to_part(device) container_of((device), struct hd_struct, __dev) +#define disk_to_dev(disk) (&((disk)->__dev)) +#define part_to_dev(part) (&((part)->__dev)) extern struct device_type part_type; extern struct kobject *block_depr; @@ -88,7 +90,7 @@ struct disk_stats { struct hd_struct { sector_t start_sect; sector_t nr_sects; - struct device dev; + struct device __dev; struct kobject *holder_dir; int policy, partno; #ifdef CONFIG_FAIL_MAKE_REQUEST @@ -139,7 +141,7 @@ struct gendisk { int flags; struct device *driverfs_dev; // FIXME: remove - struct device dev; + struct device __dev; struct kobject *holder_dir; struct kobject *slave_dir; @@ -163,7 +165,7 @@ struct gendisk { static inline struct gendisk *part_to_disk(struct hd_struct *part) { if (likely(part)) - return dev_to_disk((part)->dev.parent); + return dev_to_disk(part_to_dev(part)->parent); return NULL; } @@ -174,12 +176,12 @@ static inline int disk_max_parts(struct gendisk *disk) static inline dev_t disk_devt(struct gendisk *disk) { - return disk->dev.devt; + return disk_to_dev(disk)->devt; } static inline dev_t part_devt(struct hd_struct *part) { - return part->dev.devt; + return part_to_dev(part)->devt; } extern struct hd_struct *disk_get_part(struct gendisk *disk, int partno); @@ -187,7 +189,7 @@ extern struct hd_struct *disk_get_part(struct gendisk *disk, int partno); static inline void disk_put_part(struct hd_struct *part) { if (likely(part)) - put_device(&part->dev); + put_device(part_to_dev(part)); } /* -- cgit v1.2.3-70-g09d2 From b5d0b9df0ba5d9a044f3a21e7544f53d90bd1465 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 3 Sep 2008 09:06:42 +0200 Subject: block: introduce partition 0 genhd and partition code handled disk and partitions separately. All information about the whole disk was in struct genhd and partitions in struct hd_struct. However, the whole disk (part0) and other partitions have a lot in common and the data structures end up having good number of common fields and thus separate code paths doing the same thing. Also, the partition array was indexed by partno - 1 which gets pretty confusing at times. This patch introduces partition 0 and makes the partition array indexed by partno. Following patches will unify the handling of disk and parts piece-by-piece. This patch also implements disk_partitionable() which tests whether a disk is partitionable. With coming dynamic partition array change, the most common usage of disk_max_parts() will be testing whether a disk is partitionable and the number of max partitions will become much less important. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/genhd.c | 40 +++++++++++++++++++++++----------------- block/ioctl.c | 4 ++-- fs/block_dev.c | 2 +- fs/partitions/check.c | 12 ++++++------ include/linux/genhd.h | 11 +++++++++-- 5 files changed, 41 insertions(+), 28 deletions(-) (limited to 'fs/partitions') diff --git a/block/genhd.c b/block/genhd.c index 0a2f16bd54b..65b7386c26d 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -54,10 +54,10 @@ struct hd_struct *disk_get_part(struct gendisk *disk, int partno) { struct hd_struct *part; - if (unlikely(partno < 1 || partno > disk_max_parts(disk))) + if (unlikely(partno < 0 || partno >= disk_max_parts(disk))) return NULL; rcu_read_lock(); - part = rcu_dereference(disk->__part[partno - 1]); + part = rcu_dereference(disk->__part[partno]); if (part) get_device(part_to_dev(part)); rcu_read_unlock(); @@ -85,8 +85,10 @@ void disk_part_iter_init(struct disk_part_iter *piter, struct gendisk *disk, if (flags & DISK_PITER_REVERSE) piter->idx = disk_max_parts(piter->disk) - 1; - else + else if (flags & DISK_PITER_INCL_PART0) piter->idx = 0; + else + piter->idx = 1; piter->flags = flags; } @@ -114,7 +116,10 @@ struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter) /* determine iteration parameters */ if (piter->flags & DISK_PITER_REVERSE) { inc = -1; - end = -1; + if (piter->flags & DISK_PITER_INCL_PART0) + end = -1; + else + end = 0; } else { inc = 1; end = disk_max_parts(piter->disk); @@ -177,7 +182,7 @@ struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, sector_t sector) { int i; - for (i = 0; i < disk_max_parts(disk); i++) { + for (i = 1; i < disk_max_parts(disk); i++) { struct hd_struct *part = rcu_dereference(disk->__part[i]); if (part && part->start_sect <= sector && @@ -669,7 +674,7 @@ static int show_partition(struct seq_file *seqf, void *v) char buf[BDEVNAME_SIZE]; /* Don't show non-partitionable removeable devices or empty devices */ - if (!get_capacity(sgp) || (!disk_max_parts(sgp) && + if (!get_capacity(sgp) || (!disk_partitionable(sgp) && (sgp->flags & GENHD_FL_REMOVABLE))) return 0; if (sgp->flags & GENHD_FL_SUPPRESS_PARTITION_INFO) @@ -742,7 +747,7 @@ static ssize_t disk_ext_range_show(struct device *dev, { struct gendisk *disk = dev_to_disk(dev); - return sprintf(buf, "%d\n", disk_max_parts(disk) + 1); + return sprintf(buf, "%d\n", disk_max_parts(disk)); } static ssize_t disk_removable_show(struct device *dev, @@ -998,7 +1003,7 @@ dev_t blk_lookup_devt(const char *name, int partno) if (strcmp(dev->bus_id, name)) continue; - if (partno < 0 || partno > disk_max_parts(disk)) + if (partno < 0 || partno >= disk_max_parts(disk)) continue; if (partno == 0) @@ -1045,21 +1050,22 @@ struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) GFP_KERNEL | __GFP_ZERO, node_id); if (disk) { int tot_minors = minors + ext_minors; + int size = tot_minors * sizeof(struct hd_struct *); if (!init_disk_stats(disk)) { kfree(disk); return NULL; } - if (tot_minors > 1) { - int size = (tot_minors - 1) * sizeof(struct hd_struct *); - disk->__part = kmalloc_node(size, - GFP_KERNEL | __GFP_ZERO, node_id); - if (!disk->__part) { - free_disk_stats(disk); - kfree(disk); - return NULL; - } + + disk->__part = kmalloc_node(size, GFP_KERNEL | __GFP_ZERO, + node_id); + if (!disk->__part) { + free_disk_stats(disk); + kfree(disk); + return NULL; } + disk->__part[0] = &disk->part0; + disk->minors = minors; disk->ext_minors = ext_minors; rand_initialize_disk(disk); diff --git a/block/ioctl.c b/block/ioctl.c index a5f672ad55f..64e7c67a64b 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -30,7 +30,7 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user if (bdev != bdev->bd_contains) return -EINVAL; partno = p.pno; - if (partno <= 0 || partno > disk_max_parts(disk)) + if (partno <= 0 || partno >= disk_max_parts(disk)) return -EINVAL; switch (a.op) { case BLKPG_ADD_PARTITION: @@ -102,7 +102,7 @@ static int blkdev_reread_part(struct block_device *bdev) struct gendisk *disk = bdev->bd_disk; int res; - if (!disk_max_parts(disk) || bdev != bdev->bd_contains) + if (!disk_partitionable(disk) || bdev != bdev->bd_contains) return -EINVAL; if (!capable(CAP_SYS_ADMIN)) return -EACCES; diff --git a/fs/block_dev.c b/fs/block_dev.c index a02df22f37c..c982a910797 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -892,7 +892,7 @@ int check_disk_change(struct block_device *bdev) if (bdops->revalidate_disk) bdops->revalidate_disk(bdev->bd_disk); - if (disk_max_parts(bdev->bd_disk)) + if (disk_partitionable(bdev->bd_disk)) bdev->bd_invalidated = 1; return 1; } diff --git a/fs/partitions/check.c b/fs/partitions/check.c index ac0df3acdcd..b60699c271a 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -173,7 +173,7 @@ check_partition(struct gendisk *hd, struct block_device *bdev) if (isdigit(state->name[strlen(state->name)-1])) sprintf(state->name, "p"); - state->limit = disk_max_parts(hd) + 1; + state->limit = disk_max_parts(hd); i = res = err = 0; while (!res && check_part[i]) { memset(&state->parts, 0, sizeof(state->parts)); @@ -329,12 +329,12 @@ void delete_partition(struct gendisk *disk, int partno) { struct hd_struct *part; - part = disk->__part[partno-1]; + part = disk->__part[partno]; if (!part) return; blk_free_devt(part_devt(part)); - rcu_assign_pointer(disk->__part[partno-1], NULL); + rcu_assign_pointer(disk->__part[partno], NULL); kobject_put(part->holder_dir); device_del(part_to_dev(part)); @@ -359,7 +359,7 @@ int add_partition(struct gendisk *disk, int partno, const char *dname; int err; - if (disk->__part[partno - 1]) + if (disk->__part[partno]) return -EBUSY; p = kzalloc(sizeof(*p), GFP_KERNEL); @@ -413,7 +413,7 @@ int add_partition(struct gendisk *disk, int partno, /* everything is up and running, commence */ INIT_RCU_HEAD(&p->rcu_head); - rcu_assign_pointer(disk->__part[partno - 1], p); + rcu_assign_pointer(disk->__part[partno], p); /* suppress uevent if the disk supresses it */ if (!ddev->uevent_suppress) @@ -467,7 +467,7 @@ void register_disk(struct gendisk *disk) disk_sysfs_add_subdirs(disk); /* No minors to use for partitions */ - if (!disk_max_parts(disk)) + if (!disk_partitionable(disk)) goto exit; /* No such device (e.g., media were just removed) */ diff --git a/include/linux/genhd.h b/include/linux/genhd.h index e4e18c509ac..9e866a2aee5 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -127,12 +127,13 @@ struct gendisk { char disk_name[32]; /* name of major driver */ - /* Array of pointers to partitions indexed by partno - 1. + /* Array of pointers to partitions indexed by partno. * Protected with matching bdev lock but stat and other * non-critical accesses use RCU. Always access through * helpers. */ struct hd_struct **__part; + struct hd_struct part0; struct block_device_operations *fops; struct request_queue *queue; @@ -171,7 +172,12 @@ static inline struct gendisk *part_to_disk(struct hd_struct *part) static inline int disk_max_parts(struct gendisk *disk) { - return disk->minors + disk->ext_minors - 1; + return disk->minors + disk->ext_minors; +} + +static inline bool disk_partitionable(struct gendisk *disk) +{ + return disk_max_parts(disk) > 1; } static inline dev_t disk_devt(struct gendisk *disk) @@ -197,6 +203,7 @@ static inline void disk_put_part(struct hd_struct *part) */ #define DISK_PITER_REVERSE (1 << 0) /* iterate in the reverse direction */ #define DISK_PITER_INCL_EMPTY (1 << 1) /* include 0-sized parts */ +#define DISK_PITER_INCL_PART0 (1 << 2) /* include partition 0 */ struct disk_part_iter { struct gendisk *disk; -- cgit v1.2.3-70-g09d2 From 80795aefb76d10c5d698e60c7e7750b5330787da Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:07 +0900 Subject: block: move capacity from disk to part0 Move disk->capacity to part0->nr_sects and convert all users who directly accessed the field to use {get|set}_capacity(). This is done early to allow the __dev field to be moved. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- drivers/block/aoe/aoeblk.c | 2 +- drivers/block/aoe/aoecmd.c | 4 ++-- drivers/block/aoe/aoedev.c | 2 +- fs/partitions/check.c | 2 +- include/linux/genhd.h | 5 ++--- 5 files changed, 7 insertions(+), 8 deletions(-) (limited to 'fs/partitions') diff --git a/drivers/block/aoe/aoeblk.c b/drivers/block/aoe/aoeblk.c index 3edb6cb7d68..aa69556c348 100644 --- a/drivers/block/aoe/aoeblk.c +++ b/drivers/block/aoe/aoeblk.c @@ -276,7 +276,7 @@ aoeblk_gdalloc(void *vp) gd->first_minor = d->sysminor * AOE_PARTITIONS; gd->fops = &aoe_bdops; gd->private_data = d; - gd->capacity = d->ssize; + set_capacity(gd, d->ssize); snprintf(gd->disk_name, sizeof gd->disk_name, "etherd/e%ld.%d", d->aoemajor, d->aoeminor); diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c index 17eed8c025d..934800f979c 100644 --- a/drivers/block/aoe/aoecmd.c +++ b/drivers/block/aoe/aoecmd.c @@ -645,7 +645,7 @@ aoecmd_sleepwork(struct work_struct *work) unsigned long flags; u64 ssize; - ssize = d->gd->capacity; + ssize = get_capacity(d->gd); bd = bdget_disk(d->gd, 0); if (bd) { @@ -707,7 +707,7 @@ ataid_complete(struct aoedev *d, struct aoetgt *t, unsigned char *id) if (d->flags & (DEVFL_GDALLOC|DEVFL_NEWSIZE)) return; if (d->gd != NULL) { - d->gd->capacity = ssize; + set_capacity(d->gd, ssize); d->flags |= DEVFL_NEWSIZE; } else d->flags |= DEVFL_GDALLOC; diff --git a/drivers/block/aoe/aoedev.c b/drivers/block/aoe/aoedev.c index a1d813ab0d6..6a8038d115b 100644 --- a/drivers/block/aoe/aoedev.c +++ b/drivers/block/aoe/aoedev.c @@ -91,7 +91,7 @@ aoedev_downdev(struct aoedev *d) } if (d->gd) - d->gd->capacity = 0; + set_capacity(d->gd, 0); d->flags &= ~DEVFL_UP; } diff --git a/fs/partitions/check.c b/fs/partitions/check.c index b60699c271a..902b95f1f9d 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -586,7 +586,7 @@ void del_gendisk(struct gendisk *disk) disk_part_iter_exit(&piter); invalidate_partition(disk, 0); - disk->capacity = 0; + set_capacity(disk, 0); disk->flags &= ~GENHD_FL_UP; unlink_gendisk(disk); disk_stat_set_all(disk, 0); diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 9e866a2aee5..1cf828148ec 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -138,7 +138,6 @@ struct gendisk { struct block_device_operations *fops; struct request_queue *queue; void *private_data; - sector_t capacity; int flags; struct device *driverfs_dev; // FIXME: remove @@ -411,11 +410,11 @@ static inline sector_t get_start_sect(struct block_device *bdev) } static inline sector_t get_capacity(struct gendisk *disk) { - return disk->capacity; + return disk->part0.nr_sects; } static inline void set_capacity(struct gendisk *disk, sector_t size) { - disk->capacity = size; + disk->part0.nr_sects = size; } #ifdef CONFIG_SOLARIS_X86_PARTITION -- cgit v1.2.3-70-g09d2 From e56105214943ce5f0901d20e972a7cfd0d1d0656 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:09 +0900 Subject: block: unify sysfs size node handling Now that capacity and __dev are moved to part0, part0 and others can share the same method. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/genhd.c | 10 +--------- fs/partitions/check.c | 4 ++-- include/linux/genhd.h | 3 +++ 3 files changed, 6 insertions(+), 11 deletions(-) (limited to 'fs/partitions') diff --git a/block/genhd.c b/block/genhd.c index 36b9f1bdd91..c70db35076a 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -760,14 +760,6 @@ static ssize_t disk_ro_show(struct device *dev, return sprintf(buf, "%d\n", disk->policy ? 1 : 0); } -static ssize_t disk_size_show(struct device *dev, - struct device_attribute *attr, char *buf) -{ - struct gendisk *disk = dev_to_disk(dev); - - return sprintf(buf, "%llu\n", (unsigned long long)get_capacity(disk)); -} - static ssize_t disk_capability_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -835,7 +827,7 @@ static DEVICE_ATTR(range, S_IRUGO, disk_range_show, NULL); static DEVICE_ATTR(ext_range, S_IRUGO, disk_ext_range_show, NULL); static DEVICE_ATTR(removable, S_IRUGO, disk_removable_show, NULL); static DEVICE_ATTR(ro, S_IRUGO, disk_ro_show, NULL); -static DEVICE_ATTR(size, S_IRUGO, disk_size_show, NULL); +static DEVICE_ATTR(size, S_IRUGO, part_size_show, NULL); static DEVICE_ATTR(capability, S_IRUGO, disk_capability_show, NULL); static DEVICE_ATTR(stat, S_IRUGO, disk_stat_show, NULL); #ifdef CONFIG_FAIL_MAKE_REQUEST diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 902b95f1f9d..24d2c56d7d2 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -208,8 +208,8 @@ static ssize_t part_start_show(struct device *dev, return sprintf(buf, "%llu\n",(unsigned long long)p->start_sect); } -static ssize_t part_size_show(struct device *dev, - struct device_attribute *attr, char *buf) +ssize_t part_size_show(struct device *dev, + struct device_attribute *attr, char *buf) { struct hd_struct *p = dev_to_part(dev); return sprintf(buf, "%llu\n",(unsigned long long)p->nr_sects); diff --git a/include/linux/genhd.h b/include/linux/genhd.h index ff293ec8b3f..9cb8380cf0e 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -591,6 +591,9 @@ extern void blk_register_region(dev_t devt, unsigned long range, void *data); extern void blk_unregister_region(dev_t devt, unsigned long range); +extern ssize_t part_size_show(struct device *dev, + struct device_attribute *attr, char *buf); + #else /* CONFIG_BLOCK */ static inline void printk_all_partitions(void) { } -- cgit v1.2.3-70-g09d2 From b7db9956e57c8151b930d5e5fe5c766e6aad3ff7 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:10 +0900 Subject: block: move policy from disk to part0 Move disk->policy to part0->policy. Implement and use get_disk_ro(). Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/genhd.c | 16 +++++----------- drivers/ide/ide-cd.c | 2 +- drivers/md/dm-ioctl.c | 2 +- fs/partitions/check.c | 2 +- include/linux/genhd.h | 6 +++++- 5 files changed, 13 insertions(+), 15 deletions(-) (limited to 'fs/partitions') diff --git a/block/genhd.c b/block/genhd.c index c70db35076a..70358f3c742 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -757,7 +757,7 @@ static ssize_t disk_ro_show(struct device *dev, { struct gendisk *disk = dev_to_disk(dev); - return sprintf(buf, "%d\n", disk->policy ? 1 : 0); + return sprintf(buf, "%d\n", get_disk_ro(disk) ? 1 : 0); } static ssize_t disk_capability_show(struct device *dev, @@ -1090,10 +1090,7 @@ EXPORT_SYMBOL(put_disk); void set_device_ro(struct block_device *bdev, int flag) { - if (bdev->bd_contains != bdev) - bdev->bd_part->policy = flag; - else - bdev->bd_disk->policy = flag; + bdev->bd_part->policy = flag; } EXPORT_SYMBOL(set_device_ro); @@ -1103,8 +1100,8 @@ void set_disk_ro(struct gendisk *disk, int flag) struct disk_part_iter piter; struct hd_struct *part; - disk->policy = flag; - disk_part_iter_init(&piter, disk, DISK_PITER_INCL_EMPTY); + disk_part_iter_init(&piter, disk, + DISK_PITER_INCL_EMPTY | DISK_PITER_INCL_PART0); while ((part = disk_part_iter_next(&piter))) part->policy = flag; disk_part_iter_exit(&piter); @@ -1116,10 +1113,7 @@ int bdev_read_only(struct block_device *bdev) { if (!bdev) return 0; - else if (bdev->bd_contains != bdev) - return bdev->bd_part->policy; - else - return bdev->bd_disk->policy; + return bdev->bd_part->policy; } EXPORT_SYMBOL(bdev_read_only); diff --git a/drivers/ide/ide-cd.c b/drivers/ide/ide-cd.c index f16bb466723..03c2cb6a58b 100644 --- a/drivers/ide/ide-cd.c +++ b/drivers/ide/ide-cd.c @@ -1113,7 +1113,7 @@ static ide_startstop_t cdrom_start_rw(ide_drive_t *drive, struct request *rq) if (write) { /* disk has become write protected */ - if (cd->disk->policy) { + if (get_disk_ro(cd->disk)) { cdrom_end_request(drive, 0); return ide_stopped; } diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c index c3de311117a..5b919159f08 100644 --- a/drivers/md/dm-ioctl.c +++ b/drivers/md/dm-ioctl.c @@ -548,7 +548,7 @@ static int __dev_status(struct mapped_device *md, struct dm_ioctl *param) */ param->open_count = dm_open_count(md); - if (disk->policy) + if (get_disk_ro(disk)) param->flags |= DM_READONLY_FLAG; param->event_nr = dm_get_event_nr(md); diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 24d2c56d7d2..ace6d03602c 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -375,7 +375,7 @@ int add_partition(struct gendisk *disk, int partno, p->start_sect = start; p->nr_sects = len; p->partno = partno; - p->policy = disk->policy; + p->policy = get_disk_ro(disk); dname = dev_name(ddev); if (isdigit(dname[strlen(dname) - 1])) diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 9cb8380cf0e..4411bdd671d 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -145,7 +145,6 @@ struct gendisk { struct kobject *slave_dir; struct timer_rand_state *random; - int policy; atomic_t sync_io; /* RAID */ unsigned long stamp; @@ -403,6 +402,11 @@ extern struct block_device *bdget_disk(struct gendisk *disk, int partno); extern void set_device_ro(struct block_device *bdev, int flag); extern void set_disk_ro(struct gendisk *disk, int flag); +static inline int get_disk_ro(struct gendisk *disk) +{ + return disk->part0.policy; +} + /* drivers/char/random.c */ extern void add_disk_randomness(struct gendisk *disk); extern void rand_initialize_disk(struct gendisk *disk); -- cgit v1.2.3-70-g09d2 From 4c46501d1659475dc6c89554af6ce7fe6ecf615c Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:11 +0900 Subject: block: move holder_dir from disk to part0 Move disk->holder_dir to part0->holder_dir. Kill now mostly superflous bdev_get_holder(). While at it, kill superflous kobject_get/put() around holder_dir, slave_dir and cmd_filter creation and collapse disk_sysfs_add_subdirs() into register_disk(). These serve no purpose but obfuscating the code. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/cmd-filter.c | 9 ++------- fs/block_dev.c | 10 +--------- fs/partitions/check.c | 15 +++------------ include/linux/genhd.h | 1 - 4 files changed, 6 insertions(+), 29 deletions(-) (limited to 'fs/partitions') diff --git a/block/cmd-filter.c b/block/cmd-filter.c index 79c14996ac1..e669aed4c6b 100644 --- a/block/cmd-filter.c +++ b/block/cmd-filter.c @@ -211,14 +211,10 @@ int blk_register_filter(struct gendisk *disk) { int ret; struct blk_cmd_filter *filter = &disk->queue->cmd_filter; - struct kobject *parent = kobject_get(disk->holder_dir->parent); - if (!parent) - return -ENODEV; - - ret = kobject_init_and_add(&filter->kobj, &rcf_ktype, parent, + ret = kobject_init_and_add(&filter->kobj, &rcf_ktype, + &disk_to_dev(disk)->kobj, "%s", "cmd_filter"); - if (ret < 0) return ret; @@ -231,7 +227,6 @@ void blk_unregister_filter(struct gendisk *disk) struct blk_cmd_filter *filter = &disk->queue->cmd_filter; kobject_put(&filter->kobj); - kobject_put(disk->holder_dir->parent); } EXPORT_SYMBOL(blk_unregister_filter); #endif diff --git a/fs/block_dev.c b/fs/block_dev.c index c982a910797..57d57264285 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -548,14 +548,6 @@ static struct kobject *bdev_get_kobj(struct block_device *bdev) return kobject_get(&disk_to_dev(bdev->bd_disk)->kobj); } -static struct kobject *bdev_get_holder(struct block_device *bdev) -{ - if (bdev->bd_contains != bdev) - return kobject_get(bdev->bd_part->holder_dir); - else - return kobject_get(bdev->bd_disk->holder_dir); -} - static int add_symlink(struct kobject *from, struct kobject *to) { if (!from || !to) @@ -608,7 +600,7 @@ static int bd_holder_grab_dirs(struct block_device *bdev, if (!bo->sdev) goto fail_put_hdev; - bo->hdir = bdev_get_holder(bdev); + bo->hdir = kobject_get(bdev->bd_part->holder_dir); if (!bo->hdir) goto fail_put_sdev; diff --git a/fs/partitions/check.c b/fs/partitions/check.c index ace6d03602c..f0f604950ff 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -305,16 +305,6 @@ struct device_type part_type = { .release = part_release, }; -static inline void disk_sysfs_add_subdirs(struct gendisk *disk) -{ - struct kobject *k; - - k = kobject_get(&disk_to_dev(disk)->kobj); - disk->holder_dir = kobject_create_and_add("holders", k); - disk->slave_dir = kobject_create_and_add("slaves", k); - kobject_put(k); -} - static void delete_partition_rcu_cb(struct rcu_head *head) { struct hd_struct *part = container_of(head, struct hd_struct, rcu_head); @@ -464,7 +454,8 @@ void register_disk(struct gendisk *disk) return; } #endif - disk_sysfs_add_subdirs(disk); + disk->part0.holder_dir = kobject_create_and_add("holders", &ddev->kobj); + disk->slave_dir = kobject_create_and_add("slaves", &ddev->kobj); /* No minors to use for partitions */ if (!disk_partitionable(disk)) @@ -592,7 +583,7 @@ void del_gendisk(struct gendisk *disk) disk_stat_set_all(disk, 0); disk->stamp = 0; - kobject_put(disk->holder_dir); + kobject_put(disk->part0.holder_dir); kobject_put(disk->slave_dir); disk->driverfs_dev = NULL; #ifndef CONFIG_SYSFS_DEPRECATED diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 4411bdd671d..2c0e1b597ab 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -141,7 +141,6 @@ struct gendisk { int flags; struct device *driverfs_dev; // FIXME: remove - struct kobject *holder_dir; struct kobject *slave_dir; struct timer_rand_state *random; -- cgit v1.2.3-70-g09d2 From 0762b8bde9729f10f8e6249809660ff2ec3ad735 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:12 +0900 Subject: block: always set bdev->bd_part Till now, bdev->bd_part is set only if the bdev was for parts other than part0. This patch makes bdev->bd_part always set so that code paths don't have to differenciate common handling. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/blk-core.c | 2 +- drivers/md/md.c | 5 +--- fs/block_dev.c | 67 ++++++++++++++++++++++++--------------------------- fs/partitions/check.c | 7 +----- include/linux/genhd.h | 2 +- 5 files changed, 35 insertions(+), 48 deletions(-) (limited to 'fs/partitions') diff --git a/block/blk-core.c b/block/blk-core.c index e0a5ee36849..a4a7c08d2f2 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -1274,7 +1274,7 @@ __setup("fail_make_request=", setup_fail_make_request); static int should_fail_request(struct bio *bio) { if ((bio->bi_bdev->bd_disk->flags & GENHD_FL_FAIL) || - (bio->bi_bdev->bd_part && bio->bi_bdev->bd_part->make_it_fail)) + bio->bi_bdev->bd_part->make_it_fail) return should_fail(&fail_make_request, bio->bi_size); return 0; diff --git a/drivers/md/md.c b/drivers/md/md.c index 96e9fccd2ea..2bd9cf41612 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -1464,10 +1464,7 @@ static int bind_rdev_to_array(mdk_rdev_t * rdev, mddev_t * mddev) if ((err = kobject_add(&rdev->kobj, &mddev->kobj, "dev-%s", b))) goto fail; - if (rdev->bdev->bd_part) - ko = &part_to_dev(rdev->bdev->bd_part)->kobj; - else - ko = &disk_to_dev(rdev->bdev->bd_disk)->kobj; + ko = &part_to_dev(rdev->bdev->bd_part)->kobj; if ((err = sysfs_create_link(&rdev->kobj, ko, "block"))) { kobject_del(&rdev->kobj); goto fail; diff --git a/fs/block_dev.c b/fs/block_dev.c index 57d57264285..c3fa19bd64d 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -540,14 +540,6 @@ EXPORT_SYMBOL(bd_release); * /sys/block/sda/holders/dm-0 --> /sys/block/dm-0 */ -static struct kobject *bdev_get_kobj(struct block_device *bdev) -{ - if (bdev->bd_contains != bdev) - return kobject_get(&part_to_dev(bdev->bd_part)->kobj); - else - return kobject_get(&disk_to_dev(bdev->bd_disk)->kobj); -} - static int add_symlink(struct kobject *from, struct kobject *to) { if (!from || !to) @@ -596,7 +588,7 @@ static int bd_holder_grab_dirs(struct block_device *bdev, if (!bo->hdev) goto fail_put_sdir; - bo->sdev = bdev_get_kobj(bdev); + bo->sdev = kobject_get(&part_to_dev(bdev->bd_part)->kobj); if (!bo->sdev) goto fail_put_hdev; @@ -919,7 +911,6 @@ static int __blkdev_put(struct block_device *bdev, int for_part); static int do_open(struct block_device *bdev, struct file *file, int for_part) { - struct module *owner = NULL; struct gendisk *disk; struct hd_struct *part = NULL; int ret; @@ -941,25 +932,27 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) ret = -ENXIO; file->f_mapping = bdev->bd_inode->i_mapping; + lock_kernel(); + disk = get_gendisk(bdev->bd_dev, &partno); - if (!disk) { - unlock_kernel(); - bdput(bdev); - return ret; - } - owner = disk->fops->owner; + if (!disk) + goto out_unlock_kernel; + part = disk_get_part(disk, partno); + if (!part) + goto out_unlock_kernel; mutex_lock_nested(&bdev->bd_mutex, for_part); if (!bdev->bd_openers) { bdev->bd_disk = disk; + bdev->bd_part = part; bdev->bd_contains = bdev; if (!partno) { struct backing_dev_info *bdi; if (disk->fops->open) { ret = disk->fops->open(bdev->bd_inode, file); if (ret) - goto out_first; + goto out_clear; } if (!bdev->bd_openers) { bd_set_size(bdev,(loff_t)get_capacity(disk)<<9); @@ -975,31 +968,32 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) whole = bdget_disk(disk, 0); ret = -ENOMEM; if (!whole) - goto out_first; + goto out_clear; BUG_ON(for_part); ret = __blkdev_get(whole, file->f_mode, file->f_flags, 1); if (ret) - goto out_first; + goto out_clear; bdev->bd_contains = whole; - part = disk_get_part(disk, partno); bdev->bd_inode->i_data.backing_dev_info = whole->bd_inode->i_data.backing_dev_info; if (!(disk->flags & GENHD_FL_UP) || !part || !part->nr_sects) { ret = -ENXIO; - goto out_first; + goto out_clear; } - bdev->bd_part = part; bd_set_size(bdev, (loff_t)part->nr_sects << 9); } } else { + disk_put_part(part); put_disk(disk); - module_put(owner); + module_put(disk->fops->owner); + part = NULL; + disk = NULL; if (bdev->bd_contains == bdev) { if (bdev->bd_disk->fops->open) { ret = bdev->bd_disk->fops->open(bdev->bd_inode, file); if (ret) - goto out; + goto out_unlock_bdev; } if (bdev->bd_invalidated) rescan_partitions(bdev->bd_disk, bdev); @@ -1012,20 +1006,24 @@ static int do_open(struct block_device *bdev, struct file *file, int for_part) unlock_kernel(); return 0; -out_first: + out_clear: bdev->bd_disk = NULL; + bdev->bd_part = NULL; bdev->bd_inode->i_data.backing_dev_info = &default_backing_dev_info; if (bdev != bdev->bd_contains) __blkdev_put(bdev->bd_contains, 1); bdev->bd_contains = NULL; - put_disk(disk); - disk_put_part(part); - module_put(owner); -out: + out_unlock_bdev: mutex_unlock(&bdev->bd_mutex); + out_unlock_kernel: unlock_kernel(); - if (ret) - bdput(bdev); + + disk_put_part(part); + if (disk) + module_put(disk->fops->owner); + put_disk(disk); + bdput(bdev); + return ret; } @@ -1110,11 +1108,8 @@ static int __blkdev_put(struct block_device *bdev, int for_part) put_disk(disk); module_put(owner); - - if (bdev->bd_contains != bdev) { - disk_put_part(bdev->bd_part); - bdev->bd_part = NULL; - } + disk_put_part(bdev->bd_part); + bdev->bd_part = NULL; bdev->bd_disk = NULL; bdev->bd_inode->i_data.backing_dev_info = &default_backing_dev_info; if (bdev != bdev->bd_contains) diff --git a/fs/partitions/check.c b/fs/partitions/check.c index f0f604950ff..87298c0fc8c 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -134,12 +134,7 @@ char *disk_name(struct gendisk *hd, int partno, char *buf) const char *bdevname(struct block_device *bdev, char *buf) { - int partno = 0; - - if (bdev->bd_part) - partno = bdev->bd_part->partno; - - return disk_name(bdev->bd_disk, partno, buf); + return disk_name(bdev->bd_disk, bdev->bd_part->partno, buf); } EXPORT_SYMBOL(bdevname); diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 2c0e1b597ab..45a3682b5d8 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -412,7 +412,7 @@ extern void rand_initialize_disk(struct gendisk *disk); static inline sector_t get_start_sect(struct block_device *bdev) { - return bdev->bd_contains == bdev ? 0 : bdev->bd_part->start_sect; + return bdev->bd_part->start_sect; } static inline sector_t get_capacity(struct gendisk *disk) { -- cgit v1.2.3-70-g09d2 From eddb2e26b5ee3c5da68ba4bf1921ba20e2097bff Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:13 +0900 Subject: block: kill GENHD_FL_FAIL and use part0->make_it_fail GENHD_FL_FAIL for disk is what make_it_fail is for parts. Kill it and use part0->make_it_fail. Sysfs node handling is unified too. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/blk-core.c | 5 +++-- block/genhd.c | 30 +----------------------------- fs/partitions/check.c | 10 +++++----- include/linux/genhd.h | 8 +++++++- 4 files changed, 16 insertions(+), 37 deletions(-) (limited to 'fs/partitions') diff --git a/block/blk-core.c b/block/blk-core.c index a4a7c08d2f2..505ec61067d 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -1273,8 +1273,9 @@ __setup("fail_make_request=", setup_fail_make_request); static int should_fail_request(struct bio *bio) { - if ((bio->bi_bdev->bd_disk->flags & GENHD_FL_FAIL) || - bio->bi_bdev->bd_part->make_it_fail) + struct hd_struct *part = bio->bi_bdev->bd_part; + + if (part_to_disk(part)->part0.make_it_fail || part->make_it_fail) return should_fail(&fail_make_request, bio->bi_size); return 0; diff --git a/block/genhd.c b/block/genhd.c index 70358f3c742..06a252f2b96 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -795,34 +795,6 @@ static ssize_t disk_stat_show(struct device *dev, jiffies_to_msecs(disk_stat_read(disk, time_in_queue))); } -#ifdef CONFIG_FAIL_MAKE_REQUEST -static ssize_t disk_fail_show(struct device *dev, - struct device_attribute *attr, char *buf) -{ - struct gendisk *disk = dev_to_disk(dev); - - return sprintf(buf, "%d\n", disk->flags & GENHD_FL_FAIL ? 1 : 0); -} - -static ssize_t disk_fail_store(struct device *dev, - struct device_attribute *attr, - const char *buf, size_t count) -{ - struct gendisk *disk = dev_to_disk(dev); - int i; - - if (count > 0 && sscanf(buf, "%d", &i) > 0) { - if (i == 0) - disk->flags &= ~GENHD_FL_FAIL; - else - disk->flags |= GENHD_FL_FAIL; - } - - return count; -} - -#endif - static DEVICE_ATTR(range, S_IRUGO, disk_range_show, NULL); static DEVICE_ATTR(ext_range, S_IRUGO, disk_ext_range_show, NULL); static DEVICE_ATTR(removable, S_IRUGO, disk_removable_show, NULL); @@ -832,7 +804,7 @@ static DEVICE_ATTR(capability, S_IRUGO, disk_capability_show, NULL); static DEVICE_ATTR(stat, S_IRUGO, disk_stat_show, NULL); #ifdef CONFIG_FAIL_MAKE_REQUEST static struct device_attribute dev_attr_fail = - __ATTR(make-it-fail, S_IRUGO|S_IWUSR, disk_fail_show, disk_fail_store); + __ATTR(make-it-fail, S_IRUGO|S_IWUSR, part_fail_show, part_fail_store); #endif static struct attribute *disk_attrs[] = { diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 87298c0fc8c..60592d9f43b 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -238,17 +238,17 @@ static ssize_t part_stat_show(struct device *dev, } #ifdef CONFIG_FAIL_MAKE_REQUEST -static ssize_t part_fail_show(struct device *dev, - struct device_attribute *attr, char *buf) +ssize_t part_fail_show(struct device *dev, + struct device_attribute *attr, char *buf) { struct hd_struct *p = dev_to_part(dev); return sprintf(buf, "%d\n", p->make_it_fail); } -static ssize_t part_fail_store(struct device *dev, - struct device_attribute *attr, - const char *buf, size_t count) +ssize_t part_fail_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) { struct hd_struct *p = dev_to_part(dev); int i; diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 45a3682b5d8..3d15b42dc35 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -112,7 +112,6 @@ struct hd_struct { #define GENHD_FL_CD 8 #define GENHD_FL_UP 16 #define GENHD_FL_SUPPRESS_PARTITION_INFO 32 -#define GENHD_FL_FAIL 64 struct gendisk { /* major, first_minor, minors and ext_minors are input @@ -596,6 +595,13 @@ extern void blk_unregister_region(dev_t devt, unsigned long range); extern ssize_t part_size_show(struct device *dev, struct device_attribute *attr, char *buf); +#ifdef CONFIG_FAIL_MAKE_REQUEST +extern ssize_t part_fail_show(struct device *dev, + struct device_attribute *attr, char *buf); +extern ssize_t part_fail_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count); +#endif /* CONFIG_FAIL_MAKE_REQUEST */ #else /* CONFIG_BLOCK */ -- cgit v1.2.3-70-g09d2 From 074a7aca7afa6f230104e8e65eba3420263714a5 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:14 +0900 Subject: block: move stats from disk to part0 Move stats related fields - stamp, in_flight, dkstats - from disk to part0 and unify stat handling such that... * part_stat_*() now updates part0 together if the specified partition is not part0. ie. part_stat_*() are now essentially all_stat_*(). * {disk|all}_stat_*() are gone. * part_round_stats() is updated similary. It handles part0 stats automatically and disk_round_stats() is killed. * part_{inc|dec}_in_fligh() is implemented which automatically updates part0 stats for parts other than part0. * disk_map_sector_rcu() is updated to return part0 if no part matches. Combined with the above changes, this makes NULL special case handling in callers unnecessary. * Separate stats show code paths for disk are collapsed into part stats show code paths. * Rename disk_stat_lock/unlock() to part_stat_lock/unlock() While at it, reposition stat handling macros a bit and add missing parentheses around macro parameters. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/blk-core.c | 84 ++++++++++-------------- block/blk-merge.c | 12 ++-- block/genhd.c | 97 +++++++-------------------- drivers/block/aoe/aoecmd.c | 12 ++-- drivers/md/dm.c | 27 ++++---- drivers/md/linear.c | 9 +-- drivers/md/md.c | 4 +- drivers/md/multipath.c | 9 +-- drivers/md/raid0.c | 9 +-- drivers/md/raid1.c | 9 +-- drivers/md/raid10.c | 9 +-- drivers/md/raid5.c | 9 +-- fs/partitions/check.c | 12 ++-- include/linux/genhd.h | 159 +++++++++++++-------------------------------- 14 files changed, 165 insertions(+), 296 deletions(-) (limited to 'fs/partitions') diff --git a/block/blk-core.c b/block/blk-core.c index 505ec61067d..98138f00252 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -61,21 +61,17 @@ static void drive_stat_acct(struct request *rq, int new_io) if (!blk_fs_request(rq) || !rq->rq_disk) return; - cpu = disk_stat_lock(); + cpu = part_stat_lock(); part = disk_map_sector_rcu(rq->rq_disk, rq->sector); if (!new_io) - all_stat_inc(cpu, rq->rq_disk, part, merges[rw], rq->sector); + part_stat_inc(cpu, part, merges[rw]); else { - disk_round_stats(cpu, rq->rq_disk); - rq->rq_disk->in_flight++; - if (part) { - part_round_stats(cpu, part); - part->in_flight++; - } + part_round_stats(cpu, part); + part_inc_in_flight(part); } - disk_stat_unlock(); + part_stat_unlock(); } void blk_queue_congestion_threshold(struct request_queue *q) @@ -983,8 +979,22 @@ static inline void add_request(struct request_queue *q, struct request *req) __elv_add_request(q, req, ELEVATOR_INSERT_SORT, 0); } -/* - * disk_round_stats() - Round off the performance stats on a struct +static void part_round_stats_single(int cpu, struct hd_struct *part, + unsigned long now) +{ + if (now == part->stamp) + return; + + if (part->in_flight) { + __part_stat_add(cpu, part, time_in_queue, + part->in_flight * (now - part->stamp)); + __part_stat_add(cpu, part, io_ticks, (now - part->stamp)); + } + part->stamp = now; +} + +/** + * part_round_stats() - Round off the performance stats on a struct * disk_stats. * * The average IO queue length and utilisation statistics are maintained @@ -998,36 +1008,15 @@ static inline void add_request(struct request_queue *q, struct request *req) * /proc/diskstats. This accounts immediately for all queue usage up to * the current jiffies and restarts the counters again. */ -void disk_round_stats(int cpu, struct gendisk *disk) -{ - unsigned long now = jiffies; - - if (now == disk->stamp) - return; - - if (disk->in_flight) { - disk_stat_add(cpu, disk, time_in_queue, - disk->in_flight * (now - disk->stamp)); - disk_stat_add(cpu, disk, io_ticks, (now - disk->stamp)); - } - disk->stamp = now; -} -EXPORT_SYMBOL_GPL(disk_round_stats); - void part_round_stats(int cpu, struct hd_struct *part) { unsigned long now = jiffies; - if (now == part->stamp) - return; - - if (part->in_flight) { - part_stat_add(cpu, part, time_in_queue, - part->in_flight * (now - part->stamp)); - part_stat_add(cpu, part, io_ticks, (now - part->stamp)); - } - part->stamp = now; + if (part->partno) + part_round_stats_single(cpu, &part_to_disk(part)->part0, now); + part_round_stats_single(cpu, part, now); } +EXPORT_SYMBOL_GPL(part_round_stats); /* * queue lock must be held @@ -1567,11 +1556,10 @@ static int __end_that_request_first(struct request *req, int error, struct hd_struct *part; int cpu; - cpu = disk_stat_lock(); + cpu = part_stat_lock(); part = disk_map_sector_rcu(req->rq_disk, req->sector); - all_stat_add(cpu, req->rq_disk, part, sectors[rw], - nr_bytes >> 9, req->sector); - disk_stat_unlock(); + part_stat_add(cpu, part, sectors[rw], nr_bytes >> 9); + part_stat_unlock(); } total_bytes = bio_nbytes = 0; @@ -1758,19 +1746,15 @@ static void end_that_request_last(struct request *req, int error) struct hd_struct *part; int cpu; - cpu = disk_stat_lock(); + cpu = part_stat_lock(); part = disk_map_sector_rcu(disk, req->sector); - all_stat_inc(cpu, disk, part, ios[rw], req->sector); - all_stat_add(cpu, disk, part, ticks[rw], duration, req->sector); - disk_round_stats(cpu, disk); - disk->in_flight--; - if (part) { - part_round_stats(cpu, part); - part->in_flight--; - } + part_stat_inc(cpu, part, ios[rw]); + part_stat_add(cpu, part, ticks[rw], duration); + part_round_stats(cpu, part); + part_dec_in_flight(part); - disk_stat_unlock(); + part_stat_unlock(); } if (req->end_io) diff --git a/block/blk-merge.c b/block/blk-merge.c index d926a24bf1f..c77196d5589 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -390,17 +390,13 @@ static int attempt_merge(struct request_queue *q, struct request *req, struct hd_struct *part; int cpu; - cpu = disk_stat_lock(); + cpu = part_stat_lock(); part = disk_map_sector_rcu(req->rq_disk, req->sector); - disk_round_stats(cpu, req->rq_disk); - req->rq_disk->in_flight--; - if (part) { - part_round_stats(cpu, part); - part->in_flight--; - } + part_round_stats(cpu, part); + part_dec_in_flight(part); - disk_stat_unlock(); + part_stat_unlock(); } req->ioprio = ioprio_best(req->ioprio, next->ioprio); diff --git a/block/genhd.c b/block/genhd.c index 06a252f2b96..e1cb96fb883 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -176,7 +176,7 @@ EXPORT_SYMBOL_GPL(disk_part_iter_exit); * while preemption is disabled. * * RETURNS: - * Found partition on success, NULL if there's no matching partition. + * Found partition on success, part0 is returned if no partition matches */ struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, sector_t sector) { @@ -189,7 +189,7 @@ struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, sector_t sector) sector < part->start_sect + part->nr_sects) return part; } - return NULL; + return &disk->part0; } EXPORT_SYMBOL_GPL(disk_map_sector_rcu); @@ -580,24 +580,24 @@ void __init printk_all_partitions(void) * numbers in hex - the same format as the root= * option takes. */ - printk("%s %10llu %s", - bdevt_str(disk_devt(disk), devt_buf), - (unsigned long long)get_capacity(disk) >> 1, - disk_name(disk, 0, name_buf)); - if (disk->driverfs_dev != NULL && - disk->driverfs_dev->driver != NULL) - printk(" driver: %s\n", - disk->driverfs_dev->driver->name); - else - printk(" (driver?)\n"); + disk_part_iter_init(&piter, disk, DISK_PITER_INCL_PART0); + while ((part = disk_part_iter_next(&piter))) { + bool is_part0 = part == &disk->part0; - /* now show the partitions */ - disk_part_iter_init(&piter, disk, 0); - while ((part = disk_part_iter_next(&piter))) - printk(" %s %10llu %s\n", + printk("%s%s %10llu %s", is_part0 ? "" : " ", bdevt_str(part_devt(part), devt_buf), (unsigned long long)part->nr_sects >> 1, disk_name(disk, part->partno, name_buf)); + if (is_part0) { + if (disk->driverfs_dev != NULL && + disk->driverfs_dev->driver != NULL) + printk(" driver: %s\n", + disk->driverfs_dev->driver->name); + else + printk(" (driver?)\n"); + } else + printk("\n"); + } disk_part_iter_exit(&piter); } class_dev_iter_exit(&iter); @@ -674,12 +674,7 @@ static int show_partition(struct seq_file *seqf, void *v) return 0; /* show the full disk and all non-0 size partitions of it */ - seq_printf(seqf, "%4d %7d %10llu %s\n", - MAJOR(disk_devt(sgp)), MINOR(disk_devt(sgp)), - (unsigned long long)get_capacity(sgp) >> 1, - disk_name(sgp, 0, buf)); - - disk_part_iter_init(&piter, sgp, 0); + disk_part_iter_init(&piter, sgp, DISK_PITER_INCL_PART0); while ((part = disk_part_iter_next(&piter))) seq_printf(seqf, "%4d %7d %10llu %s\n", MAJOR(part_devt(part)), MINOR(part_devt(part)), @@ -768,40 +763,13 @@ static ssize_t disk_capability_show(struct device *dev, return sprintf(buf, "%x\n", disk->flags); } -static ssize_t disk_stat_show(struct device *dev, - struct device_attribute *attr, char *buf) -{ - struct gendisk *disk = dev_to_disk(dev); - int cpu; - - cpu = disk_stat_lock(); - disk_round_stats(cpu, disk); - disk_stat_unlock(); - return sprintf(buf, - "%8lu %8lu %8llu %8u " - "%8lu %8lu %8llu %8u " - "%8u %8u %8u" - "\n", - disk_stat_read(disk, ios[READ]), - disk_stat_read(disk, merges[READ]), - (unsigned long long)disk_stat_read(disk, sectors[READ]), - jiffies_to_msecs(disk_stat_read(disk, ticks[READ])), - disk_stat_read(disk, ios[WRITE]), - disk_stat_read(disk, merges[WRITE]), - (unsigned long long)disk_stat_read(disk, sectors[WRITE]), - jiffies_to_msecs(disk_stat_read(disk, ticks[WRITE])), - disk->in_flight, - jiffies_to_msecs(disk_stat_read(disk, io_ticks)), - jiffies_to_msecs(disk_stat_read(disk, time_in_queue))); -} - static DEVICE_ATTR(range, S_IRUGO, disk_range_show, NULL); static DEVICE_ATTR(ext_range, S_IRUGO, disk_ext_range_show, NULL); static DEVICE_ATTR(removable, S_IRUGO, disk_removable_show, NULL); static DEVICE_ATTR(ro, S_IRUGO, disk_ro_show, NULL); static DEVICE_ATTR(size, S_IRUGO, part_size_show, NULL); static DEVICE_ATTR(capability, S_IRUGO, disk_capability_show, NULL); -static DEVICE_ATTR(stat, S_IRUGO, disk_stat_show, NULL); +static DEVICE_ATTR(stat, S_IRUGO, part_stat_show, NULL); #ifdef CONFIG_FAIL_MAKE_REQUEST static struct device_attribute dev_attr_fail = __ATTR(make-it-fail, S_IRUGO|S_IWUSR, part_fail_show, part_fail_store); @@ -836,7 +804,7 @@ static void disk_release(struct device *dev) kfree(disk->random); kfree(disk->__part); - free_disk_stats(disk); + free_part_stats(&disk->part0); kfree(disk); } struct class block_class = { @@ -873,28 +841,11 @@ static int diskstats_show(struct seq_file *seqf, void *v) "\n\n"); */ - cpu = disk_stat_lock(); - disk_round_stats(cpu, gp); - disk_stat_unlock(); - seq_printf(seqf, "%4d %7d %s %lu %lu %llu %u %lu %lu %llu %u %u %u %u\n", - MAJOR(disk_devt(gp)), MINOR(disk_devt(gp)), - disk_name(gp, 0, buf), - disk_stat_read(gp, ios[0]), disk_stat_read(gp, merges[0]), - (unsigned long long)disk_stat_read(gp, sectors[0]), - jiffies_to_msecs(disk_stat_read(gp, ticks[0])), - disk_stat_read(gp, ios[1]), disk_stat_read(gp, merges[1]), - (unsigned long long)disk_stat_read(gp, sectors[1]), - jiffies_to_msecs(disk_stat_read(gp, ticks[1])), - gp->in_flight, - jiffies_to_msecs(disk_stat_read(gp, io_ticks)), - jiffies_to_msecs(disk_stat_read(gp, time_in_queue))); - - /* now show all non-0 size partitions of it */ - disk_part_iter_init(&piter, gp, 0); + disk_part_iter_init(&piter, gp, DISK_PITER_INCL_PART0); while ((hd = disk_part_iter_next(&piter))) { - cpu = disk_stat_lock(); + cpu = part_stat_lock(); part_round_stats(cpu, hd); - disk_stat_unlock(); + part_stat_unlock(); seq_printf(seqf, "%4d %7d %s %lu %lu %llu " "%u %lu %lu %llu %u %u %u %u\n", MAJOR(part_devt(hd)), MINOR(part_devt(hd)), @@ -1000,7 +951,7 @@ struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) int tot_minors = minors + ext_minors; int size = tot_minors * sizeof(struct hd_struct *); - if (!init_disk_stats(disk)) { + if (!init_part_stats(&disk->part0)) { kfree(disk); return NULL; } @@ -1008,7 +959,7 @@ struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) disk->__part = kmalloc_node(size, GFP_KERNEL | __GFP_ZERO, node_id); if (!disk->__part) { - free_disk_stats(disk); + free_part_stats(&disk->part0); kfree(disk); return NULL; } diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c index 934800f979c..961d29a53ca 100644 --- a/drivers/block/aoe/aoecmd.c +++ b/drivers/block/aoe/aoecmd.c @@ -758,15 +758,15 @@ diskstats(struct gendisk *disk, struct bio *bio, ulong duration, sector_t sector struct hd_struct *part; int cpu; - cpu = disk_stat_lock(); + cpu = part_stat_lock(); part = disk_map_sector_rcu(disk, sector); - all_stat_inc(cpu, disk, part, ios[rw], sector); - all_stat_add(cpu, disk, part, ticks[rw], duration, sector); - all_stat_add(cpu, disk, part, sectors[rw], n_sect, sector); - all_stat_add(cpu, disk, part, io_ticks, duration, sector); + part_stat_inc(cpu, part, ios[rw]); + part_stat_add(cpu, part, ticks[rw], duration); + part_stat_add(cpu, part, sectors[rw], n_sect); + part_stat_add(cpu, part, io_ticks, duration); - disk_stat_unlock(); + part_stat_unlock(); } void diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 637806695bb..327de03a5bd 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -381,10 +381,10 @@ static void start_io_acct(struct dm_io *io) io->start_time = jiffies; - cpu = disk_stat_lock(); - disk_round_stats(cpu, dm_disk(md)); - disk_stat_unlock(); - dm_disk(md)->in_flight = atomic_inc_return(&md->pending); + cpu = part_stat_lock(); + part_round_stats(cpu, &dm_disk(md)->part0); + part_stat_unlock(); + dm_disk(md)->part0.in_flight = atomic_inc_return(&md->pending); } static int end_io_acct(struct dm_io *io) @@ -395,12 +395,13 @@ static int end_io_acct(struct dm_io *io) int pending, cpu; int rw = bio_data_dir(bio); - cpu = disk_stat_lock(); - disk_round_stats(cpu, dm_disk(md)); - disk_stat_add(cpu, dm_disk(md), ticks[rw], duration); - disk_stat_unlock(); + cpu = part_stat_lock(); + part_round_stats(cpu, &dm_disk(md)->part0); + part_stat_add(cpu, &dm_disk(md)->part0, ticks[rw], duration); + part_stat_unlock(); - dm_disk(md)->in_flight = pending = atomic_dec_return(&md->pending); + dm_disk(md)->part0.in_flight = pending = + atomic_dec_return(&md->pending); return !pending; } @@ -899,10 +900,10 @@ static int dm_request(struct request_queue *q, struct bio *bio) down_read(&md->io_lock); - cpu = disk_stat_lock(); - disk_stat_inc(cpu, dm_disk(md), ios[rw]); - disk_stat_add(cpu, dm_disk(md), sectors[rw], bio_sectors(bio)); - disk_stat_unlock(); + cpu = part_stat_lock(); + part_stat_inc(cpu, &dm_disk(md)->part0, ios[rw]); + part_stat_add(cpu, &dm_disk(md)->part0, sectors[rw], bio_sectors(bio)); + part_stat_unlock(); /* * If we're suspended we have to queue diff --git a/drivers/md/linear.c b/drivers/md/linear.c index 00cbc8e4729..c80ea90593d 100644 --- a/drivers/md/linear.c +++ b/drivers/md/linear.c @@ -325,10 +325,11 @@ static int linear_make_request (struct request_queue *q, struct bio *bio) return 0; } - cpu = disk_stat_lock(); - disk_stat_inc(cpu, mddev->gendisk, ios[rw]); - disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); - disk_stat_unlock(); + cpu = part_stat_lock(); + part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]); + part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw], + bio_sectors(bio)); + part_stat_unlock(); tmp_dev = which_dev(mddev, bio->bi_sector); block = bio->bi_sector >> 1; diff --git a/drivers/md/md.c b/drivers/md/md.c index 2bd9cf41612..0a3a4bdcd4a 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -5546,8 +5546,8 @@ static int is_mddev_idle(mddev_t *mddev) rcu_read_lock(); rdev_for_each_rcu(rdev, mddev) { struct gendisk *disk = rdev->bdev->bd_contains->bd_disk; - curr_events = disk_stat_read(disk, sectors[0]) + - disk_stat_read(disk, sectors[1]) - + curr_events = part_stat_read(&disk->part0, sectors[0]) + + part_stat_read(&disk->part0, sectors[1]) - atomic_read(&disk->sync_io); /* sync IO will cause sync_io to increase before the disk_stats * as sync_io is counted when a request starts, and diff --git a/drivers/md/multipath.c b/drivers/md/multipath.c index 182f5a94cdc..8bb8794129b 100644 --- a/drivers/md/multipath.c +++ b/drivers/md/multipath.c @@ -159,10 +159,11 @@ static int multipath_make_request (struct request_queue *q, struct bio * bio) mp_bh->master_bio = bio; mp_bh->mddev = mddev; - cpu = disk_stat_lock(); - disk_stat_inc(cpu, mddev->gendisk, ios[rw]); - disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); - disk_stat_unlock(); + cpu = part_stat_lock(); + part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]); + part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw], + bio_sectors(bio)); + part_stat_unlock(); mp_bh->path = multipath_map(conf); if (mp_bh->path < 0) { diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c index e26030fa59a..f52f442a735 100644 --- a/drivers/md/raid0.c +++ b/drivers/md/raid0.c @@ -406,10 +406,11 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio) return 0; } - cpu = disk_stat_lock(); - disk_stat_inc(cpu, mddev->gendisk, ios[rw]); - disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); - disk_stat_unlock(); + cpu = part_stat_lock(); + part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]); + part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw], + bio_sectors(bio)); + part_stat_unlock(); chunk_size = mddev->chunk_size >> 10; chunk_sects = mddev->chunk_size >> 9; diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index babb13036f9..b9764429d85 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -804,10 +804,11 @@ static int make_request(struct request_queue *q, struct bio * bio) bitmap = mddev->bitmap; - cpu = disk_stat_lock(); - disk_stat_inc(cpu, mddev->gendisk, ios[rw]); - disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); - disk_stat_unlock(); + cpu = part_stat_lock(); + part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]); + part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw], + bio_sectors(bio)); + part_stat_unlock(); /* * make_request() can abort the operation when READA is being diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index 5ec80da0a9d..5f990133f5e 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -844,10 +844,11 @@ static int make_request(struct request_queue *q, struct bio * bio) */ wait_barrier(conf); - cpu = disk_stat_lock(); - disk_stat_inc(cpu, mddev->gendisk, ios[rw]); - disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bio)); - disk_stat_unlock(); + cpu = part_stat_lock(); + part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]); + part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw], + bio_sectors(bio)); + part_stat_unlock(); r10_bio = mempool_alloc(conf->r10bio_pool, GFP_NOIO); diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 5899f211515..ae16794bef2 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -3396,10 +3396,11 @@ static int make_request(struct request_queue *q, struct bio * bi) md_write_start(mddev, bi); - cpu = disk_stat_lock(); - disk_stat_inc(cpu, mddev->gendisk, ios[rw]); - disk_stat_add(cpu, mddev->gendisk, sectors[rw], bio_sectors(bi)); - disk_stat_unlock(); + cpu = part_stat_lock(); + part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]); + part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw], + bio_sectors(bi)); + part_stat_unlock(); if (rw == READ && mddev->reshape_position == MaxSector && diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 60592d9f43b..f517869e8d1 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -210,15 +210,15 @@ ssize_t part_size_show(struct device *dev, return sprintf(buf, "%llu\n",(unsigned long long)p->nr_sects); } -static ssize_t part_stat_show(struct device *dev, - struct device_attribute *attr, char *buf) +ssize_t part_stat_show(struct device *dev, + struct device_attribute *attr, char *buf) { struct hd_struct *p = dev_to_part(dev); int cpu; - cpu = disk_stat_lock(); + cpu = part_stat_lock(); part_round_stats(cpu, p); - disk_stat_unlock(); + part_stat_unlock(); return sprintf(buf, "%8lu %8lu %8llu %8u " "%8lu %8lu %8llu %8u " @@ -575,8 +575,8 @@ void del_gendisk(struct gendisk *disk) set_capacity(disk, 0); disk->flags &= ~GENHD_FL_UP; unlink_gendisk(disk); - disk_stat_set_all(disk, 0); - disk->stamp = 0; + part_stat_set_all(&disk->part0, 0); + disk->part0.stamp = 0; kobject_put(disk->part0.holder_dir); kobject_put(disk->slave_dir); diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 3d15b42dc35..c90e1b4fbe5 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -145,13 +145,6 @@ struct gendisk { struct timer_rand_state *random; atomic_t sync_io; /* RAID */ - unsigned long stamp; - int in_flight; -#ifdef CONFIG_SMP - struct disk_stats *dkstats; -#else - struct disk_stats dkstats; -#endif struct work_struct async_notify; #ifdef CONFIG_BLK_DEV_INTEGRITY struct blk_integrity *integrity; @@ -232,46 +225,18 @@ extern struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, * internal use only. */ #ifdef CONFIG_SMP -#define disk_stat_lock() ({ rcu_read_lock(); get_cpu(); }) -#define disk_stat_unlock() do { put_cpu(); rcu_read_unlock(); } while (0) - -#define disk_stat_add(cpu, gendiskp, field, addnd) \ - (per_cpu_ptr(gendiskp->dkstats, cpu)->field += addnd) - -#define disk_stat_read(gendiskp, field) \ -({ \ - typeof(gendiskp->dkstats->field) res = 0; \ - int i; \ - for_each_possible_cpu(i) \ - res += per_cpu_ptr(gendiskp->dkstats, i)->field; \ - res; \ -}) - -static inline void disk_stat_set_all(struct gendisk *gendiskp, int value) -{ - int i; - - for_each_possible_cpu(i) - memset(per_cpu_ptr(gendiskp->dkstats, i), value, - sizeof(struct disk_stats)); -} +#define part_stat_lock() ({ rcu_read_lock(); get_cpu(); }) +#define part_stat_unlock() do { put_cpu(); rcu_read_unlock(); } while (0) -#define part_stat_add(cpu, part, field, addnd) \ - (per_cpu_ptr(part->dkstats, cpu)->field += addnd) - -#define all_stat_add(cpu, gendiskp, part, field, addnd, sector) \ -({ \ - if (part) \ - part_stat_add(cpu, part, field, addnd); \ - disk_stat_add(cpu, gendiskp, field, addnd); \ -}) +#define __part_stat_add(cpu, part, field, addnd) \ + (per_cpu_ptr((part)->dkstats, (cpu))->field += (addnd)) #define part_stat_read(part, field) \ ({ \ - typeof(part->dkstats->field) res = 0; \ + typeof((part)->dkstats->field) res = 0; \ int i; \ for_each_possible_cpu(i) \ - res += per_cpu_ptr(part->dkstats, i)->field; \ + res += per_cpu_ptr((part)->dkstats, i)->field; \ res; \ }) @@ -284,109 +249,73 @@ static inline void part_stat_set_all(struct hd_struct *part, int value) sizeof(struct disk_stats)); } -#else /* !CONFIG_SMP */ -#define disk_stat_lock() ({ rcu_read_lock(); 0; }) -#define disk_stat_unlock() rcu_read_unlock() - -#define disk_stat_add(cpu, gendiskp, field, addnd) \ - (gendiskp->dkstats.field += addnd) -#define disk_stat_read(gendiskp, field) (gendiskp->dkstats.field) - -static inline void disk_stat_set_all(struct gendisk *gendiskp, int value) +static inline int init_part_stats(struct hd_struct *part) { - memset(&gendiskp->dkstats, value, sizeof (struct disk_stats)); + part->dkstats = alloc_percpu(struct disk_stats); + if (!part->dkstats) + return 0; + return 1; } -#define part_stat_add(cpu, part, field, addnd) \ - (part->dkstats.field += addnd) - -#define all_stat_add(cpu, gendiskp, part, field, addnd, sector) \ -({ \ - if (part) \ - part_stat_add(cpu, part, field, addnd); \ - disk_stat_add(cpu, gendiskp, field, addnd); \ -}) - -#define part_stat_read(part, field) (part->dkstats.field) - -static inline void part_stat_set_all(struct hd_struct *part, int value) +static inline void free_part_stats(struct hd_struct *part) { - memset(&part->dkstats, value, sizeof(struct disk_stats)); + free_percpu(part->dkstats); } -#endif /* CONFIG_SMP */ - -#define disk_stat_dec(cpu, gendiskp, field) \ - disk_stat_add(cpu, gendiskp, field, -1) -#define disk_stat_inc(cpu, gendiskp, field) \ - disk_stat_add(cpu, gendiskp, field, 1) -#define disk_stat_sub(cpu, gendiskp, field, subnd) \ - disk_stat_add(cpu, gendiskp, field, -subnd) - -#define part_stat_dec(cpu, gendiskp, field) \ - part_stat_add(cpu, gendiskp, field, -1) -#define part_stat_inc(cpu, gendiskp, field) \ - part_stat_add(cpu, gendiskp, field, 1) -#define part_stat_sub(cpu, gendiskp, field, subnd) \ - part_stat_add(cpu, gendiskp, field, -subnd) +#else /* !CONFIG_SMP */ +#define part_stat_lock() ({ rcu_read_lock(); 0; }) +#define part_stat_unlock() rcu_read_unlock() -#define all_stat_dec(cpu, gendiskp, field, sector) \ - all_stat_add(cpu, gendiskp, field, -1, sector) -#define all_stat_inc(cpu, gendiskp, part, field, sector) \ - all_stat_add(cpu, gendiskp, part, field, 1, sector) -#define all_stat_sub(cpu, gendiskp, part, field, subnd, sector) \ - all_stat_add(cpu, gendiskp, part, field, -subnd, sector) +#define __part_stat_add(cpu, part, field, addnd) \ + ((part)->dkstats.field += addnd) -/* Inlines to alloc and free disk stats in struct gendisk */ -#ifdef CONFIG_SMP -static inline int init_disk_stats(struct gendisk *disk) -{ - disk->dkstats = alloc_percpu(struct disk_stats); - if (!disk->dkstats) - return 0; - return 1; -} +#define part_stat_read(part, field) ((part)->dkstats.field) -static inline void free_disk_stats(struct gendisk *disk) +static inline void part_stat_set_all(struct hd_struct *part, int value) { - free_percpu(disk->dkstats); + memset(&part->dkstats, value, sizeof(struct disk_stats)); } static inline int init_part_stats(struct hd_struct *part) { - part->dkstats = alloc_percpu(struct disk_stats); - if (!part->dkstats) - return 0; return 1; } static inline void free_part_stats(struct hd_struct *part) { - free_percpu(part->dkstats); } -#else /* CONFIG_SMP */ -static inline int init_disk_stats(struct gendisk *disk) -{ - return 1; -} +#endif /* CONFIG_SMP */ -static inline void free_disk_stats(struct gendisk *disk) -{ -} +#define part_stat_add(cpu, part, field, addnd) do { \ + __part_stat_add((cpu), (part), field, addnd); \ + if ((part)->partno) \ + __part_stat_add((cpu), &part_to_disk((part))->part0, \ + field, addnd); \ +} while (0) -static inline int init_part_stats(struct hd_struct *part) +#define part_stat_dec(cpu, gendiskp, field) \ + part_stat_add(cpu, gendiskp, field, -1) +#define part_stat_inc(cpu, gendiskp, field) \ + part_stat_add(cpu, gendiskp, field, 1) +#define part_stat_sub(cpu, gendiskp, field, subnd) \ + part_stat_add(cpu, gendiskp, field, -subnd) + +static inline void part_inc_in_flight(struct hd_struct *part) { - return 1; + part->in_flight++; + if (part->partno) + part_to_disk(part)->part0.in_flight++; } -static inline void free_part_stats(struct hd_struct *part) +static inline void part_dec_in_flight(struct hd_struct *part) { + part->in_flight--; + if (part->partno) + part_to_disk(part)->part0.in_flight--; } -#endif /* CONFIG_SMP */ /* drivers/block/ll_rw_blk.c */ -extern void disk_round_stats(int cpu, struct gendisk *disk); extern void part_round_stats(int cpu, struct hd_struct *part); /* drivers/block/genhd.c */ @@ -595,6 +524,8 @@ extern void blk_unregister_region(dev_t devt, unsigned long range); extern ssize_t part_size_show(struct device *dev, struct device_attribute *attr, char *buf); +extern ssize_t part_stat_show(struct device *dev, + struct device_attribute *attr, char *buf); #ifdef CONFIG_FAIL_MAKE_REQUEST extern ssize_t part_fail_show(struct device *dev, struct device_attribute *attr, char *buf); -- cgit v1.2.3-70-g09d2 From 540eed5637b766bb1e881ef744c42617760b4815 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:15 +0900 Subject: block: make partition array dynamic disk->__part used to be statically allocated to the maximum possible number of partitions. This patch makes partition array allocation dynamic. The added overhead is minimal as only real change is one memory dereference changed to RCU one. This saves both a bit of memory and cpu cycles iterating through unoccupied slots and makes increasing partition limit easier. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/genhd.c | 129 ++++++++++++++++++++++++++++++++++++++++++-------- block/ioctl.c | 2 +- fs/partitions/check.c | 31 ++++++++++-- include/linux/genhd.h | 19 +++++++- 4 files changed, 154 insertions(+), 27 deletions(-) (limited to 'fs/partitions') diff --git a/block/genhd.c b/block/genhd.c index e1cb96fb883..c2b14aa69d5 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -52,14 +52,21 @@ static struct device_type disk_type; */ struct hd_struct *disk_get_part(struct gendisk *disk, int partno) { - struct hd_struct *part; + struct hd_struct *part = NULL; + struct disk_part_tbl *ptbl; - if (unlikely(partno < 0 || partno >= disk_max_parts(disk))) + if (unlikely(partno < 0)) return NULL; + rcu_read_lock(); - part = rcu_dereference(disk->__part[partno]); - if (part) - get_device(part_to_dev(part)); + + ptbl = rcu_dereference(disk->part_tbl); + if (likely(partno < ptbl->len)) { + part = rcu_dereference(ptbl->part[partno]); + if (part) + get_device(part_to_dev(part)); + } + rcu_read_unlock(); return part; @@ -80,17 +87,24 @@ EXPORT_SYMBOL_GPL(disk_get_part); void disk_part_iter_init(struct disk_part_iter *piter, struct gendisk *disk, unsigned int flags) { + struct disk_part_tbl *ptbl; + + rcu_read_lock(); + ptbl = rcu_dereference(disk->part_tbl); + piter->disk = disk; piter->part = NULL; if (flags & DISK_PITER_REVERSE) - piter->idx = disk_max_parts(piter->disk) - 1; + piter->idx = ptbl->len - 1; else if (flags & DISK_PITER_INCL_PART0) piter->idx = 0; else piter->idx = 1; piter->flags = flags; + + rcu_read_unlock(); } EXPORT_SYMBOL_GPL(disk_part_iter_init); @@ -105,13 +119,16 @@ EXPORT_SYMBOL_GPL(disk_part_iter_init); */ struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter) { + struct disk_part_tbl *ptbl; int inc, end; /* put the last partition */ disk_put_part(piter->part); piter->part = NULL; + /* get part_tbl */ rcu_read_lock(); + ptbl = rcu_dereference(piter->disk->part_tbl); /* determine iteration parameters */ if (piter->flags & DISK_PITER_REVERSE) { @@ -122,14 +139,14 @@ struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter) end = 0; } else { inc = 1; - end = disk_max_parts(piter->disk); + end = ptbl->len; } /* iterate to the next partition */ for (; piter->idx != end; piter->idx += inc) { struct hd_struct *part; - part = rcu_dereference(piter->disk->__part[piter->idx]); + part = rcu_dereference(ptbl->part[piter->idx]); if (!part) continue; if (!(piter->flags & DISK_PITER_INCL_EMPTY) && !part->nr_sects) @@ -180,10 +197,13 @@ EXPORT_SYMBOL_GPL(disk_part_iter_exit); */ struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, sector_t sector) { + struct disk_part_tbl *ptbl; int i; - for (i = 1; i < disk_max_parts(disk); i++) { - struct hd_struct *part = rcu_dereference(disk->__part[i]); + ptbl = rcu_dereference(disk->part_tbl); + + for (i = 1; i < ptbl->len; i++) { + struct hd_struct *part = rcu_dereference(ptbl->part[i]); if (part && part->start_sect <= sector && sector < part->start_sect + part->nr_sects) @@ -798,12 +818,86 @@ static struct attribute_group *disk_attr_groups[] = { NULL }; +static void disk_free_ptbl_rcu_cb(struct rcu_head *head) +{ + struct disk_part_tbl *ptbl = + container_of(head, struct disk_part_tbl, rcu_head); + + kfree(ptbl); +} + +/** + * disk_replace_part_tbl - replace disk->part_tbl in RCU-safe way + * @disk: disk to replace part_tbl for + * @new_ptbl: new part_tbl to install + * + * Replace disk->part_tbl with @new_ptbl in RCU-safe way. The + * original ptbl is freed using RCU callback. + * + * LOCKING: + * Matching bd_mutx locked. + */ +static void disk_replace_part_tbl(struct gendisk *disk, + struct disk_part_tbl *new_ptbl) +{ + struct disk_part_tbl *old_ptbl = disk->part_tbl; + + rcu_assign_pointer(disk->part_tbl, new_ptbl); + if (old_ptbl) + call_rcu(&old_ptbl->rcu_head, disk_free_ptbl_rcu_cb); +} + +/** + * disk_expand_part_tbl - expand disk->part_tbl + * @disk: disk to expand part_tbl for + * @partno: expand such that this partno can fit in + * + * Expand disk->part_tbl such that @partno can fit in. disk->part_tbl + * uses RCU to allow unlocked dereferencing for stats and other stuff. + * + * LOCKING: + * Matching bd_mutex locked, might sleep. + * + * RETURNS: + * 0 on success, -errno on failure. + */ +int disk_expand_part_tbl(struct gendisk *disk, int partno) +{ + struct disk_part_tbl *old_ptbl = disk->part_tbl; + struct disk_part_tbl *new_ptbl; + int len = old_ptbl ? old_ptbl->len : 0; + int target = partno + 1; + size_t size; + int i; + + /* disk_max_parts() is zero during initialization, ignore if so */ + if (disk_max_parts(disk) && target > disk_max_parts(disk)) + return -EINVAL; + + if (target <= len) + return 0; + + size = sizeof(*new_ptbl) + target * sizeof(new_ptbl->part[0]); + new_ptbl = kzalloc_node(size, GFP_KERNEL, disk->node_id); + if (!new_ptbl) + return -ENOMEM; + + INIT_RCU_HEAD(&new_ptbl->rcu_head); + new_ptbl->len = target; + + for (i = 0; i < len; i++) + rcu_assign_pointer(new_ptbl->part[i], old_ptbl->part[i]); + + disk_replace_part_tbl(disk, new_ptbl); + return 0; +} + static void disk_release(struct device *dev) { struct gendisk *disk = dev_to_disk(dev); kfree(disk->random); - kfree(disk->__part); + disk_replace_part_tbl(disk, NULL); free_part_stats(&disk->part0); kfree(disk); } @@ -948,22 +1042,16 @@ struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) disk = kmalloc_node(sizeof(struct gendisk), GFP_KERNEL | __GFP_ZERO, node_id); if (disk) { - int tot_minors = minors + ext_minors; - int size = tot_minors * sizeof(struct hd_struct *); - if (!init_part_stats(&disk->part0)) { kfree(disk); return NULL; } - - disk->__part = kmalloc_node(size, GFP_KERNEL | __GFP_ZERO, - node_id); - if (!disk->__part) { - free_part_stats(&disk->part0); + if (disk_expand_part_tbl(disk, 0)) { + free_part_stats(&disk->part0); kfree(disk); return NULL; } - disk->__part[0] = &disk->part0; + disk->part_tbl->part[0] = &disk->part0; disk->minors = minors; disk->ext_minors = ext_minors; @@ -973,6 +1061,7 @@ struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) device_initialize(disk_to_dev(disk)); INIT_WORK(&disk->async_notify, media_change_notify_thread); + disk->node_id = node_id; } return disk; } diff --git a/block/ioctl.c b/block/ioctl.c index 64e7c67a64b..38bee321e1f 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -30,7 +30,7 @@ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user if (bdev != bdev->bd_contains) return -EINVAL; partno = p.pno; - if (partno <= 0 || partno >= disk_max_parts(disk)) + if (partno <= 0) return -EINVAL; switch (a.op) { case BLKPG_ADD_PARTITION: diff --git a/fs/partitions/check.c b/fs/partitions/check.c index f517869e8d1..772b2ed8d23 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -312,14 +312,18 @@ static void delete_partition_rcu_cb(struct rcu_head *head) void delete_partition(struct gendisk *disk, int partno) { + struct disk_part_tbl *ptbl = disk->part_tbl; struct hd_struct *part; - part = disk->__part[partno]; + if (partno >= ptbl->len) + return; + + part = ptbl->part[partno]; if (!part) return; blk_free_devt(part_devt(part)); - rcu_assign_pointer(disk->__part[partno], NULL); + rcu_assign_pointer(ptbl->part[partno], NULL); kobject_put(part->holder_dir); device_del(part_to_dev(part)); @@ -341,10 +345,16 @@ int add_partition(struct gendisk *disk, int partno, dev_t devt = MKDEV(0, 0); struct device *ddev = disk_to_dev(disk); struct device *pdev; + struct disk_part_tbl *ptbl; const char *dname; int err; - if (disk->__part[partno]) + err = disk_expand_part_tbl(disk, partno); + if (err) + return err; + ptbl = disk->part_tbl; + + if (ptbl->part[partno]) return -EBUSY; p = kzalloc(sizeof(*p), GFP_KERNEL); @@ -398,7 +408,7 @@ int add_partition(struct gendisk *disk, int partno, /* everything is up and running, commence */ INIT_RCU_HEAD(&p->rcu_head); - rcu_assign_pointer(disk->__part[partno], p); + rcu_assign_pointer(ptbl->part[partno], p); /* suppress uevent if the disk supresses it */ if (!ddev->uevent_suppress) @@ -487,7 +497,7 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) struct disk_part_iter piter; struct hd_struct *part; struct parsed_partitions *state; - int p, res; + int p, highest, res; if (bdev->bd_part_count) return -EBUSY; @@ -511,6 +521,17 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) /* tell userspace that the media / partition table may have changed */ kobject_uevent(&disk_to_dev(disk)->kobj, KOBJ_CHANGE); + /* Detect the highest partition number and preallocate + * disk->part_tbl. This is an optimization and not strictly + * necessary. + */ + for (p = 1, highest = 0; p < state->limit; p++) + if (state->parts[p].size) + highest = p; + + disk_expand_part_tbl(disk, highest); + + /* add partitions */ for (p = 1; p < state->limit; p++) { sector_t size = state->parts[p].size; sector_t from = state->parts[p].from; diff --git a/include/linux/genhd.h b/include/linux/genhd.h index c90e1b4fbe5..ecf649c3dee 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -113,6 +113,21 @@ struct hd_struct { #define GENHD_FL_UP 16 #define GENHD_FL_SUPPRESS_PARTITION_INFO 32 +#define BLK_SCSI_MAX_CMDS (256) +#define BLK_SCSI_CMD_PER_LONG (BLK_SCSI_MAX_CMDS / (sizeof(long) * 8)) + +struct blk_scsi_cmd_filter { + unsigned long read_ok[BLK_SCSI_CMD_PER_LONG]; + unsigned long write_ok[BLK_SCSI_CMD_PER_LONG]; + struct kobject kobj; +}; + +struct disk_part_tbl { + struct rcu_head rcu_head; + int len; + struct hd_struct *part[]; +}; + struct gendisk { /* major, first_minor, minors and ext_minors are input * parameters only, don't use directly. Use disk_devt() and @@ -131,7 +146,7 @@ struct gendisk { * non-critical accesses use RCU. Always access through * helpers. */ - struct hd_struct **__part; + struct disk_part_tbl *part_tbl; struct hd_struct part0; struct block_device_operations *fops; @@ -149,6 +164,7 @@ struct gendisk { #ifdef CONFIG_BLK_DEV_INTEGRITY struct blk_integrity *integrity; #endif + int node_id; }; static inline struct gendisk *part_to_disk(struct hd_struct *part) @@ -503,6 +519,7 @@ extern void blk_free_devt(dev_t devt); extern dev_t blk_lookup_devt(const char *name, int partno); extern char *disk_name (struct gendisk *hd, int partno, char *buf); +extern int disk_expand_part_tbl(struct gendisk *disk, int target); extern int rescan_partitions(struct gendisk *disk, struct block_device *bdev); extern int __must_check add_partition(struct gendisk *, int, sector_t, sector_t, int); extern void delete_partition(struct gendisk *, int); -- cgit v1.2.3-70-g09d2 From 689d6fac40b41c7bf154f362deaf442548e4dc81 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:16 +0900 Subject: block: replace @ext_minors with GENHD_FL_EXT_DEVT With previous changes, it's meaningless to limit the number of partitions. Replace @ext_minors with GENHD_FL_EXT_DEVT such that setting the flag allows the disk to have maximum number of allowed partitions (only limited by the number of entries in parsed_partitions as determined by MAX_PART constant). This kills not-too-pretty alloc_disk_ext[_node]() functions and makes @minors parameter to alloc_disk[_node]() unnecessary. The parameter is left alone to avoid disturbing the users. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/genhd.c | 16 +--------------- drivers/ide/ide-disk.c | 14 +++++--------- drivers/scsi/sd.c | 9 ++------- fs/partitions/check.h | 4 +--- include/linux/genhd.h | 16 ++++++++-------- 5 files changed, 17 insertions(+), 42 deletions(-) (limited to 'fs/partitions') diff --git a/block/genhd.c b/block/genhd.c index c2b14aa69d5..eedab5b4685 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -1024,18 +1024,9 @@ struct gendisk *alloc_disk(int minors) { return alloc_disk_node(minors, -1); } +EXPORT_SYMBOL(alloc_disk); struct gendisk *alloc_disk_node(int minors, int node_id) -{ - return alloc_disk_ext_node(minors, 0, node_id); -} - -struct gendisk *alloc_disk_ext(int minors, int ext_minors) -{ - return alloc_disk_ext_node(minors, ext_minors, -1); -} - -struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) { struct gendisk *disk; @@ -1054,7 +1045,6 @@ struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) disk->part_tbl->part[0] = &disk->part0; disk->minors = minors; - disk->ext_minors = ext_minors; rand_initialize_disk(disk); disk_to_dev(disk)->class = &block_class; disk_to_dev(disk)->type = &disk_type; @@ -1065,11 +1055,7 @@ struct gendisk *alloc_disk_ext_node(int minors, int ext_minors, int node_id) } return disk; } - -EXPORT_SYMBOL(alloc_disk); EXPORT_SYMBOL(alloc_disk_node); -EXPORT_SYMBOL(alloc_disk_ext); -EXPORT_SYMBOL(alloc_disk_ext_node); struct kobject *get_disk(struct gendisk *disk) { diff --git a/drivers/ide/ide-disk.c b/drivers/ide/ide-disk.c index a072df5053a..29c8ae75268 100644 --- a/drivers/ide/ide-disk.c +++ b/drivers/ide/ide-disk.c @@ -41,16 +41,12 @@ #include #include -#define IDE_DISK_PARTS (1 << PARTN_BITS) - #if !defined(CONFIG_DEBUG_BLOCK_EXT_DEVT) -#define IDE_DISK_MINORS IDE_DISK_PARTS +#define IDE_DISK_MINORS (1 << PARTN_BITS) #else #define IDE_DISK_MINORS 1 #endif -#define IDE_DISK_EXT_MINORS (IDE_DISK_PARTS - IDE_DISK_MINORS) - struct ide_disk_obj { ide_drive_t *drive; ide_driver_t *driver; @@ -1161,8 +1157,7 @@ static int ide_disk_probe(ide_drive_t *drive) if (!idkp) goto failed; - g = alloc_disk_ext_node(IDE_DISK_MINORS, IDE_DISK_EXT_MINORS, - hwif_to_node(drive->hwif)); + g = alloc_disk_node(IDE_DISK_MINORS, hwif_to_node(drive->hwif)); if (!g) goto out_free_idkp; @@ -1189,9 +1184,10 @@ static int ide_disk_probe(ide_drive_t *drive) drive->attach = 1; g->minors = IDE_DISK_MINORS; - g->ext_minors = IDE_DISK_EXT_MINORS; g->driverfs_dev = &drive->gendev; - g->flags = drive->removable ? GENHD_FL_REMOVABLE : 0; + g->flags |= GENHD_FL_EXT_DEVT; + if (drive->removable) + g->flags |= GENHD_FL_REMOVABLE; set_capacity(g, idedisk_capacity(drive)); g->fops = &idedisk_ops; add_disk(g); diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c index 280d231a86e..6598024531d 100644 --- a/drivers/scsi/sd.c +++ b/drivers/scsi/sd.c @@ -86,16 +86,12 @@ MODULE_ALIAS_SCSI_DEVICE(TYPE_DISK); MODULE_ALIAS_SCSI_DEVICE(TYPE_MOD); MODULE_ALIAS_SCSI_DEVICE(TYPE_RBC); -#define SD_PARTS 64 - #if !defined(CONFIG_DEBUG_BLOCK_EXT_DEVT) #define SD_MINORS 16 #else #define SD_MINORS 1 #endif -#define SD_EXT_MINORS (SD_PARTS - SD_MINORS) - static int sd_revalidate_disk(struct gendisk *); static int sd_probe(struct device *); static int sd_remove(struct device *); @@ -1811,7 +1807,7 @@ static int sd_probe(struct device *dev) if (!sdkp) goto out; - gd = alloc_disk_ext(SD_MINORS, SD_EXT_MINORS); + gd = alloc_disk(SD_MINORS); if (!gd) goto out_free; @@ -1856,7 +1852,6 @@ static int sd_probe(struct device *dev) gd->major = sd_major((index & 0xf0) >> 4); gd->first_minor = ((index & 0xf) << 4) | (index & 0xfff00); gd->minors = SD_MINORS; - gd->ext_minors = SD_EXT_MINORS; gd->fops = &sd_fops; if (index < 26) { @@ -1880,7 +1875,7 @@ static int sd_probe(struct device *dev) blk_queue_prep_rq(sdp->request_queue, sd_prep_fn); gd->driverfs_dev = &sdp->sdev_gendev; - gd->flags = GENHD_FL_DRIVERFS; + gd->flags = GENHD_FL_EXT_DEVT | GENHD_FL_DRIVERFS; if (sdp->removable) gd->flags |= GENHD_FL_REMOVABLE; diff --git a/fs/partitions/check.h b/fs/partitions/check.h index 17ae8ecd9e8..98dbe1a8452 100644 --- a/fs/partitions/check.h +++ b/fs/partitions/check.h @@ -5,15 +5,13 @@ * add_gd_partition adds a partitions details to the devices partition * description. */ -enum { MAX_PART = 256 }; - struct parsed_partitions { char name[BDEVNAME_SIZE]; struct { sector_t from; sector_t size; int flags; - } parts[MAX_PART]; + } parts[DISK_MAX_PARTS]; int next; int limit; }; diff --git a/include/linux/genhd.h b/include/linux/genhd.h index ecf649c3dee..04524c213de 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -58,6 +58,8 @@ enum { UNIXWARE_PARTITION = 0x63, /* Same as GNU_HURD and SCO Unix */ }; +#define DISK_MAX_PARTS 256 + #include #include #include @@ -112,6 +114,7 @@ struct hd_struct { #define GENHD_FL_CD 8 #define GENHD_FL_UP 16 #define GENHD_FL_SUPPRESS_PARTITION_INFO 32 +#define GENHD_FL_EXT_DEVT 64 /* allow extended devt */ #define BLK_SCSI_MAX_CMDS (256) #define BLK_SCSI_CMD_PER_LONG (BLK_SCSI_MAX_CMDS / (sizeof(long) * 8)) @@ -129,15 +132,13 @@ struct disk_part_tbl { }; struct gendisk { - /* major, first_minor, minors and ext_minors are input - * parameters only, don't use directly. Use disk_devt() and - * disk_max_parts(). + /* major, first_minor and minors are input parameters only, + * don't use directly. Use disk_devt() and disk_max_parts(). */ int major; /* major number of driver */ int first_minor; int minors; /* maximum number of minors, =1 for * disks that can't be partitioned. */ - int ext_minors; /* number of extended dynamic minors */ char disk_name[32]; /* name of major driver */ @@ -180,7 +181,9 @@ static inline struct gendisk *part_to_disk(struct hd_struct *part) static inline int disk_max_parts(struct gendisk *disk) { - return disk->minors + disk->ext_minors; + if (disk->flags & GENHD_FL_EXT_DEVT) + return DISK_MAX_PARTS; + return disk->minors; } static inline bool disk_partitionable(struct gendisk *disk) @@ -527,9 +530,6 @@ extern void printk_all_partitions(void); extern struct gendisk *alloc_disk_node(int minors, int node_id); extern struct gendisk *alloc_disk(int minors); -extern struct gendisk *alloc_disk_ext_node(int minors, int ext_minrs, - int node_id); -extern struct gendisk *alloc_disk_ext(int minors, int ext_minors); extern struct kobject *get_disk(struct gendisk *disk); extern void put_disk(struct gendisk *disk); extern void blk_register_region(dev_t devt, unsigned long range, -- cgit v1.2.3-70-g09d2 From 3e1a7ff8a0a7b948f2684930166954f9e8e776fe Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 25 Aug 2008 19:56:17 +0900 Subject: block: allow disk to have extended device number Now that disk and partition handlings are mostly unified, it's easy to allow disk to have extended device number. This patch makes add_disk() use extended device number if disk->minors is zero. Both sd and ide-disk are updated to use this. * sd_format_disk_name() is implemented which can generically determine the drive name. This removes disk number restriction stemming from limited device names. * If sd index goes over SD_MAX_DISKS (which can be increased now BTW), sd simply doesn't initialize minors letting block layer choose extended device number. * If CONFIG_DEBUG_EXT_DEVT is set, both sd and ide-disk always set minors to 0 and use extended device numbers. Signed-off-by: Tejun Heo Signed-off-by: Jens Axboe --- block/genhd.c | 25 ++++++++++++++++- drivers/ide/ide-disk.c | 2 +- drivers/scsi/sd.c | 74 ++++++++++++++++++++++++++++++++++++-------------- fs/partitions/check.c | 1 + include/linux/genhd.h | 3 +- 5 files changed, 82 insertions(+), 23 deletions(-) (limited to 'fs/partitions') diff --git a/block/genhd.c b/block/genhd.c index eedab5b4685..d9de3e482d1 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -478,14 +478,37 @@ static int exact_lock(dev_t devt, void *data) * * This function registers the partitioning information in @disk * with the kernel. + * + * FIXME: error handling */ void add_disk(struct gendisk *disk) { struct backing_dev_info *bdi; + dev_t devt; int retval; + /* minors == 0 indicates to use ext devt from part0 and should + * be accompanied with EXT_DEVT flag. Make sure all + * parameters make sense. + */ + WARN_ON(disk->minors && !(disk->major || disk->first_minor)); + WARN_ON(!disk->minors && !(disk->flags & GENHD_FL_EXT_DEVT)); + disk->flags |= GENHD_FL_UP; - disk_to_dev(disk)->devt = MKDEV(disk->major, disk->first_minor); + + retval = blk_alloc_devt(&disk->part0, &devt); + if (retval) { + WARN_ON(1); + return; + } + disk_to_dev(disk)->devt = devt; + + /* ->major and ->first_minor aren't supposed to be + * dereferenced from here on, but set them just in case. + */ + disk->major = MAJOR(devt); + disk->first_minor = MINOR(devt); + blk_register_region(disk_devt(disk), disk->minors, NULL, exact_match, exact_lock, disk); register_disk(disk); diff --git a/drivers/ide/ide-disk.c b/drivers/ide/ide-disk.c index 29c8ae75268..33ea8c04871 100644 --- a/drivers/ide/ide-disk.c +++ b/drivers/ide/ide-disk.c @@ -44,7 +44,7 @@ #if !defined(CONFIG_DEBUG_BLOCK_EXT_DEVT) #define IDE_DISK_MINORS (1 << PARTN_BITS) #else -#define IDE_DISK_MINORS 1 +#define IDE_DISK_MINORS 0 #endif struct ide_disk_obj { diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c index 6598024531d..bcb04b2a767 100644 --- a/drivers/scsi/sd.c +++ b/drivers/scsi/sd.c @@ -89,7 +89,7 @@ MODULE_ALIAS_SCSI_DEVICE(TYPE_RBC); #if !defined(CONFIG_DEBUG_BLOCK_EXT_DEVT) #define SD_MINORS 16 #else -#define SD_MINORS 1 +#define SD_MINORS 0 #endif static int sd_revalidate_disk(struct gendisk *); @@ -1769,6 +1769,52 @@ static int sd_revalidate_disk(struct gendisk *disk) return 0; } +/** + * sd_format_disk_name - format disk name + * @prefix: name prefix - ie. "sd" for SCSI disks + * @index: index of the disk to format name for + * @buf: output buffer + * @buflen: length of the output buffer + * + * SCSI disk names starts at sda. The 26th device is sdz and the + * 27th is sdaa. The last one for two lettered suffix is sdzz + * which is followed by sdaaa. + * + * This is basically 26 base counting with one extra 'nil' entry + * at the beggining from the second digit on and can be + * determined using similar method as 26 base conversion with the + * index shifted -1 after each digit is computed. + * + * CONTEXT: + * Don't care. + * + * RETURNS: + * 0 on success, -errno on failure. + */ +static int sd_format_disk_name(char *prefix, int index, char *buf, int buflen) +{ + const int base = 'z' - 'a' + 1; + char *begin = buf + strlen(prefix); + char *end = buf + buflen; + char *p; + int unit; + + p = end - 1; + *p = '\0'; + unit = base; + do { + if (p == begin) + return -EINVAL; + *--p = 'a' + (index % unit); + index = (index / unit) - 1; + } while (index >= 0); + + memmove(begin, p, end - p); + memcpy(buf, prefix, strlen(prefix)); + + return 0; +} + /** * sd_probe - called during driver initialization and whenever a * new scsi device is attached to the system. It is called once @@ -1821,8 +1867,8 @@ static int sd_probe(struct device *dev) if (error) goto out_put; - error = -EBUSY; - if (index >= SD_MAX_DISKS) + error = sd_format_disk_name("sd", index, gd->disk_name, DISK_NAME_LEN); + if (error) goto out_free_index; sdkp->device = sdp; @@ -1849,24 +1895,12 @@ static int sd_probe(struct device *dev) get_device(&sdp->sdev_gendev); - gd->major = sd_major((index & 0xf0) >> 4); - gd->first_minor = ((index & 0xf) << 4) | (index & 0xfff00); - gd->minors = SD_MINORS; - gd->fops = &sd_fops; - - if (index < 26) { - sprintf(gd->disk_name, "sd%c", 'a' + index % 26); - } else if (index < (26 + 1) * 26) { - sprintf(gd->disk_name, "sd%c%c", - 'a' + index / 26 - 1,'a' + index % 26); - } else { - const unsigned int m1 = (index / 26 - 1) / 26 - 1; - const unsigned int m2 = (index / 26 - 1) % 26; - const unsigned int m3 = index % 26; - sprintf(gd->disk_name, "sd%c%c%c", - 'a' + m1, 'a' + m2, 'a' + m3); + if (index < SD_MAX_DISKS) { + gd->major = sd_major((index & 0xf0) >> 4); + gd->first_minor = ((index & 0xf) << 4) | (index & 0xfff00); + gd->minors = SD_MINORS; } - + gd->fops = &sd_fops; gd->private_data = &sdkp->driver; gd->queue = sdkp->device->request_queue; diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 772b2ed8d23..0e411603fdf 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -593,6 +593,7 @@ void del_gendisk(struct gendisk *disk) disk_part_iter_exit(&piter); invalidate_partition(disk, 0); + blk_free_devt(disk_to_dev(disk)->devt); set_capacity(disk, 0); disk->flags &= ~GENHD_FL_UP; unlink_gendisk(disk); diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 04524c213de..206cdf96c3a 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -59,6 +59,7 @@ enum { }; #define DISK_MAX_PARTS 256 +#define DISK_NAME_LEN 32 #include #include @@ -140,7 +141,7 @@ struct gendisk { int minors; /* maximum number of minors, =1 for * disks that can't be partitioned. */ - char disk_name[32]; /* name of major driver */ + char disk_name[DISK_NAME_LEN]; /* name of major driver */ /* Array of pointers to partitions indexed by partno. * Protected with matching bdev lock but stat and other -- cgit v1.2.3-70-g09d2 From 9bc3ffbfbdf71fefda8a261ef8d6fdc388a29b42 Mon Sep 17 00:00:00 2001 From: Andrew Patterson Date: Thu, 4 Sep 2008 14:27:30 -0600 Subject: Check for device resize when rescanning partitions Check for device resize in the rescan_partitions() routine. If the device has been resized, the bdev size is set to match. The rescan_partitions() routine is called when opening the device and when calling the BLKRRPART ioctl. Signed-off-by: Andrew Patterson Signed-off-by: Jens Axboe --- fs/partitions/check.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs/partitions') diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 0e411603fdf..7408227c49c 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -504,7 +504,6 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) res = invalidate_partition(disk, 0); if (res) return res; - bdev->bd_invalidated = 0; disk_part_iter_init(&piter, disk, DISK_PITER_INCL_EMPTY); while ((part = disk_part_iter_next(&piter))) @@ -513,6 +512,8 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) if (disk->fops->revalidate_disk) disk->fops->revalidate_disk(disk); + check_disk_size_change(disk, bdev); + bdev->bd_invalidated = 0; if (!get_capacity(disk) || !(state = check_partition(disk, bdev))) return 0; if (IS_ERR(state)) /* I/O error reading the partition table */ -- cgit v1.2.3-70-g09d2