summaryrefslogtreecommitdiff
path: root/include/linux
diff options
context:
space:
mode:
authorVivek Goyal <vgoyal@redhat.com>2012-08-01 12:24:18 +0200
committerJens Axboe <axboe@kernel.dk>2012-08-01 12:24:18 +0200
commitc83f6bf98dc1f1a194118b3830706cebbebda8c4 (patch)
treeea8fbd925584f784164617964a9f025bda16ed15 /include/linux
parent4638a83e8615de9c16c39dfed234951d0f468cf1 (diff)
downloadlwn-c83f6bf98dc1f1a194118b3830706cebbebda8c4.tar.gz
lwn-c83f6bf98dc1f1a194118b3830706cebbebda8c4.zip
block: add partition resize function to blkpg ioctl
Add a new operation code (BLKPG_RESIZE_PARTITION) to the BLKPG ioctl that allows altering the size of an existing partition, even if it is currently in use. This patch converts hd_struct->nr_sects into sequence counter because One might extend a partition while IO is happening to it and update of nr_sects can be non-atomic on 32bit machines with 64bit sector_t. This can lead to issues like reading inconsistent size of a partition. Sequence counter have been used so that readers don't have to take bdev mutex lock as we call sector_in_part() very frequently. Now all the access to hd_struct->nr_sects should happen using sequence counter read/update helper functions part_nr_sects_read/part_nr_sects_write. There is one exception though, set_capacity()/get_capacity(). I think theoritically race should exist there too but this patch does not modify set_capacity()/get_capacity() due to sheer number of call sites and I am afraid that change might break something. I have left that as a TODO item. We can handle it later if need be. This patch does not introduce any new races as such w.r.t set_capacity()/get_capacity(). v2: Add CONFIG_LBDAF test to UP preempt case as suggested by Phillip. Signed-off-by: Vivek Goyal <vgoyal@redhat.com> Signed-off-by: Phillip Susi <psusi@ubuntu.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Diffstat (limited to 'include/linux')
-rw-r--r--include/linux/blkpg.h1
-rw-r--r--include/linux/genhd.h57
2 files changed, 58 insertions, 0 deletions
diff --git a/include/linux/blkpg.h b/include/linux/blkpg.h
index faf8a45af210..a8519446c111 100644
--- a/include/linux/blkpg.h
+++ b/include/linux/blkpg.h
@@ -40,6 +40,7 @@ struct blkpg_ioctl_arg {
/* The subfunctions (for the op field) */
#define BLKPG_ADD_PARTITION 1
#define BLKPG_DEL_PARTITION 2
+#define BLKPG_RESIZE_PARTITION 3
/* Sizes of name fields. Unused at present. */
#define BLKPG_DEVNAMELTH 64
diff --git a/include/linux/genhd.h b/include/linux/genhd.h
index 017a7fb5a1fc..b88723b81b3d 100644
--- a/include/linux/genhd.h
+++ b/include/linux/genhd.h
@@ -98,7 +98,13 @@ struct partition_meta_info {
struct hd_struct {
sector_t start_sect;
+ /*
+ * nr_sects is protected by sequence counter. One might extend a
+ * partition while IO is happening to it and update of nr_sects
+ * can be non-atomic on 32bit machines with 64bit sector_t.
+ */
sector_t nr_sects;
+ seqcount_t nr_sects_seq;
sector_t alignment_offset;
unsigned int discard_alignment;
struct device __dev;
@@ -648,6 +654,57 @@ static inline void hd_struct_put(struct hd_struct *part)
__delete_partition(part);
}
+/*
+ * Any access of part->nr_sects which is not protected by partition
+ * bd_mutex or gendisk bdev bd_mutex, should be done using this
+ * accessor function.
+ *
+ * Code written along the lines of i_size_read() and i_size_write().
+ * CONFIG_PREEMPT case optimizes the case of UP kernel with preemption
+ * on.
+ */
+static inline sector_t part_nr_sects_read(struct hd_struct *part)
+{
+#if BITS_PER_LONG==32 && defined(CONFIG_LBDAF) && defined(CONFIG_SMP)
+ sector_t nr_sects;
+ unsigned seq;
+ do {
+ seq = read_seqcount_begin(&part->nr_sects_seq);
+ nr_sects = part->nr_sects;
+ } while (read_seqcount_retry(&part->nr_sects_seq, seq));
+ return nr_sects;
+#elif BITS_PER_LONG==32 && defined(CONFIG_LBDAF) && defined(CONFIG_PREEMPT)
+ sector_t nr_sects;
+
+ preempt_disable();
+ nr_sects = part->nr_sects;
+ preempt_enable();
+ return nr_sects;
+#else
+ return part->nr_sects;
+#endif
+}
+
+/*
+ * Should be called with mutex lock held (typically bd_mutex) of partition
+ * to provide mutual exlusion among writers otherwise seqcount might be
+ * left in wrong state leaving the readers spinning infinitely.
+ */
+static inline void part_nr_sects_write(struct hd_struct *part, sector_t size)
+{
+#if BITS_PER_LONG==32 && defined(CONFIG_LBDAF) && defined(CONFIG_SMP)
+ write_seqcount_begin(&part->nr_sects_seq);
+ part->nr_sects = size;
+ write_seqcount_end(&part->nr_sects_seq);
+#elif BITS_PER_LONG==32 && defined(CONFIG_LBDAF) && defined(CONFIG_PREEMPT)
+ preempt_disable();
+ part->nr_sects = size;
+ preempt_enable();
+#else
+ part->nr_sects = size;
+#endif
+}
+
#else /* CONFIG_BLOCK */
static inline void printk_all_partitions(void) { }