[Libguestfs] [PATCH] file: Zero support for block devices and NFS 4.2
Nir Soffer
nirsof at gmail.com
Sun Jul 29 12:04:10 UTC 2018
If we may not trim, we tried ZERO_RANGE, but this is not well supported
yet, for example it is not available on NFS 4.2. ZERO_RANGE and
PUNCH_HOLE are supported now on block devices, but not on RHRL 7, so we
fallback to slow manual zeroing there.
Change the logic to support block devices on RHEL 7, and file systems
that do not support ZERO_RANGE.
The new logic:
- If we may trim, try PUNCH_HOLE
- If we can zero range, Try ZERO_RANGE
- If we can punch hole and fallocate, try fallocate(PUNCH_HOLE) followed
by fallocate(0).
- If underlying file is a block device, try ioctl(BLKZEROOUT)
- Otherwise fallback to manual zeroing
The handle keeps now the underlying file capabilities, so once we
discover that an operation is not supported, we never try it again.
Here are examples runs on a server based on Intel(R) Xeon(R) CPU E5-2630
v4 @ 2.20GHz, using XtremIO storage via 4G FC HBA and 4 paths to
storage.
$ export SOCK=/tmp/nbd.sock
$ export BLOCK=/dev/e30bfac2-8e13-479d-8cd6-c6da5e306c4e/c9864222-bc52-4359-80d7-76e47d619b15
$ src/nbdkit -f plugins/file/.libs/nbdkit-file-plugin.so file=$BLOCK -U $SOCK
$ time qemu-img convert -n -f raw -O raw /var/tmp/fedora-27.img nbd:unix:$SOCK
real 0m2.741s
user 0m0.224s
sys 0m0.634s
$ time qemu-img convert -n -f raw -O raw -W /var/tmp/fedora-27.img nbd:unix:$SOCK
real 0m1.920s
user 0m0.163s
sys 0m0.735s
Issues:
- ioctl(BLKZEROOUT) will fail if offset or count are not aligned to
logical sector size. I'm not sure if nbdkit or qemu-img ensure this.
- Need testing with NFS
---
plugins/file/file.c | 126 ++++++++++++++++++++++++++++++++++++--------
1 file changed, 103 insertions(+), 23 deletions(-)
diff --git a/plugins/file/file.c b/plugins/file/file.c
index fb20622..bce2ed1 100644
--- a/plugins/file/file.c
+++ b/plugins/file/file.c
@@ -33,6 +33,7 @@
#include <config.h>
+#include <stdbool.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
@@ -42,6 +43,8 @@
#include <sys/stat.h>
#include <errno.h>
#include <linux/falloc.h> /* For FALLOC_FL_* on RHEL, glibc < 2.18 */
+#include <sys/ioctl.h>
+#include <linux/fs.h>
#include <nbdkit-plugin.h>
@@ -116,6 +119,10 @@ file_config_complete (void)
/* The per-connection handle. */
struct handle {
int fd;
+ bool is_block_device;
+ bool can_punch_hole;
+ bool can_zero_range;
+ bool can_fallocate;
};
/* Create the per-connection handle. */
@@ -123,6 +130,7 @@ static void *
file_open (int readonly)
{
struct handle *h;
+ struct stat statbuf;
int flags;
h = malloc (sizeof *h);
@@ -144,6 +152,23 @@ file_open (int readonly)
return NULL;
}
+ if (fstat (h->fd, &statbuf) == -1) {
+ nbdkit_error ("fstat: %s: %m", filename);
+ free (h);
+ return NULL;
+ }
+
+ h->is_block_device = S_ISBLK(statbuf.st_mode);
+
+ /* These flags will disabled if an operation is not supported. */
+#ifdef FALLOC_FL_PUNCH_HOLE
+ h->can_punch_hole = true;
+#endif
+#ifdef FALLOC_FL_ZERO_RANGE
+ h->can_zero_range = true;
+#endif
+ h->can_fallocate = true;
+
return h;
}
@@ -164,27 +189,29 @@ static int64_t
file_get_size (void *handle)
{
struct handle *h = handle;
- struct stat statbuf;
- if (fstat (h->fd, &statbuf) == -1) {
- nbdkit_error ("stat: %m");
- return -1;
- }
-
- if (S_ISBLK (statbuf.st_mode)) {
+ if (h->is_block_device) {
+ /* Block device, so st_size will not be the true size. */
off_t size;
- /* Block device, so st_size will not be the true size. */
size = lseek (h->fd, 0, SEEK_END);
if (size == -1) {
nbdkit_error ("lseek (to find device size): %m");
return -1;
}
+
return size;
- }
+ } else {
+ /* Regular file. */
+ struct stat statbuf;
+
+ if (fstat (h->fd, &statbuf) == -1) {
+ nbdkit_error ("fstat: %m");
+ return -1;
+ }
- /* Else regular file. */
- return statbuf.st_size;
+ return statbuf.st_size;
+ }
}
static int
@@ -250,33 +277,86 @@ file_pwrite (void *handle, const void *buf, uint32_t count, uint64_t offset)
static int
file_zero (void *handle, uint32_t count, uint64_t offset, int may_trim)
{
-#if defined(FALLOC_FL_PUNCH_HOLE) || defined(FALLOC_FL_ZERO_RANGE)
struct handle *h = handle;
-#endif
int r = -1;
#ifdef FALLOC_FL_PUNCH_HOLE
- if (may_trim) {
+ /* If we can and may trim, punching hole is our best option. */
+ if (h->can_punch_hole && may_trim) {
r = do_fallocate (h->fd, FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE,
offset, count);
- if (r == -1 && errno != EOPNOTSUPP) {
+ if (r == 0)
+ return 0;
+
+ if (errno != EOPNOTSUPP) {
nbdkit_error ("zero: %m");
+ return r;
}
- /* PUNCH_HOLE is older; if it is not supported, it is likely that
- ZERO_RANGE will not work either, so fall back to write. */
- return r;
+
+ h->can_punch_hole = false;
}
#endif
#ifdef FALLOC_FL_ZERO_RANGE
- r = do_fallocate (h->fd, FALLOC_FL_ZERO_RANGE, offset, count);
- if (r == -1 && errno != EOPNOTSUPP) {
- nbdkit_error ("zero: %m");
+ /* ZERO_RANGE is not well supported yet, but it the next best option. */
+ if (h->can_zero_range) {
+ r = do_fallocate (h->fd, FALLOC_FL_ZERO_RANGE, offset, count);
+ if (r == 0)
+ return 0;
+
+ if (errno != EOPNOTSUPP) {
+ nbdkit_error ("zero: %m");
+ return r;
+ }
+
+ h->can_zero_range = false;
}
-#else
+#endif
+
+#ifdef FALLOC_FL_PUNCH_HOLE
+ /* If we can punch hole but may not trim, we can combine punching hole and
+ fallocate to zero a range. This is much more efficient than writing zeros
+ manually. */
+ if (h->can_punch_hole && h->can_fallocate) {
+ r = do_fallocate (h->fd, FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE,
+ offset, count);
+ if (r == 0) {
+ r = do_fallocate(h->fd, 0, offset, count);
+ if (r == 0)
+ return 0;
+
+ if (errno != EOPNOTSUPP) {
+ nbdkit_error ("zero: %m");
+ return r;
+ }
+
+ h->can_fallocate = false;
+ } else {
+ if (errno != EOPNOTSUPP) {
+ nbdkit_error ("zero: %m");
+ return r;
+ }
+
+ h->can_punch_hole = false;
+ }
+ }
+#endif
+
+ /* For block devices, we can use BLKZEROOUT.
+ NOTE: count and offset must be aligned to logical block size. */
+ if (h->is_block_device) {
+ uint64_t range[2] = {offset, count};
+
+ r = ioctl(h->fd, BLKZEROOUT, &range);
+ if (r == 0)
+ return 0;
+
+ nbdkit_error("zero: %m");
+ return r;
+ }
+
/* Trigger a fall back to writing */
errno = EOPNOTSUPP;
-#endif
return r;
}
--
2.17.1
More information about the Libguestfs
mailing list