On 2/27/20 1:29 AM, Denis Plotnikov wrote:
The patch adds some preparation parts for incompatible compression type
feature to Qcow2 that indicates which allow to use different compression
to qcow2, allowing the use of different
methods for image clusters (de)compressing.
It is implied that the compression type is set on the image creation and
can be changed only later by image conversion, thus compression type
defines the only compression algorithm used for the image, and thus,
for all image clusters.
The goal of the feature is to add support of other compression methods
to qcow2. For example, ZSTD which is more effective on compression than ZLIB.
The default compression is ZLIB. Images created with ZLIB compression type
are backward compatible with older qemu versions.
Signed-off-by: Denis Plotnikov <dplotni...@virtuozzo.com>
---
block/qcow2.c | 105 ++++++++++++++++++++++++++++++++++++++
block/qcow2.h | 31 ++++++++---
include/block/block_int.h | 1 +
qapi/block-core.json | 22 +++++++-
4 files changed, 150 insertions(+), 9 deletions(-)
diff --git a/block/qcow2.c b/block/qcow2.c
index 3c754f616b..2ccb2cabd1 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -1242,6 +1242,50 @@ static int qcow2_update_options(BlockDriverState *bs,
QDict *options,
return ret;
}
+static int validate_compression_type(BDRVQcow2State *s, Error **errp)
+{
+ /*
+ * Sanity check
+ * according to qcow2 spec, the compression type is 1-byte field
+ * but in BDRVQcow2State the compression_type is enum sizeof(int)
+ * so, the max compression_type value is 255.
+ */
+ if (s->compression_type > 0xff) {
+ error_setg(errp, "qcow2: compression type value is too big");
+ return -EINVAL;
+ }
Hmm - I think it may be worth a tweak to qcow2.txt to call out:
104: compression_type
105 - 111: padding, must be 0
or else call out:
104-111: compression type
and just blindly use all 8 bytes for the value even though really only 1 or two
values will ever be defined. Of course, that moves the byte in question from
104 to 111, thanks to our big endian encoding, but as this series is the first
one installing a non-zero value in those 8 bytes, and as we just finished
documenting that the header length must be a multiple of 8, there is no real
impact - we can make such tweaks up until the 5.0 release.
+
+ switch (s->compression_type) {
+ case QCOW2_COMPRESSION_TYPE_ZLIB:
+ break;
+
+ default:
+ error_setg(errp, "qcow2: unknown compression type: %u",
+ s->compression_type);
+ return -ENOTSUP;
+ }
Having two checks feels redundant, compared to just letting the default catch
all unrecognized values in that field.
+
+ /*
+ * if the compression type differs from QCOW2_COMPRESSION_TYPE_ZLIB
+ * the incompatible feature flag must be set
+ */
+ if (s->compression_type == QCOW2_COMPRESSION_TYPE_ZLIB) {
+ if (s->incompatible_features & QCOW2_INCOMPAT_COMPRESSION_TYPE) {
+ error_setg(errp, "qcow2: Compression type incompatible feature "
+ "bit must not be set");
+ return -EINVAL;
+ }
+ } else {
+ if (!(s->incompatible_features & QCOW2_INCOMPAT_COMPRESSION_TYPE)) {
+ error_setg(errp, "qcow2: Compression type incompatible feature "
+ "bit must be set");
+ return -EINVAL;
+ }
+ }
Matches what we documented in the spec.
+
+ return 0;
+}
+
/* Called with s->lock held. */
static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options,
int flags, Error **errp)
@@ -1357,6 +1401,26 @@ static int coroutine_fn qcow2_do_open(BlockDriverState
*bs, QDict *options,
s->compatible_features = header.compatible_features;
s->autoclear_features = header.autoclear_features;
+ /*
+ * Handle compression type
+ * Older qcow2 images don't contain the compression type header.
+ * Distinguish them by the header length and use
+ * the only valid (default) compression type in that case
+ */
+ if (header.header_length > offsetof(QCowHeader, compression_type)) {
+ /*
+ * don't deal with endians since compression_type is 1 byte long
+ */
+ s->compression_type = header.compression_type;
Changes if you go with my suggestion of just making the compression_type field
occupy 8 bytes in the qcow2 header. (And if you want to keep it 1 byte, I
still think the spec should call out explicit padding bytes).
+ } else {
+ s->compression_type = QCOW2_COMPRESSION_TYPE_ZLIB;
+ }
+
+ ret = validate_compression_type(s, errp);
+ if (ret) {
+ goto fail;
+ }
+
if (s->incompatible_features & ~QCOW2_INCOMPAT_MASK) {
void *feature_table = NULL;
qcow2_read_extensions(bs, header.header_length, ext_end,
@@ -2720,6 +2784,12 @@ int qcow2_update_header(BlockDriverState *bs)
total_size = bs->total_sectors * BDRV_SECTOR_SIZE;
refcount_table_clusters = s->refcount_table_size >> (s->cluster_bits - 3);
+ ret = validate_compression_type(s, NULL);
+
+ if (ret) {
+ goto fail;
+ }
+
*header = (QCowHeader) {
/* Version 2 fields */
.magic = cpu_to_be32(QCOW_MAGIC),
@@ -2742,6 +2812,7 @@ int qcow2_update_header(BlockDriverState *bs)
.autoclear_features = cpu_to_be64(s->autoclear_features),
.refcount_order = cpu_to_be32(s->refcount_order),
.header_length = cpu_to_be32(header_length),
+ .compression_type = (uint8_t) s->compression_type,
Is the cast necessary?
};
/* For older versions, write a shorter header */
@@ -2839,6 +2910,11 @@ int qcow2_update_header(BlockDriverState *bs)
.bit = QCOW2_COMPAT_LAZY_REFCOUNTS_BITNR,
.name = "lazy refcounts",
},
+ {
+ .type = QCOW2_FEAT_TYPE_INCOMPATIBLE,
Ordering: please group all the incompatible bits side-by-side (this should come
before the lazy refcount bit).
+ .bit = QCOW2_INCOMPAT_COMPRESSION_TYPE_BITNR,
+ .name = "compression type",
This change breaks iotests; at least 31, 36, and 61 need updates (I've got a
similar patch pending which fixes the fact that we forgot the autoclear bit
[1]). You'll need to squash in fixes for those at the same time.
[1] https://lists.gnu.org/archive/html/qemu-devel/2020-01/msg08069.html
+ },
};
ret = header_ext_add(buf, QCOW2_EXT_MAGIC_FEATURE_TABLE,
@@ -3401,6 +3477,7 @@ qcow2_co_create(BlockdevCreateOptions *create_options,
Error **errp)
.refcount_table_offset = cpu_to_be64(cluster_size),
.refcount_table_clusters = cpu_to_be32(1),
.refcount_order = cpu_to_be32(refcount_order),
+ .compression_type = (uint8_t) QCOW2_COMPRESSION_TYPE_ZLIB,
Is the cast necessary?
.header_length = cpu_to_be32(sizeof(*header)),
};
@@ -3420,6 +3497,26 @@ qcow2_co_create(BlockdevCreateOptions *create_options,
Error **errp)
cpu_to_be64(QCOW2_AUTOCLEAR_DATA_FILE_RAW);
}
+ if (qcow2_opts->has_compression_type &&
+ qcow2_opts->compression_type != QCOW2_COMPRESSION_TYPE_ZLIB) {
+
+ if (qcow2_opts->compression_type > 0xff) {
+ error_setg_errno(errp, -EINVAL, "Too big compression type value");
+ goto out;
+ }
+
+ switch (qcow2_opts->compression_type) {
+ default:
+ error_setg_errno(errp, -EINVAL, "Unknown compression type");
+ goto out;
+ }
This should probably be an assert that qcow2_opts->compression_type is in
range, rather than a switch statement and error_setg. Callers of qcow2_co_create
should not be handing us unknown values.
+
+ header->compression_type = (uint8_t) qcow2_opts->compression_type;
Why the cast?
+
+ header->incompatible_features |=
+ cpu_to_be64(QCOW2_INCOMPAT_COMPRESSION_TYPE);
+ }
+
ret = blk_pwrite(blk, 0, header, cluster_size, 0);
g_free(header);
if (ret < 0) {
@@ -3602,6 +3699,7 @@ static int coroutine_fn qcow2_co_create_opts(const char
*filename, QemuOpts *opt
+++ b/block/qcow2.h
@@ -146,6 +146,12 @@ typedef struct QCowHeader {
uint32_t refcount_order;
uint32_t header_length;
+
+ /* Additional fields */
+ uint8_t compression_type;
+
+ /* header must be a multiple of 8 */
+ uint8_t padding[7];
} QEMU_PACKED QCowHeader;
You're changing the size of this struct, which WILL break iotests (and even
more than just the 3 I pointed out above for the feature name table).
/me looks ahead
Aha - you even noticed it: patch 7/8 fixes test 80. That fix needs to be
squashed in here, where the change is made.
typedef struct QEMU_PACKED QCowSnapshotHeader {
@@ -213,16 +219,20 @@ enum {
/* Incompatible feature bits */
enum {
- QCOW2_INCOMPAT_DIRTY_BITNR = 0,
- QCOW2_INCOMPAT_CORRUPT_BITNR = 1,
- QCOW2_INCOMPAT_DATA_FILE_BITNR = 2,
- QCOW2_INCOMPAT_DIRTY = 1 << QCOW2_INCOMPAT_DIRTY_BITNR,
- QCOW2_INCOMPAT_CORRUPT = 1 << QCOW2_INCOMPAT_CORRUPT_BITNR,
- QCOW2_INCOMPAT_DATA_FILE = 1 << QCOW2_INCOMPAT_DATA_FILE_BITNR,
+ QCOW2_INCOMPAT_DIRTY_BITNR = 0,
+ QCOW2_INCOMPAT_CORRUPT_BITNR = 1,
+ QCOW2_INCOMPAT_DATA_FILE_BITNR = 2,
+ QCOW2_INCOMPAT_COMPRESSION_TYPE_BITNR = 3,
+ QCOW2_INCOMPAT_DIRTY = 1 << QCOW2_INCOMPAT_DIRTY_BITNR,
+ QCOW2_INCOMPAT_CORRUPT = 1 << QCOW2_INCOMPAT_CORRUPT_BITNR,
+ QCOW2_INCOMPAT_DATA_FILE = 1 <<
QCOW2_INCOMPAT_DATA_FILE_BITNR,
+ QCOW2_INCOMPAT_COMPRESSION_TYPE =
+ 1 << QCOW2_INCOMPAT_COMPRESSION_TYPE_BITNR,
Uggh. I hate realigning = just because we added a longer name, especially when
you then can't even keep things on one line because of length. If it were me,
I'd leave the existing lines alone, and/or switch everything to just use 'BITNR
= ' rather than trying to align =.
Bikeshedding - since the new name is so long, can you get by with the shorter
QCOW2_INCOMPAT_COMPRESSION_BITNR (drop the _TYPE)?
