Should we note somehow in qcow2 spec that we use streamed version of zstd with
specific end byte?
23.03.2020 17:25, Denis Plotnikov wrote:
zstd significantly reduces cluster compression time.
It provides better compression performance maintaining
the same level of the compression ratio in comparison with
zlib, which, at the moment, is the only compression
method available.
The performance test results:
Test compresses and decompresses qemu qcow2 image with just
installed rhel-7.6 guest.
Image cluster size: 64K. Image on disk size: 2.2G
The test was conducted with brd disk to reduce the influence
of disk subsystem to the test results.
The results is given in seconds.
compress cmd:
time ./qemu-img convert -O qcow2 -c -o compression_type=[zlib|zstd]
src.img [zlib|zstd]_compressed.img
decompress cmd
time ./qemu-img convert -O qcow2
[zlib|zstd]_compressed.img uncompressed.img
compression decompression
zlib zstd zlib zstd
------------------------------------------------------------
real 65.5 16.3 (-75 %) 1.9 1.6 (-16 %)
user 65.0 15.8 5.3 2.5
sys 3.3 0.2 2.0 2.0
Both ZLIB and ZSTD gave the same compression ratio: 1.57
compressed image size in both cases: 1.4G
Signed-off-by: Denis Plotnikov <dplotni...@virtuozzo.com>
QAPI part:
Acked-by: Markus Armbruster <arm...@redhat.com>
---
[..]
+
+/*
+ * qcow2_zstd_compress()
+ *
+ * Compress @src_size bytes of data using zstd compression method
+ *
+ * @dest - destination buffer, @dest_size bytes
+ * @src - source buffer, @src_size bytes
+ *
+ * Returns: compressed size on success
+ * -ENOMEM destination buffer is not enough to store compressed data
+ * -EIO on any other error
+ */
+static ssize_t qcow2_zstd_compress(void *dest, size_t dest_size,
+ const void *src, size_t src_size)
+{
+ size_t ret;
+ ZSTD_outBuffer output = { dest, dest_size, 0 };
+ ZSTD_inBuffer input = { src, src_size, 0 };
+ ZSTD_CCtx *cctx = ZSTD_createCCtx();
+
+ if (!cctx) {
+ return -EIO;
+ }
+ /*
+ * ZSTD spec: "You must continue calling ZSTD_compressStream2()
+ * with ZSTD_e_end until it returns 0, at which point you are
+ * free to start a new frame"
+ */
+ {
+ /*
+ * zstd simple interface requires the exact compressed size.
+ * zstd stream interface reads the comressed size from
+ * the compressed stream frame.
+ * Instruct zstd to compress the whole buffer and write
+ * the frame which includes the compressed size.
+ * This allows as to use zstd streaming semantics and
+ * don't store the compressed size for the zstd decompression.
+ */
+ ret = ZSTD_compressStream2(cctx, &output, &input, ZSTD_e_end);
+ if (ZSTD_isError(ret)) {
+ ret = -EIO;
+ goto out;
+ }
+ /* Dest buffer isn't big enough to store compressed content */
+ if (output.pos + ret > output.size) {
+ ret = -ENOMEM;
+ goto out;
+ }
+ } while (ret);
+
+ /* if no error, the input data must be fully consumed */
+ assert(input.pos == input.size);
+ /* make sure we can safely return compressed buffer size with ssize_t *//z
+ assert(output.pos <= SSIZE_MAX);
+ ret = output.pos;
+
+out:
+ ZSTD_freeCCtx(cctx);
+ return ret;
+}
+
+/*
+ * qcow2_zstd_decompress()
+ *
+ * Decompress some data (not more than @src_size bytes) to produce exactly
+ * @dest_size bytes using zstd compression method
+ *
+ * @dest - destination buffer, @dest_size bytes
+ * @src - source buffer, @src_size bytes
+ *
+ * Returns: 0 on success
+ * -EIO on any error
+ */
+static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size,
+ const void *src, size_t src_size)
+{
+ size_t ret = 0;
+ ZSTD_outBuffer output = { dest, dest_size, 0 };
+ ZSTD_inBuffer input = { src, src_size, 0 };
+ ZSTD_DCtx *dctx = ZSTD_createDCtx();
+
+ if (!dctx) {
+ return -EIO;
+ }
+
+ {
+ ret = ZSTD_decompressStream(dctx, &output, &input);
+ if (ZSTD_isError(ret)) {
+ ret = -EIO;
+ goto out;
+ }
+ /*
+ * Dest buffer size is the image cluster size.
+ * It should be big enough to store uncompressed content.
+ * There shouldn't be any cases when the decompressed content
+ * size is greater then the cluster size.
+ */
+ if (output.pos + ret > output.size) {
+ ret = -EIO;
+ goto out;
+ }
+ } while (ret);
Hmm. Unfortunately, zstd spec is not enough verbose to understand how to use
these functions :).
But I found this in comment in
https://github.com/facebook/zstd/blob/dev/examples/streaming_decompression.c :
/* The return code is zero if the frame is complete, but there may
* be multiple frames concatenated together.
so, I think it would be safer to move to "while (input.pos < input.size)" loop
like in example.
and drop next assertion. And possibly do same for compression to be consistent
(and safer?).
+
+ /*
+ * If decompression went fine we must have the compressed
+ * cluster fully consumed and flushed
+ */
+ if (output.pos != output.size) {
+ ret = -EIO;
+ }
+out:
+ ZSTD_freeDCtx(dctx);
+ return ret;
+
+}
+#endif
+
static int qcow2_compress_pool_func(void *opaque)
{
Qcow2CompressData *data = opaque;
@@ -217,6 +341,11 @@ qcow2_co_compress(BlockDriverState *bs, void *dest, size_t
dest_size,
[..]
--
Best regards,
Vladimir
