Skip to content

Commit

Permalink
lib/lzo: implement run-length encoding
Browse files Browse the repository at this point in the history 
Patch series "lib/lzo: run-length encoding support", v5.

Following on from the previous lzo-rle patchset:

  https://lkml.org/lkml/2018/11/30/972

This patchset contains only the RLE patches, and should be applied on
top of the non-RLE patches ( https://lkml.org/lkml/2019/2/5/366 ).

Previously, some questions were raised around the RLE patches.  I've
done some additional benchmarking to answer these questions.  In short:

 - RLE offers significant additional performance (data-dependent)

 - I didn't measure any regressions that were clearly outside the noise

One concern with this patchset was around performance - specifically,
measuring RLE impact separately from Matt Sealey's patches (CTZ & fast
copy).  I have done some additional benchmarking which I hope clarifies
the benefits of each part of the patchset.

Firstly, I've captured some memory via /dev/fmem from a Chromebook with
many tabs open which is starting to swap, and then split this into 4178
4k pages.  I've excluded the all-zero pages (as zram does), and also the
no-zero pages (which won't tell us anything about RLE performance).
This should give a realistic test dataset for zram.  What I found was
that the data is VERY bimodal: 44% of pages in this dataset contain 5%
or fewer zeros, and 44% contain over 90% zeros (30% if you include the
no-zero pages).  This supports the idea of special-casing zeros in zram.

Next, I've benchmarked four variants of lzo on these pages (on 64-bit
Arm at max frequency): baseline LZO; baseline + Matt Sealey's patches
(aka MS); baseline + RLE only; baseline + MS + RLE.  Numbers are for
weighted roundtrip throughput (the weighting reflects that zram does
more compression than decompression).

  https://drive.google.com/file/d/1VLtLjRVxgUNuWFOxaGPwJYhl_hMQXpHe/view?usp=sharing

Matt's patches help in all cases for Arm (and no effect on Intel), as
expected.

RLE also behaves as expected: with few zeros present, it makes no
difference; above ~75%, it gives a good improvement (50 - 300 MB/s on
top of the benefit from Matt's patches).

Best performance is seen with both MS and RLE patches.

Finally, I have benchmarked the same dataset on an x86-64 device.  Here,
the MS patches make no difference (as expected); RLE helps, similarly as
on Arm.  There were no definite regressions; allowing for observational
error, 0.1% (3/4178) of cases had a regression > 1 standard deviation,
of which the largest was 4.6% (1.2 standard deviations).  I think this
is probably within the noise.

  https://drive.google.com/file/d/1xCUVwmiGD0heEMx5gcVEmLBI4eLaageV/view?usp=sharing

One point to note is that the graphs show RLE appears to help very
slightly with no zeros present! This is because the extra code causes
the clang optimiser to change code layout in a way that happens to have
a significant benefit.  Taking baseline LZO and adding a do-nothing line
like "__builtin_prefetch(out_len);" immediately before the "goto next"
has the same effect.  So this is a real, but basically spurious effect -
it's small enough not to upset the overall findings.

This patch (of 3):

When using zram, we frequently encounter long runs of zero bytes.  This
adds a special case which identifies runs of zeros and encodes them
using run-length encoding.

This is faster for both compression and decompresion.  For high-entropy
data which doesn't hit this case, impact is minimal.

Compression ratio is within a few percent in all cases.

This modifies the bitstream in a way which is backwards compatible
(i.e., we can decompress old bitstreams, but old versions of lzo cannot
decompress new bitstreams).

Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Dave Rodgman <[email protected]>
Cc: David S. Miller <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Cc: Herbert Xu <[email protected]>
Cc: Markus F.X.J. Oberhumer <[email protected]>
Cc: Matt Sealey <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Nitin Gupta <[email protected]>
Cc: Richard Purdie <[email protected]>
Cc: Sergey Senozhatsky <[email protected]>
Cc: Sonny Rao <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
  • Loading branch information
@daverodgman @torvalds
daverodgman authored and torvalds committed Mar 8, 2019
1 parent 761b323 commit 5ee4014
Show file tree
Hide file tree
Showing 5 changed files with 181 additions and 43 deletions.
35 changes: 28 additions & 7 deletions Documentation/lzo.txt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -78,16 +78,30 @@ Description
is an implementation design choice independent on the algorithm or
encoding.

Versions

0: Original version
1: LZO-RLE

Version 1 of LZO implements an extension to encode runs of zeros using run
length encoding. This improves speed for data with many zeros, which is a
common case for zram. This modifies the bitstream in a backwards compatible way
(v1 can correctly decompress v0 compressed data, but v0 cannot read v1 data).

Byte sequences
==============

First byte encoding::

0..17 : follow regular instruction encoding, see below. It is worth
noting that codes 16 and 17 will represent a block copy from
the dictionary which is empty, and that they will always be
0..16 : follow regular instruction encoding, see below. It is worth
noting that code 16 will represent a block copy from the
dictionary which is empty, and that it will always be
invalid at this place.

17 : bitstream version. If the first byte is 17, the next byte
gives the bitstream version. If the first byte is not 17,
the bitstream version is 0.

18..21 : copy 0..3 literals
state = (byte - 17) = 0..3 [ copy <state> literals ]
skip byte
Expand Down Expand Up @@ -140,6 +154,11 @@ Byte sequences
state = S (copy S literals after this block)
End of stream is reached if distance == 16384

In version 1, this instruction is also used to encode a run of zeros if
distance = 0xbfff, i.e. H = 1 and the D bits are all 1.
In this case, it is followed by a fourth byte, X.
run length = ((X << 3) | (0 0 0 0 0 L L L)) + 4.

0 0 1 L L L L L (32..63)
Copy of small block within 16kB distance (preferably less than 34B)
length = 2 + (L ?: 31 + (zero_bytes * 255) + non_zero_byte)
Expand All @@ -165,7 +184,9 @@ Authors
=======

This document was written by Willy Tarreau <[email protected]> on 2014/07/19 during an
analysis of the decompression code available in Linux 3.16-rc5. The code is
tricky, it is possible that this document contains mistakes or that a few
corner cases were overlooked. In any case, please report any doubt, fix, or
proposed updates to the author(s) so that the document can be updated.
analysis of the decompression code available in Linux 3.16-rc5, and updated
by Dave Rodgman <[email protected]> on 2018/10/30 to introduce run-length
encoding. The code is tricky, it is possible that this document contains
mistakes or that a few corner cases were overlooked. In any case, please
report any doubt, fix, or proposed updates to the author(s) so that the
document can be updated.
2 changes: 1 addition & 1 deletion include/linux/lzo.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@
#define LZO1X_1_MEM_COMPRESS (8192 * sizeof(unsigned short))
#define LZO1X_MEM_COMPRESS LZO1X_1_MEM_COMPRESS

#define lzo1x_worst_compress(x) ((x) + ((x) / 16) + 64 + 3)
#define lzo1x_worst_compress(x) ((x) + ((x) / 16) + 64 + 3 + 2)

/* This requires 'wrkmem' of size LZO1X_1_MEM_COMPRESS */
int lzo1x_1_compress(const unsigned char *src, size_t src_len,
Expand Down
100 changes: 89 additions & 11 deletions lib/lzo/lzo1x_compress.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -20,7 +20,7 @@
static noinline size_t
lzo1x_1_do_compress(const unsigned char *in, size_t in_len,
unsigned char *out, size_t *out_len,
size_t ti, void *wrkmem)
size_t ti, void *wrkmem, signed char *state_offset)
{
const unsigned char *ip;
unsigned char *op;
Expand All @@ -35,27 +35,85 @@ lzo1x_1_do_compress(const unsigned char *in, size_t in_len,
ip += ti < 4 ? 4 - ti : 0;

for (;;) {
const unsigned char *m_pos;
const unsigned char *m_pos = NULL;
size_t t, m_len, m_off;
u32 dv;
u32 run_length = 0;
literal:
ip += 1 + ((ip - ii) >> 5);
next:
if (unlikely(ip >= ip_end))
break;
dv = get_unaligned_le32(ip);
t = ((dv * 0x1824429d) >> (32 - D_BITS)) & D_MASK;
m_pos = in + dict[t];
dict[t] = (lzo_dict_t) (ip - in);
if (unlikely(dv != get_unaligned_le32(m_pos)))
goto literal;

if (dv == 0) {
const unsigned char *ir = ip + 4;
const unsigned char *limit = ip_end
< (ip + MAX_ZERO_RUN_LENGTH + 1)
? ip_end : ip + MAX_ZERO_RUN_LENGTH + 1;
#if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) && \
defined(LZO_FAST_64BIT_MEMORY_ACCESS)
u64 dv64;

for (; (ir + 32) <= limit; ir += 32) {
dv64 = get_unaligned((u64 *)ir);
dv64 |= get_unaligned((u64 *)ir + 1);
dv64 |= get_unaligned((u64 *)ir + 2);
dv64 |= get_unaligned((u64 *)ir + 3);
if (dv64)
break;
}
for (; (ir + 8) <= limit; ir += 8) {
dv64 = get_unaligned((u64 *)ir);
if (dv64) {
# if defined(__LITTLE_ENDIAN)
ir += __builtin_ctzll(dv64) >> 3;
# elif defined(__BIG_ENDIAN)
ir += __builtin_clzll(dv64) >> 3;
# else
# error "missing endian definition"
# endif
break;
}
}
#else
while ((ir < (const unsigned char *)
ALIGN((uintptr_t)ir, 4)) &&
(ir < limit) && (*ir == 0))
ir++;
for (; (ir + 4) <= limit; ir += 4) {
dv = *((u32 *)ir);
if (dv) {
# if defined(__LITTLE_ENDIAN)
ir += __builtin_ctz(dv) >> 3;
# elif defined(__BIG_ENDIAN)
ir += __builtin_clz(dv) >> 3;
# else
# error "missing endian definition"
# endif
break;
}
}
#endif
while (likely(ir < limit) && unlikely(*ir == 0))
ir++;
run_length = ir - ip;
if (run_length > MAX_ZERO_RUN_LENGTH)
run_length = MAX_ZERO_RUN_LENGTH;
} else {
t = ((dv * 0x1824429d) >> (32 - D_BITS)) & D_MASK;
m_pos = in + dict[t];
dict[t] = (lzo_dict_t) (ip - in);
if (unlikely(dv != get_unaligned_le32(m_pos)))
goto literal;
}

ii -= ti;
ti = 0;
t = ip - ii;
if (t != 0) {
if (t <= 3) {
op[-2] |= t;
op[*state_offset] |= t;
COPY4(op, ii);
op += t;
} else if (t <= 16) {
Expand Down Expand Up @@ -88,6 +146,17 @@ lzo1x_1_do_compress(const unsigned char *in, size_t in_len,
}
}

if (unlikely(run_length)) {
ip += run_length;
run_length -= MIN_ZERO_RUN_LENGTH;
put_unaligned_le32((run_length << 21) | 0xfffc18
| (run_length & 0x7), op);
op += 4;
run_length = 0;
*state_offset = -3;
goto finished_writing_instruction;
}

m_len = 4;
{
#if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) && defined(LZO_USE_CTZ64)
Expand Down Expand Up @@ -170,7 +239,6 @@ lzo1x_1_do_compress(const unsigned char *in, size_t in_len,

m_off = ip - m_pos;
ip += m_len;
ii = ip;
if (m_len <= M2_MAX_LEN && m_off <= M2_MAX_OFFSET) {
m_off -= 1;
*op++ = (((m_len - 1) << 5) | ((m_off & 7) << 2));
Expand Down Expand Up @@ -207,6 +275,9 @@ lzo1x_1_do_compress(const unsigned char *in, size_t in_len,
*op++ = (m_off << 2);
*op++ = (m_off >> 6);
}
*state_offset = -2;
finished_writing_instruction:
ii = ip;
goto next;
}
*out_len = op - out;
Expand All @@ -221,6 +292,12 @@ int lzo1x_1_compress(const unsigned char *in, size_t in_len,
unsigned char *op = out;
size_t l = in_len;
size_t t = 0;
signed char state_offset = -2;

// LZO v0 will never write 17 as first byte,
// so this is used to version the bitstream
*op++ = 17;
*op++ = LZO_VERSION;

while (l > 20) {
size_t ll = l <= (M4_MAX_OFFSET + 1) ? l : (M4_MAX_OFFSET + 1);
Expand All @@ -229,7 +306,8 @@ int lzo1x_1_compress(const unsigned char *in, size_t in_len,
break;
BUILD_BUG_ON(D_SIZE * sizeof(lzo_dict_t) > LZO1X_1_MEM_COMPRESS);
memset(wrkmem, 0, D_SIZE * sizeof(lzo_dict_t));
t = lzo1x_1_do_compress(ip, ll, op, out_len, t, wrkmem);
t = lzo1x_1_do_compress(ip, ll, op, out_len,
t, wrkmem, &state_offset);
ip += ll;
op += *out_len;
l -= ll;
Expand All @@ -242,7 +320,7 @@ int lzo1x_1_compress(const unsigned char *in, size_t in_len,
if (op == out && t <= 238) {
*op++ = (17 + t);
} else if (t <= 3) {
op[-2] |= t;
op[state_offset] |= t;
} else if (t <= 18) {
*op++ = (t - 3);
} else {
Expand Down
75 changes: 52 additions & 23 deletions lib/lzo/lzo1x_decompress_safe.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -46,11 +46,23 @@ int lzo1x_decompress_safe(const unsigned char *in, size_t in_len,
const unsigned char * const ip_end = in + in_len;
unsigned char * const op_end = out + *out_len;

unsigned char bitstream_version;

op = out;
ip = in;

if (unlikely(in_len < 3))
goto input_overrun;

if (likely(*ip == 17)) {
bitstream_version = ip[1];
ip += 2;
if (unlikely(in_len < 5))
goto input_overrun;
} else {
bitstream_version = 0;
}

if (*ip > 17) {
t = *ip++ - 17;
if (t < 4) {
Expand Down Expand Up @@ -154,32 +166,49 @@ int lzo1x_decompress_safe(const unsigned char *in, size_t in_len,
m_pos -= next >> 2;
next &= 3;
} else {
m_pos = op;
m_pos -= (t & 8) << 11;
t = (t & 7) + (3 - 1);
if (unlikely(t == 2)) {
size_t offset;
const unsigned char *ip_last = ip;
NEED_IP(2);
next = get_unaligned_le16(ip);
if (((next & 0xfffc) == 0xfffc) &&
((t & 0xf8) == 0x18) &&
likely(bitstream_version)) {
NEED_IP(3);
t &= 7;
t |= ip[2] << 3;
t += MIN_ZERO_RUN_LENGTH;
NEED_OP(t);
memset(op, 0, t);
op += t;
next &= 3;
ip += 3;
goto match_next;
} else {
m_pos = op;
m_pos -= (t & 8) << 11;
t = (t & 7) + (3 - 1);
if (unlikely(t == 2)) {
size_t offset;
const unsigned char *ip_last = ip;

while (unlikely(*ip == 0)) {
ip++;
NEED_IP(1);
}
offset = ip - ip_last;
if (unlikely(offset > MAX_255_COUNT))
return LZO_E_ERROR;
while (unlikely(*ip == 0)) {
ip++;
NEED_IP(1);
}
offset = ip - ip_last;
if (unlikely(offset > MAX_255_COUNT))
return LZO_E_ERROR;

offset = (offset << 8) - offset;
t += offset + 7 + *ip++;
NEED_IP(2);
offset = (offset << 8) - offset;
t += offset + 7 + *ip++;
NEED_IP(2);
next = get_unaligned_le16(ip);
}
ip += 2;
m_pos -= next >> 2;
next &= 3;
if (m_pos == op)
goto eof_found;
m_pos -= 0x4000;
}
next = get_unaligned_le16(ip);
ip += 2;
m_pos -= next >> 2;
next &= 3;
if (m_pos == op)
goto eof_found;
m_pos -= 0x4000;
}
TEST_LB(m_pos);
#if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)
Expand Down
12 changes: 11 additions & 1 deletion lib/lzo/lzodefs.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,12 @@
*/


/* Version
* 0: original lzo version
* 1: lzo with support for RLE
*/
#define LZO_VERSION 1

#define COPY4(dst, src) \
put_unaligned(get_unaligned((const u32 *)(src)), (u32 *)(dst))
#if defined(CONFIG_X86_64) || defined(CONFIG_ARM64)
Expand All @@ -28,6 +34,7 @@
#elif defined(CONFIG_X86_64) || defined(CONFIG_ARM64)
#define LZO_USE_CTZ64 1
#define LZO_USE_CTZ32 1
#define LZO_FAST_64BIT_MEMORY_ACCESS
#elif defined(CONFIG_X86) || defined(CONFIG_PPC)
#define LZO_USE_CTZ32 1
#elif defined(CONFIG_ARM) && (__LINUX_ARM_ARCH__ >= 5)
Expand All @@ -37,7 +44,7 @@
#define M1_MAX_OFFSET 0x0400
#define M2_MAX_OFFSET 0x0800
#define M3_MAX_OFFSET 0x4000
#define M4_MAX_OFFSET 0xbfff
#define M4_MAX_OFFSET 0xbffe

#define M1_MIN_LEN 2
#define M1_MAX_LEN 2
Expand All @@ -53,6 +60,9 @@
#define M3_MARKER 32
#define M4_MARKER 16

#define MIN_ZERO_RUN_LENGTH 4
#define MAX_ZERO_RUN_LENGTH (2047 + MIN_ZERO_RUN_LENGTH)

#define lzo_dict_t unsigned short
#define D_BITS 13
#define D_SIZE (1u << D_BITS)
Expand Down

0 comments on commit 5ee4014

Please sign in to comment.