rpms/zlib
5
0
Fork 0
Code Issues Pull Requests Releases Activity

Inflate small window optimization for IBM z15

Browse Source 
Resolves: #2154775
This commit is contained in:
Ilya Leoshkevich 2022-12-19 14:57:25 +01:00
parent 1e2f66900e
commit 9006047624
2 changed files with 306 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

299
zlib-1.2.11-IBM-Z-hw-accelrated-inflate-small-window.patch Normal file
View File

@ -0,0 +1,299 @@
--- a/contrib/s390/dfltcc.c
+++ b/contrib/s390/dfltcc.c
@@ -539,10 +539,6 @@ int ZLIB_INTERNAL dfltcc_can_inflate(strm)
struct inflate_state FAR *state = (struct inflate_state FAR *)strm->state;
struct dfltcc_state FAR *dfltcc_state = GET_DFLTCC_STATE(state);
- /* Unsupported compression settings */
- if (state->wbits != HB_BITS)
- return 0;
-
/* Unsupported hardware */
return is_bit_set(dfltcc_state->af.fns, DFLTCC_XPND) &&
is_bit_set(dfltcc_state->af.fmts, DFLTCC_FMT0);
@@ -606,8 +602,6 @@ dfltcc_inflate_action ZLIB_INTERNAL dfltcc_inflate(strm, flush, ret)
/* Translate stream to parameter block */
param->cvt = state->flags ? CVT_CRC32 : CVT_ADLER32;
param->sbb = state->bits;
- param->hl = state->whave; /* Software and hardware history formats match */
- param->ho = (state->wnext - state->whave) & ((1 << HB_BITS) - 1);
if (param->hl)
param->nt = 0; /* Honor history for the first block */
param->cv = state->flags ? ZSWAP32(state->check) : state->check;
@@ -621,8 +615,6 @@ dfltcc_inflate_action ZLIB_INTERNAL dfltcc_inflate(strm, flush, ret)
strm->msg = oesc_msg(dfltcc_state->msg, param->oesc);
state->last = cc == DFLTCC_CC_OK;
state->bits = param->sbb;
- state->whave = param->hl;
- state->wnext = (param->ho + param->hl) & ((1 << HB_BITS) - 1);
strm->adler = state->check = state->flags ? ZSWAP32(param->cv) : param->cv;
if (cc == DFLTCC_CC_OP2_CORRUPT && param->oesc != 0) {
/* Report an error if stream is corrupted */
@@ -644,11 +636,52 @@ int ZLIB_INTERNAL dfltcc_was_inflate_used(strm)
return !param->nt;
}
+/*
+ Rotates a circular buffer.
+ The implementation is based on https://cplusplus.com/reference/algorithm/rotate/
+ */
+local void rotate OF((Bytef *start, Bytef *pivot, Bytef *end));
+local void rotate(start, pivot, end)
+ Bytef *start;
+ Bytef *pivot;
+ Bytef *end;
+{
+ Bytef *p = pivot;
+ Bytef tmp;
+
+ while (p != start) {
+ tmp = *start;
+ *start = *p;
+ *p = tmp;
+
+ start++;
+ p++;
+
+ if (p == end)
+ p = pivot;
+ else if (start == pivot)
+ pivot = p;
+ }
+}
+
+#define MIN(x, y) ({ \
+ typeof(x) _x = (x); \
+ typeof(y) _y = (y); \
+ _x < _y ? _x : _y; \
+})
+
+#define MAX(x, y) ({ \
+ typeof(x) _x = (x); \
+ typeof(y) _y = (y); \
+ _x > _y ? _x : _y; \
+})
+
int ZLIB_INTERNAL dfltcc_inflate_disable(strm)
z_streamp strm;
{
struct inflate_state FAR *state = (struct inflate_state FAR *)strm->state;
struct dfltcc_state FAR *dfltcc_state = GET_DFLTCC_STATE(state);
+ struct dfltcc_param_v0 *param = &dfltcc_state->param;
if (!dfltcc_can_inflate(strm))
return 0;
@@ -660,6 +693,9 @@ int ZLIB_INTERNAL dfltcc_inflate_disable(strm)
return 1;
/* DFLTCC was not used yet - decompress in software */
memset(&dfltcc_state->af, 0, sizeof(dfltcc_state->af));
+ /* Convert the window from the hardware to the software format */
+ rotate(state->window, state->window + param->ho, state->window + HB_SIZE);
+ state->whave = state->wnext = MIN(param->hl, state->wsize);
return 0;
}
@@ -830,9 +866,9 @@ voidpf ZLIB_INTERNAL dfltcc_alloc_window(strm, items, size)
voidpf p, w;
/* To simplify freeing, we store the pointer to the allocated buffer right
- * before the window.
+ * before the window. Note that DFLTCC always uses HB_SIZE bytes.
*/
- p = ZALLOC(strm, sizeof(voidpf) + items * size + PAGE_ALIGN,
+ p = ZALLOC(strm, sizeof(voidpf) + MAX(items * size, HB_SIZE) + PAGE_ALIGN,
sizeof(unsigned char));
if (p == NULL)
return NULL;
@@ -841,6 +877,14 @@ voidpf ZLIB_INTERNAL dfltcc_alloc_window(strm, items, size)
return w;
}
+void ZLIB_INTERNAL dfltcc_copy_window(dest, src, n)
+ void *dest;
+ const void *src;
+ size_t n;
+{
+ memcpy(dest, src, MAX(n, HB_SIZE));
+}
+
void ZLIB_INTERNAL dfltcc_free_window(strm, w)
z_streamp strm;
voidpf w;
@@ -951,6 +995,24 @@ local void append_history(param, history, buf, count)
}
}
+local void get_history OF((struct dfltcc_param_v0 FAR *param,
+ const Bytef *history,
+ Bytef *buf));
+local void get_history(param, history, buf)
+ struct dfltcc_param_v0 FAR *param;
+ const Bytef *history;
+ Bytef *buf;
+{
+ if (param->ho + param->hl <= HB_SIZE)
+ /* Circular history buffer does not wrap - copy one chunk */
+ memcpy(buf, history + param->ho, param->hl);
+ else {
+ /* Circular history buffer wraps - copy two chunks */
+ memcpy(buf, history + param->ho, HB_SIZE - param->ho);
+ memcpy(buf + HB_SIZE - param->ho, history, param->ho + param->hl - HB_SIZE);
+ }
+}
+
int ZLIB_INTERNAL dfltcc_deflate_set_dictionary(strm, dictionary, dict_length)
z_streamp strm;
const Bytef *dictionary;
@@ -975,20 +1037,43 @@ int ZLIB_INTERNAL dfltcc_deflate_get_dictionary(strm, dictionary, dict_length)
struct dfltcc_state FAR *dfltcc_state = GET_DFLTCC_STATE(state);
struct dfltcc_param_v0 FAR *param = &dfltcc_state->param;
- if (dictionary) {
- if (param->ho + param->hl <= HB_SIZE)
- /* Circular history buffer does not wrap - copy one chunk */
- zmemcpy(dictionary, state->window + param->ho, param->hl);
- else {
- /* Circular history buffer wraps - copy two chunks */
- zmemcpy(dictionary,
- state->window + param->ho,
- HB_SIZE - param->ho);
- zmemcpy(dictionary + HB_SIZE - param->ho,
- state->window,
- param->ho + param->hl - HB_SIZE);
- }
+ if (dictionary)
+ get_history(param, state->window, dictionary);
+ if (dict_length)
+ *dict_length = param->hl;
+ return Z_OK;
+}
+
+int ZLIB_INTERNAL dfltcc_inflate_set_dictionary(strm, dictionary, dict_length)
+ z_streamp strm;
+ const Bytef *dictionary;
+ uInt dict_length;
+{
+ struct inflate_state *state = (struct inflate_state *)strm->state;
+ struct dfltcc_state *dfltcc_state = GET_DFLTCC_STATE(state);
+ struct dfltcc_param_v0 *param = &dfltcc_state->param;
+
+ if (inflate_ensure_window(state)) {
+ state->mode = MEM;
+ return Z_MEM_ERROR;
}
+
+ append_history(param, state->window, dictionary, dict_length);
+ state->havedict = 1;
+ return Z_OK;
+}
+
+int ZLIB_INTERNAL dfltcc_inflate_get_dictionary(strm, dictionary, dict_length)
+ z_streamp strm;
+ Bytef *dictionary;
+ uInt *dict_length;
+{
+ struct inflate_state *state = (struct inflate_state *)strm->state;
+ struct dfltcc_state *dfltcc_state = GET_DFLTCC_STATE(state);
+ struct dfltcc_param_v0 *param = &dfltcc_state->param;
+
+ if (dictionary && state->window)
+ get_history(param, state->window, dictionary);
if (dict_length)
*dict_length = param->hl;
return Z_OK;
--- a/contrib/s390/dfltcc.h
+++ b/contrib/s390/dfltcc.h
@@ -11,6 +11,8 @@ void ZLIB_INTERNAL dfltcc_copy_state OF((voidpf dst, const voidpf src,
void ZLIB_INTERNAL dfltcc_reset OF((z_streamp strm, uInt size));
voidpf ZLIB_INTERNAL dfltcc_alloc_window OF((z_streamp strm, uInt items,
uInt size));
+void ZLIB_INTERNAL dfltcc_copy_window OF((void *dest, const void *src,
+ size_t n));
void ZLIB_INTERNAL dfltcc_free_window OF((z_streamp strm, voidpf w));
#define DFLTCC_BLOCK_HEADER_BITS 3
#define DFLTCC_HLITS_COUNT_BITS 5
@@ -44,11 +46,18 @@ dfltcc_inflate_action ZLIB_INTERNAL dfltcc_inflate OF((z_streamp strm,
int flush, int *ret));
int ZLIB_INTERNAL dfltcc_was_inflate_used OF((z_streamp strm));
int ZLIB_INTERNAL dfltcc_inflate_disable OF((z_streamp strm));
+int ZLIB_INTERNAL dfltcc_inflate_set_dictionary OF((z_streamp strm,
+ const Bytef *dictionary,
+ uInt dict_length));
+int ZLIB_INTERNAL dfltcc_inflate_get_dictionary OF((z_streamp strm,
+ Bytef *dictionary,
+ uInt* dict_length));
#define ZALLOC_STATE dfltcc_alloc_state
#define ZFREE_STATE ZFREE
#define ZCOPY_STATE dfltcc_copy_state
#define ZALLOC_WINDOW dfltcc_alloc_window
+#define ZCOPY_WINDOW dfltcc_copy_window
#define ZFREE_WINDOW dfltcc_free_window
#define TRY_FREE_WINDOW dfltcc_free_window
#define INFLATE_RESET_KEEP_HOOK(strm) \
@@ -77,5 +86,15 @@ int ZLIB_INTERNAL dfltcc_inflate_disable OF((z_streamp strm));
do { \
if (dfltcc_was_inflate_used((strm))) return Z_STREAM_ERROR; \
} while (0)
+#define INFLATE_SET_DICTIONARY_HOOK(strm, dict, dict_len) \
+ do { \
+ if (dfltcc_can_inflate(strm)) \
+ return dfltcc_inflate_set_dictionary(strm, dict, dict_len); \
+ } while (0)
+#define INFLATE_GET_DICTIONARY_HOOK(strm, dict, dict_len) \
+ do { \
+ if (dfltcc_can_inflate(strm)) \
+ return dfltcc_inflate_get_dictionary(strm, dict, dict_len); \
+ } while (0)
#endif
\ No newline at end of file
diff --git a/inflate.c b/inflate.c
index 3750152..a0e2169 100644
--- a/inflate.c
+++ b/inflate.c
@@ -93,6 +93,7 @@
#define ZFREE_STATE ZFREE
#define ZCOPY_STATE zmemcpy
#define ZALLOC_WINDOW ZALLOC
+#define ZCOPY_WINDOW zmemcpy
#define ZFREE_WINDOW ZFREE
#define INFLATE_RESET_KEEP_HOOK(strm) do {} while (0)
#define INFLATE_PRIME_HOOK(strm, bits, value) do {} while (0)
@@ -101,6 +102,8 @@
#define INFLATE_NEED_UPDATEWINDOW(strm) 1
#define INFLATE_MARK_HOOK(strm) do {} while (0)
#define INFLATE_SYNC_POINT_HOOK(strm) do {} while (0)
+#define INFLATE_SET_DICTIONARY_HOOK(strm, dict, dict_len) do {} while (0)
+#define INFLATE_GET_DICTIONARY_HOOK(strm, dict, dict_len) do {} while (0)
#endif
#ifdef MAKEFIXED
@@ -1330,6 +1333,8 @@ uInt *dictLength;
if (inflateStateCheck(strm)) return Z_STREAM_ERROR;
state = (struct inflate_state FAR *)strm->state;
+ INFLATE_GET_DICTIONARY_HOOK(strm, dictionary, dictLength);
+
/* copy dictionary */
if (state->whave && dictionary != Z_NULL) {
zmemcpy(dictionary, state->window + state->wnext,
@@ -1365,6 +1370,8 @@ uInt dictLength;
return Z_DATA_ERROR;
}
+ INFLATE_SET_DICTIONARY_HOOK(strm, dictionary, dictLength);
+
/* copy dictionary to window using updatewindow(), which will amend the
existing dictionary if appropriate */
ret = updatewindow(strm, dictionary + dictLength, dictLength);
@@ -1529,8 +1536,7 @@ z_streamp source;
}
copy->next = copy->codes + (state->next - state->codes);
if (window != Z_NULL) {
- wsize = 1U << state->wbits;
- zmemcpy(window, state->window, wsize);
+ ZCOPY_WINDOW(window, state->window, 1U << state->wbits);
}
copy->window = window;
dest->state = (struct internal_state FAR *)copy;

8
zlib.spec
View File

@ -2,7 +2,7 @@
Name: zlib
Version: 1.2.11
Release: 35%{?dist}
Release: 36%{?dist}
Summary: Compression and decompression library
# /contrib/dotzlib/ have Boost license
License: zlib and Boost
@ -46,6 +46,8 @@ Patch20: zlib-1.2.11-cve-2022-37434_2.patch
# Fix setting strm.adler on z15
Patch21: zlib-1.2.11-IBM-Z-hw-accelrated-deflate-strm-adler-fix.patch
# Optimization for z15
Patch22: zlib-1.2.11-IBM-Z-hw-accelrated-inflate-small-window.patch
BuildRequires: make
BuildRequires: automake, autoconf, libtool
@ -119,6 +121,7 @@ developing applications which use minizip.
%patch19 -p1
%patch20 -p1
%patch21 -p1
%patch22 -p1
iconv -f iso-8859-2 -t utf-8 < ChangeLog > ChangeLog.tmp
@ -196,6 +199,9 @@ find $RPM_BUILD_ROOT -name '*.la' -delete
%changelog
* Mon Dec 19 2022 Ilya Leoshkevich <iii@linux.ibm.com> - 1.2.11-36
- Inflate small window optimization for IBM z15 rhbz#2154775
* Wed Oct 12 2022 Ilya Leoshkevich <iii@linux.ibm.com> - 1.2.11-35
- Fix for IBM strm.adler rhbz#2134074