fix tsw: add constraints: m < n - 3

fallback to search_small then.
Fixes one crash in GH #13

asan.lst

@@ -35,7 +35,7 @@ fjs
 hash3
 hash5
 hash8
-#tsw
+tsw
 #graspm
 #ssef
 aut

sanitizer.log

@@ -38,46 +38,7 @@ hash3
 hash5
 hash8
 tsw
-AddressSanitizer:DEADLYSIGNAL
-=================================================================
-==3586985==ERROR: AddressSanitizer: SEGV on unknown address 0x7f2c9a15efff (pc 0x000000403d73 bp 0x7ffe6e399280 sp 0x7ffe6e317dc0 T0)
-==3586985==The signal is caused by a READ memory access.
-    #0 0x403d73 in search source/algos/tsw.c:72
-    #1 0x4017ba in main source/algos/include/main.h:125
-    #2 0x7f2c9d046149 in __libc_start_call_main (/lib64/libc.so.6+0x28149) (BuildId: 7ea8d85df0e89b90c63ac7ed2b3578b2e7728756)
-    #3 0x7f2c9d04620a in __libc_start_main_impl (/lib64/libc.so.6+0x2820a) (BuildId: 7ea8d85df0e89b90c63ac7ed2b3578b2e7728756)
-    #4 0x401e84 in _start (/home/rurban/Software/smart/bin/asan/tsw+0x401e84) (BuildId: 1ae7608d7493a33edb8c4e9881f69c7ba7bee3b0)
-
-AddressSanitizer can not provide additional info.
-SUMMARY: AddressSanitizer: SEGV source/algos/tsw.c:72 in search
-==3586985==ABORTING
-AddressSanitizer:DEADLYSIGNAL
-=================================================================
-==3587157==ERROR: AddressSanitizer: SEGV on unknown address 0x7f101455efff (pc 0x000000403dbf bp 0x7ffeaf8d81b0 sp 0x7ffeaf856cf0 T0)
-==3587157==The signal is caused by a READ memory access.
-    #0 0x403dbf in search source/algos/tsw.c:72
-    #1 0x4017ba in main source/algos/include/main.h:125
-    #2 0x7f1017446149 in __libc_start_call_main (/lib64/libc.so.6+0x28149) (BuildId: 7ea8d85df0e89b90c63ac7ed2b3578b2e7728756)
-    #3 0x7f101744620a in __libc_start_main_impl (/lib64/libc.so.6+0x2820a) (BuildId: 7ea8d85df0e89b90c63ac7ed2b3578b2e7728756)
-    #4 0x401e84 in _start (/home/rurban/Software/smart/bin/asan/tsw+0x401e84) (BuildId: 1ae7608d7493a33edb8c4e9881f69c7ba7bee3b0)
-
-AddressSanitizer can not provide additional info.
-SUMMARY: AddressSanitizer: SEGV source/algos/tsw.c:72 in search
-==3587157==ABORTING
-AddressSanitizer:DEADLYSIGNAL
-=================================================================
-==3587228==ERROR: AddressSanitizer: SEGV on unknown address 0x7f0d11d5efff (pc 0x000000403dbf bp 0x7ffee9699e10 sp 0x7ffee9618950 T0)
-==3587228==The signal is caused by a READ memory access.
-    #0 0x403dbf in search source/algos/tsw.c:72
-    #1 0x4017ba in main source/algos/include/main.h:125
-    #2 0x7f0d14c46149 in __libc_start_call_main (/lib64/libc.so.6+0x28149) (BuildId: 7ea8d85df0e89b90c63ac7ed2b3578b2e7728756)
-    #3 0x7f0d14c4620a in __libc_start_main_impl (/lib64/libc.so.6+0x2820a) (BuildId: 7ea8d85df0e89b90c63ac7ed2b3578b2e7728756)
-    #4 0x401e84 in _start (/home/rurban/Software/smart/bin/asan/tsw+0x401e84) (BuildId: 1ae7608d7493a33edb8c4e9881f69c7ba7bee3b0)
-
-AddressSanitizer can not provide additional info.
-SUMMARY: AddressSanitizer: SEGV source/algos/tsw.c:72 in search
-==3587228==ABORTING
-graspm
+raspm
 AddressSanitizer:DEADLYSIGNAL
 =================================================================
 ==3587252==ERROR: AddressSanitizer: SEGV on unknown address 0x7f1bfacfefff (pc 0x000000403927 bp 0xffffffffffffffff sp 0x7ffd58dd27a0 T0)

source/algorithms.h

@@ -319,7 +319,7 @@ const struct algo ALGOS[] = {
   // L. Vespa and N. Weng, SWM: Simplified Wu-Manber for GPU-based Deep Packet
   // Inspection, in Proceedings of the 2012 International Conference on
   // Security and Management, 2012.
-  [_TSW] = {_TSW, OK, "tsw", "Two-Sliding-Window", 0, 0},
+  [_TSW] = {_TSW, OK, "tsw", "Two-Sliding-Window", 0, 0}, // m < n - 3
   [_BMH2] = {_BMH2, 0, "bmh2", "Boyer-Moore-Horspool with q-grams (BMH2)", 2, MISSING},
   [_BMH4] = {_BMH4, 0, "bmh4", "Boyer-Moore-Horspool with q-grams (BMH4)", 4, MISSING},
   [_GRASPm] = {_GRASPm, OK, "graspm", "Genomic Rapid Algorithm for String Pattern-match", 0, 0},

source/algos/tsw.c

@@ -20,10 +20,14 @@
  * in A. Hudaib and R. Al-Khalid and D. Suleiman and M. Itriq and A. Al-Anani.
  * A Fast Pattern Matching Algorithm with Two Sliding Windows (TSW). J. Comput.
  * Sci., vol.4, n.5, pp.393--401, (2008).
+ *
+ * Undocumented constraints: m < n - 3
  */
 
+#include <assert.h>
 #include "include/define.h"
 #include "include/main.h"
+#include "include/search_small.h"
 
 void preBrBc(unsigned char *x, int m, int brBc[SIGMA][SIGMA]) {
   int a, b, i;
@@ -43,11 +47,14 @@ int search(unsigned char *x, int m, unsigned char *y, int n) {
   int i, a, b;
   int count;
   unsigned char x1[XSIZE];
-  for (i = m - 1, j = 0; i >= 0; i--, j++)
-    x1[j] = x[i];
+  if (m >= n - 3)
+    return search_small(x, m, y, n);
+  assert(m < n - 2);
 
   /* Preprocessing */
   BEGIN_PREPROCESSING
+  for (i = m - 1, j = 0; i >= 0; i--, j++)
+    x1[j] = x[i];
   preBrBc(x, m, brBc_left);
   preBrBc(x1, m, brBc_right);
   count = 0;
@@ -68,6 +75,10 @@ int search(unsigned char *x, int m, unsigned char *y, int n) {
     if (b >= m && j < a)
       OUTPUT(a);
 
+    assert(j + m + 1 < n);
+    assert(a - 1 < n);
+    //if (a - 2 < 0) fprintf(stderr, "%s %d %x %d\n", x, m, y, n);
+    assert(a - 2 >= 0);
     j += brBc_left[y[j + m]][y[j + m + 1]];
     a -= brBc_right[y[a - 1]][y[a - 2]];
   }