From d706f90a75d472a1dc2fe53083525d937ece472f Mon Sep 17 00:00:00 2001
From: Jens Steube <jens.steube@gmail.com>
Date: Tue, 10 Mar 2020 15:42:55 +0100
Subject: [PATCH] Dictstat: On Windows, the st_ino attribute in the stat struct
 is not set which can lead to invalid cache hits. Added the filename to the
 database entry.

---
 docs/changes.txt   |  1 +
 include/dictstat.h |  2 +-
 include/types.h    |  2 ++
 src/dictstat.c     |  4 ++++
 src/wordlist.c     | 16 +++++++++++++++-
 5 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/docs/changes.txt b/docs/changes.txt
index 34f830347..160004738 100644
--- a/docs/changes.txt
+++ b/docs/changes.txt
@@ -108,6 +108,7 @@
 - Building: Fix for library compilation failure due to multiple defenition of sbob_xx64()
 - Building: Updated BUILD.md
 - Cracking bcrypt and Password Safe v2: Use a feedback from the compute API backend to dynamically find out optimal thread count
+- Dictstat: On Windows, the st_ino attribute in the stat struct is not set which can lead to invalid cache hits. Added the filename to the database entry.
 - Documents: Added README on how to build hashcat on MSYS2
 - File handling: Print a truncation warning when an oversized line is detected
 - My Wallet: Added additional plaintext pattern used in newer versions
diff --git a/include/dictstat.h b/include/dictstat.h
index 4e79169b1..bfe8fe7b0 100644
--- a/include/dictstat.h
+++ b/include/dictstat.h
@@ -18,7 +18,7 @@
 #define MAX_DICTSTAT 100000
 
 #define DICTSTAT_FILENAME "hashcat.dictstat2"
-#define DICTSTAT_VERSION  (0x6863646963743200 | 0x01)
+#define DICTSTAT_VERSION  (0x6863646963743200 | 0x02)
 
 int sort_by_dictstat (const void *s1, const void *s2);
 
diff --git a/include/types.h b/include/types.h
index a17f9d10f..5ae4f2b0e 100644
--- a/include/types.h
+++ b/include/types.h
@@ -1656,6 +1656,8 @@ typedef struct dictstat
   char encoding_from[64];
   char encoding_to[64];
 
+  u8 hash_filename[16];
+
 } dictstat_t;
 
 typedef struct hashdump
diff --git a/src/dictstat.c b/src/dictstat.c
index 15830c68b..03736e74b 100644
--- a/src/dictstat.c
+++ b/src/dictstat.c
@@ -17,6 +17,10 @@ int sort_by_dictstat (const void *s1, const void *s2)
   const dictstat_t *d1 = (const dictstat_t *) s1;
   const dictstat_t *d2 = (const dictstat_t *) s2;
 
+  const int rc_hash = memcmp (d1->hash_filename, d2->hash_filename, 16);
+
+  if (rc_hash != 0) return rc_hash;
+
   const int rc_from = strcmp (d1->encoding_from, d2->encoding_from);
 
   if (rc_from != 0) return rc_from;
diff --git a/src/wordlist.c b/src/wordlist.c
index 380dfcf4d..651ff4d80 100644
--- a/src/wordlist.c
+++ b/src/wordlist.c
@@ -13,6 +13,7 @@
 #include "rp_cpu.h"
 #include "shared.h"
 #include "wordlist.h"
+#include "emu_inc_hash_sha1.h"
 
 size_t convert_from_hex (hashcat_ctx_t *hashcat_ctx, char *line_buf, const size_t line_len)
 {
@@ -340,7 +341,7 @@ int count_words (hashcat_ctx_t *hashcat_ctx, HCFILE *fp, const char *dictfile, u
 
   dictstat_t d;
 
-  d.cnt = 0;
+  memset (&d, 0, sizeof (d));
 
   if (fstat (hc_fileno (fp), &d.stat))
   {
@@ -378,6 +379,19 @@ int count_words (hashcat_ctx_t *hashcat_ctx, HCFILE *fp, const char *dictfile, u
     return 0;
   }
 
+  const size_t dictfile_len = strlen (dictfile);
+
+  u32 *dictfile_padded = (u32 *) hcmalloc (dictfile_len + 64); // padding required for sha1_update()
+
+  sha1_ctx_t sha1_ctx;
+  sha1_init   (&sha1_ctx);
+  sha1_update (&sha1_ctx, dictfile_padded, dictfile_len);
+  sha1_final  (&sha1_ctx);
+
+  hcfree (dictfile_padded);
+
+  memcpy (d.hash_filename, sha1_ctx.h, 16);
+
   const u64 cached_cnt = dictstat_find (hashcat_ctx, &d);
 
   if (run_rule_engine (user_options_extra->rule_len_l, user_options->rule_buf_l) == 0)