blob: 3d92689541db892289a6042ec196e63ada6b4a2f (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
|
From: Göktürk Yüksek <gokturk@binghamton.edu>
Subject: [PATCH] Fix array indexing in Zcav::Read when max_loop > 1 and start_offset > 0
In method Zcav::Read, the variable 'i' holds the block index. It is
also used as an index to arrays of measurement values (read times and
block counts) when (max_loops > 0). However, the blocks array and
measurements arrays will be out of sync if some initial blocks are to
be skipped (by having start_offset > 0). Using the same index value
for arrays of different sizes causes segfaults. Fix it by substracting
the start_offset properly when accessing the measurements arrays.
X-Gentoo-Bug-URL: https://bugs.gentoo.org/show_bug.cgi?id=309319
--- a/zcav_io.cpp
+++ b/zcav_io.cpp
@@ -108,7 +108,7 @@
double total_read_time = 0.0;
bool nextLoop = false;
for( ; !nextLoop && (!max_size || i < max_size)
- && (loops == 0 || (m_times[i] && m_times[i][0] != -1.0))
+ && (loops == 0 || (m_times[i - start_offset] && m_times[i - start_offset][0] != -1.0))
&& (!max_size || i < max_size); i++)
{
double read_time = access_data(i ? skip_rate - 1 : 0);
@@ -135,8 +135,8 @@
m_times.push_back(new double[max_loops]);
m_count.push_back(0);
}
- m_times[i][loops] = read_time;
- m_count[i]++;
+ m_times[i - start_offset][loops] = read_time;
+ m_count[i - start_offset]++;
}
} // end loop for reading blocks
|
