[orc] branch master updated: ORC-496: [C++] Fix integer overflow in skip of column reader (#389)

[mdeepak]

This is an automated email from the ASF dual-hosted git repository.

mdeepak pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/orc.git


The following commit(s) were added to refs/heads/master by this push:
     new 0d13ddb  ORC-496: [C++] Fix integer overflow in skip of column reader 
(#389)
0d13ddb is described below

commit 0d13ddb28894ce861350c6239606ce6b0a34bb91
Author: Gang Wu <gan...@alibaba-inc.com>
AuthorDate: Tue Apr 30 11:08:41 2019 -0700

    ORC-496: [C++] Fix integer overflow in skip of column reader (#389)
    
    skip() function is the StringDirectColumnReader and DoubleColumnReader
    use size_t to compute total size to skip; however, the value is casted
    to int when calling underlying input stream.
---
 c++/src/ColumnReader.cc | 17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

diff --git a/c++/src/ColumnReader.cc b/c++/src/ColumnReader.cc
index 01b9925..4e36f4d 100644
--- a/c++/src/ColumnReader.cc
+++ b/c++/src/ColumnReader.cc
@@ -417,8 +417,14 @@ namespace orc {
         bytesPerValue * numValues) {
       bufferPointer += bytesPerValue * numValues;
     } else {
-      inputStream->Skip(static_cast<int>(bytesPerValue * numValues -
-          static_cast<size_t>(bufferEnd - bufferPointer)));
+      size_t sizeToSkip = bytesPerValue * numValues -
+                          static_cast<size_t>(bufferEnd - bufferPointer);
+      const size_t cap = static_cast<size_t>(std::numeric_limits<int>::max());
+      while (sizeToSkip != 0) {
+        size_t step = sizeToSkip > cap ? cap : sizeToSkip;
+        inputStream->Skip(static_cast<int>(step));
+        sizeToSkip -= step;
+      }
       bufferEnd = nullptr;
       bufferPointer = nullptr;
     }
@@ -667,7 +673,12 @@ namespace orc {
     } else {
       // move the stream forward after accounting for the buffered bytes
       totalBytes -= lastBufferLength;
-      blobStream->Skip(static_cast<int>(totalBytes));
+      const size_t cap = static_cast<size_t>(std::numeric_limits<int>::max());
+      while (totalBytes != 0) {
+        size_t step = totalBytes > cap ? cap : totalBytes;
+        blobStream->Skip(static_cast<int>(step));
+        totalBytes -= step;
+      }
       lastBufferLength = 0;
       lastBuffer = nullptr;
     }