In PL_DCT add option to limit the size of uncompressed corrupt data

Also, apply limit in dct_fuzzer

fuzz/dct_fuzzer.cc

@@ -27,7 +27,7 @@ void
 FuzzHelper::doChecks()
 {
     Pl_Discard discard;
-    Pl_DCT p("decode", &discard);
+    Pl_DCT p("decode", &discard, 20'000'000);
     p.write(const_cast<unsigned char*>(data), size);
     p.finish();
 }

include/qpdf/Pl_DCT.hh

@@ -34,6 +34,11 @@ class QPDF_DLL_CLASS Pl_DCT: public Pipeline
     QPDF_DLL
     Pl_DCT(char const* identifier, Pipeline* next);
 
+    // Constructor for decompressing image data. If corrupt_data_limit is non-zero and the data is
+    // corrupt, only attempt to uncompress if the uncompressed size is less than corrupt_data_limit.
+    QPDF_DLL
+    Pl_DCT(char const* identifier, Pipeline* next, size_t corrupt_data_limit);
+
     class QPDF_DLL_CLASS CompressConfig
     {
       public:
@@ -80,26 +85,30 @@ class QPDF_DLL_CLASS Pl_DCT: public Pipeline
         ~Members() = default;
 
       private:
+        // For compression
         Members(
-            action_e action,
-            char const* buf_description,
-            JDIMENSION image_width = 0,
-            JDIMENSION image_height = 0,
-            int components = 1,
-            J_COLOR_SPACE color_space = JCS_GRAYSCALE,
-            CompressConfig* config_callback = nullptr);
+            JDIMENSION image_width,
+            JDIMENSION image_height,
+            int components,
+            J_COLOR_SPACE color_space,
+            CompressConfig* config_callback);
+        // For decompression
+        Members(size_t corrupt_data_limit);
         Members(Members const&) = delete;
 
         action_e action;
         Pl_Buffer buf;
 
-        // Used for compression
-        JDIMENSION image_width;
-        JDIMENSION image_height;
-        int components;
-        J_COLOR_SPACE color_space;
+        // Used for decompression
+        size_t corrupt_data_limit{0};
 
-        CompressConfig* config_callback;
+        // Used for compression
+        JDIMENSION image_width{0};
+        JDIMENSION image_height{0};
+        int components{1};
+        J_COLOR_SPACE color_space{JCS_GRAYSCALE};
+
+        CompressConfig* config_callback{nullptr};
     };
 
     std::shared_ptr<Members> m;

libqpdf/Pl_DCT.cc

@@ -1,5 +1,6 @@
 #include <qpdf/Pl_DCT.hh>
 
+#include "qpdf/QPDFLogger.hh"
 #include <qpdf/QIntC.hh>
 #include <qpdf/QTC.hh>
 
@@ -31,16 +32,21 @@ error_handler(j_common_ptr cinfo)
     longjmp(jerr->jmpbuf, 1);
 }
 
+Pl_DCT::Members::Members(size_t corrupt_data_limit) :
+    action(a_decompress),
+    buf("DCT compressed image"),
+    corrupt_data_limit(corrupt_data_limit)
+{
+}
+
 Pl_DCT::Members::Members(
-    action_e action,
-    char const* buf_description,
     JDIMENSION image_width,
     JDIMENSION image_height,
     int components,
     J_COLOR_SPACE color_space,
     CompressConfig* config_callback) :
-    action(action),
-    buf(buf_description),
+    action(a_compress),
+    buf("DCT uncompressed image"),
     image_width(image_width),
     image_height(image_height),
     components(components),
@@ -50,8 +56,13 @@ Pl_DCT::Members::Members(
 }
 
 Pl_DCT::Pl_DCT(char const* identifier, Pipeline* next) :
+    Pl_DCT(identifier, next, 0)
+{
+}
+
+Pl_DCT::Pl_DCT(char const* identifier, Pipeline* next, size_t corrupt_data_limit) :
     Pipeline(identifier, next),
-    m(new Members(a_decompress, "DCT compressed image"))
+    m(new Members(corrupt_data_limit))
 {
 }
 
@@ -64,14 +75,7 @@ Pl_DCT::Pl_DCT(
     J_COLOR_SPACE color_space,
     CompressConfig* config_callback) :
     Pipeline(identifier, next),
-    m(new Members(
-        a_compress,
-        "DCT uncompressed image",
-        image_width,
-        image_height,
-        components,
-        color_space,
-        config_callback))
+    m(new Members(image_width, image_height, components, color_space, config_callback))
 {
 }
 
@@ -311,16 +315,22 @@ Pl_DCT::decompress(void* cinfo_p, Buffer* b)
 
     (void)jpeg_read_header(cinfo, TRUE);
     (void)jpeg_calc_output_dimensions(cinfo);
-
     unsigned int width = cinfo->output_width * QIntC::to_uint(cinfo->output_components);
-    JSAMPARRAY buffer =
-        (*cinfo->mem->alloc_sarray)(reinterpret_cast<j_common_ptr>(cinfo), JPOOL_IMAGE, width, 1);
+    if (cinfo->err->num_warnings == 0 || m->corrupt_data_limit == 0 ||
+        (width * QIntC::to_uint(cinfo->output_height)) < m->corrupt_data_limit) {
+        // err->num_warnings is the number of corrupt data warnings emitted.
+        // err->msg_code could also be the code of an informational message.
+        JSAMPARRAY buffer = (*cinfo->mem->alloc_sarray)(
+            reinterpret_cast<j_common_ptr>(cinfo), JPOOL_IMAGE, width, 1);
 
-    (void)jpeg_start_decompress(cinfo);
-    while (cinfo->output_scanline < cinfo->output_height) {
-        (void)jpeg_read_scanlines(cinfo, buffer, 1);
-        this->getNext()->write(buffer[0], width * sizeof(buffer[0][0]));
+        (void)jpeg_start_decompress(cinfo);
+        while (cinfo->output_scanline < cinfo->output_height) {
+            (void)jpeg_read_scanlines(cinfo, buffer, 1);
+            getNext()->write(buffer[0], width * sizeof(buffer[0][0]));
+        }
+        (void)jpeg_finish_decompress(cinfo);
+    } else {
+        *QPDFLogger::defaultLogger()->getError() << "corrupt JPEG data ignored" << "\n";
     }
-    (void)jpeg_finish_decompress(cinfo);
-    this->getNext()->finish();
+    getNext()->finish();
 }