LibCoreDump+CrashDaemon: Compress coredumps

Most coredumps contain large amounts of consecutive null bytes and as
such are a prime candidate for compression.

This commit makes CrashDaemon compress files once the kernel finishes
emitting them, as well as adds the functionality needed in LibCoreDump
to then parse them.

3 files changed, 20 insertions, 7 deletions

diff --git a/Userland/Libraries/LibCoreDump/CMakeLists.txt b/Userland/Libraries/LibCoreDump/CMakeLists.txt
index f6a9457d30..d1b05fed07 100644
--- a/Userland/Libraries/LibCoreDump/CMakeLists.txt
+++ b/Userland/Libraries/LibCoreDump/CMakeLists.txt
@@ -4,4 +4,4 @@ set(SOURCES
 )
 
 serenity_lib(LibCoreDump coredump)
-target_link_libraries(LibCoreDump LibC LibCore LibDebug)
+target_link_libraries(LibCoreDump LibC LibCompress LibCore LibDebug)
diff --git a/Userland/Libraries/LibCoreDump/Reader.cpp b/Userland/Libraries/LibCoreDump/Reader.cpp
index 24a97293fc..fee4fd2f6a 100644
--- a/Userland/Libraries/LibCoreDump/Reader.cpp
+++ b/Userland/Libraries/LibCoreDump/Reader.cpp
@@ -26,6 +26,7 @@
 
 #include <AK/JsonObject.h>
 #include <AK/JsonValue.h>
+#include <LibCompress/Gzip.h>
 #include <LibCoreDump/Reader.h>
 #include <signal_numbers.h>
 #include <string.h>
@@ -37,12 +38,12 @@ OwnPtr<Reader> Reader::create(const String& path)
     auto file_or_error = MappedFile::map(path);
     if (file_or_error.is_error())
         return {};
-    return adopt_own(*new Reader(file_or_error.release_value()));
+    return adopt_own(*new Reader(file_or_error.value()->bytes()));
 }
 
-Reader::Reader(NonnullRefPtr<MappedFile> coredump_file)
-    : m_coredump_file(move(coredump_file))
-    , m_coredump_image(m_coredump_file->bytes())
+Reader::Reader(ReadonlyBytes coredump_bytes)
+    : m_coredump_buffer(decompress_coredump(coredump_bytes))
+    , m_coredump_image(m_coredump_buffer.bytes())
 {
     size_t index = 0;
     m_coredump_image.for_each_program_header([this, &index](auto pheader) {
@@ -56,6 +57,16 @@ Reader::Reader(NonnullRefPtr<MappedFile> coredump_file)
     VERIFY(m_notes_segment_index != -1);
 }
 
+ByteBuffer Reader::decompress_coredump(const ReadonlyBytes& raw_coredump)
+{
+    if (!Compress::GzipDecompressor::is_likely_compressed(raw_coredump))
+        return ByteBuffer::copy(raw_coredump); // handle old format core dumps (uncompressed)
+    auto decompressed_coredump = Compress::GzipDecompressor::decompress_all(raw_coredump);
+    if (!decompressed_coredump.has_value())
+        return ByteBuffer::copy(raw_coredump); // if we didnt manage to decompress it, try and parse it as decompressed core dump
+    return decompressed_coredump.value();
+}
+
 Reader::~Reader()
 {
 }
diff --git a/Userland/Libraries/LibCoreDump/Reader.h b/Userland/Libraries/LibCoreDump/Reader.h
index 7fd3aaf715..8cf111d759 100644
--- a/Userland/Libraries/LibCoreDump/Reader.h
+++ b/Userland/Libraries/LibCoreDump/Reader.h
@@ -70,7 +70,9 @@ public:
     HashMap<String, String> metadata() const;
 
 private:
-    Reader(NonnullRefPtr<MappedFile>);
+    Reader(ReadonlyBytes);
+
+    static ByteBuffer decompress_coredump(const ReadonlyBytes&);
 
     class NotesEntryIterator {
     public:
@@ -92,7 +94,7 @@ private:
     // as getters with the appropriate (non-JsonValue) types.
     const JsonObject process_info() const;
 
-    NonnullRefPtr<MappedFile> m_coredump_file;
+    ByteBuffer m_coredump_buffer;
     ELF::Image m_coredump_image;
     ssize_t m_notes_segment_index { -1 };
 };