From 3d9db48dbe79715fe8ad9db8696e4d8a7c53c18a Mon Sep 17 00:00:00 2001
From: Aleksa Sarai <asarai@suse.de>
Date: Wed, 8 Nov 2017 00:05:46 +1100
Subject: [PATCH] tar: asm: store padding in chunks to avoid memory exhaustion

Previously, we would read the entire padding in a given archive into
memory in order to store it in the packer. This would cause memory
exhaustion if a malicious archive was crafted with very large amounts of
padding. Since a given SegmentType is reconstructed losslessly, we can
simply chunk up any padding into large segments to avoid this problem.
Use a reasonable default of 1MiB to avoid changing the tar-split.json of
existing archives that are not malformed.

Fixes: CVE-2017-14992
Signed-off-by: Aleksa Sarai <asarai@suse.de>
---
 tar/asm/disassemble.go | 43 +++++++++++++++++++++++++++---------------
 1 file changed, 28 insertions(+), 15 deletions(-)

diff --git a/tar/asm/disassemble.go b/tar/asm/disassemble.go
index 54ef23a..009b3f5 100644
--- a/tar/asm/disassemble.go
+++ b/tar/asm/disassemble.go
@@ -2,7 +2,6 @@ package asm
 
 import (
 	"io"
-	"io/ioutil"
 
 	"github.com/vbatts/tar-split/archive/tar"
 	"github.com/vbatts/tar-split/tar/storage"
@@ -119,20 +118,34 @@ func NewInputTarStream(r io.Reader, p storage.Packer, fp storage.FilePutter) (io
 			}
 		}
 
-		// it is allowable, and not uncommon that there is further padding on the
-		// end of an archive, apart from the expected 1024 null bytes.
-		remainder, err := ioutil.ReadAll(outputRdr)
-		if err != nil && err != io.EOF {
-			pW.CloseWithError(err)
-			return
-		}
-		_, err = p.AddEntry(storage.Entry{
-			Type:    storage.SegmentType,
-			Payload: remainder,
-		})
-		if err != nil {
-			pW.CloseWithError(err)
-			return
+		// It is allowable, and not uncommon that there is further padding on
+		// the end of an archive, apart from the expected 1024 null bytes. We
+		// do this in chunks rather than in one go to avoid cases where a
+		// maliciously crafted tar file tries to trick us into reading many GBs
+		// into memory.
+		const paddingChunkSize = 1024 * 1024
+		var paddingChunk [paddingChunkSize]byte
+		for {
+			var isEOF bool
+			n, err := outputRdr.Read(paddingChunk[:])
+			if err != nil {
+				if err != io.EOF {
+					pW.CloseWithError(err)
+					return
+				}
+				isEOF = true
+			}
+			_, err = p.AddEntry(storage.Entry{
+				Type:    storage.SegmentType,
+				Payload: paddingChunk[:n],
+			})
+			if err != nil {
+				pW.CloseWithError(err)
+				return
+			}
+			if isEOF {
+				break
+			}
 		}
 		pW.Close()
 	}()