diff --git a/mfbt/double-conversion/bignum.cc b/mfbt/double-conversion/bignum.cc
index 747491a0897..dc8a2a63e77 100644
--- a/mfbt/double-conversion/bignum.cc
+++ b/mfbt/double-conversion/bignum.cc
@@ -501,8 +501,8 @@ uint16_t Bignum::DivideModuloIntBignum(const Bignum& other) {
   // Start by removing multiples of 'other' until both numbers have the same
   // number of digits.
   while (BigitLength() > other.BigitLength()) {
-    // This naive approach is extremely inefficient if the this divided other
-    // might be big. This function is implemented for doubleToString where
+    // This naive approach is extremely inefficient if `this` divided by other
+    // is big. This function is implemented for doubleToString where
     // the result should be small (less than 10).
     ASSERT(other.bigits_[other.used_digits_ - 1] >= ((1 << kBigitSize) / 16));
     // Remove the multiples of the first digit.
@@ -755,7 +755,6 @@ void Bignum::SubtractTimes(const Bignum& other, int factor) {
     Chunk difference = bigits_[i] - borrow;
     bigits_[i] = difference & kBigitMask;
     borrow = difference >> (kChunkSize - 1);
-    ++i;
   }
   Clamp();
 }
diff --git a/mfbt/double-conversion/double-conversion.cc b/mfbt/double-conversion/double-conversion.cc
index a79fe92d225..febba6cd782 100644
--- a/mfbt/double-conversion/double-conversion.cc
+++ b/mfbt/double-conversion/double-conversion.cc
@@ -577,7 +577,7 @@ double StringToDoubleConverter::StringToIeee(
     const char* input,
     int length,
     int* processed_characters_count,
-    bool read_as_double) {
+    bool read_as_double) const {
   const char* current = input;
   const char* end = input + length;
 
diff --git a/mfbt/double-conversion/double-conversion.h b/mfbt/double-conversion/double-conversion.h
index c62b16b6758..0900ba0060b 100644
--- a/mfbt/double-conversion/double-conversion.h
+++ b/mfbt/double-conversion/double-conversion.h
@@ -503,7 +503,7 @@ class StringToDoubleConverter {
   // in the 'processed_characters_count'. Trailing junk is never included.
   double StringToDouble(const char* buffer,
                         int length,
-                        int* processed_characters_count) {
+                        int* processed_characters_count) const {
     return StringToIeee(buffer, length, processed_characters_count, true);
   }
 
@@ -512,7 +512,7 @@ class StringToDoubleConverter {
   // due to potential double-rounding.
   float StringToFloat(const char* buffer,
                       int length,
-                      int* processed_characters_count) {
+                      int* processed_characters_count) const {
     return static_cast<float>(StringToIeee(buffer, length,
                                            processed_characters_count, false));
   }
@@ -527,7 +527,7 @@ class StringToDoubleConverter {
   double StringToIeee(const char* buffer,
                       int length,
                       int* processed_characters_count,
-                      bool read_as_double);
+                      bool read_as_double) const;
 
   DISALLOW_IMPLICIT_CONSTRUCTORS(StringToDoubleConverter);
 };
diff --git a/mfbt/double-conversion/update.sh b/mfbt/double-conversion/update.sh
index 43091e31253..7bb417c3364 100755
--- a/mfbt/double-conversion/update.sh
+++ b/mfbt/double-conversion/update.sh
@@ -3,7 +3,7 @@
 # Copies the needed files from a directory containing the original
 # double-conversion source that we need.
 
-# This was last updated with git rev e5b34421b763f7bf7e4f9081403db417d5a55a36.
+# This was last updated with git rev 04cae7a8d5ef3d62ceffb03cdc3d38f258457a52.
 
 set -e
 
diff --git a/mfbt/double-conversion/use-StandardInteger.patch b/mfbt/double-conversion/use-StandardInteger.patch
index 6cd49b3bcec..755dd1b9e3f 100644
--- a/mfbt/double-conversion/use-StandardInteger.patch
+++ b/mfbt/double-conversion/use-StandardInteger.patch
@@ -2,7 +2,7 @@ diff --git a/mfbt/double-conversion/utils.h b/mfbt/double-conversion/utils.h
 index cd3e330..bdc7d4b 100644
 --- a/mfbt/double-conversion/utils.h
 +++ b/mfbt/double-conversion/utils.h
-@@ -68,23 +68,7 @@
+@@ -74,23 +74,7 @@
  #endif
  
  
diff --git a/mfbt/double-conversion/utils.h b/mfbt/double-conversion/utils.h
index 0eec2d91775..6d67729b23f 100644
--- a/mfbt/double-conversion/utils.h
+++ b/mfbt/double-conversion/utils.h
@@ -55,10 +55,12 @@
 #if defined(_M_X64) || defined(__x86_64__) || \
     defined(__ARMEL__) || defined(__avr32__) || \
     defined(__hppa__) || defined(__ia64__) || \
-    defined(__mips__) || defined(__powerpc__) || \
+    defined(__mips__) || \
+    defined(__powerpc__) || defined(__ppc__) || defined(__ppc64__) || \
     defined(__sparc__) || defined(__sparc) || defined(__s390__) || \
     defined(__SH4__) || defined(__alpha__) || \
-    defined(_MIPS_ARCH_MIPS32R2)
+    defined(_MIPS_ARCH_MIPS32R2) || \
+    defined(_AARCH64EL_)
 #define DOUBLE_CONVERSION_CORRECT_DOUBLE_OPERATIONS 1
 #elif defined(_M_IX86) || defined(__i386__) || defined(__i386)
 #if defined(_WIN32)