diff --git a/compiler/rustc_codegen_llvm/src/back/command_line_args.rs b/compiler/rustc_codegen_llvm/src/back/command_line_args.rs
index b14713969b34c..1d914c7bcda64 100644
--- a/compiler/rustc_codegen_llvm/src/back/command_line_args.rs
+++ b/compiler/rustc_codegen_llvm/src/back/command_line_args.rs
@@ -7,31 +7,41 @@ mod tests;
 /// The result is intended to be informational, for embedding in debug metadata,
 /// and might not be properly quoted/escaped for actual command-line use.
 pub(crate) fn quote_command_line_args(args: &[String]) -> String {
-    // Start with a decent-sized buffer, since rustc invocations tend to be long.
-    let mut buf = String::with_capacity(128);
+    // The characters we care about quoting are all ASCII, so we can get some free
+    // performance by performing the quoting step on bytes instead of characters.
+    //
+    // Non-ASCII bytes will be copied as-is, so the result is still UTF-8.
+
+    // Calculate an adequate buffer size, assuming no escaping will be needed.
+    // The `+ 3` represents an extra space and pair of quotes per arg.
+    let capacity_estimate = args.iter().map(|arg| arg.len() + 3).sum::<usize>();
+    let mut buf = Vec::with_capacity(capacity_estimate);
 
     for arg in args {
         if !buf.is_empty() {
-            buf.push(' ');
+            buf.push(b' ');
         }
 
         print_arg_quoted(&mut buf, arg);
     }
 
-    buf
+    // Converting back to String isn't strictly necessary, since the bytes are
+    // only passed to LLVM which doesn't care, but validating should be cheap
+    // and it's nice to have some assurance that we didn't mess up.
+    String::from_utf8(buf).expect("quoted args should still be UTF-8")
 }
 
 /// Equivalent to LLVM's `sys::printArg` with quoting always enabled
 /// (see llvm/lib/Support/Program.cpp).
-fn print_arg_quoted(buf: &mut String, arg: &str) {
+fn print_arg_quoted(buf: &mut Vec<u8>, arg: &str) {
     buf.reserve(arg.len() + 2);
 
-    buf.push('"');
-    for ch in arg.chars() {
-        if matches!(ch, '"' | '\\' | '$') {
-            buf.push('\\');
+    buf.push(b'"');
+    for &byte in arg.as_bytes() {
+        if matches!(byte, b'"' | b'\\' | b'$') {
+            buf.push(b'\\');
         }
-        buf.push(ch);
+        buf.push(byte);
     }
-    buf.push('"');
+    buf.push(b'"');
 }