bpf: use bitCast instead of intCast in ld_imm_impl

Any 32 bit immediate is allowed in a BPF instruction, including those
greater than the largest positive i32 value.

Author: Feyorsh

lib/std/os/linux/bpf.zig

@@ -642,7 +642,7 @@ pub const Insn = packed struct {
             .dst = @intFromEnum(dst),
             .src = @intFromEnum(src),
             .off = 0,
-            .imm = @as(i32, @intCast(@as(u32, @truncate(imm)))),
+            .imm = @as(i32, @bitCast(@as(u32, @truncate(imm)))),
         };
     }
 
@@ -652,7 +652,7 @@ pub const Insn = packed struct {
             .dst = 0,
             .src = 0,
             .off = 0,
-            .imm = @as(i32, @intCast(@as(u32, @truncate(imm >> 32)))),
+            .imm = @as(i32, @bitCast(@as(u32, @truncate(imm >> 32)))),
         };
     }