PHP安全三板斧:转义篇Blade引擎避免XSS攻击原理
PHP 转义实现
把输出渲染成网页或API响应时,一定要转义输出,这也是一种防护措施,能避免渲染恶意代码,造成XSS攻击,还能防止应用的用户无意中执行恶意代码。
我们可以使用前面提到的htmlentities函数转移输出,该函数的第二个参数一定要使用ENT_QUOTES,让这个函数转义单引号和双引号,而且,还要在第三个参数中指定合适的字符编码(通常是UTF-8),下面的例子演示了如何在渲染前转义HTML输出:
';echo htmlentities($output, ENT_QUOTES, ‘UTF-8');
如果不转义直接输出,会弹出提示框:
转义之后输出变成:
现代PHP支持许多模板引擎,这些模板引擎在底层已经为了做好了转义处理,比如现在流行的twig/twig和smarty/smarty都会自动转义输出。这种默认处理方式很赞,为PHP Web应用提供了有力的安全保障。
Blade 模板引擎避免XSS攻击原理
Laravel使用的模板引擎是Blade,关于Blade的使用可以参考其官方文档,这里我们简单探讨下Laravel底层如何对输出进行转义处理。
一般我们在Laravel中返回视图内容会这么做:
return view(’test’, [‘data’=>$data]);
这是一个很简单的例子,意味着我们会在resources/views目录下找到test.blade.php视图文件,然后将$data变量传入其中,并将最终渲染结果作为响应的内容返回给用户。那么这一过程经历了哪些底层源码的处理,如果$data变量中包含脚本代码(如JavaScript脚本),又该怎么去处理呢?接下来我们让来一窥究竟。
首先我们从辅助函数view入手,当然这里我们也可以使用View:make,但是简单起见,我们一般用view函数,该函数定义在Illuminate\Foundation\helpers.php文件中:
function view($view = null, $data = [], $mergeData = []){ $factory = app(ViewFactory::class); if (func_num_args() === 0) { return $factory; } return $factory->make($view, $data, $mergeData);}
该函数中的逻辑是从容器中取出视图工厂接口ViewFactory对应的实例$factory(该绑定关系在Illuminate\View\ViewServiceProvider的register方法中注册,此外这里还注册了模板引擎解析器EngineResolver,包括PhpEngine和载入BladeCompiler的CompilerEngine,以及视图文件查找器FileViewFinder,一句话,这里注册了视图解析所需的所有服务),如果传入了参数,则调用$factory上的make方法:
public function make($view, $data = [], $mergeData = []){ if (isset($this->aliases[$view])) { $view = $this->aliases[$view]; } $view = $this->normalizeName($view); $path = $this->finder->find($view); $data = array_merge($mergeData, $this->parseData($data)); $this->callCreator($view = new View($this, $this->getEngineFromPath($path), $view, $path, $data)); return $view;}
这个方法位于Illuminate\View\Factory,这里所做的事情是获取视图文件的完整路径,合并传入变量,$this->getEngineFromPath会通过视图文件后缀获取相应的模板引擎,比如我们使用.blade.php结尾的视图文件则获得到的是CompilerEngine(即Blade模板引擎),否则将获取到PhpEngine,然后我们根据相应参数实例化View(Illuminate\View\View)对象并返回。需要注意的是View类中重写了__toString方法:
public function __toString(){ return $this->render();}
所以当我们打印$view实例的时候,实际上会调用View类的render方法,所以下一步我们理所应当研究render方法做了些什么:
public function render(callable $callback = null){ try { $contents = $this->renderContents(); $response = isset($callback) ? call_user_func($callback, $this, $contents) : null; // Once we have the contents of the view, we will flush the sections if we are // done rendering all views so that there is nothing left hanging over when // another view gets rendered in the future by the application developer. $this->factory->flushSectionsIfDoneRendering(); return ! is_null($response) ? $response : $contents; } catch (Exception $e) { $this->factory->flushSections(); throw $e; } catch (Throwable $e) { $this->factory->flushSections(); throw $e; }}
这里重点是$this->renderContents()方法,我们继续深入研究View类中的renderContents方法:
protected function renderContents(){ // We will keep track of the amount of views being rendered so we can flush // the section after the complete rendering operation is done. This will // clear out the sections for any separate views that may be rendered. $this->factory->incrementRender(); $this->factory->callComposer($this); $contents = $this->getContents(); // Once we've finished rendering the view, we'll decrement the render count // so that each sections get flushed out next time a view is created and // no old sections are staying around in the memory of an environment. $this->factory->decrementRender(); return $contents;}
我们重点关注$this->getContents()这里,进入getContents方法:
protected function getContents(){ return $this->engine->get($this->path, $this->gatherData());}
我们在前面已经提到,这里的$this->engine对应CompilerEngine(Illuminate\View\Engines\CompilerEngine),所以我们进入CompilerEngine的get方法:
public function get($path, array $data = []){ $this->lastCompiled[] = $path; // If this given view has expired, which means it has simply been edited since // it was last compiled, we will re-compile the views so we can evaluate a // fresh copy of the view. We'll pass the compiler the path of the view. if ($this->compiler->isExpired($path)) { $this->compiler->compile($path); } $compiled = $this->compiler->getCompiledPath($path); // Once we have the path to the compiled file, we will evaluate the paths with // typical PHP just like any other templates. We also keep a stack of views // which have been rendered for right exception messages to be generated. $results = $this->evaluatePath($compiled, $data); array_pop($this->lastCompiled); return $results;}
同样我们在之前提到,CompilerEngine使用的compiler是BladeCompiler,所以$this->compiler也就是Blade编译器,我们先看$this->compiler->compile($path);这一行(首次运行或者编译好的视图模板已过期会进这里),进入BladeCompiler的compile方法:
public function compile($path = null){ if ($path) { $this->setPath($path); } if (! is_null($this->cachePath)) { $contents = $this->compileString($this->files->get($this->getPath())); $this->files->put($this->getCompiledPath($this->getPath()), $contents); }}
这里我们做的事情是先编译视图文件内容,然后将编译好的内容存放到视图编译路径(storage\framework\views)下对应的文件(一次编译,多次运行,以提高性能),这里我们重点关注的是$this->compileString方法,该方法中使用了token_get_all函数将视图文件代码分割成多个片段,如果片段是数组的话则循环调用$this->parseToken方法:
protected function parseToken($token){ list($id, $content) = $token; if ($id == T_INLINE_HTML) { foreach ($this->compilers as $type) { $content = $this->{"compile{$type}"}($content); } } return $content;}
来到这里,我们已经很接近真相了,针对HTML代码(含Blade指令代码),循环调用compileExtensions、compileStatements、compileComments和compileEchos方法,我们重点关注输出方法compileEchos,Blade引擎默认提供了compileRawEchos、compileEscapedEchos和compileRegularEchos三种输出方法,对应的指令分别是{!! !!}、{{{ }}}和{{ }},顾名思义,compileRawEchos对应的是原生输出:
protected function compileRawEchos($value){ $pattern = sprintf('/(@)?%s\s*(.+?)\s*%s(\r?\n)?/s', $this->rawTags[0], $this->rawTags[1]); $callback = function ($matches) { $whitespace = empty($matches[3]) ? '' : $matches[3].$matches[3]; return $matches[1] ? substr($matches[0], 1) : 'compileEchoDefaults($matches[2]).'; ?>'.$whitespace; }; return preg_replace_callback($pattern, $callback, $value);}
即Blade视图中以{!! !!}包裹的变量会原生输出HTML,如果要显示图片、链接,推荐这种方式。
{{{}}}对应的CompileEscapedEchos,这个在Laravel 4.2及以前版本中用于转义,现在已经替换成了{{}},即调用compileRegularEchos方法:
protected function compileRegularEchos($value){ $pattern = sprintf('/(@)?%s\s*(.+?)\s*%s(\r?\n)?/s', $this->contentTags[0], $this->contentTags[1]); $callback = function ($matches) { $whitespace = empty($matches[3]) ? '' : $matches[3].$matches[3]; $wrapped = sprintf($this->echoFormat, $this->compileEchoDefaults($matches[2])); return $matches[1] ? substr($matches[0], 1) : ''.$whitespace; }; return preg_replace_callback($pattern, $callback, $value);}
其中$this->echoFormat对应e(%s),无独有偶,compileEscapedEchos中也用到这个方法:
protected function compileEscapedEchos($value){ $pattern = sprintf('/(@)?%s\s*(.+?)\s*%s(\r?\n)?/s', $this->escapedTags[0], $this->escapedTags[1]); $callback = function ($matches) { $whitespace = empty($matches[3]) ? '' : $matches[3].$matches[3]; return $matches[1] ? $matches[0] : 'compileEchoDefaults($matches[2]).'); ?>'.$whitespace; }; return preg_replace_callback($pattern, $callback, $value);}
辅助函数e()定义在Illuminate\Support\helpers.php中:
function e($value){ if ($value instanceof Htmlable) { return $value->toHtml(); } return htmlentities($value, ENT_QUOTES, 'UTF-8', false);}
其作用就是对输入的值进行转义。
经过这样的转义,视图中的{{ $data }}或被编译成,最终如何将$data传入视图输出,我们再回到CompilerEngine的get方法,看这一段:
$results = $this->evaluatePath($compiled, $data);
evaluatePath中传入了编译后的视图文件路径和传入的变量$data,该方法定义如下:
protected function evaluatePath($__path, $__data){ $obLevel = ob_get_level();ob_start(); extract($__data, EXTR_SKIP); // We'll evaluate the contents of the view inside a try/catch block so we can // flush out any stray output that might get out before an error occurs or // an exception is thrown. This prevents any partial views from leaking. try { include $__path; } catch (Exception $e) { $this->handleViewException($e, $obLevel); } catch (Throwable $e) { $this->handleViewException(new FatalThrowableError($e), $obLevel); } return ltrim(ob_get_clean());}
这里面调用了PHP系统函数extract将传入变量从数组中导入当前符号表(通过include $__path引入),其作用也就是将编译后视图文件中的变量悉数替换成传入的变量值(通过键名映射)。
好了,这就是Blade视图模板从渲染到输出的基本过程,可以看到我们通过{{}}来转义输出,从而达到避免XSS攻击的目的。