Linux 版本的 Windows“非分页池”这样的事存在吗?
我一直在使用一个 Windows 应用程序,它从“非分页池”读取数据以提高性能。在这种情况下,非分页池是网络驱动程序在从线路上抓取数据时写入数据的内存区域。
Linux 如何处理需要高速独占访问 RAM 的网络驱动程序(或其他驱动程序)的内存,以及问题“如何直接从非分页池读取?”当应用于 Linux 时甚至有意义吗?
非常感谢
I have been working with an Windows application which reads from the 'nonpaged pool' to increase performance. In this case the nonpaged pool is the area of memory where the network drivers write data as they grab it off the wire.
How does Linux handle memory which network drivers (or other drivers) which require high speed exclusive access to RAM and does the question 'how do I read directly from nonpaged pool?' even make sense when applied to Linux?
Many thanks
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
一些网络(例如 Infiniband)支持 RDMA,这需要能够防止对进程中的某些页面进行分页。请参阅 mlock()、mlockall()、munlock()、munlockall() 函数。
除此之外,我认为本身不存在“非分页池”的概念。一般来说,内核内存是不可分页的,但所有用户内存(除了用 mlock() 等锁定的内存)都是可分页的。
Some networks like Infiniband support RDMA, which requires being able to prevent paging for some of the pages in a process. See the mlock(), mlockall(), munlock(), munlockall() functions.
Other than that, I don't think there is a concept of "nonpaged pool", per se. Generally, kernel memory is AFAIK not pageable, but all user memory except that locked with mlock() or such is.