У нас вы можете посмотреть бесплатно 主題分享: How to use KASAN to debug memory corruption in OpenStack environment - (Gavin Guo) или скачать в максимальном доступном качестве, видео которое было загружено на ютуб. Для загрузки выберите вариант из формы ниже:
Если кнопки скачивания не
загрузились
НАЖМИТЕ ЗДЕСЬ или обновите страницу
Если возникают проблемы со скачиванием видео, пожалуйста напишите в поддержку по адресу внизу
страницы.
Спасибо за использование сервиса ClipSaver.ru
Gavin Guo 為我們介紹 Linux Kernel Address Sanitizer 以及其在 OpenStack 上的除錯經驗。 Slides: https://www.slideshare.net/GavinGuo3/... Notes: https://docs.google.com/document/d/1L... Gavin: 這份簡報完整的介紹了KASan的實做機制。是當下介紹KASan最詳細的文件之一。建議聽到第二個小時的內容,有非常完善和詳盡的 Slub & KASan 細節講解。 之後有可能會再更完善,上傳 slideshare 後再更新link。 **KASAN在OpenStack雲計算平台上的調試經驗** 講者:Gavin Guo/陳亮 Ubuntu 社區在OpenStack雲計算平台上的出現,起因於當今NOVA計算節點產生大量KVM(Kernel Virtual Machine) 虛擬機造成的負載,導致系統出現難以複製與重現的use-after-free/out-of-bound內存崩潰。在KASAN被backport LTS kernel v3.13 之前,調試諸如此類臭蟲,是場夢靨。KASAN(Kernel Address Sanitizer)在v4.0-rc1被導入後,利用目前調試內存崩潰最先進的技術之一---GCC ASAN(Address space sanitizer),讓噩夢終結。在這個簡報裡面,講者會詳盡的介紹如何backport KASAN到Ubuntu LTS Trusty v3.13 內核並且在OpenStack雲計算平台上面進行調試,同時分享如何解決KASAN捕獲的NUMA process migration隱藏之use-after-free臭蟲。 How to use KASAN to debug the memory corruption on the OpenStack platform. Speakers: Gavin Guo/Liang Chen Since the emerging of the OpenStack cloud computing platform in the Ubuntu community, increasing number of public/private cloud service providers choose to deploy it all over the world. The tremendous loading generated by the VMs running on OpenStack triggers much hard-to-debug memory corruptions on the system. Not until KASAN was backported to the Ubuntu Trusty v3.13 kernel, it's a nightmare to debug this sort of bugs. After KASAN, using the state-of-the-art technology--ASAN(Address Space Sanitizer), was introduced in the v4.0-rc1, the nightmare was terminated. In this presentation, the speaker will introduce how to backport the KASAN to the Ubuntu Trusty v3.13 kernel and share the experience how KASAN helps debugging the memory corruption on OpenStack environment. Finally, we will also give a glance how we fixed the NUMA bug found through KASAN. **講者介紹** Canonical Gavin Guo Canonical Linux 社區內核開發者,專長解決Ubuntu OpenStack雲計算平台內核相關疑難雜症,特別針對內存架構NUMA(non-uniform memory access)、SLUB分配器,與KASAN(Kernel Address Sanitizer)都有著通盤透徹的理解與無數的實作經驗 。在此之前任職於台灣自有處理器架構Andes公司內核開發團隊,負責內核處理器架構層的維護與實作。更早期曾任職BIOS與Linux驅動開發。在豐富的系統開發經驗累積之下,在內核、編譯器工具鏈,與處理器架構系統級別的協同開發與調試中扮演不可或缺的重要角色。 Gavin Guo is a Linux kernel developer in the Ubuntu community. He is now working for Canonical in the Customer Success division. He is responsible for the kernel stability of the OpenStack platform especially in the NUMA(Non-uniform Memory Access) and SLUB allocator. He is also the one who introduces KASAN into the team to investigate kernel issues on OpenStack platform and that ended a lot of nightmares. Before joining Canonical, he worked for Andes technology and was responsible for maintaining the Andes-created NDS32 architecture kernel implementation. Canonical 陳亮/Liang Chen Canonical OpenStack兼內核開發者,早期OpenStack Grizzly即加入的活躍社區開發者,曾在IBM OpenStack upstream團隊。在過去幾年內在OpenStack貢獻了可觀的代碼量,特別是在Heat, Cinder, Keystone元件。近來集中在Nova以及libvirt/qemu/kvm virtualization技術。過去曾任職於TripAdvisor,是Java開發者,熟捻前端/後端技術。早年在校主修OS方向,而後保持著對OS的關注研究。由於全方面的知識面工作經驗,鑄造了陳亮這位少見的generalist,目前在Canonical負責OpenStack跨多組件全棧內核系統級別的問題分析。 Liang Chen works as a full stack engineer at Canonical. His primary responsibility is to maintain OpenStack and related packages. As a secondary focus, he also spends substantial time on Linux kernel. At the time he was a designated OpenStack developer at IBM, he contributed quite a lot of work upstream across various components. Recent years he leans more towards virtualization technologies, such as KVM, VirtIO, etc. He was also an experienced Java engineer at his early career path, and it was the university years that let him extended his reach into OS kernel. All his professional experiences together make him a generalist and capable of working on crossing userspace and kernel issues.