5月8号，今年度的Linux Storage, Filesystem, Memory Management & BPF Summit已经拉开帷幕，Linux存储、文件系统、内存管理以及BPF领域的年度峰会又一次到来。此次峰会聚集了Linux最重要的开发专家以及内核子系统维护者，以规划和探索改进相关模块的发展方向，这些技术点将在未来24-48个月内进入主流内核和Linux发行版。

本次会议持续三天，分为IO、FS、MM、BPF 四个分会场，相关topic演讲者通过线下和线上的方式对所在领域的前沿技术点进行分享和谈论，其中包含了性能优化、稳定性、系统安全等核心技术，值得关注和学习。

IO分会场 topic：

1. Block Stack Performance
2. bpf iterator for file-system
3. Cloud storage optimizations
4. Zoned storage support in filesystems
5. kdevops: Advances with automation of testing with fstests and blktests
6. Sunsetting buffer heads
7. State of the Page
8. CXL Specification Working Group Update
9. Small Block Layer Segments (<4KB)
10. NVMe Boot
11. NVMe Multipath vs. NUMA
12. Network Namespaces and iSCSI Initiator
13. Non-blocking I/O (io_uring)
14. Host FTL
15. Cryptographic Material Handling
16. SSDFS + ZNS SSD

FS分会场topic：

1. NFSD BoF
2. Improving i_version support in filesystems
3. IOMAP conversion status
4. Fanotify HSM API
5. Mount info/mount notification
6. Fuse passthrough solutions and status
7. Fuse uring request allocation and communication
8. Image-based read-only filesystems (EROFS, maybe userns discussion)
9. Composefs vs erofs+overlay
10. Online repair of filesystems: what next?
11. Filesystem backporting to stable
12. Better data structures for mount propagation Atomically replacing a mount
13. Deletegated mounts
14. Final hurdles with procfs magic link hardening
15. Removal of kthread freezer next steps
16. Bcachefs status update

MM分会场topic：

1. CXL Fabric Manager
2. Physr discussion
3. State Of The Page
4. Lazy RCU memory reclaim
5. mm docs
6. Reducing direct map fragmentation
7. DAMON Updates and Future Plans
8. Using hardware counters to determine hot/cold pages
9. Scalable Pagefaults
10. Virtual Machine Memory Passthrough
11. Single Owner Memory
12. SMDK inspired MM changes for CXL
13. Flexible orders for anonymous folios
14. VM Memory Overcommit
15. Eliminate vmap/vmalloc lock contention
16. Userspace control of memory management
17. HGM for hugetlbfs
18. SLOB+SLAB allocators removal and future SLUB improvements
19. VM live migration over CXL memory
20. Reducing zombie memcgs
21. Swap abstraction
22. The future of memory tiering
23. Memory profiling using code tagging

BPF分会场topic：

1. Verifier refactoring, specifically unifying kfunc + helper checking
2. Per-arg kfunc flags: Enabling kfuncs to specify modifiers per-arg, rather than per-kfunc
3. Multi-kfunc sets: Allowing different sets of kfuncs to be allowed for different bpf_struct_ops operations
4. Local-storage userspace mapping: Allowing user space to map entries in local storage (task, cgroup, etc) map types
5. Resurrection of /dev/bpf discussion for implementing trusted unpriv BPF
6. BPF signing using fsverity and LSM gatekeeper
7. BPF Turing machine
8. BPF Static Keys Support
9. Update on wildcard map including use cases and restrictions
10. Supporting sketch data structures in eBPF (e.g., for percentile computation)
11. BPF Data Structures, specifically the verifier changes necessary to implement them
12. Update on the new module_alloc
13. Make bpf memory allocator more robust
14. BPF memory accounting
15. BPF_ID namespace
16. Update on generic BPF multi-attach API & initial implementation for tc BPF (and future XDP)
17. Update on meta netdevices for tc BPF programs
18. XDP metadata for TX
19. XDP/BPF in RDMA(RXE)
20. Namespace unaware BPF (sockets) iterator
21. BPF cgroup helper enhancements for Kubernetes like environments
22. BPF and firewall: kernel support to ease more complexe packets filtering
23. Cgroup-bpf production hic-hups
24. FUSE BPF
25. Multi uprobe link
26. SFrame-based user space stack unwinder in the kernel
27. Batteries-included symbolization with blazesym
28. Monitoring SRIOV with BPF
29. Kernel lock performance analysis using BPF
30. BPF iterator for file-system
31. Sched_ext
32. Can BPF-based scheduling help ChromeOS and Android scheduling issues (sched_ext) ?
33. Extend bpf-helpers to better support Kubernetes
34. Tetragon a k8s observability and security tool
35. BPF subsystem CI system
36. Vmtest: reusable virtual machine testing infrastructure
37. BPF CI: A year later
38. BPF CI discussion
39. Applicability of BPF in a Confidential Computing context
40. BPF documentation/standardization progress
41. BPF program allow/deny policy mechanisms (Not Presenting)
42. XDP TX hook (see Stanislav’s proposal). Note that I can only attend virtually (Merged with Stanislav’s talk)

会议正在进行中，感兴趣的话可以查看详细Docs文档，后续将对一些热点topic进行跟踪和分析。
https://docs.google.com/spreadsheets/d/1tIDYHgLhhcetoXtgyvcoM6YZWWHcVLdNYipBq2dH-_k/edit#gid=0