Kernel Space: BPF, Persistent Memory, Raptor Lake, and Vulkan/OpenGL


-
A memory allocator for BPF code [LWN.net]
Loading a BPF program into the kernel involves a lot of steps, including verification, permissions checking, linking to in-kernel helper functions, and compilation to the native instruction format. Underneath all of that, though, lies one other simple task: allocating some memory to store the compiled BPF program in the kernel's address space. It turns out that this allocation can be somewhat wasteful of memory in current kernels, especially on systems where large numbers of BPF programs are loaded. This patch set from Song Liu seeks to remedy this problem by introducing yet another specialized memory allocator into the kernel.
The kernel allows BPF programs to be reasonably large, but most of them are, in practice, quite small. BPF can do quite a bit with a few hundred bytes of code space. But the kernel allocates space for BPF programs in units of full base (usually 4KB) pages, with all of the space past the end of each program being wasted for as long as that program remains loaded. For a small program, most of the page allocated to hold the code is unused; if there are many programs loaded, the amount of wasted memory can start to add up.
Liu's patch set adds a new "bpf_prog_pack" allocator to address this problem. It would be, on the surface, one of the simplest memory allocators in the system. It maintains a list of huge pages to hold BPF programs, allocating new ones as needed. Whenever space is needed for a new BPF program, a simple, first-fit algorithm allocates the amount needed. A bitmap associated with each huge page (or "pack") tracks the free space (in 64B chunks), thus automatically coalescing chunks returned to the allocator when programs are unloaded. It is not written for speed, but it does not need to be; even on a system that makes heavy use of BPF, allocation and free operations will be relatively rare.
The advantages of this simple allocator are reasonably clear. Because multiple programs can be packed into a single page, the memory that is wasted due to internal fragmentation will be greatly reduced. Performance will be helped by putting BPF programs into huge pages, which reduces translation lookaside buffer (TLB) pressure. But it is worth asking why yet another allocator is needed when the kernel already has memory-management code that has been extensively developed and tuned over the years. The answer is that BPF brings some special needs that cannot be met with existing allocators.
-
Stray-write protection for persistent memory [LWN.net]
Persistent memory has a number of advantages; it is fast, CPU-addressable, available in large quantities and, of course, persistent. But it also, arguably, poses a higher risk of suffering corruption as a result of bugs in the kernel. Protecting against this possibility is the objective of this patch set from Ira Weiny, which makes use of Intel's "protection keys supervisor" (PKS) feature to make it harder for the kernel to inadvertently write to persistent memory.
-
Intel Core “Raptor Lake” gets advanced Linux support
The Raptor Lake processor family is the 13th generation of Intel Core, which is scheduled for launch later this year.
-
Intel confirms standalone Arc Graphics due in Q2, also 'Project Endgame' this year
Intel has confirmed some updates to their tech roadmap today and the good news is that Arc Graphics are on the way.
For notebooks, Intel Arc Alchemist will be arriving in the first quarter of this year. We already mostly knew this though, as they confirmed that before. At least now though, we're getting clearer info on standalone cards for desktops - which will be due out in Q2 (April - June). Intel think they will manage to ship "4 million" discrete GPUs in 2022 which would be across laptops, consumer desktops and workstations.
-
6 years on the Vulkan API has enabled some amazing things | GamingOnLinux
Can you actually believe that the Vulkan API has just crossed over the 6 year mark? Sometimes it feels like it was only yesterday. It truly has helped to create some amazing stuff while it steadily takes over from OpenGL as the cross-platform open API to use.
-

- Login or register to post comments
Printer-friendly version- 1686 reads
PDF version
More in Tux Machines
- Highlights
- Front Page
- Latest Headlines
- Archive
- Recent comments
- All-Time Popular Stories
- Hot Topics
- New Members
digiKam 7.7.0 is released
After three months of active maintenance and another bug triage, the digiKam team is proud to present version 7.7.0 of its open source digital photo manager. See below the list of most important features coming with this release.
|
Dilution and Misuse of the "Linux" Brand
|
Samsung, Red Hat to Work on Linux Drivers for Future Tech
The metaverse is expected to uproot system design as we know it, and Samsung is one of many hardware vendors re-imagining data center infrastructure in preparation for a parallel 3D world.
Samsung is working on new memory technologies that provide faster bandwidth inside hardware for data to travel between CPUs, storage and other computing resources. The company also announced it was partnering with Red Hat to ensure these technologies have Linux compatibility.
|
today's howtos
|








.svg_.png)
Content (where original) is available under CC-BY-SA, copyrighted by original author/s.

Recent comments
1 year 11 weeks ago
1 year 11 weeks ago
1 year 11 weeks ago
1 year 11 weeks ago
1 year 11 weeks ago
1 year 11 weeks ago
1 year 11 weeks ago
1 year 11 weeks ago
1 year 11 weeks ago
1 year 11 weeks ago