64-bit ARM Changes For Linux 4.19 Has "A Bunch Of Good Stuff"
Will Deacon submitted the 64-bit ARM (ARM64/AArch64) changes on Tuesday for the Linux 4.19 kernel merge window.
The 64-bit ARM space on Linux remains as busy as ever. For this next kernel cycle, Deacon characterized the changes as "a bunch of good stuff in here." That good 64-bit ARM stuff for Linux 4.19 includes:
- ARM64 support for the new GCC STACKLEAK plug-in that was merged into gcc-plugins for Linux 4.19. The STACKLEAK compiler plug-in is able to fend off possible flaws/attacks pertaining to uninitialized stack usage, stack content leaking, and stack exhaustion/guard-page skipping. This mainline kernel STACKLEAK was ported from old GrSecurity/PaX code.
- Support for the Restartable Sequences system call. This new system call was originally added in Linux 4.18 and "RSEQ" allows for faster user-space operations on per-CPU data by providing a shared data structure ABI between each user-space thread and the kernel.
- A rewrite of their syscall entry code in C in order to zero out the GPR registers on entry from user-space.
- Kexec and Kdump now work on systems started without ACPI support.
- Qspinlock to replace their old ticket lock code.
- Support for chained PMU counters.
- Re-enabled support for huge vmalloc/IO mappings.
The complete list of patches can be found via the kernel mailing list.
The 64-bit ARM space on Linux remains as busy as ever. For this next kernel cycle, Deacon characterized the changes as "a bunch of good stuff in here." That good 64-bit ARM stuff for Linux 4.19 includes:
- ARM64 support for the new GCC STACKLEAK plug-in that was merged into gcc-plugins for Linux 4.19. The STACKLEAK compiler plug-in is able to fend off possible flaws/attacks pertaining to uninitialized stack usage, stack content leaking, and stack exhaustion/guard-page skipping. This mainline kernel STACKLEAK was ported from old GrSecurity/PaX code.
- Support for the Restartable Sequences system call. This new system call was originally added in Linux 4.18 and "RSEQ" allows for faster user-space operations on per-CPU data by providing a shared data structure ABI between each user-space thread and the kernel.
- A rewrite of their syscall entry code in C in order to zero out the GPR registers on entry from user-space.
- Kexec and Kdump now work on systems started without ACPI support.
- Qspinlock to replace their old ticket lock code.
- Support for chained PMU counters.
- Re-enabled support for huge vmalloc/IO mappings.
The complete list of patches can be found via the kernel mailing list.
Add A Comment