Messages in this thread | | | Date | Sun, 15 Mar 2015 20:25:50 -0700 | From | Josh Triplett <> | Subject | Re: [RFC PATCH] sys_membarrier(): system/process-wide memory barrier (x86) (v12) |
| |
On Sun, Mar 15, 2015 at 03:24:19PM -0400, Mathieu Desnoyers wrote: > Here is an implementation of a new system call, sys_membarrier(), which > executes a memory barrier on either all running threads of the current > process (MEMBARRIER_PRIVATE_FLAG) or calls synchronize_sched() to issue > a memory barrier on all threads running on the system. It can be used to > distribute the cost of user-space memory barriers asymmetrically by > transforming pairs of memory barriers into pairs consisting of > sys_membarrier() and a compiler barrier. For synchronization primitives > that distinguish between read-side and write-side (e.g. userspace RCU, > rwlocks), the read-side can be accelerated significantly by moving the > bulk of the memory barrier overhead to the write-side.
From a quick review, this seems quite reasonable (as it did 5 years ago).
One request: Could you please add a config option (default y) in the EXPERT menu to disable this? You actually seem to already have it marked as a cond_syscall.
Also, a very minor nit: flags in kernel APIs aren't typically named with a _FLAG suffix.
With the syscall made optional, and with or without that naming nit fixed: Reviewed-by: Josh Triplett <josh@joshtriplett.org>
- Josh Triplett
| |