Messages in this thread | | | Date | Tue, 18 Feb 2014 14:29:13 -0500 | From | Waiman Long <> | Subject | Re: [PATCH v4 1/3] qspinlock: Introducing a 4-byte queue spinlock implementation |
| |
On 02/18/2014 02:30 AM, Peter Zijlstra wrote: > On Mon, Feb 17, 2014 at 03:41:22PM -0500, Waiman Long wrote: >> +/* >> + * The queue node structure >> + * >> + * This structure is essentially the same as the mcs_spinlock structure >> + * in mcs_spinlock.h file. This structure is retained for future extension >> + * where new fields may be added. >> + */ >> +struct qnode { >> + u32 wait; /* Waiting flag */ >> + struct qnode *next; /* Next queue node addr */ >> +}; >> + >> +struct qnode_set { >> + int node_idx; /* Current node to use */ >> + struct qnode nodes[MAX_QNODES]; >> +}; >> + >> +/* >> + * Per-CPU queue node structures >> + */ >> +static DEFINE_PER_CPU(struct qnode_set, qnset) ____cacheline_aligned >> + = { 0 }; > You really didn't pay attention did you. > > That should be DEFINE_PER_CPU_ALIGNED() > > Furthermore, your structure is bigger than 1 cacheline; your struct > qnode is 16 bytes, 4*16=64, and then you add that int.
Thank for letting me know about that. This is a minor problem that I will fix in the next version.
-Longman
| |