Messages in this thread | | | Date | Fri, 28 Sep 2012 10:53:01 -0700 | From | "Paul E. McKenney" <> | Subject | Re: [PATCH 0/7] rcu: optimize RCU trace on the huge platform |
| |
On Thu, Sep 20, 2012 at 08:51:01AM +0800, Michael Wang wrote: > This patch set optimize the RCU trace to make it perform better on the > platform with thousands of cpus. > > The optimization including: > -- the cpu info will be separated by each flavor of rcu. > -- implement the 'cpu units sequence reading' to avoid losing data. > > Michael Wang (7): > [PATCH 1/7] rcu: create directory for each flavor of rcu > [PATCH 2/7] rcu: fundamental facility for 'cpu units sequence reading' > [PATCH 3/7] rcu: optimize the 'rcudata' for RCU trace > [PATCH 4/7] rcu: optimize the 'rcudata.csv' for RCU trace > [PATCH 5/7] rcu: optimize the 'rcu_pending' for RCU trace > [PATCH 6/7] rcu: replace the old interface with the new one > [PATCH 7/7] rcu: remove the interface "rcudata.csv" > > Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com> > --- > b/kernel/rcutree_trace.c | 8 + > kernel/rcutree_trace.c | 368 ++++++++++++++++++++--------------------------- > 2 files changed, 170 insertions(+), 206 deletions(-)
Queued for 3.8, thank you Michael!
Thanx, Paul
| |