Content-Type: multipart/mixed; boundary="===============8668022520072172688==" MIME-Version: 1.0 From: Huang Ying To: lkp@lists.01.org Subject: Re: [mm] c8c06efa8b5: -7.6% unixbench.score Date: Thu, 08 Jan 2015 16:24:04 +0800 Message-ID: <1420705444.6201.103.camel@linux.intel.com> In-Reply-To: <1420703154.12346.10.camel@stgolabs.net> List-Id: --===============8668022520072172688== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On Wed, 2015-01-07 at 23:45 -0800, Davidlohr Bueso wrote: > On Thu, 2015-01-08 at 10:27 +0800, Huang Ying wrote: > Cc'ing Peter. > = > > FYI, we noticed the below changes on > > = > > commit c8c06efa8b552608493b7066c234cfa82c47fcea ("mm: convert i_mmap_mu= tex to rwsem") > = > Same exact everything, except for the lock type. No sharing going on. > = > > testbox/testcase/testparams: lituya/unixbench/performance-execl > > = > > 83cde9e8ba95d180 c8c06efa8b552608493b7066c2 = > > ---------------- -------------------------- = > > %stddev %change %stddev > > \ | \ = > > 721721 =C2=B1 1% +303.6% 2913110 =C2=B1 3% unixbench.time.= voluntary_context_switches > > 11767 =C2=B1 0% -7.6% 10867 =C2=B1 1% unixbench.score > = > And this workload appears to be from execl, right? Make sense with some > of those numbers!! Yes. The test we run for unixbench is execl. > > 2.323e+08 =C2=B1 0% -7.2% 2.157e+08 =C2=B1 1% unixbench.time.= minor_page_faults > > 207 =C2=B1 0% -7.0% 192 =C2=B1 1% unixbench.time.= user_time > > 4923450 =C2=B1 0% -5.7% 4641672 =C2=B1 0% unixbench.time.= involuntary_context_switches > > 584 =C2=B1 0% -5.2% 554 =C2=B1 0% unixbench.time.= percent_of_cpu_this_job_got > > 948 =C2=B1 0% -4.9% 902 =C2=B1 0% unixbench.time.= system_time > > 0 =C2=B1 0% +Inf% 672942 =C2=B1 2% latency_stats.h= its.call_rwsem_down_write_failed.vma_adjust.__split_vma.split_vma.mprotect_= fixup.SyS_mprotect.system_call_fastpath > = > What does this "hits" thing mean exactly? Since I assume both before and > after runs have the same level of concurrency when pounding on mmap > operations, I doubt it means that its the amount of calls into the > slowpath... in addition the lock is obviously contended so we can forget > about anything in the fastpath. I think because you changed mutex to rwsem so there was no rwsem related statistics data for the parent commit. Best Regards, Huang, Ying > So this is a call_rwsem_down_write_failed() vs __mutex_lock_common() > issue. > = > = > Thanks, > Davidlohr > = > _______________________________________________ > LKP mailing list > LKP(a)lists.01.org > https://lists.01.org/mailman/listinfo/lkp --===============8668022520072172688==--