Messages in this thread | | | Date | Thu, 12 Sep 2013 20:44:13 -0700 | From | Greg KH <> | Subject | Re: increased vmap_area_lock contentions on "n_tty: Move buffers into n_tty_data" |
| |
On Fri, Sep 13, 2013 at 11:38:04AM +0800, Fengguang Wu wrote: > On Thu, Sep 12, 2013 at 08:17:00PM -0700, Greg KH wrote: > > On Fri, Sep 13, 2013 at 08:51:33AM +0800, Fengguang Wu wrote: > > > Hi Peter, > > > > > > FYI, we noticed much increased vmap_area_lock contentions since this > > > commit: > > > > What does that mean? What is happening, are we allocating/removing more > > memory now? > > // leave this question to Peter and Tejun. :) > > > What type of load were you running that showed this problem? > > The increased contentions and lock hold/wait time showed up in a > number of test cases. > > The second email has the details, and this section of data is most > illustrating. > > 8cb06c983822103da1cf 20bafb3d23d108bc0a89 > ------------------------ ------------------------ > 3470.31 +1631.0% 60070.49 fat/micro/dd-write/1HDD-cfq-ext4-10dd > 3339.57 +1676.9% 59340.71 fat/micro/dd-write/1HDD-cfq-ext4-1dd > 2848.79 +1489.1% 45269.54 lkp-a04/micro/netperf/120s-200%-TCP_CRR > 3563.01 +1515.2% 57548.50 lkp-a04/micro/netperf/120s-200%-TCP_MAERTS > 2678.64 +1346.0% 38733.80 lkp-a04/micro/netperf/120s-200%-TCP_RR > 2839.84 +1450.2% 44022.08 lkp-a04/micro/netperf/120s-200%-TCP_SENDFILE > 3417.37 +1571.4% 57116.34 lkp-a04/micro/netperf/120s-200%-TCP_STREAM > 2558.59 +1450.8% 39677.58 lkp-a04/micro/netperf/120s-200%-UDP_RR > 3737.24 +1558.0% 61963.62 lkp-a04/micro/netperf/120s-200%-UDP_STREAM > 20219.50 +1488.7% 321218.02 lkp-a06/crypto/tcrypt/2s-200-204 > 21017.17 +1457.1% 327257.41 lkp-a06/crypto/tcrypt/2s-205-210 > 22109.84 +1240.3% 296346.33 lkp-a06/crypto/tcrypt/2s-401-417 > 17909.60 +1467.3% 280693.71 lkp-a06/micro/dbench/100% > 489739.50 +978.5% 5281916.05 lkp-ne04/micro/aim7/shell_rtns_1 > 1601675.63 +906.7% 16123642.52 lkp-snb01/micro/aim7/exec_test > 12105.00 +2453.6% 309110.42 nhm-white/micro/aim7/dbase > 822461.02 +1585.0% 13858430.62 nhm-white/micro/aim7/exec_test > 9858.11 +2715.9% 277595.41 nhm-white/micro/aim7/fork_test > 3452.91 +1685.5% 61650.74 nhm-white/micro/aim7/fserver > 300.14 +2621.5% 8168.53 nhm-white/micro/aim7/misc_rtns_1 > 345479.21 +1624.5% 5957828.25 nhm-white/micro/aim7/shell_rtns_1 > 2694.48 +1974.4% 55894.19 nhm-white/sysbench/oltp/100%-600-1000000 > 4415.67 +1202.2% 57501.52 nhm8/micro/dbench/100% > 2284.65 +1505.2% 36672.75 snb-drag/crypto/tcrypt/2s-200-204 > 2446.02 +1537.1% 40042.87 snb-drag/crypto/tcrypt/2s-205-210 > 2484.11 +1599.6% 42219.71 snb-drag/crypto/tcrypt/2s-500-504 > 2118.55 +1155.8% 26604.99 vpx/crypto/tcrypt/2s-200-204 > 2713.48 +1198.5% 35234.77 vpx/crypto/tcrypt/2s-205-210 > 2711.31 +973.8% 29114.07 vpx/crypto/tcrypt/2s-301-319 > 2369.23 +940.3% 24648.12 vpx/crypto/tcrypt/2s-401-417 > 2620.64 +1428.7% 40060.71 vpx/crypto/tcrypt/2s-500-504 > 1713.98 +1624.3% 29553.72 vpx/crypto/tcrypt/2s-505-509 > 3423353.12 +1184.9% 43985148.08 TOTAL lock_stat.vmap_area_lock.holdtime-total > > The format of report is > > parent commit commit > ------------------------ ------------------------ > number %change number testbox/testcase/test-params > ... > ... > ... > total number %change total number TOTAL perf-metric-name being compared
That's a lot of slowdowns, especially for such a simple patch.
Peter, any ideas?
| |