Bug#832931: (no subject)

2017-01-27 Thread Otto Kekäläinen 
2017-01-27 9:18 GMT+02:00 Otto Kekäläinen :
> Please also note the related
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=852728 and PR at
> https://github.com/ottok/mariadb-10.1/pull/8

Actually Bug#852728 has nothing to do with this, my mistake. But the
PR is relevant.

(I seem to be making stupid mistakes way too much now, maybe I should
take a break from this topic and continue with fresh eyes later.)

Bug#832931: (no subject)

2017-01-26 Thread Otto Kekäläinen 
Please also note the related
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=852728 and PR at
https://github.com/ottok/mariadb-10.1/pull/8

Bug#832931: (no subject)

2017-01-26 Thread Daniel Black 
As Thadeu Lima de Souza Cascardo said, the build time detected page size
isn't correct, perhaps using --with-lg-page=/--with-lg-page-sizes at
compile time.

This is available in the 3.6.0 version of jemalloc. (upstream currently
at 4.4.0). I've got part way though an upstream patch to change this to
runtime detection, no idea when it will be complete, and I have been
told that upstream won't risk performance regressions.