On Thu, Apr 1, 2021 at 9:09 PM Xunlei Pang <xlpang@xxxxxxxxxxxxxxxxx> wrote: > > On 3/26/21 5:44 PM, Xunlei Pang wrote: > > Add the following knobs in PATCH 1~3: > > /sys/kernel/mm/page_reporting/reported_kbytes > > /sys/kernel/mm/page_reporting/refault_kbytes > > /sys/kernel/mm/page_reporting/reporting_factor > > > > Fix unexpected user OOM in PATCH 4. > > > > Xunlei Pang (4): > > mm/page_reporting: Introduce free page reported counters > > mm/page_reporting: Introduce free page reporting factor > > mm/page_reporting: Introduce "page_reporting_factor=" boot parameter > > mm/page_reporting: Fix possible user allocation failure > > > > Documentation/admin-guide/kernel-parameters.txt | 3 + > > include/linux/mmzone.h | 3 + > > mm/page_alloc.c | 6 +- > > mm/page_reporting.c | 268 ++++++++++++++++++++++-- > > 4 files changed, 260 insertions(+), 20 deletions(-) > > > > Hi guys, > > Looks "Alexander Duyck <alexander.h.duyck@xxxxxxxxxxxxxxx>" was not > available, so Cced more, any comment? > > Thanks! Yes, my Intel account has been offline since October. If you need to reach me, my gmail is the best way to go. As far as the patch series itself I am not exactly thrilled with it. There seems to be a number of spots where things are being changed such that the CPU overhead will be much more significant. The cover page should actually say what the patch set is attempting to accomplish. In the patch descriptions you have told us what you are doing, but the why isn't completely clear. For example I am not sure if the issue addressed in patch 4 was present before patches 1-3 were introduced.
