[openib-general] sysfs exposure of port counters useless?

[Michael Newton]

On Tue May 9 05:06:13 PDT 2006, Leonid Arsh leonida at voltaire.com
posted a patch under the Subject
[openib-general][RFC][PATCH] core/sysfs.c: ability to reset port counters
in which /sys/class/infiniband/*/ports/*/counters/* were made writeable,
so that they could be reset by writing zero

Michael S. Tsirkin mst at mellanox.co.il
Tue May 9 06:17:13 PDT 2006
replied with some constructive amendments..

However this is where the thread appears to have finished:

Roland Dreier rdreier at cisco.com
Wed May 10 08:55:22 PDT 2006
>    Leonid> A user space application is an option too, although I
>    Leonid> think it's nice to have a 'built in' kernel feature.
>
>As Hal pointed out, there already is an app to do this.  So I don't
>see much need to put it into the kernel.

..ie the ability to reset using perfquery was felt to be sufficient,
and the patch wasnt adopted.

I wonder though who would ever use the sysfs exposure of port counters, as
they stand?

These are 32 bit counters. The rcv/xmit_data counters count 32-bit
blocks. Also, these counts do not wrap: they peg at all 1s.
At infiniband speeds, these counts can peg out very quickly indeed,
to the point they can really only be of use if they can be reset each time
there read. Now if anyone who wants to use them has to go the CLI to reset
them, and theres little point in reading them without reset, why would
anyone read them via sysfs? so why have them?

Dr.Michael("Kimba")Newton  kimbrr at sgi.com