Thanks for the review. Find my answers below.
/Jan
> -Original Message-
> From: O Mahony, Billy [mailto:billy.o.mah...@intel.com]
> Sent: Friday, 19 January, 2018 12:31
> To: Jan Scheurich ; d...@openvswitch.org
> Cc: ktray...@redhat.com; Stokes, Ian ;
> i.maxim...@samsung.com
> Subject:
> >>> I am afraid this little "ugliness" is unavoidable.It is the explicit
> >>> purpose of the
> >>> PMD performance metrics and supervision feature to monitor only the
> >>> vhostuser queue length:
> >>> 1. Physical port rx queues are much longer (2K default vs 128 with stock
> >>> Qemu and virt
> >> We are iterating over a sorted list of pmd pointers from cmap dp-
> >>> poll_threads under the protection of the global dp_netdev_mutex lock.
> >> Each pmd struct in this cmap (except for the NON_PMD_CORE_ID) is running
> >> a pmd thread because the thread is created immediately after insertin
On 19.01.2018 14:04, Ilya Maximets wrote:
> On 18.01.2018 20:21, O Mahony, Billy wrote:
>> Hi All,
>>
>> The thread gets updated before I can finish my review so I'm just going to
>> jump in to this thread rather than reply to the o/p with a fresh thread. I
>> hope that's ok.
>>
>> I've run the p
A few things I didn't come across until reading 3/3 but are not related to the
other atomic/volatile/mutex discussion..
All these are suggestions so use only if felt they offer improvements in
clarity.
> -Original Message-
> From: Jan Scheurich [mailto:jan.scheur...@ericsson.com]
> Sent
On 18.01.2018 20:21, O Mahony, Billy wrote:
> Hi All,
>
> The thread gets updated before I can finish my review so I'm just going to
> jump in to this thread rather than reply to the o/p with a fresh thread. I
> hope that's ok.
>
> I've run the patch and any issues with previous revisions are f
Hi All,
The thread gets updated before I can finish my review so I'm just going to jump
in to this thread rather than reply to the o/p with a fresh thread. I hope
that's ok.
I've run the patch and any issues with previous revisions are fixed.
Perhaps add an error message for stats-clear when t
Hi Ilya,
Thanks for your in-depth review and comments. Please find my answers below.
BR, Jan
> -Original Message-
> From: Ilya Maximets [mailto:i.maxim...@samsung.com]
> Sent: Wednesday, 17 January, 2018 13:13
> To: Jan Scheurich ; d...@openvswitch.org
> Cc: ktray...@redhat.com; ian.sto.
Not a full review.
Comments inline.
On 16.01.2018 04:51, Jan Scheurich wrote:
> This patch instruments the dpif-netdev datapath to record detailed
> statistics of what is happening in every iteration of a PMD thread.
>
> The collection of detailed statistics can be controlled by a new
> Open_vSwi
This patch instruments the dpif-netdev datapath to record detailed
statistics of what is happening in every iteration of a PMD thread.
The collection of detailed statistics can be controlled by a new
Open_vSwitch configuration parameter "other_config:pmd-perf-metrics".
By default it is disabled. T
10 matches
Mail list logo