[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <9102a18d-399c-c00d-501a-448cbabd2d10@intel.com>
Date: Wed, 1 Nov 2017 00:40:16 -0700
From: "Nambiar, Amritha" <amritha.nambiar@...el.com>
To: Stephen Hemminger <stephen@...workplumber.org>
Cc: netdev@...r.kernel.org, alexander.h.duyck@...el.com,
jiri@...nulli.us, jhs@...atatu.com, alexander.duyck@...il.com,
xiyou.wangcong@...il.com
Subject: Re: [iproute2 PATCH] tc/mqprio: Offload mode and shaper options in
mqprio
On 10/31/2017 10:20 AM, Stephen Hemminger wrote:
> On Thu, 26 Oct 2017 17:02:42 -0700
> Amritha Nambiar <amritha.nambiar@...el.com> wrote:
>
>> This patch was previously submitted as RFC. Submitting this as
>> non-RFC now that the tc/mqprio changes are accepted in net-next.
>>
>> Adds new mqprio options for 'mode' and 'shaper'. The mode
>> option can take values for offload modes such as 'dcb' (default),
>> 'channel' with the 'hw' option set to 1. The new 'channel' mode
>> supports offloading TCs and other queue configurations. The
>> 'shaper' option is to support HW shapers ('dcb' default) and
>> takes the value 'bw_rlimit' for bandwidth rate limiting. The
>> parameters to the bw_rlimit shaper are minimum and maximum
>> bandwidth rates. New HW shapers in future can be supported
>> through the shaper attribute.
>>
>> # tc qdisc add dev eth0 root mqprio num_tc 2 map 0 0 0 0 1 1 1 1\
>> queues 4@0 4@4 hw 1 mode channel shaper bw_rlimit\
>> min_rate 1Gbit 2Gbit max_rate 4Gbit 5Gbit
>>
>> # tc qdisc show dev eth0
>>
>> qdisc mqprio 804a: root tc 2 map 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0
>> queues:(0:3) (4:7)
>> mode:channel
>> shaper:bw_rlimit min_rate:1Gbit 2Gbit max_rate:4Gbit 5Gbit
>>
>> Signed-off-by: Amritha Nambiar <amritha.nambiar@...el.com>
>
>
> Please build test your patch with compiler checks enabled....
>
> CC q_mqprio.o
> In file included from /usr/include/string.h:635:0,
> from q_mqprio.c:20:
> In function ‘memcpy’,
> inlined from ‘mqprio_print_opt’ at q_mqprio.c:237:2:
> /usr/include/x86_64-linux-gnu/bits/string3.h:53:10: warning: call to __builtin___memcpy_chk will always overflow destination buffer
> return __builtin___memcpy_chk (__dest, __src, __len, __bos0 (__dest));
> ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
I will send v2 fixing this.
Powered by blists - more mailing lists