***SPAM*** Re: [ofa-general] Re: [PATCH 1/3 v2] opensm: Added io_guid_file and max_reverse_hops options

[Nicolas Morey-Chaisemartin]

Sasha Khapyorsky a écrit :
> On 10:24 Fri 13 Feb     , Nicolas Morey Chaisemartin wrote:
>   
>> Signed-off-by: Nicolas Morey-Chaisemartin 
>> <nicolas.morey-chaisemartin at ext.bull.net>
>>     
>
> Applied (I will push it to main stream tomorrow). Thanks.
>
> All your patches are whitespace mangled - non-diff lines are started
> from two spaces. I fixed it with "sed -e 's/^  / /'", but please check
> your mailer.
>
> Also small note below.
>   
Thanks for applying and sorry for the indentation.
I tried to put my patches inline as Yevgeni advised me but it seems
thunderbird messes things up though I directly output git format-patch
into a thunderbird draft file.
I guess I'll stick to attachment from now on...
>   
>> ---
>> Reposted as io_guid_file and max_reverse_hops were missing from the opt_tbl 
>> and wouldn't be read from the cached option file.
>>
>>  opensm/include/opensm/osm_subnet.h |    6 ++++++
>>  opensm/opensm/main.c               |   26 +++++++++++++++++++++++++-
>>  opensm/opensm/osm_subnet.c         |   14 ++++++++++++++
>>  3 files changed, 45 insertions(+), 1 deletions(-)
>>
>> diff --git a/opensm/include/opensm/osm_subnet.h 
>> b/opensm/include/opensm/osm_subnet.h
>> index 8863e47..671b51f 100644
>> --- a/opensm/include/opensm/osm_subnet.h
>> +++ b/opensm/include/opensm/osm_subnet.h
>> @@ -190,6 +190,8 @@ typedef struct osm_subn_opt {
>>  	char *lfts_file;
>>  	char *root_guid_file;
>>  	char *cn_guid_file;
>> +	char *io_guid_file;
>> +       uint16_t max_reverse_hops;
>>     
>
> Why should max_reverse_hops be 16 bits long? In IB max hops value is 64.
>   

In OpenSM Fat-tree max height is 8.
So except on really irregular topology, max_reverse_hops shouldn't be
more than one byte. For security reasons I chose 2 bytes so it should
never overflow. Anyway more than 2^16 reverse hops is a really bad idea
I guess.


Nicolas