On Mon, Mar 31, 2014 at 9:14 AM, Dmitry Bazhenov <dim...@pigeonpoint.com> wrote: > Hello, Zdenek, > > I think there should be no such checks inside these callbacks. > However, I guess there should be a check inside thr > ipmi_intf_set_max_request/response_data_size > functions which guarantee that the minimum value will be not less than 25 > bytes (required by IPMI spec). > > Could you please add such check or is it better for me to provide a new > patch revision? > > Regards, > Dmitry >
Dmitry, I don't have access to any IPMI capable hardware, so I'm afraid it's either up to you or somebody else. I'm sorry. Best regards, Z. > 31.03.2014 13:07, Zdenek Styblik пишет: > >> On Thu, Mar 20, 2014 at 6:54 AM, Zdenek Styblik >> <zdenek.styb...@gmail.com> wrote: >>> >>> On Wed, Mar 19, 2014 at 8:33 AM, Dmitry Bazhenov <dim...@pigeonpoint.com> >>> wrote: >>> [...] >>>>> >>>>> I got a bit "scared" by solution applied to >>>>> ipmi_intf_get_max_request_data_size() and >>>>> ipmi_intf_get_max_response_data_size(). But then I've tried to compile >>>>> just this one function with all kinds of switches and compiler didn't >>>>> comply, so I guess it's ok. >>>>> I wonder, shouldn't be the same logic applied to >>>>> ipmi_lanp_set_max_rq_data_size() and ipmi_lanp_set_max_rp_data_size() >>>>> as well? >>>> >>>> [DB] Calculations in the ipmi_intf_get_max_request_data_size() are >>>> required >>>> for the case if the target IPMC device is accessed via IPMI bridging. >>>> Since >>>> we can not deduce the target channel maximum message size, we use the >>>> minimum required size. These calculations are not needed for direct IPMC >>>> device access. >>>> [DB] Set max size functions are required if maximum message size over >>>> the >>>> chosen interface must be somehow modified from the value received from >>>> the >>>> interface properties. This is the case for the encrypted RMCP+ payload >>>> where >>>> maximum message size must be reduced by the confidentiality >>>> header/trailer >>>> sizes. Other interface types do not even implement these callbacks. >>>> >>> What I meant is whether under/over-flow shouldn't be checked in those >>> functions as well. >>> >> Ping? >> >> Z. > > ------------------------------------------------------------------------------ _______________________________________________ Ipmitool-devel mailing list Ipmitool-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ipmitool-devel
