On Tue, Jan 30, 2018 at 05:39:58PM +0200, Andy Shevchenko wrote: > On Tue, Jan 30, 2018 at 5:35 PM, <Mario.Limonciello@xxxxxxxx> wrote: > > >> > dell_set_arguments(0x2, 0, 0, 0); > >> > ret = dell_send_request(CLASS_INFO, SELECT_RFKILL); > >> > >> Hi! I'm looking at this code, and do we need shared global buffer with > >> mutex protection at all? Is not buffer allocated on stack enough? > > > > Oh you mean rather than create buffer mutex to just remove global > > buffer and allocate in each function? That seems like a workable > > approach to me too. > > > > I'm fine with either way. > > > > Andy or Darren, what's your preference in this area? > > It reminds me USB stuff where buffer for transfer is allocated on heap > before performing communication. > So, it looks similar to some extent and I have no objection on that > kind of approach. Late to the party it seems, but FWIW: I don't see a significant advantage of a global buffer. It doesn't *need* to be global, and the locking just adds complexity. The heap solution seems much preferable to me. -- Darren Hart VMware Open Source Technology Center
