On Tue, 23 Jan 2018 10:55:33 +0000, Stachura, Mariusz wrote: > Hello Kuba,
Hi Mariusz! > First of all sorry for a long response. > > Did you have a chance to try it using instruction I've sent? Unfortunately, I can't take the machine with Fortville down ATM to swap kernels. Is this patch already applied to your out-of-tree packaged driver? Could I try that? > As for review: > 1) nit: reverse christmas tree > I'm fixing it. > 2) Are abilities guaranteed to be filled in by this function? Otherwise you > will use uninitialized stack memory. > Good point, I'm changing dev_dbg to dev_err and returning immediately > after printing it > 3) nit: looks like a GENMASK_ULL() coded as a loop > You are right, but I just realized this is not what I wanted :). I wanted > to include all enum values, however not all values between enums. In example, > there is a gap between: > I40E_PHY_TYPE_UNSUPPORTED = 0xF, > I40E_PHY_TYPE_100BASE_TX = 0x11, > And I did not want to include 0x10. I am adding #define that inludes all > ORed enums (in i40e_adminq_cmd.h, just after enum i40e_aq_phy_type > definition). I will assign this define to the mask, instead of using > GENMASK_ULL. What about the error handling? That was my main worry. That the communication with FW will fail on open and the link will remain disabled.
