Hi Vabhav, > -----Original Message----- > From: Vabhav Sharma <vabhav.sharma@xxxxxxx> > Sent: Monday, April 18, 2022 7:11 PM > To: Fabio Estevam <festevam@xxxxxxxxx>; Varun Sethi <V.Sethi@xxxxxxx> > Cc: herbert@xxxxxxxxxxxxxxxxxxx; Horia Geanta <horia.geanta@xxxxxxx>; > Gaurav Jain <gaurav.jain@xxxxxxx>; linux-crypto@xxxxxxxxxxxxxxx; Fabio > Estevam <festevam@xxxxxxx> > Subject: RE: [EXT] [PATCH v2] crypto: caam - fix i.MX6SX entropy delay value > > > > > -----Original Message----- > > From: Fabio Estevam <festevam@xxxxxxxxx> > > Sent: Sunday, April 17, 2022 12:41 AM > > To: Varun Sethi <V.Sethi@xxxxxxx> > > Cc: herbert@xxxxxxxxxxxxxxxxxxx; Vabhav Sharma > > <vabhav.sharma@xxxxxxx>; Horia Geanta <horia.geanta@xxxxxxx>; Gaurav > > Jain <gaurav.jain@xxxxxxx>; linux-crypto@xxxxxxxxxxxxxxx; Fabio > > Estevam <festevam@xxxxxxx> > > Subject: Re: [EXT] [PATCH v2] crypto: caam - fix i.MX6SX entropy delay > > value > > > > Caution: EXT Email > > > > Hi Varun, > > > > On Sat, Apr 16, 2022 at 3:00 PM Varun Sethi <V.Sethi@xxxxxxx> wrote: > > > > > > Hi Fabio, > > > Vabhav is working on a fix for the Linux driver. He would be > > > introducing a > > new property in the CAAM device tree node, which would be used for > > specifying the entropy delay value. This would make the solution generic. > > This property is optional. > > > > Unfortunately, a devicetree property solution via optional property > > would not work. > > > > Such a solution would not be backported to stable kernels and people > > running old devicetree with new kernels would still face the problem. > Please elaborate and specify the version for kernel, devicetree > > > > This problem is seen since kernel 5.10, so we need a kernel-only fix. > Kernel 5.10 support devicetree, Do you mean customer using kernel without > device tree? Fabio's concern is about using new kernel image with an old device tree. Fabio, We feel that it would be better to have a provision to provide entropy delay parameter via device tree. This offers more flexibility. For i.M6SX we can follow the approach proposed by you but at the same time we can still have the device tree provision. Regards Varun
