On Wed, 2021-04-21 at 14:26 +0200, Hans de Goede wrote: > Hi, > > On 4/21/21 2:17 PM, Vaittinen, Matti wrote: > > On Wed, 2021-04-21 at 13:58 +0200, Hans de Goede wrote: > > > Hi, > > > > > > On 4/21/21 9:51 AM, Matti Vaittinen wrote: > > > > On Tue, 2021-03-23 at 15:19 +0100, Greg KH wrote: > > > > > On Tue, Mar 23, 2021 at 02:58:28PM +0100, Hans de Goede > > > > > wrote: > > > > > > Hi, > > > > > > > > > > > > On 3/23/21 2:56 PM, Matti Vaittinen wrote: > > > > > > > Devm helper header containing small inline helpers was > > > > > > > added. > > > > > > > Hans promised to maintain it. > > > > > > > > > > > > > > Add Hans as maintainer and myself as designated reviewer. > > > > > > > > > > > > > Ultimately this is up to Greg though, so lets wait and see > > > > > > what > > > > > > Greg has to say about this. > > > > > > > > > > Can we move some of the devm_* calls in include/device.h into > > > > > here as > > > > > well so that you all can be in charge of them instead of me? ... > > but what comes to my (not always so humble) opinion - most of > > the devm functions in device.h are tightly related to the device > > interface or devres. Thus the device.h feels like appropriate place > > for > > most of those. > > I agree with you that most devm_ functions in device.h are probably > left there. Moving them will also mean modifying all the drivers > which use them to include the new devm-helpers.h include file > which seems like needless churn. > > > OTOH, the kmalloc/kfree related functions, strdub and > > kmemdub might be candidates for move - those are not really "device > > things". > > I'm certainly open to moving some functions to devm-helpers.h, but > also see above about needless churn. I agree. Whether the move is worth all the hassle depends on how much Greg _really_ wishes to get rid of those devm functions. Especially the allocs are used _a lot_. Best Regards --Matti