On Thu, Sep 24, 2020 at 04:50:01PM -0300, Jason Gunthorpe wrote: > On Sat, Sep 19, 2020 at 12:09:28PM +0300, Leon Romanovsky wrote: > > On Fri, Sep 18, 2020 at 08:31:52PM -0300, Jason Gunthorpe wrote: > > > On Mon, Sep 07, 2020 at 03:21:53PM +0300, Leon Romanovsky wrote: > > > > From: Leon Romanovsky <leonro@xxxxxxxxxxxx> > > > > > > > > Restrack stores the IB objects in the internal xarray and insertion > > > > where can fail. As long as restrack was helper tool for the > > > > debuggability, such failure were ignored. > > > > > > > > In the following patches, the ib_core will be changed to manage allocated > > > > IB objects in restrack DB for the proper memory lifetime. It requires to > > > > ensure that insertion errors are not ignored. > > > > > > Why? This looks like it is all about removing valid, not sure what the > > > kref has to do with it.. > > > > This DB is going to be main source of all HW objects and their memory > > allocations. We want to be sure that everything there is valid. > > Not really, what has happened here is no_track replaces valid. valid > used to mean the entry was in the xarray, now no_track means the same > thing. The patches in between had both because of how the conversion > ended up > > This commit message should just explain that valid is no longer needed > and no_track now indicates if the entry is in the xarray or not so > destruction knows what to do. er I mean the message of patch #14 This commit message should explain that this adds error handling for the rdma_restrack_add(). As far as I can tell this isn't strictly required as a failing add could just set no_track and things would still be fine. Jason
