Matthew Wilcox wrote on Wed, Jul 11, 2018: > On Wed, Jul 11, 2018 at 02:40:38PM +0200, Dominique Martinet wrote: > > > + idr_preload(GFP_KERNEL); > > > + spin_lock_irq(&clnt->lock); > > > + ret = idr_alloc_u32(&clnt->fids, fid, &fid->fid, UINT_MAX, GFP_NOWAIT); > > > > There's also a P9_NOFID that we shouldn't use, so the max here should be > > P9_NOFID - 1 > > Happy to fix that. It shouldn't actually happen, of course. I can't > imagine us having 4 billion FIDs in use at once. Oh, I said that assuming that it was random.. But I guess we might as well fix it. > > That aside this introduces a change of behaviour that fid used to be > > alloc'd linearily from 0 which no longer holds true, that breaks one > > serveur (nfs-ganesha just returns ERANGE) but others seem to handle this > > just fine so they'll just need to fix that server. > > max aside this looks good. > > I don't understand your assertion that this is a change of behaviour. > The implementation of p9_idpool_get() uses idr_alloc(), not > idr_alloc_cyclic(), so I don't believe I've changed which FID would > be allocated. Hmm, I just tried mounting something with ganesha and that broke because it received fid=1730858632 in a TATTACH request, so something in the patch series made some big fid happens. If you say this isn't intented though this needs debugging, I'm glad I brought this up :P (Note that it really should be fine if it is random within the pool, I have notified the current developpers of the problem) -- Dominique
