On 27.09.2021 10:20, Juergen Gross wrote: > On 27.09.21 10:13, Jan Beulich wrote: >> On 24.09.2021 17:04, Juergen Gross wrote: >>> The interface definition of PV USB devices is lacking the specification >>> of possible values of the status filed in a response. Those are >>> negative errno values as used in Linux, so they might differ in other >>> OS's. Specify them via appropriate defines. >> >> What if new errno values got used by the driver? Would we alter the >> public header every time? Or is the likelihood of further values ever >> getting used vanishingly small? In how far would it be possible to tie >> these to Xen's public/errno.h instead? > > Those are the current values returned by the backend. Other ones used > internally in the backend should IMO tried to be mapped to the values > defined in the interface specification.
In which case I'd like to reword my question: Is the set of values added sufficiently rich? There's e.g. no ENOMEM equivalent. One other observation: Why is this header using tab indentation? This is by far the largest (albeit sadly not the only) of the style violators in xen/include/public/io/. Jan
